Background.A high rate of congenital cytomegalovirus (CMV) has been documented in human immunodeficiency virus (HIV)-exposed infants in industrialized settings, both in the pre-and post-highly active antiretroviral therapy (HAART) era. Only limited data on the birth prevalence of congenital CMV among infants of HIV-infected women on prenatal antiretroviral (ARV) prophylaxis are available from sub-Saharan Africa, despite a high prevalence of both infections. We evaluated the prevalence of congenital CMV in HIV-exposed infants in the Western Cape, South Africa. Methods.HIV-infected mothers were recruited in the immediate postnatal period at a referral maternity hospital between April and October 2012. Maternal and infant clinical data and newborn saliva swabs were collected. Saliva swabs were assayed by real-time polymerase chain reaction for CMV. Data were analyzed using univariate and multivariate logistic regression analyses to determine specific demographic, maternal, and newborn characteristics associated with congenital CMV. Results.CMV was detected in 22 of 748 newborn saliva swabs (2.9%; 95% confidence interval [CI], 1.9%-4.4%). Overall, 96% of mothers used prenatal ARV prophylaxis (prenatal zidovudine, 43.9%; HAART, 52.1%). Maternal age, gestational age, prematurity (<37 weeks™ gestation), type of ARV prophylaxis, length of ARV prophylaxis, birth weight, small for gestational age, and infant feeding choice were not significantly different between CMV-infected and-uninfected infants. Maternal CD4 count <200 cells/Î1/4L during pregnancy was independently associated with congenital CMV (adjusted odds ratio, 2.9; 95% CI, 1.2-7.3). A negative correlation between CMV load in saliva and maternal CD4 count was observed (r = â̂'0.495, n = 22, P =. 019). Conclusions.The birth prevalence of congenital CMV was high despite prenatal ARV prophylaxis, and was associated with advanced maternal immunosuppression. © 2014 The Author.