Estimation of pregnancy losses attributable to exposure to ambient fine particles in south Asia: an epidemiological case-control study.

BACKGROUND:Maternal exposure to ambient particulate matter (PM2·5) is associated with pregnancy loss (ie, stillbirth and miscarriage). South Asia has the highest burden of pregnancy loss globally and is one of the most PM2·5 polluted regions in the world. However, knowledge of the relevant exposure-response function for mothers is insufficient. METHODS:In this epidemiological case-control study, we collected data from Demographic and Health Surveys from India, Pakistan, and Bangladesh for the period 1998-2016 for women who reported at least one pregnancy loss and one or more livebirths. We assessed ambient exposure during gestation with satellite-based PM2·5 measurements for the period. To derive the exposure-response function, we did a self-compared case-control study in which each case of pregnancy loss was compared with a successful livebirth control or controls by the same mother. Using the estimated exposure-response function, we quantified pregnancy losses attributable to PM2·5 in the region for the period 2000-16 using a standard risk assessment approach. FINDINGS:We assessed data for 34 197 mothers, who had at least one pregnancy loss and one or more livebirth, of whom 26 282 (76·9%) were from India, 4228 (12·4%) were from Pakistan, and 3687 (10·8%) were from Bangladesh. The 34 197 cases of pregnancy loss were matched with 76 282 livebirth controls. The mean level of PM2·5 exposure was 56·00 μg/m3 (SD 30·82) for cases and 54·57 μg/m3 (31·73) for controls. After adjustment for maternal age, non-linear terms for temperature and humidity, seasonal variation, and long-term trends, each 10 μg/m3 increment in PM2·5 was associated with an odds ratio for pregnancy loss of 1·03 (95% CI 1·02-1·05). According to the estimated non-linear exposure-response function by age and urban residence versus rural residence, for the period 2000-16, an estimated 349 681 (95% CI 152 932-489 493) pregnancy losses per year were attributed to ambient air exposure of more than 40 μg/m3 (the annual PM2·5 standard of India), accounting for 7·1% (95% CI 3·3-31·2) of the total annual pregnancy loss burden in south Asia for this period. However, our estimates could be biased because of the limitations of the data (eg, misclassification of induced and spontaneous pregnancy losses). INTERPRETATION:Our findings add to epidemiological evidence of the association between pregnancy loss and PM2·5. Suboptimal air quality contributes to a considerable fraction of total pregnancy loss in south Asia. Controlling PM2·5 pollution will promote maternal health in south Asia. FUNDING:Chinese Natural Science Foundation and Ministry of Science and Technology of China.