Background: We evaluated the association between maternal cytomegalovirus (CMV) viremia during pregnancy and adverse birth and infant health outcomes in HIV-infected mothers and their HIV-exposed uninfected infants.

Methods: HIV-positive women and their infants were followed prospectively from pregnancy through 2 years postpartum in the "Tshipidi" study in Botswana. We analyzed the association between detectable CMV DNA in maternal blood at delivery and adverse birth outcomes (stillbirth, preterm delivery, small for gestational age, or birth defect), as well as infant hospitalization and mortality through 24 months.

Results: We measured CMV DNA in blood samples from 350 (77.1%) of 454 HIV-positive women from the Tshipidi study. The median maternal CD4 count was 422 cells/mL, and median HIV-1 RNA at entry was 3.2 log10 copies/mL. Fifty-one (14.6%) women had detectable CMV DNA. In unadjusted analyses, detectable CMV DNA was associated with higher maternal HIV-1 RNA [odds ratio (OR) 1.4, 95% confidence interval (CI): 1.1 to 1.9], presence of a birth defect (OR 9.8, 95% CI: 1.6 to 60.3), and occurrence of any adverse birth outcome (OR 2.0, 95% CI: 1.04 to 3.95). In multivariable analysis, we observed a trend toward association between detectable maternal CMV DNA and occurrence of any adverse birth outcome (adjusted OR 1.9, 95% CI: 0.96 to 3.8). Maternal CMV viremia was not associated with infant hospitalization and/or death by 24 months.

Conclusions: Approximately 1 in 6 HIV-positive women in Botswana had detectable CMV DNA in blood at delivery. The presence of maternal CMV viremia had a borderline association with adverse birth outcomes but not with 24-month morbidity or mortality in HIV-exposed uninfected children.