M A J O R A R T I C L E H I V A I D S

Contribution of Genetic Background TraditionalRisk Factors and HIV-Related Factors toCoronary Artery Disease Events in HIV-PositivePersons

Margalida Rotger1a Tracy R Glass2a Thomas Junier34a Jens Lundgren56 James D Neaton57 Estella S Poloni11

Angeacutelique B van rsquot Wout8b Rubin Lubomirov1c Sara Colombo1 Raquel Martinez1 Andri Rauch913 HuldrychF Guumlnthard912 Jacqueline Neuhaus57 Deborah Wentworth57 Danielle van Manen8 Luuk A Gras8

Hanneke Schuitemaker8 Laura Albini14 Carlo Torti1415 Lisa P Jacobson16 Xiuhong Li16 Lawrence A Kingsley16

Federica Carli17 Giovanni Guaraldi17 Emily S Ford18 Irini Sereti18 Colleen Hadigan18 Esteban Martinez19

Mireia Arnedo19 Lander Egantildea-Gorrontildeo19 Jose M Gatell19 Matthew Law20 Courtney Bendall20 Kathy Petoumenos20

Juumlrgen Rockstroh21 Jan-Christian Wasmuth21 Kabeya Kabamba22 Marc Delforge22 Stephane De Wit22 Florian Berger23

Stefan Mauss23 Mariana de Paz Sierra24 Marcelo Losso24 Waldo H Belloso24 Maria Leyes25 Antoni Campins25

Annalisa Mondi26 Andrea De Luca2628 Ignacio Bernardino27 Moacutenica Barriuso-Iglesias27 Ana Torrecilla-Rodriguez27

Juan Gonzalez-Garcia27 Joseacute R Arribas27 Iuri Fanti28 Silvia Gel29 Jordi Puig29 Eugenia Negredo29 Mar Gutierrez30

Pere Domingo30 Julia Fischer31 Gerd Faumltkenheuer31 Carlos Alonso-Villaverde32 Alan Macken33 James Woo34

Tara McGinty35 Patrick Mallon33 Alexandra Mangili36 Sally Skinner36 Christine A Wanke36 Peter Reiss8

Rainer Weber912 Heiner C Bucher29 Jacques Fellay149 Amalio Telenti19 and Philip E Tarr910 for the MAGNIFICENTConsortiumd INSIGHTe and the Swiss HIV Cohort Study1Institute of Microbiology University Hospital Center University of Lausanne 2Basel Institute for Clinical Epidemiology and Biostatistics UniversityHospital Basel 3Swiss Institute of Bioinformatics Lausanne and 4Eacutecole Polytechnique Feacutedeacuterale de Lausanne Switzerland 5International Network forStrategic Initiatives in Global HIV Trials 6University of Copenhagen Denmark 7University of Minnesota Minneapolis 8AIDS Therapy Evaluation in theNetherlands Amsterdam 9Swiss HIV Cohort Study 10Kantonsspital Baselland Bruderholz University of Basel 11Laboratory of Anthropology Departmentof Genetics and Evolution University of Geneva 12Division of Infectious Diseases and Hospital Epidemiology University Hospital Zurich University ofZurich and 13Division of Infectious Diseases University Hospital Bern Switzerland 14Department of Infectious Diseases University of Brescia15Infectious Diseases Unit University ldquoMagna Graeciardquo Catanzaro Italy 16Multicenter AIDS Cohort Study 17University of Modena and Reggio EmiliaItaly 18Intramural National Institute of Allergy and Infectious Diseases Cohort Bethesda Maryland 19Hospital Clinic-IDIBAPS Barcelona Spain20Australian HIV Observational Database The Kirby Institute University of New South Wales Darlinghurst Australia 21University Hospital Bonn Germany22Saint-Pierre University Hospital Brussels Belgium 23Center HIV Hepatogastroenterology Duumlsseldorf Germany 24Cohorte LATINA (Argentiniancomponent) Buenos Aires Argentina 25University Hospital Son Espases Palma de Mallorca Spain 26Universitagrave Cattolica del Sacro Cuore Cohort RomeItaly 27University Hospital La Paz and IdiPAZ Biobanco Madrid Spain 28Italian Cohort Naive for Antiretrovirals Foundation Italy 29University HospitalGermans Trias i Pujol Badalona and 30University Hospital de la Santa Creu i Sant Pau Cohort Autonomous University Barcelona Spain 31UniversityHospital Cologne Germany 32Hospital Santa Tecla i Sant Pau Tarragona Spain 33School of Medicine and Medical Science University College Dublin34St Jamesrsquos Hospital and 35Mater Misericordiae University Hospital Dublin Ireland and 36Tufts University School of Medicine Boston Massachusetts

Background Persons infected with human immunodeficiency virus (HIV) have increased rates of coronaryartery disease (CAD) The relative contribution of genetic background HIV-related factors antiretroviral medica-tions and traditional risk factors to CAD has not been fully evaluated in the setting of HIV infection

Methods In the general population 23 common single-nucleotide polymorphisms (SNPs) were shown tobe associated with CAD through genome-wide association analysis Using the Metabochip we genotyped 1875

Received 21 December 2012 accepted 14 March 2013aM R T R G and T J contributed equally to this workbPresent affiliation Crucell Holland BV Leiden The NetherlandscPresent affiliation Clinical Pharmacology Center Pharmacology and Therapeu-

tics Department La Paz University Hospital School of Medicine UniversidadAutoacutenoma de Madrid IdiPAZ Madrid Spain

dThe Myocardial Infarction Assessment of Antiretroviral and Genetic Factors inHuman Immunodeficiency Virus Infection Consortium

eInternational Network for Strategic Initiatives in Global HIV Trials

Clinical Infectious Diseasescopy The Author 2013 Published by Oxford University Press on behalf of the InfectiousDiseases Society of America All rights reserved For Permissions please e-mailjournalspermissionsoupcomDOI 101093cidcit196

Correspondence Philip E Tarr MD Infectious Diseases Service KantonsspitalBaselland University of Basel 4101 Bruderholz Switzerland (philiptarrunibasch)

HIVAIDS bull CID bull 1

Clinical Infectious Diseases Advance Access published April 22 2013 at U

niversiteit van Am

sterdam on M

ay 7 2013httpcidoxfordjournalsorg

Dow

nloaded from

HIV-positive white individuals enrolled in 24 HIV observational studies including 571 participants with a first CAD event duringthe 9-year study period and 1304 controls matched on sex and cohort

Results A genetic risk score built from 23 CAD-associated SNPs contributed significantly to CAD (P = 29times10minus4) In the finalmultivariable model participants with an unfavorable genetic background (top genetic score quartile) had a CAD odds ratio (OR)of 147 (95 confidence interval [CI] 105ndash204) This effect was similar to hypertension (OR = 136 95 CI 106ndash173) hypercho-lesterolemia (OR = 151 95 CI 116ndash196) diabetes (OR = 166 95 CI 110ndash249) ge1 year lopinavir exposure (OR = 136 95CI 106ndash173) and current abacavir treatment (OR = 156 95 CI 117ndash207) The effect of the genetic risk score was additive tothe effect of nongenetic CAD risk factors and did not change after adjustment for family history of CAD

Conclusions In the setting of HIV infection the effect of an unfavorable genetic background was similar to traditional CADrisk factors and certain adverse antiretroviral exposures Genetic testing may provide prognostic information complementary tofamily history of CAD

Keywords HIV infection coronary artery disease genetics traditional risk factors antiretroviral therapy

A major long-term concern in HIV-positive persons includesincreased rates and premature onset of coronary artery disease(CAD) stroke and peripheral vascular disease compared tothe general population [1ndash6] The pathogenesis of CAD in HIVis incompletely understood a high prevalence of smokingproinflammatory and procoagulant mechanisms in the contextof immunosuppression [7ndash9] adverse viral effects on endothe-lial and other cells and deleterious metabolic effects such asdyslipidemia and insulin resistance after exposure to certain an-tiretroviral treatments have been implicated [2 10ndash12]

CAD has a strong hereditary component [13 14] Genome-wide association studies (GWAS) have identified commongenetic variants that contribute to the risk of CAD in thegeneral population [15 16] The Myocardial Infarction Assess-ment of Antiretroviral and Genetic Factors in Human Immu-nodeficiency Virus Infection (MAGNIFICENT) Consortiumwas established with the aim of assessing the relative contribu-tion of traditional risk factors HIV-related factors antiretrovi-ral regimen and genetic background to CAD in HIV-positivepersons We report here on 571 white HIV-positive personswho experienced a first CAD event and 1304 HIV-positivematched controls without CAD events in 24 HIV observationalstudies This represents the most comprehensive geneticsndashCADstudy undertaken in HIV-positive persons

METHODS

Study Population Inclusion CriteriaThe MAGNIFICENT Consortium includes 24 HIV observa-tional studies from Europe the United States Australia andArgentina (Supplementary Data) Participants gave written in-formed consent for genetic testing The ethics committee ofeach study center approved the study Applying a case-controldesign we defined cases as HIV positive with a first CAD eventduring the study period (1 April 2000 through 31 March 2009)Controls were HIV positive and event free during the studyperiod For each case we aimed to select 3 controls from the

same cohort using risk-set sampling [17] Controls werematched only on sex to allow analysis of the effect of relevantnongenetic factors Participants with cardiovascular eventsprior to the study period were excluded Because most previousCAD GWAS in the general population were conducted in pop-ulations of European descent [16] the present report is restrict-ed to participants of European descent

CAD EventsCAD events were validated by the treating physician anddefined according to the Data Collection on Adverse Events ofAnti-HIV Drugs (DAD) study and the MONICA Project ofthe World Health Organization [2 18] CAD events includeddefinite myocardial infarction (MI) possible MI or unstableangina percutaneous coronary intervention including coronaryangioplasty and stenting coronary artery bypass surgery andfatal CAD which required evidence of CAD before death AllCAD events in participating cohorts that occurred during thestudy period were included

Power Calculation Genotyping and Quality ControlWe interrogated 23 single-nucleotide polymorphisms (SNPs)with known CAD association in GWAS meta-analysis in thegeneral population [16]Using the ESPRESSO-CC Power Calcu-lator [19] with projected 600 cases and 1800 controls the studyhad an 80 power to capture the effect of SNPs with minorallele frequency (MAF) ge01 and CAD odds ratio (OR) ge125Genotyping was performed on the Metabochip (IlluminaEindhoven the Netherlands and Broad Institute HarvardUniversityMassachusetts Institute of Technology Boston MA)a custom array of 196 725 SNPs from gene regions associatedwith multiple metaboliccardiovascular traits in GWAS [20]The Metabochip was developed by leader groups in the fieldto facilitate affordable genotyping of (1) recognized SNPs and(2) the genetic regions that carried them with the goal ofdiscovering causal variants associated with the recognized tagSNPs However the study was not designed or powered for aMetabochip-wide association study which would require a

2 bull CID bull HIVAIDS

at Universiteit van A

msterdam

on May 7 2013

httpcidoxfordjournalsorgD

ownloaded from

significance threshold of P lt 25 times 10ndash7 (ie P = 05 divided bythe number of SNPs interrogated [196 725])

Participants were filtered based on gender check (heterozy-gosity testing) and cryptic relatedness We used a modified Ei-genstrat approach to identify and exclude population outliersand to control for the possibility of spurious associations result-ing from residual population stratification [21] This methodderives the principal components of the correlations amongcommon (MAF gt 5) gene variants which reflect populationancestry and corrects for those correlations in the subsequentassociation tests by integrating the coordinates of the signifi-cant principal component axes as covariates (Eigenstrat covari-ates) in the models

Nongenetic CAD Risk FactorsCovariates were selected a priori based on published CAD effectand included in the final model regardless of statistical signifi-cance high total cholesterol (gt62 mmolL [22] or being onlipid-lowering medication) low high-density lipoprotein (HDL)cholesterol (lt104 mmolL) [22] diabetes mellitus (confirmedplasma glucose level ge70 mmolL [fasting] or ge111 mmolL[nonfasting] or taking antidiabetic medication) [23 24]hypertension (systolic blood pressure ge140 mm Hg or diastolic

blood pressure ge90 mm Hg or taking antihypertensive medica-tion) smoking (never past or current) family history of CADand age (per 5-year increments [25]) HIV-related covariateswere defined a priori based on their contribution to CAD inthe DAD study [25] CD4+ count and HIV RNA value (closestto the event date) current antiretroviral therapy exposurecurrent abacavir exposure and cumulative exposure to lopina-vir and indinavir Because few patients had ge2 years exposureand the CAD effect of 1 year and ge2 years of treatment wasequivalent these drug exposures were considered as binarycovariates (ie lt or ge1 year)

Missing DataCertain covariates were unavailable or had gt20 missing data(Supplementary Table 3) Mostly these data were systemati-cally missing in entire cohorts and therefore assumed to bemissing at random This assumption was further checked bycomparing summary statistics on nonmissing values acrosscohorts There was no evidence that cohorts differed signifi-cantly in the distribution of important confounders Thereforesingle imputation using predictive mean matching was per-formed to replace missing data for glucose total and HDLcholesterol blood pressure smoking family history duration

Figure 1 Summary of the models applied and sensitivity analyses performed Abbreviations CAD coronary artery disease HIV human immunodeficien-cy virus

HIVAIDS bull CID bull 3

at Universiteit van A

msterdam

on May 7 2013

httpcidoxfordjournalsorgD

ownloaded from

of lopinavir and indinavir exposure HIV RNA and CD4+

count Missing values were imputed using models with thepredictors age sex abacavir at time of event region and casecontrol status [26] Primary analyses utilized the imputeddataset sensitivity analyses utilized the nonimputed dataset(Supplementary Figures 1B and 3)

Genetic Association AnalysesWe built 2 a priori defined genetic risk scores using 23 SNPs(or a proxy with r2 gt 08) with known CAD association [16](Supplementary Table 2) (1) additive genetic score (number ofCAD risk alleles heterozygous = 1 homozygous = 2 that isscores ranged from 0 to 46 higher scores indicate a higherCAD risk (2) additive weighted genetic score that takes intoaccount the effect size reported in the reference paper [16] Fora SNP with for example a CAD OR of 12 reference allele = 0heterozygous = 02 homozygous risk allele = 04 The numbersobtained for each of the 23 SNPs were added to create an indi-vidual weighted genetic risk score

Statistical AnalysisA summary of the models applied and sensitivity analyses per-formed is provided in Figure 1 First we tested the associationsof nongenetic factors using a conditional logistic regressionmodel [27] Then we tested the weighted genetic score plus the5 Eigenstrat covariates and added them to the model by divid-ing study participants into 4 genetic score quartiles We made apost hoc test for an interaction between genetic score and tradi-tional risk factors plus factors that contributed to CAD in theDAD study [25] The pseudo-r2 from each conditional logisticregression model was used as an estimate of the percentage ofexplained CAD variability in the study population Analyseswere done using PLINK R SAS version 92 (SAS CorporationCary NC) and Stata version 120 (StataCorp LP CollegeStation TX)

Sensitivity AnalysesTo assess the robustness of results we repeated the final modelin participants with (1) complete (nonimputed) data for all co-variates (2) stringent case definition (definite MI coronaryartery bypass surgery and fatal CAD plus corresponding con-trols) (3) definite MI plus corresponding controls (4) familyhistory of CAD excluded from the model

Exploratory Genetic Association AnalysesFirst all 196 725 SNPs present on the Metabochip were sepa-rately tested for association with CAD by conditional logisticregression Second to search for additional weaker genetic as-sociations in the regions containing known CAD-associatedgenes or variants we considered as a group all SNPs locatedinnear (plusmn5 kb) the 23 CAD-associated genes [16] and as a sep-arate group the SNPs mapping to genes associated with traits

indirectly related to CAD (total low-density lipoprotein andHDL cholesterol diabetes mellitus fasting glucose level bodymass index [20]) The distribution of association P values wascompared between these groups and all other SNPs genotypedon the Metabochip using the 2-sample Kolmogorov-Smirnovtest Third we evaluated a potential association of CAD eventswith mitochondrial DNA (mtDNA) haplogroups

Results

Study PopulationWe received DNA specimens from 702 cases and 1849 controlsTwenty-one cases and 158 controls were excluded because ofregistration in the cohort of the control after the event date oftheir matched case (n = 124) insufficient DNA quantity orquality (n = 42) sample administrative error (n = 7) nonwhiteself-reported origin (n = 4) event occurred after study ended(n = 1) or missing genetic consent (n = 1) After genotypingquality control 97 cases were excluded because they were popu-lation outliers in the Eigenstrat analysis (n = 89) or geneticallyrelated with another participant (n = 8) corresponding con-trols were also excluded The final study population included1875 participants (571 cases and 1304 controls)

Among the 571 cases there were 273 definite MI 48 possibleMI or unstable angina 179 percutaneous coronary interven-tions 32 coronary artery bypass surgeries and 39 fatal CADCharacteristics of participants are shown in Table 1 Themedian age at first CAD event was 50 years Cases were olderthan controls and more likely to be smokers and to have elevat-ed cholesterol and glucose levels a family history of CAD andcurrent treatment with abacavir

Nongenetic Factors Contributing to CADAll covariates were significantly associated with the OR of afirst CAD event except low HDL cholesterol (P = 29) being onantiretroviral therapy at time of CAD event (P = 12) CD4+

count (P = 44) and HIV viremia (P = 88) (Figure 2) In thecomplete case analysis (participants without missing covariatedata) the sample size was 720 individuals (183 cases 537 con-trols) For the imputed models the sample size was 1875 (571cases 1304 controls) Conditional logistic regression models forthe imputed dataset were consistent with the complete caseanalysis as regards direction and effect size of individual covari-ates (Supplementary Figures 1B and 3) Therefore the finalmodel and the results presented hereafter are based on theimputed dataset

CAD Odds Ratio According to Genetic Risk ScoreIn unadjusted analysis (Table 2) participants in the third andfourth genetic risk score quartiles had an increased CAD ORcompared to the first quartile (OR = 134 95 confidence

4 bull CID bull HIVAIDS

at Universiteit van A

msterdam

on May 7 2013

httpcidoxfordjournalsorgD

ownloaded from

Table 1 Characteristics of the Cases and Controls at the Matching Date

Characteristic Cases Controls

Total No 571 1304

Male sexa 911 913Age y median (range) 500 (22ndash855) 450 (165ndash813)

Smoking

Never 228 314Past 233 216

Current 539 469

Hypercholesterolemia 455 318Low HDL cholesterol 431 393

Diabetes mellitus 194 136

Arterial hypertension 436 311Family history of coronary artery disease 257 154

Receiving antiretroviral therapy 877 793

Currently on abacavir 256 176Duration of treatment with indinavir y median (range) 0 (0ndash82) 0 (0ndash113)

Duration of treatment with lopinavir y median (range) 0 (0ndash80) 0 (0ndash87)

CD4+ T-cell count cellsμL median (range) 497 (11ndash1688) 500 (10ndash1905)HIV RNA log copiesmL median (range) 38 (0ndash146) 39 (0ndash136)

HIV RNA

lt50 copiesmL 632 602lt400 copiesmL 741 682

All values are percentages unless otherwise specified

Abbreviations HDL high-density lipoprotein HIV human immunodeficiency virusa Cases and controls were matched by sex and cohort

Figure 2 Contribution of traditional coronary artery disease (CAD) risk factors HIV-related factors and weighted genetic score to CAD risk in multivari-able analysis Results are represented as the estimated effect and 95 confidence interval on the odds ratio of a first CAD event for genetic riskscore quartile (black dots) HIV-related variables (gray triangles) and traditional CAD risk factors (gray squares) Results for the final fully adjusted model(Supplementary Table 1A) and for the weighted genetic risk score (see Methods section) are shown Abbreviations ART antiretroviral therapy CAD coro-nary artery disease CI confidence interval HDL high-density lipoprotein HIV human immunodeficiency virus

HIVAIDS bull CID bull 5

at Universiteit van A

msterdam

on May 7 2013

httpcidoxfordjournalsorgD

ownloaded from

interval [CI] 1ndash179 P = 05 and OR = 159 95 CI 119ndash213 P lt 01 respectively) In the final multivariable model theadditive and the weighted genetic score were associated withthe CAD odds (P = 13 times 10minus4 and P = 29 times 10minus4 respectively)Cases were more likely to be in the upper 2 genetic score quar-tiles compared to controls (P = 01 Supplementary Figure 2)The effect of the weighted genetic score (CAD OR = 147 forthe fourth quartile compared to first quartile 95 confidence

interval [CI] 106ndash204 P = 02) was similar to the effect of es-tablished CAD risk factors and certain antiretroviral medica-tions (Figure 2 and Supplementary Table 1A) This includedfamily history (OR = 205 95 CI 154ndash274) hypertension(OR = 136 95 CI 106ndash173) hypercholesterolemia (OR =151 95 CI 116ndash196) diabetes (OR = 166 95 CI 110ndash249) current smoking (OR = 248 95 CI 185ndash332) ge1 yearlopinavir exposure (OR = 136 95 CI 106ndash173) and currentabacavir treatment (OR = 156 95 CI 117ndash207) An unfa-vorable genetic background had an additive effect on the CADodds without a significant interaction effect (P = 60 Figure 3and Supplementary Table 1B)

Relative Contribution of Clinical HIV-Related and GeneticFactorsIn the final model 75 of the CAD odds ratio variability wasexplained by age 31 by current smoking 19 by familyhistory and 09 by genetic score Smaller percentages wereexplained by traditional and HIV-related risk factors forexample 07 each by hypercholesterolemia ge1 year lopinaviror current abacavir treatment 05 each by diabetes or hyper-tension (Figure 4) Addition of the genetic score to the clinicalmodel improved the fit of the model (χ2 = 628 P = 01)

Sensitivity AnalysesModels restricted to participants with nonimputed data strin-gent case definition (344 cases 806 corresponding controls)and definite MIs (273 cases 651 corresponding controls)showed similar results except for a widening of confidence in-tervals due to reduced sample size (Supplementary Figures 3ndash5)Removing family history did not change the estimates for thegenetic score (Table 2 Supplementary Figure 6)

Exploratory Metabochip-wide Analyses mtDNAVariantsNone of the 196 725 SNPs on the array were associated withCAD events in a metabochip-wide analysis after correction formultiple testing (Supplementary Figure 7) The global

Table 2 Odds Ratio for Coronary Artery Disease According to Weighted Genetic Risk Score Quartile

Genetic RiskScore Quartile

Genetic Risk Score and5 Eigenstrat Covariates

Unadjusted forNongenetic Covariates

Final ModelWith Family

History of CADa

Final ModelWithout FamilyHistory of CADb

Quartile 2 vs quartile 1 127 (95ndash169) P= 11 103 (74ndash144) P= 84 104 (75ndash144) P= 82

Quartile 3 vs quartile 1 134 (1ndash179) P= 05 125 (90ndash174) P= 18 125 (90ndash172) P= 18

Quartile 4 vs quartile 1 159 (119ndash213) Plt 01 147 (106ndash204) P= 02 147 (106ndash203) P= 02

Data in parentheses are 95 confidence intervals

Abbreviation CAD coronary artery diseasea See Figure 2b See Supplementary Figure 7

Figure 3 Coronary artery disease (CAD) risk according to genetic scorequartile and number of non-genetic CAD risk factors (odds ratio and 95confidence interval) Participants are stratified into 12 groups by weightedgenetic score quartile (quartile 1 2 3 and 4) and by the number of nonge-netic risk factors (0ndash2 3ndash4 or gt4 nongenetic CAD risk factors) The firstgroup is the reference group (odds ratio = 1) ie participants with 0ndash2 non-genetic risk factors who are in genetic risk score quartile 1 The sum of allnongenetic CAD risk factors is considered (presence of risk factor = 1absence of risk factor = 0) including traditional risk factors and additionalfactors that contributed significantly to CAD risk in the DAD study [25]ie age past smoking exposure ge1 year to lopinavir exposure ge1 year toindinavir current exposure to abacavir Abbreviations CAD coronaryartery disease CI confidence interval HIV human immunodeficiencyvirus Q quartile

6 bull CID bull HIVAIDS

at Universiteit van A

msterdam

on May 7 2013

httpcidoxfordjournalsorgD

ownloaded from

distribution of association P values was not significantly differ-ent between the 5070 SNPs within CAD-associated genes the36 863 SNPs within genes with a potential indirect CAD associ-ation [16] and the rest of the SNPs on the Metabochip indicat-ing an absence of enrichment of potentially interestingassociation signals in the first 2 groups The mtDNA coverageof the Metabochip was found to be unreliable with an excess ofseemingly polymorphic (heteroplasmic) mtDNA SNPs 80 ofthe participants displayed ge1 heteroplasmic SNP out of 135SNPs mapped to mtDNA (total 5604 heteroplasmic callsSupplementary Results Supplementary Figure 8) Results werealso inconsistent with the complete mtDNA genomes of 2 par-ticipants that had been previously Sanger sequenced [28] with47 unmatched SNPs (64135 positions SupplementaryTable 4)

DISCUSSION

This is the first large-scale analysis of clinical HIV-related andgenetic risk factors that contribute to CAD in HIV-positivepersons Our findings suggest that the effect of an unfavorablegenetic background on CAD events is comparable to well-established traditional risk factors and certain antiretroviral

regimens The genetic risk score which was defined a prioriand captures the joint effect of 23 common SNPs with knownCAD association in the general population [16] remained in-dependently associated with CAD after considering multiplenongenetic factors and in sensitivity analyses suggesting thatthe effect is robust

In this HIV-positive study population genetic backgroundexplained a larger proportion of the CAD variability than diddiabetes hypertension or dyslipidemia but a smaller propor-tion than age or current smoking Our exploratory analysesusing the metabochip did not provide any novel insight aboutthe genetics of CAD in HIV-positive persons This was expect-ed as the study was designed to assess a panel of candidateSNPs with validated CAD association and was not powered formetabochip-wide discovery of novel gene variants

Family history and genetic risk score contributed to CAD toa similar degree and the effect of the genetic score did notchange after adjusting for family history This suggests thatfamily history which may reflect genetic background but alsoenvironmental social and lifestyle factors shared among familymembers [29 30] and assessment of common genetic variantscapture independent complementary effects on CAD in HIV-positive persons This is consistent with the results by Ripatti

Figure 4 Coronary artery disease (CAD) variability explained by traditional risk factors human immunodeficiency virusndashrelated factors and genetic back-ground Variability in the CAD odds ratio explained by the final model 211 Of this age 75 current smoking 31 past smoking 04 high totalcholesterol 07 hypertension 05 diabetes 05 low high-density lipoprotein cholesterol 01 family history of CAD 19 genetic risk score09 current antiretroviral therapy 02 current abacavir 07 lopinavir (ge1 year) 07 indinavir (ge1 year) 03 HIV load 0 CD4+ count 0Abbreviations ABC abacavir CAD coronary artery disease HDL high-density lipoprotein IDV indinavir LPV lopinavir

HIVAIDS bull CID bull 7

at Universiteit van A

msterdam

on May 7 2013

httpcidoxfordjournalsorgD

ownloaded from

and colleagues a large study in the general population that as-sessed a similar genetic score [31]

An unfavorable genetic background had an effect on CADcomparable to certain antiretroviral agents known to increasecardiovascular risk Because the magnitude of the CAD effect ofcertain drugs (eg abacavir) is unresolved [32] we based selectionof drug exposure covariates on the DAD study the largestongoing consortium of observational HIV studies [2 25] The in-creased CAD risk associated with lopinavir and indinavir is con-sistent with their previously recorded metabolic effects [2 33]Low CD4+ count or detectable HIV viremia at the time of theCAD event were not associated with CAD in our dataset con-sistent with the DAD study [34] Other authors have notedadverse effects of immunosuppression on CAD risk [35 36]

Strengths of this study include the assembly of a large studypopulation of HIV-positive persons who experienced a first CADevent during a 9-year study period rigorous quality control of thegenotyping data exclusion of population outliers and correctionfor residual population stratification physician validation of allCAD events analysis of only SNPs and nongenetic covariates withestablished CAD association and robust results in sensitivity anal-yses Our study was limited by the effort required to establish theMAGNIFICENT Consortium Even though HIV-positive popula-tions are aging [37] the number of HIV-positive persons whohave experienced CAD events is limited and not all studies includeconsent for genetic testing Because demonstration of the CADeffect of 13 of the 23 SNPs required meta-analysis of gt86 000 par-ticipants from multiple GWASs in the general population [16]additional CAD-associated SNPs with modest effect sizes mayemerge from HIV-positive study populations larger than theMAGNIFICENT consortium At the time of study designGWAS-based CAD associations were essentially limited to whitepopulations so we restricted the present analysis to white partici-pants our findings may not be applicable to other populations

Our findings suggest that genetic testing may provide prog-nostic information complementary to that afforded by familyhistory traditional risk factors and antiretroviral regimen Par-ticularly in high-risk patients knowledge of a deleterious geneticCAD predisposition might further emphasize the rationale foraggressive risk factor modification and selection of a CAD-neutral antiretroviral regimen to achieve HIV control The clini-cal value of genetic testing will rely on demonstration of im-proved CAD risk stratification in prospective studies as shownby Ripatti in the general population [31 38] This was beyond thescope of the case-control design of our consortium Areas forfuture investigation include addition of genetic score to forexample Framingham or DAD score in prospective HIV studysettings comparison of genetic CAD prediction in HIV-positiveversus HIV-negative populations and integration of geneticbackground and plasma biomarkers of inflammation coagula-tion and endothelial function to predict CAD in HIV [7 39]

Supplementary Data

Supplementary materials are available at Clinical Infectious Diseases online(httpcidoxfordjournalsorg) Supplementary materials consist of dataprovided by the author that are published to benefit the reader The postedmaterials are not copyedited The contents of all supplementary data are thesole responsibility of the authors Questions or messages regarding errorsshould be addressed to the author

Notes

Acknowledgments The authors acknowledge the contributions of thedifferent research institutions and the effort and commitment of investiga-tors study nurses laboratory personnel and participants We thank theCardio-Metabochip Consortium for designing the Metabochip and makingit available at a competitive price and the participants of the IdiPAZBiobank integrated in the Spanish Hospital Biobanks Network for clinicalsamplesFinancial support This work was supported by the Swiss National

Science Foundation (SNF project 324730_1276311) the Swiss HIVCohort Study (SNF grant 33CS30_134277 SHCS project 599) an INFEC-TIGEN grant from the Universities of Geneva and Lausanne (to P E T)the Intramural Research Program of the National Institute of Allergy andInfectious Diseases and unrestricted grants from Gilead Sciences andMerck Sharp amp Dohme Switzerland to the SHCS research foundation TheIdiPAZ Biobank is supported by Instituto de Salud Carlos III SpanishHealth Ministry (RETIC RD09007600073) and Farmaindustria throughthe Cooperation Program in Clinical and Translational Research of theCommunity of Madrid Red de Investigacioacuten en SIDA grant (ISCIII-RETICRD0600061017 ISCIII-MA060164)Potential conflicts of interest A B W currently is Manager Scientific

Affairs with the Crucell Vaccine Institute the present work was initiatedwhile she was working full-time at the Academic Medical Center (AMC) atUniversity of Amsterdam and has been completed through her continuedaffiliation with the AMC H F G has been adviser andor consultant forGlaxoSmithKline (GSK) Abbott Gilead Novartis Boehringer IngelheimRoche Tibotec Pfizer and Bristol-Myers Squibb (BMS) and his institutionhas received funding from Roche Abbott BMS Gilead Astra-ZenecaGlaxoSmithKline and MSD C T has received lecture fees from Viiv andtravel expenses from Viiv and Gilead L J has been a consultant forBMS E M has received consultancy and lecture fees from Abbott BIBMS Gilead GSK MSD Theratechnologies Tibotec and ViiV J R has re-ceived advisorylectureconsulting fees from Abbott Bionor BI BMSGilead GSK Merck Novartis Janssen Pfizer Vertex and ViiV and his in-stitution has received funding from Abbott MSD and Roche J-C W hasreceived advisorylecture fees from Abbott BMS BI Pfizer and Tibotecand travel fees from Gilead S D has received advisory fees from BMS andhis institution has received consultancy fees from Viiv and travel fees fromJanssen Viiv Gilead Abbvie and BMS S M has received consultingadvi-sory fees from AbbVie BI BMS Gilead Janssen and Roche and travelgrants from the same companies plus MSD J G has received consultingfees and his institution has received funding from Gilead Abbott BMSMSD and Janssen The institution of C B K P and M L has receivedfunding from BI BMS Gilead GSK Janssen MSD Pfizer andRoche M L has received payments for serving on data and safety monitor-ing boards for Sirtex Pty Ltd A D has received advisoryconsultant feesfrom Janssen ViiV Abbott and Gilead travel grants from ViiV andAbbott and funding from ViiV and Janssen A D has received advisoryfees from Gilead and Viiv consulting fees from Janssen and Abbott andtravel grants from Abbott and Viiv and his institution has received fundingfrom Viiv and Janssen I B has received research funding and consultinglecture fees from Abbott Gilead BMS ViiV and Janssen J G G has re-ceived consultancy fees from Abbott Janssen MSD BMS Gilead ViiVand Roche J R A has received advisoryspeaker fees and grant supportfrom Viiv Tibotec Janssen Abbott BMS Gilead and MSD E N has

8 bull CID bull HIVAIDS

at Universiteit van A

msterdam

on May 7 2013

httpcidoxfordjournalsorgD

ownloaded from

received consultant fees from Gilead BI MSD Abbott Tibotec GSK andBMS P D has received advisory consultant andor lecture fees and hasbeen a data safety monitoring board member for Gilead Abbott JanssenBI BMS MSD Theratechnologies ViiV and Ferrer G F has received con-sultancylecture fees and funding from AbbVie Janssen Gilead BMSMSD Pfizer Roche and ViiV A M currently is Medical Director HIVEndocrinology with EMD Serono the present work was initiated while shewas working full-time at Tufts University and has been completed throughher continued affiliation with Tufts R Wrsquos institution has received travelgrants from Abbott BI BMS Gilead GSK MSD Pfizer Roche TRBChemedica and Tibotec P R has been adviser for GSK Gilead andJanssen and his institution has received funding andor travel support fromGilead ViiV MSD Janssen BMS and BI He has served on data safetymonitoring boards and endpoint adjudication committees for Janssen andhis institution has received honoraria for development of educational pre-sentations from Gilead H C Brsquos institution has received travel grants hon-oraria and unrestricted research grants from GSK BMS Gilead JanssenRoche Abbott Tibotec BI and ViiV and is supported by unrestrictedgrants from Santeacutesuisse and the Gottfried and Julia Bangerter-Rhyner-Foundation P E T has received advisory fees from Janssen consultancyfees from Gilead and honoraria from Viiv and his institution has receivedadvisory fees from Gilead and MSD honoraria from Viiv and travel ex-penses from MSD and Viiv All other authors report no potential conflictsAll authors have submitted the ICMJE Form for Disclosure of Potential

Conflicts of Interest Conflicts that the editors consider relevant to thecontent of the manuscript have been disclosed

References

1 drsquoArminio A Sabin CA Phillips AN et al Cardio- and cerebrovascularevents in HIV-infected persons AIDS (London) 2004 181811ndash7

2 Friis-Moller N Reiss P Sabin CA et al Class of antiretroviraldrugs and the risk of myocardial infarction N Engl J Med 20073561723ndash35

3 Grinspoon SK Grunfeld C Kotler DP et al State of the science confer-ence initiative to decrease cardiovascular risk and increase quality ofcare for patients living with HIVAIDS executive summary Circula-tion 2008 118198ndash210

4 Obel N Thomsen HF Kronborg G et al Ischemic heart disease inHIV-infected and HIV-uninfected individuals a population-basedcohort study Clin Infect Dis 2007 441625ndash31

5 Periard D Cavassini M Taffe P et al High prevalence of peripheralarterial disease in HIV-infected persons Clin Infect Dis 2008 46761ndash7

6 Triant VA Lee H Hadigan C Grinspoon SK Increased acute myocar-dial infarction rates and cardiovascular risk factors among patientswith human immunodeficiency virus disease J Clin Endocrinol Metab2007 922506ndash12

7 Ford ES Greenwald JH Richterman AG et al Traditional risk factorsand D-dimer predict incident cardiovascular disease events in chronicHIV infection AIDS 2010 241509ndash17

8 Kuller LH Tracy R Belloso W et al Inflammatory and coagulationbiomarkers and mortality in patients with HIV infection PLoS Med2008 5e203

9 Strategies for Management of Antiretroviral Therapy Study GEl-SadrWMLundgren JD et al CD4+ count-guided interruption of antiretro-viral treatment N Engl J Med 2006 3552283ndash96

10 Sabin CA Worm SW Weber R et al Use of nucleoside reverse tran-scriptase inhibitors and risk of myocardial infarction in HIV-infectedpatients enrolled in the DAD study a multi-cohort collaborationLancet 2008 3711417ndash26

11 Riddler SA Smit E Cole SR et al Impact of HIV infection andHAART on serum lipids in men JAMA 2003 2892978ndash82

12 Stein JH Klein MA Bellehumeur JL et al Use of human immunodefi-ciency virus-1 protease inhibitors is associated with atherogenic

lipoprotein changes and endothelial dysfunction Circulation 2001104257ndash62

13 Lloyd-Jones DM Nam BH DrsquoAgostino RB Sr et al Parental cardio-vascular disease as a risk factor for cardiovascular disease in middle-aged adults a prospective study of parents and offspring JAMA 20042912204ndash11

14 Murabito JM Pencina MJ Nam BH et al Sibling cardiovasculardisease as a risk factor for cardiovascular disease in middle-aged adultsJAMA 2005 2943117ndash23

15 Clarke R Peden JF Hopewell JC et al Genetic variants associated withLp(a) lipoprotein level and coronary disease N Engl J Med 20093612518ndash28

16 Schunkert H Konig IR Kathiresan S et al Large-scale association anal-ysis identifies 13 new susceptibility loci for coronary artery disease NatGenet 2011 43333ndash8

17 Essebag V Genest J Jr Suissa S Pilote L The nested case-control studyin cardiology Am Heart J 2003 146581ndash90

18 World Health Organization MONICA Manual Part IV Event Regis-tration Section 1 Coronary Event Registration Data Component 1999Available at wwwktlfipublicationsmonicamanualpart4iv-1htmAccessed 17 December 2012

19 Burton PR Hansell AL Fortier I et al Size matters just how big isBIG Quantifying realistic sample size requirements for humangenome epidemiology Int J Epidemiol 2009 38263ndash73

20 Voight BF Kang HM Ding J et al The metabochip a custom genotyp-ing array for genetic studies of metabolic cardiovascular and anthro-pometric traits PLoS Genet 2012 8e1002793

21 Price AL Patterson NJ Plenge RM Weinblatt ME Shadick NA ReichD Principal components analysis corrects for stratification in genome-wide association studies Nat Genet 2006 38904ndash9

22 Executive Summary of the Third Report of the National CholesterolEducation Program (NCEP) Expert Panel on Detection Evaluationand Treatment of High Blood Cholesterol In Adults (Adult TreatmentPanel III) JAMA 2001 2852486ndash97

23 Genuth S Alberti KG Bennett P et al Follow-up report on the diagno-sis of diabetes mellitus Diabetes Care 2003 263160ndash7

24 Rotger M Gsponer T Martinez R et al Impact of singlenucleotide polymorphisms and of clinical risk factors on new-onset di-abetes mellitus in HIV-infected individuals Clin Infect Dis 2010 511090ndash8

25 Friis-Moller N Thiebaut R Reiss P et al Predicting the risk of cardio-vascular disease in HIV-infected patients the data collection onadverse effects of anti-HIV drugs study Eur J Cardiovasc Prev Rehabil2010 17491ndash501

26 White IR Royston P Wood AM Multiple imputation usingchained equations Issues and guidance for practice Stat Med 201130377ndash99

27 Clayton D Hills M Statistical models in epidemiology New York NYOxford University Press 1993

28 Ortiz M Poloni ES Furrer H et al No longitudinal mitochondrialDNA sequence changes in HIV-infected individuals with and withoutlipoatrophy J Infect Dis 2011 203620ndash4

29 Berg AO Baird MA Botkin JR et al National Institutes of HealthState-of-the-Science Conference Statement Family History and Im-proving Health Ann Intern Med 2009 151872ndash7

30 Chow CK Islam S Bautista L et al Parental history and myocardial in-farction risk across the world the INTERHEART Study J Am CollCardiol 2011 57619ndash27

31 Ripatti S Tikkanen E Orho-Melander M et al A multilocus geneticrisk score for coronary heart disease case-control and prospectivecohort analyses Lancet 2010 3761393ndash400

32 Triant VA HIV infection and coronary heart disease an intersection ofepidemics J Infect Dis 2012 205(suppl 3)S355ndash61

33 Periard D Telenti A Sudre P et al Atherogenic dyslipidemia in HIV-infected individuals treated with protease inhibitors The Swiss HIVCohort Study Circulation 1999 100700ndash5

HIVAIDS bull CID bull 9

at Universiteit van A

msterdam

on May 7 2013

httpcidoxfordjournalsorgD

ownloaded from

34 Sabin C Worm SW Law M et al on behalf of the DAD study groupAssociation between markers of immunosuppression and the risk of car-diovascular disease (CVD) the DAD study In 19th Conference on Ret-roviruses and Opportunistic Infections Seattle WA 2012 Poster 822

35 Klein DL Leyden WA Xu L et al Contribution of immunodeficiencyto CHD cohort study of HIV+ and HIVndash Kaiser Permanentemembers In 18th Conference on Retroviruses and OpportunisticInfections Boston MA 2011 Poster 810

36 Lang S Mary-Krause M Simon A et al HIV replication andimmune status are independent predictors of the risk of myocardial

infarction in HIV-infected individuals Clin Infect Dis 201255600ndash7

37 Hasse B Ledergerber B Furrer H et al Morbidity and aging in HIV-in-fected persons the Swiss HIV cohort study Clin Infect Dis 2011531130ndash9

38 Thanassoulis G Vasan RS Genetic cardiovascular risk prediction willwe get there Circulation 2010 1222323ndash34

39 Duprez DA Neuhaus J Kuller LH et al Inflammation coagulationand cardiovascular disease in HIV-infected individuals PLoS One2012 7e44454

10 bull CID bull HIVAIDS

at Universiteit van A

msterdam

on May 7 2013

httpcidoxfordjournalsorgD

ownloaded from