Time to seroreversion among HIV exposed infants in Lagos, Nigeria: a prospective cohort study

Introduction The efforts towards an HIV-free generation could be undermined by the burden of vertical transmission (VT) in sub-Saharan Africa, compounded by dwindling donor support, erratic availability of early infant diagnosis supplies, and a fragile health system. Our study evaluated the time to seroreversion and identified associated factors among perinatally HIV exposed infants (HEIs) in Lagos, Nigeria. Methods and materials A prospective cohort involving HEIs and their mothers attending the PMTCT clinic at the Nigerian Institute of Medical Research, Yaba, Lagos. Demographic and clinical characteristics of mother-infant pair, including the infant's age at enrolment, gender, birth weight, place of delivery, mode of delivery, gestational age at delivery, duration of rupture of membrane, ART prophylaxis, mode of feeding, and maternal characteristics (educational status, occupation, type of ART, and recent HIV RNA viral load and CD4 counts) were collected. Infant HIV serology was done at 6–8 weeks, 3 months, 6 months, 9 months, 12 months, 15 months, and 18 months. A rapid test for infants is discontinued on at least two consecutive visits with a negative result. Time to seroreversion was estimated with Kaplan–Meier methods. Associated factors were assessed with a single multivariable Cox proportional hazards model over the full 18-month follow-up, with an interval-censored proportional hazards model as a check and Benjamini-Hochberg correction for multiple testing. Results Of 130 mother-infant pairs screened, 111 were enrolled; 103 infants seroreverted and 8 were censored (7 lost to follow-up, 1 death). The median age at seroreversion was 24 weeks (95% CI 24–32), and the cumulative probability of seroreversion was 20.0% by 12 weeks, 54.6% by 24 weeks and 87.8% by 36 weeks. Mean ( ± SD) age at seroreversion was 21.7 ( ± 11.5) weeks and 27.7 ( ± 11.9) weeks among infants on breast milk substitutes (BMS) and exclusively breastfed infants, respectively ( p = 0.013). In the single multivariable Cox model, BMS feeding was associated with earlier seroreversion (adjusted HR 2.77, 95% CI 1.67–4.61; Benjamini-Hochberg adjusted p < 0.001), and the interval-censored model gave a consistent estimate (adjusted HR 2.22, 95% CI 1.34–3.68). The association did not differ across age periods ( p for heterogeneity = 0.53). Conclusion This study provides valuable insight into the time to seroreversion among HEIs in the era of test-and-treat. The finding can help strengthen diagnostic algorithms, ensure sustainable early detection, and improve paediatric HIV outcomes in Nigeria.