1Department of International Health, Johns Hopkins University Bloomberg School of Public Health, and 2University of Maryland School of Medicine, Baltimore, and 3University of Maryland Biotechnology Institute, Baltimore, and Department of Cell Biology and Molecular Genetics, University of Maryland, College Park; 4International Center for Diarrheal Disease Research, Dhaka, Bangladesh; 5Rollins School of Public Health, Emory University, Atlanta, Georgia
Received 29 May 2002; revised 6 September 2002; electronically published 13 December 2002.
How Vibrio cholerae spreads around the world and what determines its seasonal peaks in endemic areas are not known. These features of cholera have been hypothesized to be primarily the result of environmental factors associated with aquatic habitats that can now be identified. Since 1997, fortnightly surveillance in 4 widely separated geographic locations in Bangladesh has been performed to identify patients with cholera and to collect environmental data. A total of 5670 patients (53% <5 years of age) have been studied; 14.3% had cholera (10.4% due to V. cholerae O1 El Tor, 3.8% due to O139). Both serogroups were found in all locations; outbreaks were seasonal and often occurred simultaneously. Water-use patterns showed that bathing and washing clothes in tube-well water was significantly protective in two of the sites. These data will be correlated with environmental factors, to develop a model for prediction of cholera outbreaks.
Presented in part: Annual U.S.-Japan Cooperative Medical Science Program, Shonan Village, Japan, 30 November–
2 December 1998; Annual U.S.-Japan Cooperative Medical Science Program, Osaka, Japan, 17–
19 January 2001; Annual Meeting of the Infectious Diseases Society of America, San Francisco, California, 25–
28 October (abstract 871).
Informed consent was obtained from the patients or their guardians. The research was approved by the Committee on Human Research of the Johns Hopkins University Bloomberg School of Public Health, and its guidelines were followed in the conduct of the clinical research.
Their were no conflicts of interest of any of the authors.
Financial support: National Institutes of Health (grant AI39129.01-05).
Reprints or correspondence: Dr. R. Bradley Sack, Dept. of International Health, Johns Hopkins University, Bloomberg School of Public Health, 624 N. Broadway, Rm. 200, Baltimore, MD 21205
Cholera, whose "traditional home" has been the Ganges Delta and Southeast Asia, is one of man's oldest scourges. Over the past 40 years, as part of the 7th pandemic, cholera has spread worldwide and is now a major public health problem in >75 countries in Asia, Africa, and South America [1]. Vibrio cholerae is one of a few pathogenic bacteria that have the potential for pandemic spread. More recently, in the latter part of 1992, a new serogroup, O139 (the first known serogroup other than O1 to cause epidemic cholera) appeared in areas surrounding the Bay of Bengal [2,4], produced major epidemics in India and Bangladesh, spread to neighboring countries, and continues to cause epidemic cholera in many of these areas. Unfortunately, in spite of many years of study, it is still not known what determines the seasonal pattern of epidemic peaks of V. cholerae, how it has spread around the world, and what determined the evolution of the new, O139 strain [5].
Cholera is well known as the classical "water-borne" disease, and the importance of aquatic ecology in its epidemiology is suggested by the close association between V. cholerae and surface water [5–
17]. Changes in aquatic ecology that result in increased numbers of these pathogenic bacteria result in their being more easily transmitted to humans.
Cholera remains a major public health problem with significant mortality wherever it occurs, despite the fact that it is a simple and inexpensive disease to treat when adequate medical supplies and knowledgeable, committed health workers are available. Unfortunately, mortality rates remain unacceptably high in many areas of the developing world [18–
20].
Specifically, cholera continues to be a major public health problem in Bangladesh. In Dhaka, the capital city, and in the well-studied, rural population of Matlab [21–
23], cholera occurs year-round, but with well-described peaks of disease during the spring and fall (before and after the monsoons) [21, 23, 24]. Little is known of the patterns of cholera in the rest of the country, however, because of lack of surveillance [23, 24].
Here we report 4.5 years of year-round bimonthly clinical cholera surveillance involving 4 widely separated areas of rural Bangladesh. These data will provide the basis for the investigation of possible associations between cholera outbreaks and changes in environmental factors relating to water ecology; the latter data are presented in a separate publication.
MATERIALS AND METHODS
Surveillance
Establishing the surveillance sites. The 4 cholera surveillance areas—
Bakerganj, Matlab, Chhatak, and Chaugachha —
were selected to represent different geographical areas of Bangladesh. Bakerganj is on the upper edge of the estuarine area of the southern coastal region of Bangladesh. Chhatak is situated in the flood plains of the river Brahmaputra. Chaugachha is on the western edge of the tidal plains of the river Ganges. Criteria for choosing these areas were as follows: (1) each had a known history of past cholera outbreaks (on the basis of information from the Epidemic Control Preparedness Programme of the International Center for Diarrheal Disease Research, Bangladesh (ICDDR,B), based in Dhaka, which collaborates with the Health Service of the Government of Bangladesh in the investigations and intervention of major cholera outbreaks); (2) travel to each site was possible throughout the year, in a timely fashion; (3) except for the Matlab diarrhea hospital, which is an ICDDR,B facility, each had a Thana Health facility (of the Government of Bangladesh) that was recognized as the treatment hospital for diarrheal diseases in that area and had a medical staff that was willing to facilitate the clinical surveillance; and (4) each had, for the hospital, a catchment area of ∼
140,000–
200,000 persons.
fig.ommitted
Figure 1. Four epidemiological surveillance sites in Bangladesh. Dhaka is the site of the laboratories; Chittagong is listed only for reference.
Matlab, a riverine area that has a well-defined population of ∼
200,000, has been under cholera surveillance by the ICDDR,B for the past 36 years and is known to be a highly cholera-endemic area [21]. Chhatak and Bakerganj had had a history of regular cholera outbreaks; Chaugaucha, on the other hand, had had no history of cholera outbreaks during the past 10 years and was designated as the control area .
Once the clinical areas of study had been selected, 4 stations (in rivers, ponds, and lakes) from each surveillance area were chosen to determine physical, microbiological and plankton data for surface-water samples.
Surveillance methods. Clinical and environmental surveillance every 15 days was begun at 2 sites, Matlab and Chhatak, in March 1997 and at the remaining 2 sites, Bakerganj and Chaugaucha, in June 1997. No sampling periods were missed. Surveillance continued through December 2001, except in the case of Chaugachha, where, for financial reasons, it was discontinued in June 2001.
Clinical surveillance. Each of the surveillance hospitals was visited by a physician who stayed for a 3-day period, during which time he personally saw all patients presenting with acute watery diarrhea. After each patient had given informed consent, a standardized history (including a questionnaire regarding water use) was taken, a physical exam was given, standard treatment (consisting mainly of fluid replacement) for watery diarrhea was administered, and a rectal swab, which was placed in Cary Blair media for culture of V. cholerae, was done. Rectal swabs from the patients in Matlab were cultured on the same day, in the Matlab laboratory; rectal swabs from the other 3 sites were transported, within 72 h, to the Environmental Microbiology Laboratory at ICDDR,B. Standard bacteriological methods were employed [25]. Rectal swabs were cultured both directly and after a 6-h enrichment in alkaline peptone water, on selective media (thiosulfate citrate bile-salts sucrose agar and taurocholate tellurite gelatin agar) at 37°C. Suspect colonies were agglutinated with antisera specific for V. cholerae O1 (Inaba and Ogawa serotypes) and for V. cholerae O139. Positive colonies were confirmed as V. cholerae by biochemical testing, and the two biotypes of O1 serogroup (either classical or El Tor) were differentiated by standard methods [25].
Statistical Methods
Clinical data collected through December 2001 were used for the analyses in this report.
Descriptive statistics. The observed number of cases of cholera (O1 and O139) and diarrhea were plotted over time, for each study area.
Association between age and occurrence of cholera. The age distributions of patients with and without cholera were compared, as were those of patients with cholera serogroups O1 and O139, by 2 tests of association. [26].
Association between water source used and occurrence of cholera. The clinical surveillance data included information on water sources (tube wells, rivers, ponds, and canals) and uses (drinking, bathing, and washing of clothes and utensils.) Age-adjusted associations between the occurrence of cholera and the water-source and water-use variables were investigated by use of multivariable logistic regressions. [26] Odds-ratio estimates and 95% confidence intervals were computed for significant risk factors in the logistic regressions.
RESULTS
Clinical Surveillance
Through December 2001, a total of 5670 patients had been seen at clinical surveillance sites. The numbers of patients with the bacteriological diagnosis of cholera are given in . V. cholerae O1 El Tor, Ogawa, was the predominant organism, although both serotypes and serogroups of V. cholerae were identified. The cholera attack rate in Matlab, the only surveillance area for which we had detailed demographic information, was 3.28/1000/year.
fig.ommitted Table 1. Patients with bacteriological diagnosis of cholera.
Epidemiological information from the 4 surveillance sites—
that is, the total number of patients with acute diarrhea who were seen every 15 days and the number with V. cholerae O1 or O139 isolated from their rectal swabs—
is shown in . The findings are as follows: Most cases of acute diarrhea were caused by agents other than V. cholerae. Matlab had by far the greatest number of cases of cholera, whereas the designated control area (Chaugachha) had the least. Although O1 vibrios were predominant in all 4 areas, O139 vibrios also were isolated at all sites at some time during the surveillance. In 2 periods, fall of 1998 and fall of 2000, all 4 sites had simultaneous outbreaks of cholera. The seasonal patterns of cholera were similar, although not identical, in all areas.
fig.ommitted
Figure 2. Number of patients with acute watery diarrhea, including cholera, during the period of surveillance. Surveillance in Matlab and Chhatak began in March 1997; surveillance in Bakerganj and Chaugachha began in June 1997.
The age distribution of patients, divided according to whether they had cholera diarrhea or noncholera diarrhea, is shown in . Most of the patients were children; 53% were <5 years of age. The age distribution of patients with cholera was significantly different (P < .01) from that of patients without cholera. Noncholera diarrhea occurred primarily in very young patients, whereas cholera diarrhea occurred more frequently in older children and adults.
fig.ommitted
Figure 3. Age of patients and occurrence of diarrhea associated with cholera (due to either V. cholerae O1 or V. cholerae O139) and of diarrhea not associated with cholera (n = 5670). The age distributions of the 2 groups are significantly different (P < .01).
The age distribution of patients with cholera serogroup O1 was significantly different (P < .01) from that of patients with cholera serogroup O139, as shown in . V. cholerae O1 occurred in a significantly greater number of younger patients, whereas V. cholerae O139 occurred primarily in older patients, who are presumed to have little immunity to this relatively new organism. Similar age differences were seen in all of the 4 surveillance areas. These differences in age distribution are similar to those previously reported for hospitalized patients in Bangladesh [2].
fig.ommitted
Figure 4. Age of patients with cholera caused by either Vibrio cholerae O1 or V. cholerae O139 (n = 808). The age distributions of the 2 groups are significantly different (P < .01).
Associations between Cases of Cholera and Water Use
Logistic regressions with cholera as the outcome and with water-source and water-use variables as the risk factors were performed for each study area, with adjustment for age. Bathing with tube-well water had a statistically significant protective effect in Bakerganj: the age-adjusted odds of cholera for patients who exclusively used tube-well water for bathing were 0.4 times as large (95% confidence interval 0.2–
0.8) as those for patients who used pond and/or river and/or canal water . In Matlab, washing of clothes and utensils with tube-well water was protective. The age-adjusted odds of cholera for people who used tube-well water exclusively for washing of clothes and utensils were 0.5 times as large (95% confidence interval 0.3–
0.9) as those for people who used a different water source.
fig.ommitted
Table 2. Significant Associations between Water Source and Occurrence of Cholera
The potential protective effect of drinking tube-well water could not be assessed, because very few people reported using any other source of drinking water.
DISCUSSION
This study is the first to investigate the epidemiology of cholera in rural Bangladesh, outside Matlab. It shows cholera to be widely distributed and present in all 4 areas studied, including the designated control area. Each area had a different pattern of cholera outbreaks, although there were clear similarities between the 4 areas.
In all of the areas, most of the cholera outbreaks occurred during either the pre- or the post-monsoon season. (The monsoon season is approximately June–
September, throughout the country). The fall peak, however, appeared to be more predictable in 2 areas, Chhatak and Matlab, whereas the spring peak was more common in Bakerganj.
Matlab clearly had the largest number of cases of cholera, and these occurred throughout the year, except for the year 2000, when cases of cholera were very few. Cases of cholera also decreased during this period in Bakerganj and Chaugachha; only Chhatak had a large outbreak during the fall of 2000.
V. cholerae O1 and V. cholerae O139 were found in all surveillance sites, although V. cholerae O1 was more frequently seen. On several occasions, however, V. cholerae O139 clearly predominated, and, on one occasion, it was the sole organism causing the epidemic (June 1999 in Chhatak). It should be noted that V. cholerae O139 has been isolated with V. cholerae O1 in India [3] and that, for short periods of time, each serogroup has been temporarily displaced by the other.
It was of considerable interest that, on two occasions (fall of 1998 and fall of 2000), all 4 sites had simultaneous outbreaks. This suggests that the synchrony in these widely scattered areas was determined by common factors, most likely environmental.
It was noted that there were many periods of "epidemic" diarrhea, when the number of patients with cholera was extremely high, suggesting that this organism caused most of the illness; but there also were outbreaks of diarrhea in which few or no cases of cholera were seen (April 1998 and June 2000 in Chattak; December 1999 in Chaugachha), suggesting that other etiologic agents were responsible. In this study, we identified only V. cholerae and, therefore, can only speculate on what caused these outbreaks of noncholera diarrhea. Because, in Matlab and Dhaka, as noted in a previous study [27], rotavirus and enterotoxigenic Escherichia coli are, respectively, the most frequent viral and bacterial causes of acute diarrhea in children, they are the most likely responsible agents.
The patients seen in this study were primarily children <5 years of age. When the ages of patients with cholera were compared with those of patients without cholera, there were marked differences, as would be expected. In children <2 years of age, noncholera diarrhea occurred much more often than did cholera diarrhea. In the older children and young adults, however, cholera diarrhea was seen more frequently than noncholera diarrhea. This age-related phenomenon may suggest a severity of the illness for which this age group sought treatment. In previous studies in Matlab, in adults, enterotoxigenic E. coli was found to be more than twice as frequent as cholera [28].
There were also clear age differences when V. cholerae O1 and V. cholerae O139 were compared: V. cholerae O1 occurred more frequently in the young children, whereas V. cholerae O139 was seen in all age groups, but more frequently in older adults. These findings are compatible with earlier studies that showed that, in a nonimmune population, V. cholerae O139 occurred with similar attack rates at all ages [2]. These data suggest that V. cholerae O139 has not yet become endemic in those areas where V. cholerae O1 has been heavily endemic for many years.
There were significant associations of cholera and water sources used for bathing and washing. Since nearly all patients used tube wells for drinking water, we could not determine the protective effect that this practice may have had. Other studies have had difficulty in showing a protective effect of tube wells [29, 30], because other sources of drinking water also were used. In the present study, the use of tube-well water for bathing and washing, however, was shown to be protective against cholera in Bakerganj and Matlab, respectively. It should be pointed out that these comparisons were made between patients with cholera diarrhea and those with noncholera diarrhea; there was no control group without diarrhea. Since our control group consisted of patients with diarrhea who also may have had some similar risk factors, the differences seen here are even more significant.
The clinical surveillance of cholera in 4 widely separated areas of Bangladesh provides a basis for determining which environmental factors may play a role in defining the outbreaks of cholera. The clinical observations in this study indicate that cholera is both seasonal in all areas, sometimes being synchronous in all areas, and is related to water use in some areas, and all suggest that environmental factors play a significant role. Clearly, V. cholerae O139 is widespread in Bangladesh and, although not yet endemic, appears to be competing with the O1 serogroup for an environmental niche. In a separate publication, these clinical surveillance data will be correlated with the environmental data collected from the same geographic areas to determine which factors are associated with cholera outbreaks and to develop a model that can be used to predict outbreaks, thus facilitating the early introduction of preventative and treatment measures.
Acknowledgments
This work was supported by the International Centre for Diarrheal Disease Research, Bangladesh (ICDDR,B). The ICDDR,B is supported by countries and agencies that share its concern for health problems of developing countries. We would like to thank the physicians and staff at the ICDDR,B for their dedicated work in supporting this project.
References
1.
Barua D. History of cholera. In: Barua D, Greenough WB, eds. Cholera. New York: Plenum Medical Book, 1992:1–
5.
2.
Cholera Working Group, International Centre for Diarrhoeal Diseases Research, Bangladesh. Large epidemic of cholera-like disease in Bangladesh caused by Vibrio cholerae O139 synonym Bengal. Lancet 1993; 342:387–
90.
3.
Nair GB, Ramamurthy T, Bhattacharya SK, et al. Spread of Vibrio cholerae O139 Bengal in India. J Infect Dis 1994; 169:1029–
34.
4.
Ramamurthy T, Garg S, Sharma R, et al. Emergence of novel strain of Vibrio cholerae with epidemic potential in southern and eastern India. Lancet 1993; 341:703–
4.
5.
Craig JP. Cholera: outlook for the twenty-first century. Caduceus 1996; 12:25–
42.
6.
Ward HB, Whipple CC. In: Freshwater biology. 2nd ed. New York and London: John Wiley & Sons, 1959.
7.
Tison DL. Vibrios. In: Manual of clinical microbiology. 7th ed. Murray PR, Baron EJ, Pfaller MA, Tenover FC, Yolken RH, eds. Washington DC: American Society for Microbiology, 1999:497–
506.
8.
Kogure K, Simidu U, Taga N. A tentative direct microscopic method for counting living marine bacteria. Can J Microbiol 1979; 25:415–
20.
9.
Huq A, Colwell RR, Rahman R, et al. Detection of Vibrio cholerae 01 in the aquatic environment by fluorescent monoclonal antibody and culture methods. Appl Environ Microbiol 1990; 56:2370–
3.
10.
Huq A, Parveen S, Qadri F, Sack DA. Comparison of Vibrio cholerae serotype 01 strains isolated from patients and the aquatic environment. J Trop Med Hyg 1993; 96:86–
92.
11.
Islam MS, Drasar BS, Albert MJ, Sack RB, Huq A, Colwell RR. Toxigenic Vibrio cholerae in the environment: a minireview. Trop Dis Bull 1997; 94:R1–
R11.
12.
Colwell RR, Huq A. Vibrios in the environment: viable but nonculturable Vibrio cholerae. In: Vibrio cholerae and cholera: molecular to global perspectives. Wachsmuth IK, Blake PA, Olsvik O, eds. Washington, DC: American Society for Microbiology, 1994:117–
33.
13.
Islam MS, Drassar BS, Sack RB. Ecology of Vibrio cholerae: role of aquatic fauna and flora. In: Cholera and the Ecology of Vibrio cholerae. Drassar BS, Forrest BD, eds. London: Chapman & Hall, 1996.
14.
Islam MS, Miah MA, Hasan MK, Sack RB, Albert MJ. Detection of non-culturable Vibrio cholerae 01 associated with a cyanobacterium from an aquatic environment in Bangladesh. Trans R Soc Trop Med Hyg 1994; 88:298–
9.
15.
Pollitzer R. Cholera. Geneva: World Health Organization, 1959.
16.
Koch R. An address on cholera and its bacillus. Br Med J 1884; 2:403–
7, 453–
9.
17.
Cockburn TA, Cassanos JG. Epidemiology of endemic cholera. Pub Health Rep 1960; 75:791–
804.
18.
Gunnlaugsson G, Angulo FJ, Einarsdottir J, Passa A, Tauxe RV. Epidemic cholera in Guinea-Bissau: the challenge of preventing deaths in rural West Africa. Int J Infect Dis 2000; 4:8–
13.
19.
La Raja M. Cholera-like diarrhoeal disease and rivers in rural Mozambique. Trop Doct 2000; 30:53–
4.
20.
Germani Y, Quilici ML, Glaziou P, Mattera D, Morvan J, Fournier JM. Emergence of cholera in the Central African Republic. Eur J Clin Microbiol Infect Dis 1998; 17:888–
90.
21.
Glass RI, Becker S, Huq ML, et al. Endemic cholera in rural Bangladesh, 1966–
1980. Am J Epidemiol 1982; 116:959–
70.
22.
Samadi AR, Chowdhury MK, Huq MI, Khan MU. Seasonality of classical and El Tor cholera in Dhaka, Bangladesh: 17-year trends. Trans R Soc Trop Med Hyg 1983; 77:853–
6.
23.
Siddique AK, Zaman K, Baqui AH, et al. Cholera epidemics in Bangladesh: 1985-1991. J Diarrhoeal Dis Res 1992; 10:79–
86.
24.
Longini IM, Yunus M, Zaman K, Siddique AK, Sack RB, Nizam A. Epidemic and endemic cholera trends over a 33-year period in Bangladesh. J Infect Dis 2002; 186:246–
51.
25.
World Health Organization Guidelines for the laboratory diagnosis of cholera, Geneva: WHO Bacterial Disease Unit, 1974.
26.
Kleinbaum DG, Kupper LL, Muller KE, Nizam A. Applied regression analysis and multivariable methods. Pacific Grove, CA: Duxbury Press, 1988.
27.
Albert MJ, Garuque ASG, Faruque SM, Sack RB , Mahalanabis D. Case-control study of enteropathogens associated with childhood diarrhea in Dhaka, Bangladesh. J Clin Microbiol 1999; 37:3458–
64.
28.
Black RE, Merson, MH, Rahman ASMM, et al. A two-year study of bacterial, viral, and parasitic agents associated with diarrhea in rural Bangladesh. J Infect Dis 1980; 142:660–
4.
29.
Sommer A, Woodward WE. The influence of protected water supplies on the spread of classical Inaba and El Tor Ogawa in rural East Bengal. Lancet 1972; 2:985–
7.
30.
Levine RJ, Khan MR, D'Souza S, et al. Failure of sanitary wells to protect against cholera and other diarrhoeas in Bangladesh, 1976. Lancet 1976; 2:869. |
(R. Bradley Sack A. Kasem Siddique Ira M. Longini Jr. Azhar Nizam Md. Yunus M. Sirajul Islam J. Glenn)
|
