BACKGROUND: The rate of mother-to-child transmission (MTCT) of human immunodeficiency virus (HIV) type 1 is as low as 0.5% in non-breast-feeding mothers who delivered at term while receiving antiretroviral therapy with a plasma viral load <500 copies/mL. This situation accounted for 20% of the infected children born during the period 1997-2006 in the French Perinatal Cohort. We aimed to identify factors associated with such residual transmission risk. METHODS: We performed a case-control study nested in the aforementioned subpopulation of the French Perinatal Cohort. RESULTS: Nineteen case patients (transmitters) and 60 control subjects (nontransmitters) were included. Case patients and control subjects did not differ by geographical origin, gestational age at HIV diagnosis, type of antiretroviral therapy received, or elective Cesarean delivery. Case patients were less often receiving treatment at the time that they conceived pregnancy than control subjects (16% vs 45%; P=.017). A lower proportion of case patients had a viral load <500 copies/mL, compared with control subjects, at 14 weeks (0% vs 38.1%; P=.02), 28 weeks (7.7% vs 62.1%; P=.005), and 32 weeks: (21.4% vs 71.1%; P=.004). The difference remained significant when we restricted analysis to the 10 of 16 intrapartum transmission cases. In a multivariate analysis at 30+/-4 weeks adjusted for viral load, CD4(+) T cell count, and time at antiretroviral therapy initiation, viral load was the only factor independently associated with MTCT of HIV (adjusted odds ratio, 23.2; 95% confidence interval, 3.5-553; P<.001). CONCLUSIONS: Early and sustained control of viral load is associated with a decreasing residual risk of MTCT of HIV-1. Guidelines should take into account not only CD4(+) T cell count and risk of preterm delivery, but also baseline HIV-1 load for deciding when to start antiretroviral therapy during pregnancy.