Malaria prevention in pregnancy, birthweight, and neonatal mortality: a meta-analysis of 32 national cross-sectional datasets in Africa

Abstract

Low birthweight is a significant risk factor for neonatal and infant death. A prominent cause of low birthweight is infection with Plasmodium falciparum during pregnancy. Antimalarial intermittent preventive therapy in pregnancy (IPTp) and insecticide-treated mosquito nets (ITNs) significantly reduce the risk of low birthweight in regions of stable malaria transmission. We aimed to assess the effectiveness of malaria prevention in pregnancy (IPTp or ITNs) at preventing low birthweight and neonatal mortality under routine programme conditions in malaria endemic countries of Africa. We used a retrospective birth cohort from national cross-sectional datasets in 25 African countries from 2000-10. We used all available datasets from multiple indicator cluster surveys, demographic and health surveys, malaria indicator surveys, and AIDS indicator surveys that were publically available as of 2011. We tried to limit confounding bias through exact matching on potential confounding factors associated with both exposure to malaria prevention (ITNs or IPTp with sulfadoxine-pyrimethamine) in pregnancy and birth outcomes, including local malaria transmission, neonatal tetanus vaccination, maternal age and education, and household wealth. We used a logistic regression model to test for associations between malaria prevention in pregnancy and low birthweight, and a Poisson model for the outcome of neonatal mortality. Both models incorporated the matched strata as a random effect, while accounting for additional potential confounding factors with fixed effect covariates. We analysed 32 national cross-sectional datasets. Exposure of women in their first or second pregnancy to full malaria prevention with IPTp or ITNs was significantly associated with decreased risk of neonatal mortality (protective efficacy [PE] 18%, 95% CI 4-30; incidence rate ratio [IRR] 0·820, 95% CI 0·698-0·962), compared with newborn babies of mothers with no protection, after exact matching and controlling for potential confounding factors. Compared with women with no protection, exposure of pregnant women during their first two pregnancies to full malaria prevention in pregnancy through IPTp or ITNs was significantly associated with reduced odds of low birthweight (PE 21%, 14-27; IRR 0·792, 0·732-0·857), as measured by a combination of weight and birth size perceived by the mother, after exact matching and controlling for potential confounding factors. Malaria prevention in pregnancy is associated with substantial reductions in neonatal mortality and low birthweight under routine malaria control programme conditions. Malaria control programmes should strive to achieve full protection in pregnant women by both IPTp and ITNs to maximise their benefits. Despite an attempt to mitigate bias and potential confounding by matching women on factors thought to be associated with access to malaria prevention in pregnancy and birth outcomes, some level of confounding bias possibly remains.