Journal of Infection and Public Health (2011) 4, 12—21

A comparison of disease caused by Shigella andCampylobacter species: 24 months communitybased surveillance in 4 slums of Karachi, Pakistan

Sajid Bashir Soofia,∗, Muhammad Atif Habiba, Lorenz von Seidleinc,Muhammad Jawed Khana, Shah Muhammada, Naveed Bhuttoa,Mohammad Imran Khana, Shahid Rasoola, Afia Zafarb,John D. Clemensc, Qamaruddin Nizamia, Zulfiqar A. Bhuttaa

a Department of Paediatrics & Child Health, Aga Khan University, Pakistanb Department of Microbiology, Aga Khan University, Pakistanc International Vaccine Institute, Seoul, Republic of Korea

Received 26 March 2009; received in revised form 10 October 2010; accepted 15 October 2010

KEYWORDSDiarrhea;Campylobacteriosis;Shigellosis;Children;Epidemiology

Summary Despite the efforts of the international community diarrheal dis-eases still pose a major threat to children in children less than five years ofage. Bacterial diarrhea has also emerged as a public health concern due to theproliferation of drug resistant species in many parts of the world. There is apaucity of population-based data about the incidence of shigellosis and Campy-lobacter infections in Pakistan. We report country specific results for Shigelladiarrhea that were derived from a multicenter study conducted in six Asian coun-tries. Disease surveillance was conducted over a 24 month period in urban slumsof Karachi, Pakistan, a city with a population of 59,584. Cases were detectedthrough passive detection in study treatment centers. Stool specimens or rectalswabs were collected from all consenting patients. Between January 2002 andDecember 2003 10,540 enteric infection cases were detected. The incidence rateof treated diarrhea in children under 5 was 488/1000/year. In children, 5 years

and older, the diarrhea rate was 22/1000/year. 576 (7%) Campylobacter isolateswere detected. The pre-dominant Campylobacter species was C. jenuni with an increase of 29/1000 year in children under 5 years. Shigella species were iso-lated from 394 of 8032 children under 5 years of age. Shigella flexneri was thedominant species (10/1000/year in children under 5 years) followed by Shigellasonnei (3.9/1000/year), Shigella boydii (2.0/1000/year) and Shigella dysenteriae(1.3/1000/year). Shigellosis and Campylobacter infection rates peaked during the

∗ Corresponding author at: Department of Pediatrics & Child Health, Aga Khan University, P.O. Box 3500, Stadium Road, Karachi74800, Pakistan. Tel.: +92 21 4864798.

E-mail address: sajid.soofi@aku.edu (S.B. Soofi).

1876-0341/$ — see front matter © 2010 King Saud Bin Abdulaziz University for Health Sciences. Published by Elsevier Ltd. All rights reserved.

doi:10.1016/j.jiph.2010.10.001

A comparison of disease caused by Shigella and Campylobacter species 13

second year of life. The incidence rate of shigellosis increased in old age but sucha trend was not observed in Campylobacter infections. Of 394 shigellosis patients123 (31%) presented with dysentery in contrast to only 54 (9%) of 576 patients withCampylobacter infections (p < 0.001). Both Campylobacter infections and shigellosisare common in community settings of Pakistan but shigellosis presented more fre-quently with abdominal pain and dysentery than Campylobacter infections indicatingthat shigellosis may be a more severe illness than Campylobacter infections. Due tothe increased and disease severity, drug resistant shigella have become a significanthealth problem; moreover it is a disease of poor and impoverished people who do nothave the access to standard water and sanitary conditions, health care services oroptimal treatment. In the face of these facts it is empirically important to develop alow cost effective vaccine that can protect these populations for a longer duration.

dulaziz University for Health Sciences. Published by Elsevier

I

Daae[ibmiSSi

maicldudrdaTitcaucaa

odtn

asrcics[lea

aagdnbrCpacct

atpopitlergs and colleagues [17].

The Campylobacter disease burden is also quite

© 2010 King Saud Bin AbLtd. All rights reserved.

ntroduction

espite efforts of the global community to controlnd prevent the diarrheal diseases, they still remainmajor cause of morbidity and mortality in childrenspecially in impoverished areas of South East Asia1,2] moreover the incidence of dysentery (whichs an aggravated diarrheal condition characterizedy frequent passage of stool containing blood anducus) has also increased. Much of the dysentery

s considered to be of bacillary origin species ofhigella (S. flexneri, S. sonnei, S. dysenteriae, and. boydii) and Shigellosis has become a public healthssue [3].

On a global level, it is estimated that about 165illion cases of diarrhea due to Shigella occurs

nnually, 99% occur in developing countries, andn developing countries 69% of episodes occur inhildren under five years of age. About 1.1 mil-ion deaths are attributed to Shigella infections ineveloping countries, 60% of which occur in thender-five age group [3]. A multicenter study con-ucted in six Asian countries illustrates incidenceates of Shigellosis up to 13% [4]. A combination ofisease severity and antibiotic resistance is posingreal dilemma especially in young children [5,6].

he recognition of extended range beta lactamasesn Shigella strains may well jeopardize the effec-iveness of antibiotics including third-generationephalosporins [7—9]. The appearance of resistancegainst the majority of antimicrobials has addedrgency to the development of vaccines whichan protect against shigellosis while to date onlysingle vaccine to protect against shigellosis is

ccredited worldwide [10].Campylobacter infection is also a common cause

f acute gastroenteritis both in the developed andeveloping countries and like other bacterial infec-ions incidence is much higher in the developingations due to poor water and sanitary conditions

hBig

nd mishandling of food [11,12]. Based on thetudies from Thailand and Mexico the incidenceates are estimated to be 40,000 per 100,000 forhildren younger than five years [11]. In develop-ng countries, Campylobacter infections in childrenommonly occur under the age of two years and canometimes leads to death and other complications12]. In several community-based surveys Campy-obacter jejuni was the most commonly isolatednteropathogen in diarrheal stools from childrenged less than 5 years [13].

Like shigellosis Campylobacter infections arelso wide spread and C. jejuni is considered to becommon cause of acute bacterial gastroenteritis

lobally. The organism becomes more seriousue to its association with serious post-infectioneurological complications. A study conductedy Clarence et al. reveals a much increasedisk of Gullien Barre Syndrome (GBS) amongampylobacter enteritis patients compared withreviously reported studies [14]. Campylobacterre widespread in both developed and developingountries but due to increasing incidence andomplication much attention is required to addresshe problem in developing countries [15].

Drug resistance is also an important factor whileddressing Campylobacter infections. Addressinghese infections in many developed countries is ariority due to the emergence of increased levelsf macrolide resistance [16]. Further the infectionsortray the same situation of drug resistance evenn the developing countries which is evident fromhe studies conducted in Thailand by Serichanta-

igh in the south Asian region. Studies fromangladesh, India and Pakistan reveals that it

s one of the most frequent organisms causingastroenteritis [18]. In Pakistan the presence of

TQmmsmaaaOetlbeefiadof

tptEhdcttaofp

rwwmmahndddwsp

14

Campylobacter organisms are quite high in foodcommodities which are usually used by the commonman [19], which in turn results in an increased inci-dence of Campylobacter infections (12%) in all agegroups [20]. Another study conducted at Rawalpindireveals that C. jejuni is the most frequent causeof childhood diarrhea and dysentery in that setting[21].

Due to its epidemiological spread, Pakistan is apotential site for future vaccine trial and in ground-ing of future vaccine trial sites it is crucial torecognize the shigellosis incidence as shigellosis ismuch more notorious than the other enteric organ-isms. Our study is the foremost community basedstudy to estimate the factual incidence of shigel-losis on a population level. Earlier hospital basedstudies do not provide a denominator essential forincidence estimates [22—29]. Not only is it signif-icant to appreciate the shigellosis incidence in astudy area it is evenly imperative to understandthe distribution of Shigella serogroups, types, andsubtypes because the knowledge about cross pro-tection between Shigella serotypes and subtypesis very limited at best [30]. We have consequentlyconducted shigellosis surveillance in 4 urban slumsof Karachi over a 24 months period. The unexpect-edly high detection rates of Campylobacter specieshave allowed us to evaluate and contrast the epi-demiology as well as the clinical presentation ofshigellosis and Campylobacter infections.

Methods

This study was done as a part of the Multicenterstudy for Shigella diarrhea in six Asian countries[4] to compare shigellosis with campylobacterio-sis in four slum communities, in Karachi, Pakistan(Rehri Goth, Sherpao Colony, Sultanabad and HijratColony). Pashto and Hazara migrants from theNWFP Province of Pakistan are the major populationgroups in the impoverished slum areas of Sherpaocolony, Hijrat colony and Sultanabad. Rehri Goth isa much older fishing community mainly consisting ofSindhi speaking settlers. A census was conducted ineach of the communities prior to the surveillancethat demonstrated a population of 59,584 indi-viduals including 8381 children less than 5 years.Karachi is a tropical area with hot summers last-ing from April through October and more moderatetemperatures for the rest of the year, rains are

expected during the second half of each summerbut can skip one or more years.

The health care system of these slums con-sisted of 79 private practitioners, among them 32were the graduates from medical colleges [31].

ihmss

S.B. Soofi et al.

he quackery medical practice was common anduacks or ‘‘village doctors’’ who have no for-al training or degree used to prescribe allopathicedications even antibiotics. Untrained pharmacy

hopkeepers and registered pharmacists provideedication based on the description of the signs

nd symptoms of the patient over the counter. Inddition, a variety of traditional healers (Hakims)nd homoeopaths practiced in the study areas.nce these primary health care providers have beenxhausted the residents of the study areas can seekreatment from hospitals. A study clinic was estab-ished in each of the 4 slums for more than a yearefore the surveillance started, these clinics werequipped with trained doctors, medicines and rel-vant laboratory equipment. The clinics providedree medical services for residents. The availabil-ty of free healthcare in the study clinics had beennnounced during community meetings when resi-ents were asked to attend the study clinics in casef diarrhea. The presence of the study clinics wasurther publicized during the census.

We estimated the burden of enteric diseaseshrough a population-based, surveillance system. Aassive case detection system was used which cap-ured patients when they visited the study clinics.ach household was visited weekly and the house-old head or a representative was asked aboutiarrhea cases in the preceding 7 days. If a diarrheaase was reported the patient was asked to come tohe study clinic. The study followed a generic pro-ocol, which was adapted to local needs by staffnd collaborating centers [32]. Verbal consent wasbtained from each participant (parent or guardianor children) following an explanation of the pur-ose of the study.

Consenting patients of all age groups with diar-hea or dysentery presenting to the study clinicsere eligible to participate in the study. Diarrheaas defined as three or more loose bowel move-ents during a 24-h period, dysentery as one orore loose bowel movements with visible blood,

nd fever as an axillary temperature of 37.5 ◦C origher. Diarrhea following three days or more oformal bowel movements was considered a newiarrhea episode. For every patient presenting withiarrhea, a case report form (CRF), describingemographics and medical history & examinationas completed and two rectal/stool swabs or a

tool specimens were obtained. All the cases wererovided with the standard treatments as outlined

n the protocols. All consenting patients with aistory of dysentery or diarrhea for three days orore were eligible to participate in the study. The

tudy received approval from the Aga Khan Univer-ity Ethics Review Committee and the Secretariat

A pylobacter species 15

CW

dropwtftPpatwcvToAocr

MBwLrlaf(awCmibfThSusT

iiMmaunt

10540 Diarrhea Patients presented *

90 Revisits70 Not eligible†

2339 didn’t provide a specimen8032 patients enrolles

576 campylobacter Isolates¶

547 C. Jejuni24 C.Coli5 C. Larid

394 Shigella Isolates‡

242 S. Flexerni72 S. Sonnei42 S. Boydii37 S. dysenteriae

340 Other Isolates§

129 V. Cholerae56 E.Coli98 Salmonella Spp.57 Aeromonas Spp.

6756 No organism Isolated

*During the study period 1/1/2002 to 12/31/03†Didn’t fulfill the inclusion criteria‡From 22 patients more than one organism was isolated¶From 25 patients more than one organism was isolated§From 21 patients more than one organism was isolated

Figure 1 Case assembly. *During the study period1/1/2002 to 12/31/03. †Did not fulfill the inclusion cri-t ‡

iii

sStddct[wadeiiPw

luTrptw

comparison of disease caused by Shigella and Cam

ommittee for Research Involving Human Subjects,HO, Geneva, Switzerland.To isolate the exact organism we set up a stan-

ard laboratory at every clinic. We obtained threeectal swabs one for buffered glycerol saline (BGS),ne for alkaline peptone water APW and one forhosphate buffered saline PBS and along with thate ask the patient to bring a stool sample in a con-

ainer for stool culture. All these medias were usedor the enrichment and transport of the isolates tohe Central Lab at Aga Khan University, Karachi,akistan. The collected stool samples were alsolated immediately onto XLD-, SS-, MacConkey’s-gar and incubated at 37 ◦C for 18—24 h beforeransport at the center. All swabs and stool samplesere stored refrigerated and transported daily in aool box to the Central Laboratory at Aga Khan Uni-ersity, Karachi following the standard protocols.he Central Laboratory at Aga Khan University isne of the state of the art laboratories in the Southsian region. To ensure the robust implementationf laboratory procedures, the doctors of the healthenters received an extensive training for all labo-atory procedures.

On arrival specimens in BGS were plated onacConkey’s agar and Salmonella-Shigella agar.iochemical reactions of lactose negative coloniesere evaluated with Triple Sugar Iron agar and

ysine Indole Motility medium isolates producingeactions consistent with shigella were sero-ogically confirmed by slide agglutination withppropriate group-specific polyvalent antisera,ollowed by type-specific monovalent antiseraDenka-Seiken, Tokyo, Japan). In cases where nogglutination occurred with live bacteria, the testas repeated with boiled suspensions of bacteria.ampylobacter was identified after 48 h by colonyorphology, oxidase test, Gram staining, and motil-

ty. The organisms were identified to species levely a positive catalase test, a negative urease test,ailure to produce H2S, and non-fermentation inSI, resistance to cephalotin, nitrate reaction andippurate hydrolysis. Although in the main study ofhigella diarrhea the investigators look for Shigellasing PCR but that was not done in Pakistan. All PCRtudies were conducted at the USAFRIMS, Bangkok,hailand [4].

All Case Record Forms were archived seriallyn the box files and then were double enterednto a custom-made data entry programs (FoxPro,icrosoft, Redmond, WA, USA). The data manage-

ent programs included error as well as consistency

nd range check programs. Chi-square tests weresed for binary data analysis. For the analysis ofon-normally distributed data, Wilcoxon rank sumest was applied. Data were analyzed with SAS

R

D1

eria. From 22 patients more than one organism wassolated. ¶From 25 patients more than one organism wassolated. §From 21 patients more than one organism wassolated.

oftware (SAS Institute Inc., Cary, NC, USA) andtata/SE 8 (Stata Corp., Texas, USA). A p-value lesshan 0.05 (2-tailed) was considered significant. Inci-ence rates were calculated by using age-specificenominators of the population residing in theatchment area in 2001. We calculated 95% CI forhe incidence rates by the Wilson score method33]. As the observation period was 24 months,e assumed that each person residing in the studyrea contributed 24 months of person time to theenominator. The number of age-specific diseasepisodes was used as numerator. For the compar-son of clinical presentations only patients werencluded from whom a single organism was isolated.atients with multiple infections (co-infections)ere excluded.Through out the surveillance period, the senior

evel staff did regular ongoing monitoring and eval-ation of various process and outcome indicators.he data management unit produced an indicatoreport on weekly basis that reflected the flow of theroject activities. To ensure the proper implemen-ations of the project, regular refresher trainingsere provided to the staff involved in the project.

esults

uring the 24 month study period, which startedst January 2002 and ended on 31st December

16 S.B. Soofi et al.

0

100

200

300

400

500

600

700

800

900

1000

Dec-03Nov-03Oct-03Sep-03Aug-03Jul-03Jun-03May-03Apr-03Mar-03Feb-03Jan-03

no. o

f cas

esm

m o

f pre

cipi

tatio

n

0

5

10

15

20

25

30

35

o Cel

sius

Diarrhea

Dysentery

Temp (°C)

Precip (mm)

0

10

20

30

40

50

60

70

80

90

Dec-03Nov-03Oct-03Sep-03Aug-03Jul-03n-03

no. o

f cas

eso C

elsi

us

0

50

100

150

200

250

300

mm

of p

reci

pita

tion

shigellosis

campylobacter

Precip (mm)

Temp (°C)

a

b

lation to weather changes in Karachi, Pakistan in 2003.

serotypes were 2a (38 of 89 isolates; 16%) followedby S. flexneri 6 (37 isolates; 15%), S. flexneri 1b (25isolates; 10%), and S. flexneri 2b (23 isolates; 10%).

JuMay-03Apr-03Mar-03Feb-03Jan-03

Figure 2 (a and b) Seasonality of enteric infections in re

2003 10,540 enteric infection cases were detected(Fig. 1). The incidence rate of treated diarrhea inchildren under 5 years was 488/1000/year com-pared to 22/1000/year in the older population. Thishigh incidence in young children is a clear evi-dence that the enteric infections still prevail in theyounger populations.

Enteric infections (diarrhea, dysentery) andspecifically Campylobacter infections and shigel-losis increased during the warmer months of theyear (Fig. 2a and b). There was no correlationbetween changes in enteric infections rate andrainfall in 2003.

From 8032 stool specimens 394 (5%) Shigellaspecies were isolated. S. flexneri was the dom-

inant species (10/1000/year in children under5 years) followed by S. sonnei (3.9/1000/year),S. boydii (2.0/1000/year) and S. dysenteriae(1.3/1000/year). The dominant S. flexneri

Figure 3 Incidence of children under 5 years presentingwith dysentery in Karachi, Pakistan.

A comparison of disease caused by Shigella and Campylobacter species 17

0

10

20

30

40

50

60

70

10th9th8th7th6th5th4th3rd2nd1st

year of life

inci

den

ce r

ate

/ 100

0 / y

ear

shigellosis

campylobacter

0

0.5

1

1.5

2

2.5

70+60 to 6950 to 5940 to 49 30 to 3920 to 2910 to 19

years of life

inci

den

ce r

ate

/ 100

0 / y

ear

shigellosis

campylobacter

a

b

Cam

ti(lciF

tsolTaytti

sStwi(

wpTpt(5oawpo

D

SrCs

Figure 4 (a and b) Age distribution of Shigella and

During the study period 576 (7%) Campylobac-er isolates were detected. The most frequentlysolated Campylobacter species was C. jejuni29/1000/year in children under 5 years) fol-owed by Campylobacter coli (1.4/1000/year inhildren under 5 years) and Campylobacter lar-dis (0.3/1000/year in children under 5 years) (seeig. 3).

There were significant differences betweenhe presentation of Campylobacter infections andhigellosis. Shigellosis patients were significantlylder (mean age 6.3 years) compared to Campy-obacter patients (mean age 2.8 years, p < 0.001;able 2). As illustrated in Fig. 4a both shigellosisnd Campylobacter infections peaked in the 2ndear of life and dropped below 10/1000/year byhe 5th year of life. In contrast to Campylobac-er infection rates, shigellosis rates increased withncreasing age after age 30 (Fig. 4b).

Of the 10,540 enteric infections 1123 (11%) pre-ented with dysentery. Out of which 123 (11%) werehigella species and 54 (5%) were Campylobac-

er species respectively. The dysentery incidenceas highest in S. flexneri patients (3.5/1000/year

n children under 5 years) followed by C. jejuni2.5/1000/year; Fig. 5).

Cfltsa

pylobacter infected patients in Karachi, Pakistan.

Of 394 shigellosis patients 123 (31.2%) presentedith dysentery in contrast only 54 (9.3%) of 576atients with Campylobacter infections (Table 1).he percentage of Campylobacter and shigellosisatients presenting with dysentery varied withhe age of the patient and the infecting speciesFig. 5). Of 173 children with S. flexneri infections8 (34%) presented with dysentery in contrast to 31f 69 (45%) adults (p = 0.05; Fig. 5). The percent-ges of Campylobacter infected patients presentingith dysentery were identical to the percentagesatients presenting with dysentery from whom norganism could be isolated.

iscussion

urveillance in impoverished areas of Karachi foundelatively high shigellosis and even higher rates ofampylobacter infection. Consistent with previoustudies the dominant Campylobacter species was

. jejuni and the dominant Shigella species was S.exneri [34]. The infection rates were highest inhe extremes of age. Both infections peaked in theecond year of life. Newborns may be protectedgainst shigellosis and Campylobacter infection by

18 S.B. Soofi et al.

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

no organismC.laridC.jejuniC.coliS.boydiiS.dysentriaeS.sonneiS.flexneri

% p

atie

nts

pres

entin

g w

ith d

ysen

tery

% of children < 5 years

% of individuals 5 years and older

Figure 5 The percentage of patients with Shigella or Campylobacter infections presenting with dysentery (the 95%confidence intervals are shown as whiskers). *No organism isolated.

Table 1 Total population, number of cases over a 24 months observation period, incidence rates and 95% confidenceintervals in Karachi, Pakistan.

<60 monthsn = 8381a

≥60 monthsn = 51,203a

Total n = 59,584a

Cases IR/1000/year Cases IR/1000/year Cases IR/1000/year

Diarrhea 8174 488 (477—498) 2276 22 (21—24) 10,450 88 (86—90)Dysentery 761 45 (41—50) 362 4 (3—4) 1123 9 (9—10)Campylobacter infections 515 31 (27—35) 61 0.6 (0.4—0.9) 576 5 (4—5)Shigellosis 292 17 (15—21) 102 1 (0.7—1.3) 394 3 (2.9—3.90)Shigellosis dysentery 82 5 (4—7) 41 0.4 (0.2—0.6) 123 1.0 (0.8—1.3)Campylobacter dysentery 45 3 (2—4) 9 0.1 (0.0—0.2) 54 0.5 (0.3—0.7)

a Denominator.

Table 2 Patient characteristics.

Shigellosis patients % Age Campylobacter patients % Age

na 372a 551b

Male 207 56% 320 58%Age (mean) 6.3 — 2.8 —Age (median) 2 — 1.5 —Dysentery 119 32% 50 9%Ever watery stools 236 63% 413 75%Mucous stools 95 26% 125 23%Abdominal pain 46 12% 43 8%Rectal prolapse 1 0.3% 2 0.4%Fever 135 36% 224 41%Vomiting 46 12% 70 13%Dehydration 17 5% 27 5%Number of stools in the last 24 h

5 or less 139 37% 201 37%6—10 118 32% 141 26%11—30 10 3% 9 2%Too many to count 105 28% 200 36%a 22 patients with co-infections have been excluded.b 25 patients with co-infections have.

A pyl

thptAwiswt

itFmCtCdiltadyP

aspwawptwotmdeAete

srspvsplst

bwt

tpCmoabTji

tK3ps2Ssr7

iaeoashAclbiisHv

arrwpeat

comparison of disease caused by Shigella and Cam

rans placental antibodies, breast milk, and per-aps lower exposure rates [35—39]. Following aeak in the first years of life the rates for both infec-ions dropped by more than an order of magnitude.fter age 30 a trend for increasing shigellosis ratesas observed but not for Campylobacter. This find-

ng may be explained by waning immunity againsthigellosis with increasing age. There is no apparentaning of immunity against Campylobacter infec-

ions.Alternatively transmission of shigellosis may

ncrease with increasing age. Shigellosis is thoughto be mostly transmitted from person to person.lies may in some places play a role in the trans-ission of this disease [40]. The transmission ofampylobacter is less well understood. It is thoughthat most poultry and many household pets areampylobacter carriers in Pakistan [41]. The ageistribution of C. jejuni infections is very differentn the USA where after a peak in the first years ofife a second peak of infections is observed duringhe third decade of life [42]. Not surprisingly thege distribution of C. jejuni cases in Bangladeshescribed by Glass and coworkers more than 20ears ago was strikingly similar to our findings inakistan [43].

Shigellosis and Campylobacter infections werelso significantly different in their presentation. Aignificantly larger proportion of shigellosis patientsresented with dysentery compared to patientsith Campylobacter infections. Indeed the percent-ge of Campylobacter infected patients presentingith dysentery was similar to the percentage inatients from whom no organism could be iden-ified. It cannot be excluded that co-infectionith agents responsible for the background ratesf dysentery also cause dysentery in Campylobac-er infection patients. Older patients presentedore frequently with dysentery compared to chil-ren less than 5 years. This finding may be bestxplained by differences in health care utilization.dult patients may only seek care for severe dis-ase including dysentery. Children may be broughto health care providers for more trivial diseasepisodes.

Two thirds of the shigellosis patients had atome stage of their diarrhea episode watery diar-hea. The proportion of patients who had atome stage watery diarrhea was even higher inatients with Campylobacter infections. An obser-ation that challenges the notion which associates

higellosis exclusively with dysentery. Abdominalain was more frequently associated with shigel-osis than with Campylobacter infections. Whilehigellosis appears to cause more pain and dysen-ery. C. jejuni infections caused more frequent

a

Pst

obacter species 19

owel movements. These findings are in agreementith findings from Bangladesh as well as other coun-

ries [42,43].Shigellosis and Campylobacter infections appear

o have a similar seasonality. Consistent withrevious reports from Karachi, shigellosis andampylobacteriosis rates increased in the hot sum-er months and dropped during the colder months

f the year [44]. The notion that ‘‘the seasonal vari-tion of C. jejuni infection appears to be influencedy rainfall’’ was not supported by our findings [45].he increase in all enteric infections including C.ejuni started well before the onset of rains in 2003n Karachi.

Several hospital-based studies of enteric infec-ions have been conducted in Pakistan includingarachi. The studies found that from between% and 19% of stool specimens were Shigellaositive [22—24,26,44]. In hospital based foundtudies Campylobacter was isolated from 17% to5% of stool specimens [41,42,46—48]. In our studyhigella species were isolated was from 5% of stoolpecimens that is at the lower end of the expectedange. Campylobacter species were isolated from% of specimens which is below the expected range.

Our study differs from previous reports in thatt was community-based and included patients ofll ages hence patients with a potentially differ-nt disease severity and prevalence. Furthermoreur study was set up to detect shigellosis casesnd methods were optimized to detect Shigellapecies. Campylobacter unlike Shigella species areighly fastidious and may have escaped detection.

cohort study conducted in Mexico found thathildren under 5 years had 2.1 episodes of Campy-obacter infections per year [48]. In a hospitalased study in the US Campylobacter spp. weresolated from 5% of stool specimens [42]. A studyn Bangladesh C. jejuni in 14% of outpatients pre-enting with diarrhea to a research hospital [43].owever C. jejuni was isolated less frequently fromillage children with diarrhea (5%).

The findings of our study were consistent withnother done in Peru [49]. The Peruvian studyeveals that C. jejuni was the leading cause of diar-hea and is more frequent in malnourished childrenith low socioeconomic background. The partici-ants in our study also belonged to a low socioconomic stratum of society. The Peruvian studylso found an excess risk of GBS among Campylobac-er enteritis patients [14]. We followed our cases

nd did not find such trends.

Our study made use of passive surveillance.atients were captured when they presented totudy clinics. This method has the advantage thatrivial diarrhea cases that usually would not seek

C

F

R

[

[

[[

[

[

20

treatment are not included. Passive surveillancehowever has the disadvantage that patients whoseek care outside the study clinics will escape thesurveillance. Furthermore 23% of diarrhea patientswho were captured by the surveillance did notprovide a stool specimen. For these reasons itseems likely that our study underestimated the trueShigellosis and Campylobacter infection rates.

Shigellosis and Campylobacter infection inci-dence rates were highest in the second year of life.Decreasing rates after the second year of life couldbe due to acquired immunity. The trend for increas-ing shigellosis rates in old age which was observedfor shigellosis but not for Campylobacter infectionscould be due waning of immunity or differences inexposure. Shigellosis presented significantly moreoften frequently with abdominal pain and dysen-tery than Campylobacter infections indicating thatshigellosis may be a more severe illness and haveworse clinical consequences than Campylobacterinfections.

Conclusion

It is established that Campylobacteriosis and Shigel-losis are the most common organisms causingdiarrhea. The shigellosis is comparatively a severeform of enteric infection that needs instant atten-tion. There has also been a perceptible increasein ciprofloxacin resistance in our centre as also inother parts of country, which could be attributedto unrestricted administration of antibiotics fordiarrhea irrespective of aetiology. Many casesthe diarrhea are empirically treated with fluo-roquinolones particularly ciprofloxacin in generalpractice clinics.

Due to increased disease burden, increasedsymptom severity and proliferating drug resis-tance both campylobacteriosis and shigellosis havebecome significant health problems; moreover theyare the diseases of poor and impoverished peoplewho do not have access to clean water and sani-tary conditions, health care services and optimaltreatment. In the face of these facts it is empiri-cally important to develop a low cost vaccine thatwill protect the affected populations for a longerduration. However the barrier to develop a vac-cine against campylobacteriosis and shigellosis foruse in developing countries is the poor responsetherefore further research studies of promising

campylobacteriosis and shigellosis vaccine candi-dates are immediately desirable after confirmationof protection, immunogenicity, and safety in volun-teer settings.

[

S.B. Soofi et al.

onflict of interest statement

unding: No funding sources.Competing interests: None declared.Ethical approval: Not required.

eferences

[1] Bern C, Martines J, de Zoysa I, Glass RI. The magnitudeof the global problem of diarrhoeal disease: a ten-yearupdate. Bull World Health Organ 1992;70:705—14.

[2] Kosek M, Bern C, Guerrant RL. The global burdenof diarrhoeal disease, as estimated from studies pub-lished between 1992 and 2000. Bull World Health Organ2003;81:197—204.

[3] Kotloff KL, Winickoff JP, Ivanoff B, Clemens JD, Swerd-low DL, Sansonetti PJ, et al. Global burden of Shigellainfections: implications for vaccine development andimplementation of control strategies. Bull World HealthOrgan 1999;77:651—66.

[4] von Seidlein L, Kim DR, Ali M, Lee H, Wang XY, ThiemVD, et al. A multicentre study of Shigella diarrhoea in sixAsian countries: disease burden, clinical manifestations,and microbiology. PLoS Med 2006;3:e353.

[5] Dutta S, Dutta P, Matsushita S, Bhattacharya SK, YoshidaS. Shigella dysenteriae serotype 1, Kolkata, India. EmergInfect Dis 2003;9:1471—4.

[6] Talukder KA, Khajanchi BK, Islam MA, Dutta DK, Islam Z,et al. Genetic relatedness of ciprofloxacin-resistant Shigelladysenteriae type 1 strains isolated in south Asia. J Antimi-crob Chemother 2004;54:730—4.

[7] Naheed A, Kalluri P, Talukder KA, et al. Fluoroquinolone-resistant Shigella dysenteriae type 1 in northeasternBangladesh. Lancet Infect Dis 2004;4:607—8.

[8] Rahman M, Shoma S, Rashid H, Siddique AK, Nair GB, SackDA. Extended-spectrum beta-lactamase-mediated third-generation cephalosporin resistance in Shigella isolates inBangladesh. J Antimicrob Chemother 2004;54:846—7.

[9] Kim S, Kim J, Kang Y, Park Y, Lee B. Occurrence ofextended-spectrum {beta}-lactamases in members of thegenus Shigella in the Republic of Korea. J Clin Microbiol2004;42:5264—9.

10] Tu G, Cui C, Wang J, et al. Double-blind field trial oforal live F2a-Sonnei (FS) dysentery vaccine. J Biol Prod1999;12:178—80.

11] Oberhelman RA, Taylor D. Campylobacter infections indeveloping countries. In: Blaser MJ, Nachamkin I, edi-tors. Campylobacter. Washington, DC: American Society forMicrobiology Press; 1999. p. 139—54.

12] WHO. Campylobacter fact sheet, N255, Geneva; 2000.13] Castillo AP. Campylobacter infection in community. Indian

J Pediatr 1989;56(2 (March)).14] Tam CC, O’Brien SJ, Petersen I, Islam A, Hayward A,

Rodriguez LC. Guillain-Barré syndrome and preceding infec-tion with Campylobacter, influenza and Epstein—Barr virusin the general practice research database. PLoS ONE2007;2(4):e344.

15] Coker AO, Isokpehi RD, Thomas BN, Amisu KO, Obi CL.

Human campylobacteriosis in developing countries. EmergInfect Dis 2002;8(3 (March)):237—44.

16] Gibreel A, Taylor DE. Macrolide resistance in Campylobac-ter jejuni and Campylobacter coli. J Antimicrob Chemother2006;58(2):243—55.

A pyl

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[R. Cohort study of intestinal infection with Campylobacter

comparison of disease caused by Shigella and Cam

17] Serichantalergs O, Dalsgaard A, Bodhidatta L, KrasaesubS, Pitarangsi C, Srijan A, et al. Emerging fluoroquinolonesand macrolide resistance of Campylobacter jejuni andCampylobacter coli isolates and their serotypes in Thaichildren from 1991 to 2000. Epidemiol Infect 2007;135(8(November)):1299—306.

18] Haq JA, Rahman KM. Campylobacter jejuni as a cause ofacute diarrhoea in children: a study at an urban hospital inBangladesh. J Trop Med Hyg 1991;94(1 (February)):50—4.

19] Hussain I, Shahid Mahmood M, Akhtar M, Khan A. Preva-lence of Campylobacter species in meat, milk and otherfood commodities in Pakistan. Food Microbiol 2007;24(3(May)):219—22.

20] Khalil K, Lindblom G-B, Mazhar K, Khan SR, Kajiser B. Earlychild health in Lahore, Pakistan. VIII. Microbiology. ActaPaediatr 1993;82 [article first published online: 21 January2008].

21] Ali AM, Qureshi AH, Rafi S, Roshan E, Khan I, Malik AM, et al.Frequency of Campylobacter jejuni in diarrhoea/dysenteryin children in Rawalpindi and Islamabad. JPMA 2003;53:517.

22] Mubashir M, Khan A, Baqai R, et al. Causative agents ofacute diarrhoea in the first 3 years of life: hospital-basedstudy. J Gastroenterol Hepatol 1990;5:264—70.

23] Ahmed K, Shakoori FR, Shakoori AR. Aetiology of shigellosisin northern Pakistan. J Health Popul Nutr 2003;21:32—9.

24] Khalil K, Khan SR, Mazhar K, Kaijser B, Lindblom GB. Occur-rence and susceptibility to antibiotics of Shigella speciesin stools of hospitalized children with bloody diarrhea inPakistan. Am J Trop Med Hyg 1998;58:800—3.

25] Huilan S, Zhen LG, Mathan MM, et al. Etiology of acutediarrhoea among children in developing countries: a mul-ticentre study in five countries. Bull World Health Organ1991;69:549—55.

26] Khan MM, Iqbal J, Ghafoor A, Burney MI. Aetiologic agents ofdiarrhoeal diseases in hospitalised children in Rawalpindi,Pakistan. J Diarrhoeal Dis Res 1988;6:228—31.

27] Karamat KA. A study of enteropathogenic microbial agentsof diarrhoea disease in Rawalpindi/Islamabad areas. Pak JMed Res 1993;32:139—42.

28] Taj Y, Memon HI, Memon I. Etiological agents of dysen-tery in children under five years of age in Karachi. JCPSP2000;10:477—80.

29] Tanwani AK. Aetiological agents identified in children withdiarrhoeal disease. Pak J Pathol 1995;6:15—9.

30] Rasolofo-Razanamparany V, Cassel-Beraud AM, Roux J, San-sonetti PJ, Phalipon A. Predominance of serotype-specificmucosal antibody response in Shigella flexneri-infectedhumans living in an area of endemicity. Infect Immun2001;69:5230—4.

31] Nizami SQ, Khan IA, Bhutta ZA. Drug prescribing practicesof general practitioners and paediatricians for childhooddiarrhoea in Karachi. Pak Soc Sci Med 1996;42:1133—9.

32] Clemens JD, Kotloff K, Kay B. Generic protocol to estimatethe burden of Shigella diarrhoea and dysenteric mortality.

Geneva: World Health Organization; 1999. Available from:www.who.int/gpv-documents/.

33] Newcombe RG. Interval estimation for the differencebetween independent proportions: comparison of elevenmethods. Stat Med 1998;17:873—90.

[

Available online at www.

obacter species 21

34] Allos BM. Campylobacter jejuni Infections: update onemerging issues and trends. Clin Infect Dis 2001;32:1201—6.

35] Shahid M, Amjad M, Kazmi SU. Breast feeding may protectchildren against Campylobacter jejuni associated diar-rhoea. J Coll Physicians Surg Pak 1996;6:43—6.

36] Clemens JD, Stanton B, Stoll B, Shahid NS, Banu H, Chowd-hury AK. Breast feeding as a determinant of severity inshigellosis. Evidence for protection throughout the firstthree years of life in Bangladeshi children. Am J Epidemiol1986;123:710—20.

37] Ahmed F, Clemens JD, Rao MR, Sack DA, Khan MR, HaqueE. Community-based evaluation of the effect of breast-feeding on the risk of microbiologically confirmed orclinically presumptive shigellosis in Bangladeshi children.Pediatrics 1992;90:406—11.

38] Ahmed F, Clemens JD, Rao MR, Khan MR, Haque E. Initi-ation of food supplements and stopping of breast-feedingas determinants of weanling shigellosis. Bull World HealthOrgan 1993;71:571—8.

39] Ahmed F, Clemens JD, Rao MR, Ansaruzzaman M, Haque E.Epidemiology of shigellosis among children exposed to casesof Shigella dysentery: a multivariate assessment. Am J TropMed Hyg 1997;56:258—64.

40] Khalil K, Lindblom GB, Mazhar K, Kaijser B. Flies and wateras reservoirs for bacterial enteropathogens in urban andrural areas in and around Lahore, Pakistan. Epidemiol Infect1994;113:435—44.

41] Khalil K, Lindblom GB, Mazhar K, Sjogren E, Kaijser B. Fre-quency and enterotoxigenicity of Campylobacter jejuni andC. coli in domestic animals in Pakistan as compared toSweden. J Trop Med Hyg 1993;96:35—40.

42] Blaser MJ, Wells JG, Feldman RA, Pollard RA, Allen JR.Campylobacter enteritis in the United States. A multicenterstudy. Ann Intern Med 1983;98:360—5.

43] Glass RI, Stoll BJ, Huq MI, Struelens MJ, Blaser M, Kib-riya AK. Epidemiologic and clinical features of endemicCampylobacter jejuni infection in Bangladesh. J Infect Dis1983;148:292—6.

44] Alam M, Akhtar YN, Ali SS, et al. Seasonal variation in bac-terial pathogens isolated from stool samples in Karachi,Pakistan. J Pak Med Assoc 2003;53:125—9.

45] Mandell G, Bennett J, Dolin R, editors. Mandell, Douglas,and Bennett’s, Campylobacter. Principles and practice ofinfectious diseases. 5th ed. Philadelphia: Churchill Liv-ingston; 2000.

46] Ibrahim NG, Zafar A, Hasan R. Evaluation of frequency ofisolation and trends in antibiotic resistance among Campy-lobacter isolates over 11 year period. J Pak Med Assoc2004;54:291—4.

47] Hafiz S, Syed Y, Rauf U, Qadr B, Shahabuddin M. Etiologyand management of diarrhoeal diseases in Karachi. J PakMed Assoc 1991;41:211—3.

48] Calva JJ, Ruiz-Palacios GM, Lopez-Vidal AB, Ramos A, Bojalil

in Mexican children. Lancet 1988;1:503—6.49] Murga HL, Huicho, Guevara G. Acute diarrhoea and Campy-

lobacter in Peruvian children: a clinical and epidemiologicapproach. J Trop Pediatr 1993;39(6):338—41.

sciencedirect.com