Tiffany S. Chang, Jeffrey Wiener, Sheila C. Dollard, Minal M. Amin, Sascha Ellington, Charles Chasela, Dumbani Kayira, Gerald Tegha, Deborah Kamwendo, Denise J. Jamieson, Charlie van der Horst, Athena P. Kourtis

Cytomegalovirus (CMV) infection can be acquired in utero or postnatally through horizontal transmission and breastfeeding. The effect of postnatal CMV infection on postnatal HIV transmission is unknown. The Breastfeeding, Antiretrovirals and Nutrition study, conducted in Malawi, randomized 2369 mothers and their infants to three antiretroviral prophylaxis arms - mother (triple regimen), infant (nevirapine), or neither - for 28 weeks of breastfeeding, followed by weaning. Stored plasma and peripheral blood mononuclear cell specimens were available for 492 infants at 24 weeks and were tested with CMV PCR. Available samples from infants who were CMV PCR-positive at 24 weeks were also tested at birth (N = 242), and from infants PCR-negative at 24 weeks were tested at 48 weeks (N = 96). Cox proportional-hazards models were used to determine if CMV infection was associated with infant morbidity, mortality, or postnatal HIV acquisition. At 24 weeks of age, CMV DNA was detected in 345/492 infants (70.1%); the estimated congenital CMV infection rate was 2.3%, and the estimated rate of CMV infection at 48 weeks was 78.5%. CMV infection at 24 weeks was associated with subsequent HIV acquisition through breastfeeding or infant death between 24 and 48 weeks of age (hazard ratio 4.27, P = 0.05). Most breastfed infants of HIV-infected mothers in this resource-limited setting are infected with CMV by 24 weeks of age. Early CMV infection may be a risk factor for subsequent infant HIV infection through breastfeeding, pointing to the need for comprehensive approaches in order to achieve elimination of breastfeeding transmission of HIV.