LAUSR

OBJECTIVE To describe changes in maternal viral control over time in South African women living with HIV (WLHIV) using surveillance data from the National Health Laboratory Service's Corporate Data Warehouse (NHLS CDW). DESIGN A retrospective cohort analysis of maternal viral load during pregnancy and up to 15 months postpartum was performed amongst WLHIV (15-49 years) within the public-health sector between 2016 and 2017. METHODS HIV and pregnancy-related test data were used to create a synthetic cohort of pregnant WLHIV from the NHLS CDW. Syphilis-screening, in association with ward type and/or postpregnancy cervical screening and/or birth HIV test and/or positive β-hCG, was used as a proxy for pregnancy. The syphilis-screening date marked the first antenatal care visit (fANC). Fractional polynomial models described viral load evolution from fANC up to 15 months postdelivery. Piecewise linear regression models determined factors associated with viral load decline. FINDINGS Among 178 319 pregnant WLHIV, 345 174 viral load tests were performed [median = 2 (IQR: 2-3) per woman]. At fANC, 85 545 (48%) women were antiretroviral therapy (ART) experienced; 88 877 (49.8%) were not and 3897 (2.2%) unknown. Proportions of viraemia (viral load ≥50 copies/ml) were 39 756 (53.6%) at first viral load performed during pregnancy, 14 780 (36.9%) at delivery and 24 328 (33.5%) postpartum. Maternal age at least 25 years, CD4+ cell count at least 500 cells/μl and viral load less than 50 copies/ml at baseline predicted sustained viral load suppression during follow-up. CONCLUSION Despite high-ART coverage among pregnant women in South Africa, only 63% of WLHIV achieved viral load less than 50 copies/ml at delivery. Maternal viral load monitoring requires prioritization for maternal health and eMTCT.