Drinking Water Salinity and Maternal Health in Coastal Bangladesh: Implications of Climate Change


Background: Drinking water from natural sources in coastal Bangladesh has become contaminated by varying degrees of salinity due to saltwater intrusion from rising sea levels, cyclone and storm surges, and upstream withdrawal of freshwater.

Objective: Our objective was to estimate salt intake from drinking water sources and examine environmental factors that may explain a seasonal excess of hypertension in pregnancy.

Methods: Water salinity data (1998–2000) for Dacope, in rural coastal Bangladesh, were obtained from the Centre for Environment and Geographic Information System in Bangladesh. Information on drinking water sources, 24-hr urine samples, and blood pressure was obtained from 343 pregnant Dacope women during the dry season (October 2009 through March 2010). The hospital-based prevalence of hypertension in pregnancy was determined for 969 pregnant women (July 2008 through March 2010).

Results: Average estimated sodium intakes from drinking water ranged from 5 to 16 g/day in the dry season, compared with 0.6–1.2 g/day in the rainy season. Average daily sodium excretion in urine was 3.4 g/day (range, 0.4–7.7 g/day). Women who drank shallow tube-well water were more likely to have urine sodium > 100 mmol/day than women who drank rainwater [odds ratio (OR) = 2.05; 95% confidence interval (CI), 1.11–3.80]. The annual hospital prevalence of hypertension in pregnancy was higher in the dry season (OR = 12.2%; 95% CI, 9.5–14.8) than in the rainy season (OR = 5.1%; 95% CI, 2.91–7.26).

Conclusions: The estimated salt intake from drinking water in this population exceeded recommended limits. The problem of saline intrusion into drinking water has multiple causes and is likely to be exacerbated by climate change–induced sea-level rise.

Keywords: climate change, hypertension, maternal health, pregnancy, salinity intrusion

Although scientists do not know with certainty the full extent of the effects of climate change, there is a growing concern that one of the first and most critical impacts will be on the world’s freshwater resources. In coastal Bangladesh, natural drinking water sources, such as rivers and groundwater, are threatened by saltwater intrusion from the Bay of Bengal. According to the Intergovernmental Panel on Climate Change (IPCC), groundwater, crop soils, and many rivers are likely to become increasingly saline from higher tidal waves and storm surges, as a result of climate change impacts (Parry et al. 2007).

The coastal area of Bangladesh is a part of the flat Ganges Delta, which is intersected by large tidal rivers discharging into the Bay of Bengal. The saline front along the 720-km coastline has encroached > 100 km inland into domestic ponds, groundwater supplies, and agricultural land through various estuaries and water inlets, which are interlinked with the major rivers (Allison et al. 2003; Rahman and Bhattacharya 2006). Levels of water salinity have a clear seasonal pattern (Rahman and Ravenscroft 2003) due to rainfall patterns and upstream withdrawal of freshwater (owing to the operation of the Farakka Barrage, which the Indian government uses to regulate flow on the Ganges) during the drier months. Since 1948, river salinity in the southern districts of Patuakhali, Pirojpur, Barguna, Satkhira, Bagerhat, and Khulna has risen by 45% (Integrated Regional Information Networks 2007). Salinity intrusion is likely to increase in the future because of further reduced river flows, increased upstream withdrawal, and longer term climate change–induced decreases in dry season rainfall and sea-level rise.

The coastal population of Bangladesh relies heavily on rivers, tube wells (groundwater), and ponds for washing, bathing, and obtaining drinking water. Domestic ponds, which take up 10% of the total land area (excluding rice paddies), are primarily rain fed but can also mix with saline water from rivers, soil runoff, and shallow groundwater (Rahman and Ravenscroft 2003). Approximately 20 million people living along the coast are affected by varying degrees of salinity in drinking water obtained from various natural sources [Ministry of Environment and Forest (MOEF) 2006].

Guidelines for dietary salt intake have been established by the World Health Organization (WHO), but no guidelines have been released for safe salinity levels in drinking water, except that sodium levels > 0.2 g/L are unacceptable to taste (WHO 2008). High salinity levels in drinking water may have numerous direct and indirect impacts on health. In 2002 the WHO recognized health impacts of consumption of highly saline waters as a priority for investigation under its public health initiatives (WHO 2003). In a survey conducted in 2008, higher rates of (pre)eclampsia and gestational hypertension in pregnant women living in the southwestern coast of Bangladesh, compared with noncoastal pregnant women, were hypothesized to be caused by saline contamination of drinking water (Khan et al. 2008).

In this descriptive study, we estimated salt intake of the population in Dacope, situated in southwestern coastal Bangladesh, and investigated the potential role of salinity in explaining the unusual seasonal pattern of hypertension in pregnancy among the same population.

This study was approved by the Bangladesh Medical Research Council. The research was conducted in accordance with the principles of the Declaration of Helsinki, and it complied with all relevant national, state, and local regulations.

Materials and Methods

Study area and populations. The area included in our study is Dacope Upazilla (subdistrict), situated under the Khulna district in Bangladesh’s southwest coastal region. Dacope is divided into nine administrative unions that comprise 107 villages, with a total population of 157,500 people. It is intersected by a river network, the Passur River being the largest. This region’s rivers are tidal (ranging between 2 and 4.5 m), with semidiurnal, fortnightly, and seasonal variation in water levels.

Salinity in drinking water: indirect estimates. The main sources of drinking water in the area include shallow and deep tube wells, ponds, rivers, and rainwater. Water may or may not be filtered before drinking. Rainwater, collected from roofs and stored in large numbers of relatively small containers, is assumed to have negligible levels of salinity.

Indirect estimates of individual salinity intake from groundwater and river water were determined based on salinity data for 1998–2000 from the Centre for Environment and Geographic Information System (CEGIS) in Bangladesh. These data included monthly measurements of salinity in shallow and deep groundwater tube wells at various sites in the Khulna region and in the Passur River. All measurements were converted from decisiemens per meter, a measure of electrical conductivity, into an equivalent concentration of parts per thousand (ppt), assuming that 1 ppt is approximately equivalent to 0.64 dS/m (Ayers and Westcot 1985). We used spatially and temporally (monthly) averaged river and shallow tube-well salinities to estimate average levels of salt consumption from river water and shallow groundwater, respectively. For our estimates, we assumed a conservative water intake of 2 L/day per person.

Sample of pregnant women and measurement of urinary sodium. A network of health assistants, under the supervision of the Upazilla Health Complex (UHC) in Dacope, regularly identify and monitor all pregnant women between the ages of 13 and 45 years, in every village of the subdistrict, with a measurement of blood pressure at gestational week 20. Three hundred forty-three pregnant women, a random sample from the monitored women, were enrolled in the study between October 2009 and March 2010 (dry season). The women were sampled in the context of the pilot phase of a larger study that is being planned, where the sampling frame is based on the census. Participants of the study provided written or oral informed consent before participation in the research. The women were all monitored and interviewed at home, using a questionnaire where they indicated their drinking water sources. Direct estimates of salt water intake were made based on salt excretion measured in 24-hr urine samples collected by the women according to verbal instructions provided by research staff. The importance of the completeness of collection was strongly emphasized. Each woman was asked to record the starting time and the finishing time of urine collection. The total volume of urine was recorded, and concentration of sodium was measured in a laboratory at the International Centre for Diarrhoeal Disease Research in Bangladesh.

Women referred for hypertension in pregnancy. We also analyzed data from the UHC, the only hospital in the subdistrict. Pregnant women are tested for proteinuria (a symptom of preeclampsia) by health workers during routine pregnancy monitoring at week 20 and if they have high blood pressure or edema. Women diagnosed with mild hypertension with or without preeclampsia are treated at home but are referred to the hospital for more severe or refractory conditions. We determined the prevalence of hypertension in pregnancy among 969 pregnant women (13–45 years of age) who visited the UHC for antenatal care or pregnancy-related complications or were referred to the UHC for severe hypertension between July 2008 and March 2010. Data were collected from medical records on patient diagnosis, age, residence (village), and distance from the UHC. Women were classified as having “hypertension in pregnancy” if they were diagnosed with gestational hypertension (systolic blood pressure > 140 mmHg or diastolic blood pressure > 90 mmHg after the 20th week of pregnancy and not before pregnancy), preeclampsia (high blood pressure in pregnancy, with significant proteinuria), and eclampsia (preeclampsia accompanied by convulsions that could not be attributed to other causes).

Cutoff point for urinary sodium and blood pressure. A cutoff point of 100 mmol/day was chosen to classify high salt excretion as an outcome. This threshold was selected based on its association with increased systolic and diastolic blood pressure in the Intersalt study (Elliott et al. 1996; Intersalt Cooperative Research Group 1988). We chose cutoff values for diastolic and systolic blood pressure levels of > 85 mmHg and > 130 mmHg, respectively, because levels higher than those are considered “hypertensive.”

Statistical analysis. Using logistic regression, we estimated odds ratios (ORs) and 95% confidence intervals (CIs) for urinary salt excretion > 100 mmol/day according to water source, and for diastolic blood pressure > 85 mmHg and systolic blood pressure > 130 mmHg according to quartiles of urinary sodium concentration. p-Values for trend were also estimated by logistic regression. We used a linear regression model to estimate associations between urinary sodium (modeled as a continuous variable) and water sources. We estimated the hospital-based prevalence of hypertension in pregnancy by dividing the number of cases by the total number of pregnant women that visited the UHC and estimated the 95% CI based on normal approximation.


Indirect estimates of salt intake from drinking water source and seasonality of salinity levels. The average level of river salinity in Dacope was 8.21 ppt (range, 1.35–12.9 ppt) during the dry season and 0.64 ppt (0.19–3.90 ppt) during the monsoon season; shallow groundwater salinity averaged 2.6 ppt during the dry season (0.4–11.4 ppt) and 0.60 ppt (0.4–3.8 ppt) during the monsoon season (1998–2000 data from various sites in the Khulna region, obtained from CEGIS) (Table 1). Assuming an average daily water consumption of 2 L, estimated salt intake from drinking water was 5–16 g/day during the dry season (depending on the water source) and 1.2 g/day during the monsoon season.

The following two tabs change content below.