Placental malaria, air pollution, and birth outcomes in a cohort of infants in Ghana

Abstract

Background: Household air pollution (HAP) exposure and malarial infection during pregnancy have both been associated with adverse birth outcomes. Previous studies on the impact of HAP on birth outcomes have not assessed the potential modifying effect of placental malaria (PMal).Methods: Placental tissue from mothers enrolled in the Ghana Randomized Air Pollution and Health Study (GRAPHS) was tested for evidence of malarial infection using histopathology. Associations between carbon monoxide (CO) exposure during pregnancy and birth outcomes were assessed using multivariable regression models with statistical interaction terms to assess effect modification by PMal.Results: 286 of 1170 placental samples (23.2%) showed evidence of PMal. CO was not independently associated with PMal (OR 0.94 [95% CI: 0.81, 1.11]) but PMal was associated with reduced birth weight (-175.0g [95% CI: -235.4, -114.6]), birth length (-0.7cm [-1.2, -0.2], head circumference (-0.6cm [-0.9, -0.2], and gestational age (-3.2 days [-4.9, -1.5]) in unadjusted analyses. An interaction (p-value 0.03) was observed between CO and PMal on birth weight. In the malaria-negative group, a 1ppm increase in prenatal CO was associated with -62.7g lower infant weight at birth [95% CI: -95.8, -29.6] adjusting for infant sex and maternal characteristics. No significant association was seen among mothers testing positive for PMal (1ppm increase in CO associated with 10.7g [95% CI -48.5, 70.0] difference in birth weight). Similar but non-significant interactions by PMal were observed for birth length, head circumference, and gestational age.Conclusions: PMal is a known risk factor for reduced infant size at birth, and demonstrated in our cohort. The negative impact of malaria during pregnancy may overwhelm the effect of HAP on birth weight and other infant anthropometrics at birth. In settings where malaria is endemic, associations between air pollution and birth outcomes may be masked or attenuated if the effect of malaria is not assessed.