Prenatal exposure to particulate matter (PM2.5) from biomass fuel and low birth weight in a Sri Lankan birth cohort

Background Prenatal exposure to particulate matter of size 2.5 μm (PM2.5) emissions from biomass fuel may be associated with low birth weight (LBW) (<2.5 kg). We present results examining the association between PM2.5 and LBW from a birth cohort of 545 pregnant women followed from first trimester until delivery, in Sri Lanka. Methods Exposure to household air pollution (HAP) from biomass smoke was assessed using a detailed questionnaire; two-hour measurements of kitchen PM2.5 were collected in a subset of households (n, 303, 56%). Data from questionnaire and measured PM2.5 were used to estimate two-hour kitchen PM2.5 concentrations in unmeasured households. Fetal growth including weight and gestational age were assessed at birth. Linear and logistic regression were used to evaluate the association between HAP and birth weight. Results About 78% of the households used wood (n, 423). In linear regression models, we found an inverse association between a 10-unit increase in PM2.5 and birth weight (β±SE, −0.03±0.02; p, 0.02) adjusted for covariates. Similarly, households primarily using wood (>50%) were negatively associated with birth weight as compared to LPG users (β, −0.13±0.06; p, 0.03). In logistic regression models, a 10-unit increase in PM2.5 was associated with increased odds for LBW (OR, 1.26; 95%CI, 1.02-1.55; p, 0.04). The odds for LBW were highest among >50% wood users (OR, 2.82; 95%CI, 1.18-6.73; p, 0.01), compared to those using >50% LPG with wood (OR, 1.44; 95%CI, 0.57-3.63) and 100% LPG users (referent). The association between HAP exposure and birth weight/LBW were consistent among term-births (n,486). Discussion The finding of a significant association between prenatal PM2.5 exposure and LBW in a low-middle income country (LMIC) setting where competing risk factors are minimal fills a gap in the body of evidence linking HAP to LBW and underscore need to prevent and reduce of HAP exposure in LMICs.