Non-Malarial Fevers and Antibacterial Use in a Cohort of Ugandan Pregnant Women Enrolled in a Malaria Chemoprevention Trial: A Causal Mediation Analysis

Background
Intermittent preventive treatment of malaria in pregnancy (IPTp) with dihydroartemisinin-piperaquine has been shown to be more effective in reducing the burden of malaria in pregnancy compared to sulfadoxine-pyrimethamine, the current standard of care in sub- Saharan Africa. However, IPTp-DP has not translated into better birth outcomes compared with IPTp-SP in malaria-endemic settings. We aimed to determine whether non-malarial pathways impacted by IPTp regimens influence birth outcomes.

Methods
We leveraged participant-level data from 654 HIV-uninfected pregnant women living in Busia District, Uganda, and enrolled in a randomized clinical trial comparing IPTp with monthly sulfadoxine-pyrimethamine or monthly dihydroartemisinin-piperaquine. Women were enrolled at 12-16 weeks of gestational age and followed through delivery. We quantified the effect of our exposure (randomized IPTp arm) and mediators (incident non-malarial febrile illnesses (NMFIs) and antibacterial prescriptions excluding and including the number of doses of sulfadoxine-pyrimethamine) on the primary outcomes of birthweight and birthweight-for-gestational-age Z-scores (both continuous measures). We conducted causal mediation analyses to decompose the total effect of IPTp regimens on birth outcomes into the effect of IPTp on birth outcomes that is mediated by the number of NMFIs or the number of antibacterial prescriptions (indirect effect) and the effect of IPTp on birth outcomes that cannot be explained by the same mediators (direct effect), accounting for confounding.

Findings
Mean birthweight-for-gestational-age Z-scores were significantly higher among neonates of women randomly assigned to sulfadoxine-pyrimethamine compared to those of women randomly assigned to dihydroartemisinin-piperaquine (mean difference 0.19, 95% CI 0.04 to 0.33). Women in the dihydroartemisinin-piperaquine group had more incident NMFIs (IRR 1.34, 95% CI 1.05 to 1.72) and were exposed to fewer antibacterials in the form of study drug doses and clinician prescriptions as compared to women in the sulfadoxine-pyrimethamine group (incidence rate ratio [IRR] 0.25, 95% CI 0.23 to 0.27). Mediation analyses with incident NMFIs as the mediator showed that sulfadoxine-pyrimethamine conferred a greater direct effect on birthweight-for-gestational-age Z-scores and the risk for having a small-for-gestational-age baby than did dihydroartemisinin-piperaquine (mean difference 0.20, 95% CI 0.07 to 0.34 and risk difference 5.95%, 95% CI 0.14% to 11.57%, respectively). Mediation analyses with antibacterial prescriptions excluding the number of doses of sulfadoxine-pyrimethamine as the mediator showed similar effects. However, mediation analyses with antibacterial prescriptions including the number of doses of sulfadoxine-pyrimethamine as the mediator showed that sulfadoxine-pyrimethamine conferred a greater indirect effect on birthweight-for-gestational-age Z-scores (mean difference 0.18, 95% CI 0.04 to 0.32) and birthweight (mean difference 74 g, 95% CI 15 g to 133 g).

Interpretation
Improving birthweight-for-gestational-age Z-scores and neonatal birthweight seems to arise from the combined effect of antibacterial prescriptions and monthly sulfadoxine-pyrimethamine. Future analyses and studies are needed to understand the exact mechanisms by which sulfadoxine-pyrimethamine improve birth outcomes in malaria-endemic settings.