Vol II: The Role of El Niño – Southern Oscillation in Shaping Cholera Disease Dynamics: The Case of Bangladesh

According to the WHO (2011), cholera incidence is underestimated in the Indian subcontinent and Southeast Asia because countries do not report their cholera cases. Although there are no reported cases from Bangladesh, experts estimate that there may be 1 million cases a year in Bangladesh (WHO, 2011). Therefore, WHO (2011) estimates the Burden of disease in Bangladesh is much higher than is reflected among cases reported to the WHO. Cholera estimates can be made, since the International Centre for Diarrheal Disease Research in Bangladesh (ICDDRB) operates hospitals in Matlab and Dhaka Bangladesh and keeps records of cholera patients (Emch, 2010). It is suggested that the number of patients arriving to hospitals in Bangladesh in shock due to cholera is increasing (WHO, 2011). Moreover, there is a large body of literature discussing the role of ENSO events shaping cholera disease dynamics (Patz et al., 2002; Pascaul et al., 2000; Koelle et al., 2005; Cash et al., 2010). Given the quantity and quality of work on the subject in Bangladesh and the fact that the disease is becoming more virulent, this provides an informative case study for how ENSO affects cholera outbreaks over time.

There exists a clear climatic sensitivity to cholera epidemics in Bangladesh. Many researchers suggest a bimodal seasonal pattern to the outbreaks (Cash et al., 2010; Kovats, 2000; Kovats et al., 2003). That is, cholera incidence is at its maximum during fall, following the abatement of summer monsoons, and reaches another maximum in the spring, before summer monsoons begin (Cash et al., 2010). ENSO is emerging as a significant predictor of cholera risk because it is observed that ENSO events alter the traditional bimodal distribution (Patz et al., 2005). Cholera in Bangladesh has varied with climatic fluctuations and SSTs influenced by ENSO phenomena over multi-decadal time

periods (Patz et al., 2005). An association between fall cholera incidence and winter ENSO events has been observed since cholera incidence has been shown to increase following winter El Niño events (Cash et al., 2010). The physical basis for this is that Bangladesh summer rainfall increases as a result of warm winter SST anomalies in the tropical pacific initiated by winter El Niño events. Cholera incidence, consequently, rises through increased flooding and breakdown in sanitation (Cash et. al., 2010). Outbreaks also increase because rainfall improves insoluble iron levels, which advances the survival of V. cholerae in marine ecosystems suggesting that increased rainfall associated with ENSO events may improve conditions for the bacteria and increase food and water contamination (Robo et al., 2005).

The correlation between cholera and ENSO in Bangladesh has become more pronounced in recent years. An analysis of historical data in Bengal from 1890-1940 indicates that effects of El Niño on cholera was historically confined to coastal regions, and may have caused spring epidemics, a shift from the usual seasonal outbreaks (Kovats et al., 2003). Trends have become more pronounced since the 1980s in Bangladesh as the ENSO events have become stronger (Rodo et al., 2002; Pascual et al., 2000) examined the association between cholera and ENSO using 18 year records from Bangladesh. Monthly time series data for cholera incidence between January 1980 and March 1998 and SST anomaly data in the equatorial Pacific was used to create an ENSO index during the same period. A significant correlation was observed between cholera and ENSO, with strong associations happening during more pronounced ENSO fluctuations (Pascual et al., 2000). This is largely due to a relationship between cholera cases in Bangladesh and sea surface temperature in the Bay of Bengal. Copepods bloom in response to warming sea surface temperatures that are associated with ENSO (Patz et al., 2005; Kovats, 2003). However, the possible mechanism by which SST disease transmission increased beyond the norm is not greatly understood.

On the other hand, understanding the role of environmental determinants of cholera requires a more holistic approach that includes a combined analysis of intrinsic host immunity to the disease because, “evidence that climate variability drives these inter-annual cycles has been highly controversial, chiefly because it is difficult to isolate the contribution of environmental forcing while taking into account nonlinear epidemiological dynamics generated by host immunity” (Koelle et al., 2005, p. 696). Koelle et al (2005) consider the interplay of these factors and find there is evidence for the role of climate variability in the transmission of cholera. It is suggested that inter-annual variability of cholera is strongly correlated to ENSO, SSTs in the Bay of Bengal over short time spans, during refractory periods, when population susceptibility is considered low. That is, under suitable climate conditions, such as ENSO, epidemics are likely to occur only if a sufficient proportion of the population is non-immune.

III. Forecasting the risk of a cholera outbreak

ENSO events are becoming easier to predict with dynamical and statistical climate models that are used to indicate the likelihood of typical ENSO temperature and precipitation anomalies. Although, the magnitude of an event has high levels of uncertainty and the severity of impacts is hard to predict (McPhaden et al., 2006). Pascual et al. (2008) attempted to predict cholera outbreaks with ENSO data retrospectively. The results illustrate that ENSO is a key covariate and confirm its interplay with immunity levels. Moreover, it is suggested that when v. cholera is considered in predictive models, a robust early warning system for cholera in endemic regions can be developed (Constantin de Magny et al., 2008). The feasibility of using a model as a forecasting tool is established and it is argued that, regardless of model limitations, valuable information for the development of risk management strategies for climate sensitive diseases can be provided (McPhaden et al., 2006). Therefore, the environmental determinants of cholera epidemics can help provide a foundation to build predictive mechanisms for cholera outbreaks and reduce regional vulnerability.

Early warning systems for enhancing public health measures in Bangladesh and other at risk areas have been proposed (Kovats et al., 2000; Constantin de Magny et al., 2008). Seasonal forecasts are becoming sufficiently useful for health purposes on the seasonal level, because major climate trends can be predicted. They are also useful at the short term level because weather can be more accurately predicted approximately one week in advance (Kovats et al., 2003). Seasonal forecasts indicate areas where the probability of some deviation from the climate mean is increased, such as wet or dry or warmer or cold conditions (Kovats et al., 2003). For operational use, however, seasonal climate trends are not sufficient, as our climate system is chaotic. Therefore local weather conditions are needed to provide early warning of epidemic risk (Patz et al., 2005). In addition, it is argued that seasonal forecasts are more accurate during ENSO events than other times (Kovats, 2000; Kovats et al., 2003). This is likely because it is known that ENSO SSTs, in the second half of the year, persist for two seasons and the evolution can be monitored quite accurately proving useful for decision makers (McPhaden et al., 2006). Early warning system can help limit health impacts because it can influence response times. Figure 4 depicts the time-scale of El Niño early warning systems with both weather and climate.

Given that cholera occurs from ingesting contaminated food or water, it is not always possible to provide adequate treatment in a timely manner once the epidemic has begun (Cash et al., 2010). Therefore, it is suggested that better methods of forecasting cholera risk would be a great benefit to societies that face cholera outbreaks, because medical supplies could be prepared in advance. For example, in 1997, the regional WHO Cholera Surveillance Team was aware of an El Niño related drought that had been forecast for South-East Africa and were able to institute measures to help decrease the severity of the outbreak by preparing health care institutions and increasing supplies (WHO, 1999). Thus, it is suggested that disease surveillance and forecasting systems can be effective in reducing vulnerability to climate extremes as the ability to predict high or low transmission seasons can help target the timing and place of public health interventions.

Discussion