Clinical, epidemiological, and spatial characteristics of Vibrio parahaemolyticus diarrhea and cholera in the urban slums of Kolkata, India

The study was conducted in urban slum communities in Kolkata, the capital of the state of West Bengal. Kolkata, the third largest city in India, has 14 million inhabitants living within an area of 1,450 km², making it one of the world’s most densely populated cities. The Kolkata Municipal Corporation consists of 141 civic administrative units called wards, with each ward having an office responsible for public health supervised by a medical officer. The study site comprises three contiguous wards (29, 30, and 33) with about 100,000 residents. These residents live in homes tightly-spaced together along winding sewage-littered pathways, and they rely on shared toilets and drinking-water [10]. The area has a high population density. Through the years, extensive subletting has resulted in overcrowding as more and more people are squeezed into available housing [11]. Sufficient water supply and sanitary facilities are unavailable in the area. Several households share one or two latrines and water taps. Most sewage is collected in open drainage gutters which tend to overflow during the rainy season, flooding adjacent homes. Kolkata has three seasons, the cool dry months from November to February, the hot dry period from March to May, and the monsoon season from June to October. Seventy percent of the people are Hindus, and the rest are predominately Muslims.

This study used the data of a cluster-randomized, double-blind, placebo-controlled trial of a killed oral cholera vaccine [12]. The clusters were dwellings, which were randomly assigned to receive either vaccine or placebo, so that individuals living in the same dwelling (cluster) received the same agent (vaccine or placebo). A dwelling was defined as a hut, a group of huts, or a multistory building with several households using shared water pipes, bathrooms, and latrines as assigned by the Kolkata Municipal Corporation. There were 3933 clusters of which 1966 clusters were assigned to receive vaccine and 1967 clusters were assigned to receive placebo. Residents were eligible to receive a study intervention if they were aged one year or older and were not pregnant.

Nine clinics were established in the community to undertake diarrhea surveillance in addition to the two referral hospitals (Infectious Diseases Hospital and B.C. Roy Children’s Hospital). Private medical practitioners were encouraged to refer patients with diarrhea to the clinics. Patients from the study area were identified by use of household identification cards and a computerized database. Study physicians recorded pertinent clinical details on a structured clinical data form. Rectal swabs were obtained from all patients presenting with history of loose stools and transported in Cary-Blair media to a laboratory at the National Institute of Cholera and Enteric Diseases (NICED) within 8 h of specimen collection. At the laboratory, rectal swabs were examined for V. parahaemolyticus and V. cholerae by use of conventional methods [13].

The disease surveillance data of four years from January 1, 2007 to December 31, 2010 were used in this analysis. A diarrheal visit was defined as a visit by a patient who had, in the 24 h before presentation, three or more loose or liquid stools; or at least one loose or liquid stool with blood; or, if one to two or an indeterminate number of loose or liquid stools were reported, at least some evidence of dehydration, according to WHO criteria [14]. The onset of a diarrheal visit was the day on which the patient first experienced loose or liquid stools. Diarrheal visits for which the date of onset was less than or equal to 7 days from the date of discharge for the previous visit were grouped into the same diarrheal episode. A cholera episode was one in which V. cholerae O1 or O139 was isolated from a fecal specimen during any component diarrheal visit; an episode of V. parahaemolyticus diarrhea was one in which V. parahaemolyticus was isolated.

To detect potential geographic areas of high risk for the diseases, the spatial scan test has been widely used in recent times [15‐18]. The spatial scan test is also suitable for uneven geographic distribution of cases and population density [19]. We used the spatial scan test implemented through SaTScan® [16] to identify unique non-random spatial clusters that are higher risk for cholera and for V. parahaemolyticus. We assumed that the incidences of two diseases followed a Poisson distribution. Under the null hypothesis, the incidence of disease in a particular location is proportional to the number of residents in that location [19]. Using SaTScan®, we estimated the probability that the frequency of disease at each peak surpasses that expected by chance. We set the space limitations to 50% population at risk, which allowed us to scan for both large and small clusters of disease risk. We took into account the observed number of cases inside and outside the circle when calculating the highest likelihood for each circle. This circle was the most probable cluster and had a rate that was the least likely to happen by chance alone. The statistical significance of possible clusters was calculated using 999 Monte Carlo simulations [20]. Purely spatial analysis was performed using circular windows. The output from SaTScan® was imported into ArcGIS (Version 9.2, California, USA) to map significant (p<0.01) clusters of higher risk.

This is an observational population based study in which the population was classified in clusters of residential dwelling units. The study participants were geographically identified based on their residence, and this spatial component was included in the study. Several demographic and socioeconomic variables that were thought to be independently associated with the risk for the diseases were evaluated in the study. However, this is an exploratory analysis of a secondary data source. To avoid distortions due to the impact of the cholera vaccination on the outcomes, we limited our analysis to residential dwellings assigned to placebo. Comparison of individual characteristics was performed using generalized estimating equations (GEE) with logit link function adjusting for the design effect of the clusters used to allocated subjects to the two agents in the randomized trial [21]. To analyze the risk factors of V. parahaemolyticus diarrhea or cholera, we used survival analyses of the time to the first episode of the disease, censoring the follow-up of individuals who died or migrated out [22]. We fitted unadjusted and covariate adjusted Cox proportional hazard regression models verifying first that the proportionality assumption was satisfied for all independent variables [23‐25]. In both covariate-unadjusted and covariate-adjusted analyses of the risk for V. parahaemolyticus diarrhea or cholera, we accounted for the design effect induced by cluster randomization by use of robust sandwich variance estimates [25]. These estimates enabled inferences about risk of the disease at the individual level, adjusting for the design effect. Final adjusted risk estimates were obtained from the model significant at p<0·05 in a forward selection algorithm.

The study received approval from the Health Ministry Screening Committee of the Government of India, Scientific Advisory Committee and Institutional Ethics Committee of NICED, Institutional Review board of International Vaccine Institute and also from the Secretariat Committee for Research Involving Human Subjects, World Health Organization Geneva, Switzerland.