Flood exposure and pregnancy loss in 33 developing countries

Floods, exacerbated by climate change, are a major global health concern. While immediate impacts like injuries and disease are well-documented, the indirect effects on vulnerable populations, especially pregnant women in developing countries, remain understudied. This study addresses this gap by investigating the association between flood exposure during pregnancy and pregnancy loss. The heightened vulnerability of pregnant women in developing countries stems from factors such as limited healthcare access, pre-existing health conditions, and unsafe living conditions that are further strained by flood events. Unsafe labor and delivery conditions, exacerbated scarcity of essential resources, and long-term consequences like mental health issues and socioeconomic hardship compound the risks. Prior research on this topic is limited by small sample sizes, focused on isolated incidents or specific regions. This study aims to provide a more comprehensive understanding by utilizing a large multi-region dataset and sophisticated statistical methods to analyze the relationship between gestational flood exposure and pregnancy loss across 33 developing countries and quantify the resulting excess pregnancy losses.

Existing literature highlights the devastating immediate consequences of floods, including injuries, psychological stress, and infectious diseases. Studies have also shown the indirect health impacts of floods on ecological determinants like access to clean drinking water and food security. Previous research has indicated a link between flood exposure and various health problems in the general population such as psychological disorders, infectious diseases, and gastrointestinal illnesses. However, there's a significant knowledge gap regarding the effects of flood exposure on specific vulnerable groups, particularly pregnant women. While some studies have hinted at increased pregnancy loss risk due to maternal flood exposure during pregnancy, these often suffer from limitations in sample size, geographic scope, and methodological rigor, hindering definitive conclusions. The existing literature underscores the need for a large-scale, multi-regional study to address the knowledge gaps and provide reliable evidence regarding the impact of flood exposure on pregnancy outcomes.

This study employed a matched case-control design using data from Demographic and Health Surveys (DHS) across 33 developing countries. The dataset contained individual-level records of pregnancy loss (miscarriages and stillbirths) and successful pregnancies. These data were spatially matched with flood data from the Dartmouth Flood Observatory (DFO) database, using spatial analysis tools to identify gestational flood exposure. The primary analysis used a conditional logistic regression model to examine the association between gestational flood exposure (defined as flooding overlapping with the pregnancy period) and pregnancy loss, adjusting for maternal age, year and month of conception, previous pregnancy losses, and climatic factors (temperature and precipitation). Sensitivity analyses were conducted using an alternative flood database (Global Flood Database), excluding specific regions, and reducing covariates. Subgroup analyses explored effect modification by maternal age, pregnancy stage, socioeconomic factors (wealth and education), and living conditions (water source and floor material). To estimate excess pregnancy losses, the study combined pregnancy data from the WorldPop database with the DFO flood data. A formula was used to calculate excess losses for each country, incorporating relative risk (represented by the odds ratio from the regression models) and the number of flood events and pregnancies in each region. The study also analyzed pre-conception flood exposure, categorized into different time windows before pregnancy.

The study found a statistically significant association between gestational flood exposure and pregnancy loss (OR 1.08, 95% CI: 1.04-1.11). This risk was more pronounced in several subgroups: women outside the peak reproductive age (under 21 or over 35), during mid and late pregnancy, women relying on surface water, those with lower socioeconomic status (lower wealth and education), and those living in houses with rudimentary flooring. Prolonged flood exposure (≥6 days) was associated with a substantially higher risk of pregnancy loss (OR 2.00, 95% CI: 1.83-2.13). The estimated excess pregnancy losses attributable to gestational flood exposure during the 2010s ranged from 53,944 to 148,345 annually across the 33 countries. South Asia had the highest number of excess losses. A significant upward trend in annual excess pregnancy losses was observed from 2010 to 2020. Heavy rainfall and monsoon rains contributed to the largest proportion of flood-related pregnancy losses. Sensitivity analyses generally supported the primary findings.

The study's findings strongly support the hypothesis that gestational flood exposure significantly increases the risk of pregnancy loss in developing countries. This risk is not uniform but is amplified by a variety of factors, including age, pregnancy stage, socioeconomic status, and access to safe water and sanitation. The observed increased risk among women outside peak reproductive age is consistent with known increased vulnerabilities at these life stages, which are further exacerbated by stressful events like floods. The findings highlight the importance of considering indirect health consequences of climate-related disasters in public health planning. The results underscore the need for targeted interventions to protect vulnerable pregnant women in flood-prone regions. Such interventions should include improved access to healthcare, infrastructure development focusing on water and sanitation, and social support systems to reduce the vulnerability of women during and after flood events.

This study provides strong evidence of the significant association between gestational flood exposure and pregnancy loss in 33 developing countries. The findings highlight the disproportionate impact of floods on maternal and child health in vulnerable populations. Future research could explore the underlying biological mechanisms linking flood exposure to pregnancy loss, develop more precise risk prediction models, and evaluate the effectiveness of targeted interventions to mitigate these risks. Furthermore, research is needed to better understand the long-term consequences of flood exposure on the health of mothers and children.

The study acknowledges several limitations. The spatial data from the DHS clusters were scrambled to protect privacy, potentially introducing some inaccuracy in determining flood exposure. Residual confounding may exist from unmeasured maternal factors that vary over time. The analysis used uniform odds ratios across countries in the excess pregnancy loss estimation due to insufficient sample size for country-specific estimations. The assumption of constant annual pregnancies in each grid from 2010 to 2020 might also affect the accuracy of the estimates.