Cholera in Bahrain: epidemiological characteristics of an outbreak*

Bulnofte Bulletin of the World Health Organization, 59 (1): 61-66 (1981) Cholera in Bahrain: epidemiological characteristics of an outbreak* ROBERT A. GUNN,1 ANN M. KIMBALL,1 P. P. MATHEW,2 S. R. DUTTA,3 & A. H. M. RIFAAT, 3 In theperiod 1OAugust 1978-23 January 1979, 913 culture-confirmed cases ofcholera caused by Vibrio cholerae, biotype El Tor, serotype Ogawa, occurred in Bahrain. After discovery ofthe initial cases, others occurred sporadically, and the incidence reached apeak of25-35 cases per day during the seventh week of the outbreak (16-22 September). The overall attack rate (27per10 000) was low and the outbreaksubsided without mass immuniz- ation campaigns or rigorous border control ofpersons and imports. Investigation of 746 culture-confirmed cases that occurred in the period 10 August-13 October 1978, showed that cases occurred throughout most areas ofthe country and mainly affected infants, young children, and adult working-age males. Symptoms were very mild; fewer than 20% of patients required specific rehydration therapy. The highest attack rate (84 per 10 000) occurred in infants less than 1 year ofage. No common vehicle or mode oftransmission was identified. A matched-pairstudy of35 cases and controlsshowed that adult cases were more likely than controls to have consumedfood or beverage outside ofthe home before becoming ill. V. cholerae was isolatedfrom stored drinking water in the houses of8 cases but notfrom numerous samples of food and tap-water. It was presumed that cholera transmission occurred through a complex interaction ofmildand asymptomatically infectedpersons with food, water, and the environment. The El Tor biotype of Vibrio cholerae O-Group 1 (01) was first recognized in South-West Asia in 1965, when Bahrain reported one case. Since that time El Tor cholera has continued to spread throughout South-West Asia and westward into Africa and southern Europe (1). In 1972 and 1973, Bahrain ex- perienced limited outbreaks caused by V. cholerae El Tor serotype Inaba, with 74 cases in 1972 and 37 cases in 1973. Routine cultures of stool, food, and water since 1973 have been negative for V. cholerae 01. In this report we describe an investigation of an out- * Requests for reprints should be addressed to: Enteric Diseases Branch, Bacterial Diseases Division, Bureau of Epidemiology, Centers for Diseas Control (CDC), Atlanta, Georgia, 30333, USA. I Medical Epidemiologist, Bureau of Epidemiology, Centers for Disease Control, Atlanta, GA 30333, USA. (WHO Consultant). 2Ministry of Health, Bahrain. break of 913 culture-confirmed cases of V. cholerae, biotype El Tor, serotype Ogawa, that occurred in Bahrain in the period 10 August 1978-23 January 1979. The epidemiological characteristics of the out- break and the results of a case-control study designed to determine the vehicle and mode of transmission are reported. OUTBREAK Bahrain is a small island state situated in the Arabian Gulf with an area of 670 km2 and a popu- lation of 341 000 (1978 estimate). Municipal drinking water is provided by approximately 60 deep wells and 10 desalinization plants. Desalinated water is pre- ferred by most people for drinking and tea making. 4031 - 61 R. A. GUNN ET AL. Well water is not chlorinated, whereas desalinated water obtained from the larger plants is reportedly chlorinated to not less than 0.2 mg/litre. Samples of water from deep wells and desalinated water are routinely monitored for coliforms and since 1973 also for V. cholerae 01. Chlorine residuals are monitored at the distribution point of desalination plants. Most households have piped-in well water, electricity, and refrigeration. Drinking water is often stored in 20-litre screw-top plastic containers. On 10 August 1978, V. cholerae, biotype El Tor, serotype Ogawa, was isolated from a rice-water-stool specimen from a man residing in Manama City, the capital of Bahrain. The following day, 2 additional cases were reported among his immediate contacts. All 3 patients had recently had close contact with visitors from a country that in the subsequent week reported cholera cases. Despite quarantine of the patients and their contacts, other cases occurred sporadically in most regions of the country, some- times in people who had had no known contact with cholera patients. The number of cases increased markedly in the week following the Eid religious holiday (3-5 September) and then gradually increased to a peak of 25-35 per day in the seventh week of the outbreak (16-22 September). There was no evidence that the patients had been exposed to a common vehicle. Almost all cases occurred in the poorer people, with a disproportionately large number (10.9%) in infants less than 1 year of age. The signs and symptoms were, in general, very mild and severe dehydration was rare; only 2 deaths occurred (a 6-week-old infant and an 83-year-old man). Control measures focused on improving community and restaurant sanitation and on active surveillance of new cases. Case households were visited by a health inspector within 24 hours of bacteriological confirmation of the first case; a prophylactic antibiotic (doxycycline-approximately 6 mg/kg body weight daily for 3 days) was adminis- tered to all household contacts. Throughout the country, efforts to monitor water supplies and health education activities were intensified. A single cholera immunization was provided only to those who requested it and no significant changes in immigration and import regulations were instituted. Because cases were continuing despite initial control measures and the vehicle(s) and mode of trans- mission were unknown, the investigation was ex- panded and a case-control study was carried out. MATERIALS AND METHODS Case data were obtained from a log book on all symptomatic, culture-positive persons treated at a special, centrally located diarrhoea/cholera clinic during the period 10 August- 13 October 1978. A case-control study was conducted in a recently constructed (inhabited in 1%8) urban community of 11 000 people situated 8 km south of the capital city. Houses were of concrete block construction, had piped water, and were connected to a municipal sewer- age system. They were inhabited by people of all socio- economic levels, except for the very poor and very rich. The cases (n = 35) included in the study were all identified culture-confirmed symptomatic patients living in the community and the surrounding villages whose illness began between the beginning of the out- break and 25 September; however, most of the cases occurred in the 2 weeks before 26 September, when the study began. For each case, a control subject matched by age and sex was systematically selected from the neighbourhood. The control subjects were selected from households in which no cases of cholera or afebrile diarrhoea had occurred in the preceding 2 months. Laboratory investigations included culturing faecal swabs, food, and water specimens (450 ml). Each specimen was enriched in alkaline peptone water for 6-8 hours, after which the enrichment culture was streaked on thiosulfate-citrate-bile salts-sucrose (TCBS) agar. After overnight incubation, suspicious colonies were agglutinated with polyvalent and type- specific antisera. V. cholera 01 isolates from water and randomly selected isolates from stools were bio- chemically identified. The 27 isolates of V. cholerae 01 obtained both during the early and later stages of the outbreak, including rough strains isolated in the last week of September, were phage typed (2). RESULTS All cases In the period 10 August-13 October 1978, 746 culture-confirmed cases of V. cholerae, biotype El Tor, serotype Ogawa, were identified (Fig. 1). The distribution of infant cases was similar to that of other cases. Incidence was highest (84 per 10 000) among infants less than 1 year of age (Table 1), and Fig. 2 shows that males 20-39 years old also had a relatively high attack rate. In the case-control study there were 17 cases in males and 18 in females; 15 (43%) were more than 12 years old, 17 (49%) were between 1 and 11 years of age, and 3 (8%) were less than 1 year old. Cases were distributed throughout the community without evi- dence of clustering. Of the 35 cases, 23 patients (66%) were admitted to hospital primarily for isolation from the community rather than for treatment; only 6 (17%) required rehydration therapy. 62 CHOLERA IN BAHRAIN ALL CASES (IN 74) * INFANS CI YEAR (NuSI) Ix lS 25 1 . 00 lx La z 8 iS 22 29 6 MALE --- FEMALE AGE Fig. 2. Number of bacteriologically confirmed, clinical cases of cholera per 10 000 population, by age (years) and sex, 5 August-6 October 1978 (based on 1976 population estimate). WEEKNO. 2 3 4 a 6 7 a 9 10 Fig. 1. Number of bacteriologically confirmed, clinical cases of cholera, by week of onset, Bahrain, 5 August- 10 October 1978 (week no. 10 = 4 days only). Table 1. Cholera cases,' age-specific attack rates, Bahrain, 10 August-6 October 1978 Age group Populationb Attack rateC 0-2 months 2 421 45.4 3-5 months 2 383 50.4 84.4 6-8 months 2 357 132.1 9-11 months 2 321 116.8 1 year 8 893 78.7 2 years 8 359 50.2 42.8 3 years 7 858 25.5 4 years 7386 20.3 5-9 years 41 530 12.8 10-19 years 72 970 16.7 20-29 years 50 750 36.7 30-39 years 28 190 25.9 40-49 years 21 240 14.6 50-59 years 13 550 6.6 >60 years 10 070 14.9 280 278 25.3 a Bacteriologically confirmed clinical illness. b 1976 estimated population. c Per 10000 population. The case-control study did not identify significant differences between cases and controls as regards most of the variables that were examined: persons per household (cases 9.8/controls 8.6), number of persons per household less than 15 years of age (cases 4.8/controls 4.0), number of persons per bedroom (cases 3.5/controls 3.3), standard of education of case or head of household, occupation of head of house- hold, source of drinking water, intrahousehold handling and storage of drinking water, individual consumption of water per day, use of ice, ingestion of raw foods, subjective sanitary rating, or close contact with visitors from other cholera infected countries. Although we did not identify a common restaurant or vehicle of transmission, the study did show that adult cases (12 years or older, n = 15) were significantly more likely than controls to have eaten or drunk in "restaurants" (8 cases:0 controls) during the 5 days before they became ill (P<0.02, McNemar's matched- pair test). Infant cases Eighty-one cases of cholera in infants less than 1 year of age were identified during the investigation period. Of these, 34 (42%) were male and 47 (58%) were female; 11 (13.6%) of the infants were less than 3 months old, 12 (14.8%) were 3-5 months old and 58 (71.6%) were 6-11 months old. The attack rates by age group, calculated from crude population estimates derived from a geometric progression of annual survival rates, show a markedly elevated rate in the 6-11-month age group (Table 1). Infant cases were not clustered by area or place of residence. 220 200 ISO 140 i4O ta 120 41 100 U00 60 40 20 WEIEK ENDING 4 I 63 -- F-.,, .%F %,%,%, FWF%A.UL1%.F11, M U11%, l Vulp, U11%AMMAIIIIIM R. A. GUNN ET AL. From medical records it was determined that of 43 hospitalized infants, 23 (53qo) received intravenous rehydration, although only 19 were reported to be dehydrated. Dehydration was reported as moderate to severe in 11 cases (26%o). Serum bicarbonate levels were studied in 15 infants and 6 (40lo) were less than 20 mmoUlitre; however, the lowest being 17 mmoUlitre. Body temperature was recorded as >38.3 IC in 14 infants (33%) but this was not related to the presence of dehydration. The significance of the observed temperature elevation was often not commented upon in the medical notes. One 6-week- old infant died after approximately 11 hours of hospi- talization. The mean hospital stay was 4.4 days (range 2-10 days). Laboratory During a one-month period (9 September-8 October) 1629 stool specimens from food handlers were submitted to the Public Health Department of Bahrain for routine examination and during the out- break approximately 4000 additional specimens were submitted. Of the food handlers tested, 25 were infected with V. cholerae 01 (infection rate approxi- mately 0.4%). Of these, 7 had symptomatic cholera. During August and September, 236 food and bever- age samples were cultured and all were negative for V. cholerae 01. In the same period, more than 700 routine water samples obtained from bore-hole wells or their nearest distribution tap and from various de- salination plants were also negative for V. cholerae 01. No samples showed grossly elevated coliform counts; in a small percentage of samples, low total coliform counts (less than 20 per 100 ml) were noted, but in these samples the counts soon returned to normal. In contrast, V. cholerae 01 was isolated from stored drinking water (either tap or distilled water) from 20-litre storage containers, thermos jugs, or refriger- ated drinking pitchers in the houses of 8 cases. In 4 out of the 5 specimens examined for total coliforms, the counts were greater than 180 per 100 ml (most prob- able number). Samples obtained from tap-water in the same house and in some instances from neigh- bouring houses were all negative for coliforms and V. cholerae 01. Of 27 confirmed isolates of V. cholerae 01 that were phage typed, all but 2 were El Tor, phage type IV, defined as sensitive to El Tor phages 1, 2, and 5. DISCUSSION The 1978 El Tor cholera outbreak in Bahrain, al- though considerably larger than outbreaks in 1972 and 1973, still affected a very small percentage (0.3%) of the inhabitants. Cases occurred sporadically through- out the country and clusters around common sources were not identified. The outbreak gradually reached a peak during the seventh week, quickly declined to only a few cases a day, and finally terminated on 23 January 1979. Although cholera had spread through- out all districts of Bahrain, the relatively small number of cases identified shows that a country such as Bahrain that has protected municipal water supplies, a moderate level of sanitation, and an active public health department can contain cholera and can do so without employing stringent border restrictions, as regards people and goods, and/or mass immunization campaigns. One of the striking characteristics of the outbreak was the mild symptoms manifested by the majority of patients. Only 2 deaths occurred (case: fatality ratio 0.2%) and less than 20% of the cases required re- hydration fluids. An outbreak of V. cholerae, El Tor, Ogawa, that occurred in the eastern provinces of Saudi Arabia in 1978, also demonstrated a similar pattern of very mild disease with low morbidity and mortality (Y. Watanbe, personal communication, 1978). Although the ratio of asymptomatic and mild cases to severe cases has been shown to be much greater with El Tor cholera than with classical cholera (3, 4), the former can also cause dehydration and death; for example, recently in Africa and the Maldive Islands severe cases and deaths occurred (5, and CDC, unpublished data, 1978). The reason for the mild disease pattern observed in Bahrain and Saudi Arabia is unknown. In this investigation, incidence was calculated on the basis of cases identified among persons who had attended a diarrhoea/cholera clinic rather than a random population survey which might have been more accurate. Even though the clinic was centrally located and transportation facilities were adequate, certain population groups may have been more likely to seek diagnosis and treatment than others. This may have accounted for the high incidence recorded in children 0-4 years of age. Other data from this study of infant cholera cases showed that bottle feeding was a significant risk factor for infants acquiring symptomatic cholera (6), and this is borne out by the very high attack rates in the 6-11-month age group, the age at which infants are weaned in this population. Numerous other studies have shown that diarrhoeal disease incidence in children is highest during and shortly after the weaning period (7). The high attack rate in males in the working-age group in Bahrain suggests that their mobility outside the home may have increased their exposure to contaminated food or water. Investigation of this outbreak failed to identify a common vehicle or mode of transmission. Although 64 CHOLERA IN BAHRAIN 65 in past outbreaks, case-control studies were successful in epidemiologically implicating a vehicle of infection when likely vehicles were initially suspected (8-10), when this is not the case they are probably of limited value. In this outbreak, the case-control study suggested that adults may have acquired their infec- tion from food or beverages consumed outside the home. Transmission among other family members may have occurred from contamination of food and water within the household. The occurrence of intra- household transmission is supported by the isolation of V. cholerae 01 from stored drinking water within 8 affected households, while concomitant examination of public piped-in water was negative for coliforms and V. cholerae 01. Studies in Bangladesh showed that safe public water supplies (tube wells) did not prevent the spread of El Tor cholera because the population had contact with alternative unsafe water sources that were frequently used for food preparation, bathing, and dishwashing (11, 12). Since infection with El Tor cholera, in contrast to classical cholera, results in a longer period of Vibrio shedding and a higher incidence ofasympto- matic infection, and because the El Tor organism survives longer in the extra-intestinal environment, El Tor cholera may in some instances be transmitted through a complex interaction of contaminated food, water, and environment rather than through public drinking water supplies (12). Transmission of cholera in Bahrain probably occurred in this manner. Continued detailed microbiological and epidemio- logical investigations are needed to further delineate the precise mode of El Tor cholera transmission. ACKNOWLEDGEMENTS We gratefully acknowledge the support of Dr F. Partow and Dr V. Parisi of the WHO Regional Office for the Eastern Mediterranean, Alexandria, Egypt, under whose auspices this investigation was conducted. We also wish to thank Dr Ali Fakhro, Minister of Health; Dr Akbar Mohsin, Chairman, Department of Pediatrics; Dr Ahmad A. Ahmad, Chief of Staff, Salmaniya Hospital; and Dr Sameer Khalfan, Ebrahim Yousef, Iftikhar Nasimi, Khairia Asiri, Lila Philipose, and other members of the Public Health Directorate and Ministry of Health in Bahrain, who assisted in this investigation. Dr Charlotte Parker, Department of Microbiology, University of Texas, Austin, Texas, USA, performed the cholera phage typing. RESUML LE CHOLERA A BAHREYN: CARACTERES EPIDEMIOLOGIQUES D'UNE EPIDEMIE Entre le 10 aofit 1978 et le 23 janvier 1979, on a observe A Bahrein 913 cas de cholera, confirmes par culture, dus A Vibrio cholerae, biotype El Tor, serotype Ogawa. Apres la d6couverte des cas initiaux, d'autres se sont presentes de maniere sporadique, atteignant un pic de 25-35 cas par jour durant la septieme semaine de l'epidemie (16-22 septembre). Le taux d'atteinte global (27 pour 10 000) 6tait faible et l'epidemie a regresse sans campagnes de vaccination de masse ni contrbles rigoureux aux frontieres des personnes et des importations. Les enquetes portant sur 746 cas survenus au cours de la periode 10 aofit-13 octobre 1978 et confirmes par culture ont montre que la plupart des regions du pays ont Wtt atteintes et que la maladie a surtout affecte les nourrissons, les jeunes enfants et les hommes adultes en Age de travailler. Les symptbmes etaient tres discrets; moins de 20% des malades ont eu besoin d'une rehydratation spcifique. Le taux d'atteinte le plus N1ev6 (84 pour 10 000) concernait les nourrissons de moins d'un an. Aucun vehicule ou mode de transmission courant n'a W identifie. Une etude de 35 cas et temoins par paires assorties a montre que les cas adultes avaient plus probablement que les temoins consomme de la nourriture ou des boissons hors de leur domicile avant de tomber malades. V. cholerae a ete isole de 1'eau de boisson conserv6e au foyer de 8 cas, mais non dans de nombreux echantillons d'aliments et d'eau du robinet. On a suppose que la trans- mission du cholera s'est faite grace a l'interaction complexe de divers facteurs: personnes atteintes d'infections benignes et asymptomatiques, aliments, eau et environnement. REFERENCES 1. KAMAL, A. M. The seventh pandemic of cholera. In: Barua, D. & Burrows, W., ed., Cholera, Philadelphia, Saunders, 1974, pp. 1-14. 2. BASU, S. & MUKERJEE, S. Bacteriophage typing of Vibrio eltor. Experientia, 24: 299-300 (1968). 3. BART, K. J. ET AL. Seroepidemiologic studies during a simultaneous epidemic of infection with El Tor Ogawa and classical Inaba Vibrio cholerae. Journal of infec- tious diseases, 121 (Suppl.): 17-24 (1970). 66 R. A. GUNN ET AL. 4. MCCORMACK, W. M. ET AL. A community study of inapparent cholera infections. American journal of epidemiology, 89: 658-664 (1968). 5. GOODGAME, R. W. ET AL. Cholera in Africa: a message for the West. Annals of internal medicine, 82: 101-106 (1975). 6. GUNN, R. A. ET AL. Bottle-feeding as a risk factor in cholera in infants. Lancet, 2: 730-732 (1979). 7. GORDON, J. E. ET AL. Weanling diarrhea. American journal of medical sciences, 248: 345-377 (1964). 8. BLAKE, P. A. ET AL. Cholera in Portugal, 1974. I. Modes of transmission. American journal of epide- miology, 105: 337-3 (1977). 9. BAINE, W. B. ET AL. Epidemiology of cholera in Italy in 1973. Lancet, 2: 1370-1381 (1974). 10. MCINTYRE, R. C. ET AL. Modes of transmission of cholera in a newly infected population on an atoll and implications for control measures. Lancet, 1: 311-314 (1979). 11. LEVINE, R. J. ET AL. Failure of sanitary wells to protect against cholera and other diarrhoeas in Bangladesh. Lancet, 2: 86-89 (1976). 12. SOMMER, A. & WOODWARD, W. E. The influence of protected water supplies on the spread of classical! Inaba and El Tor/Ogawa cholera in rural East Bengal. Lancet, 2: 985-987 (1972).