Impact of antenatal antiretroviral drug exposure on the growth of children who are HIV‑exposed uninfected : the national South African Prevention of Mother to Child Evaluation cohort study

ADDITIONAL FILE 1: BOX 1. Anthropometry data cleaning criteria. FIGURE S1. Directed Acyclic Graph representing the hypothesized relationships, 2012–2014, South Africa. FIGURE S2. Study cohort profile of HIV exposed uninfected infants from 6-weeks to 18-months postpartum, 2012–2014, South Africa. TABLE S1. Proportion of underweight children from 6-weeks to 18-months postpartum by in-utero antiretroviral exposure status, 2012–2014, South Africa. TABLE S2. Proportion of stunted children from 6-weeks to 18-months postpartum by in-utero antiretroviral exposure status, 2012–2014, South Africa. TABLE S3. Frequency of maternal antiretroviral treatment over time by baseline maternal antiretroviral categories, 2012–2014, South Africa ; BACKGROUND : The relationship between in-utero antiretroviral (ARV) drug exposure and child growth needs further study as current data provide mixed messages. We compared postnatal growth in the first 18-months of life between children who are HIV-exposed uninfected (CHEU) with fetal exposure to ARV drugs (prophylaxis or triple-drug therapy (ART)) and CHEU not exposed to ARVs. We also examined other independent predictors of postnatal growth. METHODS : We analysed data from a national prospective cohort study of 2526 CHEU enrolled at 6-weeks and followed up 3-monthly till 18-months postpartum, between October 2012 and September 2014. Infant anthropometry was measured, and weight-for-age (WAZ) and length-for-age (LAZ) Z-scores calculated. Generalized estimation equation models were used to compare Z-scores between groups. RESULTS : Among 2526 CHEU, 617 (24.4%) were exposed to ART since -pregnancy (pre-conception ART), 782 (31.0%) to ART commencing post-conception, 879 (34.8%) to maternal ARV prophylaxis (Azidothymidine (AZT)), and 248 (9.8%) had no ARV exposure. In unadjusted analyses, preterm birth rates were higher among CHEU with no ARV exposure than in other groups. Adjusting for infant age, the mean WAZ profile was lower among CHEU exposed to preconception ART [-0.13 (95% confidence ...