Long-term effects of cooking with liquefied petroleum gas or biomass on linear growth trajectories from birth to the pre-school years in Puno, Peru: a prospective cohort study

Background Household air pollution (HAP) is a major global health risk. Observational studies link HAP exposure to impaired child growth, but randomized controlled trial (RCT) evidence is inconsistent. Methods We followed children born during an RCT of an 18-month liquefied petroleum gas (LPG) intervention among 800 pregnant women in Puno, Peru. We measured personal exposures to fine particulate matter (PM 2.5 ) and carbon monoxide (CO) three times during pregnancy and three times during infancy. We measured length quarterly between birth and 12 months and height once between age 2–4 years. We assessed the effect of the LPG intervention on growth trajectories and evaluated exposure-response associations between height-for-age z-score (HAZ) and PM 2.5 or CO exposures. Findings We revisited 683 children (mean age 34.0 ± 6.6 months, 49.3% male, 52.3% intervention). Mean HAZ at age 2–4 years was −0.92 ± 0.83 SDs in intervention children and −1.00 ± 0.80 SDs in controls (p = 0.33). In intention-to-treat analysis, the HAZ difference between groups was 0.08 SDs (95% CI −0.04 to 0.21) favoring the intervention. Neither prenatal nor postnatal PM 2.5 or CO exposures were associated with HAZ. A 10 μg/m 3 difference in prenatal and postnatal PM 2.5 corresponded to a HAZ difference of −0.003 SDs (−0.011 to 0.005) and −0.001 SDs (−0.005 to 0.007), respectively. A 1 ppm difference in prenatal or postnatal CO corresponded to −0.009 SDs (−0.025 to 0.008) and 0.000 (−0.011 to 0.012), respectively. Interpretation Children of mothers randomized to LPG were not taller than controls. Personal PM 2.5 or CO exposures did not influence child growth. Funding US National Institutes of Health; Bill & Melinda Gates Foundation.