Background: We evaluated data from prospectively identified twins to understand better the mechanisms and covariates of mother-to-infant transmission of human immunodeficiency virus (HIV).
Methods: Using data obtained from an international collaboration and multivariate quasilikelihood modeling, we assessed concordance, birth order, route of delivery, and other factors for HIV infection in 115 prospectively studied twin pairs born to HIV-infected women. Actuarial methods were used to evaluate overall survival and survival free of acquired immunodeficiency syndrome for HIV-infected twins.
Results: Infection with HIV occurred in 35% of vaginally delivered firstborn (A) twins, 16% of cesarean-delivered A twins, 15% of vaginally delivered second-born (B) twins, and 8% of cesarean-delivered B twins. In a multivariate model, the adjusted odds ratios for HIV infection were 11.8 (confidence interval: 3.1 to 45.3) for concordance of infection with the co-twin, 2.8 (confidence interval: 1.6 to 5.0) for A versus B twins, and 2.7 (confidence interval: 1.1 to 6.6) for vaginally delivered versus cesarean-delivered twins. Among A twins, 52% (lower confidence limit: 6%) of the transmission risk was related to vaginal delivery. Comparing vaginally delivered A twins (infants most exposed to vaginal mucus and blood) to cesarean-delivered B twins (infants least exposed), 76% (lower confidence limit: 48%) of the transmission risk was related to vaginal exposure. Infected B twins had slightly reduced Quetelet indexes and more rapid development of illnesses related to acquired immunodeficiency syndrome.
Conclusions: These results indicate that HIV infection of B twins occurs predominantly in utero, whereas infection of A twins (and, by implication, singletons) occurs predominantly intrapartum. We propose that intrapartum transmission is responsible for the majority of pediatric HIV infections and that reducing exposure to HIV in the birth canal may reduce transmission of the virus from mother to infant.