THÈSE DOCTORAT DE L'UNIVERSITÉ BORDEAUX 2

1

Année 2013 Thèse n°2112

THÈSE pour le

DOCTORAT DE L’UNIVERSITÉ BORDEAUX 2

Mention Sociétés, Politique, Santé Publique Spécialité : Santé Publique

Option : Epidémiologie

Thèse préparée dans le cadre du Réseau doctoral en santé publique animé par l’Ecole des Hautes Etudes en Santé Publique (EHESP)

Présentée et soutenue publiquement

Le 20 décembre 2013

Par

Sophie DESMONDE

Née le 07 août 1986 à Epsom, Royaume-Uni

La prise en charge du VIH pédiatrique avant et après traitement antirétroviral en Afrique de l’Ouest

Contribution au développement d’une modélisation multi-états

Care of HIV-infected children before and after antiretroviral therapy initiation in West Africa

Contribution towards the development of a multi-state model

Membres du Jury M Rodolphe THIEBAUT, PU-PH, Université Bordeaux Segalen, France Président Mme Laurence MEYER, PU-PH, Université Paris-Sud, France Rapporteur Mme Diana GIBB, Professor, MRC Clinical Trials Unit, Londres, Royaume-Uni Rapporteur Mme Madeleine AMORISSANI-FOLQUET, Professeur, Université Cocody, Côte d’Ivoire Examinatrice M Xavier ANGLARET, DR2 Inserm U897, Université Bordeaux Segalen, France Examinateur Mme Andrea CIARANELLO, Assistant Professor, Harvard Medical School, Boston, USA Examinatrice Mme Valériane LEROY, DR2 Inserm U897, Université Bordeaux Segalen, France Directrice

2

To my father with love,

3

"In the final analysis, the questions of why bad things happen to good people transmutes

itself into some very different questions, no longer asking why something happened, but

asking how we will respond, what we intend to do now that it happened "

Pierre Teilhard de Chardin

4

Acknowlegements

To Professor Rodolphe Thiébaut,

I do not know where to start to thank you for all the support and help you have given me since

I set foot in the ISPED. You have influenced me greatly and inspired me to work in a rigorous

way, always respecting the fact that what mattered to me above all was the subject of my

research. I feel privileged to have benefitted from your teaching and am very honoured you

accepted to preside over this jury, thank you.

To Professor Laurence Meyer,

You accepted to be one of the two rapporteurs of this work, and I thank you for that.Your

experience in the care of HIV in African settings in both adults and children will no doubt

provide me with key elements to improve this work.

To Professor Diana Gibb,

Thank you for reviewing this PhD. I am honoured to have the opportunity to benefit from

feedback of someone as experienced as you in the field of paediatric HIV. I am grateful for

your comments.

To Professor Madeleine Amorissani-Folquet,

Thank you for agreeing to judge this work, I am very honoured. It has been a pleasure

working with you during this PhD. Your advice and input have been valuable and I look

forward to continuing in this path.

To Doctor Xavier Anglaret,

Thank you for taking the time to judge my work. I would also like to thank you for your trust

and support. You allowed me to work so easily within the PAC-CI programme in Abidjan,

and advised me and provided input all along the development of this PhD. I thank you for

everything.

To Doctor Andrea Ciaranello,

Thank you for being a part of this jury. Your expertise, understanding, and patience, have

added considerably to my experience. I am very grateful for your support and the welcome

you gave me when I came to Boston. I look forward to continuing working with you.

To Doctor Valériane Leroy

I could not have asked for a better role model. Your patience, flexibility, genuine caring and

concern, and faith in me over the past four years have enabled me to attend to life while also

conducting my Ph.D. Working with you has been on occasion challenging, but most of all fun

and very rewarding. I look forward to continuing on this path. I hope that one day I can pass

on the research values and the dreams that you have given to me. Thank you.

5

To all the staff members of the PAC-CI programme, the CePReF and the MTCT Plus

programme in Abidjan, Côte d’Ivoire

I could not possibly mention every single one of you individually, but thank you all for your

warm welcome on both of my visits.

This work was made possible because of your devotion. During the post-electoral crisis in

2011, each of you courageously continued to serve the sick, the wounded, the displaced, the

refugees and each other – all victims of the war and to bring them medicine, food, water,

support and humanity.

To all the staff members of the IeDEA West Africa Collaboration

My profound gratitude and respect go towards all the investigators of the paediatric West

Africa Database collaboration.

To all the staff members of the CEPAC group, Boston, USA

I would particularly like to thank Ken Freedberg and Rochelle Walensky in addition to

Andrea Ciaranello for giving me such a warm welcome in Boston. During those three months

I worked in a fantastic environment and was truly amazed by the respect, professionalism and

humanity that reigns in your research institute. I would also like to thank by name Julia

Maxwell for all her help settling in and Katie Doherty for her very valuable help in mastering

the CEPAC model! I could not end this section without thanking all the RAs with whom I had

some really good times.

To the HIV, cancer and global health in resource-limited settings research team,

Thank you all for your help, input, advice, comments. I have learnt from every single one of

you and feel privileged to be working with you. Your support, advice and genuine help and

concern mean a lot to me.

I am particularly grateful to François Dabis: thank you for your help and encouragements

regarding my work, but also for your support and guidance prior to the EHESP audition that

allowed me getting the grant to conduct this research.

To all the ISPED staff,

My deepest appreciation goes towards the ISPED teaching staff, in particular Ahmadou

Alioum and Virginie Rondeau for their availability and methodological support. Professor

Rachid Salmi, I feel privileged to have benefitted from your teaching throughout my master’s

degree. You have taught me to be thorough, consistent and organised. Thank you.

Thank you to the CREDIM and more importantly the hotliners for their numerous expeditions

retrieving lost documents – I swear I will no longer work on attached files nor will I forget

where I have saved them!

I would also like to acknowledge the library staff for their rapidity and availability.

6

To those I have met along the way,

To my friends in Abidjan, Zama, Grace, Valentin and Denise and to your baby girl I look

forward to meeting! You opened your community to me; made sure I was never alone and

invited me to share your lives, respecting my differences. I experienced with you authentic

daily life in Abidjan and appreciated every minute of it. Thank you for everything.

To Gary, Matti, and Marta, my Bostonian flatmates and friends. It seems like yesterday we

saw the 4th of July fireworks in the pouring rain, the mad road trips to Cape Cod or

Connecticut, and the best of all: BC’s opening football game! Thank you for such a fantastic

summer. Matti, your painting is still in my living-room!

To all the fellow doctoral students and post-docs whom I have shared these past three years

with at the ISPED, particularly Riccardo and Boris, I have had tremendous fun with you two.

Even if I am undoubtedly the world’s worst badminton player, that weekly hour of torture was

well worth the beer and burgers that followed! Boris, I’m glad you overcame living “en

Province” and hope you stick around. And Riccardo: thank you for returning to Bordeaux

armed with much missed Twirls and Hobnobs!

And thank you to those with whom I shared practically everything: Julie, Alex, Ana, Camille,

Mélanie, Marion and Mathieu.

Julie, you were one of the first friendly faces to greet me when I arrived in Bordeaux.

You have always been there, advised me in many ways and stopped me making huge

mistakes. You are a true friend, thank you. I will forever remember the wasted hours

browsing PhD Comics or PhDelirium and our coffee breaks putting the world of research to

rights.

Alex, despite the fact that I hold you solely responsible for my addiction to One

Direction, you have managed to become my friend. It is going to be weird not having you

around at the ISPED anymore, but I’m sure you’re going to be a rock star at primary school

teaching.

Ana, we hit it off as soon as we met, and we have shared a lot. You have been there, in

good and bad times, including the flooding of my flat! You have never judged or pushed,

simply listened. Thank you. Hopefully, you’ll stop planning barbecues: they are clearly

correlated to the bad weather!

Marion, the queen of good deals! We have certainly had our differences, but also

shared a lot of mad stuff, including getting lost in the rain and mud in the middle of the

Minho! I am convinced that whatever you engage, you will succeed.

Camille, my colleague (!), neighbour and most importantly my friend. Thank you for

your support (and your Sangria).

7

Mélanie, ma petite boulette! Your vision of the world and of research is refreshing and

your optimism is contagious! Having said that, I will never get on the same plane as you…

And considering recent events, it is not advised to travel by train with you either!

Mathieu, you hooked us all on weird board games and card games. Playing a round of

tarot over lunch has now become a routine. It’s just a shame you are no longer to here to share

these moments with us. I wish you a lot of success for the future. As the dog said, 'If I fall

down for you and you fall down for me, it is playing’.

Last but not least, thank you to all those I have not yet mentioned but mean a lot to me,

To Ghislaine, Camille, Emilie, Alexis and all the rest of my friends with whom I have spent

time, crises and cake whenever it was needed.

To Virginia (aka Raroon) and all those of you who have read and sweat over parts of this

dissertation or even the whole thing for the bravest (including my mother)! I still don’t master

hyphenation but have learnt that data is plural. As promised, a special mention for the mince

pie chase, worthy of the great fruit gum robbery! Thank you all for being a part of this,

including Waitrose’s welcome desk cast starring Ross!

To my grandparents who were convinced that public health was a fancy and more modern

way of saying “flitgun operating” and kept asking me when I’d get a proper job. To my

father, whom I can still hear saying “And there at the very back of the store was….”

All three have died recently. There are few days where I do not think about them.

To Milly who once asked me why we weren’t like “normal” sisters, why what we know about

each other’s whereabouts is reduced to our respective Facebook pages. It’s not about what we

know and don’t know, it’s about being there when needed, and you have never failed. Thank

you.

To my mother, undoubtedly the most amazing and strongest person I have ever known. You

believed in me and supported me in every decision I made. For ever you have been involved

in everything I do and made sure I succeeded - I was probably the only teenager in France

whose mother made playing cards to learn dates for the history baccalaureate! You taught me

how to read, write and spell in the English language (so all mistakes are on you now!). Thank

you for making me bilingual, I hope one day I will be able to offer the same privilege to my

children.

It is not possible to summarize Vincent and his influence on me in just a couple of sentences.

You have completely changed my life and changed me for the best. You are the most

balanced and well-adjusted person I have ever known and your love, intelligence and

goodness make me smile every day. You have supported me in hundreds of ways throughout

the development and writing of this PhD and I am forever grateful.

8

This work was conducted within the doctoral network of the

Ecoles des Hautes Etudes en Santé Publique (EHESP)

From October 2010 to September 2013, I was financed by the French Ministry of Higher Education and Research via the doctoral network of the EHESP.

From October 2013 to December 2013, I was financed by the IeDEA West Africa

collaboration

9

Abstract Access to prevention of mother-to-child transmission (PMTCT) interventions is limited in

West Africa and mothers continue to transmit HIV disease to their children. Children are

diagnosed at an advanced age and stage of the disease, and their antiretroviral treatment is

often initiated too late. Important questions on early HIV diagnosis and early antiretroviral

therapy (ART) for children in resource-limited settings remain unanswered. Traditional

clinical research methods, such as observational studies and randomised trials, provide

information essential to answering these questions; however these studies are often conducted

in specific settings and represent a small amount of selected children. Computer simulation

models can provide helpful information to project long-term patient outcomes and inform

health policy. Although simulation models are computationally sophisticated, the usefulness

of the results of modelling studies depends on the quality and accuracy of the data on which

they are based. The main objective of the following work was to provide accurate and up-to-

date data on mortality, severe morbidity and healthcare resource utilisation in HIV-infected

children enrolled in care, before and after ART initiation in the context of the access to ART

roll-out since 2004 in West Africa.

Our findings suggest mortality rates comparable to those of other studies, reaching 5.5% by

18 months of follow-up in children enrolled in cohorts at a median age of 5 years who had not

yet initiated ART. Severe morbidity rates were high, in both ART-treated and untreated

children. We found that one hospitalisation in three was caused by an infectious disease,

avoidable by cotrimoxazole prophylaxis, a simple and efficient intervention that is still not

accessible to all in West Africa. We also reported substantial rates of healthcare resource

utilisations associated with this severe morbidity. However, in both untreated and ART-

treated children, healthcare resource utilisation was lower in the sickest, most

immunodeficient children, as well as in children who were cared for by their mother alone

compared to those cared for by both parents and father alone. Access to healthcare remains

limited and one of the explanations we put forward are the costs borne by the families.

Finally, children on ART remain initiated at a too later stage to be able to restore a normal

immunity for age; this is even less likely in those who initiated ART after 5 years compared

to younger children.

Overall, this work underlines the need for an effective early HIV diagnosis and treatment.

This necessarily involves the strengthening of the continuum of care from antenatal

consultations in PMTCT services to the long-term follow-up of HIV-infected children, on

ART in paediatric care. In middle HIV prevalence settings, tracing HIV-exposed infants is

essential. Optimising this requires interventions at multiple levels of the healthcare system

and no single approach is likely to be effective. Furthermore, lifetime treatment costs will

need to be assessed as HIV becomes a chronic disease leading to greater healthcare resource

utilisation. Integrating these data in computer simulation models will assist healthcare

providers and policy-makers to identify the most effective and cost-effective strategies for

diagnosis, treatment and monitoring of paediatric HIV in low income countries.

Keywords: HIV; children; mortality; severe morbidity; healthcare resource utilisation;

immune recovery; West Africa

10

Table of contents

Acknowlegements ..................................................................................................................... 4

Abstract ..................................................................................................................................... 9

List of tables ............................................................................................................................ 13

List of figures .......................................................................................................................... 14

List of abbreviations ............................................................................................................... 15

Introduction ............................................................................................................................ 16

1. FACTUAL BACKGROUND ........................................................................................ 18

1.1. Situation analysis: the burden of paediatric HIV disease in West Africa .......... 19

1.1.1. Prevention of mother-to-child transmission (PMTCT) programme

cascade…………………………………………………………………………………..19

1.1.2. The operational failure of prevention of mother-to-child transmission (PMTCT) programmes ...................................................................................................................... 20

1.1.2.1. Efficient PMTCT regimens but low uptake ................................................. 20

1.1.2.2. From virtual elimination of paediatric HIV to the operational reality in

Western Africa .............................................................................................................. 23

1.1.3. The paediatric HIV epidemic ............................................................................. 25

1.1.4. Diagnosing HIV infection in children is a major challenge in Africa. .............. 27

1.1.5. Access to care of HIV-infected children ............................................................ 31

1.1.5.1. Cotrimoxazole prophylaxis ......................................................................... 31

1.1.5.2. Antiretroviral treatment .............................................................................. 32

1.2. Research gaps and many unanswered questions in promoting and assessing

paediatric HIV care in Africa ....................................................................................... 38 1.2.1. Clinical research gaps ......................................................................................... 38 1.2.2. Operational research gaps .................................................................................. 40

1.3. Which study design when addressing the above questions? ................................ 41 1.3.1. Observational cohorts and clinical trials gaps .................................................... 41 1.3.2. Computer simulation models comprise important adjuncts to more traditional research methods .............................................................................................................. 42 1.3.3. An example of a computer simulation model: the Cost Effectiveness of Preventing AIDS Complications (CEPAC)-Paediatric model ......................................... 43

1.3.3.1. The CEPAC model ...................................................................................... 43

1.3.3.2. The CEPAC-Paediatric model .................................................................... 45

2. HYPOTHESES AND OBJECTIVES ........................................................................... 48

2.1. Conceptual framework ............................................................................................ 49

2.2. Hypotheses ................................................................................................................ 51

2.3. My PhD Objectives .................................................................................................. 51

11

3. SETTINGS ...................................................................................................................... 53

3.1. Implicated research infrastructures ...................................................................... 54 3.1.1. The Bordeaux School of Public Health & the Inserm U897 research centre “Epidemiology & Biostatistics” ....................................................................................... 54 3.1.2. The PAC-CI programme .................................................................................... 55 3.1.3. The CEPAC group ............................................................................................. 56

3.2. Data platforms ......................................................................................................... 56 3.2.1. The International epidemiologic Databases to Evaluate AIDS (IeDEA), West Africa 56 3.2.2. Côte d’Ivoire ...................................................................................................... 59

3.2.2.1. Geography of Côte d’Ivoire ....................................................................... 59

3.2.2.2. The HIV pandemic in Côte d’Ivoire ............................................................ 59

3.2.2.3. The National response to HIV in Côte d’Ivoire .......................................... 60

3.2.2.4. Presentation of ACONDA ........................................................................... 61

3.2.2.5. The Centre de Prise en Charge, de Recherche, et de Formation .............. 62

3.2.2.6. The MTCT – Plus programme: Mother to Child Transmission Prevention

programme ................................................................................................................... 62

4. WHAT WAS THE NATURAL EVOLUTION OF PAEDIATRIC HIV BEFORE

ART INITIATION DURING THE ROLL-OUT OF ART IN ABIDJAN, COTE

D’IVOIRE? ............................................................................................................................. 63

4.1. Severe morbidity and mortality in untreated HIV-infected children in a

paediatric care programme in Abidjan, Côte d’Ivoire, 2004-2009 ............................ 64 4.1.1. Research question and background .................................................................... 64 4.1.2. Desmonde et al – BMC Infectious Diseases – November 2011 ........................ 64 4.1.3. Principal results and discussion .......................................................................... 77

4.2. Healthcare resource utilisation in untreated HIV-infected children in a

paediatric care programme, Abidjan, Côte d’Ivoire, 2004-2009 ............................... 78 4.2.1. Research question and background .................................................................... 78 4.2.2. Desmonde et al – JAIDS – January 2013 .......................................................... 78 4.2.3. Principal results and discussion .......................................................................... 87

5. REASONS FOR HOSPITALISATION IN HIV-INFECTED CHILDREN, ART-

TREATED OR NOT, IN WEST AFRICA .......................................................................... 89

5.1. Research question and context ............................................................................... 90

5.2. Dicko et al – Journal of International AIDS Society – in review ........................ 90

5.3. Principal results and discussion ........................................................................... 107

6. WHAT IS THE EVOLUTION OF HIV IN CHILDREN AFTER THE

INITIATION OF ANTIRETROVIRAL THERAPY? ..................................................... 108

6.1. Severe morbidity and healthcare resource utilisation in HIV-infected children

on ART in Abidjan, Côte d’Ivoire .............................................................................. 109 6.1.1. Research question and background .................................................................. 109 6.1.2. Desmonde et al – JAIDS – in press .................................................................. 109 6.1.3. Principal results and discussion ........................................................................ 130

12

6.1.4. Description of direct costs of care in HIV-infected children on ART ............. 130

6.2. Immune response to antiretroviral therapy according to age at treatment

initiation in HIV-infected children in West Africa ................................................... 134 6.2.1. Research question and background .................................................................. 134 6.2.2. Desmonde et al – AIDS – in revision .............................................................. 134 6.2.3. Principal results and discussion ........................................................................ 154

7. THE INDEPENDENT “ART-EFFECT” ................................................................... 156

7.1. Background and objective .................................................................................... 157

7.2. Which data to answer the ART-effect question in the Abidjan context? ......... 157 7.3. Pointers to address the biases ............................................................................... 160

8. DISCUSSION AND PERSPECTIVES ....................................................................... 161

8.1. Work synthesis ....................................................................................................... 162

8.2. Methodological considerations ............................................................................. 162

8.3. Public health implications to improve HIV paediatric care .............................. 166

8.4. Research perspectives ............................................................................................ 167 8.4.1. Addressing costs of care and identifying cost-effective strategies .................. 167 8.4.2. How to promote early infant diagnosis in the West African context? ............. 168 8.4.3. How to improve retention in care programmes ................................................ 171 8.4.4. How to monitor ART response? What are the switching strategies when treating HIV-infected on the long-term? ..................................................................................... 172

Conclusion ............................................................................................................................. 174

References ............................................................................................................................. 176

Appendices ............................................................................................................................ 194

13

List of tables Table 1: Programme options for antiretroviral therapy for Prevention of Mother-To-Child Transmission (WHO, 2012) ..................................................................................................... 21 Table 2: Summary of recommendations of when to start antiretroviral therapy in children ... 35 Table 3: Summary of estimates in preventing mother to child transmission and providing paediatric care and treatment in sub-Saharan Africa for selected regions and countries in 2005, 2009 and 2012. ............................................................................................................... 37 Table 4: Amount of drugs prescribed in the context of outpatient care by type of severe morbid event in HIV-infected children on ART included in the CePReF, Abidjan, Côte d’Ivoire, 2004-2009 ................................................................................................................ 131 Table 5: Amount of times each drug was prescribed in the context of outpatient care per event, by type of severe event in HIV-infected children on ART included in the CePReF, Abidjan, Côte d’Ivoire, 2004-2009 ........................................................................................ 132 Table 6: Mean costs ($) spent by the families of drugs prescribed in the context of outpatient care per event, by type of severe event in HIV-infected children on ART included in the CePReF, Abidjan, Côte d’Ivoire, 2004-2009 ......................................................................... 133

14

List of figures Figure 1: “Cascade” of PMTCT and postnatal HIV care. ........................................................ 19 Figure 2: Percentage of pregnant women who received an HIV test in low- and middle-income countries by region, 2005 and 2008-2009. .................................................................. 22 Figure 3: Antiretroviral coverage to prevent new infections among children, and the final HIV transmission rate from mother to child, including breastfeeding ............................................. 23 Figure 4: New HIV infections through mother-to-child transmission (MTCT) for different scenarios, 25 countries, 2009-15. ............................................................................................. 24 Figure 5: Number of children living with HIV globally, 2012 ................................................ 26 Figure 6: HIV RNA levels and anti-HIV antibody responses among HIV-exposed infants with and without HIV infection. ....................................................................................................... 28 Figure 7: Number of children born to HIV-infected mothers tested for HIV in 2008-2009 .... 30 Figure 8: Early infant diagnosis cascade .................................................................................. 30 Figure 9: Strategies to identify unknown HIV infected children. ............................................ 31 Figure 10: Percentage of eligible children (0–14 years old) receiving antiretroviral therapy, 2012 .......................................................................................................................................... 33 Figure 11: Selection bias in observational cohorts of HIV-infected children .......................... 42 Figure 12: Basic CEPAC model structure ................................................................................ 44 Figure 13: Basic CEPAC-Paediatric model structure .............................................................. 45 Figure 14: Internal validation of clinical event risk outcomes: CEPAC-Paediatric model results compared to IeDEA data. .............................................................................................. 47 Figure 15: Conceptual framework of the care of paediatric HIV ............................................ 50 Figure 16: Map of West Africa ................................................................................................ 58 Figure 17: Timing of events X and Y relative to observation window A-B.. ........................ 164 Figure 18: Selection bias in observational cohorts of HIV-infected children. ....................... 165

15

List of abbreviations

3TC Lamivudine

95%CI 95% Confidence Interval

ANRS French Agency for Research on HIV/AIDS and Hepatitis

ARROW AntiRetroviral Research fOr Watato

ART Antiretroviral therapy

AZT Zidovudine

CD4 Glycoprotein found on the surface of immune cells

CEPAC Cost-Effectiveness for Preventing AIDS Complications

CHER Children with HIV Early antiRetroviral therapy trial

CM Child-Months

CY Child-Years

DBS Dried Blood Spot

DNA Deoxyribonucleic acid

EID Early Infant Diagnosis

GDP Gross Domestic Product

ICER Incremental Cost-Effectiveness Ratio

IeDEA International epidemiological Database to Evaluate AIDS

IMPAACT International Maternal Paediatric Adolescents AIDS Clinical Trials group

IQR InterQuartile Range

LPV/r Lopinavir/ritonavir

LTFU Lost To Follow-Up

MTCT/PMTCT Mother-to Child-transmission / Prevention of mother to child transmission

NIH National Institutes of Health

NNRTI Non-Nucleoside Reverse-Transcriptase Inhibitor

NRTI Nucleoside Reverse-Transcriptase Inhibitor

NVP Nevirapine

OI Opportunistic Infections

PACTG Paediatric AIDS Clinical Trial Group

PENTA Paediatric European Network for Treatment of AIDS

PI Protease Inhibitor

PNPEC National fight against AIDS programme in Côte d’Ivoire

POC Point of Care

PREDICT Paediatric Randomised Early vs. Deferred Initiation in Cambodia and Thailand

pWADA Paediatric West African Database on AIDS

RNA RiboNucleic Acid

UNAIDS Joint United Nations Programme on AIDS

UNICEF The United Nations Child Fund

WHO World Health Organisation

YLS Year of Life Saved

16

Introduction

By the end of 2012, an estimated 35.3 million people were living with HIV in the world,

including 3.3 million children less than 15 years (1). In 2011, it was estimated that 330 000

children were newly infected with HIV, almost half the number of children in 2003 and 24%

lower than in 2009 (2). Indeed, substantial progress has been made over the past decade in

reducing the number of children infected with HIV including the increase of the availability

of prevention of mother-to-child transmission (PMTCT) interventions that demonstrated their

efficacy.

However, the access to PMTCT interventions remains a challenge, in particular in sub-

Saharan Africa (3-6): of the 3.3 million children infected with HIV globally in 2012, 90% live

in this region.

In HIV-infected children, the natural evolution of HIV is rapid, leading to an early and high

mortality. Increased access to cotrimoxazole prophylaxis and antiretroviral treatment (ART)

for those who are infected has led to a global decline of HIV related deaths from 320 000 in

2004 to 230 000 in 2011 (7), but the scaling up of such care remains a challenge in resource-

limited settings (8, 9) and children continue to die. The number of HIV-related deaths among

children globally was estimated to be 210 000 in 2012 (1). Although the number of children

receiving antiretroviral therapy in 2012 increased, the scale-up was substantially slower than

for adults. In priority countries defined by UNAIDS, only three in 10 children receive HIV

treatment. The failure to expand access in many settings to early infant diagnosis (EID) is an

important reason explaining why HIV treatment coverage remains much lower for children

than for adults. In five priority countries, coverage of less than 5% was reported for early

infant diagnostic services in 2012 (1).

Although many reports demonstrate the feasibility and efficacy of ART for children in

resource-limited settings, important questions about diagnosis and therapy remain unanswered

to scale-up ART access in HIV-infected children. How can linkage between PMTCT services

and EID be improved? How can access to EID be improved? When to start ART in HIV-

infected children? What ART drug regimens to start with? What interventions would retain

children in care while waiting for ART initiation? What is the natural progression of HIV

infection before ART access? What is the impact of HIV-care interventions (ART,

cotrimoxazole)? What interventions would retain children in care after ART initiation? How

does HIV infection evolve in ART-treated children? While traditional clinical research

methods, such as observational studies and randomised clinical trials, provide information

essential to answering these questions, these studies cannot integrate all of the currently

available data, nor can they project long-term clinical and economic outcomes over the

lifetime of infected children. Computer simulation modelling is a useful adjunct to trials and

observational data in informing health policy. With this aim, in 1996, the Cost Effectiveness

of Preventing AIDS Complications (CEPAC) model was developed to evaluate strategies for

the management of HIV infection in the United States then internationally (10-12). In 2009,

the CEPAC investigators proposed developing a simulation model of HIV disease in infants

17

and children <5 years of age, the CEPAC-Paediatric model, to address policy questions

related to prevention, diagnosis and treatment of paediatric HIV.

Thus, my PhD research is built around informing on the care of HIV-infected children in the

West African setting, allowing investigators to derive input parameters for the CEPAC-

Paediatric model, but no model analyses will be presented in this work. Based on data issued

from two operational cohorts of HIV-infected children (Aconda and IeDEA West Africa)

implemented in the context of scaling up paediatric ART from 2004 and presented

subsequently, my PhD work articulates in four parts useful to contribute to the model: (i)

mortality, severe morbidity and healthcare resource utilisation prior to ART initiation, (ii)

reasons for hospitalisation in HIV-infected children, (iii) severe morbidity, healthcare

resource utilisation and treatment response after ART initiation, and (iv) the independent

ART-effect.

First, we will present a situation analysis of the paediatric HIV infection epidemic in West

Africa as well as the challenges in optimising paediatric HIV care. We will also present the

interest of computer simulation models in answering research questions and more specifically

describe the CEPAC-Paediatric model. In the following two chapters, we will describe the

conceptual framework of this research, the main objectives as well as the settings and the two

research projects which provided the data. The fourth, fifth and sixth chapters present the

results of our original research. We will present two studies investigating mortality, severe

morbidity and healthcare resource utilisation prior to ART initiation in children. We will then

further present reasons for hospitalisation related to paediatric HIV, and then two studies

investigating severe morbidity and healthcare resource utilisation and immune response after

ART initiation. We will then present the concept of the independent ART-effect and its

relevance in children. The last part of this work is dedicated to a general discussion on the

challenges in the management of HIV-infected children in West Africa and the perspectives

for my future researches.

18

1.FACTUAL BACKGROUND

Challenges raised by the paediatric HIV epidemic in

West-Africa

19

1.1.Situation analysis: the burden of paediatric HIV disease in West

Africa

1.1.1. Prevention of mother-to-child transmission (PMTCT) programme cascade

Paediatric HIV infection is the result of mother-to-child transmission (MTCT) of the virus,

nosocomial infections or, in older children, sexual transmission. It is estimated that 90% of

infections are the result of MTCT during pregnancy, the perinatal period or breastfeeding (1).

By the end of 2011, of the 34 million adults infected with HIV worldwide, 16 million were

women (2); in sub-Saharan Africa young women aged 15–24 were more than twice as likely

to be infected than their male counterparts in 2010. This feminisation of the epidemic is all

the more concerning in this region, with a high HIV prevalence among pregnant women.

Several interventions could prevent MTCT (1, 13):

1. Prevent new HIV sero-conversion among women of reproductive age;

2. Avoid unwanted pregnancies among HIV-infected women;

3. Offer universal access to counselling and HIV testing among pregnant women;

4. Offer antiretroviral PMTCT interventions for all HIV-infected pregnant women during

pregnancy, delivery and their new-born until the end of breastfeeding;

5. Offer Early Infant Diagnosis (EID) from six week of age in all HIV-exposed children.

6. Provide HIV care, treatment and support for women, children living with HIV and

their families.

Of these, all but EID were formally recommended in the Global Plan towards the elimination

of paediatric HIV before 2015.

The diagramme below (Figure 1) shows the number of different stages that a woman must

progress through to complete a PMTCT programme. Opportunities to maximize the

effectiveness of PMTCT interventions may be lost at each step in the pathway.

Source: Ciaranello et al. Plos One,2011;6:e20224 (6)

Figure 1: “Cascade” of PMTCT and postnatal HIV care.

20

1.1.2. The operational failure of prevention of mother-to-child transmission

(PMTCT) programmes

1.1.2.1. Efficient PMTCT regimens but low uptake

Before PMTCT intervention, the prevention of unwanted pregnancies among HIV-infected

women would prevent the paediatric epidemic. However, the unmet need for family planning

services among women living with HIV continues to undermine efforts to eliminate new HIV

infections among children (1). In East Africa and West Africa for instance, more than 20% of

women had an unmet need for family planning services, with no reduction in unmet need

reported for 1990–2010. This means that more than one in five women in the region express

the desire to delay or stop childbearing, but are not using contraception (1).

In the absence of any PMTCT intervention, the MTCT rate varies between 13% - 48% (14)

and is highest in African contexts mainly because of the long duration of breastfeeding. In

1994, the AIDS Clinical Trials Group Protocol 076 Study Group (PATCG 076 and ANRS

024 trial) demonstrated the efficacy of zidovudine in reducing MTCT of HIV from 24% to

9% for the first time in non-breastfed infants (15), leading to a significant reduction of

perinatal HIV in high-income countries. Since, in resource-limited settings, further research

has demonstrated the feasibility, efficacy and tolerance of PMTCT interventions during the

perinatal period in African breastfed children in the late 1990s (16-20). The success of these

studies led to the establishment of World Health Organisation (WHO) guidelines for PMTCT,

for the first time in 2000 (21).

Maternal combined short ART combined with postnatal interventions demonstrated their

efficiency in reducing substantially the risk of MTCT in African breastfed children, to less

than 5% such as in Côte d’Ivoire (22, 23). Research on PMTCT is a dynamic and rapidly

changing field; in 2010, guidelines took a major step towards more efficacious regimens,

providing life-long ART to eligible women and two short term prophylactic options for

women who were not yet eligible for ART (Options A and B – table 1). These

recommendations were revised in 2012 to include a third option: to provide life-long ART to

all HIV-infected pregnant women, regardless of their eligibility criteria according to their

clinical staging or CD4 cell count (Option B+). A number of countries are already adopting

or considering this approach (24). PMTCT recommendations were most recently updated in

2013 (25): “all pregnant and breastfeeding women with HIV should initiate triple ART,

which should be maintained at least for the duration of mother-to-child transmission risk.

Women meeting treatment eligibility criteria should continue lifelong ART”.

21

Table 1: Programme options for antiretroviral therapy for Prevention of Mother-To-

Child Transmission (WHO, 2012)

Source: Programmatic update: use of antiretroviral drugs for treating pregnant women and preventing HIV

infection in infants (24)

The overall PMTCT coverage in 2012 was 56% in the 21 sub-Saharan African countries with

a generalised epidemic prioritized by the Global Plan (1). But its distribution was

heterogeneous. Several countries have advanced efforts to prevent MTCT. Botswana and

Zambia have achieved more than 95% coverage of PMTCT services in 2012. Ghana is the

only West African country with a PMTCT coverage higher than 80%. However, elsewhere in

Western Africa, the coverage levels are much lower; in Côte d’Ivoire it was estimated to be

67%. In 2012, the lowest estimated PMTCT coverage was in Nigeria and the Democratic

Republic of Congo where it did not exceed 15% (26).

The first obstacle to PMTCT coverage is access to HIV diagnosis and antenatal care for

pregnant women (Figure 2). In 2012, only 44% of pregnant women in resource-limited

settings accessed HIV diagnosis and counselling overall (27). This is substantially higher than

22

the 35% reported in 2010 (7), but still remains insufficient. The proportion of pregnant

women receiving HIV testing and counselling varies greatly between regions, from more than

95% in South Africa and Zambia to 6% in Chad; in Côte d’Ivoire, less than 50% of pregnant

women received either an HIV test or counselling in 2010 (28).

The percentage of women receiving an HIV test in 2009 was estimated to be 21% in Western

and Central Africa; it was more than twice as high in Eastern and Southern Africa (Figure 2).

Source: WHO. Towards Universal access: Scaling up priority HIV/AIDS interventions in the health sector.

Progress Report 2010 (28)

Figure 2: Percentage of pregnant women who received an HIV test in low- and middle-

income countries by region, 2005 and 2008-2009.

In addition to HIV counselling and testing, HIV-infected mothers must be provided with

effective antiretroviral regimens from pregnancy throughout breastfeeding to prevent MTCT.

Those eligible for ART for their own health should also receive their treatment. In 2010,

approximately 51% of pregnant women who were tested positive for HIV were assessed for

their eligibility to receive ART for their own health (28). Access to antenatal and postnatal

services is another major obstacle to the success of PMTCT programmes.

Overall, in 2012, 61% of pregnant women received antiretroviral therapy to prevent MTCT of

HIV during pregnancy. However, this estimated coverage dropped to 30% during the

breastfeeding period (29). It is estimated that on average, nearly half of all children who are

newly infected with HIV are acquiring the disease during breastfeeding, because of low

antiretroviral coverage during this period. Overall, five out of ten women and children were

receiving antiretroviral therapy to prevent MTCT during breastfeeding in 2012; however, in

some countries coverage is higher than others. In Botswana, almost all eligible women were

23

receiving ART to prevent MTCT perinatally and seven out of ten women and children were

receiving antiretroviral drugs to prevent transmission during breastfeeding. In Southern and

Eastern Africa, many other countries report same coverage levels but in West Africa, this

PMTCT step coverage during breastfeeding lags. In Côte d’Ivoire, only one out of ten women

and children were receiving antiretroviral drugs to prevent transmission during breastfeeding

in 2012 (26). Providing antiretroviral prophylaxis to HIV-exposed infants during

breastfeeding is a critical component but also the strongest barrier to the success of PMTCT

interventions (Figure 3).

Source: 2013 progress report on the global plan towards the elimination of new HIV infections among children

by 2015 and keeping their mothers alive. UNAIDS.

Figure 3: Antiretroviral coverage to prevent new infections among children, and the

final HIV transmission rate from mother to child, including breastfeeding

1.1.2.2. From virtual elimination of paediatric HIV to the operational reality

in Western Africa

In 2009, the WHO called for the "virtual elimination" of paediatric HIV and launched in 2011

the Global Plan towards to the elimination of new HIV infections among children by 2015

and keeping their mothers alive (13). This plan included (Figure 4):

Reducing the number of new HIV infections among children by 90%

Reducing the number of AIDS-related maternal deaths by 50%.

24

Indeed, more effective antiretroviral regimens could reduce the MTCT rate to less than 1%

(30, 31). Furthermore, model-based analyses have proved that these new strategies are cost-

effective, both improving maternal and infant outcomes and saving money (32-34).

Source: Mahy et al. Sex Transm Infect 2010;86(Suppl 2):ii48eii55 (30)

Figure 4: New HIV infections through mother-to-child transmission (MTCT) for

different scenarios, 25 countries, 2009-15.

However, achieving a 90% reduction in new paediatric HIV infections by 2015 is an

ambitious target and requires leadership, commitment, and strategic deployment of resources

(35). Even when the most effective antiretroviral regimens are available for free through

public programmes, many women and infants are lost at different stages in the “PMTCT

cascade” (36). To reach virtual elimination of paediatric HIV, PMTCT programmes must

expand coverage and propose full packages including primary prevention of HIV among

women of childbearing age, family-planning, prevention of MTCT and the provision of

appropriate treatment, care and support. However, rapid expansion of delivery of effective

advances in preventing MTCT is being held back by inadequate access to antenatal and

postnatal services.

Routine, HIV testing and counselling of all pregnant women is the key entry point for

PMTCT services. In some high-burden countries of sub-Saharan Africa (Botswana, Namibia,

South Africa and Zambia) HIV testing rates in pregnant women are above 80% (35). But most

countries are still grappling with the best way to identify the greatest number of pregnant

women living with HIV.

Furthermore, providing the appropriate treatment to prevent MTCT requires immunological

assessment of all HIV-positive pregnant women for their ART eligibility through the use of

CD4 counts. In 2010, only 51% of HIV-infected pregnant women were assessed for ART

25

eligibility (7) and overall antiretroviral coverage still remains low. According to a recent

multi-level PMTCT assessment in 15 sub-Saharan countries, overall NVP coverage, the

simplest of all PMTCT strategies, reached only 44% (37). One of the main reasons is the lack

of human resources and infrastructures to deliver the services. Many studies have shown that

task-shifting allows increasing the number of antiretroviral, improving PMTCT coverage (38,

39). Further evidence suggests that the integration of maternal HIV testing into childhood

immunization clinics can significantly increase the identification of at-risk HIV-exposed

infants previously missed by traditional PMTCT models (40). Countries must now extend and

integrate PMTCT services at the primary care level in maternal and child health services (41)

Efforts to improve access to PMTCT services, retain women in care, and support medication

adherence throughout pregnancy and breastfeeding are needed to reach a MTCT risk of less

than 5%, as modelled in Zimbabwe (6).

Effective PMTCT services and implementation of the new WHO guidelines require pregnant

women to use existing maternal and child health services earlier and more frequently.

However, the availability of effective PMTCT services does not guarantee that women will

take advantage of them. Globally, 79% of pregnant women have at least one antenatal care

visit; this proportion drops through the continuum of care, and fewer women have at least four

antenatal care visits or deliver at health facilities. Various social, cultural and economic

factors still restrict the use of PMTCT services in resource-limited settings. Obstacles include

the high cost of antenatal and delivery care services, long waiting time, inability to pay for

transportation to health centres, lack of partner support and HIV-related stigma and

discrimination (36, 37, 40, 42).

Finally, the elimination of MTCT has significant financial implications. Despite considerable

efforts to increase domestic funding, many sub-Saharan countries remain highly dependent on

international aid to support their national AIDS response.

1.1.3. The paediatric HIV epidemic

PMTCT programmes have proved to reduce substantially the number of new infections in

children. However, despite the success of PMTCT in some settings, the failure to implement

programmes on an appropriate scale has resulted in thousands of new infections. It is

estimated that each day, one thousand children are newly infected with HIV in the world in

2009 (8). Of the 3.3 million children infected with HIV globally, in 2012 90% live in sub-

Saharan Africa (Figure 5) (1) .

26

Source: UNAIDS report of the global AIDS epidemic 2013(1)

Figure 5: Number of children living with HIV globally, 2012

Between 2009 and 2012, the overall number of new paediatric infections decreased by 38%;

however the pace of this decline has been slower in some regions compared to others. In

South Africa, the rate of new infections declined by 63% and in Botswana, Ethiopia and

Malawi it reached 50%. However, in West Africa the rate of decline in new infections lags; in

Côte d’Ivoire, it was estimated to be less than 30% over 2009-2012 (26).

In 2012 the WHO estimated that 210,000 children aged less than 15 years died of HIV/AIDS

worldwide (1); a meta-analysis reported that in 2008, 90% of these deaths occurred in the

African region (43). HIV/AIDS is an increasingly important cause of under-five mortality in

Africa (44) and caused 4% of under-five mortality in sub-Saharan Africa (43), however trends

are geographically disparate. In 2012, Côte d’Ivoire reported a 3% under-five mortality rate

caused by HIV, but this reached 28% in more prevalent settings, such as South Africa (29).

HIV infection progresses more rapidly in children compared to adults. HIV-related deaths

peak within the first six months of life, followed by a more gradual progression towards AIDS

in the surviving children. The natural history of HIV has been well described in European and

American cohorts. Before the introduction of ART, 26% of HIV-infected infants developed

clinical events within the first year of life and 17% died (45). In Africa, without treatment,

these rates are even higher: for instance in Rwanda, the estimated cumulative risk of

developing AIDS in HIV-infected children was 28% and 35% at 2 and 5 years of age,

respectively. The estimated risk of death among infected children at 2 and 5 years of age was

45% and 62%, respectively; these rates were 21 times higher than those observed in

uninfected children (46).

Bimodal expression of HIV-1 paediatric disease is encountered in Africa, as in industrialised

countries, but its prognosis is poorer: it was estimated that one third of children born with

HIV die before their first birthday; 50% die before they turn two (47, 48). The risks of death

27

are all the more important as children are infected early (49, 50). In the absence of appropriate

treatment, HIV mortality is high, varying between 49% - 59% in low-income countries (46,

51). The net survival at 5 years, in the absence of treatment, is 33% in perinatally infected

children and 60% in those infected during breastfeeding (50).

Paediatric HIV is characterised by early and frequent infections: diarrhoea is the most

common cause of death in infancy, either because of rotavirus or because of other

opportunistic infections (52, 53), such as pneumonia, common in African children, and often

considered the leading cause of mortality (54, 55). Furthermore, children with HIV infection

are more at risk of malnutrition because of anorexia, oral infection and malabsorption (56).

Co-infection with other organisms is also frequent; HIV increases the risk of tuberculosis,

which is more progressive and severe in these infected children (57). HIV-infected children

living in malarial endemic areas also have more frequent malarial episodes (58, 59). Other

predictors for early mortality include advanced maternal HIV disease, maternal anaemia,

delivery complications, early growth faltering, formula feeding and low infant CD4 (48, 60-

66).

Early paediatric HIV diagnosis is thus critical to allow the timely start of appropriate

treatment, reduce morbidity and mortality and improve the quality of life of HIV-infected

children in Africa (67). The WHO has advocated for increased attention to diagnosing HIV in

infants as a critical final step for PMTCT programmes, which cannot prevent all infections.

However, identifying and diagnosing HIV-exposed infants remains challenging in Africa, for

both logistical and social reasons.

1.1.4. Diagnosing HIV infection in children is a major challenge in Africa.

Techniques for infant HIV diagnosis are not always available on a routine basis

The usual HIV testing algorithm in adults is a combination of two tests for HIV antibodies:

the ELISA method followed by a Western Blot in USA or a combination of two HIV rapid

tests. However, for the infant diagnosis and because of the trans-placental passage of maternal

HIV antibodies to the virus, these tests lack in specificity and sensitivity when differentiating

maternal and infant antibodies in infants up to 18 months of age, according to figure 6 (68).

Consequently, virologic assays, including HIV-1 DNA or RNA assays and p24 antigen assays

represent the gold standard for diagnostic testing of infants and children younger than 18

months (69-71). With such testing, the diagnosis of HIV-1 infection can be established within

the first several weeks of life among non-breastfed infants (72) and throughout the

breastfeeding period, by 4-6 weeks after exposure, for others (68, 73, 74).

28

Source: Ciaranello et al. BMC Medicine 2011,9:59 (75)

Figure 6: HIV RNA levels and anti-HIV antibody responses among HIV-exposed infants

with and without HIV infection. The horizontal axis shows infant age in months. The left vertical axis

shows mean HIV-1 RNA level on a logarithmic scale, and corresponds to the green lines on each graph. The

right vertical axis shows the proportion of infants for whom an HIV antibody test would likely return positive,

and corresponds to the red lines on each graph.

29

While early infant diagnosis is increasing in West Africa, there are ongoing challenges of

access and return of results in infants testing positive. Virological tests are expensive and

require sophisticated laboratory facilities. The costs of creating a laboratory with appropriate

quality control and assurance to perform virological testing are substantial, as are the costs of

test reagents (73, 75-78). Logistical barriers persist as the adequate facilities are often

unavailable outside large cities resulting in difficulties in the transport to and from clinical

sites reducing acceptability of testing and causing results to go unclaimed (75, 77).

Furthermore, venous blood sampling from infants is invasive, often difficult and involves

conditioning in the laboratory before any analyses can be carried out (75).

For some years now, the use of dried blood spots (DBS) has improved considerably the access

to early infant diagnosis. DBS consists of whole blood obtained via heel stick or finger prick

that is then dried on filter paper. DBS samples are non-infectious and can be transported

easily via mail or courier reducing the cost of transport as they do not require to be kept cool

nor immediate transportation to the laboratory (79). DBS-based testing has been evaluated in

many settings, is both highly specific and sensitive (79-82) and is thought to be more

acceptable for parents (83). The use of DBS based testing has improved testing networks, as

in South Africa, but still needs to be generalised in many African countries (79).

Social acceptability of EID remains low in West Africa

Alternative EID strategies have been proposed to reach infants with both known and unknown

HIV-exposure, for example offering EID to all infants at postnatal points of care. However

challenges to these programmes include parental and healthcare staff acceptability of EID,

that will also affect the efficiency of these strategies (84). High acceptance rates of EID have

been reported in high-prevalent settings, reaching 90% (85), but post-test return rate remains

low, reaching only 57%. In low-intermediate prevalent countries, such as Côte d’Ivoire a

study has reported much lower acceptance rates, around 58% for mothers and 15% for the

father (86).

Finally, diagnosing infants at the earliest stage requires active investment in key areas such as

training and support for providers. In Côte d’Ivoire a recent study reported that no staff in

immunization services and only 26% of staff in paediatric services were trained in counselling

and testing (87). Infant diagnosis at a primary care level is often limited by lack of appropriate

training of healthcare workers.

Coverage of EID remains low in West Africa

Despite the 2006 WHO recommendations to promote EID (88), many HIV-infected children

remain undiagnosed at infancy (Figure 7). In 2010, WHO reported that only 28% of HIV-

exposed children were tested and this rate did not exceed 10% in West and Central Africa en

2009 (7).

While the EID coverage increased substantially in Southern Africa (89-92), it still remains

low in West Africa (83).

30

Figure 7: Number of children born to HIV-infected mothers tested for HIV in 2008-2009

Despite the expanding access to EID and the feasibility and acceptability in resource-limited

settings, the cascade of care required for effective programmes is complex and many infants

are lost from care at each step (Figure 8). As a consequence, most HIV-infected infants are

diagnosed at a much later age as part of workup for recurrent or severe ill health. This means

that nearly 50% of HIV-infected infants die before their HIV status is known (48).

Source: Ciaranello et al. BMC Medicine 2011,9:59 (75)

Figure 8: Early infant diagnosis cascade

Efficient PMTCT regimens exist and programmes are effective in some settings, going

towards the eradication of paediatric HIV, as in Rwanda. However, in West Africa, the

operational implementation of PMTCT is a failure and many children continue to be infected.

These children need to be identified and enrolled in HIV care programmes the earliest

possible.

31

Figure 9 shows the difficulty in identifying children who are born to mothers who were never

enrolled into PMTCT programmes. These women contribute the majority of infected children

in the community and more needs to be done to identify, test and enrol those children by

implementing routine approaches to testing in settings where children are accessing care

services. These strategies are context-specific and depend highly on HIV prevalence (84).

Source : Kellerman & Essajee, PLos Med, 2010(84)

Figure 9: Strategies to identify unknown HIV infected children.

1.1.5. Access to care of HIV-infected children

1.1.5.1. Cotrimoxazole prophylaxis

Cotrimoxazole prophylaxis is an important component of the HIV care and treatment package

for children. Cotrimoxazole is a combination of two antibiotics, trimethoprim and

sulfamethoxazole. It is safe, inexpensive and has a broad-spectrum prophylactic action against

common bacterial pathogens such Pneumocystis jirovecii, responsible for pneumonia in HIV-

infected children (93). It has also proved to act against protozoa such as Plasmodium spp, thus

also having a protective effect against malaria (94, 95). Several studies report this drug as

highly effective in reducing morbidity and mortality among HIV-infected children (93-99).

Data from a randomised clinical trial in Zambia demonstrated that this drug reduced mortality

in children by 43% and hospital admission rates by 23% compared with matched placebo.

The benefits of cotrimoxazole were demonstrated at all ages beyond 12 months and all levels

of CD4 percentage (100).

32

Because of the difficulty in diagnosing early HIV infection in infants, cotrimoxazole

prophylaxis is recommended for all HIV-exposed children born to mothers living with HIV

starting at 4–6 weeks after birth and continuing until HIV infection has been excluded and the

infant is no longer at risk of acquiring HIV through breastfeeding in Africa (101).

However, WHO 2010 estimates reveal that only 23% of infants needing cotrimoxazole

prophylaxis had been started on it by two months of age (7). The main reason for the low

uptake of cotrimoxazole prophylaxis appears to be largely dependent on the lack of

understanding of guidelines by the health service providers (102). Furthermore, because of the

weak links between PMTCT programmes and paediatric HIV care programmes, there are

many missed opportunities to identify HIV-exposed and infected children in need of

cotrimoxazole prophylaxis. Other barriers include limited policy support, insufficient training

of healthcare workers and lack of understanding of the importance of such prophylaxis and

insufficient quantities of cotrimoxazole made available (103).

1.1.5.2. Antiretroviral treatment

Low coverage of ART in children in West Africa

Clinical care and access to antiretroviral therapy for HIV infected children in African contexts

is a complex problem to analyse, made difficult by operational issues. Access to antiretroviral

therapy has highly improved, reaching 34% in 2012, but remains insufficient with only one

third of eligible children in sub-Saharan Africa who access treatment (26). This is much lower

than the 65% coverage of antiretroviral therapy in adults in 2012 (1). The number of eligible

children receiving antiretroviral therapy has increased in all countries in sub-Saharan Africa,

but in some more than others. Botswana and Namibia have already achieved universal access,

with >80% of the eligible children receiving antiretroviral treatment and more than half the

eligible children are receiving ART in South Africa. However, in Western and Central Africa,

despite the increase in numbers of eligible children receiving treatment, coverage remains

unacceptably low, between 9% and 25% (Figure 10) (26).

Providing a continuum of care remains challenging and children who have a chance to access

antiretroviral therapy are frequently treated at an advanced age (median age >5 years) and at

an advanced stage of the disease, where clinical symptoms have appeared and therefore when

they are already immunodeficient (104-107). This delay in antiretroviral treatment is

associated with high mortality around the start of treatment; in a pooled analysis of more than

2000 children treated in Africa, an abnormally high and early mortality rate was observed in

the first months after the beginning of treatment, more frequent in immunodeficient children

compared to children at less advanced stage of immunodeficiency (105). Many studies have

investigated the determinants of this high mortality; immune reconstitution inflammatory

syndrome previously described in Thailand and in South Africa (108, 109) is a reason, and

could be controlled by an earlier start of antiretroviral therapy, at a less advanced stage of

immunodeficiency. Other determinants of early mortality are high viral load at ART initiation

(110), choice of first line ART regimen (111), low hemoglobin levels (104, 112, 113) and age

33

(104, 114). More recently, in a multiregional analysis of 13 611 children on ART, from Asia,

East Africa, Southern Africa, and West Africa, age < 24 months, WHO stage 4, CD4 < 10%,

attending a private sector clinic, larger cohort size, and living in West Africa were

independently associated with poorer survival (115).

Source: 2013 progress report on the global plan towards the elimination of new HIV infections among children

by 2015 and keeping their mothers alive. UNAIDS (26)

Figure 10: Percentage of eligible children (0–14 years old) receiving antiretroviral

therapy, 2012

Another challenge that health programmes face is ensuring that all children who are tested

positive for HIV are successfully linked to and retained in a paediatric HIV care programme

so they can initiate ART the earliest possible (106, 116). However, early ART initiation is

often limited by competing death and losses to follow-up (LTFU) in the interval between

diagnosis and initiation of ART (117, 118). Prior to ART initiation, CD4 cell count is

determined, ART eligibility is assessed and in some settings, caregivers go through a process

of pre-treatment counselling. Children are LTFU at each step of this pre-ART cascade (117).

Few data exist on loss to programme between HIV testing and ART initiation in children. In

2011, a systematic review reported on retention in pre-ART care in HIV-infected children

(117): the percentage of children diagnosed with HIV who benefited from a CD4 cell

measurement ranged from 78% to 97% at the median age of 2.2 to 6.5 years. Among these

children, the proportion of those eligible for ART ranged from 63% to 89%, and the

proportion of children actually initiating treatment ranged from 39% to 99%. Because of the

considerable challenges in early infant diagnosis, many children are only tested for HIV

because they are symptomatic. These children are at an advanced stage of clinical disease and

34

immunodeficiency and consequently, have high pre-treatment mortality rates (118-121).

Furthermore, children follow-up in ART care are strongly influenced by their

maternal/caregiver lifestyle (122) and consequently socio-economic obstacles are major

reasons for poor retention in pre-ART care programmes. Long queues, overcrowding,

negative staff attitudes, and provider difficulties with paediatric counselling are often cited as

reasons for not attending appointments (123). Transport to and from the clinic is also a factor,

though many programmes reimburse these costs. Finally, the fear of social rejection from the

disclosure of the child’s status is also a reason for LTFU during this pre-ART period (124,

125).

All findings raised overall concerns about delayed ART initiation, low access to free HIV

services for children, and increased workload on programme retention in lower-income

countries. Universal free access to ART services and innovative approaches are urgently

needed to improve paediatric outcomes at the programme level.

WHO recommendations towards an earlier ART initiation

Many factors show the benefits of early antiretroviral care in HIV-infected children at an

asymptomatic stage with low immunodeficiency on improving survival and immuno-

virological response (126). This would also likely reduce toxicity when starting treatment

(127, 128).

In 2009, the Children with HIV Early Antiretroviral therapy (CHER) trial in South Africa

demonstrated the effectiveness at 12 months of early antiretroviral treatment initiated from 12

weeks of age in all HIV-infected children. A very significant reduction by 76% in infant

mortality among children treated immediately at 12 weeks compared to those treated at a later

age, according to the 2006 WHO recommendations was observed: 4% versus 16% (129).

These very convincing results incline towards the initiation of early antiretroviral therapy for

children diagnosed with HIV within the first weeks of life. The paediatric advisory board

consultation for the WHO revised treatment recommendations during the technical

consultation in April 2008, so that systematic antiretroviral treatment would be initiated for all

infected children under the age of 12 months and systematic early infant HIV diagnosis

carried out for all exposed children from 6 weeks of age. These recommendations were

revised in 2010, extending immediate ART initiation to all children less than 24 months of

age (130); treatment was initiated in older children according to clinical or immunological

criteria (WHO clinical stage 3 or 4 disease or CD4 350 cells/mm3). Very recently, in June

2013, the WHO revised simplified recommendations to initiate ART in all HIV-infected

children up to five years of age and to increase the CD4 count threshold for ART initiation to

500 cells/mm³ in children older than 5 years (Table 2) (25).

35

Table 2: Summary of recommendations of when to start antiretroviral therapy in

children

Source: Consolidated guidelines on the use of antiretroviral drugs for treating and preventing HIV

infection. 2013. (25)

Operational difficulties in the roll-out of early ART

There are many difficulties in the long-term follow-up of these children and in ensuring the

safety and efficacy of the antiretroviral treatment. To this are added drug adherence issues and

social difficulties encountered by children’s families.

South African and Rwandese national programmes are already introducing immediate ART

for children younger than five years but improving access to ART remains a major challenge

in most African countries where:

Expanding ART coverage requires more effective approaches to cover HIV testing in

children. Integrating HIV testing into routine health services has helped to increase the

uptake of testing but not sufficiently. Appropriate infrastructures are rare, especially in

rural settings, and the organisation of the transport of blood samples to laboratories is

precarious. The follow-up of HIV-exposed infants needs urgently improving, mainly

in low/middle HIV prevalent countries.

Insufficient human and material resources are an important factor in the delay of

treatment in eligible children. The scaling up of ART necessitates treatment

implementation in a large number of children in a short amount of time. Staff

shortage, lack of training for existing staff and inadequate facilities for storage and

roll-out limit access to ART.

Treatment options are limited. Child-friendly drug formulations are rare, particularly

for the youngest and most vulnerable children.

As more people start antiretroviral therapy, concerns are growing about the possible

increase in HIV drug resistance. Routine viral load and CD4 monitoring is difficult to

implement. Detecting virological failure and/or mutations and resistance in HIV-

infected children is difficult, limiting treatment options.

36

Antiretroviral therapy alone is not sufficient for the care of infected children. In order

to provide optimal care, clinics must ensure that the children are receiving a whole

care package, including immunisations, nutritional supplementation and prevention of

malaria. Furthermore, psychosocial elements relative to the attitudes and practices of

the caregiver must be considered in order to maximise treatment observance.

Appropriate HIV care of children caregivers is also challenging.

Finally, the expansion of ART coverage for young children will likely represent an

increased burden on the already fragile health systems in resource-limited settings

(131).

Table 3 summarises the estimates in preventing mother to child transmission and providing

paediatric care and treatment in sub-Saharan Africa for selected regions and countries in

2005, 2009 and 2012.

Table 3: Summary of estimates in preventing mother to child transmission and providing paediatric care and treatment in sub-Saharan Africa for selected

regions and countries in 2005, 2009 and 2012.

Sources of data: UNAIDS, WHO, UNICEF, UNGASS, Global Bank, World Health Statistics, Health Demographic Survey and CHAI

��

��

��

��

��

��

��

��

��

�� !��"

��!��#� ��"��!��#� ��$��

$��

��

��

��

%� "%�#��

$!��

��

%��#��$!��

��

&��$!��

'��

'�(��)��

�*��+��

,#��

��

%��

��

38

1.2. Research gaps and many unanswered questions in promoting and

assessing paediatric HIV care in Africa

Although many reports demonstrate the feasibility and efficacy of, HIV diagnosis and early

ART for children in resource-limited settings, important questions remain unanswered. Below

is a non-exhaustive list of research gaps in the care of HIV-infected children. I will consider

some more specifically in the West-African context, in my PhD manuscript.

1.2.1. Clinical research gaps

What is the paediatric natural history of HIV? HIV-1-infected infants develop disease

manifestations early in life. The bimodal evolution that is observed in northern settings is

also observed in Africa, but HIV disease progression seems to be more severe in these

resource-limited settings with a higher residual morbidity and despite a higher frequency

of late acquisition of infection related to breastfeeding (46). Factors associated with pre-

ART mortality include younger age, immunodeficiency, anaemia and low weight-for-age

(120). Since the introduction of ART in 2004 in Africa, few studies have investigated the

natural history of HIV infection before ART initiation and in a context where the standard

of care may have improved. These base case data are nevertheless useful to assess the

effects of HIV care interventions.

When to start ART in HIV-infected children? The CHER trial has proved the benefits of

early ART in children aged < 12 months. However there is limited evidence on the

optimal timing in older children. Observational studies of African birth cohorts have

demonstrated that untreated HIV-infected children aged > 12 years continue to

experience high mortality (48); however no trials support this evidence. The Paediatric

Randomised Early versus Deferred Initiation in Cambodia and Thailand (PREDICT) trial

enrolled 300 children aged 1-12 years with initial CD4 percentage (CD4%) varying from

15-24% and without CDC Stage C disease. These children were randomly assigned to

start ART immediately or defer therapy until their CD4% <15% or a CDC Stage C event

occurred. There was no difference between both arms in the 3-year risk of mortality,

AIDS-free survival, occurrence of stage B and C events, neurodevelopmental outcomes,

rates of hospital admission or drug related adverse events (132). Another study, based on

an observational cohort also reported no survival benefit from early treatment in children

aged 24-59 months (133). Clinical trials are difficult to perform. The PREDICT trial like

the CHER trial in the past provide important insights for the management of children

living with HIV but cannot address all the questions needed for shaping the public health

approach and the clinical guidelines of paediatric HIV infection (134). Despite limited

clinical evidence, WHO recommends early ART in all children aged > 5 years to simplify

the guidelines (24). However, what are the clinical and financial implications of such

strategies? Results should be completed by cost-effectiveness modelling to better inform

the competing demands for limited funds and resources to optimise efficiency of ART

paediatric programmes.

39

What ART regimen to start? Children who are exposed to Non-Nucleoside Reverse-

Transcriptase Inhibitors (NNRTI) used for PMTCT are subject to viral resistance (135),

which compromises the response to first-line ART based on NVP (136, 137). For this

reason, the 2010 WHO guidelines recommended the use of Protease Inhibitors (PI) with a

LPV/r-based treatment in children younger than 24 months of age previously perinatally

exposed to NNRTIs. For infants not exposed to NNRTIs during their perinatal period or

whose status was unknown, an NVP-based regimen was recommended (130). The

IMPAACT P106 trial brought extra evidence suggesting the superiority of a LPV/r-based

regimen regardless of PMTCT exposure (138, 139). In 2013, the WHO revised guidelines

as follows: “a LPV/r-based regimen should be used as first-line ART for all children

infected with HIV younger than three years (36 months) of age, regardless of NNRTI

exposure” (25). However, providing an LPV/r-based regimen to infants and in resource-

limited settings remains challenging. The current LPV/r syrup formulation has cold-chain

requirements until the point of dispensing, is costly and the syrup is unpalatable, which

strongly affects its observance in older children. Furthermore, PIs are mostly used as

second-line treatment, limiting treatment options in case of failure in resource-limited

settings. Clinical trials are currently ongoing to study the feasibility of initiating an LPV/r-

based regimen and then switching to NNRTIs (140, 141). Such PI-sparing strategies could

avoid costs and lead to better tolerability. This approach has been proved effective (142,

143) but could add complexity to treatment programmes.

Can planned treatment interruptions be envisaged in children? ART is a lifelong

treatment and patients are at risk of developing viral mutations and drug resistance;

furthermore, prolonged ART can lead to toxicity. Planned treatment interruptions after

maximal viral load suppression has been achieved are a possible strategy, though it has

been proved non-viable in adults (144, 145). The PENTA 11 trial investigated this, and

although children who had interrupted ART were recommended to restart at the end of the

trial, there were no clinical consequences observed in the next two years following this

suggesting that so far this strategy is viable in children (146). This questions was also

within the CHER trial, which also compared early time-limited ART with deferred ART.

Results have been very recently published, reporting better clinical and immunological

outcomes in the early time-limited ART group compared to those on deferred ART.

Furthermore, investigators concluded longer time on primary ART permits longer

subsequent interruption, with marginally better outcomes (147). Nevertheless, further

research in resource-limited settings is needed to investigate this.

What is the long-term evolution of HIV disease in children on ART? Short-term

mortality has been widely investigated in children initiating ART in Africa. However, few

studies have investigated the overall severe morbidity in children on ART, or their

healthcare resource utilisation and associated costs. Furthermore, what is the long-term

immune response to ART and do children fully recover their immune system and when

after treatment initiation? Finally, what is the long-term tolerance of ART regimens?

40

1.2.2. Operational research gaps

Which public health approaches to access to HIV-exposed infants? Too often, people

who take HIV tests do not return to collect the results. Integrating universal HIV

screening based on DBS in routine paediatric health services has increased accessibility

and uptake. However in low HIV prevalent settings, where coverage of PMTCT

interventions is lower, this strategy is insufficient.

Which public health approaches and what services linkages to optimise early HIV

diagnosis in children and increase their access to early ART initiation? To enrol in

EID and access early ART, an infant must be brought by a caregiver to a healthcare

facility. In Côte d’Ivoire only 58% of mothers accepted EID. Among these, only 15%

were tested (83). Furthermore, a large proportion of test results never reach the intended

health facility nor result in mother notification (148). Improving the health system is a

necessity in order to track and link HIV-infected infants to care (149). Furthermore,

technology innovations are a possibility such as SMS interventions; for instance, the

implementation of the internet-based programme HITSystem in Kenya has significantly

increased the proportion of HIV-infected infants started on early ART (150). Indeed,

mobile phones have proved their effectiveness in health communication in other contexts

such as improving immunisation coverage in Burkina Faso (151) or increasing skilled

delivery attendance among women in Zanzibar (152). Mobile phone messaging is poorly

described in the context of paediatric HIV: in 2012, a systematic review reported on

effectiveness, acceptability and feasibility of the use of mobile phone messaging for HIV

infection prevention, treatment and care: of the 21 studies included, none addressed EID

(153).

What interventions to improve retention in HIV care? The magnitude of LTFU and

death between HIV diagnosis and start of ART in children are not well defined. Most

children who present for HIV care meet eligibility criteria for ART; however ART

initiation is not immediate. Concomitant severe infections such as tuberculosis are often

reasons for delaying ART initiation (154), co-treatment options need further

investigations. Retaining HIV-exposed or HIV-infected children in care also presents

unique challenges because of their dependence on a caregiver. LTFU is particularly high

along the continuum of care, with retention especially challenging for children who are in

HIV care but not yet eligible for ART.

41

What monitoring of ART response in children on ART? When to switch to second line

ART and what to switch to? For children who develop virological failure on first-line

ART, prompt switching to a second-line regimen can suppress viral load and restore

immune function, reducing the risk of mortality (25, 130). However, when virological

assays are not available, as is often the case in resource-limited settings, providers rely on

a combination of clinical and immunological criteria. Studies have shown that these

algorithms have low sensitivity and specificity, leading to misclassification as virological

success and delayed switching (155-159). There is, however, evidence on the

effectiveness of the clinical monitoring of children on NNRTIs, reported by the ARROW

trial (160). The cost and feasibility of such strategies remain to be investigated.

1.3.Which study design when addressing the above questions?

1.3.1. Observational cohorts and clinical trials gaps

Traditional clinical research methods, such as observational studies and randomised trials,

provide information essential to partly answering the above questions; however these studies

are also very difficult and time-consuming to implement, often conducted in specific settings

and represent a small amount of selected children.

Indeed, among all children aged 0 to 15 years in 2013, only a selected amount access through

PMTCT services. In 2009, in Côte d’Ivoire, approximately 10% of HIV-exposed infants,

regardless of their perinatal exposure to PMTCT interventions, were diagnosed, among which

a small proportion initiated early ART and an even smaller proportion before the age of 2

years as recommended by WHO before 2013. Furthermore, among those who do not access

EID services, and who survive early infancy, HIV diagnosis is indicated because of clinical

symptomatology suggesting HIV infection. Consequently, the participation of HIV-infected

children in a clinical trial is conditioned by their survival until their effective access to HIV

diagnosis and care which is now to occur late in the early 2010s (115). Consequently, when

keeping in mind the early mortality peak observed in HIV-infected children during the first

two years of life (48, 161), it is acknowledged that a left truncation bias will affect the cohorts

that will have to be analysed cautiously (Figure 11) . Second, clinical research outcomes are

limited on their generalizability, often more representative of urban settings rather than rural

settings (115).

42

Figure 11: Selection bias in observational cohorts of HIV-infected children

1.3.2. Computer simulation models comprise important adjuncts to more

traditional research methods

Traditional clinical research methods, such as observational studies and randomised trials

cannot integrate all of the currently available data, nor can they project long-term clinical and

economic outcomes over lifetime of HIV-infected children. Simulation models can integrate

available data, project long-term clinical and economic outcomes beyond study periods,

identify influential parameters for which additional data are needed, and be helpful to inform

care and treatment guidelines (10, 11, 162-165).

Simulation modelling is an increasingly well-established and essential tool for many

epidemiological investigations, including studies of population disease dynamics and

evaluation of mechanisms of disease control (166, 167). They present an alternative to

mathematically oriented modelling approaches, often deterministic, as they attempt to mimic

the actual processes that occur within a system. In deterministic models outcomes are

precisely determined through known relationships among states and events, without any room

for random variation. In such models, a given input will always produce the same output. In

comparison, stochastic models use ranges of values for variables in the form of probability

distributions (168). Several approaches to disease modelling exist including chain binomial,

Markov chain, and state transition models (169-172). Simulation models also have different

levels of complexity and can include spatial and temporal concepts, as well as economic

aspects of a disease (173). Epidemiological models may take any of several forms; there is no

clear set of rules that distinguishes models from one type to another and there exist models

that defy the very broad and brief characterizations above.

Exposed to PMTCT

HIVdiagnosis

ART

All HIV-infected children

aged 0-15 years

43

To date, only three published analyses have reported on simulation models of HIV-infected

children:

A Markov model used to evaluate the cost-effectiveness of cotrimoxazole prophylaxis

in HIV-infected children in Zambia. Authors built a probabilistic decision analytical

model of HIV/AIDS progression in children, based on CD4 cell percentage. Cost-

effectiveness was calculated across a number of different scenarios at tertiary and

primary healthcare centres (174).

A decision-analytic model of strategies for early infant diagnosis. Two models were

constructed to estimate the effectiveness and cost-effectiveness of conventional and

modified testing algorithms in different age groups and to assess the performance of

these algorithms if used routinely (78).

A stochastic agent-based simulation model used to evaluate the cost-effectiveness of

laboratory monitoring of ART in Thailand. The model simulated costs and clinical

outcomes over time according to different frequencies of viral load monitoring and

initiation of second-line treatment when appropriate (175).

1.3.3. An example of a computer simulation model: the Cost Effectiveness of

Preventing AIDS Complications (CEPAC)-Paediatric model

1.3.3.1. The CEPAC model

When the complexity of a clinical problem involves competing choices and the information

required for some components of the decision is incomplete, decision-analysis methods offer

an opportunity to synthesizing existing data and quantifying the trade-offs for alternative

options using sensitivity analysis. With this aim, since 1994, the Cost-Effectiveness of

Preventing AIDS Complications (CEPAC) team, led by Pr. Kenneth Freedberg at Harvard

Medical School, has used simulation models to evaluate clinical outcomes, costs, and cost-

effectiveness of strategies for treating HIV/AIDS and its complications. The CEPAC model is

a validated, individual patient-level (Monte Carlo) simulation of HIV disease initially

developed in adults that has informed HIV testing and treatment policy in the United States ,

as well as investigators from Brigham and Women’s Hospital, the Harvard School of Public

Health, Weill Cornell Medical College, and Yale University (10, 165, 176-178). The US

CEPAC team collaborates internationally with research teams in Côte d’Ivoire, France, India,

South Africa, and Zimbabwe.

The adult CEPAC model is a computer-based simulation model of the progression and

outcomes of HIV disease in a cohort of hypothetical patients. Although the patients are

44

hypothetical, the data used to define their characteristics are collected or derived from real

patient data input (issued from clinical trials, cohort studies, patient health databases, etc.).

Disease progression of each simulated patient is characterised by a sequence of monthly

transitions from one “health state” to another (“state-transition” model). The model defines

three general categories of health states: chronic, acute, and death. Normally, patients reside

in one of the chronic states, where progression of disease and immune system deterioration

occur. Patients who develop an acute complication temporarily move to an acute health state

in the month of the event. In this state, quality of life is lower and healthcare resource

utilisation and mortality rates are higher. Death can occur in a chronic or acute health state,

and can be attributed to a particular opportunistic infection (OI), chronic AIDS (e.g. wasting),

or non-AIDS-related causes (e.g. getting hit by a bus). The basic structure of the adult

CEPAC model is represented in Figure 12 below.

Source: CEPAC model user’s guide (available from:

http://web2.research.partners.org/cepac/documents/Cepac%20User%20Guide.pdf)

Figure 12: Basic CEPAC model structure

Important determinants of a patient’s state at any given point in time include age, gender,

current and past CD4 cell count and viral load, history of OI(s), HIV infection status, and

currently administered therapies (including OI prophylaxis and ART). The model uses the

Monte Carlo method (random number generator) to generate one hypothetical patient at a

time and draw from a set of transition probabilities to determine the occurrence of clinical

events and health-state transitions experienced by the patient. Each simulated patient’s

clinical course is tracked from time of entry into the model until death. Upon the patient’s

death, summary statistics are recorded and a new patient enters the model. The simulation run

is complete when the last patient in the cohort has passed through the model. Cohorts of 10

million patients are simulated to generate stable model outcomes. The programme maintains

tallies of clinical events, duration of time spent in each health state, monthly life expectancy,

quality-adjusted life expectancy, and costs.

For each evaluated strategy of care, we describe key clinical outcomes (survival and life

expectancy), economic outcomes (costs over 1, 5, 10-year and lifetime horizons), and

incremental cost-effectiveness ratios (ICERs). An ICER equals the difference in lifetime costs

45

for two strategies divided by the difference in life expectancy, in US dollars per year of life

saved ($/YLS) (162). There is no universal definition of a threshold ratio above which an

intervention would not be considered cost-effective. The WHO recommends that an

intervention with ICERs that are less than one times a country’s per-capita gross domestic

product (GDP)/YLS be considered “very cost-effective”, and less than three times the per

capita GDP be considered “cost-effective” (179).

1.3.3.2. The CEPAC-Paediatric model

Building on this adult CEPAC model platform, the CEPAC Team developed a simulation

model of HIV disease in infants and children < 5 years of age, the CEPAC-Paediatric model.

Infants enter the model at birth, after HIV-infection acquired in utero or during delivery. A

random number generator is used to draw from user-specified distributions of CD4% and HIV

RNA level at birth, maternal CD4 and receipt of ART, PMTCT exposure, and breastfeeding

or replacement feeding. In the absence of ART, each simulated child's CD4% declines

monthly at a specified rate until age five, when immunologic status becomes characterised by

absolute CD4 count instead of CD4%. At age 13, children enter an adolescent model, similar

in structure to, but with different data inputs than, the adult CEPAC model (Figure 13).

Figure 13: Basic CEPAC-Paediatric model structure

Like the adult model, disease progression in the CEPAC-Paediatric model is characterised by

monthly transitions between health states, including chronic HIV infection, acute illness, and

46

death. In each month of the simulation, random numbers determine transitions between these

health states, based on probabilities that are specified as model inputs. Transition probabilities

depend on current age and current CD4%. Simulated patients face monthly risks of acute

"clinical events"; model analyses simulate 3 mutually exclusive categories of clinical events:

WHO Stage 3 (excluding pulmonary and lymph node tuberculosis), WHO Stage 4 (excluding

extrapulmonary tuberculosis), and tuberculosis (at any anatomic site). In addition, the CEPAC-Paediatric model simulates three causes of mortality:

Children with no history of acute clinical event face a monthly risk of HIV-related

death ("chronic HIV mortality"), stratified by current age and CD4%.

Children who experience a clinical event face "acute mortality" risks in the 30 days

post-event; after this 30-day period, children with a prior event face increased monthly

risks of "chronic HIV mortality."

The model also includes a monthly risk of "non-HIV-related mortality," derived from

UNAIDS age- and sex-adjusted, country-specific mortality rates that exclude the

impact of HIV (for example, dying in a car crash).

The model also simulates the benefits and harms (toxicity, resistance) of ART. Simulated

patients start ART once they meet user-specified age, immunological, virological, and/or

clinical criteria. The model can incorporate up to 10 discrete ART regimens. Each ART

regimen is modelled to confer unique efficacies (probability of viral suppression), monthly

ART costs, and monthly gains in CD4% for children with suppressed HIV RNA. Children

who initially suppress their RNA at 24 weeks face a monthly risk of virological failure after

this time (“late failure”). The model can accommodate an independent benefit of ART on

mortality and OI risks, in addition to the effects of suppressive ART on CD4 and RNA. After

failed ART, CD4% remains stable for a user-specified amount of time (usually 12 months),

and then declines at the rate assigned for untreated children. The user assigns a monitoring

strategy (nature and frequency of laboratory and clinical assessment) and the criteria by which

ART failure is detected: virological (no RNA decrease to <400c/ml at 24 weeks),

immunological (decline to CD4% <10%), clinical (Stage 3/4 OI), or any combination. After

observed failure, patients can switch to the next available line of therapy. LTFU may occur at

user-specified monthly rates, at which time decline in CD4% and OI risk revert to off-ART

rates. Like the adult model, the CEPAC-Paediatric model tracks clinical events, changes in

CD4 and HIV RNA, time in each health state and healthcare costs associated with each state.

After an individual simulated patient has died, the next patient enters the model. Summary

statistics are tallied for the entire cohort and key clinical and economic outcomes are

described, as well as ICERs.

The CEPAC-Paediatric model has been internally validated by comparing model outputs to

existing data used to derive model inputs and to confirm the accuracy of the model structure

(180). In these simulations, clinical event and mortality data came from an international

consortium of AIDS care and treatment centres in the East African region (IeDEA East

Africa). The difference between model-generated results and observed data ranged from 2-

12% (Figure 14) which was below the 10-15% threshold accepted as “good fitting”(10). The

47

model was then calibrated to find the CD4% and age-stratified mortality risks necessary to fit

survival data from more than 1,300 untreated, perinatally HIV-infected children in six sub-

Saharan African countries (49). These calibrated input parameters are used for all policy

analyses.

Source: Ciaranello et al, PLosONE, in press (180)

Figure 14: Internal validation of clinical event risk outcomes: CEPAC-Paediatric model

results compared to IeDEA data. Thus, this validated paediatric CEPAC model will include primary data outcomes issued from

HIV-infected children in HIV-care in Côte d’Ivoire and gathered since 2004, the beginning of

the roll-out of ART access. My research work conducted there will precisely inform this

paediatric model.

15.88

28.92

61.54

7.34

19.85

66.16

17.93

32.83

67.89

8.23

21.41

67.53

0 20 40 60 80

TB

WHO Stage 4

WHO Stage 3

TB

WHO Stage 4

WHO Stage 3

Clinical event risk/100 PY

IeDEA

CEPAC- Pediatric Model

48

2. HYPOTHESES AND OBJECTIVES

49

2.1.Conceptual framework

The success of paediatric HIV care requires an upstream efficient health system that can offer

a comprehensive package of care, accessible and acceptable to the targeted population. Care

of HIV-infected children should theoretically start during the in utero period and should be

continued through to ART initiation and pursued during post-ART follow-up. Each step of

the cascade of care constitutes in itself a complex process that needs investigating in order to

identify the most effective and cost-effective care strategies.

Although we will not specifically address these in this PhD, specific indicators to measure the

field success of a care intervention need to be defined:

Access to care. This is a complex concept to define; Gulliford et al describe several

dimensions. First, having access to care denotes a potential to utilise a health service if

required. This means that there is an adequate supply of health services available.

Second, people in need may have access to these services, yet this access may not be

realised. This is influenced by the acceptability, affordability and accommodation of

services. Patients may not feel the need to seek care, or are discouraged by costs or

long waiting times. Finally, access to care is also defined as providing the right service

at the right time in the right place thus leading to the best possible outcome (181).

Efficiency. According to the Collins dictionary, efficiency is defined as “the

comparison of what is actually produced or performed with what can be achieved with

the same consumption of resources”. When measuring the efficiency of a care

programme, we must account for several indicators:

o Acceptability. Acceptability refers to the social and cultural perceptions of

health services, which are understood and therefore accepted and utilised by

potential patients.

o Efficacy. In medicine, efficacy indicates the capacity for a beneficial effect of

a given health intervention (drug, medical device, surgical procedure, or

a public health intervention). If efficacy is established, an intervention is likely

to be at least as good as other available interventions, to which it will have

been compared. Comparisons of this type are typically made in explicative

randomised controlled trials.

o Effectiveness. Effectiveness relates to how well an intervention works in in the

field conditions. Often, health services exist, but organisational barriers to

access (transportation costs, long queues…) may result in delays in treatment

and lead to worse clinical outcomes.

o Tolerance or Safety: the ability of a treatment or intervention health to not

affect or harm the health of the individual.

50

Impact. The WHO defines health impacts as « the overall effects, direct or indirect, of

a health policy, strategy, programme or project at the population level». Measuring

the impact of an intervention consists in comparing what actually happened with

would have happened if the intervention had not taken place. This must be

distinguished from outcome monitoring, which is examining whether targets have

been achieved.

The following conceptual framework (Figure 15) treats every step of the cascade of healthcare

conceivable for HIV-infected children but focuses primarily on identifying HIV-infected

children through diagnosis up to their enrolment and monitoring in appropriate HIV care. My

personal work is focused on the steps circled in red.

Figure 15: Conceptual framework of the care of paediatric HIV

51

2.2. Hypotheses

We supposed that the benefits of early ART have been proved throughout the CHER trial

(129) and suggested by many observational cohort studies mentioned in the first chapter of

this PhD. However, identifying and initiating ART in HIV-infected children early enough,

(i.e. before two years of age, according to the 2010 WHO recommendations) remains a

challenge, particularly in West Africa. New testing strategies and approaches are enabling

earlier diagnosis of HIV in children, and new, more affordable technologies for monitoring

children receiving ART are becoming available. However, many difficulties persist in the

implementation and adaptation of WHO guidelines to resource-limited settings. There is a

need to improve knowledge on access to care and the organisation of paediatric HIV care for

infected children, from PMTCT programmes through to antiretroviral treatment initiation and

follow-up. Furthermore, children who access care are those who have survived early infancy.

As children are identified at an earlier stage, access to healthcare will be greater as will be the

associated costs. This will have to be compared to a base case scenario of pre-ART treatment

costs and those encountered by children diagnosed at an advanced stage of the disease in

order to highlight the cost-effectiveness of PMTCT, EID and early ART in resource-limited

settings.

2.3. My PhD Objectives

Although disease models are computationally sophisticated, the quality of the results of

modelling studies depends highly on the quality and accuracy of the data on which they are

based, and on the conceptual soundness and validity of the models themselves. The main

objective of my research PhD work is to provide accurate and up to date data on the

care of HIV-infected children enrolled in care, either waiting for ART initiation or

already on ART in the West African context, mainly in Côte d’Ivoire.

Specific objectives are as follow:

- To document better the natural evolution of paediatric HIV disease in children who

have not yet initiated ART in the context of the roll-out of ART in Abidjan, Côte

d’Ivoire:

o To describe their mortality and severe morbidity;

o To measure their healthcare resource utilisation.

o To analyse respective associated factors

- To document better the course of paediatric HIV disease in children after ART

initiation in the same above context:

o To describe their mortality and severe morbidity and associated factors;

o To measure their healthcare resource utilisation and associated factors;

o To describe causes of hospitalisation in both treated and untreated HIV-

infected children;

o To investigate the 24-month immune response to ART and investigate the

effect of age at ART initiation on this response.

52

These data will directly inform the CEPAC-model and will be able to contribute to informing

healthcare providers and policy makers about many of the most critical current question in

paediatric HIV care.

53

3. SETTINGS

54

3.1. Implicated research infrastructures

3.1.1. The Bordeaux School of Public Health & the Inserm U897 research centre

“Epidemiology & Biostatistics”

The Bordeaux School of Public Health (Institut de Santé Publique, d’Epidémiolgie et de

Développement – ISPED) within Bordeaux Segalen University, is the first in France to be

closely connected to a medical faculty yet remain distinct. Created by Pr Roger Salamon and

currently directed by Pr Louis Rachid Salmi, the ISPED aims to meet contemporary public

health challenges: increasing life expectancy, the reform of health systems, the resurgence of

infectious diseases worldwide, the impact of industrialisation and globalisation on the

environment and population health. The ISPED’s main strength is the availability of a team of

highly qualified researchers, attached to the Inserm U897 research centre “Epidemiology and

Biostatistics”, and of hospital-based and academic research teams.

Created in 2008, the Inserm U897 research centre was directed by Pr Roger Salamon until

September 2013; Pr Christophe Tzourio has recently been named as his successor. Many

research teams within this centre have won national and international recognition in their

fields. They have led diversified projects covering public health issues of most concern, using

innovative methods in biostatistics, ageing and neurodegenerative diseases, HIV infection,

cancer, nutrition, environmental health and trauma prevention.

The research team in which I conducted this PhD was the “HIV, cancer and global health in

resource-limited settings” team, directed by Professor François Dabis. This team has been

conducting its research within the ISPED for the past 20 years.

Our main objective is to optimise the biomedical approach of HIV care and treatment of

adults and children in sub-Saharan Africa, to prevent the overall risk of mother-to child

transmission of HIV-1 and to apply expertise to other research themes, either for the control

of the HIV pandemic or for other public health priorities contributing to global health in sub-

Saharan Africa. This objective is built around four main research areas:

• Care of HIV-infected adults (led by Xavier Anglaret);

• PMTCT and maternal health (led by Renaud Becquet and François Dabis);

• Cancer prevention (led by Annie Sasco);

• Care of HIV-infected children (led by Valériane Leroy)

The scientific rationale for the “Care of HIV-infected children” axis is based on the scale-up

on ART in children in West Africa. The research conducted within this group is divided into

another two research areas:

55

• Clinical research around new strategies for optimising HIV care: how to identify the

earliest possible HIV-infected infants in order to enrol them into care (Pedi-Test

ANRS 12165 Study)? Which tools best diagnose routinely HIV-infected children in

Côte d’Ivoire (PediTest ANRS 12183 Study)? Can ART be simplified in HIV-

infected children whose treatment is initiated before the age of 2 years after virological

success at 12 months (Monod ANRS 12206 Trial)?

• Operational research based on observational cohorts from ART roll-out programmes

in West Africa supported by the team, and within the IeDEA project.

I conducted most of my PhD work hosted at the Inserm unit team since 2011 and was

supervised by Valériane Leroy.

3.1.2. The PAC-CI programme

The PAC-CI programme is a Franco-Ivorian research programme on HIV/AIDS based in

Abidjan, Côte d’Ivoire. This research programme was created in 1996, based on an agreement

between the Ministry of Public Health in Côte d’Ivoire and the French Agency for Research

on HIV/AIDS and Hepatitis (ANRS). The two main objectives in the creation of this research

programme were the training of healthcare workers for research activities around HIV/AIDS

and the implementation of medical research activities around the disease and associated

pathologies (tuberculosis, malaria or cancer).

The PAC-CI programme is currently led by Professor Thérèse Ndri-Yoman (South coordinator) and Dr Xavier Anglaret (North coordinator). The PAC-CI programme is above all a partnership between the Inserm U897 (University

Bordeaux Segalen), the department of virology at the Necker Hospital and many different

sites in Côte d’Ivoire:

University Hospital of Treichville (department of infectious diseases and the CEDRES

laboratory centre for HIV diagnosis and associated diseases)

University Hospital of Yopougon (department of paediatrics, general medicine and

gynaecology)

University hospital of Cocody (department of paediatrics, and medical school)

USAC (Unité de soins ambulatoires et de conseil) – a centre for HIV care, diagnosis

and counseling

CePReF (Centre de prise en charge et de formation) - a centre for research, training

and treatment of HIV in both adults and children

CIRBA (Centre intégré de recherches biocliniques à Abidjan) – a centre for biological

and clinical research

CNTS (Centre national de transfusion sanguine) – the national blood transfusion

centre.

Since 2008, the PAC-CI programme has also been collaborating with the CEPAC team in

Boston, presented subsequently.

56

The programme carries out research in three major fields: (i) care of HIV-infected adults, (ii)

reducing MTCT and (iii) care of HIV-infected children. Within these fields of research, there

have been many breakthroughs and new recommendations relative to operational research

have been issued consequently. Since it was created the PAC-CI programme has implemented

large care programmes in which thousands of HIV-patients are now followed-up.

The success of the PAC-CI programme has allowed it to apply now to become an Inserm

“International Associate Laboratory” and evolve towards an Ivorian centre of research

specialised in clinical and epidemiological research in infectious diseases.

Since 2010, I have been several times to Abidjan, to conduct my research work with the PAC-

CI team: in 2010 I stayed there for three months to conduct a study and collect data using a

retrospective data collection instrument I had previously designed and analyse data on

morbidity in HIV-infected children before their ART initiation from 2004 to 2009, presented

elsewhere (182, 183). I returned in 2012 to train and supervise an MD student to collect and

analyse the same data in children on ART during the same period (184).

3.1.3. The CEPAC group

The CEPAC team has been presented in the first chapter. They have been collaborating with

the PAC-CI programme since 2005, initially on adult HIV care and on paediatric care since

2009. The paediatric CEPAC model is under the auspices of Dr Andrea Ciaranello.

As a research fellow, I spent three months in 2012 at the Massachusetts General Hospital in

Boston to work on the use and the structure of the CEPAC paediatric model described in the

first chapter of this work. A manuscript describing the validation and calibration of this model

has recently been accepted for publication in PLosOne.

3.2. Data platforms

3.2.1. The International epidemiologic Databases to Evaluate AIDS (IeDEA),

West Africa The IeDEA network is an international research consortium established in 2005 by the

National Institute of Allergy and Infectious Diseases to provide a rich resource for globally

diverse HIV/AIDS data. Sites in various regions throughout the world collaborate to collect

and define key variables, harmonise data, and implement methodology to pool data

effectively as a cost-effective means of generating large datasets to address the high priority

research questions on HIV/AIDS. IeDEA collects HIV/AIDS data from seven international

regional data centres, including four in Africa (Central, East, Southern, and West Africa), and

one each in the Asia-Pacific region, the Central/South America/Caribbean region, and North

57

America. This type of data and resource pooling allows researchers to address unique and

evolving research questions that individual cohorts are unable to answer.

The WADA (West Africa Data Base on Antiretroviral Therapy) collaboration was launched

in July 2006 and renewed in 2011, its commitment to continue and expand the IeDEA

programme in the West African region, for five additional years. The WADA collaboration is

a consortium of leading clinicians and clinical epidemiologists. WADA is coordinated by the

WADA Coordinating Centre at the Bordeaux School of Public Health with a WADA

Regional Office hosted by the PAC-CI programme in Côte d’Ivoire. WADA is a unique

collaboration among clinical centres in West Africa with more than five years of experience.

Its mission is to conduct hypothesis-driven epidemiological research on the prognosis and

outcomes of HIV-1 and HIV-2-infected people, including adults, pregnant mothers, and

children in care and generally on treatment, as well as HIV-exposed children. The overall

aims of WADA are:

To measure the effectiveness of combination antiretroviral treatment in HIV-infected

individuals in West Africa in the long-term, and to assess factors that influence these

long-term outcomes;

To document the morbidity burden in HIV-infected individuals in care in HIV

programmes throughout West Africa.

WADA is constituted of five working groups: paediatrics, cancer, HIV-2, morbidity and

hepatitis. The principal aim of the paediatric working group (pWADA) is to document better

the operational access to HIV care and its long-term outcomes among HIV-exposed and HIV-

infected children. This is built around four specific aims:

Early infant diagnosis (EID) and early access to ART in HIV-infected children

Retention (to be conducted in consistency with the adult research agenda)

Documentation of long-term outcomes

Improving the care of HIV-infected adolescents on ART

The pWADA collaboration is constituted of 11 clinical sites in seven countries in West

Africa: Benin, Burkina Faso, Côte d’Ivoire, Ghana, Mali, Senegal and Togo. Nigeria will join

the paediatric group in 2014.

58

Figure 16: Map of West Africa

The participating clinical site in Benin is the Centre National Hospitalier Universitaire

(CNHU), situated in Cotonou, the largest city and economic capital of Benin. According to

WHO estimates, in 2011, 64,000 people were living with AIDS in Benin, of which 52% were

women. The prevalence of HIV among pregnant women was 1.7% and the overall PMTCT

coverage was 26.5% (185).

The pWADA site in Burkina Faso is located in Ouagadougou, the capital city, and is the

CHU Pédiatrique de Charles de Gaulle. In 2011, in Burkina Faso, 1,200,000 people were

living with HIV, of which a large proportion were women of a child-bearing age (47%). The

estimated prevalence of HIV among pregnant women in 2010 was 1.6% and 96% of

healthcare centres offered PMTCT services in 2012. However, in 2011, only 1,776 children of

the 23,000 living with HIV were on ART (186).

Côte d’Ivoire hosts five participating paediatric clinical sites, all of which are located in

Abidjan, the economic capital. We will further describe the epidemiology of the HIV

epidemic in Côte d’Ivoire in the next section. The MTCT-Plus programme and the Centre de

prise en charge, de recherche et de formation (CePReF-Enfants) are two of the participating

clinical sites which also contribute to other care programmes. The paediatric wards of the

University Hospitals of Yopougon and Cocody are two additional participating sites. The

Centre Intégré de Recherches Biocliniques d’Abidjan (CIRBA), also part of the PAC-CI

programme, is the fifth participating site.

The participating clinical site in Ghana is the Korle-Bu Teaching Hospital in Accra. It is the

premier healthcare facility in the country and the only tertiary hospital in the southern part of

Ghana. In 2011, according to WHO estimates, 230,000 adults were living with HIV of whom

55% were women. Ghana is one the 21 priority countries in the Global Plan to Eliminate

AIDS and within a very short time span it tripled its coverage of antiretroviral medicines for

pregnant women living with HIV, resulting in a 76% reduction in the number of new HIV

infections among children, the highest reduction observed among the countries with a high

burden of pregnant women living with HIV (27). However, paediatric services remain poor;

59

the percentage of infants born to HIV-positive women receiving a virological test for HIV

within 2 months of birth was 18% (187).

The paediatric ward of the University Hospital Gabriel Toure in Bamako is the pWADA

participating site in Mali. According to WHO estimates, in 2011, 110,000 people were living

with HIV yielding to an overall prevalence of 1.9%. Overall, 88% of pregnant women were

offered HIV-testing of which 35% accepted and 2.2% were positive. Among them, 96%

initiated ART to prevent MTCT (188).

The participating pWADA site in Senegal is the Albert Royer Children’s Hospital in Dakar.

In Senegal, the prevalence of HIV is one of the lowest in West Africa, estimated to be 0.7% in

2011. The ratio of infected women over men was 1.6 which underlines the feminisation of the

epidemic. Approximately 41% of pregnant women accessed an HIV diagnosis and 23%

received ART for PMTCT. The estimated MTCT rate was 5.1% in 2011 (189).

The last participating site is the paediatric ward of the Tokoin University Hospital, based in

Lomé, the capital city of Togo. The overall prevalence of HIV in Togo is 3.4% and is

estimated to be 3.5% among pregnant women. In 2011, PMTCT coverage among HIV-

infected pregnant women was 60.5% and the MTCT rate was estimated to be 6.5% (190).

3.2.2. Côte d’Ivoire

3.2.2.1. Geography of Côte d’Ivoire

Côte d’Ivoire has an area of 322,462 square kilometres and borders Liberia, Guinea, Mali,

Burkina Faso and Ghana. Its southern boundary is along the Gulf of Guinea. The political

capital of Côte d’Ivoire is Yamoussoukro, situated in the centre of the country, 300 kilometres

from Abidjan, the economic capital city. The country has the largest economy in the West

African Economic and Monetary Union. Côte d’Ivoire is the world's largest exporter of cocoa,

and the fourth largest exporter of goods, in general, in sub-Saharan Africa. Côte d’Ivoire has,

for the region, a relatively high GDP per capita of $1,800 in 2012, and plays a key role in

transit trade for neighbouring, landlocked countries. Health expenditures represent 5.3% of

GDP (191). In 2013, it was estimated that the population of Côte d’Ivoire reached 22,400,835;

approximately 40% were children aged 0-14 years and the total fertility rate was 3.73 children

born per woman. Overall maternal and infant mortality remains high reaching 400 deaths per

100,000 live births and 61.66 deaths per 1,000 live births, respectively (191).

3.2.2.2. The HIV pandemic in Côte d’Ivoire

Côte d’Ivoire is one of the countries the most affected by the HIV/AIDS epidemic in West

Africa. In 2009, the prevalence of HIV among adults was estimated to be 3.4%, ranking Côte

d’Ivoire as the 19th most HIV-prevalent country in the world (191). In 2011, the WHO

estimated there to be 360,000 people living with HIV, of whom 17% were children and 47%

women. With two women infected for one man, the feminisation of the infection is to be

60

noted (6.4% for women against 2.9% for men). Among young pregnant women (15-24 years)

living in Abidjan, HIV prevalence is estimated at 5.2% (35). Furthermore, Côte d’Ivoire is

one of the 21 priority countries in the Global Plan towards the elimination of new infections;

between 2009 and 2012, women’s access to ART increased by 57%. However, the decline in

new infections is one of the slowest, reaching only 25% between 2009 and 2012. The

moderate levels of coverage of antiretroviral medicines may not provide adequate protection

during breastfeeding, keeping HIV transmission rates high ; in 2012 the MTCT rate was 25%

(26).

Côte d’Ivoire bears a heavy burden in terms of death related to HIV/AIDS. In 2011, 23,000

people died of HIV/AIDS; 17% of pregnancy-related deaths were attributable to HIV/AIDS,

as was 4.4% of infant mortality (26, 191). According to the UNICEF, 540,000 Ivorian

children are orphaned and vulnerable due to HIV/AIDS, 40,000 currently live with HIV, and

over 16,000 are infected each year. Despite this alarming situation, only 4,000 children are

under follow-up for HIV/AIDS and 2,000 benefit from ART (35).

Côte d’Ivoire is also marked by a renewed outbreak of tuberculosis, the most frequent

opportunistic infection in HIV-infected patients and the first cause of death (32%). In addition

to tuberculosis, hepatitis, malaria and cervical cancer add to the burden of the disease (192).

3.2.2.3. The National response to HIV in Côte d’Ivoire

For the past decade, the Ivoirian government has made substantial efforts in its fight against

HIV/AIDS. This initially took the form of the creation of many institutions, including a

Ministry of Fight Against AIDS in 2000 and the national programme for people living with

HIV in 2001, the PNPEC.

The national response to HIV relies essentially on a decentralisation and multi-sectorial

strategy. Committees for the fight against HIV/AIDS have been created on regional and

communal levels, allowing the coordination of prevention and care interventions all over the

country. Almost all areas of activity are prompted to take part in and organise this national

response as HIV is now considered a major developmental challenge and security problem.

Throughout national and international protocols, and with the help of many key players, the

country has been able to provide and improve services for the care of HIV/AIDS. Despite the

lack of human resources and infrastructures, with only 0.14 physicians per 1,000 population

and 0.4 hospital beds per 1,000 population (191), efforts in the response to HIV/AIDS are

noticeable: since 2008, access to ART is free for all patients, as are cotrimoxazole

prophylaxis, PMTCT services and counselling. After the post-electoral crisis in 2011, DBS

testing at 6 weeks of HIV-exposed children was implemented at the national level, after an

initial assessment was performed (193).

61

3.2.2.4. Presentation of ACONDA

ACONDA-VS-CI is a non-governmental association comprised of health professionals

specialised in the care of HIV-infected patients and has been advocating for ART access since

2001 in Côte d’Ivoire. In June 2004, ACONDA, in partnership with the Bordeaux School of

Public Health, launched a five-year programme of access to HIV care and treatment. This

programme was funded by the United States President’s Emergency Plan for AIDS Relief

(PEPFAR), through the Elizabeth Glaser Paediatric AIDS Foundation (EGPAF, Washington

DC, USA) (194) and the French state-run entity, ESTHER. Since 2004, ACONDA is one of

the three existing HIV-implementers in Côte d’Ivoire with EGPAF and the International

Centre of AIDS Care and Treatment Programmes (I-CAP) at Columbia University’s Mailman

School of Public Health.

The ACONDA programme was implemented with the main objective of favouring clinical

research and care interventions for the HIV-infected, taking a comprehensive family-based

approach to care. The association has set up different health facilities in the poorest areas of

Abidjan, such as Yopougon or Abobo, where different clinical trials and healthcare

programmes are led.

Within the programme, healthcare workers are trained to offer HIV testing to every pregnant

woman attending antenatal clinics and to encourage HIV-infected mothers to bring their

children and partners with them for testing and counselling. Each pregnant woman with an

HIV infection is immediately referred for an adult consultation. She then receives PMTCT

therapy, either a short regimen or ART depending on her clinical and immunological status,

and undergoes assessment at both antenatal and adult clinics. Children aged 15 years enter

the Aconda programme in one of two ways:

After referral for HIV testing at the age of 6 weeks or more, because their mother has

been diagnosed with HIV infection and has received ART.

After HIV testing at a paediatric clinic following presentation with AIDS-related

symptoms, even if they have not been previously identified as HIV-exposed or if their

parents do not participate in the programme.

The interventions led by the paediatric services within the Aconda programme are focused on

family care and cover four main components: psychological care (discussion groups, school

support, psychiatric support…), clinical care (diagnosis and treatment including ART and

prophylaxis), prevention (HIV screening) and clinical research (clinical trials and cohort

studies).

For paediatric care, the Aconda programme relies on a number of public and private

healthcare facilities. In these centres, the programme trains personnel in standardized

procedures of HIV care and treatment, and implements computerised data management

systems. In addition to these centres, Aconda relies on two sites entirely dedicated to HIV

62

care in Abidjan: the CePReF and the MTCT-Plus programme. The first paediatric HIV care

outcomes reported after the scaling up of ART access in children in Côte d’Ivoire were issued

from the paediatric Aconda programme (119).

3.2.2.5. The Centre de Prise en Charge, de Recherche, et de Formation

(CePReF)

The CePReF is Aconda’s largest health facility entirely dedicated to HIV care. It is situated in

the district of Yopougon-Attié in North West Abidjan. The centre is focused on three areas of

intervention: the care of HIV-infected adults, the care of HIV-infected children and the

training of health professionals. This centre follows up one of the biggest active paediatric

ART files in Abidjan.

3.2.2.6. The MTCT – Plus programme: Mother to Child Transmission

Prevention programme

The MTCT-Plus programme was coordinated by the ACONDA-VS-CI organisation and was

implemented in two health facilities in the north of Abidjan; Abobo-Avocatier and Yopougon

Niangon-Sud. The MTCT-Plus initiative aimed to enrol HIV-infected pregnant and

postpartum women in comprehensive HIV care and treatment for themselves and their

families and to promote an early and clinical care and follow-up of children born to HIV-

infected mothers, starting from pregnancy. The programme offered care to the whole family

of the HIV-positive patient.

63

4. WHAT WAS THE NATURAL

EVOLUTION OF PAEDIATRIC HIV

BEFORE ART INITIATION DURING

THE ROLL-OUT OF ART IN ABIDJAN,

COTE D’IVOIRE?

Contribution to model inputs

64

4.1.Severe morbidity and mortality in untreated HIV-infected children

in a paediatric care programme in Abidjan, Côte d’Ivoire, 2004-

2009

4.1.1. Research question and background

The paediatric HIV epidemic is fuelled by the high prevalence of the disease among women

of a child-bearing age, who, despite PMTCT interventions, continue to transmit the virus to

their children. The coverage of early infant diagnosis and access to care remained insufficient

in Côte d’Ivoire in 2013 leading to early severe morbidity and mortality in HIV-infected

children. It is crucial to improve the neglected management of HIV-infected children to

reduce paediatric mortality in Africa. Indeed, child mortality related to HIV/AIDS occurs

early in Africa because of the rapid development of disease manifestations within the first two

years of life (46). Mortality rates were estimated to be 35.2% by 12 months and 52.5% by 24

months of age among HIV-infected children (48).

Few data are available documenting the natural evolution of HIV-related severe morbidity

and mortality in children who have not yet initiated ART in the Ivoirian context. Among 98

perinatally HIV-infected children followed-up in Abidjan until two years of age, infant

morbidity and mortality before access to ART remained high despite systematic

cotrimoxazole prophylaxis (195). In fact, these data are needed to assess comparatively the

effect of different ART strategies implemented in paediatric care since the roll-out of

paediatric ART in 2004 in Abidjan. Therefore, detailed and accurate data on the clinical

evolution of HIV-infected children not yet on ART are essential in order to evaluate the effect

of HIV care and treatment now being provided and to measure its effectiveness in the field as

ART is being scaled up. These data will be useful input contributing to the paediatric CEPAC

model.

The main objective of this study was thus to describe the clinical evolution, assessed by

severe morbidity (defined as WHO Stage 3 or 4 event or any event leading to either

hospitalisation or death) and mortality and their determinants, in HIV-infected children from

their entry in the Aconda care programme until their ART initiation.

We conducted a retrospective study within a prospective cohort formed by HIV-infected

children who had not yet undergone ART followed up in the two model centres of the Aconda

programme in Abidjan (CePReF and MTCT-Plus programme), between 2004 and 2009.

4.1.2. Desmonde et al – BMC Infectious Diseases – November 2011

The results of this study were presented at the 18th Conference on Retroviruses and

Opportunistic Infections in 2011, and were subsequently published in BMC Infectious

Diseases.

RESEARCH ARTICLE Open Access

Severe morbidity and mortality in untreated HIV-infected children in a paediatric care programmein Abidjan, Côte d’Ivoire, 2004-2009Sophie Desmonde1, Patrick Coffie2, Edmond Aka3, Clarisse Amani-Bosse4, Eugène Messou3, François Dabis1,

Ahmadou Alioum1, Andrea Ciaranello5 and Valériane Leroy1*

Abstract

Background: Clinical evolution of HIV-infected children who have not yet initiated antiretroviral treatment (ART) is

poorly understood in Africa. We describe severe morbidity and mortality of untreated HIV-infected children.

Methods: All HIV-infected children enrolled from 2004-2009 in a prospective HIV programme in two health

facilities in Abidjan, Côte d’Ivoire, were eligible from their time of inclusion. Risks of severe morbidity (the first

clinical event leading to death or hospitalisation) and mortality were documented retrospectively and estimated

using cumulative incidence functions. Associations with baseline characteristics were assessed by competing risk

regression models between outcomes and antiretroviral initiation.

Results: 405 children were included at a median age of 4.5 years; at baseline, 66.9% were receiving cotrimoxazole

prophylaxis, and 27.7% met the 2006 WHO criteria for immunodeficiency by age. The risk of developing a severe

morbid event was 14% (95%CI: 10.7 - 17.8) at 18 months; this risk was lower in children previously exposed to any

prevention of mother-to-child-transmission (PMTCT) intervention (adjusted subdistribution hazard ratio [sHR]: 0.16,

95% CI: 0.04 - 0.71) versus those without known exposure. Cumulative mortality reached 5.5% (95%CI: 3.5 - 8.1) at

18 months. Mortality was associated with immunodeficiency (sHR: 6.02, 95% CI: 1.28-28.42).

Conclusions: Having benefited from early access to care minimizes the severe morbidity risk for children who

acquire HIV. Despite the receipt of cotrimoxazole prophylaxis, the risk of severe morbidity and mortality remains

high in untreated HIV-infected children. Such evidence adds arguments to promote earlier access to ART in HIV-

infected children in Africa and improve care interventions in a context where treatment is still not available to all.

BackgroundIn 2008, UNAIDS reported 14% of new human immu-

nodeficiency virus (HIV) infections occurring in children

below 15 years of age throughout the world [1]. By the

end of 2008, it was estimated that 91% of HIV-infected

children lived in sub-Saharan Africa [1]. Before ART

was introduced, infant mortality of HIV-infected chil-

dren on the African continent was high and premature,

reaching 35% among children less than 12 months of

age and 52% among children aged less than 24 months

[2]. While the rate of new paediatric infections has

slowly declined, only 38% of infected children have

access to antiretroviral therapy (ART), with a huge het-

erogeneity between world regions [3]. In Côte d’Ivoire,

the paediatric pandemic is sustained by a high preva-

lence of HIV among pregnant women (6.4%) [1], who,

despite increasing access to mother-to-child transmis-

sion prevention (PMTCT) programmes, transmit the

disease to their children [4].

In addition, providing a continuum of care between

postnatal HIV diagnosis and paediatric care remains

challenging, especially in Côte d’Ivoire [3-6], and the

uptake of early infant diagnosis remains fairly poor.

Thus, children are diagnosed belatedly, and when the

HIV diagnosis is eventually made, the antiretroviral

needs are not sufficiently met, estimated to reach only

* Correspondence: [email protected], U897 & Institut de Santé Publique, Epidémiologie et

Développement (ISPED), Université Victor Segalen Bordeaux 2, Bordeaux,

France

Full list of author information is available at the end of the article

Desmonde et al. BMC Infectious Diseases 2011, 11:182

http://www.biomedcentral.com/1471-2334/11/182

© 2011 Desmonde et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the CreativeCommons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, andreproduction in any medium, provided the original work is properly cited.

15% in Côte d’Ivoire [1], leading to a severe morbidity

and premature mortality before any access to ART

[7-10].

In Africa, data on the disease progression of HIV-

infected children before access to ART are rare, frag-

mented and most often obtained from selected children

who have survived early opportunistic pathologies and

who are therefore selected for these reasons [7-14]. In

addition, the scarce existing data were collected during

time periods when ART access was limited or nonexis-

tent, when practices of care may have differed from the

current standards of care in Côte d’Ivoire[11].

Detailed and accurate data on clinical evolution of

HIV-infected children not on ART are essential in order

to evaluate the effect of HIV care and treatment now

being provided and to measure its effectiveness in the

field as ART access is being scaled up [15]. In this

paper, we describe the clinical evolution of HIV-infected

children who have not yet initiated ART, assessed by

severe morbidity and mortality and its determinants

before access to ART, in a paediatric HIV-care pro-

gramme in Abidjan, Côte d’Ivoire.

MethodsStudy design and participants

In 2001, the Aconda Programme, a non-governmental

association comprised mainly of health professionals,

was implemented in Côte d’Ivoire. The main objective

of this programme is to provide HIV and antiretroviral

care in HIV-infected patients (adults and children)

[3,16]. In 2004, Aconda, in partnership with the Institute

of Public Health, Epidemiology and Development

(ISPED, Bordeaux, France), launched a five-year pro-

gramme of access to HIV care and treatment. ART was

delivered according the 2006 WHO guidelines. The pae-

diatric Aconda programme relies mostly on the two fol-

lowing HIV care centres: the paediatric CePReF site

(Centre de Prise en charge, de Recherche et de Forma-

tion) which follows-up one of the largest active paedia-

tric ART cohorts in Abidjan [3] and the MTCT-Plus

programme, entirely dedicated to paediatric care follow-

ing the enrolment of HIV-infected pregnant and post-

partum women. Children included in the latter are, as a

consequence, diagnosed earlier than in other sites [17].

Our work was a retrospective study conducted within

the prospective assembly of a cohort of HIV-infected

children who had not yet initiated ART and who were

being followed-up in one of the two model paediatric

health facilities in the Aconda care programme in Abid-

jan. All children under 15 years of age with a medical

record who had not yet initiated any form of ART other

than PMTCT interventions and who were enrolled in

the programme between 1st January 2004 and 31st

December 2009 after a confirmed HIV diagnosis by

positive PCR at any age - or by positive serology if aged

over 18 months - were included in this study.

Data collection

The data collection took place retrospectively, from the

CePReF and MTCT pre-existing databases. However,

the CePReF database was not thorough and most of the

data were extracted from the medical files. A standar-

dised data collection instrument was issued for this pur-

pose. The following variables were recorded into a

database: date of inclusion, sex, date of birth, parental

HIV status, parental vital status, date and method of

diagnosis, HIV-diagnosis, PMTCT history, feeding his-

tory; date, height and weight; dates and causes of hospi-

talisation; dates and reason for treatment for in-patient

daycare; dates and AIDS-defining morbid events; dates

and receipts for cotrimoxazole, date and motif for can-

cellation of the prophylaxis; date of ART initiation;

dates and results of haematology and CD4 cell count

and percentage; and current status (transferred, LTFU,

on ART, deceased or still known to be alive and not

treated by ART).

Event classification during follow-up

Data regarding all clinical examinations, in-patient day-

care, and hospitalisations were extracted from the

patients’ medical records. The morbid events were

defined and classified in the following ways: cases of

respiratory tract infections, malnutrition, diarrhoea,

tuberculosis, chronic HIV-associated lung disease, anae-

mia, encephalopathy and other AIDS-defining events

were defined according to the WHO case definitions of

HIV surveillance [18]. Other frequent paediatric infec-

tious diseases such as varicella-zoster infections,

measles, hepatitis B and malaria were defined according

to the Red Book of Paediatrics [19]. Clinical events

belonging to stages 1&2 of the WHO classification were

not considered, neither were traumatic injuries. We

defined any morbid event as severe if it led to one of

the following: hospitalisation, in-patient daycare for

more than 48 hours, or death.

Statistical analysis

Baseline categorical data are presented as frequencies

(%), and continuous data as median (interquartile range,

(IQR)). Baseline differences between age groups and

immunodeficiency status, according to the 2006 WHO

recommendations (<15% if aged < 5 years; <200/mm3 if

aged ≥ 5 years) [19], were compared with Chi-2 tests,

Fisher’s exact tests, t-tests or Wilcoxon’s tests where

appropriate. We used the 2006 guideline definitions to

be consistent with the study period, 2004-2009. The fol-

low-up period was defined as the time between enrol-

ment in the Aconda programme (time origin) and the



Page 2 of 12

date of ART initiation or the closeout date (date of

death or date of last known to be alive). Two outcomes

were assessed during this follow-up period: 1) the first

severe morbid event and 2) mortality. The distributions

of time to these outcomes were estimated using compet-

ing risk survival analysis. ART initiation is a competing

cause of failure since children under ART have a better

chance of survival [20-22]. To avoid making unrealistic

assumptions about the independence of the outcomes,

we chose to use a cumulative incidence function (CIF)

to estimate the probabilities for each [23]. We ran a

multivariate competing risks analysis, performed by the

Fine and Gray proportional subdistribution hazard

regression model. We used the package cmprsk of R sta-

tistical software version 2.11.1 (The R foundation for

Statistical computing, Vienna, Austria) following guide-

lines presented by Scrucca et al [24,25]. If patients were

lost to follow-up, data were right-censored at the date

on which patients were last known to be alive (a patient

was considered lost to follow-up after a 6 month period

since the last clinical encounter).

Two multivariate models were estimated. In addition

to age, immunodeficiency status, and gender, the other

variables were inserted in the first “full” multivariate

model if their p-value was less than 0.25. A second,

“reduced” multivariate model was obtained from these

same variables using descending stepwise procedures.

The adjusted subdistribution hazard ratios were

reported with their 95% confidence intervals (95%CI).

Variables for which fewer than 70% of the patients pro-

vided data were not included. A p-value less than 0.05

was considered statistically significant on multivariate

analysis in both “full” and “reduced” models.

Ethical aspects

The national ethics committee of Côte d’Ivoire approved

the Aconda programme data management system

ResultsOverall, 462 children fulfilled the inclusion criteria (Fig-

ure 1). Of these children, 57 (12.3%) were excluded

because no medical record could be located, leaving 405

children included in the analysis.

Characteristics of patients at baseline

The baseline characteristics of the 405 included children

are presented in Table 1. Children were diagnosed at a

median age of 4.5 years (IQR: 1.9; 7.5), and 47% were

females. The median age of enrolment was lower

amongst the children followed up in the MTCT pro-

gramme (1.7 years (IQR: 0.1;4.9) vs 5.4 years

(IQR:2.7;8.1), p <0.0001). Cotrimoxazole prophylaxis had

been prescribed to 67% of the children before enrolment

in the programme. The median weight for age z-score

(WAZ) was -2.09 (IQR: -3.45;-0.95) among the children

for whom anthropometric data were available. Approxi-

mately 50% of these children were suffering from failure

to thrive (FTT), defined as a WAZ < -2. Immunological

data (CD4 percent or count) were available for 308 chil-

dren (76.1%). Of these children, 27.7% met the 2006

WHO criteria for immunodeficiency by age. The pro-

portion of immunosuppressed children was highest

amongst the 10-15 year olds (55%, p <0.0001).

Patient follow-up

Among the 405 children included, the overall median

pre-ART follow-up time was 12 months (IQR: 1.3;30.6).

Over the study period, 240 (59.1%) initiated ART, 21

(5.2%) died, 41 (10%) were lost to follow-up, 11 (2.7%)

transferred to other healthcare facilities and 92 (23%)

were still known to be alive and untreated at the end-

point date (Figure 1).

Morbidity

Forty-two percent (170 children) presented with at least

one HIV-related condition during follow-up. Fifty-six

children (14%) presented with a first severe morbid

event during the first 18 months of follow-up, most fre-

quently malaria (37.5%) and WHO stage 3 events

(28.3%). Of these severe morbid events, 16 (28.6%) led

to death (Table 2). The observed incidence density rate

of occurrence of the first severe event was 17.8 per 100

child-years of follow-up. From the date of enrolment in

the programme, the overall median time of infection

until occurrence of the first severe event was 9.1 months

(IQR: 1.2;26.3). The probability of developing a first

severe morbid event at 18 months was high, with an

overall cumulative incidence of 11% (95%CI: 10.9;11.1).

During these first 18 months of follow-up (Figure 2),

the probability did not differ significantly by age at

enrolment nor by immune status (p = 0.51 and p = 0.87

respectively). Univariate analysis show that children in

the MTCT programme were less at risk of a first severe

event occurring than those in the CePReF programme

(sHR: 0.43, 95%CI:0.19;1), (Table 3). However, this asso-

ciation did not remain in the full multivariate model

(Table 3), due to correlation between the age and pro-

gramme variables (children in the MTCT programme

were younger). Associations between baseline character-

istics and the occurrence of the first severe event were

not seen in the full model, but the reduced model sug-

gested a lower risk in children ≥5 years of age at enrol-

ment compared to those <5 years of age (5-10 years,

sHR: 0.30, 95%CI: 0.10;0.97 and 10-15 years, SHR: 0.15,

95%CI:0.03;0.66). This association was confounded by

PMTCT interventions and immunodeficiency status,

however. The reduced model also showed a lower risk

in children previously exposed to any kind of PMTCT



Page 3 of 12

intervention (sHR: 0.16, 95%CI: 0.04;0.71, versus those

not exposed).

Mortality

By the end of the study, 21 deaths were recorded, 90.5%

of which occurred at home; the median time to death

was 12 months (IQR:1.5;30.8) and the incidence density

rate was estimated as 3.26 per 100 child-years of follow-

up (95%CI: 3.12;3.39). Fifty-two percent of the deaths

occurred in children aged less than 2 years. The

mortality rates for children aged 0-1 and 1-2 years were

9.70 and 3.41 per 100 child-years of follow-up, respec-

tively (Table 1).

The overall cumulative mortality reached 5.5% (95%CI:

5.3;5.7) at 18 months; reaching 14% at 18 months of fol-

low-up among children less than 12 months (p = 0.023)

(Figure 3A). Immunosuppressed children also had a

higher probability of death at 18 months compared to

non-immunosuppressed children (4.5%, 95%CI: 1.8;9.7,

vs 0.6%; 95%CI: 0.2;4.0). In addition, the absence of

INELIGIBLE

ART initiated before 1st January 2004

(n=143)

Unknwon diagnosis date (n=48)

Diagnosed after 1st December 2009

(n=72)

Infected with HIV-2 alone (n=20)

INELIGIBLE

ART initiated before 1st January 2004

(n=25)

ELIGIBLE CHILDREN (n= 462)

CePReF (n=370)

MTCT (n=92)

EXCLUDED

Missing medical record (n = 57)

INCLUDED IN THE STUDY (n= 405)

CePReF (N=313)

MTCT (N=92)

Initiated ART (n=186) Initiated ART (n=54)

Lost to follow-up (n=34) *

Deaths (n=15)

Lost to follow-up (n=7)*

Deaths (n=6)

Transferred out (n=11)

HIV-infected children in the

CePReF database

n= 653

HIV-infected children in the

MTCT–Plus database

n= 117

Alive and untreated (n=67) Alive and untreated (n=25)

CePReF clinic MTCT – Plus clinic

Figure 1 Cohort profile of children included in the paediatric Aconda programme. Children followed up at the CePReF and in the MTCT-

Plus programme, Abidjan, Côte d’Ivoire (January 2004 - December 2009) *Missing data = date of HIV diagnosis unknown or missing medical

record; †Lost to Follow-up: last clinical contact > 6 months.



Page 4 of 12

immunological data was informative; children with

unknown immunologic status had the highest risk of

death by 18 months: 13.6% (95%CI: 7.7;21.4; Figure 3B).

Adjusted for gender and immune status, age was no

longer associated with death (p = 0.07), but immune-

supression remained a significant predictor of death

(sHR: 6.02, 95% CI: 1.28;28.42), (Table 4).

DiscussionThis cohort study from Abidjan, Côte d’Ivoire, docu-

ments patterns of severe morbidity and mortality in

HIV-1 infected children before the initiation of ART, in

a paediatric care programme in which ART access was

being scaled-up. There are two primary findings. First,

the risk of severe morbidity before ART initiation is

high, reaching 11% at 18 months. The risk of first ser-

ious morbidity was minimised in older children but also

by having received any intervention for PMTCT. Sec-

ond, the risk of mortality before ART initiation is high

(reaching 8% at 18 months), and this risk is greater in

immunosuppressed children and children with

unavailable immunological data. Importantly, the results

from this cohort study may underestimate the true risks

of morbidity and mortality, as they were derived from a

selected population included in research programmes,

where healthcare support might have been better than

that offered outside of the Aconda context.

In addition, our results highlight the impact of age at

enrolment. The overall incidences of severe morbidity

after 18 months of follow-up do not vary by immune

status but by age, which is consistent with a rapid CD4

evolution in the younger children. As the children fol-

lowed up in the MTCT programme are significantly

younger, there is a strong correlation between age and

follow-up centre; we therefore chose, in the reduced

model, to exclude the centre variable from the analyses.

Adjusting for PMTCT, immunosupression and gender

had an effect on the age variable in the final reduced

model; children aged ≥5 were less at risk of presenting

with a first morbid event. Exposure to PMTCT inter-

ventions was more frequent among the younger children

and therefore the two variables were confounded. Based

Table 1 Characteristics by age in untreated HIV-infected children

Overall Age < 1 year Age 1 - 2years

Age 2 - 5years

Age 5 - 10years

Age 10 - 15years

BASELINE N = 405 N = 64 N = 42 N = 108 N = 142 N = 49

Age, years, median (IQR) 4.5 (1.9; 7.5) 0.3 (0.01; 0.6) 1.7 (1.6; 1.8) 3.4 (2.6; 4.1) 7 (5.9; 8.5) 12.1 (11.2;13.1)

Female, n (%) 189 (46.7) 30 (46.9) 19 (45.2) 47 (43.5) 75 (52.8) 18 (36.7)

Weight available, n (%) 383 (94.6) 59 (92.2) 37 (88.1) 100 (92.6) 138 (97.2) 49 (100.0)

Height available, n (%) 369 (91.1) 59 (92.2) 36 (85.7) 95 -88 131 (92.3) 48 (97.9)

Weight for age z-score, median(IQR)

-2.09 (-3.45;-0.95)

-1.94

(-3.01;-0.91)

-2.67 (-4.10;-1.60)

-1.93 (-3.36;-0.74)

-1.84 (-3.22;-0.75)

-2.58 (-4.05;-1.62)

Height for age z-score, median(IQR)

-1,59 (-2;69;-0.76)

- -1.34 (-1.47;-0.93)

-1.79 (-3.01;-0.74)

-1.32 (-2.44;-0.72)

-1.88 (-2.79;-1.06)

Weight for height z-score,median (IQR)

-1,11 (-2.63;-0.21)

- -1.33 (-1.32;-0.38)

-0.86 (-2.37;0.06)

-1.38 (-3.57;-0.43)

-1.67 (-1.85;-1.49)

Failure to thrive, n (%) 198 (51.7) 29 (49.2) 26 (70.3) 48 -48 63 (45.7) 32 (65.3)

CD4 count measurement available,n (%)

308 (76.1) 36 (56.3) 36 (85.7) 92 (85.2) 106 (74.6) 38 (94)

CD4 cell %, median (IQR) 19 (14; 25.5) 18.3 (14.1; 26) 17.9 (11.7;23.3)

20.7 (14; 26.2) - -

CD4 count/μL, median (IQR) 403 (151; 665) - - - 507.5 (234; 723) 213.5 (55; 413)

Immunodepressed* children*, n(%)

112 (27.7) 10 (15.6) 10 (15.6) 26 (24.1) 39 (27.5) 27 (55.1)

CDC stage C, n (%) 50 (12.3) 7 (10.9) 6 (14.3) 12 (11.1) 14 (9.9) 11 (2.1)

Baseline cotrimoxazole prophylaxisn (%)

271 (66.9) 23 (35.9) 33 (78.6) 74 (68.5) 102 (71.8) 39 (67.4)

Born to known HIV-infectedmothers

273 (67.4) 64 (100) 33 (78.6) 86 (76.6) 77 (54.2) 13 (26.5)

Diagnosis by PCR before 18 months 71 (19.2) 53 (82.8) 6 (14.3) 5 (4.6) 6 (4.2) 1 (2.1)

Mother deceased at inclusion, n (%) 97 (30.9) 1 (1.6) 1 (2.4) 15 (13.9) 52 (36.6) 28 (57.1)

Breastfeeding history, n (%) 191 (47.2) 40 (62.5) 19 (45.2) 43 (39.9) 67 (47.2) 22 (44.9)

History of PMTCT intervention 65 (18.1) 50 (78.1) 10 (23.8) 5 (4.6) - -

Children followed-up by the CePReF and the MTCT-Plus programme, Abidjan, Côte d’Ivoire (January-2004-December 2009). n = 405. * Immunodeficiency

according to the 2006 criteria;



Page 5 of 12

on previous studies reporting PMTCT interventions as

having a protective effect on severe morbidity [26,27],

we expected to observe a lower risk of severe morbidity

among the younger children. However, our study shows

higher risks among these children. We hypothesise that

the outcomes of PMTCT interventions are not the only

explanation for the observed results, but the available

data does not allow further conclusions; PMTCT is a

marker of early and different access to care for younger

children than for older children. We show here the posi-

tive effect of early access to care on reducing the risk of

occurrence of the first severe HIV-related morbid event.

These results highlight the importance of comprehen-

sive HIV care in addition to ART. Despite prophylaxis

by cotrimoxazole at baseline to prevent opportunistic

infections [28], substantial morbidity attributable to

other co-infectious disease (bacterial, viral and parasitic

infections) was observed. Optimal prevention and care

for comorbidities needs to be investigated before and

after ART access to improve paediatric care.

Mortality was moderate in this cohort, reaching 8% at

18 months. Although mortality risk was significantly

higher among infants < 12 months than among older

infants in univariate analyses, we found no overall signif-

icant effect of age in multivariate analyses. There was,

however, a 12-fold increase in mortality risk for children

who had no CD4 measurement. Infants aged < 12

months have previously been reported to have higher

risk for death and loss to follow-up than older children

[6]; among infants <12 months old in our cohort, 50%

had no available CD4% data at baseline. This correlation

between young age and unavailable CD4% partially

explains the observed increased risk of mortality in chil-

dren without CD4% data. The reasons for this lack of

CD4% availability in younger children are also critical.

Children diagnosed as HIV-infected and then enrolled

in the clinic often have to return on a different date for

CD4 measurement if no staff members are available for

phlebotomy at the initial visit. While most of these chil-

dren do return, many others are lost to follow-up. Rea-

sons for loss-to-follow-up remain a primary research

question in paediatric care, but in light of these results,

we can hypothesise that many deaths occurring at home

during the interval before the scheduled return visit, and

then are reported to the clinic at a later date when con-

tacted by a care worker. This demonstrates the critical

need for improved and expedited linkage to care

between diagnosis, CD4% measurement, and ART

initiation.

There are two important limitations inherent in our

study design. First, the timing of paediatric HIV infec-

tion could not be known with complete accuracy. In

Abidjan, like in most resource limited settings, social

barriers and stigma associated with HIV/AIDS often

lead to delays in accessing care and in receipt of HIV

diagnosis. The youngest children in our study were

Table 2 Distribution of the first severe morbid event in untreated HIV-infected children

Diagnosis Number of first events, n(%) Lethal events, n(%)

WHO Stage 2

Hepatosplenomegaly 1 (1,7) -

WHO Stage 3

Failure to thrive (FTT) 2 (3,6) 1 (50)

Unexplained persistent diarrhoea 4 (7,1) -

Unexplained persistent fever 1 (1,7) -

Tuberculosis 1 (1,7) 1 (100)

Chronic lung disease 4 (7,1) -

Severe bacterial infections 4 (7,1) 2 (50)

WHO Stage 4

Pneumonia 1 (1,7) -

Other diagnosis of interest

Malaria 21 (37,5) -

Meningitis 2 (3,6) 2 (100)

Unknown 15 (26,7) 10 (67)

Total 56 (100) 16 (28,6)

Children followed-up over 18 months by the CePReF and the MTCT-Plus programme, Abidjan, Côte d’Ivoire (January-2004-December 2009). n = 405.



Page 6 of 12

most likely infected through MTCT. Although the exact

time of infection (antepartum, intrapartum, or postnatal)

is not documented, we could reasonably estimate the

time of infection as being the date of birth. However,

older children in this cohort may have been infected

perinatally, or they may have been exposed to HIV by

blood transfusion, sexual activity or hospital-acquired

infections. These children are mostly followed up in the

CePReF programme, having been diagnosed with HIV-1

following a severe opportunistic infection; their time of

infection cannot be estimated with the available data.

Furthermore, mortality is severe amongst infants in the

absence of ART [4], which combined with the late diag-

noses issues, induces a strong left truncation phenom-

enon [29]. To address this left-censoring of the data, we

modelled infection from the date of enrolment in the

healthcare programme, stratifying by age. Thus, the

information observed over different age groups provides

a detailed estimate of clinical disease progression in pre-

valent cohorts for older children, but could be consid-

ered as the clinical disease progression of perinatally

infected children for those aged less than 5 years.

Second, morbidity and mortality may have been

incompletely ascertained. Many African families with-

hold information concerning infection and diagnosis, as

well as information relative to prior events and deaths

during follow-up [30]. In addition, missing data arose as

a result of the reality of healthcare services in the field

in Africa and the lack of follow-up in clinical examina-

tions. The data coming from the pre-existing MTCT

datasets were precise, as they derive from previous

research programmes, but most of the data from the

0 5 10 15

0.0

0.2

0.4

0.6

0.8

1.0

Time to event(months)

Pro

ba

bility fo

r firs

t se

ve

re e

ve

nt

< 1 year

1-2 years

2-5 years

5-10 years

10-15 years

Gray's test: p-value=0.51

Figure 2A

0 5 10 15

0.0

0.2

0.4

0.6

0.8

1.0

Time to first serious event (months)

Pro

ba

bility fo

r firs

t se

rio

us e

ve

nt

YES

NO

UNKNOWN

Gray's test: p-value=0.87

Figure 2B

Figure 2 Cumulative incidence functions for first severe morbidity by age (2A) and immunodeficiency status (2B). Children followed up

at the CePReF and in the MTCT-Plus programme, Abidjan, Côte d’Ivoire (January 2004 - December 2009). n = 405.



Page 7 of 12

CePReF, which represents 77% of the children, was col-

lected retrospectively by the means of a standardised

collection instrument. Record-keeping practices may

vary over time and between providers. As a conse-

quence, diagnoses could not be confirmed using stan-

dardised diagnosis procedures. We attempted to ensure

the accuracy of clinical events by requiring documenta-

tion, such as clinical findings, laboratory examinations,

and radiographic results. However, this may have led to

under-documentation of the true rate of clinical events.

This could explain, in part, the observed lower risk of

severe morbidity in the older children. Moreover, the

morbidity observed must be interpreted with caution.

Not only are the diagnoses not confirmed, but there is

no control group of non-HIV-infected children with

whom to compare specific morbidity rates. Nonetheless,

we show here a high rate of overall severe and avoidable

morbidity despite prophylaxis in HIV-infected children

who do not have access to ART.

In addition to morbidity, the infant mortality rate in

this cohort may be underestimated. Many events leading

to death remain undetermined. These unknown causes

of death are frequent because most deaths occurred at

home; these deaths were most likely reported to the

clinic only in response to a follow-up call made by a

care worker. We also suspect that a substantial propor-

tion of the children lost to follow-up are in fact

deceased. Moreover, many of these children were immu-

nosuppressed according to the 2006 WHO criteria,

which is a known predictor of death. Nevertheless, the

overall loss to follow up rate was only 11% at 18

months; even taking into account the exclusion of 12.3%

Table 3 Characteristics associated with the first severe morbidity during follow-up in untreated HIV-infected children

Univariate analysis Ajusted analysis

Full model Reduced model

Variable # patients # events Observed rate (/100 CY) sHR† 95%CI‡ sHR 95%CI P sHR 95%CI P

Overall 405 58 17.80

Age at baseline(years)

0.54 0.39 0.07

< 1 64 7 12.78 1 - 1 - 1 -

[1 - 2] 42 7 20.36 1.29 (0.47-3.58) 0.45 (0.12 - 1.64) 0.52 (0.13 - 2.01)

[2 - 5] 108 17 20.41 1.24 (0.53 - 2.91) 0.43 (0.14 - 1.31) 0.42 (0.14 - 1.32)

[5 - 10] 142 21 16.65 0.94 (0.41 - 2.15) 0.38 (0.13 - 1.14) 0.30 (0.10 - 0.97)

[10 - 15] 49 4 14.58 0.45 (0.12 - 1.70) 0.23 (0.05 - 1.02) 0.15 (0.03 - 0.66)

Immunodeficiency 0.87 0.91 0.86

No 112 29 13.61 1 - 1 - 1 -

Yes 106 15 39.24 0.84 (0.43-1.63 0.88 (0.47 - 1.65) 0.84 (0.43 - 1.63)

Unknown 98 14 18.75 0.91 (0.46 - 1.77) 1.02 (0.54 - 1.93) 1.03 (0.51 - 2.07)

Gender 0.78 0.95 0.77

Male 216 32 18.07 1 - 1 - 1 -

Female 189 26 17.46 0.93 (0.54 - 1.60) 0.91 (0.54 - 1.58) 0.91 (0.51 - 1.61)

History of a PMTCTintervention

0.11 0.19 0.05

No 267 41 21.72 1 - 1 - 1 -

Yes 65 5 7.41 0.43 (0.15 - 1. 02) 0.32 (0.06 - 1. 63) 0.16 (0.04 - 0.72)

Unknown 73 12 17.20 1.04 (0.53 - 2.02) 1.21 (0.56 - 2.59) 0.82 (0.42 - 1.59)

Mother known to bedeceased

0.13 0.26

No 308 48 18.19 1 - 1 -

Yes 97 10 16.10 1.78 (0.84 - 3.14) 1.51 (0.73 - 3.09)

Cotrimoxazole prophylaxis at baseline 0.35

No 134 17 15.60 1 -

Yes 271 41 18.89 1.33 (0.73 - 2.43)

Follow-up Centre 0.05 0.14

Cepref 131 50 21.92 1 - 1 -

MTCT 92 8 8.17 0.43 (0.18 - 1) 0.45 (0.16 - 1.29)

Children followed-up over 18 months by the CePReF and the MTCT-Plus programme, Abidjan, Côte d’Ivoire (January-2004-December 2009). n = 405. Fine and

Gray Model CY: Child Years; †sHR: subdistribution hazard ratio; ‡95% confidence interval



Page 8 of 12

of the eligible cohort for missing medical records (per-

haps also representing loss-to-follow-up), this rate of

missingness is low in the African context compared to

those of cohorts of children on ART [3,15,31].

Our study is the first to investigate pre-ART morbidity

and mortality in a large paediatric cohort with a wide

age distribution during the ART era, reflecting as well

as possible current care practices in Côte d’Ivoire

[8,32-34]. Several of our results are in keeping with pre-

vious findings, suggesting the validity of our results

despite the limitations described above. First, the mor-

tality documented in this cohort is in accord with the

bi-modal pattern of the paediatric HIV infection, mainly

characterised by early mortality. We report a mortality

rate three times higher in children aged less than 12

months than in older children, but the numbers remain

low compared to those of previous studies [34-37]. As

have shown recent studies, CD4 cell counts are impor-

tant indicators [38,39] of mortality in this analysis.

Our study first suggests low rates of death but high

rates of severe morbidity among non-ART treated HIV

infected children. Despite the efforts of implementing

antiretroviral access and PMTCT interventions in Abid-

jan, our results demonstrate still a high rate of infected

children who do not have access to ART. This justifies

the revision of the WHO guidelines recommending

ART to all infected children aged < 24 months regard-

less of CD4 count, but this is not always applied in the

Figure 3 Cumulative incidence functions for mortality by age (3A) and immunodeficiency status (3B). Children followed up at the

CePReF and in the MTCT-Plus programme, Abidjan, Côte d’Ivoire (January 2004 - December 2009). n = 405



Page 9 of 12

field. In many resource-limited settings, ART coverage

remains low mainly for operational reasons, leading to

high mortality and morbidity. Additional paediatric clin-

ical research must be undertaken in order to measure

the incidence of mortality, morbidity, and loss to pro-

gramme in infected children before and after ART

access and the use of prophylactic treatment in opportu-

nistic infections. This would allow investigators to

model the evolution of the paediatric infection and mea-

sure the cost effectiveness of interventions with a final

goal to guiding care of children infected by HIV in

Africa, as has been done successfully in the past for

adults [40-42].

ConclusionsWe describe clinical progression in untreated, HIV-

infected children in Abidjan, Côte d’Ivoire, in the con-

text of paediatric-specific HIV care. Having benefited

from early access to care minimizes the severe morbidity

risk for those children who acquire HIV. Despite the

receipt of cotrimoxazole prophylaxis, the risk of serious

morbidity and mortality remains high in ART-untreated

HIV-infected children. As such clinical events are likely

preventable by early initiation of ART to improve

immune status, these findings demonstrate the need to

promote an earlier access to ART in HIV-infected chil-

dren in Africa, in a context where early ART is officially

Table 4 Characteristics associated with mortality during follow-up in untreated HIV-infected children

Univariate analysis Ajusted analysis

Full model Reduced model

Variable #patients

#events

Observed rate (/100CY)

sHR† 95%CI‡ P sHR 95%CI P sHR 95%CI P

Overall 405 21 3.26

Age at baseline(years)

0.034 0.840 0.097

< 1 64 9 9.70 1 - 1 - 1 -

[1 - 2] 42 2 3.40 0.32 (0.07 - 1.48) 0.76 (0.08 - 6.38) 0.6 (0.11 - 3.23)

[2 - 5] 108 3 1.68 0.2 (0.05 - 0.74) 0.44 (0.07 - 2.70) 0.34 (0.08 - 1.39)

[5 - 10] 142 5 1.82 0.24 (0.08 - 0.71) 0.39 (0.06 - 2.31) 0.29 (0.10 - 0.80)

[10 - 15] 49 2 4.91 0.28 (0.06 - 1.29) 0.44 (0.05 - 4.26) 0.28 (0.06 - 1.19)

Immunodeficiency 0.001 0.003 0.0025

No 112 6 1.36 1 - 1 - 1 -

Yes 106 2 3.38 5.21 (1.00 -25.60)

6.8 (1.51 -30.57)

6.02 (1.28 -28.42)

Unknown 98 13 8.90 13.82 (3.16 -60.50)

12.93 (2.95 -56.68)

12.6 (2.94 -53.97)

Gender 0.920 0.770 0.670

Male 216 11 3.09 1 - 1 - 1 -

Female 189 10 3.45 1.05 (0.47 - 2.45) 1.15 (0.44 - 2.98) 1.22 (0.49 - 3.04)

History of a PMTCTintervention

0.019 0.590

No 267 11 2.80 1 - 1 -

Yes 65 8 6.73 2.96 (1.19 - 7.33) 0.92 (0.14 - 6.21)

Unknown 73 2 1.48 0.63 (0.14 - 2.83) 0.47 (0.11 - 2.05)

Mother known to bedeceased

0.140 0.230

No 308 19 3.59 1 - 1 -

Yes 97 2 1.72 3.01 (0.71 -12.80)

2.78 (0.53 -14.68)

Cotrimoxazole prophylaxis at baseline 0.150 0.830

No 134 11 4.93 1 - 1 -

Yes 271 10 2.36 0.53 (0.23 - 1.25) 1.13 (0.37 - 3.43)

Follow-up Centre 0.550

Cepref 131 15 3.34 1 -

MTCT 92 6 3.05 1.33 (0.52 - 3.45)

Children followed-up over 18 months by the CePReF and the MTCT-Plus programme, Abidjan, Côte d’Ivoire (January-2004-December 2009). n = 405. Fine and

Gray Model. CY: Child Years; †sHR: subdistribution hazard ratio; ‡95% confidence interval



Page 10 of 12

recommended and is being scaled up, but is still not

available to a majority of children.

Acknowledgements

This study was supported in part by grants from the US National Institutes

of Health: NIAID R0I AI058736 and 5U01AI09919-01 to 05 (IeDEA West Africa)

and K01 AI078754 (AC) and R01AI058736 (SD). The content of this

publication is solely the responsibility of the authors and does not

necessarily represent the official views of any of the institutions mentioned

above.

Sophie Desmonde is a fellow of the Ecole des Hautes Etudes en Santé

Publique (EHESP), Rennes, France. We would like to thank the children and

their families who participated in the Aconda care programmes. We are

deeply grateful to all staff members of the PACCI and ACONDA teams (the

CePReF and the MTCT-Plus programme) involved in the care of HIV/AIDS in

Abidjan. Each of these people has during the post-electoral crisis in 2011

courageously continued to serve the sick, wounded, displaced, and refugees

- all victims of the war, and to bring them medicine, food, water, support

and humanity. We would also like to thank the CEPAC group (Cost-

effectiveness for Preventing AIDS Complications) for their collaboration and

input in this study. Finally, a special thank you goes to Dr X. Anglaret

(Inserm U897 and Programme PACCI) for his helpful support.

Author details1Inserm, U897 & Institut de Santé Publique, Epidémiologie et

Développement (ISPED), Université Victor Segalen Bordeaux 2, Bordeaux,

France. 2Programme PAC-CI, CHU de Treichville, Abidjan, Côte d’Ivoire.3CePReF, ACONDA, Abidjan Côte d’Ivoire. 4MTCT-Plus Initiative, ACONDA,

Abidjan, Côte d’Ivoire. 5Massachusetts General Hospital, Boston,

Massachusetts, USA.

Authors’ contributions

EA, CB and EM contributed the necessary data from their health facilities. SD

collected the data. PC, EA, CB, EM and VL contributed to the study design.

SD, AA and VL carried out the statistical analysis and interpreted the results.

FD, AC and VL contributed to the coordination. SD and VL first drafted the

manuscript. Critical revision of the manuscript for important intellectual

content: SD, PC, FD, AA, AC and VL. All authors read and approved the final

manuscript.

Competing interests

The authors declare that they have no competing interests.

Received: 30 December 2010 Accepted: 23 June 2011

Published: 23 June 2011

References

1. World Health Organization.UNAIDS: 2009 Report on global AIDS epidemic

Geneva, UNAIDS; 2009.

2. Newell M, Coovadia H, Cortina-Borja M, Rollins N, Gaillard P, Dabis F:

Mortality of infected and uninfected infants born to HIV-infected

mothers in Africa: a pooled analysis. Lancet 2004, 364(9441):1236-1243.

3. Anaky M, Duvignac J, Wemin L, Kouakoussui A, Karcher S, Touré S, Seyler C,

Fassinou P, Dabis F, N’Dri-Yoman T, Anglaret X, Leroy V: Scaling up

antiretroviral therapy for HIV-infected children in Côte d’Ivoire:

determinants of survival and loss to programme. Bull World Health Organ

2010, 88(7):490-499.

4. Stringer EM, Ekouevi DK, Coetzee D, Tih PM, Creek TL, Stinson K, Giganti MJ,

Welty TK, Chintu N, Chi BH, Wilfert CM, Shaffer N, Dabis F, Stringer JS,

PEARL Study Team: Coverage of nevirapine-based services to prevent

mother-to-child HIV transmission in 4 African countries. JAMA 2010,

304(3):293-302.

5. The KIDS-ART-LINC Collaboration: Low risk of death, but substantial

program attrition, in pediatric HIV treatment cohorts in Sub-Saharan

Africa. J Acquir Immune Defic Syndr 2008, 49(5):523-531.

6. Dunn D, HIV Paediatric Prognostic Markers Collaborative Study Group:

Short-term risk of disease progression in HIV-1-infected children

receiving no antiretroviral therapy or zidovudine monotherapy: a meta-

analysis. Lancet 2003, 362(9396):1605-1611.

7. Lucas SB, Peacock CS, Hounnou A, Brattegaard K, Koffi K, Hondé M,

Andoh J, Bell J, De Cock KM: Disease in children infected with HIV in

Abidjan, Côte d’Ivoire. BMJ 1996, 312(7027):335-338.

8. Marinda E, Humphrey JH, Iliff PJ, Mutasa K, Nathoo KJ, Piwoz EG,

Moulton LH, Salama P, Ward BJ, ZVITAMBO Study Group: Child mortality

according to maternal and infant HIV status in Zimbabwe. Pediatr Infect

Dis J 2007, 26(6):519-526.

9. Harambat J, Fassinou P, Becquet R, Touré P, Rouet F, Dabis F, Msellati P,

Blanche S, Timité-Konan M, Salamon R, Leroy V, ANRS 1201/1202 Ditrame

Plus Study Group: 18-month occurrence of severe events among early

diagnosed HIV-infected children before antiretroviral therapy in Abidjan,

Côte d’Ivoire: a cohort study. BMC Public Health 2008, 8:169.

10. Vaz P, Elenga N, Fassinou P, Msellati P, Nicolas J, Blanche S: HIV-1 infection

in children in African countries. Med Trop (Mars) 2003, 63(4-5):465-472.

11. Bobat R, Moodley D, Coutsoudis A, Coovadia H, Gouws E: The early natural

history of vertically transmitted HIV-1 infection in African children from

Durban, South Africa. Ann Trop Paediatr 1998, 18(3):187-196.

12. Walenda C, Kouakoussui A, Rouet F, Wemin L, Anaky M, Msellati P:

Morbidity in HIV-1-Infected children treated or not treated with highly

active antiretroviral therapy (HAART), Abidjan, Cote d’Ivoire, 2000-04. J

Trop Pediatr 2009, 55(3):170-176.

13. Slogrove AL, Cotton MF, Esser MM: Severe infections in HIV-exposed

uninfected infants: clinical evidence of immunodeficiency. J Trop Pediatr

2010, 56(2):75-81.

14. Leyenaar JK, Novosad PM, Ferrer KT, Thahane LK, Mohapi EQ, Schutze GE,

Kline MW: Early clinical outcomes in children enrolled in human

immunodeficiency virus infection care and treatment in Lesotho. Pediatr

Infect Dis J 2010, 29(4):340-345.

15. Sutcliffe CG, van Dijk JH, Bolton C, Persaud D, Moss WJ: Effectiveness of

antiretroviral therapy among HIV-infected children in sub-Saharan Africa.

Lancet Infect Dis 2008, 8(8):477-489.

16. Toure S, Kouadio B, Seyler C, Traore M, Dakoury-Dogbo N, Duvignac J,

Diakite N, Karcher S, Grundmann C, Marlink R, Dabis F, Anglaret X, Aconda

Study Group: Rapid scaling-up of antiretroviral therapy in 10,000 adults

in Côte d’Ivoire: 2-year outcomes and determinants. AIDS 2008,

22(7):873-882.

17. Tonwe-Gold B, Ekouevi DK, Bosse CA, Toure S, Koné M, Becquet R, Leroy V,

Toro P, Dabis F, El Sadr WM, Abrams EJ: Implementing family-focused HIV

care and treatment: the first 2 years’ experience of the mother-to-child

transmission-plus program in Abidjan, Côte d’Ivoire. Trop Med Int Health

2009, 14(2):204-212.

18. World Health Organization: WHO case definitions for HIV surveillance and

revised clinical staging and immunological classification of HIV-related disease

in adults and children 2006.

19. American Academy of Pediatrics: Summaries of Infectious Diseases. In Red

Book: 2009 Report of the Committee on Infectious Diseases.. 28 edition. Edited

by: Pickering LK, Baker CJ, Kimberlin DW, Long SS. Elk Grove Village IL:

American Academy of Pediatrics; 2009:203-755.

20. Walenda C, Kouakoussui A, Rouet F, Wemin L, Anaky M, Msellati P:

Morbidity in HIV-1-Infected children treated or not treated with highly

active antiretroviral therapy (HAART), Abidjan, Cote d’Ivoire, 2000-04. J

Trop Pediatr 2009, 55(3):170-176.

21. Timité-Konan AM, Fassinou P, Adonis-Koffy L, N’dja BB: Use of antiretroviral

drugs in HIV infected children in a tropical area: a real benefit. Arch

Pediatr 2003, 10(9):831-832.

22. Gona P, Van Dyke RB, Williams PL, Dankner WM, Chernoff MC, Nachman SA,

Seage GR: Incidence of opportunistic and other infections in HIV-infected

children in the HAART era. JAMA 2006, 296(3):292-300.

23. Putter H, Fiocco M, Geskus RB: Tutorial in biostatistics: competing risks

and multi-state models. Stat Med 2007, 26(11):2389-2430.

24. Scrucca L, Santucci A, Aversa F: Competing risk analysis using R: an easy

guide for clinicians. Bone Marrow Transplant 2007, 40(4):381-387.

25. Scrucca L, Santucci A, Aversa F: Regression modeling of competing risk

using R: an in depth guide for clinicians. Bone Marrow Transplant 2010,

45(9):1388-1395.

26. Tchendjou P, Same-Ekobo C, Nga A, Tejiokem M, Kfutwah A, Nlend AN,

Tsague L, Bissek AC, Ekoa D, Orne-Gliemann J, Rousset D, Pouillot R,

Dabis F: Effectiveness of multidrug antiretroviral regimens to prevent

mother-to-child transmission of HIV-1 in routine public health services in

Cameroon. PLoS ONE 2010, 5(4):e10411.



Page 11 of 12

27. Ahoua L, Ayikoru H, Gnauck K, Odaru G, Odar E, Ondoa-Onama C,

Pinoges L, Balkan S, Olson D, Pujades-Rodríguez M: Evaluation of a 5-year

programme to prevent mother-to-child transmission of HIV infection in

Northern Uganda. J Trop Pediatr 2010, 56(1):43-52.

28. Chintu C, Bhat GJ, Walker AS, Mulenga V, Sinyinza F, Lishimpi K, Farrelly L,

Kaganson N, Zumla A, Gillespie SH, Nunn AJ, Gibb DM, CHAP trial team:

Co-trimoxazole as prophylaxis against opportunistic infections in HIV-

infected Zambian children (CHAP): a double-blind randomised placebo-

controlled trial. Lancet 2004, 364(9448):1865-1871.

29. Brookmeyer R: Statistical problems in epidemiologic studies of the

natural history of disease. Environ Health Perspect 1990, 87:43-49.

30. Orne-Gliemann J, Becquet R, Ekouevi DK, Leroy V, Perez F, Dabis F: Children

and HIV/AIDS: from research to policy and action in resource-limited

settings. AIDS 2008, 22(7):797-805.

31. Sauvageot D, Schaefer M, Olson D, Pujades-Rodriguez M, O’Brien DP:

Antiretroviral therapy outcomes in resource-limited settings for HIV-

infected children <5 years of age. Pediatrics 2010, 125(5):e1039-1047.

32. Perez F, Zvandaziva C, Engelsmann B, Dabis F: Acceptability of routine HIV

testing ("opt-out”) in antenatal services in two rural districts of

Zimbabwe. J Acquir Immune Defic Syndr 2006, 41(4):514-520.

33. Zijenah L, Mbizvo MT, Kasule J, Nathoo K, Munjoma M, Mahomed K,

Maldonado Y, Madzime S, Katzenstein D: Mortality in the first 2 years

among infants born to human immunodeficiency virus-infected women

in Harare, Zimbabwe. J Infect Dis 1998, 178(1):109-113.

34. Spira R, Lepage P, Msellati P, Van De Perre P, Leroy V, Simonon A, Karita E,

Dabis F: Natural history of human immunodeficiency virus type 1

infection in children: a five-year prospective study in Rwanda. Mother-

to-Child HIV-1 Transmission Study Group. Pediatrics 1999, 104(5):e56.

35. Newell M, Brahmbhatt H, Ghys PD: Child mortality and HIV infection in

Africa: a review. AIDS 2004, 18(Suppl 2):S27-34.

36. Dabis F, Elenga N, Meda N, Leroy V, Viho I, Manigart O, Dequae-

Merchadou L, Msellati P, Sombie I, DITRAME Study Group: 18-Month

mortality and perinatal exposure to zidovudine in West Africa. AIDS

2001, 15(6):771-779.

37. Taha TE, Graham SM, Kumwenda NI, Broadhead RL, Hoover DR, Markakis D,

van Der Hoeven L, Liomba GN, Chiphangwi JD, Miotti PG: Morbidity

among human immunodeficiency virus-1-infected and-uninfected

African children. Pediatrics 2000, 106(6):E77.

38. Wamalwa DC, Obimbo EM, Farquhar C, Richardson BA, Mbori-Ngacha DA,

Inwani I, Benki-Nugent S, John-Stewart G: Predictors of mortality in HIV-1

infected children on antiretroviral therapy in Kenya: a prospective

cohort. BMC Pediatr 2010, 10:33.

39. The Cross Continents Collaboration for Kids: Markers for predicting

mortality in untreated HIV-infected children in resource-limited settings:

a meta-analysis. AIDS 2008, 22(1):97-105.

40. Goldie SJ, Yazdanpanah Y, Losina E, Weinstein MC, Anglaret X, Walensky RP,

Hsu HE, Kimmel A, Holmes C, Kaplan JE, Freedberg KA: Cost-effectiveness

of HIV treatment in resource-poor settings–the case of Côte d’Ivoire. N

Engl J Med 2006, 355(11):1141-1153.

41. Walensky RP, Wolf LL, Wood R, Fofana MO, Freedberg KA, Martinson NA,

Paltiel AD, Anglaret X, Weinstein MC, Losina E, CEPAC (Cost-Effectiveness of

Preventing AIDS Complications)-International Investigators: When to start

antiretroviral therapy in resource-limited settings. Ann Intern Med 2009,

151(3):157-166.

42. Yazdanpanah Y, Losina E, Anglaret X, Goldie SJ, Walensky RP, Weinstein MC,

Toure S, Smith HE, Kaplan JE, Freedberg KA, for the Global AIDS Policy

Model Investigators: Clinical impact and cost-effectiveness of co-

trimoxazole prophylaxis in patients with HIV/AIDS in Côte d’Ivoire: a

trial-based analysis. AIDS 2005, 19(12):1299-1308.

Pre-publication history

The pre-publication history for this paper can be accessed here:

http://www.biomedcentral.com/1471-2334/11/182/prepub

doi:10.1186/1471-2334-11-182Cite this article as: Desmonde et al.: Severe morbidity and mortality inuntreated HIV-infected children in a paediatric care programme inAbidjan, Côte d’Ivoire, 2004-2009. BMC Infectious Diseases 2011 11:182.

Submit your next manuscript to BioMed Centraland take full advantage of:

• Convenient online submission

• Thorough peer review

• No space constraints or color figure charges

• Immediate publication on acceptance

• Inclusion in PubMed, CAS, Scopus and Google Scholar

• Research which is freely available for redistribution

Submit your manuscript at www.biomedcentral.com/submit



Page 12 of 12

77

4.1.3. Principal results and discussion

The results of this cohort study conducted in Abidjan between 2004 and 2009 suggest high

rates of severe morbidity and mortality in HIV-infected children followed up in care and who

had not yet initiated ART. The risk of severe morbidity reached 11% by 18 months of follow-

up into HIV care programme; it was minimised in older children and in those who were

known to have benefited from a PMTCT intervention, marker of early and improved access to

care. Mortality reached 8% by 18 months and was highest in children with no CD4

measurement available, underscoring the gaps in the linkage to care between diagnosis, CD4

measurement and ART initiation.

In this study, we investigated time to death and first severe morbid event in HIV-infected

children who had enrolled in a pre-ART care programme. Both events were strongly

dependant on ART initiation: children who initiated ART would no longer be at risk of

experiencing pre-ART mortality or morbidity. We accounted for this by using a competing

risk approach; the Fine and Gray model is a cause-specific hazards model which combines

proportional hazards models for the event of interest and the competing event. This allowed

us to present sub-distribution hazard ratios which can directly be interpreted as relative risks

for each covariate (196). Although both events were also dependant on LTFU (the events

could not be observed in LTFU children) we chose not to use the same approach but rather to

right-censor our data at date of last clinical contact. This choice was mainly based on a

parallel study which compared hazard ratios for mortality in pre-ART care, comparing a

standard Cox survival model and the Fine and Gray model. Indeed, this study concluded that

the use of competing risks for correcting for LTFU would actually underestimate the

mortality as it would consider LTFU children alive and well, and this is not necessarily the

case (197).

As discussed in the paper, these results also highlight the effect of age at enrolment on clinical

outcomes and the relative positive effect of early access to care on reducing the risks of both

severe morbidity and mortality before initiating ART. However, there was a strong selection

bias inherent in this study. While these children were probably mainly HIV-infected

perinatally, their median age of the cohort at baseline was 4.5 years. This age can be

assimilated to the median age of HIV diagnosis at inclusion into the care programme.

Consequently, children who had a chance to be enrolled in the Aconda programme were

already survivors from birth and had survived through early childhood, underestimating the

true burden of HIV. Nevertheless, this study is one of the first to investigate pre-ART

morbidity and mortality in a large paediatric cohort with a wide age distribution during the

ART era, reflecting as well as possible the clinical practices at that time in Côte d’Ivoire.

78

4.2. Healthcare resource utilisation in untreated HIV-infected children

in a paediatric care programme, Abidjan, Côte d’Ivoire, 2004-2009


In Côte d’Ivoire, as in many low-income countries in West Africa, scaling up comprehensive

care for HIV-infected children still encounters many barriers. We are reminded that the

proportion of HIV-infected children with access to ART in Côte d’Ivoire is unacceptably low,

reaching only 16% in 2012 (26). Understanding the current disease burden and the unmet

medical need of HIV-infected children awaiting ART initiation provides context for defining

care interventions to improve the health outcomes in these children. Furthermore, data on the

utilisation of healthcare resources in these children, before they access ART would be helpful

in assessing changes in paediatric healthcare utilisation before and after access to ART. It is

assumed that children who initiate ART early will have lower morbidity than those initiated at

a later age. However, as the mortality would also be lower, they will survive longer with

potentially higher cumulative healthcare resource utilisations in the long-term. These trends

should be analysed to be able to plan and meet correctly child care in this context.

Within the same cohort as the previous study, we thus described the healthcare resource

utilisation among HIV-infected children who had not yet undergone ART over the 2004-2009

scaling up period in Abidjan.

4.2.2. Desmonde et al – JAIDS – January 2013

The results of this study were presented at the 6th IAS conference on HIV pathogenesis,

treatment and prevention in 2011, before being published in JAIDS in 2013.

IMPLEMENTATION AND OPERATIONAL RESEARCH: EPIDEMIOLOGY AND PREVENTION

Health Care Resource Utilization in Untreated HIV-InfectedChildren in A Pediatric Programme, Abidjan, Cote d’Ivoire,

2004–2009

Sophie Desmonde, MSc,*† Patrick A. Coffie, MD, PhD,‡ Edmond A. Aka, MD,§

Clarisse Amani-Bosse, MD,‡k Eugène Messou, MD, PhD,§ François Dabis, MD, PhD,*†

Andrea Ciaranello, MD, MPH,¶ and Valériane Leroy, MD, PhD*†

Background: We describe health care resource utilization among

HIV-1–infected children who have not yet undergone antiretroviral

treatment (ART) in Abidjan, Côte d’Ivoire.

Methods: HIV-infected children enrolled prospectively in an HIV

care programme in 2 health facilities in Abidjan (2004–2009) were

followed up from date of inclusion until database closeout, death,

ART initiation, or loss to follow-up (no clinical contact for more than

6 months). Incidences of health care resource utilization (outpatient

care, inpatient day care, and hospitalization) were described accord-

ing to severe morbidity and mixed effect log linear models were

computed to identify associated factors.

Results: Overall, 405 children were included, entering care at

a median age of 4.5 years, 66.9% were receiving cotrimoxazole

prophylaxis, and 27.7% met 2006 WHO criteria for immunodefi-

ciency by age. The median follow-up time was 11.6 months

(interquartile range: 1.4; 30.7). Overall, 371 clinical events occurred

in 162 children yielding to an incidence rate (IR) of 60.9/100 child-

years (CY) [95% confidence interval (CI): 55.1 to 67.2]: 57% of

clinical events led to outpatient care (IR: 33/100 CY), 38% to inpatient

day care (IR: 22/100 CY), and 10% to hospitalization (IR: 5.9/100

CY). Further medical examinations were made allowing confirmed

diagnoses in 40% of those (IR: 22.4/100 CY). Outpatient care was less

common among immunodeficient children than those not (relative risk

[RR] = 0.32, 95% CI: 0.18 to 0.56), in those whose main caregivers

are both parents compared with those who are primarily cared for by

their mother only (RR = 0.34, 95% CI: 0.15 to 0.77).

Conclusion: Untreated HIV-infected children require substantial

inpatient and outpatient care in a context where ART is scaling up

but still not available to all.

Key Words: HIV, pediatrics, health care resource utilization, mor-

bidity, West Africa

(J Acquir Immune Defic Syndr 2013;62:e14–e21)

INTRODUCTIONThe implementation of programmes aimed at preventing

mother-to-child transmission (PMTCT) and providing antire-troviral therapy (ART) have dramatically improved pediatricHIV care in sub-Saharan Africa.1–4 However, in Côte d’Ivoire,as in many low-income countries in West Africa, scaling upcomprehensive care for HIV-infected children still encountersmany barriers.5–7

First, although PMTCT coverage has improved greatly,reaching 54% in 2009,8 mothers continue to transmit thedisease to their children6–9 leading to an ongoing pediatricHIV epidemic. Second, the uptake of early infant diagnosisis poor in Côte d’Ivoire and requires routine polymerase chainreaction techniques10 that are expensive and not available toall. Consequently, children are diagnosed belatedly, at anadvanced clinical and/or immunologic stage of the disease.In addition, providing a continuum of care between postnatalHIV diagnosis, pediatric care, and ART remains a challengein Côte d’Ivoire. Without ART, HIV-infected children in Côted’Ivoire also experience severe morbidity and mortality.11–13

The proportion of HIV-infected children with access to ARTin Côte d’Ivoire remains unacceptably low; it was estimatedto reach only 15% in 2010,4 despite the scaling up of ARTsince 2004.

Although many studies describe the effects of ARTboth in adults14,15 and children,5,16,17 few have assessed theutilization of health care resources in HIV-infected childrenbefore their access to ART. These data would be helpful inassessing changes in pediatric health care utilization before

Received for publication April 10, 2012; accepted August 14, 2012.From the *Centre INSERM, U897—Epidémiologie—Biostatistiques, Bor-

deaux, France; †Université Bordeaux, ISPED, Centre INSERM, U897—Epidémiologie—Biostatistiques, Bordeaux, France; ‡Programme PAC-CI,CHU de Treichville, Abidjan, Côte d’Ivoire; §Centre de Prise en Charge, deRecherche et de Formation, ACONDA, Abidjan, Côte d’Ivoire; kMTCT-Plus Initiative, ACONDA, Abidjan, Côte d’Ivoire; and ¶Division of Infec-tious Diseases, Massachusetts General Hospital, Boston, MA.

Supported in part by the US National Institutes of Health: National Instituteof Allergy and Infectious Diseases Grants R01 A1058736,5U01AI09919-01–05 (IeDEA West Africa), R01AI05873, andK01AI078754. S. Desmonde is a fellow of the Ecole des Hautes Etudesen Santé Publique, Rennes, France. The content of this publication issolely the responsibility of the authors and does not necessarily representthe official views of any of the institutions mentioned above.

Presented in part at the 6th IAS Conference on HIV pathogenesis, treatment,and prevention, July 2011, Rome, Italy; poster # MOPE129.

The authors have no conflicts of interest to disclose.Correspondence to: S. Desmonde, Centre de recherche INSERM, U897,

Equipe—VIH, cancer et santé globale dans les pays à ressources limitées,Institut de Santé Publique, d’Epidémiologie et de Développement,Université de Bordeaux 2, Case 11, 146 rue Leo Saignat, 33076 BordeauxCedex, France (e-mail: [email protected]).

Copyright © 2012 by Lippincott Williams & Wilkins

e14 | www.jaids.com J Acquir Immune Defic Syndr ! Volume 62, Number 1, January 1, 2013

and after access to ART and guiding the “when to treat”question that remains unanswered in HIV-infected childrenaged 2 years and older.18

In this study, we analyzed health care resource utilization(HCRU) according to severe morbidity among ART-untreatedHIV-1–infected children waiting for access to ART over the2004–2009 scaling up period in Abidjan, Côte d’Ivoire.

METHODS

SettingsThis study was carried out in Abidjan, Côte d’Ivoire. In

2004, the estimated HIV prevalence in this setting in pregnantwomen was 8.3%.19,20 In 2008, PMTCT coverage was 40%and the HIV prevalence in newborn infants was 2.5%.21

The ACONDA programme is a nongovernmental asso-ciation whose main objective is providing care to HIV-infectedpatients in Côte d’Ivoire. Children enter the ACONDA pro-gramme in one of two ways: (1) after an HIV diagnosis ata pediatric clinic after presentation with AIDS-related symp-toms or (2) after referral for HIV testing because their motherwas identified as HIV infected. In partnership with the Bor-deaux School of Public Health (ISPED, France), ACONDAlaunched in 2004 a 5-year programme of access to HIV careand treatment delivering ART according to the 2006 WHOguidelines. In addition to a number of public and private healthcare facilities, the programme relies mostly on 2 health carefacilities entirely dedicated to pediatric care: the Centre de Priseen Charge, de Recherche et de Formation (CePReF) Enfant andthe MTCT-Plus programme (mother-to-child-transmission pre-vention programme). The CePReF provides care for one of thelargest active pediatric ART cohorts in Abidjan5; the MTCT-Plus programme follows infants born to identified HIV-infected mothers.22

Standard of CareThe ACONDA programme delivers a comprehensive

model of pediatric HIV care covering 3 components: psycho-logical care (diagnosis disclosure), clinical care (diagnosis andtreatment, including ART and prophylaxis of opportunisticinfections), and prevention (HIV screening). In addition,clinical research studies are also conducted in the CePReFand the MTCT-Plus sites. Partly financed by the President’sEmergency Plan for AIDS Relief through the Elizabeth GlaserPediatric AIDS Foundation, ARTs, cotrimoxazole prophylaxis,and blood analyses are free of charge. However, x-rays,in-patient day care, and other medication (antibiotics and anti-malarial treatment) are only partially subsidized, whereas rou-tine medical examinations (blood smears, cultures, andmicroscopy) are still mostly paid for by patient families.

Study Design and ParticipantsEligibility criteria for inclusion in our study included all

those aged 15 years or younger, who had not initiated anyform of ART other than a PMTCT intervention and who wereenrolled in the ACONDA programme (CePReF and MTCT-

Plus) between January 1, 2004 and December 31, 2009 aftera confirmed HIV diagnosis by polymerase chain reaction ora serology if aged 18 months and older.

Data Collection and AnalysisPatient data were mainly stored in paper-based medical

records at the CePReF. The data were collected retrospec-tively using a standardized data collection instrument issuedspecifically for this purpose. A thorough description of thedata collection has been described elsewhere.11 We analyzedHCRU among children followed up at least once between2004 and 2009, from their inclusion in the ACONDA pro-gramme until ART initiation or closeout date (death or lost tofollow-up, defined as no clinical contact more than 6 months).

Events were classified as “severe morbidity” if theywere suspected to be WHO stage 3 or 4 or if they led toinpatient day care, hospitalization, or death. Because therewas no standard diagnosis validation tool, events weredefined as “definite” or “probable” according to the WHOcase definitions of HIV surveillance (2006).23 In addition,malaria was considered to have occurred if the diagnosiswas either confirmed by a positive blood smear or suspectedby the presence of a high temperature leading to a prescriptionfor an antimalarial treatment. To be consistent with the studyperiod, immunodeficiency was defined according to the WHOrecommendation issued in 2006.24

HCRU was defined as either (1) outpatient care: (1.1)medical examination with complementary diagnosis method(such as complete blood count, x-ray, and blood smear) and(1.2) any of the following drug prescription: antibiotics,antimalarial treatments, and tuberculosis treatment not involv-ing any kind of hospitalization or (2) inpatient care: (2.1)inpatient day care by periods of 24 hours and (2.2)hospitalization. All HCRU initiated within 24 hours of thediagnosis were considered.

Baseline categorical data are presented as frequencies(percentage) and continuous variables as median [interquartilerange (IQR)]. Incidence rates (IR) of HCRU (complementarydiagnosis, drug prescription, inpatient day care, and hospital-ization) occurring per 100 child-years (CY) of follow-up werecomputed with their 95% confidence interval (95% CI). IRswere described overall and according to age groups.

Factors associated with the IR of HCRU were describedwith a Poisson regression approach, allowing variance adjust-ment for nonindependence when multiple observations wereincluded for a single patient. Relative risks were estimatedusing the generalized estimating equations approach.25 Analy-ses were performed using PROC GEMNOD in SAS 9.2.

RESULTS

Baseline CharacteristicsOverall, 405 children were included in our study: 313

from the CePReF and 92 from the MTCT-plus database. Thebaseline characteristics are presented in Table 1. Briefly,children were diagnosed at a median age of 4.5 years(IQR: 1.9–7.5); 46.7% were female. Overall, 12.3% were

J Acquir Immune Defic Syndr ! Volume 62, Number 1, January 1, 2013 Health Care Resource Utilization for Children

� 2012 Lippincott Williams & Wilkins www.jaids.com | e15

classified as CDC stage C with a significantly higher pro-portion in those aged 10–15 years (22.1%, P , 0.0001).Immunologic data (CD4 percent or count) were availablefor 308 children (76.1%), of these children, 27.7% met the2006 WHO criteria for immunodeficiency by age. The pro-portion of immunodeficient children was highest among the10–15 year olds (55%, P , 0.0001). Of the 74.1% ofchildren eligible for cotrimoxazole prophylaxis accordingto the 2007 WHO recommendations for the use of cotri-moxazole,26 90.3% were under treatment. Only 36% ofchildren aged 1 year or younger were prescribed cotrimox-azole prophylaxis.

Patient Follow-Up and Severe MorbidityOverall 405 children were included and followed up for

a median length of 11.6 months (IQR: 1.4–30.7). At baseline,136 (33.6%) were eligible for ART initiation; of these chil-dren, 113 initiated ART (83.1%); after a median time of 0.9months (IQR: 0.5–3.1), 8 died (5.9%) and 6 were lost tofollow-up (4.4%). The remaining 268 children, not yet eligi-ble for ART at baseline, were followed up for a median timeof 16.1 months (IQR: 7.9–26.2); 127 (47.2%) initiated ART,13 (4.8%) died and 35 (13%) were lost to follow-up. The totalfollow-up period was 642.31 CY of observation.

During this period, 371 severe morbid events wereregistered in 162 children (40%). The median time, from thedate of enrolment until the occurrence of the first event was9.1 months (IQR: 1.2–26.3). The overall observed severemorbidity IR was 60.9 per 100 CY of follow-up (95% CI:55.1 to 67.2); this was significantly more frequent in chil-dren aged 10 years and older, varying from 29.0 per 100 CYin children aged 1 year and younger to 95.5 per 100 CY inchildren aged 10–15 years (P , 0.0001) (Table 2). The

leading cause of severe morbidity was malaria, includingclinical diagnoses of malaria (35%). Lung diseases, includ-ing bronchiectasis, and diarrhea were also frequent causesof morbidity (20% and 14%, respectively) (Table 3).

Health Care Resource Utilization

Outpatient Care

Of the 371 severe morbid events, 212 led to anyoutpatient care, either complementary diagnosis or a pre-scription, yielding an estimated IR of 33.0 per 100 CY offollow-up (95% CI: 28.9 to 37.8) (Table 2).Complementarydiagnosis IR was 22.4 per 100 CY (95% CI: 19.1 to 26.4).Radiology was the most frequent diagnosis tool (15%),mostly used to investigate probable chronic lung diseases(53%) and pulmonary tuberculosis (57%). Of the 21 sus-pected cases of pulmonary tuberculosis, only 5 diagnosesled to complementary diagnoses by sputum culture. Althoughsuspected malaria was the leading cause of severe morbidity,only 11% of the presumptive diagnoses were confirmed byblood smear. Consequently, less than half of the documentedsevere morbidity had confirmed diagnoses.

The overall estimated outpatient prescription IR was15.3 per 100 CY (95% CI: 12.5 to 18.6); antibiotics were themost frequent (IR = 10.1/100 CY, 95% CI: 7.9 to 12.9). Agespecific IRs and their 95% CI are described in Table 2.There were no records of treatment prescriptions for chil-dren aged 10 years and older. We observe the higher rates ofprescriptions among children aged 2–5 years (20.4, 11.0,and 9.9 per 100 CY for overall, antibiotics, and antimalarialtreatment, respectively).

Table 3 describes HCRU according to severe morbidevents. Of the 135 malaria cases, 69 (51%) led to outpatientcare. Chronic lung disease led to a high rate of outpatient care

TABLE 1. Baseline Characteristics by Age in 405 Untreated HIV-Infected Children in a Pediatric HIV Care Programme*

Baseline Characteristics

Overall,

N = 405

Age ,1 yr,

N = 64

Age (1–2) yrs,

N = 42

Age (2–5) yrs,

N = 108

Age (5–10) yrs,

N = 142

Age (10–15) yrs,

N = 49

Age (yrs), median (IQR) 4.5 (1.9–7.5) 0.3 (0.01– 0.6) 1.7 (1.6–1.8) 3.4 (2.6–4.1) 7 (5.9–8.5) 12.1 (11.2–13.1)

Female, n (%) 189 (46.7) 30 (46.9) 19 (45.2) 47 (43.5) 75 (52.8) 18 (36.7)

CD4 cell %, median (IQR) 19 (14–25.5) 18.3 (14.1–26) 17.9 (11.7–23.3) 20.7 (14–26.2) NA NA

CD4 count/mL, median (IQR) 403 (151–665) NA NA NA 507.5 (234–723) 213.5 (55–413)

Immunodeficient children,† n (%) 112 (27.7) 10 (15.6) 10 (15.6) 26 (24.1) 39 (27.5) 27 (55.1)

CDC stage C, n (%) 50 (12.3) 7 (10.9) 6 (14.3) 12 (11.1) 14 (9.9) 11 (22.1)

Baseline cotrimoxazoleprophylaxis n (%)

271 (66.9) 23 (35.9) 33 (78.6) 74 (68.5) 102 (71.8) 39 (67.4)

History of PMTCTintervention, n (%)

65 (18.1) 50 (78.1) 10 (23.8) 5 (4.6) NA NA

Principal caregiver, n (%)

Mother alone 40 (9.9) — 7 (16.7) 5 (4.6) 18 (12.7) 10 (20.4)

Father alone 23 (5.7) — — 4 (3.7) 16 (11.3) 3 (6.1)

Both parents 149 (36.8) 19 (29.7) 22 (52.4) 60 (55.6) 46 (32.4) 2 (4.1)

Other 193 (47.7) 45 (70.3) 13 (30.9) 33 (36.1) 62 (43.6) 34 (69.4)

*CePReF and the MTCT-Plus Programme, Abidjan, Côte d’Ivoire (January 2004–December 2009).

†According the 2006 WHO definitions.

NA, not available.

Desmonde et al J Acquir Immune Defic Syndr ! Volume 62, Number 1, January 1, 2013

e16 | www.jaids.com � 2012 Lippincott Williams & Wilkins

(96%). Moreover, the 21 suspected TB cases led to 24 dif-ferent outpatient health care resources (114.3%), correspond-ing to 20 complementary diagnoses and 4 TB treatmentprescriptions.

When adjusted for age, immunodeficiency at baselineand primary caregiver, children presenting evidence of immu-nodeficiency and children with no immunologic follow-up

were less disposed to receive outpatient care (RR = 0.32, 95%CI: 0.18 to 0.56 and RR = 0.27, 95% CI: 0.13 to 0.56, respec-tively) compared with children with no evidence of immuno-deficiency. Moreover, we observed lower outpatient HCRU inchildren primarily cared for by both parents compared withthose whose primary caregiver was their mother alone (RR:0.34, 95% CI: (0.15 to 0.77).

TABLE 3. Inpatient Care Described According to the 371 Severe Morbid Events Occurring in 162 ART-Naive Children*

Severe Morbid Event Overall Any Outpatient Care†

Day Care

HospitalizationWith IV Therapy Without IV Therapy

Malaria, n (%) 135 (36.4) 69 (51.1) 74 (54.8) 5 (3.7) 7 (5.2)

Chronic lung disease, n (%) 78 (21.1) 75 (96.2) 15 (19.2) 1 (1.3) 2 (2.6)

Diarrhea, n (%) 53 (14.3) 18 (33.9) 10 (18.9) 1 (1.9) 3 (5.7)

Bacterial infection, n (%) 24 (6.5) 18 (75.0) 5 (20.8) — 3 (12.5)

Pulmonary tuberculosis, n (%) 21 (5.7) 24 (114.3)† — — —

Failure to thrive, n (%) 15 (4.1) 3 (20.0) 1 (6.7) — 2 (13.3)

Anaemia, n (%) 7 (1.9) 3 (42.9) 4 (57.1) — 2 (28.6)

Fever, n (%) 2 (0.5) — 4 (200.0)† — —

Other non-AIDS defining events, n (%) 15 (4.1) 5 (33.3) 2 (13.3) — 2 (13.3)

Other WHO Stage 4 events, n (%) 4 (1.1) 3 (75.0) — — 1 (25.0)

Unknown, n (%) 17 (4.6) 15 (88.2) 11 (64.7) 7 (41.2) 16 (94.1)

Total 371 (100) 233 (62.8) 126 (33.9) 14 (3.8) 38 (10.2)

*CePReF and the MTCT-Plus programme. Abidjan. Côte d’Ivoire (January 2004–December 2009).

†One event can lead to more than one resource to health care, leading to percentage .100%.

TABLE 2. Overall and Age at Baseline–Specific Incidence Density Rate per 100 CY (IR*) of Severe Morbidity, ComplementaryDiagnosis, Outpatient Care, and Inpatient Care in ART-Naive Children†

Morbidity and HCRU

Overall,

N = 405

Age ,1 yr,

N = 64

Age 1–2 yrs,

N = 42

Age 2–5 yrs,

N = 108

Age 5–10 yrs,

N = 142

Age 10–15 yrs,

N = 49

Number of CY of follow-up 642.3 93.0 58.7 181.5 271.4 37.7

At least one severe event, n (%) 162 (40) 13 (20.3) 12 (28.6) 43 (39.8) 70 (49.3) 24 (49.0)

Morbidity events, n (%) 371 (100) 27 (7.3) 30 (8.1) 115 (31.0) 163 (43.9) 36 (9.7)

IR* (95% CI) 60.9 (55.1 to 67.2) 29 (20 to 42.2) 51.1 (35.9 to 72.9) 63.4 (52.8 to 76.1) 60.1 (51.5 to 70) 95.5 (69.1 to 132.2)

Overall outpatient care, n (%) 212 (100) 23 (10.8) 15 (7.1) 72 (34.0) 82 (38.7) 20 (9.4)

IR* (95% CI) 33.0 (28.9 to 37.8) 24.7 (16.5 to 37.1) 25.6 (15.6 to 42.1) 39.7 (31.5 to 50) 30.2 (24.4 to 37.5) 53.1 (34.5 to 81.9)

Complementarydiagnoses, n (%)

144 (100) 17 (11.8) 9 (6.3) 43 (29.9) 55 (38.2) 20 (13.9)

IR* (95% CI) 22.4 (19.1 to 26.4) 18.3 (11.5 to 29.3) 15.3 (8.2 to 29.1) 23.7 (17.6 to 31.9) 20.3 (15.6 to 26.4) 53.1 (34.5 to 81.9)

Outpatient prescriptions, n (%) 98 (100) 12 (12.2) 6 (6.1) 37 (37.8) 43 (43.9) —

IR* (95% CI) 15.3 (12.5 to 18.6) 12.9 (7.4 to 22.5) 10.2 (4.8 to 22.2) 20.4 (14.8 to 28.1) 15.8 (11.8 to 21.3)

Antibiotics, n (%) 65 (100) 8 (12.3) 1 (1.5) 20 (30.8) 36 (55.4) —

IR* (95% CI) 10.1 (7.9 to 12.9) 8.6 (4.4 to 16.9) 1.7 (0.4 to 9.5) 11 (7.2 to 17) 13.3 (9.6 to 18.4)

Antimalarial, n (%) 39 (100) 4 (10.3) 4 (10.3) 18 (46.2) 13 (33.3) —

IR* (95% CI) 6.1 (4.5 to 8.3) 4.3 (1.8 to 11) 6.8 (2.8 to 17.4) 9.9 (6.3 to 15.7) 4.8 (2.8 to 8.2)

TB treatment, n (%) 4 (100) 1 (25.0) — 1 (25.0) 2 (50.0)

IR* (95% CI) 0.6 (0.3 to 1.6) 1.1 (0.3 to 6) 0.6 (0.1 to 3.1) 0.8 (0.2 to 2.7)

Overall inpatient care, n (%) 178 (100) 20 (11.2) 14 (7.9) 60 (33.7) 77 (43.3) 7 (3.9)

IR* (95% CI) 27.7 (23.9 to 32.1) 21.5 (13.9 to 33.2) 23.9 (14.3 to 40) 33.1 (25.7 to 42.6) 28.4 (22.7 to 35.5) 18.6 (9.2 to 38.3)

Inpatient day care, n (%) 140 (100) 10 (7.1) 10 (7.1) 50 (35.7) 67 (47.9) 3 (2.1)

IR* (95% CI) 21.8 (18.5 to 25.7) 10.8 (5.9 to 19.8) 17 (9.4 to 31.2) 27.6 (20.9 to 36.3) 24.7 (19.5 to 31.2) 7.9 (2.3 to 23.3)

Hospitalizations, n (%) 38 (100) 10 (26.3) 4 (10.5) 10 (26.3) 10 (26.3) 4 (10.5)

IR* (95% CI) 5.9 (4.3 to 8.1) 10.8 (5.9 to 19.8) 6.8 (2.8 to 17.4) 5.5 (3.1 to 10.1) 3.7 (2.1 to 6.8) 10.6 (4.3 to 27.2)

*Incidence rate, per 100 child-years.

†CePReF and the MTCT-Plus Programme, Abidjan, Côte d’Ivoire (January 2004–December 2009); n = 405.



Inpatient CareThe overall inpatient care IR (day care and hospitali-

zation) was 27.7 per 100 CY of follow-up, 95% CI: 23.9 to32.1. The estimated inpatient day care rate was 21.8 per 100CY of follow-up; the rate of hospitalization was 5.9 per 100CY (Table 2). There were no significant differences in inpa-tient care IR in the difference age groups (P = 0.49).

Inpatient day care was mostly accompanied by intra-venous therapy and many events led to more than 12 hours ofcare (percentages .100%). Overall, it was more frequentafter suspected malaria or lung disease (55% and 4% withand without IV therapy, respectively) (Table 3).

We had records for 38 hospitalizations; malaria,bacterial infections, and severe anemia were the leadingcauses. However, 16 of these hospitalizations were forundocumented reasons and represented 94% of the unknownmorbidity, most of which had already led to outpatient care orinpatient day care.

Factors associated with inpatient care are presented inTable 4. As for outpatient care, we observe significantly lowinpatient HCRU in children whose primary caregivers areboth parents compared with those primarily cared for by theirmothers only (RR = 0.30, 95% CI: 0.14 to 0.65). On the otherhand, inpatient care was significantly highest in children withno available CD4 data compared with those with documentedCD4 suggesting no signs of immunodeficiency at baseline(RR = 2.91, 95% CI: 1.66 to 5.11).

DISCUSSIONThis cohort study documents health care resource

utilization in HIV-1–infected children who had not yet under-gone ART initiation, in Abidjan, Côte d’Ivoire and who werefollowed up in a pediatric HIV care programme between 2004and 2009. In this context, we make 3 main observations. First,the severe morbidity rate is high, and highest in older

children, reaching 95.5/100 CY in children aged 10 yearsand older. Second, the overall coverage of cotrimoxazole pro-phylaxis at baseline is very low in children aged 1 year andyounger, reaching only 36%. Third, HCRU is defined bysevere morbidity and is not systematic, as only 57% of thesevere morbid events led to further investigation and/or treat-ment. HCRU was less frequent among children whose pri-mary caregivers were both parents compared with motheralone; outpatient care was higher in children not yet immu-nodeficient, whereas we observed an inverse pattern in inpa-tient care.

Severe morbidity in untreated HIV-infected children isearly and frequent. In previous work, we showed that the riskof developing a severe morbid event was not associated withimmune status, suggesting substantial morbidity attributable toother coinfections and the need for optimal prevention and carein untreated HIV-infected children.11 Indeed, these children arevulnerable to numerous serious opportunistic infections andinfectious morbidity; our study showed high rates of probablemalaria, lung disease, diarrheal disease, and tuberculosis. Thisis in agreement with well-documented high risks in sub-Sahar-an African HIV-infected children for malaria,27 tuberculosis,28

respiratory tract infections,12,13 and diarrheal disease.29 Prophy-lactic cotrimoxazole has been shown to be effective to helpprevent each one of these diseases30–32 and is an affordableintervention. However, even in an HIV health care programme,it is still not available to all, despite the fact it is recommendedfrom 6 weeks of age in HIV-exposed infants before HIV diag-nosis.26 These findings highlight the need for a greater utiliza-tion of diagnostic and therapeutic services, and mostimportantly, the operational difficulties health care centres facein the management of pediatric HIV.

Resources for diagnosis and treatment have becomeavailable through HIV/AIDS control programmes, but thereremains a deficiency in the pre-ART care of pediatric HIV;less than 50% of the severe morbidity triggered examinations

TABLE 4. Factors Associated With Resource to Health Care (Outpatient and Inpatient Care, Respectively) in ART-Naive Children

Covariables

Outpatient Care Inpatient Care

RR 95% CI P RR 95% CI P

Age, yrs 0.448 0.279

,1 1 — 1 —

1–2 0.50 0.15 to 1.63 0.53 0.18 to 1.57

2–5 1.01 0.38 to 2.67 1.10 0.50 to 2.45

.5 0.80 0.31 to 2.10 0.86 0.40 to 1.83

Immunosuppression* 0.009 0.008

No evidence 1 — 1 —

Evidence 0.32 0.18 to 0.59 0.83 0.47 to 1.47

CD4 count unknown 0.27 0.13 to 0.56 2.91 1.66 to 5.11

Principal caregiver 0.006 0.004

Mother alone 1 — 1 —

Father alone 0.66 0.33 to 1.33 0.54 0.28 to 1.05

Both parents 0.34 0.15 to 0.77 0.30 0.14 to 0.65

Other 1.09 0.47 to 2.52 1.01 0.44 to 2.29

CePReF and the MTCT-Plus Programme, Abidjan, Côte d’Ivoire (January 2004–December 2009); n = 405. RR indicates relative risk.

*According the 2006 WHO definitions.



or treatment. Radiology and blood analyses were the mostcommonly used diagnosis tools. We explain this lack of usemainly because of the costs for the families for more elaborateexaminations that would not be subsidized. Consequently,diagnoses are made relying on clinical symptoms or approx-imate diagnosis methods. Costs of HIV care could alsoexplain, in part, the low rate of HCRU: although bothinpatient day care and HIV treatments are subsidized by theACONDA programme, there still remains a cost for thefamilies to access these treatments that they may not alwaysbe able to pay. Moreover, much of the inpatient day careexceeded 48 hours, comparable to hospitalization, andprobably occurred in a context where actual hospitalizationwould have been more favorable if had been affordable. Wesuspect this is a common option that families may selectbecause inpatient day care is subsidized by the care pro-gramme, limiting patient costs but delaying care. AlthoughART remains free, other pediatric care must be paid for in partby patients/caregivers. As such, care is often not affordable,many patients experience poor retention in care, severeavoidable morbidity, and early mortality.

Outpatient care was less frequent in children who wereimmunodeficient at baseline. These children were already atan advanced stage of disease and were more likely to initiateART, leading to less observed time at risk. However, thesechildren were already at an advanced stage of disease andfaced a higher risk of death33; they may have been more likelyto die at home before they could reach health care facilities.We hypothesize that these children may account for a largerproportion of undocumented deaths and lost to follow-up thannonimmunodeficient children.34

We can explain the inverse observation in inpatient careby the severity of the event. Indeed, children with noavailable immunologic data tend to be children with verysevere clinical conditions and who die or are transferred tohospitals for long-term care.

HCRU was lower in children cared for by both parents.Indeed, having both parents as primary caregivers implies thatwhen medical decisions are necessary, both parents shouldconsent. Recent studies have pointed out the social barriersoften encountered in pediatric HIV care programmes whendisclosing the child’s and mother’s HIV status to thefather.35,36 The results observed in the “both parents” groupcan be explained by this, and we hypothesize that the father ofmost of these children is unaware of his child’s and possiblywife’s HIV status.

Our model does not allow observing the effect ofcotrimoxazole prophylaxis on HCRU. Indeed, coverage ofcotrimoxazole prophylaxis was correlated with age andsignificantly less frequent in infants 1 year or younger, whoconstitute a too small proportion of our cohort to allow themodel to converge. Nevertheless, we report incidences ofsevere morbidity and associated HCRU, showing substantialmorbidity attributable to other coinfectious diseases and theneed for pre-ART care despite cotrimoxazole prophylaxis.

There are 2 major limitations inherent in our retrospectivestudy design. First, our cohort is exposed to a left truncationbias: the less symptomatic children are the ones more likely tostill be alive and therefore our study is based on a selected

population of the more healthy untreated HIV-infected children.Our study population is comprised of children diagnosed ata later age, having consequently survived many events and whohave been included in an HIV programme where the health caresupport might have been better than that offered outside of theACONDA context. However, this survivorship bias is existentin many other studies.37,38

Second, severe morbidity and HCRU are likely under-documented in our data set. For example, the completion ofmedical charts may vary from one pediatrician to another andmay vary over time, and diagnoses could not be routinelyconfirmed using standardized diagnosic procedures. Weacknowledge that the reliability of clinical events is question-able; however, we attempted to ensure the accuracy of severemorbidity by requiring specific documentation, such asclinical findings, laboratory examinations, and radiographicresults. Social stigmatization of HIV among the Africanpopulation also leads to withheld information concerningevents during follow-up.39

Despite these considerations, our study provides orig-inal data on pre-ART severe morbidity and health careresource utilization in a large pediatric cohort, reflecting asbest as possible the current clinical practices in Côte d’Ivoire.Managing HIV in children remains challenging in an erawhere ART is scaling up.40,41 Free cotrimoxazole prophylaxisaccording to the WHO recommendations and other diagnosticand therapeutic interventions are still not available to all,leading to avoidable severe morbidity in those HIV-infectedchildren who have not yet initiated ART. This study tookplace in 2004–2009, during the rollout of ART in Côted’Ivoire. However, because of operational limitations, suchas delays in diagnosis or presentation to care, stigma associ-ated with HIV care, limited ART availability, and the require-ment for self-pay for a number of diagnostic and therapeuticservices that we observed, many missed opportunities to puteligible children on ART.

Although the WHO has revised its guidelines nowrecommending ART for all HIV-infected children 24 monthsor younger, the coverage in resource-limited settings remainslow.8 Identifying HIV-infected infants is a challenge in WestAfrica. In addition to the sophisticated and expensive techni-ques required for early infant diagnosis of HIV, PMTCTcoverage is low; PMTCT services are estimated to havereached only 54% of pregnant HIV-infected women in Côted’Ivoire in 2009. Although PMTCT interventions are scalingup today, this emphasises the need for efforts to strengthenthe link between PMTCT and childcare programmes. Improv-ing pre-ART care remains a priority in the management of thepediatric HIV epidemic and access to cotrimoxazole mustremain a priority in such a population.

Additional research must be undertaken in a pediatricpopulation, to better understand the severe morbidity in HIV-infected children and retention in care and associated costs toguide public health interventions, as it has recently been donein developed countries.42 The data presented here representa base case scenario against which we could compare morefrequently reported outcomes for children receiving ART,and we therefore believe that these data can inform currentdebates about the impact of universal ART initiation



(regardless of CD4) for children of various ages, particularlyin those older than 2 years. Furthermore, lifetime pre-ARTtreatment costs will probably be overweighed after ARTaccess, as pediatric HIV infection becomes a chronic diseaseleading to greater health care utilizations. This will have to beassessed and will highlight the cost effectiveness of prevent-ing MTCT in low-income countries.

ACKNOWLEDGMENTSThe authors would like to thank the children and their

families who participated in the ACONDA care pro-grammes. The authors are deeply grateful to all staffmembers of the PACCI and ACONDA teams (the CePReFand the MTCT-Plus programme) involved in the care ofHIV/AIDS patients in Abidjan. Each of these people hasduring the postelectoral crisis in 2011 courageously con-tinued to serve the sick, wounded, displaced, and therefugees—all victims of the war and to bring them medicine,food, water, support, and humanity. The authors would alsolike to thank the CEPAC group (Cost-Effectiveness for Pre-venting AIDS Complications) for their collaboration andinput in this study. Finally, a special thank you goes toDr X. Anglaret (INSERM U897 and Programme PACCI)for his helpful support.

REFERENCES1. Azcoaga-Lorenzo A, Ferreyra C, Alvarez A, et al. Effectiveness of a PMTCT

programme in rural Western Kenya. AIDS Care. 2011;23:274–280.2. Timité-Konan AM, Fassinou P, Adonis-Koffy L, et al. Use of antiretro-

viral drugs in HIV infected children in a tropical area: a real benefit.Archives De Pédiatrie. 2003;10:831–832.

3. Tonwe-Gold B, Ekouevi DK, Viho I, et al. Antiretroviral treatment andprevention of peripartum and postnatal HIV transmission in West Africa:evaluation of a two-tiered approach. PLoS Med. 2007;4:e257.

4. World Health Organization. UNAIDS Report on the Global AIDS Epi-demic 2010. Geneva, Switzerland: UNAIDS; 2010.

5. Anaky MF, Duvignac J, Wemin L, et al. Scaling up antiretroviral therapyfor HIV-infected children in Côte d’Ivoire: determinants of survival andloss to programme. Bull World Health Organ. 2010;88:490–499.

6. Coffie PA, Kanhon SK, Toure H, et al. Nevirapine for the prevention ofmother-to-child transmission of HIV: a nation-wide coverage survey inCôte d’Ivoire. J Acquir Immune Defic Syndr. 2011;57:3–8.

7. Kids-ART-Linc-Collaboration. Low risk of death, but substantial pro-gram attrition, in pediatric HIV treatment cohorts in Sub-Saharan Africa.J Acquir Immune Defic Syndr. 2008;49:523–531.

8. World Health Organization. Towards Universal Access: Scaling Up Pri-ority HIV/AIDS Interventions in the Health Sector. Progress Report2010. Geneva, Switzerland: UNAIDS; 2010.

9. Stringer EM, Ekouevi DK, Coetzee D, et al. Coverage of nevirapine-based services to prevent mother-to-child HIV transmission in 4 Africancountries. JAMA. 2010;304:293–302.

10. Nielsen K, Bryson YJ. Diagnosis of HIV infection in children. PediatrClin North Am. 2000;47:39–63.

11. Desmonde S, Coffie P, Aka E, et al. Severe morbidity and mortality inuntreated HIV-infected children in a paediatric care programme in Abid-jan, Côte d’Ivoire, 2004-2009. BMC Infect Dis. 2011;11:182.

12. Harambat J, Fassinou P, Becquet R, et al. 18-month occurrence of severeevents among early diagnosed HIV-infected children before antiretroviraltherapy in Abidjan, Côte d’Ivoire: a cohort study. BMC Public Health.2008;8:169.

13. Lucas SB, Peacock CS, Hounnou A, et al. Disease in children infectedwith HIV in Abidjan, Côte d’Ivoire. BMJ. 1996;312:335–338.

14. Toure S, Kouadio B, Seyler C, et al. Rapid scaling-up of antiretroviraltherapy in 10,000 adults in Côte d’Ivoire: 2-year outcomes and determi-nants. AIDS. 2008;22:873–882.

15. de Cherif TKS, Schoeman JH, Cleary S, et al. Early severe morbidity andresource utilization in South African adults on antiretroviral therapy.BMC Infect Dis. 2009;9:205.

16. Sutcliffe CG, van Dijk JH, Bolton C, et al. Effectiveness of antiretroviraltherapy among HIV-infected children in sub-Saharan Africa. LancetInfect Dis. 2008;8:477–489.

17. Gona P, Van Dyke RB, Williams PL, et al. Incidence of opportunisticand other infections in HIV-infected children in the HAART era. JAMA.2006;296:292–300.

18. World Health Organization. Antiretroviral Therapy for HIV Infection inInfants and Children: Towards Universal Access. Recommendations fora Public Health Approach. 2010 Revision. UNAIDS; 2010.

19. Ministry of Fight Against AIDS. Survey on HIV determinants in Côted’Ivoire (EIS-CI). Abidjan; 1997.

20. Ministry of Fight Against AIDS. HIV sentinel serosurveillance in preg-nant women in Côte d’Ivoire. Survey Report; Abidjan, 2004.

21. Ekouevi DK, Stringer E, Coetzee D, et al. Health facility character-istics and their relationship to coverage of PMTCT of HIV servicesacross four African Countries: The PEARL Study. PLoS One. 2012;7:e29823.

22. Tonwe-Gold B, Ekouevi DK, Bosse CA, et al. Implementing family-focused HIV care and treatment: the first 2 years’ experience of themother-to-child transmission-plus program in Abidjan, Côte d’Ivoire.Trop Med Int Health. 2009;14:204–212.

23. World Health Organization. WHO Case Definitions of HIV for Sur-veillance and Revised Clinical Staging and Immunological Classifi-cation of HIV-Related Disease in Adults and Children. Geneva,Switzerland: UNAIDS; 2007.

24. World Health Organization. Antiretroviral Therapy of HIV Infection inInfants and Children: Towards Universal Access. Recommendations fora Public Health Approach. Geneva, Switzerland: UNAIDS; 2006.

25. Zeger SL, Liang KY, Albert PS. Models for longitudinal data:a generalized estimating equation approach. Biometrics. 1988;44:1049–1060.

26. World Health Organization. Guidelines on Co-Trimoxazole Prophylaxisfor HIV-Related Infections Among Children, Adolescents and Adults.Recommendations for a Public Health Approach. Geneva, Switzerland:UNAIDS; 2006.

27. Imani PD, Musoke P, Byarugaba J, et al. Human immunodeficiency virusinfection and cerebral malaria in children in Uganda: a case-controlstudy. BMC Pediatr. 2011;11:5.

28. Walters E, Cotton MF, Rabie H, et al. Clinical presentation and outcomeof tuberculosis in human immunodeficiency virus infected children onanti-retroviral therapy. BMC Pediatr. 2008;8:1.

29. Thea DM, St Louis ME, Atido U, et al. A prospective study of diarrheaand HIV-1 infection among 429 Zairian infants. N Engl J Med. 1993;329:1696–1702.

30. Chintu C, Bhat GJ, Walker AS, et al. Co-trimoxazole as prophylaxisagainst opportunistic infections in HIV-infected Zambian children(CHAP): a double-blind randomised placebo-controlled trial. Lancet.2004;364:1865–1871.

31. Mulenga V, Ford D, Walker AS, et al. Effect of cotrimoxazole on causesof death, hospital admissions and antibiotic use in HIV-infected children.AIDS. 2007;21:77–84.

32. Gasasira AF, Kamya MR, Ochong EO, et al. Effect of trimethoprim-sulphamethoxazole on the risk of malaria in HIV-infected Ugandan chil-dren living in an area of widespread antifolate resistance. Malar J. 2010;9:177.

33. The Cross Continents Collaboration for Kids. Markers for predictingmortality in untreated HIV-infected children in resource-limited settings:a meta-analysis. AIDS. 2008;22:97–105.

34. Braitstein P, Songok J, Vreeman RC, et al. “Wamepotea” (they havebecome lost): outcomes of HIV-positive and HIV-exposed children lostto follow-up from a large HIV treatment program in western Kenya.J Acquir Immune Defic Syndr. 2011;57:e40–e46.

35. Donahue MC, Dube Q, Dow A, et al. “They have already thrown awaytheir chicken”: barriers affecting participation by HIV-infected women incare and treatment programs for their infants in Blantyre, Malawi. AIDSCare. 2012;24:1233–1239.



36. Hardon A, Vernooij E, Bongololo-Mbera G, et al. Women’s views onconsent, counseling and confidentiality in PMTCT: a mixed-methodsstudy in four African countries. BMC Public Health. 2012;12:26.

37. 3Cs4Kids. Markers for predicting mortality in untreated HIV-infectedchildren in resource-limited settings: a meta-analysis. AIDS. 2008;22:97–105.

38. Walenda C, Kouakoussui A, Rouet F, et al. Morbidity in HIV-1-Infectedchildren treated or not treated with highly active antiretroviral therapy(HAART), Abidjan, Cote d’Ivoire, 2000-04. J Trop Pediatr. 2009;55:170–176.

39. Orne-Gliemann J, Becquet R, Ekouevi DK, et al. Children andHIV/AIDS: from research to policy and action in resource-limited set-tings. AIDS. 2008;22:797–805.

40. Bland RM. Management of HIV-infected children in Africa: progressand challenges. Arch Dis Child. 2011;96:911–915.

41. Haldar P, S Reddy DC. Challenges in providing HIV care to paediatricage group in India. Indian J Med Res. 2009;129:7–10.

42. Sansom SL, Anderson JE, Farnham PG, et al. Updated estimates ofhealthcare utilization and costs among perinatally HIV-infected children.J Acquir Immune Defic Syndr. 2006;41:521–526.



87


We showed that untreated HIV-infected children require substantial inpatient and outpatient

care in a context where ART is scaling up but still not available to all. Despite an acceptable

coverage of cotrimoxazole prophylaxis observed in children aged > 2 years, this study

reported high severe morbidity rates, and in particular in children aged 10 years at

enrolment in the care programme, where the incidence of severe morbidity reached 95.5 per

100 child-years of follow-up after inclusion in the programme. As opposed to the previous

study where the time to first severe morbid event was longest in these older surviving

children, the overall incidence rate of severe morbidity is actually higher than in younger

children. Indeed, in the absence of ART, CD4 counts decline among these surviving children

considered as “slow progressors” at the beginning and they become more and more

vulnerable to opportunistic infections and infectious diseases as their immune system is more

damaged.

Although resources for prevention, diagnosis and treatment of opportunistic infections have

become available through HIV/AIDS control programmes, the principal findings of this study

underline a deficiency in pre-ART care of paediatric HIV and the operational difficulties

healthcare centres face in the management of the epidemic.

First, cotrimoxazole prophylaxis coverage was acceptable overall, but reached only 36% in

children aged < 12 months. Cotrimoxazole is an affordable and feasible intervention,

recommended in all children from 6 weeks of age since 2000. Reasons for this low coverage

remain to be investigated but healthcare provider’s reluctance because they fear resistance to

cotrimoxazole and their lack of training are suspected explanations.

Second, only 57% of the documented severe morbidity led to any kind of healthcare resource

utilisation. Less than 50% triggered examinations, often too costly for the families; the

Aconda programme subsidises ART and cotrimoxazole but more elaborate diagnostic tools

such as cultures or X-rays remain the responsibility of the patient’s family. This led to many

probable diagnoses and inappropriate treatment. Outpatient care was associated with good

health suggesting that the sickest children die without adequate access to care. Living with

both parents was also a factor associated with outpatient care underscoring the important role

of the caregiver in the management of paediatric HIV and the psycho-social factors in

attending health clinics. Inpatient care occurred mostly in the sickest, most immunodeficient

children and can be assimilated to a hospitalisation since children often came back the next

day to pursue intra venous treatment and remain under observation. As opposed to actual

overnight hospitalisation, inpatient day-care is subsidised by the programme, limiting costs

for families. Unfortunately, direct and indirect costs of pre-ART care could not be evaluated;

however resource to healthcare in non-ART-treated children is substantial and results suggest

that the possible direct costs for patients are a major barrier to care. As pre-ART care

improves and interventions to increase retention in these care programmes are implemented,

the resource to healthcare will be greater, as will the associated costs.

88

Finally, the major difficulty we encountered when addressing this research question was the

fact that children experienced more than one event and that events were non-independent in a

same child. Indeed, children who experienced severe morbidity were the sickest and therefore

the most likely to experience a subsequent event. We addressed this question by running a

Poisson regression allowing variance adjustment for this non-independence (198). Although

this study provided original and new data on pre-ART severe morbidity and healthcare

resource utilisation, the design was not ideal and we cannot account for time between events

or those children who die or who initiate ART and the non-independence of those outcomes

with severe morbidity.

Despite the limitations inherent in the study design, discussed in the paper, this study

represents a base case scenario against which outcomes of children receiving ART can be

compared. Although lifetime costs associated with pre-ART care are substantial, they will

probably be outweighed after access to ART as HIV becomes a chronic disease. This aspect is

essential in evaluating both the cost-effectiveness of preventing MTCT and EID strategies in

this setting.

89

5. REASONS FOR HOSPITALISATION

IN HIV-INFECTED CHILDREN, ART-

TREATED OR NOT, IN WEST AFRICA

90

5.1.Research question and context

Since the large implementation of ART in resource-limited settings, late diagnosis of HIV

infection and late initiation of ART are some of the current challenges that paediatric HIV

care programmes face, leading to a high prevalence of HIV-related morbidities. The prognosis

is worsened by the cost of the diagnostic tools and treatment to children’s families, who often,

as a result, choose not to resort to care, with a high risk of mortality or loss to follow-up (115,

199). Furthermore, the limited number of available ART combinations and paediatric

formulations and the use of poorly tolerated regimens may expose to frequent adverse effects

and to unplanned treatment interruptions also resulting in high morbidity. Finally, as HIV-

infected children on ART live longer than those not on ART, they may experience a wider

range of co-morbidities, including other infectious diseases but also cancer and cardiovascular

diseases (200-202), with a distribution that may differ from that described in high-income

settings. Knowledge is limited on causes of morbidity and mortality in HIV-infected patients

followed up in HIV care programmes. Available data come mainly from research protocols

and clinical trials which differ from routine care programmes by the selection of the

population and by means engaged in diagnostic tools. We hypothesised that hospitals may be

the destination of most HIV-infected children who present a severe clinical event, either for

the first time while they are unaware of their HIV status or during their regular follow-up in

ART-programmes or after a treatment interruption.

In a 6-month prospective multicentre survey within the IeDEA West Africa collaboration, we

investigated the morbidity and mortality in HIV-infected adults and children hospitalised in

participating tertiary hospital wards, described reasons of hospitalisations and described

socio-demographic, clinical, and biological characteristics of hospitalised children.

5.2. Dicko et al – Journal of International AIDS Society – in revision

The results of this study specifically conducted among children will be presented as a poster

at the 17th International Conference AIDS and STIs in Africa in December 2013 and are

currently being revised for publication in the Journal of International AIDS Society following

reviewers’ comments.

91

Reasons for hospitalisation in HIV-infected children in West Africa

Fatoumata Dicko1 MD, Sophie Desmonde2,3, §

MSc, Sikiratou Koumakpai4 MD, Hélène Dior-

Mbodj5MD, Fla Kouéta6

MD, Novisi Baeta7 MD , Niaboula Koné1 MD, Jocelyn Akakpo4 MD,

Haby Signate Sy5MD, PhD, Diarra Ye6

MD, PhD, Lorna Renner7MD, Charlotte Lewden2, 3

MD, PhD and Valériane Leroy2,3

MD, PhD for the Paediatric IeDEA West Africa Working

Group *

1 Service Pédiatrie Centre Hospitalier Universitaire Gabriel Toure, Bamako, Mali 2 Inserm, Centre Inserm U897 – Epidémiologie – Biostatistiques, F-33000 Bordeaux, France 3 Univ. Bordeaux, ISPED, Centre Inserm U897 – Epidémiologie – Biostatistiques, F-33000 Bordeaux, France 4 Service Pédiatrie, Centre National Hospitalier Universitaire, Cotonou, Bénin 5 Hopital pour Enfants Albert Royer, Dakar, Sénégal 6 Service Pédiatrie, Hopital Général de Gaulle, Ouagadougou, Burkina Faso 7 Korle Bu Teaching Hospital, Accra, Ghana

* See appendix

Corresponding author

[email protected]

Conflicts of interest and source of funding

The authors declare not conflicts of interest.

This study was supported in part by the National Cancer Institute (NCI), the Eunice Kennedy

Shriver National Institute of Child Health & Human Development (NICHD) and the National

Institute of Allergy and Infectious Diseases (NIAID) of the U.S. National Institutes of Health

(NIH), under Award Numbers U01AI069919 (Epidemiologic Databases to Evaluate AIDS

(IeDEA) West Africa. Sophie Desmonde is a fellow of the Ecole des Hautes Etudes en Santé

Publique (EHESP), Rennes, France. The content of this publication is solely the responsibility

of the authors and does not necessarily represent the official views of any of the institutions

mentioned above.

92

Abstract:

Introduction: Current knowledge on morbidity and mortality in HIV-infected children comes

from data collected in specific research programmes, which may offer a different standard of

care compared to routine care. We described hospitalisation data within a large observational

cohort of HIV-infected children in West Africa (IeDEA West Africa collaboration).

Methods: We performed a six-month prospective multicentre survey from April to October

2010 in five HIV specialised paediatric hospital wards in Ouagadougou, Accra, Cotonou,

Dakar and Bamako. Baseline and follow-up data during hospitalisation were recorded using a

standardized clinical form, and extracted from hospitalisation files and local databases.

Diagnoses were reviewed within each centre by event validation committees. HIV-related

events were defined according to the WHO definitions.

Results: From April to October 2010, 155 HIV-infected children were hospitalised; median

age was 3 years [1 – 8]. Among them, 90 (58%) were confirmed for HIV-infection during

their stay; 138 (89%) were already receiving cotrimoxazole prophylaxis and 62 children

(40%) had initiated antiretroviral therapy (ART). The median length of stay was 13 days

(IQR: 7-23); 25 children (16%) died during hospitalisation and four (3%) were transferred

out. The leading causes of hospitalisation were WHO stage 3 opportunistic infections (37%),

non-AIDS defining events (28%), cachexia and other WHO stage 4 events (25%).

Conclusions: Overall, most causes of hospitalisations were HIV-related but one

hospitalisation in three was caused by a non AIDS-defining event, mostly in children on ART.

HIV-related fatality is also high despite the scaling up of access to ART in resource-limited

settings.

Keywords: HIV, paediatrics, hospitalisation, Africa

93

Introduction

In 2010, the HIV/AIDS epidemic led to 1.8 million deaths worldwide, 14 % of which

occurred in children below 15 years of age (1). Although substantial progress has been made

in the management of paediatric HIV infection with the implementation of mother-to-child-

prevention programmes and subsequently the introduction of antiretroviral therapy (ART), an

estimated 3.3 million children are still living with HIV, of whom 3.1 million live in sub-

Saharan Africa (1, 2). Before highly active antiretroviral therapy (ART) was introduced,

HIV-infected infant mortality was high and premature, reaching 35% in the first 12 months of

life and 52% within 2 years (3).Since 2004, following the introduction of ART, reductions in

mortality and morbidity among HIV-infected children have been reported (4-8). Nevertheless,

the management of HIV-infected children remains challenging in resource-limited settings,

particularly in West-Africa (9). First, the accurate identification of HIV-exposed children is

largely insufficient (10). Second, early infant diagnosis requires expensive and sophisticated

molecular diagnosis techniques (11); consequently, children are diagnosed at a late age, with

an advanced clinical and/or immunological stage of HIV disease. Moreover, several studies

have reported that advanced immunosuppression at ART initiation increases the risk of

immune reconstitution syndrome (12, 13). Third, paediatric formulations are limited and the

use of poorly tolerated regimens may expose to adverse events and treatment interruptions.

Fourth, severe morbidity in HIV-infected children is frequent whether they are ART-treated

or not (14, 15). Finally, in case of occurrence of morbidity, resources to healthcare are often

limited for financial reasons: besides ART, which is freely delivered to children, the costs of

diagnosis tools and treatment for co-morbid events are borne by the children’s families (16).

Current knowledge on morbidity and mortality comes mainly from data collected in specific

research programmes, which may offer a different standard of care than that offered outside of

a research context, where means for diagnosis and treatment are more limited. Hospitalisation

represents an important outcome measure, as a reliable indicator for morbidity and mortality,

as well as healthcare utilisation.

In this study, we described hospitalisation in tertiary hospital wards from a large prospective

observational cohort of HIV-infected and HIV-exposed children in the paediatric West

African IeDEA Collaboration.

Methods

Setting

The International epidemiological Database to Evaluate AIDS (IeDEA) network

(http://www.iedea.org/) was initiated in July 2006 to address evolving questions in HIV/AIDS

research, unanswerable by single cohorts, by collecting and harmonizing data from multiple

HIV/AIDS cohorts and programmes. In 2012, the paediatric IeDEA West Africa collaboration

(pWADA) involved 11 paediatric HIV/AIDS clinics spread out over seven countries in West

Africa.

Health facilities in the IeDEA network deliver a model of paediatric care covering HIV

diagnosis and treatment, including ART and prophylaxis of opportunistic infections.

94

Cotrimoxazole prophylaxis is delivered to all HIV-exposed children according to the WHO

recommendations. Children initiated ART according to the 2006 WHO recommendations,

contemporaneous with the practices at time of ART initiation (17): all HIV-infected infants

diagnosed < 12 months, irrespective of CD4 count or clinical stage, or according to clinical

and immunological criteria in older children. Furthermore, children aged < 18 months who

meet presumptive diagnosis criteria were eligible for ART initiation.

Design

We performed a six-month cross-sectional multicentre survey from April to October 2010 in

five HIV specialised paediatric hospital wards involved in the pWADA IeDEA collaboration:

the Gabriel Touré Teaching Hospital (Bamako, Mali), the National Teaching Hospital

(Cotonou, Benin), the Albert Royer Children’s Hospital (Dakar, Senegal), the Général de

Gaulle Teaching Hospital (Ouagadougou, Burkina Faso) and the Korle Bu Teaching Hospital

(Accra, Ghana). All newly hospitalised HIV-infected children 18 years, previously known

as HIV-infected patients (HIV-1, HIV-2 or HIV-1+2) or diagnosed as HIV-infected during

hospitalisation were eligible, regardless of ART.

Data collection

Baseline and follow-up information during hospitalisation were recorded using a standardized

clinical form. Data were extracted from hospitalisation medical records and local databases.

The following variables were recorded: social-demographic data (gender, data of birth, town,

parents’ vital status, schooling); data of HIV diagnosis, type and circumstances of diagnosis,

tuberculosis history; ART history (date of ART initiation, regimen, weight, height, WHO

stage, CD4 and viral load - when available – current ART regimen); current prophylaxis ; co-

morbidity history, immunization status; date of hospitalisation, cause, date of first symptoms,

WHO stage, weight, height, WHO stage, CD4 count and viral load - when available –

complete blood counts; diagnosis and type of diagnosis (clinical or presumptive), degree of

severity; date of end of hospitalisation, vital status at exit, underlying cause of death in that

event. In case of hospitalisation, the available diagnosis tools were provided to children’s

family.

Clinical diagnosis

Diagnoses were reviewed within each centre by ad-hoc event validation committees. HIV-

related events were defined according to the WHO definitions (18). Immune Reconstitution

Inflammatory Syndrome (IRIS) was defined by the following criteria: HIV-infected, receiving

ART with decreasing HIV-1 RNA levels and increasing CD4 cell counts since baseline,

presenting clinical symptoms consistent with IRIS and a clinical evolution inconsistent with

expected outcomes of previously diagnosed morbidity or drug toxicity. Causality assessment

of side effects (19) was defined with simplified categories adapted from the WHO-UMC

causality categories (20). Three categories were defined to describe the causality: (i) yes

(reasonable time relationship to drug intake); (ii) no (time to drug intake inconsistent or other

drugs can provide plausible explanations) and (iii) unknown (insufficient or contradictory

information).

95

In order to determine one underlying cause for each hospitalisation, the following priorities

were applied if a patient had multiple diagnoses: (1) WHO stage 4 opportunistic diseases, (2)

WHO stage 3 opportunistic diseases, (3) other non-AIDS defining infections, (4) other

cancerous events, (5) WHO stage 4 cachexia, (6) other diseases and (7) non-specific events

including non-specific WHO stage 3 events (moderate malnutrition, persistent diarrhoea,

persistent fever).

When a medical etiological diagnosis for hospitalisation could not be identified, we reviewed

diagnoses that comprised symptoms or clinical signs and attributed the hospitalisation to

those. Finally, we grouped the prioritised diagnoses associated with hospitalisation into the

four following categories: WHO stage 4 events (including cachexia); WHO stage 3

opportunistic infections; other non AIDS-related infections; and other diseases or non-specific

diseases (including non-specific WHO stage 3 events: moderate malnutrition, persistent

diarrhoea, persistent fever). To determine the underlying cause of death, we used the

International Classification of Diseases, Tenth Revision rules (21) and defined it as the

disease or injury that initiated the morbid event(s) leading to death.

Statistical analysis

We described the patterns and causes of hospitalisation in children according the WHO

staging, to cotrimoxazole prophylaxis and ongoing ART at the time of hospitalisation.

Categorical data are presented as frequencies (percentage) and continuous variables as median

[interquartile range (IQR)]. Proportions were compared by Fisher’s exact tests

Results

Population characteristics

From April to October 2010, 155 HIV-infected children were hospitalised once in one of the

five participating tertiary hospitals in West Africa. Their main characteristics at entry in

hospital are presented in table 1, according to the underlying cause of hospitalisation. Their

median age at hospitalisation was 3 years [interquartile range (IQR) 1-8 years] and the sex

ratio was 1. Only 65 children (42%) had been diagnosed with HIV prior to hospitalisation.

Among these children, the median time since HIV diagnosis until hospitalisation was 7

months.

Overall, 99 children (64%) had reached WHO Stage 4 of the disease, 138 (89%) were already

receiving cotrimoxazole and 62 children (40%) had initiated ART. The median CD4

percentage at ART initiation was 14% (IQR: 6-21).

Hospitalisations & diagnoses

The median length of stay was 13 days (IQR: 7-23 days); 25 children (16%) died during

hospitalisation and four (3%) were transferred out to another ward or hospital. HIV-infection

was confirmed during hospitalisation in the 90 children (58%) whose HIV-status was

unknown at entry, 19% of whom initiated ART before hospital discharge. Of these children,

50% were aged> 2 years and 64% presented AIDS-defining morbidity.

The most frequent causes of hospitalisations were: WHO stage 3 opportunistic infections

(37%), other non AIDS-defining infections (28%), WHO stage 4 events including cachexia

96

(25%) and other diseases including WHO stage 3 non-specific diseases (10%). Among the

non-AIDS defining infections, the most frequent diagnoses were the following: infectious

diarrhoea (31%), malaria (24%) and lung infectious pneumopathy (15%). Two

hospitalisations were related to suspected adverse events to ART: one child presented severe

anaemia and another, a suspected muco-cutaneous disease.

Many children were hospitalised with a second or third diagnosis associated with the main

cause for hospitalisation, yielding to a total of 305 diagnoses, as described in table 2.

Overall, 54.8% of the diagnoses were attributable to infectious diseases, 43.9% to non-

specific diseases and 1.3% to other diseases. Among infectious diseases, 32.9% were non-

AIDS defining events. When considering all 305 diagnoses, the five most frequent were the

following: pneumonia (18%), cachexia (12.5%), underweight or malnutrition (11.8%),

anaemia (9.8%) tuberculosis (6.6%); furthermore, 4.3% was malaria. Overall, 27.5% of the

morbidity concerned non AIDS-defining events.

The majority of the diagnoses were definite (81.6%). Of the 50 presumptive diagnoses, 40%

were suspected cases of tuberculosis (of the 20 reported tuberculosis events, only two were

definite according the WHO classification), and 6% were suspected HIV encephalopathy (one

case out of four was definite).

Table 3a presents the causes of hospitalisation by cotrimoxazole prophylaxis. We observed no

significant difference in the overall distribution of AIDS defining and non-defining events in

children who were on cotrimoxazole compared to those off (p = 0.21). We observed a

significantly higher proportion of children hospitalised for cachexia among those who were

not receiving cotrimoxazole (47.1%) compared to those on cotrimoxazole (21.7%) (p = 0.03).

In table 3b, we compare the causes of hospitalisation in children according to ART initiation

and its duration; we observed a higher proportion of AIDS-defining events (WHO stages 3 &

4) in children who were not on ART compared to those who had initiated ART (p= 0.010).

When studying specifically the six most frequent events, we observed higher proportions of

children hospitalised for pneumonia and for cachexia in children not on ART compared to

those who had initiated ART (p = 0.012 and p= 0.005, respectively). Inversely, there was an

overall higher proportion of non AIDS-defining events in those on ART compared to those

not; the proportion of children hospitalised for malaria was significantly higher in those who

had initiated ART beyond 6 months compared to others (p = 0.015). Furthermore, we

observed a higher proportion of cachexia in children who had initiated ART within 6 months

prior to hospitalisation, compared to others (p = 0.005).

When comparing the overall causes by treatment groups (cotrimoxazole only, ART only or

both), we observed a higher proportion of WHO stage 3 events in those on cotrimoxazole

only compared to those who had also initiated ART (p = 0.001). There was no significant

difference in the distribution of causes of hospitalisation in children on ART only compared

to those on ART and cotrimoxazole (p = 0.702).

Among the 25 children who died during hospitalisation, the most frequent underlying causes

of death were: cachexia (n=6), acute respiratory disease (n=6) and chronic diarrhoea (n=3),

(Table 4).

97

Discussion

This cross-sectional study documents the causes of hospitalisation in HIV-infected children in

West Africa. Overall, 62% of causes were AIDS-related, highlighting the advanced stage of

HIV-disease in these children at the time of hospitalisation. Furthermore, we showed that

among all the events, 54.8% were in relation with an infection, underlining a context of

residual infectious morbidity, despite cotrimoxazole prophylaxis and ART initiation. Indeed,

the proportion of AIDS-defining causes of hospitalisation was lower in children who were

both on ART and cotrimoxazole. However, these children evidently still experience a

substantial amount of non AIDS-defining events serious enough to cause hospitalisation.

Many studies in resource-limited settings described hospitalisations in HIV-infected children,

but tended to be restricted to cause-specific hospitalisations or focused on HIV-exposed rather

than HIV-infected children (22-27). In this study, we provide an accurate description of all the

severe morbidity (with 82% of definite diagnoses) leading to hospitalisation in HIV-infected

children. This survey was conducted in tertiary hospitals, which provided better specificity of

diagnoses than those obtained in smaller healthcare facilities where diagnosis tools are not

always available (15, 28).

Overall, 72.5% of events were AIDS and 27.5% non AIDS defining. Among AIDS defining

events, respiratory diseases, particularly pneumonia constituted a large proportion of the

causes of hospitalisations, significantly larger in those who were not on ART. The burden of

respiratory manifestations in HIV-infected children is well documented and our results are

consistent with those of previous studies, in both high-resource and limited-resource settings

(23, 29), including a study in Thailand where it was reported that 30% of hospitalisations

were due to pneumonia (30). In addition, a study conducted in Côte d’Ivoire showed that one

third of all morbidity events in HIV-infected children were respiratory manifestations, which

then decreased dramatically with the introduction of ART (31); in our study, we observed

lower proportions of pneumonia in children who had initiated ART. On the other hand,

despite many studies reporting a protective effect of cotrimoxazole on the occurrence of

pulmonary diseases (32), we observed no difference in those who had initiated cotrimoxazole

prophylaxis compared to those not. This could be explained by an information bias:

cotrimoxazole may have prescribed to the sickest and therefore the most vulnerable to

pulmonary diseases compared to those not on cotrimoxazole.

Furthermore, almost half the hospitalisations were caused by cachexia or underweight. This is

consistent with previously published data on HIV/AIDS and malnutrition. A cross-sectional

study led in Kenya reported the prevalence of malnutrition in HIV-infected children to reach

40% in 2008 (33). Another study, also led in Kenya, reported malnutrition to be the

underlying cause of more than half of the inpatient morbidity in HIV-infected children (34) .

We observed in our study higher proportions of cachexia in children who had initiated ART

within 6 months compared to those on ART for a longer period of time. However, the

proportion of children hospitalised for malnutrition remained highest in those who had not yet

initiated ART and who were not on cotrimoxazole either. Similar observations have been

reported elsewhere, in the ARROW open label randomised trial led in Uganda and

98

Zimbabwe, authors reported that one in nine children with advanced HIV required

hospitalisation for severe malnutrition after ART initiation (26). Furthermore, a recent study

within the CHAP randomised placebo-controlled trial, lower decreases in weight-for-age and

height-for-age scores in children receiving prophylaxis compared to placebo group were

reported (35).

The third leading cause of hospitalisation in this study is malaria. Previous studies have

reported protective effects of cotrimoxazole when studying HIV and malaria co-infection

(36); a recent systematic review reported that cotrimoxazole in HIV infected individuals

protects against malaria (37). We do not observe this in hospitalised HIV-infected children.

Also, other studies report a role in the ART regimen, and possible protective effect of

Lopinavir- based ART (38, 39). Although we observed a higher proportion of malaria among

children on ART > 6 months, there was no significant difference between those treated and

those not overall. Unfortunately we do not have the sufficient data to compare the proportion

of malaria-induced hospitalisations by ART regimen.

Overall, the distribution of causes of hospitalisation did not differ between those on

cotrimoxazole and those not. This may reflect the fact that cotrimoxazole in West Africa is

not implemented consistently despite the clinical practice guidelines. On the other hand, the

spectrum of diagnoses associated with hospitalisation changed whether the patient was on or

off ART. Previous studies have reported decreases in the number of hospitalisations due to

common AIDS-defining diseases in the ART era (30, 40), compensated by the appearance of

non AIDS-defining diseases. This indicates that patients are developing other comorbidities in

a context where residual infectious morbidity is prevalent and that ART may have reached a

limit in its effectiveness in reducing clinical morbidity.

The median age at hospitalisation was 3 years and we showed that hospitalisation was an

opportunity to diagnose 58% of HIV cases in this study. This reflects the flaws in the

prevention of mother-to-child transmission of HIV (PMTCT) services in resource-limited

settings and many missed opportunities for early HIV postnatal diagnosis and care. Indeed, a

large proportion of HIV-infected infants never enters the pathway to early HIV care and is

diagnosed at a late age, once symptomatic, initiating ART too late. This leads to severe

morbidity and mortality (7, 14, 41, 42). Although early infant diagnosis is challenging in

resource-limited settings for many reasons, including costs (43), it provides substantial

benefits to HIV-exposed children and their families (44). More efforts should be made when

implementing early infant diagnosis in West Africa (45). Care interventions aiming at

identifying the earliest possible HIV-infected infants and treating them according to current

guidelines would drastically reduce costs of HIV care. Indeed, previous studies suggest that

hospitalisation constitutes 15 – 49% of total HIV care, particularly in children at most

advanced disease stage (46, 47). Unfortunately, we do not have cost data in the current IeDEA

context to compare. However, a previous study in Côte d’Ivoire reported that severe

morbidity (diagnoses, treatments and hospital stays) represented $8.8 per patient, equivalent

to 50% of patient expenditures (16). In a context where costs are often the barrier to better

care (15), it is essential to develop effective care strategies and identify HIV-infected children

99

the earliest possible. Studies in Thailand and South Africa have shown that longer pre-ART

care or early ART initiation incur lower inpatient care costs (30, 48). Early and universal free

access to ART services would improve paediatric outcomes and costs impact at programme

level (9).

Our study has, however, several drawbacks. As we have discussed previously, social barriers

and stigma associated with HIV/AIDS often lead to delays in accessing care and in receipt of

HIV diagnosis. In the absence of ART, mortality is severe amongst infants < 2 years (3),

which combined with the late diagnoses issues, induces a selection bias in our study

population. Many perinatally infected children may have died before diagnosis and treatment,

underestimating the toll of hospitalisations we observe in this six-month period. Second, in

the case where children were diagnosed and then treated, we fail to clearly identify which

came first, ART or severe morbidity. We advise caution in the interpretation of our results in

ART-treated children; some events may not be solemnly residual morbidity in a resource-

limited setting. Finally, missing data arose as a result of the reality of healthcare services in

the field in Africa and the lack of follow-up in clinical examinations did not allow us to

confirm all diagnoses, over-estimating in some cases the burden of certain diseases.

Conclusion

In summary, our study provides a thorough description of causes of hospitalisation in HIV-

infected children, according to specific gold standard definitions. In a context where

healthcare services lack, we have described here, as best as possible, severe morbidity leading

to hospitalisation in HIV-infected children. Overall, one hospitalisation in two is related to

infectious diseases and one hospitalisation in three was caused by a non AIDS-related event;

this proportion was higher in children on ART, reflecting a context of prevalent severe

morbidity, that may be beyond the threshold of ART-effectiveness. In addition to ART,

efforts must continue to focus towards increasing the offer of cotrimoxazole prophylaxis for

all in order to reduce effectively clinical morbidity in HIV-infected children.

Acknowledgements

The authors would like to thank all the participating children and their families, as well as all

the members of the hospital teams of the sites involved in the IeDEA West Africa paediatric

cohort.

100

References

1 World Health Organisation. UNAIDS report on the global AIDS epidemic 2011. UNAIDS. 2011. 2 Stringer EM, Ekouevi DK, Coetzee D, et al. Coverage of nevirapine-based services to prevent mother-to-child HIV transmission in 4 African countries. JAMA: The Journal of the American Medical Association. 2010;304(3):293-302. 3 Newell M-L, Coovadia H, Cortina-Borja M, Rollins N, Gaillard P, Dabis F. Mortality of infected and uninfected infants born to HIV-infected mothers in Africa: a pooled analysis. Lancet. 2004;364(9441):1236-43. 4 Arrivé E, Kyabayinze DJ, Marquis B, et al. Cohort Profile: The Paediatric Antiretroviral Treatment Programmes in Lower-Income Countries (KIDS-ART-LINC) Collaboration. International Journal of Epidemiology. 2008;37(3):474-80. 5 Ciaranello AL, Chang Y, Margulis AV, et al. Effectiveness of paediatric antiretroviral therapy in resource-limited settings: a systematic review and meta-analysis. Clin Infect Dis. 2009 Dec 15;49(12):1915-27. 6 Collins IJ, Jourdain G, Hansudewechakul R, et al. Long-term survival of HIV-infected children receiving antiretroviral therapy in Thailand: a 5-year observational cohort study. Clinical Infectious Diseases: An Official Publication of the Infectious Diseases Society of America. 2010;51(12):1449-57. 7 Edmonds A, Yotebieng M, Lusiama J, et al. The Effect of Highly Active Antiretroviral Therapy on the Survival of HIV-Infected Children in a Resource-Deprived Setting: A Cohort Study. PLoS medicine. 2011;8(6):e1001044-e. 8 Sutcliffe CG, van Dijk JH, Bolton C, Persaud D, Moss WJ. Effectiveness of antiretroviral therapy among HIV-infected children in sub-Saharan Africa. The Lancet Infectious Diseases. 2008;8(8):477-89. 9 Leroy V, Malateste K, Rabie H, et al. Outcomes of antiretroviral therapy in children in Asia and Africa: a comparative analysis of the IeDEA paediatric multiregional collaboration. J Acquir Immune Defic Syndr. 2013 Feb 1;62(2):208-19. 10 Manzi M, Zachariah R, Teck R, et al. High acceptability of voluntary counselling and HIV-testing but unacceptable loss to follow up in a prevention of mother-to-child HIV transmission programme in rural Malawi: scaling-up requires a different way of acting. Trop Med Int Health. 2005 Dec;10(12):1242-50. 11 Nielsen K, Bryson YJ. Diagnosis of HIV infection in Children. Paediatric Clinics of North America. 2000;47(1):39-63. 12 Orikiiriza J, Bakeera-Kitaka S, Musiime V, Mworozi EA, Mugyenyi P, Boulware DR. The clinical pattern, prevalence, and factors associated with immune reconstitution inflammatory syndrome in Ugandan children. AIDS. 2010 Aug 24;24(13):2009-17. 13 Smith K, Kuhn L, Coovadia A, et al. Immune reconstitution inflammatory syndrome among HIV-infected South African infants initiating antiretroviral therapy. AIDS. 2009 Jun 1;23(9):1097-107. 14 Desmonde S, Coffie P, Aka E, et al. Severe morbidity and mortality in untreated HIV-infected children in a paediatric care programme in Abidjan, Côte d'Ivoire, 2004-2009. BMC infectious diseases. 2011;11(1):182-. 15 Desmonde S, Coffie PA, Aka EA, et al. Healthcare resource utilisation in untreated HIV-infected children in a paediatric programme, Abidjan, Cote d'Ivoire, 2004-2009. J Acquir Immune Defic Syndr. 2013 Jan 1;62(1):e14-21.

101

16 Beaulière A, Touré S, Alexandre P, et al. The financial burden of morbidity in HIV-infected adults on antiretroviral therapy in Côte d'Ivoire. 5th Internationnal AIDS Society (IAS) Conference on HIV Pathogenesis, Treatment and Prevention, Cape Town, South Africa. 2009;Abstract MOPED035. 17 World Health Organisation. Antiretroviral Therapy of HIV infection in infants and children: towards universal access. Recommendations for a public health approach. 2006. UNAIDS. 2006. 18 World Health Organisation. WHO case definitions of HIV for surveillance and revised clinical staging and immunological classification of HIV-related disease in adults and children. UNAIDS. 2007. 19 World Health Organisation. Safety of Medicines. A guide to detecting and reporting adverse drug reactions. Geneva, World Health Organisation. 2002. 20 The Uppsala Monitoring Centre. The use of the WHO-UMC system for standardised case causality assessment. 2009 [cited; Available from: http://www.who-umc.org/graphics/4409.pdf 21 World Health Organisation. International Classification of Diseases. Tenth Revision ed, Geneva. World Health Organisation. 1993. 22 Singh HK, Gupte N, Kinikar A, et al. High Rates of All-cause and Gastroenteritis-related Hospitalisation Morbidity and Mortality among HIV-exposed Indian Infants. BMC infectious diseases. 2011;11(1):193-. 23 Madhi SA, Kuwanda L, Cutland C, Klugman KP. Five-year cohort study of hospitalisation for respiratory syncytial virus associated lower respiratory tract infection in African children. Journal of Clinical Virology: The Official Publication of the Pan American Society for Clinical Virology. 2006;36(3):215-21. 24 Berkley JA, Lowe BS, Mwangi I, et al. Bacteremia among children admitted to a rural hospital in Kenya. The New England Journal of Medicine. 2005;352(1):39-47. 25 Chhagan MK, Kauchali S. Comorbidities and mortality among children hospitalized with diarrheal disease in an area of high prevalence of human immunodeficiency virus infection. The Paediatric Infectious Disease Journal. 2006;25(4):333-8. 26 Prendergast A, Bwakura-Dangarembizi MF, Cook AD, et al. Hospitalisation for severe malnutrition among HIV-infected children starting antiretroviral therapy. AIDS (London, England). 2011;25(7):951-6. 27 Bobossi-Serengbe G, Tembeti PJ, Mobima T, Yango F, Kassa-Kelembho E. [Tuberculosis and HIV co-infection among children hospitalized in Bangui (Central African Republic)]. Archives de pédiatrie : organe officiel de la Sociéte française de pédiatrie. 2005;12(8):1215-20. 28 Horwood C, Vermaak K, Rollins N, Haskins L, Nkosi P, Qazi S. Paediatric HIV management at primary care level: an evaluation of the integrated management of childhood illness (IMCI) guidelines for HIV. BMC Paediatrics. 2009;9:59-. 29 Graham SM, Gibb DM. HIV disease and respiratory infection in children. British Medical Bulletin. 2002;61:133-50. 30 Collins IJ, Cairns J, Jourdain G, et al. Hospitalisation trends, costs, and risk factors in HIV-infected children on antiretroviral therapy. AIDS. 2012 Sep 24;26(15):1943-52. 31 Kouakoussui a, Fassinou P, Anaky MF, et al. Respiratory manifestations in HIV-infected children pre- and post-HAART in Abidjan, the Ivory Coast. Paediatric respiratory reviews. 2004;5(4):311-5. 32 Chintu C, Bhat GJ, Walker AS, et al. Co-trimoxazole as prophylaxis against opportunistic infections in HIV-infected Zambian children (CHAP): a double-blind randomised placebo-controlled trial. Lancet. 2004;364(9448):1865-71.

102

33 Berger MR, Fields-Gardner C, Wagle A, Hollenbeck CB. Prevalence of malnutrition in human immunodeficiency virus/acquired immunodeficiency syndrome orphans in the Nyanza province of Kenya: a comparison of conventional indexes with a composite index of anthropometric failure. Journal of the American Dietetic Association. 2008;108(6):1014-7 34 Bejon P, Mohammed S, Mwangi I, et al. Fraction of all hospital admissions and deaths attributable to malnutrition among children in rural Kenya. The American Journal of Clinical Nutrition. 2008;88(6):1626-31. 35 Prendergast A, Walker aS, Mulenga V, Chintu C, Gibb DM. Improved growth and anemia in HIV-infected African children taking cotrimoxazole prophylaxis. Clinical infectious diseases : an official publication of the Infectious Diseases Society of America. 2011;52(7):953-6. 36 Gasasira AF, Kamya MR, Ochong EO, et al. Effect of trimethoprim-sulphamethoxazole on the risk of malaria in HIV-infected Ugandan children living in an area of widespread antifolate resistance. Malar J. 2010;9:177. 37 Manyando C, Njunju EM, D'Alessandro U, Van Geertruyden JP. Safety and Efficacy of Co-Trimoxazole for Treatment and Prevention of Plasmodium falciparum Malaria: A Systematic Review. PloS One. 2013;8(2):e56916. 38 ART regimen protects children from malaria. BMJ. 2012;345:e8214. 39 Achan J, Kakuru A, Ikilezi G, et al. Antiretroviral agents and prevention of malaria in HIV-infected Ugandan children. N Engl J Med. 2012 Nov 29;367(22):2110-8. 40 Sudjaritruk T, Oberdorfer P, Puthanakit T, Sirisanthana T, Sirisanthana V. Causes of first hospitalisation among 1121 HIV-infected children: comparison of the pre-Pneumocystis jiroveci pneumonia prophylaxis, pre-antiretroviral therapy and antiretroviral therapy periods. International journal of STD & AIDS. 2012 May;23(5):335-9. 41 Anaky MF, Duvignac J, Wemin L, et al. Scaling up antiretroviral therapy for HIV-infected children in Côte d'Ivoire: determinants of survival and loss to programme. Bulletin of the World Health Organization. 2010;88(7):490-9. 42 De Beaudrap P, Boulle C, Lewden C, et al. Morbidity After Antiretroviral Therapy Initiation in HIV-1 Infected Children in West Africa: Temporal Trends and Relation to CD4 Count. Pediatr Infect Dis J. 2012 Oct 24. 43 Ciaranello AL, Park J-E, Ramirez-Avila L, Freedberg KA, Walensky RP, Leroy V. Early infant HIV-1 diagnosis programs in resource limited settings: opportunities for improved outcomes and more cost-effective interventions. BMC Medicine. 2011;9(1):59-. 44 Hsiao NY, Stinson K, Myer L. Linkage of HIV-Infected Infants from Diagnosis to Antiretroviral Therapy Services across the Western Cape, South Africa. PloS One. 2013;8(2):e55308. 45 Ndondoki C, Brou H, Timite M, et al. Universal HIV Screening at Postnatal Points of Care: Which public health approach for early infant diagnosis in Côte d'Ivoire ? PLosOne. 2013;in press. 46 Sansom SL, Anderson JE, Farnham PG, et al. Updated estimates of healthcare utilisation and costs among perinatally HIV-infected children. J Acquir Immune Defic Syndr. 2006 Apr 1;41(4):521-6. 47 Wilson LS, Basu R, Christenson M, et al. Paediatric HIV costs across three treatment eras from 1986 to 2007. Paediatrics. 2010 Sep;126(3):e541-9. 48 Leisegang R, Cleary S, Hislop M, et al. Early and late direct costs in a Southern African antiretroviral treatment programme: a retrospective cohort analysis. PLoS medicine. 2009;6(12):e1000189-e.

103

Table 1 : Baseline characteristics of children hospitalised between April-October 2010 in

West Africa, overall and according to the underlying cause of hospitalisation (N= 155).

IeDEA Paediatric West-Africa.

104

Table 2: Distribution of the 273 diagnoses among HIV-infected children hospitalised in West

Africa between April and October 2010. IeDEA Paediatric West-Africa.

105

Table 3: Causes of hospitalisations overall and for the six most frequent events, according to

cotrimoxazole prophylaxis (table 3a) and to duration of antiretroviral treatment (3b) in HIV-

infected children in West Africa (N = 155). IeDEA Paediatric West-Africa.

106

Table 4: Distribution of the underlying causes of hospitalisation and death in the 155 HIV-

infected children hospitalised in West Africa. IeDEA Paediatric West-Africa.

107

5.3.Principal results and discussion

As in the previous studies, we face here a strong selection bias induced by the early mortality

of HIV-infected children before any HIV diagnosis. However, this study reflects the reality of

the 2010 West African context providing a thorough description of causes of hospitalisation

in HIV-infected children according to specific gold standard definitions. We highlight the

following findings:

First, the diagnosis of HIV infection was made during hospitalisation for 58% of children,

reflecting the difficulties in implementing successful EID programmes in West Africa in

postnatal care and linking HIV-infected children to HIV care.

Second, there were no differences in the spectrum of diagnoses associated with hospitalisation

whether the children were on or off cotrimoxazole prophylaxis. But, it is interesting to note

that cotrimoxazole is clearly not prescribed on a routine basis to all HIV-exposed infants as

recommended by the WHO (101). HIV-care programmes continue to experience barriers in

the roll-out of this prophylaxis, despite it being cheap, efficacious and feasible.

Third, one hospitalisation in two was related to infectious diseases and one hospitalisation in

three was caused by a non-AIDS defining event; this proportion was higher in ART-treated

children. This reflects that there is high residual morbidity on ART that needs to be taken into

account in the health needs of children on ART. This may become an emerging problem with

the survival of HIV-infected on ART growing up on a chronic disease mode. Indeed, the

improved survival on ART is also associated with serious metabolic complications, including

lipodystrophy, dyslipidemia, insulin resistance, lactic acidosis and bone loss. In addition, the

dyslipidemia mainly seen with protease inhibitors may increase the risk of cardiovascular

disease in adulthood and potentially in children as they mature into adults (203).

Discussions on new recommendations for cotrimoxazole prophylaxis and its utility for ART-

treated children are emerging. Some studies have questioned its utility in ART-treated

children who are expected to have higher CD4 counts and therefore be less exposed to severe

infections (204, 205). On the other hand, other studies have demonstrated the efficacy of

continuing cotrimoxazole in ART-treated children; a sub-study of the ARROW trial showed

higher rates of hospitalisation and mortality in children who had stopped cotrimoxazole (206).

Our study showed no association between cotrimoxazole and the spectrum of diagnoses;

however we also report high rates of infectious morbidity that could be avoided by such

prophylaxis, providing healthcare programmes can handle the roll-out.

National programmes must maintain efforts to focus towards increasing the offer of cotrimoxazole prophylaxis to all HIV-infected children whether they are on ART or not.

108

6.WHAT IS THE EVOLUTION OF HIV IN

CHILDREN AFTER THE INITIATION

OF ANTIRETROVIRAL THERAPY?

Contribution to model inputs

109

6.1. Severe morbidity and healthcare resource utilisation in HIV-

infected children on ART in Abidjan, Côte d’Ivoire


Since the introduction of ART in Abidjan, HIV-related mortality has significantly decreased

(119). However, few data documenting more precisely severe morbidity and healthcare

resource utilisation are available.

Having these data would allow comparing severe morbidity and healthcare resource

utilisation in ART-treated children to the base case scenario we described previously, in

section 4.2 and measuring the cost-effectiveness of strategies to improve access to EID and

ART.

We documented severe morbidity and healthcare resource utilisation in HIV-infected children

on ART, within the same study period and health clinic than the base case study and therefore

the same standard of care, with the perspective of comparing these data to those collected in

ART-untreated children and to analyse the ART-effect in the Abidjan context.

This retrospective study, conducted within the prospective cohort of HIV-infected and ART-

treated children had for principal objective to describe severe morbidity (WHO stage 3 or 4,

or any event leading to hospitalisation or death) and healthcare resource utilisation (both

outpatient and inpatient care) over the 2004-2009 period in the CePReF in Abidjan, Côte

d’Ivoire.

6.1.2. Desmonde et al – JAIDS – in press

The results of this study are currently in press for publication JAIDS for publication in early

2014.

128

SCD1 - Observed incidence rates (point with 95%CI) over time since ART initiation for overall severe morbidity, WHO Stage 3/4 events, TB,

Malaria and non AIDS-defining events.

129

SDC 2– Survival baseline function for severe morbidity and healthcare resource utilisation

since ART initiation in 332 HIV-infected children in Abidjan, Côte d’Ivoire.

130


More than two thirds of the children on ART followed up in the CePReF paediatric active file

and included in our study developed at least one incident severe morbid event over the 2.5

year median follow-up period. Most events were AIDS-defining and occurred early, within

the first six months of ART, and more frequently in children with an advanced HIV-disease at

ART initiation. However, stable and substantial rates of non-AIDS defining events

throughout the study period were also documented, as reported in the causes of hospitalisation

study. Furthermore, overall severe morbidity was reduced by more than 60% in children who

were on cotrimoxazole prophylaxis. As for healthcare resource utilisation, although 88% of

severe morbidity events led to utilisation of some healthcare services, our findings reveal

many missed opportunities for appropriate care. We hypothesise that the major barrier to care

is the additional cost for families, despite the subsidies.

In this analysis, we took into account the comments made when conducting the previous

cohort study in non-ART-treated children (chapter 4). In order to account for the recurrent

events and the non-independence between death and severe morbidity, we opted for a

completely different modelling strategy, and used frailty models. Shared frailty models

allowed accounting for recurrent events in a given patient, and joint frailty models allowed

studying the joint evolution over time of both survival processes: severe morbidity and death

(207, 208).

Overall, these data provide essential information on post-ART care in HIV-infected children

and reflect the reality of the field in West Africa. In a context where there is prevalent

infectious morbidity, it is essential to continue to dispense cotrimoxazole to these children and

to implement interventions that could improve retention in care. Children are entirely

dependent on their caregiver who themselves face many social and economic barriers. A

recent study in Kenya reported that having an unrelated caregiver or more than three different

accompanying caregivers were associated with immunological and virological failure (209).

Changing caregiver attitudes requires interventions at both an individual and a societal level.

Healthcare workers should actively promote care-seeking for children.

6.1.4. Description of direct costs of care in HIV-infected children on ART

During the above study, we also investigated the cost of care among HIV-infected children on

ART though data were not presented in the final manuscript.

Unfortunately, we were not able to assess the direct costs of hospitalisation or complementary

exams nor could we assess the overall indirect costs i.e those resulting from the functioning of

the healthcare centres, such as personnel salaries and costs related to the equipment and

infrastructures.

Only the direct costs, perceived by the patient and related to drug intake, could be estimated.

The costs of each prescribed drug (out and inpatient care) were collected by the CePReF

pharmacist (price list: 2012). An exhaustive list of the costs for each drug was not available.

131

Consequently, an average cost per type of drug (antibiotic, antimalarial, antipyretic…) was estimated using the available cost data. Table 4 presents the amount of prescribed drugs in the context of outpatient care by type of severe event, in ART-treated children included in the CePReF, Abidjan, Côte d’Ivoire. The most prescribed drugs overall were antibiotics (32.3%), antipyretics (18.2%) and antimalarials (10.8%). Table 4: Amount of drugs prescribed in the context of outpatient care by type of severe

morbid event in HIV-infected children on ART included in the CePReF, Abidjan, Côte

d’Ivoire, 2004-2009

Overall Stage 3 Stage 4 Tuberculosis Malaria Other

Number of events 697 524 2 15 113 43

Type of drug

Antibiotics 225 (32.28) 206 (39.31) 2 (100) 0 (0) 1 (0.88) 16 (37.21)

Antimalarials 75 (10.76) 17 (3.24) 0 (0) 0 (0) 52 (46.02) 6 (13.95)

Anti TB 15 (2.15) 0 (0) 0 (0) 15 (100) 0 (0) 0 (0)

Parasiticides 6 (0.86) 4 (0.76) 0 (0) 0 (0) 2 (1.77) 0 (0)

Corticoids 4 (0.57) 4 (0.76) 0 (0) 0 (0) 0 (0) 0 (0)

Nutritional supplements 54 (7.75) 50 (9.54) 0 (0) 0 (0) 1 (0.88) 3 (6.98)

Anti fungals 35 (5.02) 33 (6.30) 0 (0) 0 (0) 1 (0.88) 1 (2.33)

Antipyretic / analgesic 127 (18.22) 70 (13.36) 0 (0) 0 (0) 52 (46.02) 5 (11.63)

Anti-inflammatory drugs 7 (1.00) 3 (0.57) 0 (0) 0 (0) 2 (1.77) 2 (4.65)

ENT specific drugs 79 (11.33) 76 (14.50) 0 (0) 0 (0) 0 (0) 3 (6.98) Antiemetics / Antispasmodic 45 (6.46) 40 (7.63) 0 (0) 0 (0) 2 (1.77) 3 (6.98)

Skin care / Antiseptic 7 (1.00) 7 (1.34) 0 (0) 0 (0) 0 (0) 0 (0)

Antianeamic drugs 14 (2.01) 12 (2.29) 0 (0) 0 (0) 0 (0) 2 (4.65)

Others 4 (0.57) 2 (0.38) 0 (0) 0 (0) 0 (0) 2 (4.65)

132

Table 5 presents the mean amount of times each group of drugs was prescribed per event. For instance, for every WHO stage 3 event, antibiotics were prescribed 0.40 times (ie: four WHO stage 3 events out of ten led to the prescription of antibiotics). As described in the previous paper, the occurrence of a severe event was not systematically associated with resource to outpatient care.

Table 5: Amount of times each drug was prescribed in the context of outpatient care per

event, by type of severe event in HIV-infected children on ART included in the CePReF,

Abidjan, Côte d’Ivoire, 2004-2009

Overall Stage 3 Stage 4 TB Malaria Other

mean sd* mean sd* mean sd* mean sd* mean sd* mean sd*

Antibiotics 0.33 0.49 0.40 0.51 1.00 0.00 0.00 0.00 0.01 0.09 0.39 0.49

Antimalarials 0.11 0.31 0.03 0.18 0.00 0.00 0.00 0.00 0.46 0.50 0.15 0.36

Anti TB 0.02 0.15 0.00 0.00 0.00 0.00 1.00 0.00 0.00 0.00 0.00 0.00

Parasiticides 0.01 0.09 0.08 0.09 0.00 0.00 0.00 0.00 0.02 0.13 0.00 0.00

Corticoids 0.01 0.08 0.08 0.09 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00Nutritional supplements 0.08 0.27 0.10 0.29 0.00 0.00 0.00 0.00 0.01 0.09 0.07 0.26

Anti fungals 0.05 0.22 0.06 0.24 0.00 0.00 0.00 0.00 0.01 0.09 0.02 0.16

Antipyretic / analgesic 0.19 0.39 0.14 0.36 0.00 0.00 0.00 0.00 0.46 0.50 0.12 0.33Anti-inflammatory drugs 0.01 0.10 0.01 0.08 0.00 0.00 0.00 0.00 0.02 0.13 0.05 0.22

ENT specific drugs 0.12 0.33 0.15 0.36 0.00 0.00 0.00 0.00 0.00 0.00 0.07 0.35Antiemetics / Antispasmodic 0.07 0.25 0.08 0.28 0.00 0.00 0.00 0.00 0.02 0.13 0.07 0.26

Skin care / Antiseptic 0.01 0.10 0.01 0.12 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

Antianeamic drugs 0.02 0.16 0.02 0.16 0.00 0.00 0.00 0.00 0.00 0.00 0.05 0.31

Others 0.01 0.08 0.01 0.06 0.00 0.00 0.00 0.00 0.00 0.00 0.05 0.31

*sd: standard deviation

From this data, we derived the estimated mean cost of one of each type of events in this

population. For this we multiplied the amount of drugs prescribed (table 5) for one event by

the cost of the drug. Tuberculosis treatment is subsidised by the national plan to fight

tuberculosis in Côte d’Ivoire and therefore we excluded that event from the following

descriptive analysis. Data are presented in table 6. The overall mean cost perceived by the

patient families is $5.39, equivalent to approximately 3000FCFA; the mean daily salary in

Côte d’Ivoire varies between 4000 and 6000 FCFA. We also noted that the costs of severe

events such as WHO Stage 4 events are the most expensive for the families. Furthermore,

other infectious serious events represent substantial costs ($6.08).

133

Table 6: Mean costs ($) spent by the families of drugs prescribed in the context of

outpatient care per event, by type of severe event in HIV-infected children on ART

included in the CePReF, Abidjan, Côte d’Ivoire, 2004-2009

Type of drugs Overall Stage 3 Stage 4 Malaria Other

mean sd mean sd mean sd mean sd mean sd

Overall 5,39 5,70 8,20 4,30 6,08

Antibiotics 2,69 4,00 3,28 4,21 8,20 0,00 0,07 0,77 3,20 4,05

Antimalarials 0,61 1,73 0,18 0,99 0,00 0,00 2,55 2,77 0,81 1,98

Parasiticides 0,03 0,31 0,03 0,29 0,00 0,00 0,06 0,43 0,00 0,00

Corticoids 0,02 0,27 0,03 0,31 0,00 0,00 0,00 0,00 0,00 0,00

Nutritional supplements 0,03 0,10 0,03 0,11 0,00 0,00 0,00 0,00 0,03 0,09

Anti fungals 0,53 2,31 0,67 2,57 0,00 0,00 0,01 0,03 0,26 1,64

Antipyretic / analgesic 0,49 1,04 0,36 0,92 0,00 0,00 0,09 0,98 0,33 0,88

Anti-inflammatory drugs 0,11 1,06 0,06 0,80 0,00 0,00 1,23 1,34 0,51 2,29

ENT specific drugs 0,38 1,08 0,48 1,18 0,00 0,00 0,19 1,39 0,24 1,13Antiemetics / Antispasmodic 0,36 1,39 0,42 1,50 0,00 0,00 0,00 0,00 0,40 1,44

Skin care / Antiseptic 0,05 0,47 0,06 0,54 0,00 0,00 0,10 0,72 0,00 0,00

Antianeamic drugs 0,08 0,64 0,09 0,65 0,00 0,00 0,00 0,00 0,20 1,25

Others 0,01 0,15 0,01 0,13 0,00 0,00 0,00 0,00 0,10 0,44

Data on cost of paediatric healthcare in Côte d’Ivoire are limited (210). We present here only

the costs of the use of drugs, perceived by the families. From a patient’s point of view,

associated with these costs are the costs of transport and time not spent at work in order for

the caregiver to bring the child to the clinic as well as the costs of clinical exams and costs of

inpatient care, though partly subsidised.

However, the direct costs spent by the patient families are not the only costs that need to be

accounted for when running a cost-effectiveness analysis of any care strategy or intervention.

The costs from a policy-makers point of view are also essential: further studies are necessary

in order to assess the direct costs of HIV care programmes (antiretroviral drugs, personnel

salaries …) and indirect costs related to the use of infrastructures and already existing services. The actual cost of a severe morbid event is probably far greater than that described here. A study in South Africa reported an average cost of R120 ($12) per visit for adults (211). In Abidjan, the expenditure necessary for basic care of one HIV-infected child remains high. Although ART is provided at no charge to children at public healthcare centres in Abidjan, AIDS treatment is ‘free’ for only a handful of those on ART. Abolishing or subsidising fees is not enough to relieve the financial burden, since indirect costs can sometimes be oppressive (212). Indeed, from a policy-making point of view, ART provision costs vary widely between settings and regimens. A study in Thailand reported a 70% increase in the costs of ART in children who switched to second line therapy (213). Healthcare services should make more use of generic drugs and paediatric HIV infection provides a clear example of the benefits to be obtained by such a rational strategy for the use of scarce health resources. If ART costs are relatively low costs per capita, this would leave room for further scaling up of services and strategies to improve retention in care and survival of HIV-infected children.

134

6.2. Immune response to antiretroviral therapy according to age at

treatment initiation in HIV-infected children in West Africa


Until 2013, the recommended timing for ART in children aged > 2 years was based on clinical and immunological criteria for age groups according to WHO recommendations (130). In 2013, WHO guidelines recommended treating all children less than 5 years of age, irrespective of their immune status (25). Although the evidence on the optimal timing for ART initiation in children aged 2-5 years is still limited, the advisory board felt that the operational and programmatic advantages of this strategy were strong enough improve considerably clinical outcomes of HIV-infected children. The long-term effect of immediate ART in children aged 2-5 years is not well documented in African children. Although model-based analysis have informed on the quantitative response of CD4 count to ART, few studies have described time to immune recovery in these children (214-216). We hypothesised that earlier ART initiation was associated with a better immune response and a shorter time to immune recovery, defined as a return to normal CD4 counts for age, above WHO thresholds. Within the prospective IeDEA pWADA cohort, we investigated the effect of age on the 24-month quantitative and qualitative immune response to ART in a population of HIV-infected children who initiated ART a median age of 5 years.

6.2.2. Desmonde et al – AIDS – in revision

Data on the immune response in the IeDEA pWADA cohort were presented at the 16th International Workshop on HIV Observational Databases in 2012, the 4th International Workshop on HIV Paediatrics in 2012 and the 19th International AIDS Conference in 2012. The results of this following study are currently in revision for AIDS.

135

Association between age at antiretroviral therapy initiation and 24-month

immune response in HIV-infected children in the IeDEA paediatric West

African Cohort (pWADA).

Sophie DESMONDE 1,2, Fatoumata DICKO3, Fla KOUETA4, Tanoh EBOUA5, Eric

BALESTRE1,2, Clarisse AMANI-BOSSE6, Edmond A. AKA7, Koko LAWSON-EVI8,

Madeleine AMORISSANI-FOLQUET9, Kouadio KOUAKOU10, Siriatou KOUMAKPAI11

Lorna RENNER12, Haby SIGNATÉ SY13 and Valériane LEROY1,2 for the IeDEA West

Africa Paediatric Collaboration.

1 Univ. Bordeaux, ISPED, Centre INSERM U897-Epidémiologie-Biostatistique, F-33000 Bordeaux, France 2 INSERM, ISPED, Centre INSERM U897-Epidémiologie-Biostatistique, F-33000 Bordeaux, France 3 Hôpital Gabriel Toure, Bamako, Mali. 4 Hôpital pédiatrique, Centre Hospitalier Universitaire Charles de Gaulle, Ouagadougou, Burkina Faso 5 Centre Hospitalo-Universitaire de Yopougon, Service Pédiatrie, Abidjan, Côte d’Ivoire 6 Programme MTCT +, Abidjan, Côte d’Ivoire 7 Centre de Prise en charge, de Recherche et de Formation (CePReF), Abidjan, Côte d’Ivoire. 8 Centre Hospitalier Universitaire de Tokoin, Lome, Togo 9 Centre Hospitalo-Universitaire de Cocody, Service Pédiatrie, Abidjan, Côte d’Ivoire 10 CIRBA, Abidjan, Côte d’Ivoire 11

Centre National Hospitalier Universitaire Hubert K. Maga, Cotonou, Bénin 12 Korle Bu Hospital, Accra, Ghana 13 Hôpital des Enfants Albert-Royer, Dakar, Senegal Corresponding author

Sophie Desmonde

Centre de recherche Inserm U897

Equipe VIH, cancer et santé globale dans les pays à ressources limitées

Institut de Santé Publique, d'Epidémiologie et de Développement, Université de Bordeaux 2

Case 11, 146 rue Leo Saignat - 33076 Bordeaux Cedex – France

[email protected]

Tel: +33 (0)5 57 57 50 51 / Fax: +33 (0)5 34 57 05 44

136

Abstract (250/250)

Objective: We describe the association between age at antiretroviral therapy initiation (ART) and 24-month CD4 cell response in West-African HIV-infected children. Methods: All HIV-infected children from the IeDEA paediatric West African cohort, initiating ART, with at least two CD4 cell count measurements, including one at ART initiation (baseline) were included. CD4 gain on ART was estimated using a multivariable linear mixed model adjusted for baseline variables: age, CD4 count, sex, first-line ART regimen. Kaplan–Meier survival curves and a Cox proportional hazards regression model compared immune recovery for age within 24 months post-ART. Results: Overall, 3014 children initiated ART at a median age of 5.6 years; 61.2% were immunodeficient. Within the first 12 months, children 4 years at baseline had significantly lower CD4 gains compared to children < 2 years (p<0.001). However, by 24 months on ART, we observed higher CD4 gain in children who initiated ART between 3-4 years compared to those < 2 years (p<0.001). The 24-month CD4 gain was also strongest in immunodeficient children at baseline. Among these children, 75% reached immune recovery; rates were

highest in those aged 2–5 years within the first 12 months on ART compared to those who initiated ART < 2 years. Beyond 12 months on ART, immune recovery was significantly lower in children >5 years at baseline (aHR:0.69, 95%CI:0.56;0.86). Conclusion: These results suggest that both the initiation of ART at the earliest age < 5 years and before severe immunodeficiency is needed for long-term immune recovery on ART. Keywords: HIV, children, West Africa, immune reconstitution, CD4, ART

137

Introduction

Worldwide, in 2011, more than 3 million children lived with Human Immunodeficiency Virus (HIV), 90% of whom lived in sub-Saharan Africa [1]. When untreated, HIV-infected children experience increased risks of severe morbidity and mortality because of immunodeficiency [2-5]. Antiretroviral therapy (ART) is the standard treatment for HIV-infected children [6]. Since the introduction of ART in resource-limited settings in 2004, we have observed significant reductions in mortality and morbidity rates as CD4 cell counts increase in ART-treated children [7-9]. Until 2008, the optimal timing for ART initiation was based on clinical and immunological criteria for age groups according to the World Health Organisation [3]. In 2008, following the Children with HIV Early antiRetrovirals (CHER) trial [10], which showed that early ART before 12 months reduced infant mortality and morbidity by 76% and 75% respectively, compared to deferred until reaching the WHO criteria, WHO recommended ART initiation in all children aged < 12 months irrespective of their immune status [11]. These recommendations were extended to all HIV-infected children < 24 months in 2010, older children continuing to initiate ART according to clinical and/or immunological criteria [6]. In 2013, WHO guidelines recommend to treat all children less than 5 years of age, irrespective of their immune status [12]; the rationale of new guidelines was mainly based on

programmatic advantages and there remains limited evidence on the optimal timing of

ART initiation in children aged 2-5 years. However, there is still limited evidence on the optimal timing of ART initiation in children aged > 2 years. The Pediatric Randomised Early versus Deferred Initiation in Cambodia and Thailand (PREDICT) trial enrolled children aged 1-12 years but detected no difference in mortality and morbidity between those who initiated immediate ART and those on deferred ART [13] ; the authors concluded that the study was underpowered. While early ART may improve survival, it can also lead to earlier cumulative toxicity or failure of the ART regimen, with drug resistance if adherence is poor. Understanding treatment response is

especially important in children given the limited alternative regimen choice available in

resource-limited settings. The current access to ART is delayed in children with a median age of 5 years at ART initiation in lower-income countries in 2008-2012 [14]. Younger children are faced with a

still-developing immune system that could interfere with response to treatment. The effects of earlier ART initiation in HIV-infected children including immune recovery for age remain therefore to be explored to guide both clinicians and healthcare policy-makers and bring evidence to the recent WHO recommendations. We hypothesise that earlier ART initiation is associated with a better immune response associated with immune recovery, defined as a return to normal CD4 counts for age above the WHO thresholds for immunodeficiency. The objective of this study was to investigate the association between age at ART initiation and the 24-month immune response to ART in a large paediatric cohort in West-Africa, pWADA. Secondly, we determined rates and predictors of 24-month immune recovery on ART.

138

Methods

Study population and design

The IeDEA paediatric West African Database to evaluate AIDS is aimed to address evolving research questions in the field of HIV/AIDS care and treatment using data from multicentric HIV/AIDS adults and children cohorts in West Africa [8, 15]. This collaboration, initiated in July 2006, currently involves 11 paediatric HIV/AIDS clinics in seven countries: Benin (n=1), Côte d’Ivoire (n=5), Ghana (n=1), Mali (n=1), Senegal (n=1), Togo (n=1) and Burkina Faso (n=1). HIV-infected children are typically seen in these clinics at least every three months. All HIV-infected children (with a confirmed positive PCR test < 18 months or positive serology 18 months), aged 16 years who initiated ART were included irrespective of first-line ART regimens in the prospective pWADA cohort. Treatment initiation was based on international guidelines [6, 16]. The following data were recorded in the database: age, gender, weight and height, WHO clinical staging, date of ART initiation, pre-treatment absolute CD4 cell count and percentage, haemoglobin, cotrimoxazole, initiation and type of ART regimens. All children included in the pWADA cohort, with at least two CD4 cell count measurements, including one at ART initiation (baseline) were included in the present study. A large

proportion of children did not meet inclusion criteria. We describe these children and

compare baseline characteristics. For those included, CD4 was monitored at baseline (± 12

weeks) then every 6 months (± 12 weeks) during follow-up. For all analyses, we used absolute CD4 cell counts, which have been shown to have greater prognostic value compared to CD4% [17]. Definitions for severe and moderate immunodeficiency for age were derived from the 2006 WHO definitions for immunodeficiency. Severe immunodeficiency was defined as CD4 <1500 cells/µL, <750 cells/ µL, <350 cells/ µL and <200 cells/ µL in children aged <1 year, 1-3 years, 3-5 years and 5 years at time of CD4 measurement, respectively . Moderate immunodeficiency was defined as CD4 cell count between 1500–2000 cells/µL in children aged <1 year, 750–1500 cells/µL in children aged 1 – 3 years , 350–750 cells/µL in children aged 3–5 years and 200-350 cells/µL in children 5 years. Immune recovery was defined as a return to normal CD4 cell count for age, above these thresholds. Loss-to-follow-up was defined as any patient with last clinical contact >3 months before the database closing date. We defined anthropometric indicators at baseline to define underweight as Weight-for-Age Z-score (WAZ) <-2 Standard Deviations (SD), and stunting as Height-for-Age Z-score (HAZ) <-2 SD [18]. Statistical analysis

The evolution of CD4 cell count within 24 months after ART-initiation was described according to age and immunodeficiency at ART initiation. Since 2010, the WHO

recommends treating all HIV-infected children aged < 24 months, we therefore selected

the <2 years as reference group for all analyses. We modelled the CD4 change at 6, 12 and 24 months after ART initiation using a repeated-measures linear regression model adjusted for the following baseline variables: age, immunodeficiency, gender and first-line ART regimen. There was a change in slope

intensity at 6 months which we computed and random effects were used on both slopes to

139

account for within-subject correlation caused by repeated measures. We could not adjust our model on HAZ or WAZ because of too few available data. Fixed effects on CD4 change were estimated using Wald tests. Residual homoscedasticity and normality were checked graphically. Results from primary analysis were confirmed by performing a sensitivity analysis excluding patients who died or were lost to follow-up within the 24-month follow-up period.

Kaplan–Meier survival curves were used to compare time to immune recovery for age at 24-month on ART in immunodeficient children at baseline. We then ran a Cox proportional hazard regression model to study the associated factors with immune recovery within 24 months post-ART. This was defined as reaching a CD4 count above WHO threshold for

age at time of measurement. Analyses were adjusted for baseline age, gender, ART regimen and severity of baseline immunodeficiency (moderate vs. severe). The Cox proportional hazards assumption was checked graphically for all variables. We could not confirm this assumption for age at ART initiation and consequently computed the interaction with time: adjusted hazard ratios are expressed for the two main follow-up periods, within the first 12 months on ART, then between 12 and 24 months. For other variables, the risks were constant over time.

Results

Cohort characteristics

Overall, 4808 children initiated ART within the pWADA cohort, of which 3014 (62%) met inclusion criteria. The selection process of the study population is described in Figure 1. Baseline characteristics by age at ART initiation of included children are described in

Table 1. Their median age at ART initiation was 5.6 years [interquartile range (IQR): 2.6 – 9.0]. At baseline, 20.2% of the children had already reached WHO clinical stage 3 and 11.1% WHO clinical stage 4: this was more frequent among children aged <3 years at baseline (p<0.001). Anthropometric and clinical data were limited, available in 65.4% of cases. When data were available, both median baseline WAZ and HAZ were below average; 37.8% of the children were known to be underweight and 38% stunted, this proportion was significantly higher in children <3 years (p<0.001). We had cotrimoxazole data only for 46.5% of children, among which 72.4% were following cotrimoxazole prophylaxis. There was no significant difference in the age distribution of these children (p=0.39). Most patients (77.1%) initiated a non-nucleoside reverse transcriptase (NNRTI) ART-based regimen, mostly children aged >2 years (p <0.001). The most common ART combinations for these children were AZT+3TC+EFV (35%) and AZT+3TC+NVP (30%). The remaining children initiated protease inhibitor (PI) based ART (25%), including AZT+3TC+NFV, D4T+3TC+ NFV. Most PI-based regimens included NFV before withdrawing this drug in 2007; 24% of the PI group was on LPV/r. We compared these characteristics to those of the 1794 excluded children (Supplemental

Content 1). The age distribution of excluded children was significantly different

(p<0.001), the proportion of children initiating ART with a higher proportion of > 2

year olds initiating ART (24% vs 17%). Furthermore, there was a higher proportion of

children who had reached WHO Stage 4 in excluded children (31% vs 18%). We also

140

noted higher rates of loss-to-programme in excluded children (10% died and 40% were lost to follow-up (LTFU)). Median baseline CD4 count was 431 cells/µL [IQR: 197–762]. One thousand eight hundred and forty-six children (61.2%) initiated ART while already presenting signs of immunodeficiency; 1272 (42.2%) were severely immunodeficient, the proportion of these children was highest in children aged < 3 years old (p<0.001). After 24 months on ART, 47 children (1.6%) died and 275 (9.1%) were LTFU. There was no significant difference in the loss-to-programme rate (death or LTFU) based on age groups at ART initiation (p=0.18).

24-month immune response on ART

The median number of CD4 cell count measurements during the 24-month follow-up period was 5 [IQR: 3 – 8]. The median baseline CD4 count was 431 cells/µL [197–762] and reached 735 cells/µL [455–1135] and 784 cells/µL [494–1159] after 12 and 24 months on ART, respectively. Figure 2a illustrates the observed mean CD4 cell count evolution since ART initiation according to the different age groups at ART initiation. We observe on this graph that the younger the age at ART initiation, the stronger the CD4 gain. In all age groups there

was a significant increase of CD4 counts over the 24 month period. CD4 gain by 6

months post-ART was strongest, particularly in the younger children at baseline, aged <

3 years. Following this peak at 6 months, CD4 counts continue to increase, at a slower pace, and in a non-significant manner among children < 5 years; the observed CD4

change in children aged 5 years was much lower. When observing mean CD4 gain by immunodeficiency status at baseline, we observed a stronger increase in CD4 cell counts in those who presented moderate or severe immunodeficiency at baseline, particularly in the

first 6 months of ART. However, these children, however, do not seem to reach normal CD4 cell counts for age by 24 months (Figure 2B). Table 2 presents adjusted estimates of mean CD4 gain at specific time points. The reference group was boys aged <2 years at baseline, who initiated an NNRTI-based ART regimen, without baseline immunodeficiency for age. In this group, CD4 count increased at 6

months and then declined slightly at 12 and 24 months but remained significantly higher

than at baseline (+152 cells, 95%CI: +89;+216). In adjusted multivariate analyses, age at

ART initiation remained significantly associated with CD4 gain at 6 and 12 months but this tended to fade by 24 months. After 6 months on ART, compared to children aged < 2 years, the mean CD4 gain was significantly lower in all children older than 2 years (p<0.001). After 12 months on ART, the mean CD4 gain was significantly lower in children aged 4 years at baseline compared to those aged <2 years. After 24 month on ART, and compared to children aged< 2 years at ART initiation, the mean CD4 gain was not significantly different within age groups, except for those initiated ART aged 3-4 years compared to those aged <2 years (+88 cells/µL, 95%CI: +2 ; +173). Furthermore, after 24 months on ART, the mean CD4 gain was the highest in children with signs of moderate and severe immunodeficiency for age compared to those with no signs (p<0.001). When studying other covariates, the mean CD4 gain was consistent and significantly lower over the first 24 months of ART in children initiating a PI-based regimen compared to those on NNRTIs (p<0.001). In addition, it was significantly higher in girls compared to boys (p<0.001).

141

Since 10% of the sample was lost-to-programme and this rate was higher in children who did not meet inclusion criteria, we performed a sensitivity analysis, conducting the same linear mixed model after having removed all the children who either died or who were lost to follow-up. There were no significant differences in the mean CD4 gain in both samples (data not shown). 24-month immune recovery for age for children immunodeficient at ART initiation

Of the 1846 (61.2%) children who were either severely or moderately immunodeficient at ART initiation, 1393 (75.4%) experienced immune recovery for age by 24 months on ART. The observed overall median time to immune recovery for age was 9.4 months (IQR: 6.1–15.8). The overall 12 and 24-month cumulative probabilities for immune recovery were 53.7% (95%CI: 51.3; 56.1) and 84.7% (95%CI: 82.7;86.6), respectively. The 24-month cumulative probability for immune recovery on ART was >85% in all children aged <5 years at baseline while for older children, immune recovery was significantly lower reaching only 79.3% (95%CI: 76.1 ; 82.3), (plogrank < 0.001) (Supplemental Content 2). When examining predictors of the rate of immune recovery over the 24 months on ART and adjusted for the degree of immunodeficiency (moderate or severe), ART regimen and gender, the association with age at ART initiation differed significantly over time (Table 3). Indeed, we identify two main periods in the immune recovery: before and after 12 months (Figure 2). Table 3 describes associations between baseline covariates and immune recovery on ART by 24 months on ART. Within the first 12 months of ART, children aged 2-5 years at baseline were more likely to achieve immune recovery for age compared to those who initiated ART aged <2 years with adjusted hazard ratios (aHR) reaching 1.51 (95%CI: 1.18; 1.93) in children aged 4-5 years (Table 3). During this first period, there was no significant difference between children aged <2 years and those aged 5 years. However, between 12 and 24 months post-ART initiation, we observed a change in trends as children aged 5 years who had not yet experienced immune recovery were significantly less likely to reach an immune recovery for age compared to children aged <2 years at ART initiation (aHR: 0.69, 95%CI: 0.56;0.86). There were no differences among other age groups. For the other covariables, the risks of immune recovery for age were constant over the first 24 months on ART. Children who were severely immunodeficient at baseline were less likely to achieve immune recovery for age compared to those who were moderately immunodeficient (aHR: 0.47, 95%CI; 0.42;0.53). ART regimen and gender were not predictors of immune recovery for age in this analysis.

142

Discussion

In this large collaboration of prospective cohorts of HIV-infected children in West Africa, we made several important observations about the 24-month immune response on ART. First, severe immunodeficiency was overall highly prevalent at ART initiation, and higher in the children aged <3 years. Second, on ART, we observed a substantial increase of CD4 cell count over two time periods, with an initial rapid increase during the first six months on ART, followed by a relatively smaller increase in the following 18 months. Third, when modelling the 24-month immune response to ART, age and CD4 count at ART initiation, gender and ART regimen were all correlated with the CD4 gain after 6, 12 and 24 months of ART. In the first 12 months on ART, children aged <4 years at ART initiation had a significant and substantially higher gain of mean CD4 compared to those 4 years, a trend which then faded over time except for those aged 3-4 years at baseline who had the highest mean CD4 gain over the 24 months on ART compared to those <2 years. Fourth, for those immunodeficient children at ART initiation, nearly all children <5 years at ART initiation achieved immune recovery for their age after 24 months of ART. However children aged 5 years who had not yet experienced immune recovery in the first 12 months of ART were significantly less likely to do so afterwards, compared to children initiated ART before 2 years of age. Our study presents several limitations that need to be discussed. First, the median age of our cohort was 5.6 years, underlying the late age at which children initiate ART. Consequently, we feel our population is left-truncated as the youngest and most vulnerable children have died before any chance of inclusion in our cohort, leaving a population of survivors. The children aged <2 years included in this cohort are those who accessed care, probably following symptomatic conditions and are the sickest. The of early age at ART initiation in our cohort could therefore be underestimated compared to a more representative cohort, asCD4 gain and immune recovery in children <2 years was compared to that of older and stronger children, who have already survived many HIV-related diseases. On the other hand, the use of CD4 count as oppose to CD4% could have led to an over estimation of the association with age at ART initiation as absolute CD4 thresholds are highest in young children. Since the median age at baseline in our cohort is 5 years, the choice of CD4 cell counts was justified, especially considering their improved prognostic value over CD4% [17]. Second, the rate of LTFU reached 9% at 24 months. Although comparable to other cohort studies in Africa [7-9], this rate remains high and was even higher in children who did not meet inclusion criteria. This can consequently lead to biased estimates of CD4 gain. Indeed, there is a strong assumption that children lost to-follow-up are the sickest children, who actually are deceased. These children would consequently have low CD4 counts that are not accounted for in our analysis. Statistical methods for correcting this informative censoring bias do exist, however they are based on the assumption that the population of LTFU children is homogeneous. This however is not always true, as a recent in study in Kenya has shown: only 16% of LTFU were actually deceased, others chose to not return to clinic for various reasons [19]. However, sensitivity analyses showed no differences in CD4 gain between our study sample and a sample of only children who remained in care, conferring robustness to our results. Third, we ran a Cox proportional hazard model to analyse time to first immune recovery for age on ART, thus making the assumption that children who achieve immune recovery do not relapse. Although this is not the case in reality and many

143

children fail first-line treatment, our results provide however a correct analysis of first immune recovery following ART within a 24-month time frame. We advise caution in the interpretation of these results as further modelling would be necessary to model accurately long-term immune recovery taking into account relapse. Furthermore, we defined immune recovery for age according to WHO thresholds. A more accurate definition would have been to compare CD4 counts to a reference population in West Africa of same sex and age, using z-scores as it has been done in the past in European cohorts [20]. However, considering that 42% of our sample was severely immunodeficient, the z-scores would have been particularly low overall and thus truncated for the purpose of the analyses, underestimating the burden of true immunodeficiency in HIV-infected children at ART initiation. Finally, although we were able to control for some important confounders, such as CD4 at ART initiation or ART regimen, we were unable to adjust for potential confounding by viral load, as this measurement is not collected on a routine basis in the participating clinics. We could not adjust either for morbidity events or growth indicators either, as height and weight data were scarce in this cohort. Initiating ART at a younger age is associated with a more rapid growth recovery [21, 22], however the links between both growth and immune recovery have yet to be investigated. The children in this study experienced substantial improvements in CD4 count after ART initiation. Although children aged <2 years have higher CD4 gains within the first six months of ART, this trend fades over time and by 24 months there is no difference in overall CD4 gain. We observed, however, that children aged >4 years had a significantly lower CD4 gain than the younger children in the first 12 months on ART which is comparable to other studies in Africa [7, 20, 21, 23, 24] and many explanations have been put forward. The immune reconstitution in children is mainly via the production of naïve CD4 T cells by the thymus [25]; slower recovery in older children could be linked to an age-related decrease in the thymic activity [25, 26]. This also explains the fact we observe a decline over time in CD4 cells in these younger children despite ART, and could explain the fading of the association

with age by 24 months. We also found that CD4 gain was associated with ART regimen. Indeed children initiating a PI-based ART regimen had a lower CD4 gain than those on NNRTIs. This observation disagrees with the recent results of the IMPAACT P1060 trial, which demonstrated a superior performance of LPV/r-based therapy over NVP [27]. However, we note here some confounding since PIs were available only for the sickest children. Our second main observation was that the poor immune status at treatment initiation was associated with a stronger CD4 gain, as this has been also observed in northern cohorts [28]. In our study, most of immunodeficient children experienced immune recovery by 24 months on ART. Previous studies have described immune recovery on ART in both northern and southern settings but data are scarce. In the Pediatric AIDS Clinical Trials Group (PACTG 219/219C), only 36% of children initiating ART with baseline severe immunodeficiency reached normal CD4 levels by 5 years [28]. In a cohort of HIV-infected children in Ghana, Renner et al [23] reported 76% of children achieving immune recovery at 24 months on ART, similarly to our results. We further observed that the association between age at ART

initiation and immune recovery differed over time. Within the first year, we observed

144

higher chances for immune recovery in children aged 2-5 years. We explain this by the fact that these children are selected for having already survived HIV-related diseases during their early childhood and are naturally stronger than the younger children. Children aged 5 years showed no difference in the immune recovery in the first year of ART with children aged <2 years. However, those who did not experience immune recovery within the first year on ART are significantly less likely to recover afterwards, between 12–24 months compared to those children aged <2 years. Indeed, older children, 5 years, have been infected longer and therefore are subject to more advanced disease which could have caused permanent changes to the immune system, contributing to poorer long-term status. In summary, we show that the initiation of ART at both an early age and before severe immunodeficiency is a necessity for a long-term maintenance of normal CD4 cell counts for age. These results have public health policy implications that are in line with the revised WHO guidelines that recently recommended ART initiation in all children aged <5 years irrespective of clinical or immunological disease severity [12]. Our results on immune recovery on ART in these surviving children show that treating all children less than 5 years is associated with a significantly higher immune recovery. However, the short-term CD4 gain within 6 month was the best in those who initiated ART before the age of two years. Thus, healthcare programmes in West Africa must continue to concentrate on identifying HIV-exposed infants born to HIV-infected mothers to offer them the earliest appropriate HIV care.

145

References

1. World Health Organisation. UNAIDS report on the global AIDS epidemic. In. Edited by UNAIDS; 2011.

2. Spira R, Lepage P, Msellati P, Van de Perre P, Leroy V, Simonon a, et al. Natural History of Human Immunodefiency Virus Type 1 Infection in Children: A Five-Year Prospective Study in Rwanda. Pediatrics 1999,104:e56-e56.

3. Nathan LM, Nerlander LM, Dixon JR, Ripley RM, Barnabas R, Wholeben BE, et al. Growth, morbidity, and mortality in a cohort of institutionalized HIV-1-infected African children. Journal of acquired immune deficiency syndromes (1999)

2003,34:237-241. 4. Harambat J, Fassinou P, Becquet R, Touré P, Rouet F, Dabis F, et al. 18-month

occurrence of severe events among early diagnosed HIV-infected children before antiretroviral therapy in Abidjan, Côte d'Ivoire: a cohort study. BMC Public Health

2008,8:169-169. 5. Desmonde S, Coffie P, Aka E, Amani-Bosse C, Messou E, Dabis F, et al. Severe

morbidity and mortality in untreated HIV-infected children in a paediatric care programme in Abidjan, Côte d'Ivoire, 2004-2009. BMC infectious diseases

2011,11:182-182. 6. World Health Organisation. Antiretroviral Therapy for HIV infection in infants and

children: towards universal access. Recommendations for a public health approach. 2010 Revision. In. Edited by UNAIDS; 2010.

7. Bolton-Moore C, Mubiana-Mbewe M, Cantrell RA, Chintu N, Stringer EM, Chi BH,

et al. Clinical outcomes and CD4 cell response in children receiving antiretroviral therapy at primary health care facilities in Zambia. JAMA 2007,298:1888-1899.

8. Anaky MF, Duvignac J, Wemin L, Kouakoussui A, Karcher S, Touré S, et al. Scaling up antiretroviral therapy for HIV-infected children in Côte d'Ivoire: determinants of survival and loss to programme. Bull World Health Organ. 2010,88:490-499.

9. Davies M-A, Keiser O, Technau K, Eley B, Rabie H, van Cutsem G, et al. Outcomes for the South African National Antiretroviral Treatment Programme for children: the IeDEA Southern Africa collaboration. South African Medical Journal 2010,99:730-737.

10. Violari A, Cotton MF, Gibb DM, Babiker AG, Steyn J, Madhi SA, et al. Early antiretroviral therapy and mortality among HIV-infected infants. N Engl J Med.

2008,359:2233-2244. 11. World Health O. Report of the WHO Technical Reference Group, Paediatric

HIV/ART Care Guideline Group Meeting. Revised treatment recommendations for infants. Geneva: WHO. In; 2008.

12. World Health Organisation. The use of antiretroviral drugs for treating and preventing HIV infection. In; 2013.

13. Puthanakit T, Saphonn V, Ananworanich J, Kosalaraksa P, Hansudewechakul R, Vibol U, et al. Early versus deferred antiretroviral therapy for children older than 1 year infected with HIV (PREDICT): a multicentre, randomised, open-label trial. Lancet Infect Dis 2012,12:933-941.

14. Leroy V, Malateste K, Rabie H, Lumbiganon P, Ayaya S, Dicko F, et al. Outcomes of antiretroviral therapy in children in Asia and Africa: a comparative analysis of the IeDEA pediatric multiregional collaboration. J Acquir Immune Defic Syndr

2013,62:208-219. 15. Ekouevi DK, Azondekon A, Dicko F, Malateste K, Touré P, Eboua FT, et al. 12-

month mortality and loss-to-program in antiretroviral-treated children: The IeDEA

146

pediatric West African Database to evaluate AIDS (pWADA), 2000-2008. BMC

Public Health 2011,11. 16. World Health Organisation. Antiretroviral Therapy of HIV infection in infants and

children: towards universal access. Recommendations for a public health approach. . In. Edited by UNAIDS; 2006.

17. HPPMCS, Boyd K, Dunn DT, Castro H, Gibb DM, Duong T, et al. Discordance between CD4 cell count and CD4 cell percentage: implications for when to start antiretroviral therapy in HIV-1 infected children. AIDS 2010,24:1213-1217.

18. World Health O. WHO case definitions of HIV for surveillance and revised clinical staging and immunological classification of HIV-related disease in adults and children. UNAIDS 2007.

19. Braitstein P, Songok J, Vreeman RC, Wools-Kaloustian KK, Koskei P, Walusuna L,

et al. "Wamepotea" (they have become lost): outcomes of HIV-positive and HIV-exposed children lost to follow-up from a large HIV treatment program in western Kenya. J Acquir Immune Defic Syndr 2011,57:e40-46.

20. Lewis J, Walker AS, Castro H, De Rossi A, Gibb DM, Giaquinto C, et al. Age and CD4 Count at Initiation of Antiretroviral Therapy in HIV-Infected Children: Effects on Long-term T-Cell Reconstitution. J Infect Dis 2011.

21. De Beaudrap P, Rouet F, Fassinou P, Kouakoussui A, Mercier S, Ecochard R, et al. CD4 cell response before and after HAART initiation according to viral load and growth indicators in HIV-1-infected children in Abidjan, Cote d'Ivoire. J Acquir

Immune Defic Syndr 2008,49:70-76. 22. McGrath CJ, Chung MH, Richardson BA, Benki-Nugent S, Warui D, John-Stewart

GC. Younger age at HAART initiation is associated with more rapid growth reconstitution. AIDS (London, England) 2011,25:345-355.

23. Renner L, Prin M, Li FY, Goka B, Northrup V, Paintsil E. Time to and Predictors of CD4+ T-Lymphocytes Recovery in HIV-Infected Children Initiating Highly Active Antiretroviral Therapy in Ghana. AIDS Res Treat 2011,2011:896040.

24. Puthanakit T, Aurpibul L, Oberdorfer P, Akarathum N, Kanjanavanit S, Wannarit P, et

al. Sustained immunologic and virologic efficacy after four years of highly active antiretroviral therapy in human immunodeficiency virus infected children in Thailand. Pediatr Infect Dis J 2007,26:953-956.

25. Gibb DM, Newberry A, Klein N, de Rossi A, Grosch-Woerner I, Babiker A. Immune repopulation after HAART in previously untreated HIV-1-infected children. Paediatric European Network for Treatment of AIDS (PENTA) Steering Committee. Lancet

2000,355:1331-1332. 26. Clerici M, Saresella M, Trabattoni D, Ferrante P, Vanzulli A, Vigano A. Thymic

volume predicts long-term immune reconstitution in HIV-infected children treated with highly active antiretroviral therapy. AIDS 2002,16:2219-2221.

27. Violari A, Lindsey JC, Hughes MD, Mujuru HA, Barlow-Mosha L, Kamthunzi P, et

al. Nevirapine versus ritonavir-boosted lopinavir for HIV-infected children. N Engl J

Med 2012,366:2380-2389. 28. Patel K, Hernan MA, Williams PL, Seeger JD, McIntosh K, Dyke RB, et al. Long-

term effects of highly active antiretroviral therapy on CD4+ cell evolution among children and adolescents infected with HIV: 5 years and counting. Clin Infect Dis

2008,46:1751-1760.

150

Figure 1 – Flow diagram of the selection of the study sample. Paediatric IeDEA West Africa Cohort.

Study population (HIV-infected children in the

paedatric IeDEA West Africa Cohort, on ART)

N = 4808

4364 children

No CD4 count available (n= 474)

3890 children

Initiated ART before inclusion in IeDEA (n= 444)

3878 children

No baseline CD4 available (n=12)

< 2 CD4 measurements available (n= 864)

Study sample N = 3014

151

Figure 2– Mean plot of the observed CD4 cell counts (cells/µL) at 6, 12, 18 and 24 months post-ART by age at ART initiation (A) and immunodeficiency status at ART initiation (B) in the 3014 HIV-infected initiating ART in the IeDEA paediatric

Figure 1A Figure 1B

152

Supplemental Content 1 – Comparison of baseline characteristics among the 3014 children fitting inclusion criteria and those 1794 children excluded from the study all initiating ART in the IeDEA paediatric West Africa Cohort. Categorical data are presented as frequencies (%) and continuous data as median (interquartile range. (IQR)).

*Chi-2 tests and Wilcoxon tests were performed where appropriate.

Included Excluded

Overall ; n 3014 1794 p*

Age. years; median (IQR) 5.6 [2.6 - 9.0] 5.0 [2.0 - 8.8] <0.001

Age Category ; n(%) <0.001

<2 years 509 (17.0) 434 (24.2)

2 - 3 years 350 (11.6) 185 (10.3)

3 - 5 years 499 (16.6) 275 (15.3)

5 - 10 years 1094 (36.3) 573 (31.9)

10-15 years 562 (18.7) 327 (18.2)

Gender. Male; n(%) 1606 (53.3) 940 (52.4) 0.5703

WHO clinical stage; n(%)

Stage I 312 (17.2) 189 (16.4) <0.001

Stage II 434 (23.9) 196 (17.0)

Stage III 611 (33.7) 357 (31.0)

Stage IV 336 (18.5) 364 (31.6)

Unknown 122 (6.7) 44 (26.5)

Year of ART initiation ; n(%) <0.001

< 2006 961 (31.8) 485 (27.0)

2006 – 2008 727 (24.1) 301 (16.8)

2008 - 2010 704 (23.4) 337 (18.8)

2010 – 2012 622 (20.6) 671 (37.4)

ART Regimen 0.4087

2 NRTI + 1IP 641 (21.3) 356 (7.4)

2 NRTI + 1 NNRTI 2325 (77.1) 1407 (29.3)

Other 48 (1.6) 31 (0.6)

153

(plogrank < 0.001).

Supplemental Content 2 - Kaplan-Meier estimates and their 95% confidence intervals (95%CI) for the cumulative probabilities of HIV-infected children reaching immune recovery for age according to age group at ART initiation

Age group N Immune recovery Month-12 Month-24

< 2years 388 Probability 51.8% 85.9%

95%CI (46.6 ; 57.1) (81.5 ; 89.8)

[2-3[ years 272 Probability 60.8% 96.9%

95%CI (54.8 ; 66.8) (93.6 ; 98.8)


95%CI (56.9 ; 75.7) (79.3 ; 93.6)


95%CI (56.9 ; 74.9) (84.5 ; 96.6)

5 years 963 Probability 49.6% 79.3%

95%CI (46.4 ; 52.9) (76.1 ; 82.3)

154


In this large West African cohort, severe immunodeficiency for age at ART initiation was

highly prevalent; mainly in children aged < 3 years. Although CD4 increased during ART,

this was more rapid within the first six months. Furthermore, the gain of CD4 cells/mm3 was

stronger in children aged < 4 years compared to older ones, though this trend faded over time.

Finally, nearly all children who initiated ART before 5 years restored CD4 cells above WHO

thresholds, achieving immune recovery for age. This was more likely in children aged 2-5

years in the first 12 months. Beyond that threshold, children aged > 5 years had a lower

chance of achieving immune recovery compared to those who had initiated ART at an earlier

age.

It was in this study that we were confronted with missing data and LTFU the most. In the

IeDEA cohort, biological data such as CD4 count are supposed to be collected routinely every

six months. However, if children do not present to the laboratory or clinic and if these

measurements are not performed for operational reasons, this information is not available. We

made the assumption that CD4 gain has a linear progression and therefore selected children

who had at least 2 measurements as we were measuring CD4 gain. Consequently, there were

many missing data and we excluded a large proportion of the initial cohort. We acknowledge

that multiple imputation is known as an effective method for handling missing data. However,

it is not clear that the method will be effective when the data contain a high percentage of

missing observations on a variable and the reasons for missing data need to be understood for

proper imputation. We also right-censored children who were LTFU or deceased and

arbitrarily chose not to model the joint survival function. Further sensitivity analyses revealed

no major differences in the CD4 gains among children who remained in care and those LTFU

conferring robustness to our results. We acknowledge that this is not sufficient to address this

bias; however it is a first step towards confirming the validity of our results.

Another limitation of the modelling approach is that we made the assumption that all children

followed the same CD4 reconstitution pattern and did not allow for children to have a

different response to that modelled. However, in reality, there are qualitatively different CD4

reconstitution profiles that need to be described in order to help better understand CD4

response in children with a developing immune system and help manage paediatric HIV. This

has been recently described by the ARROW trial team (217). In our cohort of West African

children we do not observe the same patterns; however modelling work that would allow

accounting for different reconstitution patterns and uncertainty in the allocation of the child’s

pattern using latent class models is on-going.

Another major difficulty in this analysis was the absence of standardisation of CD4 count

according to age. CD4 counts are age-dependant as the thymus activity reduces considerably

in older children. Ideally we should have derived age-standardized z scores for CD4 cell

counts, using a reference population of non HIV-infected children, with available age-related

standards of total CD4 counts. These data exist in European cohorts and we could have

derived z scores for our cohort using the LMS method (218). However, the baseline level of

155

immunodeficiency for age was severe and many children would have had extreme z-scores,

making any further modelling very challenging, unless z-scores were left-truncated thus

underestimating the true burden of severe immunodeficiency at ART initiation. Building

reference z-scores and percentiles curves in a resource-limited setting has already been

investigated (219) and is currently ongoing within the IeDEA consortium. Nevertheless, this

analysis provided an accurate description of CD4 gain following 2 years of ART in HIV-

infected children. Furthermore, we also investigated the time to immune recovery and its

predictors using standard survival analyses. It is arguable that children relapse as their

experience virological failure adversely affecting our results. We feel however that this study

contributes greatly to the discussion regarding immune response to ART in a context where

virological testing is not available.

Despite these limitations, we feel our study reflects well the reality of the field in terms of

access and response to ART in West Africa. Indeed, most children aged < 2 years in this

cohort are the sickest, with clinical symptoms explaining their enrolment in care. The older

children are survivors and therefore they have a better chance of recovering full immune

function compared to those younger ones. However, within the first 12 months of ART,

children aged > 5 years had the same lowest chances for immune recovery as infants < 2

years. We explain this by the advanced stage of the disease and poorer immune function.

However, after 12 months, among those older children who had not yet achieved immune

recovery, the chances of that event became a 1.5-fold lower than infants who initiated early

ART. We hypothesise that the chances of immune recovery in children aged < 2 years would

be much higher, at any time point, if they had initiated ART regardless of their clinical

symptoms. We remind here that the sample of infants aged < 2 years were the sickest children

(58% were severely immunodeficient) and advise caution in the interpretation of results.

As long as children continue to be diagnosed at a late stage, initiating ART in all children

aged less than 5 years is beneficial and nearly all achieve immune recovery within 24 months.

However, CD4 gains are stronger in younger children, and chances for immune recovery

would be higher if these children were identified at an earlier stage. Effort and resources must

continue to focus on the early identification of HIV-infected children to offer them ART at

the earliest convenience.

156

7. THE INDEPENDENT “ART-EFFECT” Contribution to model inputs

157

7.1. Background and objective

Several studies in adults, both in Northern and Southern settings, suggest that ART provides

an additional independent protective effect in reducing HIV-related morbidity and mortality,

beyond that referable to the effect on CD4 counts (220-223). In children, the effectiveness of

ART in reducing mortality relies mainly on observational studies; these include two studies

based on large population-based cohorts of HIV-infected children, one in Italy (224) and

several in the USA (225, 226). However, to this day, no study has actually identified or

measured an additional protective effect provided by ART in HIV-infected children.

In 2013, ART availability is rapidly scaling up through extensive roll-out programmes. While

one randomised clinical trial has reported a significant reduction of infant mortality in HIV-

infected South-African children (129), it is important to assess whether ART has an effect on

incidences of severe morbidity and mortality independent of CD4 count in children. If that is

the case, further questions regarding the timing (does the effect start immediately after ART

initiation?), the strength of the effect (is it the same across CD4 and age strata?) and the

targeted morbidity (does the ART-effect only reduce infectious OIs or all OIs?) will need to

be addressed. These data will also help to improve and guide towards the most cost-effective

paediatric HIV care in resource-limited settings.

Furthermore, in the context of the development of the CEPAC-Paediatric model, quantifying

the ART-effect and integrating an “ART-effect” module in the model would allow to project

more accurately the clinical outcomes of different ART strategies as well as associated costs

and measure ICERs, thus guiding national and international guidelines in the care of

paediatric HIV.

The objective of the present “ART-effect” study was to measure the effect of ART within the

different age groups and CD4% strata on the incidence of severe morbidity and mortality, in

Abidjan. We hypothesise that mortality and morbidity are reduced on ART. We present here a

comparative analysis of our data collected in HIV-infected children given pre-ART and on

ART severe morbidity and mortality in the 2004-2009 period in Abidjan.

7.2. Which data to answer the ART-effect question in the Abidjan

context?

Both Abidjan cohorts described above appear at first sight to fit the criteria to answer this

question. Indeed, we have two cohorts of HIV-infected children, with pre- and post-ART

data, followed up in the same standard of care and in the same time span. One cohort has not

yet initiated ART while the other initiated ART for the first time during the follow-up period.

All children were followed through standardised procedures which included complete blood

counts and CD4 counts on a regular basis as well as standardised management of morbid

158

events at the clinic. Immunodeficiency was defined according to the WHO 2006 definitions,

consistent with the study period.

Thus, the initial study population consisted of HIV-infected children with a confirmed

diagnosis (either PCR or positive serology if aged > 18 months), aged > 15 years and who

were followed-up in the CePReF, in Abidjan, Côte d’Ivoire between 2004 and 2009.

The entry date was the individual date of inclusion and ART initiation for those not on ART

and those on ART respectively. The termination date for all patients was the 31st December

2009 if no terminating event occurred before then. Otherwise, patients were censored at

individual date of ART initiation (if off ART period), LTFU or death, whichever came first.

LTFU was defined as > 6 months with no clinical contact. Date of ART initiation was defined

as the date of first drug intake.

The data on mortality and severe morbidity were those collected during the previous studies

in the CePReF (see section 4.1, 4.2 and 6.1). There was no standardised diagnosis of specific

HIV-related morbidity.

Of the 405 HIV-infected children followed-up in the non-ART-treated cohort, 171 children

initiated ART during the study period (2004-2009) and contributed to the ART cohort

following treatment initiation. These children, survivors of early childhood, then became

sicker and more immunodeficient, which triggered ART initiation.

The comparability of both cohorts is questionable and we encountered several difficulties that

did not allow us to address correctly the ART-effect question:

The timing of the morbid event. It is very difficult to ascertain when the morbid

event actually occurred. Events recorded as occurring within the first months of ART

could have been misclassified and actually occurred before ART initiation and were

severe enough to trigger ART initiation. Furthermore, other events, occurring during

the early ART period, could be complications attributable to those pre-ART events.

Second, there is a strong survival bias in this analysis that we need to account for.

Children in the off-ART period are those who are well enough not to be offered ART

initiation. Consequently, they are less sick and are selected as surviving, strong

children for whom we identified less severe morbidity. Indeed, the proportion of

immunodeficiency was stronger among the children in the on-ART period compared

to the others. Third, we also observed less morbidity among children in the off-ART

period because these children are stronger, have already survived early infancy. Real

severe morbidity in these non-ART-treated children is overall underestimated as many

children die before HIV diagnosis. We find again here the left truncation phenomena

we have been describing since the introduction of this PhD and the challenges this

problem presents.

159

Data collection limitations and bias. We strongly feel that in this particular study,

the data were collected differently for both follow-up periods and that there were more

post-ART data available and the collection of data was more precise. Indeed, the data

collection instrument differed. The instrument used for the post-ART data collection

was improved since the pre-ART collection, allowing the integration of many more

data and facilitating the collection. It followed the order of the medical files, while the

first instrument was built independently from the structure of the medical records,

making the person collecting data go backwards and forwards, increasing the chances

for mistakes. Furthermore, the person collecting the data was different in each study.

While for the off-ART period, I collected the data, with no experience in the care of

HIV, the person collecting on-ART data was a medical resident, in paediatrics that I

subsequently trained after my first experience in Abidjan. We feel that this could have

created a potential bias in the data collection, improving the quality of the post-ART

data. Logistics aside, we also feel that the overall standard of care of ART-treated

children may have been better than that of children who had not yet initiated ART,

despite the same follow-up clinic and study period. Children who initiate ART have

quarterly visits for drug delivery, and CD4 counts are measured every 6 months.

Children who are not on ART have no obligation to visit the clinic, other than to

collect cotrimoxazole prophylaxis when delivered. Therefore, they present only with

severe morbidity, and as we have shown in the healthcare resource utilisation study,

this was less frequent in sicker, immunodeficient children. Consequently, we rightly

feel that many children never reported severe morbidity, they either died at home or

were hospitalised elsewhere. In the light of this, severe morbidity in children during

the off-ART period is underestimated.

Missing immunological data. Very few immunological data were available for

children off ART and we did not have regular measurements for children on ART

either. Consequently, we were unable to adjust the severe morbidity for CD4 count

and determine, for each study period and CD4 strata, the actual severe morbidity. CD4

counts increase with ART, and decrease in the absence of ART. We hypothesise that

the increases we observed in the incidence density rates during the on-ART period

compared to the off-ART period would have been less marked if adjusted for CD4 cell

count. Many methods for handling missing data are available, most of which are

unbiased under some assumptions (though biased under alternative scenarios), such as

interpolating CD4 count linearly between measurements. This, however, implied that

we had enough CD4 measurement to construct a robust model, which is not the case.

Imputation is another option, but we felt we had too many missing data.

160

7.3. Pointers to address the biases

Standard regression techniques cannot appropriately address the confounding by severity over

time that we observe in this particular cohort, leading to biased estimates (222). Recent

analytic methods based on inverse probability weighting (marginal structural models)

have been developed to appropriately adjust for this type of confounding. This method was

used to estimate the effectiveness of ART in preventing mortality, using data for 1996–2006

from a USA–based prospective cohort of HIV-infected children (226). Among 1236 children

perinatally infected with HIV followed up over a 10-year period (median duration of follow-

up, 6.3 years; interquartile range, 4.3-9.8 years), 70% of participants had initiated ART.

Eighty-five deaths were observed, and the mortality hazard ratio associated with ART,

compared with non-ART regimens, was 0.24 after adjusting for measured confounding by

severity (95% confidence interval, 0.11-0.51). However, this method has been rarely used in

an African paediatric population. A recent study within the Southern Africa IeDEA

collaboration investigated optimal timing for ART initiation in children aged 2-5 years using

g-computation to adjust for time-dependant confounding of CD4 counts and WHO staging

(133).

Estimating the effect of ART in HIV-infected individuals has been done in the past, in a

population of adults in the USA (222). In this study, all patients were ART-naïve and AIDS-

free at baseline. Patients were followed-up and time-varying variables (CD4 count and viral

load) were measured on a regular basis (every 6 months). Patients who were eligible initiated

ART and/or cotrimoxazole during the follow-up period. Authors estimated the effect of ART

and the joint effect of ART and cotrimoxazole on time to first AIDS-defining event and death.

This methodology could be applied to our initial cohort in Abidjan of untreated children, all

ART-naïve at baseline and of which 171 initiated ART during the follow-up period. However,

further discussions with clinicians are necessary in order to define pertinent outcomes of

interest. Indeed, few of these children are AIDS-free at inclusion, since in this context,

HIV/AIDS symptoms triggered inclusion in the care programme. Furthermore, methods for

estimating the parameters for the model would need to be discussed. Finally, estimating the

independent ART-effect is conditioned by the availability of CD4 data. If no immunological

data is available, we are unable to identify a possible effect of ART outside of its action on

the immune reconstitution. In the case we have a lot of missing immunological data, methods

for estimating CD4 count at different points in time before and after ART need to be

discussed as well as possible multiple imputation. Further analyses are ongoing to investigate

this question within a pooled analysis of Southern and Western African data from IeDEA.

161

8. DISCUSSION AND PERSPECTIVES

162

8.1. Work synthesis

The main objective of my PhD work was to document the evolution of paediatric HIV disease

before and after ART during the scaling up of access to treatment in paediatric programmes,

to describe the resource to healthcare and finally, to describe the immune response to

treatment in West Africa. Two large data platforms comprised of prospective cohorts allowed

us to address these specific research questions regarding paediatric HIV in field conditions,

which differ from data obtained within clinical trial ones.

We make several observations:

Children initiate ART at a late age and advanced stage of HIV disease

Both the initiation of ART at the earliest age < 5 years and before severe

immunodeficiency is needed for long-term immune recovery on ART.

Despite the receipt of cotrimoxazole prophylaxis, the risk of severe morbidity and

mortality remains high in untreated HIV-infected children. This risk is minimized in

children who access early care.

ART-untreated HIV-infected children require substantial inpatient and outpatient care.

Overall, most causes of hospitalisation in HIV-infected children are related to the

disease but the burden of non AIDS-related infectious disease remains high,

particularly in ART-treated children.

In ART-treated children, cotrimoxazole significantly reduces severe morbidity and

consequently healthcare resource utilisation.

Despite ART, HIV-infected children still require substantial utilisation of healthcare

services however this is not systematic and families often encounter strong economic

barriers.

There is a need for more effective care and management of HIV-related diseases and

other non AIDS-related infectious morbidity.

In this section, we will discuss the main methodological challenges raised conducting our

project, the public health implications of our findings and lastly, the research perspectives that

I will address in the future during my post-doctoral fellowship.

8.2. Methodological considerations

Overall, except one cross-sectional survey, all our findings were issued from prospective

cohort studies that raise three main specific methodological weaknesses that need to be

understood when conducting, analysing and interpreting such epidemiological studies:

The data quality bias (completeness and accuracy).

Although these cohorts are prospective, the data collection was retrospective in some studies,

which made them dependant on the quality of medical records. Data are missing, incomplete

or inaccurate for different reasons: visit schedules vary according to patient need, and failures

163

in either data collection or data entry result in missing information. Furthermore, data

collection from medical records is subject to inter-observer variability despite procedures.

Consequences of poor data quality include underestimation of true severe morbidity and

mortality rates. A study in Uganda which compared the rates of opportunistic infections in a

routine clinic database and a Research Cohort database comprised of the same patients

reported an overall underreporting rate of 45-50% for opportunistic infections in the routine

clinic database (227). The lack of documentation of the key data items on the summary sheet

by the healthcare worker in the setting of a busy clinic was the main reason for underreporting

rather than a failure of/ incorrect data entry from the summary sheet onto the database.

Different solutions to improve the quality of medical records and thus the health management

system could be envisaged, such as implementing computer-based records, however even this

kind of data can be inaccurate and lack completeness (228). In 2008, the findings of a multi-

regional survey within the IeDEA collaboration showed that despite the implementation of

computer-based medical records, staff lacked training to manage data and trace patients

resulting in high rates of loss-to-follow-up (229). However, these electronic medical records

could play an important role in the scale-up of ART in resource-limited settings. Indeed,

provider-based electronic medical records have been proved effective for both patients and

care providers, in particular in reducing waiting times and optimising the time spent with a

care provider (230).

Electronic medical records that could be accessible directly to clinicians when examining the

patient and completed by the care provider as opposed to data-entry clerks could improve

greatly both the quality of care and the data quality. A recent study in Uganda reported an

additional effectiveness on the quality of the data collected and significant reductions in

missing and incorrect information (231).

Inclusion in the cohort leading to a left truncation bias

Left truncation occurs when children have been at risk before entering the study. As

illustrated by figure 17 below, events of interest may occur outside of the window of

observation, before entering the cohort. This phenomenon induces a selection bias and could

adversely affect the precision of our estimates (232, 233).

164

Source: Cain et al, Am J Epidemiol, 2011 (232)

Figure 17: Timing of events X and Y relative to observation window A-B. In this

hypothetical situation, individuals are recruited into the study at age 48 years (A) and

followed until age 53 years (B). Six individuals are portrayed. In reality, A and B differ by

participant but, for simplicity of presentation, the figure shows A and B as being the same for

all 6 individuals. The individual in row 1 had milestone X at age 50 years and milestone Y at

age 52 years. Both events are within the observation window and, hence, both milestones are

observed. For the second individual, X is observed, but Y has not yet occurred when

observation ends at B, while for the third, both X and Y occur after observation ends. For the

3 remaining individuals, X occurred before study entry, and for the sixth person Y also

occurred before study entry.

Both the paediatric HIV disease progression is rapid with an early mortality in the absence of

intervention estimated to 52% at two years of age (48) and the loss to follow-up of HIV-

infected children beyond early infant diagnosis HIV testing is unacceptably high (75, 234).

We feel therefore that the left truncation bias is quite important in analysing cohorts issued

from our paediatric HIV care programmes in which access to care was quite late (115, 119).

Methods for correcting for left truncation do exist (235, 236) and we have acknowledged this

phenomenon in each of our analysis but the ideal study design in order to study the paediatric

HIV disease progression before and after ART would have been to investigate a birth cohort

of HIV-infected children. However, such cohorts do not currently exist in field conditions in

West Africa. We have illustrated this in the following figure:

165

Figure 18: Selection bias in observational cohorts of HIV-infected children.

The lost-to-follow-up bias

In the context of paediatric HIV disease in resource-limited settings, there is reason to believe

that LTFU is associated with the disease progression. Consequently, even if LTFU rates are

comparable to those reported by other studies, the validity of our results could be adversely

affected. All of the studied events (mortality, severe morbidity, healthcare resource utilisation

and time to immune recovery) are associated with LTFU. Methods to account for this do

exist; however we did not resort to them in every analysis as most often they involve making

assumptions about the outcomes of LTFU children that we were not able to document in our

context. The amount of bias in mortality estimates can be large in ART programmes with

substantial LTFU such as in paediatric West Africa (115). Programmes should routinely

report mortality among patients retained in care and the proportion of patients lost (237). A

simple nomogram can then be used to estimate mortality among all patients who started ART,

for a range of plausible mortality rates among patients lost to follow-up to address these bias

(238). A recent study in sub-Saharan Africa followed this approach to assess mortality

among LTFU in sub-Saharan Africa and reported substantial changes in the estimates of life-

years gained on ART when accounting for LTFU (239).

166

8.3. Public health implications to improve HIV paediatric care

Despite the above considerations, the work provided in this PhD documents as best as

possible the care of paediatric HIV in West Africa, and has identified some of the bottlenecks

that refrain the implementation and delivery of adequate services.

The improvement in the care of HIV-infected children in West Africa needs to be approached

holistically. Diagnosing and treating infant HIV infection the earliest possible remain critical

interventions to reduce childhood mortality in West Africa.

Promoting early infant diagnosis

The first major observation we make throughout this work is the late age at which children are

diagnosed with HIV, despite EID programmes. Even in children who have survived early

infancy, rates of severe morbidity and mortality remain high and preventable and lead to

substantial healthcare resource utilisation, with consequential costs for both the families and

from a policy-making point of view. Having benefited from early access to care minimizes

this severe morbidity adding a major argument to promote earlier access to HIV diagnosis. It

is urgent to identify effective strategies in the linkage to care of PMTCT programmes and

paediatric health services (240).

Offering a comprehensive global package of HIV care in addition to ART

In all pre-ART and post-ART cohort studies, we observed high rates of malnutrition, severe

anaemia and general immunodeficiency among children initiating ART, despite

cotrimoxazole prophylaxis. This severe morbidity persists while on ART, probably because of

late ART initiation, requiring substantial healthcare and there is a high burden of infectious

and non AIDS-defining morbidity. HIV-infected children must be offered in addition to ART,

nutritional support (241, 242), psychosocial support and access to continued cotrimoxazole

prophylaxis. Also, other infectious diseases such as tuberculosis and malaria, highly prevalent

in HIV-infected children in West Africa (243-249) should be taken into account to improve

paediatric care.

Offering ART at the earliest convenience before the age of 2 years to reduce

mortality

The immune response to ART in terms of CD4 cell gain was highest in those who initiated

treatment earlier, underlying the benefits of early access to care. Furthermore, we observed

higher probabilities of full immune recovery in children who initiated ART between 2-5

years. This puts forward the problem of early access to diagnosis and subsequent ART. Those

children aged 2-5 years are survivors; hence they have a better immune response than children

aged < 2 years, who were included in our cohort mostly because these younger children were

already symptomatic and very sick. Children aged > 5 years had the lowest CD4 gain and

167

probability for immune recovery. When analysing the immune response to ART and recovery

to a normal CD4 cell count, many children aged > 5 years do not reach immune recovery,

advocating for an earlier ART initiation. These results are in line with the new 2013

consolidated guidelines on the use of ART for treating and preventing HIV infection, which

extended the age to initiate ART regardless of immune or clinical status to 5 years (25).

Furthermore a recent study has also reported that initiating immediate ART at a later age,

even if there is a low risk of immediate disease progression can maximise CD4 reconstitution

on the long-term (217).

8.4. Research perspectives

From an operational point of view, healthcare policy makers must prioritise interventions and

address demands with a limited amount of resources. The clinical impact and cost of scaling

up paediatric HIV infection care programmes need to be further described and understood.

8.4.1. Addressing costs of care and identifying cost-effective strategies

HIV care and treatment involves HIV diagnosis issues, ongoing visits to healthcare centres,

tools to monitor health status (growth, CD4 or viral load monitoring), and medications,

including ART and cotrimoxazole prophylaxis. Furthermore, the more children access care,

the longer they will live, and the higher their resource to healthcare will be: the direct and

indirect costs of paediatric HIV care will increase for both families and clinics as HIV

becomes a chronic rather than a fatal disease (250).

In 2010, in addition to the costs borne by the families, UNAIDS estimated the global cost of

HIV control to be $15.9 billion, mostly in resource-limited settings and devoted to expanding

treatment services and providing care (7). Effective prevention, treatment, care and support

services for HIV-infected children depend on the reliability of service provision and

continuity of access. UNAIDS has called for the virtual elimination of mother-to child

transmission by 2015 making PMTCT one of its priority outcome areas. Furthermore, the

new WHO guidelines recommend that all HIV-infected children under 5 years of age start

ART immediately upon diagnosis. These recommendations will help keep children in the

system once they are diagnosed – but they also mean that many more children will require

treatment immediately (131). Implementing the new guidelines will require increased

financial resources, not only because of the larger number of children who will need

treatment, but also because the drugs that are required are more expensive and more complex

to administer (35). However, treatment programmes face constrained budgets, with both

domestic and external HIV spending under increasing pressure. As we have reported in this

work, direct costs for patient families are major economic barriers to appropriate care for

HIV-infected children and strategies to spend more effectively allocated resources need to be

urgently identified.

168

There is a rapidly growing and compelling body of literature on the efficacy and cost-

effectiveness of HIV prevention interventions in resource-limited settings. These studies

contain important lessons about strategies to limit the HIV epidemic and address UNAIDS’

targets. For instance, a recent study in South Africa reported that rapid ART initiation in

pregnancy is potentially cost-effective when used as a PMTCT strategy (251). In a context

where the global funding of HIV research and care is decreasing, it is imperative that

available funds be spent for interventions that are cost-effective and that information on cost-

effectiveness help guide resource allocation as funding is scaled up to more closely match real

needs.

As programmes gain more evidence about the cost, risks and benefits of different clinical

approaches, they will be better positioned to focus on high-impact services and streamline

other aspects of care, freeing up resources to support larger patient cohorts. In this way the

benefits of improved efficiency will be measured in the number of additional patients who can

be enrolled on treatment, and through the resulting impact on HIV patients’ health, their

families’ welfare, and the reduction in new HIV infections.

Many questions regarding the paediatric HIV infection in sub-Saharan Africa remain

unanswered, most of which have been cited in the first chapter of this work. Some of these

questions have to be investigated conducting clinical trials. However, a modelling approach

could also be a useful tool to address part of the questions and be very complementary to help

clinical research.

Modelling can be helpful in simulating the reality to put in perspective the results of clinical

trials with longer-term outcome or in different contexts. Modelling could also be useful to

investigate questions for which a clinical trial is not feasible.

With this aim, cost-effectiveness modelling such as the paediatric CEPAC model could be

helpful in addressing some of the pending research questions and guiding the decisions about

prioritising interventions aimed to improve HIV paediatric care in West Africa, as illustrated

below.

8.4.2. How to promote early infant diagnosis in the West African context?

Our work highlighted that it was essential to identify HIV-exposed children the earliest

possible, offer early diagnosis and initiate ART in HIV-infected children before the aged of

12 months ideally to reduce HIV-related infant mortality. Early initiation of ART depends on

an early infant diagnosis. It is crucial to access and identify opportunities to improve

outcomes according to the specific context. However, data on the determinants of entry into

care of HIV-infected children remains limited. More effective and efficient strategies for early

infant HIV diagnosis need to be identified and implemented and the balance of the costs and

benefits of these strategies needs to be further described, accounting for age at HIV diagnosis

and the overall cost of screening.

169

As we have described in the introduction of this PhD, EID coverage remains extremely

limited in West Africa (75, 83, 84, 252). First, most EID programmes target infants of women

identified as HIV-infected in antenatal care, but not all of these women bring their children

into to care for EID testing. In addition, many women are not even tested for HIV during

pregnancy. Focusing EID only on known HIV-exposed infants means that many children

never even enter the cascade of care. Second, among those who enter the EID cascade, long

delays at each step and wide variability in time for return of results lead EID results to be

either unavailable when mothers return or go unclaimed. As a result, 30-80% of tested infants

fail to receive test results and link to care (75).

It is essential to reach infants with both known and unknown HIV exposure. PMTCT

programmes constitute the first point of entry for HIV-exposed infants. Braun et al reported

49% coverage in Malawi (253) via PMTCT programmes. In West Africa, this rate was much

lower: the PEARL study, conducted in 2010, reported only 16% coverage (254). The second

points of entry are paediatric hospital wards. Many children are diagnosed during

hospitalisation, a study in Zambia reported 87% uptake of EID (255), however these children

are diagnosed at a late stage of the disease, with low CD4 counts and malnutrition (256).

Alternative strategies have been proposed to reach infants the earliest possible with both

known and unknown HIV exposure, such as offering EID to all infants presenting for routine

immunization visits at 6 weeks of age (85). Although the acceptability and feasibility of this

has been documented in South Africa with 90% of maternal uptake, in Côte d’Ivoire, uptake

of universal EID screening was not as encouraging. Although maternal uptake reached 58%,

parental uptake was estimated to be only 15%. These results suggested that this strategy was

not effective for EID on a community level in this low prevalent setting and priority should be

given to improving the uptake of PMTCT (83). Efforts must be concentrated on strengthening

the continuum of care from antenatal consultations in PMTCT services to the long-term

follow-up of HIV-infected children, on ART. Improving the tracing of these HIV-exposed

infants is essential as is improving retention in care. Optimising this requires interventions at

multiple levels of the healthcare system and no single approach is likely to be effective for

everyone; this area of research is specific to each country’s health system and financial

resources. For instance, we showed in our study in Abidjan that costs and distance from

healthcare centres were barriers to healthcare resource utilisation. Bringing services closer to

communities, could reduce the costs of care for the people living with HIV and their families

and improve retention in care.

Alternative EID strategies are indeed efficient in high prevalent settings such as South Africa

and current research around novel point of care (POC) testing will allow improving greatly

the number of identified HIV-infected children and initiate early ART according to guidelines

(257). However, in West Africa, the HIV prevalence is much lower among pregnant women

(3.4% vs 16-44% in South Africa) and the priority question remains “how to access these

HIV-exposed infants and guarantee a high EID coverage?”. Despite the proven performances

of tools for EID (80), social and family acceptability for EID remains low. In order to

optimise EID strategies in low prevalent settings, research must focus on:

170

The strengthening and harmonisation of the identification of HIV-infected children the

earliest possible: linkage between PMTCT services and paediatric services must be

improved. This could be by the implementation of an efficient health information

system or, as it has been demonstrated in more prevalent settings, POC EID testing

with same day results. Children who are identified as HIV-positive must be fast-

tracked to initiate ART the earliest possible.

Support for HIV-infected mothers who have many issues to address, including their

own health. Furthermore, infant feeding and disclosure are difficulties among others

that these mothers face and that constitute barriers in the success of PMTCT and EID

programmes.

Approaches to involve fathers and improve acceptability of EID and potential ART for

their child.

However, challenges to these innovative approaches are going to include costs related to

reagents for DBS, shortage of infrastructures and qualified staff and stigma. In the light of

this, the most effective and cost-effective EID strategies are yet to be evaluated.

Using the CEPAC-Paediatric model, we will address this by incorporating a comprehensive

EID module within the current model. The model will then be able to include clinical and

economic outcomes for children who are born HIV-uninfected (either born to HIV-uninfected

mothers, or born HIV-exposed but uninfected, accounting for possible infection during the

breastfeeding period). The model will then simulate each step of the cascade, allowing

varying the proportion of infants who undergo EID. Finally, the model will also incorporate

the sensitivity, sensibility and costs of the proposed EID assays. Strategies will be modelled in

two different settings: South Africa, where there is a generalised epidemic and high access to

PMTCT services and Côte d’Ivoire, where the HIV prevalence is lower, but counterbalanced

by a low access to PMTCT services. The four strategies the CEPAC group propose to model

are:

Current EID programmes: women known to be HIV-infected are asked to bring their

infants for EID at 6 weeks of age. All infants receive PCR tests.

Rapid antibody testing for all infants attending immunization clinics followed by PCR

testing for those who were positive

PCR tests in all infants attending immunization clinics

Novel POC assays for all children attending immunization clinics with same day

results

In the South African context, where HIV prevalence is much higher, the evaluation of EID

performed at birth to detect all children infected in utero compared to the six-week testing

strategy showed that the latter missed one fifth of perinatally HIV-infected infants and

delayed ARTY initiation (258).

Future research using the expanded EID module include addressing emerging questions in

EID such as the optimal timing of virological tests, the role of repeat testing and the impact of

routine antibody based tests at 18-24 months of age.

171

8.4.3. How to improve retention in care programmes?

Another research gap than deserves attention is retention in care in HIV-infected children,

before and after ART initiation. A recent literature review reported that most HIV-infected

children are eligible for ART since their inclusion in a paediatric care programme (117) but

the operational difficulties related to ART access lead to late treatment initiation and many

children die or are LTFU before appropriate care. In the absence of ART, morbidity and

mortality is severe, leading to substantial healthcare resource utilisation. Improving retention

in care programmes in children who are waiting to initiate ART could significantly reduce

morbidity and limit the associated costs. Few studies have investigated retention in pre-ART

care in children, or documented reasons for LTFU in these children, waiting to access ART. A

study in Kenya reported that the major barrier to HIV care was stigma and disclosure to the

family (259), then mortality.

West Africa is the region the least implicated in pre-ART retention of HIV-infected children:

a comparative study within the IeDEA consortium showed that only 60% of clinics chased-up

LTFU children compared to 90% in South Africa (260). Findings differed substantially across

regions but raise overall concerns about delayed ART start, low access to free HIV services

for children, and increased workload on programme retention in lower-income countries. Data

on reasons for LTFU before ART initiation would also allow identifying the most appropriate

strategies to improve pre-ART retention.

LTFU of children on ART is a major barrier to effective care programmes. Reasons for LTFU

are not well described as investigators do not know if the patient voluntarily (fear of stigma,

high costs of care) or involuntarily (death) interrupted care (259). For example, the adjusted

risk of loss to follow-up was 4.1% in Asia, 9.0% in Southern Africa, 14.0% in East Africa,

and 21.8% in West Africa (P < 0.01). Age < 12 months, NNRTI-based ART regimen, WHO

stage 4 at ART initiation, ART initiation after 2005, attending a public sector or a nonurban

clinic, having to pay for laboratory tests or antiretroviral drugs, larger cohort size, and living

in East Africa or West Africa were significantly associated with higher rates of loss-to-

follow-up (115).

Strategies to better understand and address LTFU in HIV-infected children need to be

identified as well as their associated costs. In South Africa for instance, a cost-effectiveness

analysis of a pilot intervention destined to track HIV-infected adults proved the latter not to

be cost-effective (261). On the other hand, a study in Côte d’Ivoire, a lower HIV prevalent

setting, in an adult population and based on the CEPAC model, reported all studied

interventions to be cost-effective providing their efficacy was at least 12% (262). Strategies to

improve retention in care in HIV-infected children, including financial incentives, must be

investigated in low-prevalent contexts such as West Africa. In addition to the LTFU issue, to

achieve universal access to HIV treatment, HIV treatment sites need to be decentralised to

lower-level facilities within or close to communities (reducing transportation costs); maternal

172

and child health clinics are ideal settings for this purpose. However, this requires the proper

training of providers for which costs need to be evaluated.

8.4.4. How to monitor ART response? What are the switching strategies when

treating HIV-infected on the long-term?

Since 2013, the WHO recommends initiating ART in all children aged less than 5 years.

However, the most efficient strategies for monitoring paediatric ART in view of limited

resources and the alternative first-line ART regimens remain unknown. There is a need to

identify effective routine clinical monitoring strategies in order to improve follow-up of ART-

treated children and optimise their standard of care.

Children who are failing their treatment must be identified the earliest possible and

encouraged to resort to care. In the absence of virological testing, effective routine monitoring

strategies must be identified to prevent clinical failure. This question has been investigated in

the Anti-Retroviral Research fOr Watoto (ARROW) trial, a 5 year clinical trial of anti-HIV

therapy in 1200 children with HIV or AIDS in Uganda and Zimbabwe. The trial aims to

investigate two main questions: can anti-HIV drugs be given in the absence of routine

laboratory tests, relying on clinical assessments instead; and whether starting children on 4

anti-HIV drugs for a short period of time before continuing with 3 drugs is better over the

long term than starting on the standard 3 drugs. To answer this first question, all children

included in the trial had routine laboratory tests carried out but half of them (selected at

random) did not have their results returned to their doctors throughout the trial, i.e. the clinical

examination alone was used to make health management decisions. The recently published

results of this trial have shown that children on HIV treatment can be safely monitored

without the need for expensive routine laboratory tests (160). These findings mean treatment

can be given much more cheaply, which may help more children get access to life-saving

treatment.

In a context where virological and immunological monitoring is not always available, other

indicators, such as weight or height could be used in clinical algorithms to monitor treatment

response. Indeed, many studies have reported robust increases in height and weight when on

ART (263, 264). However, no reference values for weight or height gain exist among children

receiving ART. Yotebieng et al have constructed reference curves for monitoring 6 months

gains in weight, height as well as CD4 count and CD4% in children initiating ART and these

newly developed reference distributions could be a good alternative to viral load or CD4

measurements (219). Immune, virological and growth response to ART need to be modelled

jointly and the performance of growth indicators evaluated compared to the virological gold

standard. If this is feasible, the cost-effectiveness of each strategy then needs to be evaluated.

The projected impacts of clinical, CD4 and viral load monitoring strategies for children on

ART can be examined using the CEPAC-Paediatric model. The model will simulate key

173

combinations of clinical, CD4 and viral load monitoring interventions using ART switching

criteria from the WHO guidelines. All strategies will assume available second line ART,

which needs to be prioritised in low-income settings. The impact of the switch criteria (RNA

thresholds, CD4 drops) as well as first-line ART regimen (IP vs NNRTI), variations in clinical

risks and costs of care and the availability of 3rd line ART will be examined in key sensitivity

analyses. Similarly, the cost-effectiveness of these strategies will need to be carefully

analysed, comparing age at ART initiation and the occurrence of post-ART severe morbidity

and mortality and associated costs.

Other research perspectives include a full analysis of the organisational processes involved in

the delivery of healthcare in resource-limited settings, with an aim at improving linkage

between PMTCT programmes and paediatric care programmes, as well as reference

laboratories (265, 266). The social implications of HIV must not be neglected either, and in

the era where task-shifting is expanding, knowledge, aptitude and practices of both nursing

staff and caregivers must be evaluated (267-269).

174

Conclusion This work, mostly conducted in Abidjan, Côte d’Ivoire but also in other West African urban

settings, highlights the burden of paediatric HIV and the clinical and operational difficulties in

taking care of HIV-infected children in this context. So far, this region has the lowest

coverage for HIV EID and early ART leading to high rates of mortality, morbidity and

substantial healthcare resource utilisation. Despite an effective scale up of paediatric ART in

this region since 2004, there remains a high burden of residual infectious morbidity both HIV-

related and non-related. Finally, children initiate ART at a late age and advanced stage of the

disease. Thus, not all experience immune recovery; the rates for this being highest in those

treated the earliest. Furthermore, we noted limited resource to healthcare resulting in many

untreated diseases.

This PhD work was conducted in the context of operational research in West Africa,

underlining many methodological challenges encountered in the field conditions. Although

our results describe as best as possible the situation of paediatric HIV in West Africa during

the study period, they are based mostly on children aged > 2 years, who have survived early

infancy and who are therefore selected.

Most of the barriers to early access to care that were identified throughout these studies can

easily be improved. However, there remain core elements such as the lack of human and

material resources, lack of funding and appropriate healthcare systems, which, if not

addressed correctly, could jeopardise the current processes involved in paediatric HIV care. It

is urgent to diagnose and treat HIV-infected infants the earliest possible, in order to reduce

childhood mortality and prolong life expectancy. We conclude that universal free access to

ART services and innovative approaches are urgently needed to improve both enrolment and

paediatric outcomes at the programme level. Characteristics associated with favourable

paediatric enrolment and retention should be identified, further assessed and expanded to

improve paediatric HIV treatment in Africa in HIV programmes.

In this context, we feel that simulation modelling can add to traditional research methods by

examining strategies for which trials are not feasible and, more importantly for paediatric HIV

intervention, by projecting the clinical outcomes and costs beyond the time horizon of clinical

studies. This could inform healthcare policy-makers on the most effective and cost-effective

strategies addressing the current questions in paediatric HIV care.

Many questions are pending:

Which public health approaches to access HIV-exposed infants?

Which public health approaches and what services linkages to optimise early HIV

diagnosis in children and increase their access to early ART initiation?

What interventions to improve retention in care?

175

Which long-term strategies to treat children who initiated early ART?

Which monitoring strategies for children who initiate ART?

Which are the most cost-effective interventions to free up resources and address larger

size cohorts?

There is still some way to go before the elimination of MTCT in resource-limited settings,

and although every effort should certainly be made to achieve this goal, we must not forget

the millions of children already living with HIV who urgently need improved, affordable and

appropriate treatment (8).

176

References 1. World Health Organisation. UNAIDS report on the global AIDS epidemic 2013 2013. Available from: http://www.unaids.org/en/media/unaids/contentassets/documents/epidemiology/2013/gr2013/UNAIDS_Global_Report_2013_en.pdf. 2. World Health Organisation. UNAIDS Report on the global AIDS epidemic 2012. Available from: http://www.unaids.org/en/media/unaids/contentassets/documents/epidemiology/2012/gr2012/20121120_UNAIDS_Global_Report_2012_en.pdf. 3. Torpey K, Kasonde P, Kabaso M, Weaver MA, Bryan G, Mukonka V, et al. Reducing pediatric HIV infection: estimating mother-to-child transmission rates in a program setting in Zambia. J Acquir Immune Defic Syndr 2010;54(4):415-22. 4. Tonwe-Gold B, Ekouevi DK, Bosse CA, Toure S, Koné M, Becquet R, et al. Implementing family-focused HIV care and treatment: the first 2 years' experience of the mother-to-child transmission-plus program in Abidjan, Côte d'Ivoire. Trop Med Int Health 2009;14(2):204-12. 5. Creek T, Ntumy R, Mazhani L, Moore J, Smith M, Han G, et al. Factors associated with low early uptake of a national program to prevent mother to child transmission of HIV (PMTCT): results of a survey of mothers and providers, Botswana, 2003. AIDS Behav. 2009;13(2):356-64. 6. Ciaranello AL, Perez F, Keatinge J, Park JE, Engelsmann B, Maruva M, et al. What Will It Take to Eliminate Pediatric HIV? Reaching WHO Target Rates of Mother-to-Child HIV Transmission in Zimbabwe: A Model-Based Analysis. PLoS Med. 2012;9(1):e1001156. 7. World Health Organisation. UNAIDS report on the global AIDS epidemic 2011. Available from: http://www.unaids.org/en/media/unaids/contentassets/documents/unaidspublication/2011/20111130_ua_report_en.pdf. 8. Lallemant M, Chang S, Cohen R, Pecoul B. Pediatric HIV--a neglected disease? N Engl J Med. 2011;365(7):581-3. 9. Bland RM. Management of HIV-infected children in Africa: progress and challenges. Arch Dis Child. 2011;96(10):911-5. 10. Goldie SJ, Yazdanpanah Y, Losina E, Weinstein MC, Anglaret X, Walensky RP, et al. Cost-effectiveness of HIV treatment in resource-poor settings--the case of Cote d'Ivoire. N Engl J Med. 2006;355(11):1141-53. 11. Walensky RP, Wolf LL, Wood R, Fofana MO, Freedberg KA, Martinson NA, et al. When to start antiretroviral therapy in resource-limited settings. Annals of internal medicine. 2009;151(3):157-66. 12. Yazdanpanah Y, Losina E, Anglaret X, Goldie SJ, Walensky RP, Weinstein MC, et al. Clinical impact and cost-effectiveness of co-trimoxazole prophylaxis in patients with HIV/AIDS in Côte d'Ivoire: a trial-based analysis. AIDS 2005;19(12):1299-308. 13. World Health Organisation. Global Plan towards the elimination of new HIV infections among children by 2015 and keeping their mothers alive 2011. Available from: http://www.unaids.org/en/media/unaids/contentassets/documents/unaidspublication/2011/20110609_jc2137_global-plan-elimination-hiv-children_en.pdf. 14. De Cock KM, Fowler MG, Mercier E, de Vincenzi I, Saba J, Hoff E, et al. Prevention of mother-to-child HIV transmission in resource-poor countries: translating research into policy and practice. JAMA. 2000;283(9):1175-82. 15. Connor EM, Sperling RS, Gelber R, Kiselev P, Scott G, O'Sullivan MJ, et al. Reduction of maternal-infant transmission of human immunodeficiency virus type 1 with

177

zidovudine treatment. Pediatric AIDS Clinical Trials Group Protocol 076 Study Group. N Engl J Med. 1994;331(18):1173-80. 16. Dabis F, Msellati P, Meda N, Welffens-Ekra C, You B, Manigart O, et al. 6-month efficacy, tolerance, and acceptability of a short regimen of oral zidovudine to reduce vertical transmission of HIV in breastfed children in Cote d'Ivoire and Burkina Faso: a double-blind placebo-controlled multicentre trial. DITRAME Study Group. DIminution de la Transmission Mere-Enfant. Lancet. 1999;353(9155):786-92. 17. Guay LA, Musoke P, Fleming T, Bagenda D, Allen M, Nakabiito C, et al. Intrapartum and neonatal single-dose nevirapine compared with zidovudine for prevention of mother-to-child transmission of HIV-1 in Kampala, Uganda: HIVNET 012 randomised trial. Lancet. 1999;354(9181):795-802. 18. Wiktor SZ, Ekpini E, Karon JM, Nkengasong J, Maurice C, Severin ST, et al. Short-course oral zidovudine for prevention of mother-to-child transmission of HIV-1 in Abidjan, Cote d'Ivoire: a randomised trial. Lancet. 1999;353(9155):781-5. 19. Saba J for the Petra study team. Efficacy of three short-course regimens of zidovudine and lamivudine in preventing early and late transmission of HIV-1 from mother to child in Tanzania, South Africa, and Uganda (Petra study): a randomised, double-blind, placebo-controlled trial. Lancet. 2002;359(9313):1178-86. 20. Moodley D, Moodley J, Coovadia H, Gray G, McIntyre J, Hofmyer J, et al. A multicenter randomized controlled trial of nevirapine versus a combination of zidovudine and lamivudine to reduce intrapartum and early postpartum mother-to-child transmission of human immunodeficiency virus type 1. J Infect Dis. 2003;187(5):725-35. 21. World Health Organisation. Prevention of mother-to-child transmission of HIV. Use of Nevirapine among women of unknown serostatus 2001. Available from: http://www.who.int/hiv/pub/mtct/en/isbn9241562129.pdf. 22. Leroy V, Ekouevi DK, Becquet R, Viho I, Dequae-Merchadou L, Tonwe-Gold B, et al. 18-month effectiveness of short-course antiretroviral regimens combined with alternatives to breastfeeding to prevent HIV mother-to-child transmission. PloS One. 2008;3(2):e1645-e. 23. Tonwe-Gold B, Ekouevi DK, Viho I, Amani-Bosse C, Toure S, Coffie Pa, et al. Antiretroviral treatment and prevention of peripartum and postnatal HIV transmission in West Africa: evaluation of a two-tiered approach. PLoS medicine. 2007;4(8):e257-e. 24. World Health Organisation. Programmatic update: use of antiretroviral drugs for treating pregnant women and preventing HIV infection in infants. Geneva, Switzerland2012. Available from: http://whqlibdoc.who.int/hq/2012/WHO_HIV_2012.6_eng.pdf. 25. World Health Organisation. The use of antiretroviral drugs for treating and preventing HIV infection 2013. Available from: http://apps.who.int/iris/bitstream/10665/85321/1/9789241505727_eng.pdf. 26. World Health Organisation. Progress report on the global plan towards the elimination of new HIV infections among children by 2015 and keeping their mothers alive 2013. Available from: http://www.unaids.org/en/media/unaids/contentassets/documents/unaidspublication/2013/20130625_progress_global_plan_en.pdf. 27. World Health Organisation. Global AIDS response progress reporting 2013. Available from: http://www.unaids.org/en/media/unaids/contentassets/documents/epidemiology/2013/gr2013/UNAIDS_Global_Report_2013_en.pdf. 28. World Health Organisation. Towards Universal access: Scaling up priority HIV/AIDS interventions in the health sector. Progress Report 2010. Available from: http://www.who.int/hiv/pub/2010progressreport/summary_en.pdf.

178

29. World Health Organisation. Together we will end AIDS 2012. Available from: http://www.unaids.org/en/media/unaids/contentassets/documents/epidemiology/2012/jc2296_unaids_togetherreport_2012_en.pdf. 30. Mahy M, Stover J, Kiragu K, Hayashi C, Akwara P, Luo C, et al. What will it take to achieve virtual elimination of mother-to-child transmission of HIV? An assessment of current progress and future needs. Sex Transm Infect 2010;86 Suppl 2:ii48-55. 31. Ciaranello AL, Perez F, Maruva M, Chu J, Engelsmann B, Keatinge J, et al. WHO 2010 Guidelines for Prevention of Mother-to-Child HIV Transmission in Zimbabwe: Modeling Clinical Outcomes in Infants and Mothers. PloS One. 2011;6(6):e20224-e. 32. Ciaranello AL, Perez F, Engelsmann B, Walensky RP, Mushavi A, Rusibamayila A, et al. Cost-effectiveness of World Health Organization 2010 guidelines for prevention of mother-to-child HIV transmission in Zimbabwe. Clin Infect Dis. 2013;56(3):430-46. 33. Doherty K, Ciaranello A. What is needed to eliminate new pediatric HIV infections: the contribution of model-based analyses. Curr Opin HIV AIDS 2013;8(5):0. 34. Werayingyong P, Phanuphak N, Chokephaibulkit K, Tantivess S, Kullert N, Tosanguan K, et al. Economic Evaluation of 3-Drug Antiretroviral Regimens for the Prevention of Mother-to-Child HIV Transmission in Thailand. Asia Pac J Public Health. 2013. 35. World Health Organisation. Children and AIDS : fifth stocktaking report 2010. Available from: http://www.unicef.org/publications/files/Children_and_AIDS-Fifth_Stocktaking_Report_2010_EN.pdf. 36. Tudor Car L, Van Velthoven MH, Brusamento S, Elmoniry H, Car J, Majeed A, et al. Integrating prevention of mother-to-child HIV transmission programs to improve uptake: a systematic review. PloS One. 2012;7(4):e35268. 37. Audureau E, Kahn JG, Besson MH, Saba J, Ladner J. Scaling up prevention of mother-to-child HIV transmission programs in sub-Saharan African countries: a multilevel assessment of site-, program- and country-level determinants of performance. BMC Public Health. 2013;13:286. 38. Morris MB, Chapula BT, Chi BH, Mwango A, Chi HF, Mwanza J, et al. Use of task-shifting to rapidly scale-up HIV treatment services: experiences from Lusaka, Zambia. BMC Health Serv Res. 2009;9:5. 39. Shumbusho F, van Griensven J, Lowrance D, Turate I, Weaver MA, Price J, et al. Task shifting for scale-up of HIV care: evaluation of nurse-centered antiretroviral treatment at rural health centers in Rwanda. PLoS Med. 2009;6(10):e1000163. 40. Chi BH, Bolton-Moore C, Holmes CB. Prevention of mother-to-child HIV transmission within the continuum of maternal, newborn, and child health services. Curr Opin HIV AIDS 2013;8(5):497-502. 41. Perez F, Zvandaziva C, Engelsmann B, Dabis F. Acceptability of routine HIV testing ("opt-out") in antenatal services in two rural districts of Zimbabwe. J Acquir Immune Defic Syndr. 2006;41(4):514-20. 42. Wettstein C, Mugglin C, Egger M, Blaser N, Vizcaya LS, Estill J, et al. Missed opportunities to prevent mother-to-child-transmission: systematic review and meta-analysis. AIDS. 2012;26(18):2361-73. Epub 2012/09/06. 43. Black RE, Cousens S, Johnson HL, Lawn JE, Rudan I, Bassani DG, et al. Global, regional, and national causes of child mortality in 2008: a systematic analysis. Lancet. 2010;375(9730):1969-87. 44. Walker N, Schwartländer B, Bryce J. Meeting international goals in child survival and HIV/AIDS. Lancet. 2002;360(9329):284-9. 45. Children born to women with HIV-1 infection: natural history and risk of transmission. European Collaborative Study. Lancet. 1991;337(8736):253-60.

179

46. Spira R, Lepage P, Msellati P, Van de Perre P, Leroy V, Simonon a, et al. Natural History of Human Immunodefiency Virus Type 1 Infection in Children: A Five-Year Prospective Study in Rwanda. Pediatrics. 1999;104(5):e56-e. 47. Newell M-L, Brahmbhatt H, Ghys PD. Child mortality and HIV infection in Africa: a review. AIDS 2004;18 Suppl 2:S27-34. 48. Newell M-L, Coovadia H, Cortina-Borja M, Rollins N, Gaillard P, Dabis F. Mortality of infected and uninfected infants born to HIV-infected mothers in Africa: a pooled analysis. Lancet. 2004;364(9441):1236-43. 49. Becquet R, Marston M, Dabis F, Moulton LH, Gray G, Coovadia HM, et al. Children who acquire HIV infection perinatally are at higher risk of early death than those acquiring infection through breastmilk: a meta-analysis. PloS One. 2012;7(2):e28510. 50. Marston M, Becquet R, Zaba B, Moulton LH, Gray G, Coovadia H, et al. Net survival of perinatally and postnatally HIV-infected children: a pooled analysis of individual data from sub-Saharan Africa. Int J Epidemiol. 2011;40(2):385-96. 51. Dabis F, Elenga N, Meda N, Leroy V, Viho I, Manigart O, et al. 18-Month mortality and perinatal exposure to zidovudine in West Africa. AIDS 2001;15(6):771-9. 52. Nte AR, Eneh AU. HIV infection and persistent diarrhoea: a comparative study of HIV positive and HIV negative children. Afr J Med Med Sci. 2008;37(2):149-55. 53. Thea DM, St Louis ME, Atido U, Kanjinga K, Kembo B, Matondo M, et al. A prospective study of diarrhea and HIV-1 infection among 429 Zairian infants. N Engl J Med. 1993;329(23):1696-702. 54. Lucas SB, Peacock CS, Hounnou A, Brattegaard K, Koffi K, Hondé M, et al. Disease in children infected with HIV in Abidjan, Côte d'Ivoire. BMJ 1996;312(7027):335-8. 55. Gray DM, Zar HJ. Community-acquired pneumonia in HIV-infected children: a global perspective. Curr Opin Pulm Med. 2010;16(3):208-16. 56. Berger MR, Fields-Gardner C, Wagle A, Hollenbeck CB. Prevalence of malnutrition in human immunodeficiency virus/acquired immunodeficiency syndrome orphans in the Nyanza province of Kenya: a comparison of conventional indexes with a composite index of anthropometric failure. J Am Diet Assoc. 2008;108(6):1014-7. 57. Elenga N, Kouakoussui KA, Bonard D, Fassinou P, Anaky M-F, Wemin M-L, et al. Diagnosed tuberculosis during the follow-up of a cohort of human immunodeficiency virus-infected children in Abidjan, Côte d'Ivoire: ANRS 1278 study. Pediatr Infect Dis J. 2005;24(12):1077-82. 58. Flateau C, Le Loup G, Pialoux G. Consequences of HIV infection on malaria and therapeutic implications: a systematic review. Lancet Infect Dis. 2011;11(7):541-56. 59. Kalyesubula I, Musoke-Mudido P. Effects of malaria infection in human immunodeficiency virus type 1-infected Ugandan children. Pediatr Infect Dis J. 1997;16(9):871-81. 60. Cross Continents Collaboration for Kids (3Cs4Kids). Markers for predicting mortality in untreated HIV-infected children in resource-limited settings: a meta-analysis. AIDS 2008;22(1):97-105. 61. Ayiko R, Antai D, Kulane A. Trends and determinants of under-five mortality in Uganda. East Afr J Public Health. 2009;6(2):136-40. 62. Brahmbhatt H, Kigozi G, Wabwire-Mangen F, Serwadda D, Lutalo T, Nalugoda F, et al. Mortality in HIV-infected and uninfected children of HIV-infected and uninfected mothers in rural Uganda. J Acquir Immune Defic Syndr 2006;41(4):504-8. 63. Krug A, Pattinson RC, Power DJ. Why children die: an under-5 health care survey in Mafikeng region. S Afr Med J. 2004;94(3):202-6.

180

64. Marinda E, Humphrey JH, Iliff PJ, Mutasa K, Nathoo KJ, Piwoz EG, et al. Child mortality according to maternal and infant HIV status in Zimbabwe. Pediatr Infect Dis J. 2007;26(6):519-26. 65. Nakiyingi JS, Bracher M, Whitworth JA, Ruberantwari A, Busingye J, Mbulaiteye SM, et al. Child survival in relation to mother's HIV infection and survival: evidence from a Ugandan cohort study. AIDS 2003;17(12):1827-34. 66. Obimbo EM, Mbori-Ngacha DA, Ochieng JO, Richardson BA, Otieno PA, Bosire R, et al. Predictors of early mortality in a cohort of human immunodeficiency virus type 1-infected african children. Pediatr Infect Dis J 2004;23(6):536-43. 67. Prendergast A, Tudor-Williams G, Jeena P, Burchett S, Goulder P. International perspectives, progress, and future challenges of paediatric HIV infection. Lancet. 2007;370(9581):68-80. 68. Nielsen K, Bryson YJ. Diagnosis of HIV infection in Children. Pediatr Clin North Am 2000;47(1):39-63. 69. Sherman GG, Cooper PA, Coovadia AH, Puren AJ, Jones SA, Mokhachane M, et al. Polymerase chain reaction for diagnosis of human immunodeficiency virus infection in infancy in low resource settings. Pediatr Infect Dis J. 2005;24(11):993-7. 70. Bremer JW, Lew JF, Cooper E, Hillyer GV, Pitt J, Handelsman E, et al. Diagnosis of infection with human immunodeficiency virus type 1 by a DNA polymerase chain reaction assay among infants enrolled in the Women and Infants' Transmission Study. The Journal of pediatrics. 1996;129(2):198-207. Epub 1996/08/01. 71. Parpia ZA, Elghanian R, Nabatiyan A, Hardie DR, Kelso DM. p24 antigen rapid test for diagnosis of acute pediatric HIV infection. J Acquir Immune Defic Syndr. 2010;55(4):413-9. 72. Dunn DT, Brandt CD, Krivine A, Cassol SA, Roques P, Borkowsky W, et al. The sensitivity of HIV-1 DNA polymerase chain reaction in the neonatal period and the relative contributions of intra-uterine and intra-partum transmission. AIDS. 1995;9(9):F7-11. 73. Creek TL, Sherman GG, Nkengasong J, Lu L, Finkbeiner T, Fowler MG, et al. Infant human immunodeficiency virus diagnosis in resource-limited settings: issues, technologies, and country experiences. American journal of obstetrics and gynecology. 2007;197(3 Suppl):S64-71. 74. Read JS. Diagnosis of HIV-1 infection in children younger than 18 months in the United States. Pediatrics. 2007;120(6):e1547-62. 75. Ciaranello AL, Park J-E, Ramirez-Avila L, Freedberg KA, Walensky RP, Leroy V. Early infant HIV-1 diagnosis programs in resource limited settings: opportunities for improved outcomes and more cost-effective interventions. BMC Medicine. 2011;9(1):59. 76. Creek T, Tanuri A, Smith M, Seipone K, Smit M, Legwaila K, et al. Early diagnosis of human immunodeficiency virus in infants using polymerase chain reaction on dried blood spots in Botswana's national program for prevention of mother-to-child transmission. Pediatr Infect Dis J. 2008;27(1):22-6. 77. Khamadi S, Okoth V, Lihana R, Nabwera J, Hungu J, Okoth F, et al. Rapid identification of infants for antiretroviral therapy in a resource poor setting: the Kenya experience. J Trop Pediatr. 2008;54(6):370-4. 78. Menzies NA, Homsy J, Chang Pitter JY, Pitter C, Mermin J, Downing R, et al. Cost-effectiveness of routine rapid human immunodeficiency virus antibody testing before DNA-PCR testing for early diagnosis of infants in resource-limited settings. Pediatr Infect Dis J. 2009;28(9):819-25. 79. Lofgren SM, Morrissey AB, Chevallier CC, Malabeja AI, Edmonds S, Amos B, et al. Evaluation of a dried blood spot HIV-1 RNA program for early infant diagnosis and viral load monitoring at rural and remote healthcare facilities. AIDS. 2009;23(18):2459-66.

181

80. Yapo V, Toni TD, Desmonde S, Amani-Bosse C, Oga M, Lenaud S, et al. Evaluation of dried blood spot diagnosis using HIV1-DNA and HIV1-RNA Biocentric assays in infants in Abidjan, Cote d'Ivoire. The Pedi-Test DBS ANRS 12183 Study. J Virol Methods. 2013;193(2):439-45. 81. Viljoen J, Gampini S, Danaviah S, Valea D, Pillay S, Kania D, et al. Dried blood spot HIV-1 RNA quantification using open real-time systems in South Africa and Burkina Faso. J Acquir Immune Defic Syndr. 2010;55(3):290-8. 82. Stevens W, Erasmus L, Moloi M, Taleng T, Sarang S. Performance of a novel human immunodeficiency virus (HIV) type 1 total nucleic acid-based real-time PCR assay using whole blood and dried blood spots for diagnosis of HIV in infants. Journal of clinical microbiology. 2008;46(12):3941-5. 83. Ndondoki C, Brou H, Timite-Konan M, Oga M, Amani-Bosse C, Menan H, et al. Universal HIV Screening at Postnatal Points of Care: Which Public Health Approach for Early Infant Diagnosis in Cote d'Ivoire? PloS One. 2013;8(8):e67996. 84. Kellerman S, Essajee S. HIV testing for children in resource-limited settings: what are we waiting for? PLoS Med. 2010;7(7):e1000285. 85. Rollins N, Mzolo S, Moodley T, Esterhuizen T, van Rooyen H. Universal HIV testing of infants at immunization clinics: an acceptable and feasible approach for early infant diagnosis in high HIV prevalence settings. AIDS. 2009;23(14):1851-7. 86. Ndondoki C, Brou H, Timite M, Oga MA, Amani-Bosse C, menan H, et al. Universal HIV Screening at Postnatal Points of Care: Which public health approach for early infant diagnosis in Côte d'Ivoire ? PLosOne. 2013;in press. 87. Oga MA, Ndondoki C, Brou H, Salmon A, Bosse-Amani C, Timite-Konan M, et al. Attitudes and practices of health care workers toward routine HIV testing of infants in Cote d'Ivoire: the PEDI-TEST ANRS 12165 Project. J Acquir Immune Defic Syndr. 2011;57 Suppl 1:S16-21. 88. World Health Organisation. Antiretroviral Therapy of HIV infection in infants and children: towards universal access. Recommendations for a public health approach. 2006. Available from: http://www.who.int/hiv/pub/mtct/arv_guidelines_mtct.pdf. 89. Seidenberg P, Nicholson S, Schaefer M, Semrau K, Bweupe M, Masese N, et al. Early infant diagnosis of HIV infection in Zambia through mobile phone texting of blood test results. Bull World Health Organ. 2012;90(5):348-56. 90. McCollum ED, Johnson DC, Chasela CS, Siwande LD, Kazembe PN, Olson D, et al. Superior uptake and outcomes of early infant diagnosis of HIV services at an immunization clinic versus an "under-five" general pediatric clinic in Malawi. J Acquir Immune Defic Syndr. 2012;60(4):e107-10. 91. Dube Q, Dow A, Chirambo C, Lebov J, Tenthani L, Moore M, et al. Implementing early infant diagnosis of HIV infection at the primary care level: experiences and challenges in Malawi. Bull World Health Organ. 2012;90(9):699-704. 92. Ciampa PJ, Burlison JR, Blevins M, Sidat M, Moon TD, Rothman RL, et al. Improving retention in the early infant diagnosis of HIV program in rural Mozambique by better service integration. J Acquir Immune Defic Syndr. 2011;58(1):115-9. 93. Zar HJ, Dechaboon A, Hanslo D, Apolles P, Magnus KG, Hussey G. Pneumocystis carinii pneumonia in South African children infected with human immunodeficiency virus. Pediatr Infect Dis J. 2000;19(7):603-7. 94. Manyando C, Njunju EM, D'Alessandro U, Van Geertruyden JP. Safety and Efficacy of Co-Trimoxazole for Treatment and Prevention of Plasmodium falciparum Malaria: A Systematic Review. PloS One. 2013;8(2):e56916.

182

95. Sandison TG, Homsy J, Arinaitwe E, Wanzira H, Kakuru A, Bigira V, et al. Protective efficacy of co-trimoxazole prophylaxis against malaria in HIV exposed children in rural Uganda: a randomised clinical trial. BMJ. 2011;342:d1617. 96. Zachariah R, Harries AD, Luo C, Bachman G, Graham SM. Scaling-up co-trimoxazole prophylaxis in HIV-exposed and HIV-infected children in high HIV-prevalence countries. Lancet Infect Dis 2007;7(10):686-93. 97. Graham SM. Cotrimoxazole prophylaxis for infants exposed to HIV infection. Bull World Health Organ. 2004;82(4):297-8. 98. Graham SM, Mtitimila EI, Kamanga HS, Walsh AL, Hart CA, Molyneux ME. Clinical presentation and outcome of Pneumocystis carinii pneumonia in Malawian children. Lancet. 2000;355(9201):369-73. 99. Mermin J, Lule J, Ekwaru JP, Malamba S, Downing R, Ransom R, et al. Effect of co-trimoxazole prophylaxis on morbidity, mortality, CD4-cell count, and viral load in HIV infection in rural Uganda. Lancet.364(9443):1428-34. 100. Chintu C, Bhat GJ, Walker AS, Mulenga V, Sinyinza F, Lishimpi K, et al. Co-trimoxazole as prophylaxis against opportunistic infections in HIV-infected Zambian children (CHAP): a double-blind randomised placebo-controlled trial. Lancet. 2004;364(9448):1865-71. 101. World Health Organisation. Guidelines on Co-trimoxazole prophylaxis for HIV-related infections among children, adolescents and adults. Recommendations for a public health approach 2006. Available from: http://www.who.int/hiv/pub/guidelines/ctxguidelines.pdf. 102. Moodley D, Reddy L, Mahungo W, Masha R. Factors associated with coverage of cotrimoxazole prophylaxis in HIV-exposed children in South Africa. PloS One. 2013;8(5):e63273. 103. UNICEF. Cotrimoxazole prophylaxis for HIV-exposed and HIV-infected infants and children 2009. Available from: http://www.unicef.org/aids/files/CotrimoxazoleGuide_2009.pdf. 104. Bolton-Moore C, Mubiana-Mbewe M, Cantrell RA, Chintu N, Stringer EM, Chi BH, et al. Clinical outcomes and CD4 cell response in children receiving antiretroviral therapy at primary health care facilities in Zambia. JAMA. 2007;298(16):1888-99. 105. Arrivé E, Kyabayinze DJ, Marquis B, Tumwesigye N, Kieffer M-P, Azondekon A, et al. Cohort Profile: The Paediatric Antiretroviral Treatment Programmes in Lower-Income Countries (KIDS-ART-LINC) Collaboration. Int J Epidemiol 2008;37(3):474-80. 106. Sutcliffe CG, van Dijk JH, Bolton C, Persaud D, Moss WJ. Effectiveness of antiretroviral therapy among HIV-infected children in sub-Saharan Africa. Lancet Infect Dis. 2008;8(8):477-89. 107. Mubiana-Mbewe M, Bolton-Moore C, Banda Y, Chintu N, Nalubamba-Phiri M, Giganti M, et al. Causes of morbidity among HIV-infected children on antiretroviral therapy in primary care facilities in Lusaka, Zambia. Trop Med Int Health. 2009;14(10):1190-8. 108. Puthanakit T, Oberdorfer P, Akarathum N, Wannarit P, Sirisanthana T, Sirisanthana V. Immune reconstitution syndrome after highly active antiretroviral therapy in human immunodeficiency virus-infected thai children. Pediatr Infect Dis J. 2006;25(1):53-8. 109. Smith K, Kuhn L, Coovadia A, Meyers T, Hu CC, Reitz C, et al. Immune reconstitution inflammatory syndrome among HIV-infected South African infants initiating antiretroviral therapy. AIDS. 2009;23(9):1097-107. 110. Eley B, Davies MA, Apolles P, Cowburn C, Buys H, Zampoli M, et al. Antiretroviral treatment for children. S Afr Med J. 2006;96(9 Pt 2):988-93. 111. van Kooten Niekerk NKM, Knies MM, Howard J, Rabie H, Zeier M, van Rensburg a, et al. The first 5 years of the family clinic for HIV at Tygerberg Hospital: family

183

demographics, survival of children and early impact of antiretroviral therapy. J Trop Pediatr. 2006;52(1):3-11. 112. Kids-ART-Linc-Collaboration. Low risk of death, but substantial program attrition, in pediatric HIV treatment cohorts in Sub-Saharan Africa. J Acquir Immune Defic Syndr. 2008;49(5):523-31. 113. Wamalwa DC, Obimbo EM, Farquhar C, Richardson BA, Mbori-Ngacha DA, Inwani I, et al. Predictors of mortality in HIV-1 infected children on antiretroviral therapy in Kenya: a prospective cohort. BMC Pediatrics. 2010;10:33-. 114. O'Brien DP, Sauvageot D, Zachariah R. In resource-limited settings good early outcomes can be achieved in children using adult fixed-dose combination antiretroviral therapy. AIDS 2006;20(15):1955-60. 115. Leroy V, Malateste K, Rabie H, Lumbiganon P, Ayaya S, Dicko F, et al. Outcomes of antiretroviral therapy in children in Asia and Africa: a comparative analysis of the IeDEA pediatric multiregional collaboration. J Acquir Immune Defic Syndr. 2013;62(2):208-19. Epub 2012/11/29. 116. Rosen S, Fox MP. Retention in HIV Care between Testing and Treatment in Sub-Saharan Africa: A Systematic Review. PLoS medicine. 2011;8(7):e1001056-e. 117. Mugglin C, Wandeler G, Estill J, Egger M, Bender N, Davies MA, et al. Retention in Care of HIV-Infected Children from HIV Test to Start of Antiretroviral Therapy: Systematic Review. PloS One. 2013;8(2):e56446. 118. Okomo U, Togun T, Oko F, Peterson K, Jaye A. Mortality and loss to programme before antiretroviral therapy among HIV-infected children eligible for treatment in The Gambia, West Africa. AIDS Res Ther. 2012;9(1):28. 119. Anaky MF, Duvignac J, Wemin L, Kouakoussui A, Karcher S, Touré S, et al. Scaling up antiretroviral therapy for HIV-infected children in Côte d'Ivoire: determinants of survival and loss to programme. Bull World Health Organ. 2010;88(7):490-9. 120. Sutcliffe CG, van Dijk JH, Munsanje B, Hamangaba F, Siniwymaanzi P, Thuma PE, et al. Risk Factors for Pre-Treatment Mortality among HIV-Infected Children in Rural Zambia: A Cohort Study. PLoS One. 2011;6(12):e29294. 121. Raguenaud ME, Isaakidis P, Zachariah R, Te V, Soeung S, Akao K, et al. Excellent outcomes among HIV+ children on ART, but unacceptably high pre-ART mortality and losses to follow-up: a cohort study from Cambodia. BMC Pediatr. 2009;9:54. 122. Kurewa NE, Gumbo FZ, Mapingure PM, Munjoma MW, Chirenje MZ, Rusakaniko S, et al. Predictors of Attrition among Children Born in a PMTCT Programme in Zimbabwe Followed up over 5 Years. J Trop Pediatr. 2012. 123. Yeap AD, Hamilton R, Charalambous S, Dwadwa T, Churchyard GJ, Geissler PW, et al. Factors influencing uptake of HIV care and treatment among children in South Africa - a qualitative study of caregivers and clinic staff. AIDS Care. 2010;22(9):1101-7. 124. Vreeman RC, Nyandiko WM, Ayaya SO, Walumbe EG, Marrero DG, Inui TS. The perceived impact of disclosure of pediatric HIV status on pediatric antiretroviral therapy adherence, child well-being, and social relationships in a resource-limited setting. AIDS Patient Care STDS. 2010;24(10):639-49. 125. Brouwer CN, Lok CL, Wolffers I, Sebagalls S. Psychosocial and economic aspects of HIV/AIDS and counselling of caretakers of HIV-infected children in Uganda. AIDS Care. 2000;12(5):535-40. 126. Goetghebuer T, Haelterman E, Le Chenadec J, Dollfus C, Gibb D, Judd A, et al. Effect of early antiretroviral therapy on the risk of AIDS/death in HIV-infected infants. AIDS. 2009;23(5):597-604.

184

127. Faye A, Le Chenadec J, Dollfus C, Thuret I, Douard D, Firtion G, et al. Early versus deferred antiretroviral multidrug therapy in infants infected with HIV type 1. Clin Infect Dis. 2004;39(11):1692-8. 128. Newell ML, Patel D, Goetghebuer T, Thorne C. CD4 cell response to antiretroviral therapy in children with vertically acquired HIV infection: is it associated with age at initiation? J Infect Dis. 2006;193(7):954-62. 129. Violari A, Cotton MF, Gibb DM, Babiker AG, Steyn J, Madhi SA, et al. Early antiretroviral therapy and mortality among HIV-infected infants. N Engl J Med. 2008;359(21):2233-44. 130. World Health Organisation. Antiretroviral Therapy for HIV infection in infants and children: towards universal access. Recommendations for a public health approach. 2010 Revision 2010. Available from: http://whqlibdoc.who.int/publications/2010/9789241599801_eng.pdf. 131. Penazzato M, Crowley S, Mofenson L, Franceschetto G, Nannyonga MM, Mazza A, et al. Programmatic impact of the evolution of WHO pediatric antiretroviral treatment guidelines for resource-limited countries (Tukula Fenna Project, Uganda). J Acquir Immune Defic Syndr. 2012;61(4):522-5. 132. Puthanakit T, Saphonn V, Ananworanich J, Kosalaraksa P, Hansudewechakul R, Vibol U, et al. Early versus deferred antiretroviral therapy for children older than 1 year infected with HIV (PREDICT): a multicentre, randomised, open-label trial. Lancet Infect Dis. 2012;12(12):933-41. 133. Schomaker M, Egger M, Ndirangu J, Phiri S, Moultrie H, Technau K, et al. When to start antiretroviral therapy in children aged 2-5 years: a collaborative causal modelling analysis of cohort studies from southern Africa. PLoS Med. 2013;10(11):e1001555. Epub 2013/11/22. 134. Leroy V, Dabis F. When should therapy begin for children infected with HIV? Lancet Infect Dis. 2012;12(12):900-2. 135. Arrive E, Newell ML, Ekouevi DK, Chaix ML, Thiebaut R, Masquelier B, et al. Prevalence of resistance to nevirapine in mothers and children after single-dose exposure to prevent vertical transmission of HIV-1: a meta-analysis. Int J Epidemiol. 2007;36(5):1009-21. 136. Lockman S, Shapiro RL, Smeaton LM, Wester C, Thior I, Stevens L, et al. Response to antiretroviral therapy after a single, peripartum dose of nevirapine. N Engl J Med. 2007;356(2):135-47. 137. Musiime V, Ssali F, Kayiwa J, Namala W, Kizito H, Kityo C, et al. Response to nonnucleoside reverse transcriptase inhibitor-based therapy in HIV-infected children with perinatal exposure to single-dose nevirapine. AIDS Res Hum Retroviruses. 2009;25(10):989-96. 138. Palumbo P, Lindsey JC, Hughes MD, Cotton MF, Bobat R, Meyers T, et al. Antiretroviral treatment for children with peripartum nevirapine exposure. N Engl J Med. 2010;363(16):1510-20. 139. Violari A, Lindsey JC, Hughes MD, Mujuru HA, Barlow-Mosha L, Kamthunzi P, et al. Nevirapine versus ritonavir-boosted lopinavir for HIV-infected children. N Engl J Med. 2012;366(25):2380-9. 140. NEVEREST - 3 trial. Treatment options for protease inhibitor-exposed children (NEVEREST-III). Washington, DC: ClinicalTrials.gov; [cited 2013]; Available from: http://www.clinicaltrials.gov/ct2/show/NCT01146873?term=NEVEREST&rank=1. 141. trial M. Evaluation of Simplified Antiretroviral Treatment Strategies in HIV Infected Children Treated by Antiretroviral (ARV) Before One Year of Age. Washington, DC: ClinicalTrials.gov; [cited 2013]; Available from: http://clinicaltrials.gov/show/NCT01127204.

185

142. Kuhn L, Coovadia A, Strehlau R, Martens L, Hu CC, Meyers T, et al. Switching children previously exposed to nevirapine to nevirapine-based treatment after initial suppression with a protease-inhibitor-based regimen: long-term follow-up of a randomised, open-label trial. Lancet Infect Dis. 2012;12(7):521-30. 143. Coovadia A, Abrams EJ, Stehlau R, Meyers T, Martens L, Sherman G, et al. Reuse of nevirapine in exposed HIV-infected children after protease inhibitor-based viral suppression: a randomized controlled trial. JAMA. 2010;304(10):1082-90. 144. Danel C, Moh R, Minga A, Anzian A, Ba-Gomis O, Kanga C, et al. CD4-guided structured antiretroviral treatment interruption strategy in HIV-infected adults in west Africa (Trivacan ANRS 1269 trial): a randomised trial. Lancet. 2006;367(9527):1981-9. 145. El-Sadr WM, Lundgren J, Neaton JD, Gordin F, Abrams D, Arduino RC, et al. CD4+ count-guided interruption of antiretroviral treatment. N Engl J Med. 2006;355(22):2283-96. 146. Bunupuradah T, Duong T, Compagnucci A, McMaster P, Bernardi S, Kanjanavanit S, et al. Outcomes after reinitiating antiretroviral therapy in children randomized to planned treatment interruptions. AIDS. 2013;27(4):579-89. 147. Cotton MF, Violari A, Otwombe K, Panchia R, Dobbels E, Rabie H, et al. Early time-limited antiretroviral therapy versus deferred therapy in South African infants infected with HIV: results from the children with HIV early antiretroviral (CHER) randomised trial. Lancet. 2013;382(9904):1555-63. Epub 2013/11/12. 148. COULIBALY M, MEDA N, YONABA C, BLANCHE S, VAN DE PERRE P, LEROY Valériane M. Opportunités manquées de prise en charge précoce de l’infection par le VIH du nourrisson à Ouagadougou, Burkina Faso. Société Française de Santé Publique; Bordeaux, France2013. 149. Youngleson MS, Nkurunziza P, Jennings K, Arendse J, Mate KS, Barker P. Improving a mother to child HIV transmission programme through health system redesign: quality improvement, protocol adjustment and resource addition. PloS One. 2010;5(11):e13891. 150. B. Gautney, S. Finocchario-Kessler, V. Okoth, J. Dougherty, E. Serem, K. Clark, et al. Improving turn-around time for HIV DNA PCR lab results for early infant diagnosis in Kenya. 7th IAS Conference on HIV Pathogenesis, Treatment and Prevention; 30 June - 3 July 2013; Kuala-Lampur, Malaysia2013. 151. Diallo O, Schlumberger M, Sanou C, Dicko H, Aplogan A, Drabo F. [Use of SMS to ask mothers to come to vaccination sessions in Bobo-Dioulasso]. Bull Soc Pathol Exot. 2012;105(4):291-5. Recours aux SMS pour convoquer les meres aux seances de vaccination a Bobo-Dioulasso. 152. Lund S, Hemed M, Nielsen BB, Said A, Said K, Makungu MH, et al. Mobile phones as a health communication tool to improve skilled attendance at delivery in Zanzibar: a cluster-randomised controlled trial. BJOG : an international journal of obstetrics and gynaecology. 2012;119(10):1256-64. 153. van Velthoven MH, Brusamento S, Majeed A, Car J. Scope and effectiveness of mobile phone messaging for HIV/AIDS care: a systematic review. Psychology, health & medicine. 2013;18(2):182-202. 154. Feucht UD, Kinzer M, Kruger M. Reasons for delay in initiation of antiretroviral therapy in a population of HIV-infected South African children. J Trop Pediatr. 2007;53(6):398-402. 155. Keiser O, Chi BH, Gsponer T, Boulle A, Orrell C, Phiri S, et al. Outcomes of antiretroviral treatment in programmes with and without routine viral load monitoring in Southern Africa. AIDS. 2011;25(14):1761-9. 156. Westley BP, DeLong AK, Tray CS, Sophearin D, Dufort EM, Nerrienet E, et al. Prediction of treatment failure using 2010 World Health Organization Guidelines is

186

associated with high misclassification rates and drug resistance among HIV-infected Cambodian children. Clin Infect Dis. 2012;55(3):432-40. 157. Kantor R, Diero L, Delong A, Kamle L, Muyonga S, Mambo F, et al. Misclassification of first-line antiretroviral treatment failure based on immunological monitoring of HIV infection in resource-limited settings. Clin Infect Dis. 2009;49(3):454-62. 158. Reynolds SJ, Nakigozi G, Newell K, Ndyanabo A, Galiwongo R, Boaz I, et al. Failure of immunologic criteria to appropriately identify antiretroviral treatment failure in Uganda. AIDS. 2009;23(6):697-700. 159. Rawizza HE, Chaplin B, Meloni ST, Eisen G, Rao T, Sankale JL, et al. Immunologic criteria are poor predictors of virologic outcome: implications for HIV treatment monitoring in resource-limited settings. Clin Infect Dis. 2011;53(12):1283-90. 160. Kekitiinwa A, Cook A, Nathoo K, Mugyenyi P, Nahirya-Ntege P, Bakeera-Kitaka S, et al. Routine versus clinically driven laboratory monitoring and first-line antiretroviral therapy strategies in African children with HIV (ARROW): a 5-year open-label randomised factorial trial. Lancet. 2013;381(9875):1391-403. 161. Bourne DE, Thompson M, Brody LL, Cotton M, Draper B, Laubscher R, et al. Emergence of a peak in early infant mortality due to HIV/AIDS in South Africa. AIDS (London, England). 2009;23(1):101-6. 162. Weinstein MC, Siegel JE, Gold MR, Kamlet MS, Russell LB. Recommendations of the Panel on Cost-effectiveness in Health and Medicine. JAMA. 1996;276(15):1253-8. 163. Orlando S, Marazzi MC, Mancinelli S, Liotta G, Ceffa S, Giglio P, et al. Cost-effectiveness of using HAART in prevention of mother-to-child transmission in the DREAM-Project Malawi. J Acquir Immune Defic Syndr. 2010;55(5):631-4. 164. Shah M, Johns B, Abimiku A, Walker DG. Cost-effectiveness of new WHO recommendations for prevention of mother-to-child transmission of HIV in a resource-limited setting. AIDS. 2011;25(8):1093-102. 165. Freedberg KA, Losina E, Weinstein MC, Paltiel AD, Cohen CJ, Seage GR, et al. The cost effectiveness of combination antiretroviral therapy for HIV disease. N Engl J Med. 2001;344(11):824-31. 166. Fone D, Hollinghurst S, Temple M, Round A, Lester N, Weightman A, et al. Systematic review of the use and value of computer simulation modelling in population health and health care delivery. J Public Health. 2003;25(4):325-35. 167. Brown D, Rothery P. Models in biology: mathematics, statistics and computing: John Wiley & Sons Ltd.; 1993. 168. Andersson H, Britton T. Stochastic epidemic models and their statistical analysis: Springer New York; 2000. 169. Ng J, Orav EJ. A generalized chain binomial model with application to HIV infection. Math Biosci. 1990;101(1):99-119. 170. Graham N, Zeger S, Park L, Saah A, Phair J, Detels R, et al. Effect of zidovudine and Pneumocystis carinii pneumonia prophylaxis on progression of HIV-1 infection to AIDS. Lancet. 1991;338(8762):265-9.

171. Gentleman R, Lawless J, Lindsey J, Yan P. Multi state Markov models for analysing incomplete disease history data with illustrations for hiv disease. Stat Med. 1994;13(8):805-21. 172. Walensky RP, Paltiel AD, Losina E, Mercincavage LM, Schackman BR, Sax PE, et al. The survival benefits of AIDS treatment in the United States. J Infect Dis. 2006;194(1):11-9. 173. Becker N. The uses of epidemic models. Biometrics. 1979:295-305. 174. Ryan M, Griffin S, Chitah B, Walker AS, Mulenga V, Kalolo D, et al. The cost-effectiveness of cotrimoxazole prophylaxis in HIV-infected children in Zambia. AIDS. 2008;22(6):749-57.

187

175. Schneider K, Puthanakit T, Kerr S, Law MG, Cooper Da, Donovan B, et al. Economic evaluation of monitoring virologic responses to antiretroviral therapy in HIV-infected children in resource-limited settings. AIDS. 2011;25(9):1143-51. 176. Ciaranello AL, Lockman S, Freedberg KA, Hughes M, Chu J, Currier J, et al. First-line antiretroviral therapy after single-dose nevirapine exposure in South Africa: a cost-effectiveness analysis of the OCTANE trial. AIDS 2011;25(4):479-92. 177. Rydzak CE, Cotich KL, Sax PE, Hsu HE, Wang B, Losina E, et al. Assessing the performance of a computer-based policy model of HIV and AIDS. PloS One. 2010;5(9). 178. Paltiel AD, Weinstein MC, Kimmel AD, Seage GR, 3rd, Losina E, Zhang H, et al. Expanded screening for HIV in the United States--an analysis of cost-effectiveness. N Engl J Med. 2005;352(6):586-95. 179. WHO-CHOICE. Cost-effectiveness thresholds. 2010; Available from: http://www.who.int/choice/costs/CER_thresholds/en/. 180. Ciaranello A, Morris BJ, Walensky RP, Weinstein MC, Ayaya S, Doherty K, et al. Validation and calibration of a computer simulation model of paediatric HIV infection. PLosOne. 2013;in press. 181. Gulliford M, Figueroa-Munoz J, Morgan M, Hughes D, Gibson B, Beech R, et al. What does 'access to health care' mean? Journal of health services research & policy. 2002;7(3):186-8. 182. Desmonde S, Coffie P, Aka E, Amani-Bosse C, Messou E, Dabis F, et al. Severe morbidity and mortality in untreated HIV-infected children in a paediatric care programme in Abidjan, Côte d'Ivoire, 2004-2009. BMC infectious diseases. 2011;11(1):182-. 183. Desmonde S, Coffie PA, Aka EA, Amani-Bosse C, Messou E, Dabis F, et al. Health care resource utilization in untreated HIV-infected children in a pediatric programme, Abidjan, Cote d'Ivoire, 2004-2009. J Acquir Immune Defic Syndr. 2013;62(1):e14-21. Epub 2012/12/25. 184. Desmonde S, Essanin J, Aka E, Messou E, Amorissani-Folquet M, Rondeau V, et al. Morbidity and healthcare resource utilisation in HIV-infected children following antiretroviral therapy (ART) initiation in Cote d'Ivoire, 2004-2009. J Acquir Immune Defic Syndr. 2013;in press. 185. sida. RdBCndlcl. Rapport national de suivi de la déclaration politique sur le VIH/sida. 2012. 186. Rapport d'activités sur la riposte au sida du Burkina Faso (Global AIDS Response Reporting) 2012. Available from: https://www.unaids.org/en/dataanalysis/knowyourresponse/countryprogressreports/2012countries/ce_BF_Narrative_Report%5B1%5D.pdf. 187. Ghana AIDS Comission. Ghana country AIDS Progress Report 2012. Available from: https://www.unaids.org/en/dataanalysis/knowyourresponse/countryprogressreports/2012countries/ce_GH_Narrative_Report%5B1%5D.pdf. 188. République du Mali - Haut Conseil de lutte contre le sida. Rapport national UNGASS 2012 2012. Available from: https://www.unaids.org/en/dataanalysis/knowyourresponse/countryprogressreports/2012countries/RAPPORT_UNGASS%202012%20_Mali.pdf. 189. République du Sénégal - Conseil national de lutte contre le sida. Rapport de la situation sur la riposte nationale à l'épidémie de VIH/sida au Sénégal 2010-2011 2012. Available from: https://www.unaids.org/en/dataanalysis/knowyourresponse/countryprogressreports/2012countries/ce_SN_Narrative_Report%5B1%5D.pdf. 190. IST; RtCndlclsel. Rapport de progrès sur la riposte au sida au Togo. 2012.

188

191. Central Intelligence Agency. The world factbook 2013. Available from: https://www.cia.gov/library/publications/the-world-factbook/geos/iv.html. 192. République de Côte d'Ivoire - Conseil national de lutte contre le sida. Suivi de la déclaration de politique sur le sida de juin 2011 2012. Available from: https://www.unaids.org/en/dataanalysis/knowyourresponse/countryprogressreports/2012countries/ce_CI_Narrative_Report%5B1%5D.pdf. 193. Fassinou-Ekouevi P, Fayama M, Koffi L, Lorougnon M, Oura O, Ndraman J, et al. Scaling up early infant diagnosis of HIV using dried blood spot polymerase chain reaction testing at prevention of mother-to-child transmission (PMTCT) sites: Challenges and lessons learned in Côte d’Ivoire. 6th IAS conference on HIV pathogenesis, treatment and prevention; 17-20 July 2011; Roma, Italy. 194. Toure S, Kouadio B, Seyler C, Traore M, Dakoury-Dogbo N, Duvignac J, et al. Rapid scaling-up of antiretroviral therapy in 10,000 adults in Cote d'Ivoire: 2-year outcomes and determinants. AIDS. 2008;22(7):873-82. 195. Harambat J, Fassinou P, Becquet R, Touré P, Rouet F, Dabis F, et al. 18-month occurrence of severe events among early diagnosed HIV-infected children before antiretroviral therapy in Abidjan, Côte d'Ivoire: a cohort study. BMC Public Health. 2008;8:169-. 196. Fine JP, Gray RJ. A Proportional Hazards Model for the Subdistribution of a Competing Risk. J Am Stat Assoc 1999;94(446):496-509. 197. Malateste K, Dicko F, Desmonde S, Koueta F, Eboua FT, Renner L, et al. Estimating unbiased mortality risks in paediatric antiretroviral cohorts: do we need to use a Kaplan-Meier method or a competing risks method? The IeDEA West Africa Collaboration (pWADA). 16th International Workshop on HIV Observational Databases; Athens, Grece2012. 198. Zeger SL, Liang KY, Albert PS. Models for longitudinal data: a generalized estimating equation approach. Biometrics. 1988;44(4):1049-60. 199. Brinkhof MWG, Pujades-Rodriguez M, Egger M. Mortality of patients lost to follow-up in antiretroviral treatment programmes in resource-limited settings: systematic review and meta-analysis. PloS One. 2009;4(6). 200. Lipshultz SE, Mas CM, Henkel JM, Franco VI, Fisher SD, Miller TL. HAART to heart: highly active antiretroviral therapy and the risk of cardiovascular disease in HIV-infected or exposed children and adults. Expert review of anti-infective therapy. 2012;10(6):661-74. 201. Lipshultz SE, Miller TL, Wilkinson JD, Scott GB, Somarriba G, Cochran TR, et al. Cardiac effects in perinatally HIV-infected and HIV-exposed but uninfected children and adolescents: a view from the United States of America. Journal of the International AIDS Society. 2013;16:18597. 202. Mofenson LM, Cotton MF. The challenges of success: adolescents with perinatal HIV infection. Journal of the International AIDS Society. 2013;16:18650. 203. Barlow-Mosha L, Eckard AR, McComsey GA, Musoke PM. Metabolic complications and treatment of perinatally HIV-infected children and adolescents. Journal of the International AIDS Society. 2013;16:18600. 204. Coutsoudis A, Coovadia HM, Gurpreet K. Time for new recommendations on cotrimoxazole prophylaxis for HIV-exposed infants in developing countries? Bull World Health Organ. 2010;180(6):1809-18. 205. Mulenga V, Ford D, Walker AS, Mwenya D, Mwansa J, Sinyinza F, et al. Effect of cotrimoxazole on causes of death, hospital admissions and antibiotic use in HIV-infected children. AIDS 2007;21(1):77-84. 206. Bwakura-Dangarembizi M, Kendall L, Bakeera-Kitaka S, Nahirya Ntege P, Keishanyu R, Kekitiinwa A, et al. Randomized Comparison of Stopping vs Continuing Cotrimoxazole

189

Prophylaxis among 758 HIV+ Children on Long-term ART: The Anti-Retroviral Research for Watoto Trial. CROI; Atlanta, USA2013. p. Abstract # 86. 207. Rondeau V, Gonzalez JR. frailtypack: a computer program for the analysis of correlated failure time data using penalized likelihood estimation. Computer methods and programs in biomedicine. 2005;80(2):154-64. 208. Rondeau V, Mathoulin-Pelissier S, Jacqmin-Gadda H, Brouste V, Soubeyran P. Joint frailty models for recurring events and death using maximum penalized likelihood estimation: application on cancer events. Biostatistics. 2007;8(4):708-21. 209. Sivapalasingam S, Mendillo M, Ahmed A, Mwamzuka M, Said S, Marshed F, et al. The importance of caregivers in the outcome of pediatric HIV management, Mombasa, Kenya. AIDS Care. 2013. 210. Giraudon I, Leroy V, Msellati P, Elenga N, Ramon R, Welffens-Ekra C, et al. Evaluation du coût de la prise en charge des patients infectés par le VIH à Abidjan (Côte d'Ivoire) en 1996-1997. Cahiers Santé. 1999. 211. Rosen S, Ketlhapile M, Sanne I, DeSilva MB. Cost to patients of obtaining treatment for HIV/AIDS in South Africa. S Afr Med J. 2007;97(7):524-9. Epub 2007/09/07. 212. Ridde V, Haddad S. Abolishing user fees in Africa. PLoS Med. 2009;6(1):e1000008. 213. Collins I, Cairns J, Le Coeur S, Pagdi K, Ngampiyaskul C, Layangool P, et al. Five-year trends in antiretroviral usage and drug costs in HIV-infected children in Thailand. J Acquir Immune Defic Syndr. 2013;64(1):95-102. 214. Lewis J, Walker AS, Castro H, De Rossi A, Gibb DM, Giaquinto C, et al. Age and CD4 Count at Initiation of Antiretroviral Therapy in HIV-Infected Children: Effects on Long-term T-Cell Reconstitution. J Infect Dis. 2011. Epub 2011/12/30. 215. Patel K, Hernan MA, Williams PL, Seeger JD, McIntosh K, Dyke RB, et al. Long-term effects of highly active antiretroviral therapy on CD4+ cell evolution among children and adolescents infected with HIV: 5 years and counting. Clin Infect Dis. 2008;46(11):1751-60. Epub 2008/04/23. 216. De Beaudrap P, Rouet F, Fassinou P, Kouakoussui A, Mercier S, Ecochard R, et al. CD4 cell response before and after HAART initiation according to viral load and growth indicators in HIV-1-infected children in Abidjan, Cote d'Ivoire. J Acquir Immune Defic Syndr. 2008;49(1):70-6. Epub 2008/08/01. 217. Picat MQ, Lewis J, Musiime V, Prendergast A, Nathoo K, Kekitiinwa A, et al. Predicting patterns of long-term CD4 reconstitution in HIV-infected children starting antiretroviral therapy in sub-Saharan Africa: a cohort-based modelling study. PLoS Med. 2013;10(10):e1001542. Epub 2013/11/10. 218. Cole TJ. The LMS method for constructing normalized growth standards. Eur J Clin Nutr. 1990;44(1):45-60. 219. Yotebieng M, Van Rie A, Moultrie H, Meyers T. Six-month gains in weight, height, and CD4 predict subsequent antiretroviral treatment responses in HIV-infected South African children. AIDS. 2011;24(1):139-46. 220. Jones JL, Hanson DL, Dworkin MS, DeCock KM. HIV-associated tuberculosis in the era of highly active antiretroviral therapy. The Adult/Adolescent Spectrum of HIV Disease Group. Int J Tuberc Lung Dis. 2000;4(11):1026-31. 221. Miller V, Sabin CA, Phillips AN, Rottmann C, Rabenau H, Weidmann E, et al. The impact of protease inhibitor-containing highly active antiretroviral therapy on progression of HIV disease and its relationship to CD4 and viral load. AIDS. 2000;14(14):2129-36. 222. Cole SR, Hernan MA, Robins JM, Anastos K, Chmiel J, Detels R, et al. Effect of highly active antiretroviral therapy on time to acquired immunodeficiency syndrome or death using marginal structural models. Am J Epidemiol. 2003;158(7):687-94.

190

223. Losina E, Yazdanpanah Y, Deuffic-Burban S, Wang B, Wolf LL, Messou E, et al. The independent effect of highly active antiretroviral therapy on severe opportunistic disease incidence and mortality in HIV-infected adults in Cote d'Ivoire. Antivir Ther. 2007;12(4):543-51. 224. de Martino M, Tovo PA, Balducci M, Galli L, Gabiano C, Rezza G, et al. Reduction in mortality with availability of antiretroviral therapy for children with perinatal HIV-1 infection. Italian Register for HIV Infection in Children and the Italian National AIDS Registry. JAMA. 2000;284(2):190-7. 225. Gortmaker SL, Hughes M, Cervia J, Brady M, Johnson GM, Seage GR, 3rd, et al. Effect of combination therapy including protease inhibitors on mortality among children and adolescents infected with HIV-1. N Engl J Med. 2001;345(21):1522-8. 226. Patel K, Hernan MA, Williams PL, Seeger JD, McIntosh K, Van Dyke RB, et al. Long-term effectiveness of highly active antiretroviral therapy on the survival of children and adolescents with HIV infection: a 10-year follow-up study. Clin Infect Dis. 2008;46(4):507-15. Epub 2008/01/18. 227. Kiragga AN, Castelnuovo B, Schaefer P, Muwonge T, Easterbrook PJ. Quality of data collection in a large HIV observational clinic database in sub-Saharan Africa: implications for clinical research and audit of care. Journal of the International AIDS Society. 2011;14:3. Epub 2011/01/22. 228. Hanagal D. Modeling Survival Data Using Frailty Models: Chapman & Hall / CRC; 2011. 229. Forster M, Bailey C, Brinkhof MW, Graber C, Boulle A, Spohr M, et al. Electronic medical record systems, data quality and loss to follow-up: survey of antiretroviral therapy programmes in resource-limited settings. Bull World Health Organ. 2008;86(12):939-47. Epub 2009/01/15. 230. Rotich JK, Hannan TJ, Smith FE, Bii J, Odero WW, Vu N, et al. Installing and implementing a computer-based patient record system in sub-Saharan Africa: the Mosoriot Medical Record System. Journal of the American Medical Informatics Association. 2003;10(4):295-303. 231. Castelnuovo B, Kiragga A, Afayo V, Ncube M, Orama R, Magero S, et al. Implementation of provider-based electronic medical records and improvement of the quality of data in a large HIV program in Sub-Saharan Africa. PloS One. 2012;7(12):e51631. Epub 2013/01/04. 232. Cain KC, Harlow SD, Little RJ, Nan B, Yosef M, Taffe JR, et al. Bias due to left truncation and left censoring in longitudinal studies of developmental and disease processes. Am J Epidemiol. 2011;173(9):1078-84. Epub 2011/03/23. 233. Schisterman EF, Cole SR, Ye A, Platt RW. Accuracy loss due to selection bias in cohort studies with left truncation. Paediatr Perinat Epidemiol. 2013;27(5):491-502. Epub 2013/08/13. 234. Sibanda E, Weller I, Hakim J, Cowan F. The magnitude of loss to follow-up of HIV-exposed infants along the prevention of mother-to-child HIV transmission continuum of care: a systematic review and meta-analysis. AIDS. 2013. Epub 2013/09/24. 235. Geskus RB. Cause-Specific Cumulative Incidence Estimation and the Fine and Gray Model Under Both Left Truncation and Right Censoring. Biometrics. 2010. 236. Piccorelli AV, Schluchter MD. Jointly modeling the relationship between longitudinal and survival data subject to left truncation with applications to cystic fibrosis. Stat Med. 2012;31(29):3931-45. Epub 2012/07/13. 237. Braitstein P, Katshcke A, Shen C, Sang E, Nyandiko W, Ochieng VO, et al. Retention of HIV-infected and HIV-exposed children in a comprehensive HIV clinical care programme in Western Kenya. Trop Med Int Health. 2010;15(7):833-41. Epub 2010/05/22.

191

238. Egger M, Spycher BD, Sidle J, Weigel R, Geng EH, Fox MP, et al. Correcting mortality for loss to follow-up: a nomogram applied to antiretroviral treatment programmes in sub-Saharan Africa. PLoS Med. 2011;8(1):e1000390. Epub 2011/01/27. 239. Verguet S, Lim SS, Murray CJ, Gakidou E, Salomon JA. Incorporating loss to follow-up in estimates of survival among HIV-infected individuals in sub-Saharan Africa enrolled in antiretroviral therapy programs. J Infect Dis. 2013;207(1):72-9. Epub 2012/10/27. 240. Prendergast AJ, Penazzato M, Cotton M, Musoke P, Mulenga V, Abrams EJ, et al. Treatment of young children with HIV infection: using evidence to inform policymakers. PLoS Med. 2012;9(7):e1001273. Epub 2012/08/23. 241. Young S, Wheeler AC, McCoy SI, Weiser SD. A Review of the Role of Food Insecurity in Adherence to Care and Treatment Among Adult and Pediatric Populations Living with HIV and AIDS. AIDS Behav. 2013. 242. Young SL, Mbuya MN, Chantry CJ, Geubbels EP, Israel-Ballard K, Cohan D, et al. Current Knowledge and Future Research on Infant Feeding in the Context of HIV: Basic, Clinical, Behavioral, and Programmatic Perspectives. Advances in nutrition (Bethesda, Md). 2011;2(3):225-43. Epub 2012/02/15. 243. Braitstein P, Nyandiko W, Vreeman R, Wools-Kaloustian K, Sang E, Musick B, et al. The clinical burden of tuberculosis among human immunodeficiency virus-infected children in Western Kenya and the impact of combination antiretroviral treatment. The Pediatric Infectious Disease Journal. 2009;28(7):626-32. 244. Martinson NA, Moultrie H, van Niekerk R, Barry G, Coovadia A, Cotton M, et al. HAART and risk of tuberculosis in HIV-infected South African children: a multi-site retrospective cohort. The International Journal of Tuberculosis and Lung Disease: The Official Journal of the International Union Against Tuberculosis and Lung Disease. 2009;13(7):862-7. 245. Yotebieng M, Van Rie A, Moultrie H, Cole SR, Adimora A, Behets F, et al. Effect on mortality and virological response of delaying antiretroviral therapy initiation in children receiving tuberculosis treatment. AIDS. 2010;24(9):1341-9. Epub 2010/06/19. 246. Manosuthi W, Wongsawat J. Treatment Challenges in Co-infected HIV and TB Children. Indian pediatrics. 2011;48(12):937-8. Epub 2012/01/19. 247. Onyenekwe CC, Ukibe N, Meludu SC, Ilika A, Aboh N, Ofiaeli N, et al. Prevalence of malaria as co-infection in HIV-infected individuals in a malaria endemic area of southeastern Nigeria. Journal of Vector Borne Diseases. 2007;44(4):250-4. 248. Becher H, Kynast-Wolf G, Sié A, Ndugwa R, Ramroth H, Kouyaté B, et al. Patterns of malaria: cause-specific and all-cause mortality in a malaria-endemic area of west Africa. The American Journal of Tropical Medicine and Hygiene. 2008;78(1):106-13. 249. Achan J, Kakuru A, Ikilezi G, Ruel T, Clark TD, Nsanzabana C, et al. Antiretroviral agents and prevention of malaria in HIV-infected Ugandan children. N Engl J Med. 2012;367(22):2110-8. Epub 2012/11/30. 250. Sansom SL, Anderson JE, Farnham PG, Dominguez K, Soorapanth S, Clark J, et al. Updated estimates of healthcare utilization and costs among perinatally HIV-infected children. J Acquir Immune Defic Syndr. 2006;41(4):521-6. Epub 2006/05/03. 251. Zulliger R, Black S, Holtgrave DR, Ciaranello AL, Bekker LG, Myer L. Cost-Effectiveness of a Package of Interventions for Expedited Antiretroviral Therapy Initiation During Pregnancy in Cape Town, South Africa. AIDS Behav. 2013. Epub 2013/10/15. 252. Tejiokem MC, Faye A, Penda IC, Guemkam G, Ateba Ndongo F, Chewa G, et al. Feasibility of early infant diagnosis of HIV in resource-limited settings: the ANRS 12140-PEDIACAM study in Cameroon. PloS One. 2011;6(7):e21840.

192

253. Braun M, Kabue MM, McCollum ED, Ahmed S, Kim M, Aertker L, et al. Inadequate coordination of maternal and infant HIV services detrimentally affects early infant diagnosis outcomes in Lilongwe, Malawi. J Acquir Immune Defic Syndr. 2011;56(5):e122-8. 254. Ekouevi DK, Stringer E, Coetzee D, Tih P, Creek T, Stinson K, et al. Health facility characteristics and their relationship to coverage of PMTCT of HIV services across four African countries: the PEARL study. PloS One. 2012;7(1):e29823. 255. Kankasa C, Carter RJ, Briggs N, Bulterys M, Chama E, Cooper ER, et al. Routine offering of HIV testing to hospitalized pediatric patients at university teaching hospital, Lusaka, Zambia: acceptability and feasibility. J Acquir Immune Defic Syndr. 2009;51(2):202-8. 256. Wanyenze RK, Nawavvu C, Ouma J, Namale A, Colebunders R, Kamya MR. Provider-initiated HIV testing for paediatric inpatients and their caretakers is feasible and acceptable. Trop Med Int Health. 2010;15(1):113-9. 257. Schito M, Peter TF, Cavanaugh S, Piatek AS, Young GJ, Alexander H, et al. Opportunities and challenges for cost-efficient implementation of new point-of-care diagnostics for HIV and tuberculosis. J Infect Dis. 2012;205 Suppl 2:S169-80. 258. Lilian RR, Kalk E, Technau KG, Sherman GG. Birth Diagnosis of HIV Infection on Infants to Reduce Infant Mortality and Monitor for Elimination of Mother-to-Child Transmission. Pediatr Infect Dis J. 2013. Epub 2013/04/12. 259. Braitstein P, Songok J, Vreeman RC, Wools-Kaloustian KK, Koskei P, Walusuna L, et al. "Wamepotea" (they have become lost): outcomes of HIV-positive and HIV-exposed children lost to follow-up from a large HIV treatment program in western Kenya. J Acquir Immune Defic Syndr. 2011;57(3):e40-6. 260. Sohn AH, for the IeDEA Pediatric working group. A survey of paediatric HIV programmatic and clinical management practices in Asia and sub-Saharan Africa--the International epidemiologic Databases to Evaluate AIDS (IeDEA). Journal of the International AIDS Society. 2013;16(1):17998. 261. Rosen S, Ketlhapile M. Cost of using a patient tracer to reduce loss to follow-up and ascertain patient status in a large antiretroviral therapy program in Johannesburg, South Africa. Trop Med Int Health. 2010;15 Suppl 1:98-104. 262. Losina E, Toure H, Uhler LM, Anglaret X, Paltiel AD, Balestre E, et al. Cost-effectiveness of preventing loss to follow-up in HIV treatment programs: a Cote d'Ivoire appraisal. PLoS Med. 2009;6(10):e1000173. Epub 2009/10/28. 263. Musoke PM, Mudiope P, Barlow-Mosha LN, Ajuna P, Bagenda D, Mubiru MM, et al. Growth, immune and viral responses in HIV infected African children receiving highly active antiretroviral therapy: a prospective cohort study. BMC Pediatrics. 2010;10:56-. 264. McGrath CJ, Chung MH, Richardson BA, Benki-Nugent S, Warui D, John-Stewart GC. Younger age at HAART initiation is associated with more rapid growth reconstitution. AIDS (London, England). 2011;25(3):345-55. 265. Adjorlolo-Johnson G, Wahl Uheling A, Ramachandran S, Strasser S, Kouakou J, Tindyebwa D, et al. Scaling up pediatric HIV care and treatment in Africa: clinical site characteristics associated with favorable service utilization. J Acquir Immune Defic Syndr. 2013;62(1):e7-e13. 266. Tchidjou HK, Maria Martino A, Goli LP, Diop Ly M, Zekeng L, Samba M, et al. Paediatric HIV infection in Western Africa: the long way to the standard of care. J Trop Pediatr. 2012;58(6):451-6. 267. Yaya Bocoum F, Kouanda S, Kouyate B, Hounton S, Adam T. Exploring the effects of task shifting for HIV through a systems thinking lens: the case of Burkina Faso. BMC Public Health. 2013;13(1):997.

193

268. Decroo T, Rasschaert F, Telfer B, Remartinez D, Laga M, Ford N. Community-based antiretroviral therapy programs can overcome barriers to retention of patients and decongest health services in sub-Saharan Africa: a systematic review. International health. 2013;5(3):169-79. 269. Luyirika E, Towle MS, Achan J, Muhangi J, Senyimba C, Lule F, et al. Scaling up paediatric HIV care with an integrated, family-centred approach: an observational case study from Uganda. PloS One. 2013;8(8):e69548.

194

Appendices

1. Personal contribution to each presented study .................................................................. 1952. List of publications & communications ............................................................................ 196

195

1. Personal contribution to each presented study

i. What is the natural evolution of HIV before antiretroviral therapy in Abidjan, Cote

d’Ivoire?

Both studies that we conducted to answer this question were based on a database that I built in

2010, during my master’s internship. I elaborated the data collection instrument, spent 3

months in Abidjan collecting the data and entering into a database I had previously prepared.

Once back in Bordeaux and after discussions with statisticians regarding the methodology and

the use of competing risks, I ran the analyses and wrote the paper.

ii. Reasons for hospitalisation in HIV-infected children in West Africa

The protocol of this study had previously been elaborated by the paediatricians of the

involved clinical sites; Charlotte Lewden had cleaned the database and ran preliminary

analyses. For this work, I ran the principal analyses and wrote the paper.

iii. What is the evolution of HIV in children after the initiation of antiretroviral

therapy?

Severe morbidity and healthcare resource utilisation in HIV-infected

children in Abidjan, Côte d’Ivoire

Along with paediatricians working at the CePReF, I elaborated the protocol of this study and

built the data collection instrument, adapting to the available data and order of the medical

files the first one I had constructed for the pre-ART studies. Once finalised and approved by

the paediatricians on site, I spent two weeks in Abidjan to set up this study and train a medical

resident for the data collection. I built a database with a simple interface for the resident to

enter the data and run periodical verifications during the data collection period.

Once the data collected completed, I cleaned the database, run the analyses with the support

of statisticians. I interpreted the results and wrote the paper.

Immune response to antiretroviral therapy according to age at treatment

initiation in HIV-infected children in West Africa

This analysis was conducted using the fourth merger of the IeDEA database. Once in

possession of the available data and with the support of statisticians, I ran analyses,

interpreted data and wrote the paper.

196

2. List of publications & communications

Publications & papers in review within this PhD

S. Desmonde, F. Dicko, F. Koueta, T Eboua, E. Balestre, C. Amani-Bosse, E.A Aka, K. Lawson-Evi, M. Amorissani-Folquet, K. Kouakou, S. Koumakpai, L. Renner, H. Signaté Sy

and V. Leroy for the IeDEA West Africa Paediatric Collaboration. Effect of age at antiretroviral therapy initiation on the 24-month immune response in HIV-infected children in the IeDEA paediatric West African Cohort (pWADA). AIDS. (In revision) F. Dicko, S. Desmonde S.Koumakpai, H. Dior-Mbodj, F. Kouéta, N. Baeta , N. Koné, J. Akakpo, H. Signate Sy, D. Ye, L. Renner, C. Lewden and V. Leroy for the Paediatric IeDEA West Africa Working Group. Reasons for hospitalisation in HIV-infected children in West Africa. J Int AIDS Soc. (In revision)

A.Ciaranello, B.L. Morris, R.P. Walensky, M.C. Weinstein, S. Ayaya, K. Doherty, V. Leroy, T. Hou, S. Desmonde, Z. Lu, F. Noubary, K. Patel, L. Ramirez-Avila, E. Losina, GR. Seage III, K.A. Freedberg. Validation and calibration of a computer simulation model of paediatric HIV infection. PLosONE, (in press)

S. Desmonde, JB Essanin, M. Amorissani-Folquet, EA Aka, E Messou, V. Rondeau, A. Ciaranello and V. Leroy. Severe morbidity and healthcare resource utilisation in HIV-infected children on antiretroviral therapy in a paediatric programme, Abidjan, Cote d'Ivoire, 2004-2009. J Acquir Immune Defic Syndr. (in press)

S. Desmonde, P. Coffie, EA. Aka, C. Amani-Bosse, E. Messou, F. Dabis, A. Ciaranello and V. Leroy. Healthcare resource utilisation in untreated HIV-infected children in a paediatric programme, Abidjan, Cote d'Ivoire, 2004-2009. J Acquir Immune Defic Syndr. 2013 Jan 1;62(1):e14-21. S. Desmonde, P. Coffie, EA. Aka, C. Amani-Bosse, E. Messou, F. Dabis, A. Alioum, A. Ciaranello and V. Leroy. Severe morbidity and mortality in untreated HIV-infected children in a paediatric care programme in Abidjan, Côte d'Ivoire, 2004-2009. BMC infectious diseases. 2011;11(1):182

Additional publications & papers in review, outside of the scope of this

PhD

J. Jesson, S. Koumakpaï, N.R. Diagne, M. Amorissani-Folquet, F.Kouéta, E.A. Aka, K. Lawson-Evi, F. Dicko, K. Kouakou, T. Pety, L .Renner, T. Eboua, P. Coffié, S. Desmonde and V. Leroy for the paediatric WADA IeDEA Collaboration. Effect of age at ART initiation on growth reconstitution within the first 24 months among HIV-infected children in the IeDEA West African paediatric cohort. JAIDS. Submitted

A.Mounkaïla Harouna, M. Amorissani-Folquet, F.T Eboua, S. Desmonde, E.A. Aka, K. Kouadio, B. Kouacou, K. Malateste, C. Bosse-Amani, P.A. Coffie and V. Leroy for the IeDEA paediatric West African Study Group. Incidence, severity and determinants of malaria in HIV-infected children in Abidjan, in 2012. JID. Submitted

197

Yapo V, Toni TD, Desmonde S, Amani-Bosse C, Oga M, Lenaud S, et al. Evaluation of dried blood spot diagnosis using HIV1-DNA and HIV1-RNA Biocentric assays in infants in Abidjan, Cote d'Ivoire. The Pedi-Test DBS ANRS 12183 Study. J Virol Methods

2013,193:439-445.

Oral communications S. Desmonde, JB Essanin, M. Amorissani-Folquet, EA Aka, E Messou, V. Rondeau, A. Ciaranello and V. Leroy. Severe morbidity in HIV-infected children on antiretroviral therapy in a paediatric programme, Abidjan, Cote d'Ivoire, 2004-2009. 4e Rencontres Scientifiques du Réseau Doctoral de l’EHESP, Paris, 2013 S. Desmonde, F. Dicko, F. Koueta, K. Malateste, K Kouakou, L. Renner, EA. Aka, H. Signaté Sy and V.Leroy for the IeDEA West Africa Paediatric Collaboration 24-month immune recovery according to age at initiation of antiretroviral therapy in the IeDEA paediatric West African Cohort (pWADA). 3e Rencontres Scientifiques du Réseau Doctoral de l’EHESP, Paris, 2012

S. Desmonde, P. Coffie, EA. Aka, C. Amani-Bosse, E. Messou, F. Dabis, A. Ciaranello and V. Leroy. Severe morbidity and healthcare resource utilisation in untreated HIV-infected children in Abidjan, Cote d’Ivoire, 2004-2009. Journées Scientifiques du Programme PACCI – site ANRS d’Abidjan 2012, Abidjan, Côte d’Ivoire

S. Desmonde, P. Coffie, EA. Aka, C. Amani-Bosse, E. Messou, F. Dabis, A. Alioum, A.

Ciaranello and V. Leroy. Severe morbidity and mortality in untreated HIV-infected children

in a paediatric care programme in Abidjan, Cote d’Ivoire, 2004-2009. 2e Rencontres

Scientifiques du Réseau Doctoral de l’EHESP, Paris, 2011

Poster communications

S. Desmonde, F. Dicko, F. Koueta, K. Malateste, K Kouakou, L. Renner, EA. Aka, H.

Signaté Sy and V.Leroy for the IeDEA West Africa Paediatric Collaboration. 24-month

immune recovery according to age at initiation of antiretroviral therapy in the IeDEA

paediatric West African Cohort (pWADA). XXI Conference on AIDS. Washington DC, USA,

2012


Signaté Sy and V.Leroy for the IeDEA West Africa Paediatric Collaboration. 24-month

immune recovery according to age at initiation of antiretroviral therapy in the IeDEA

paediatric West African Cohort (pWADA). 4th

HIV Paediatrics Workshop, Washington DC,

USA, 2012


Signaté Sy, R. Thiébaut and V.Leroy for the IeDEA West Africa Paediatric Collaboration.

Modelling 24-month immune recovery according to age and CD4 percent at initiation of

antiretroviral therapy in the IeDEA paediatric West African Cohort (pWADA). 16th

International Workshop on Observational HIV Databases. Athens, Greece, 2012

198

S. Desmonde, P. Coffie, EA. Aka, C. Amani-Bosse, E. Messou, F. Dabis, A. Ciaranello and

V. Leroy. Severe morbidity and healthcare resource utilisation in untreated HIV-infected

children in Abidjan, Cote d’Ivoire, 2004-2009. 6th IAS Conference on HIV pathogenesis,

treatment and prevention, Rome, Italy, 2011

S. Desmonde, P. Coffie, EA. Aka, C. Amani-Bosse, E. Messou, F. Dabis, A. Alioum, A.

Ciaranello and V. Leroy. Severe morbidity and mortality in untreated HIV-infected children

in a paediatric care programme in Abidjan, Cote d’Ivoire, 2004-2009. 18th Conference on

Retroviruses and Opportunistic Infections, Boston, USA, 2011.(Young Investigator Award)

THÈSE DOCTORAT DE L'UNIVERSITÉ BORDEAUX 2

Documents

Transcript of THÈSE DOCTORAT DE L'UNIVERSITÉ BORDEAUX 2