The early diagnosis of pediatric HIV-1 infection is a critical issue in resource-limited settings to prioritize eligibility for antiretroviral therapy among HIV-1-infected children. A case-control study was performed within the ANRS 1201/1202 Ditrame Plus cohort (Abidjan, Côte d'Ivoire) to assess the usefulness of CD4+ T-cell percentage (CD4%) and total lymphocyte count (TLC) measured early in life in African children born to HIV-1-infected mothers. Using plasma HIV-1 RNA testing at 4 weeks of life as gold standard, CD4% and TLC were determined at month 3 and 6 in all 33 children HIV-1-infected in utero or intrapartum/early postpartum (cases) born to mothers receiving peripartum antiretroviral prophylaxis. Controls were 66 HIV-1-uninfected children from the same cohort. At month 3, the median CD4% was significantly lower in HIV-1-infected children (17.7%, 95% percentiles, 7.1-27.4) than in uninfected controls (34.8%, 18.5-45.3) (P < 0.001). A comparable difference was also observed at month 6. At the same time points, no significant difference was measurable for TLCs. The best threshold differentiating HIV-infected and uninfected children at month 3 was 25% CD4+. Compared to HIV-1 RNA results, sensitivity of this marker was 87.1% (95% confidence interval, 70.2-96.4) at month 3 and 88.9% (70.8-97.6) at month 6. Specificity was 78.3% (63.6-89.0) and 88.3% (77.4-95.2), respectively. Early CD4% measurement allows one to classify adequately the vast majority of exposed children according to their HIV status. CD4% should be further evaluated under field conditions for the diagnosis of pediatric HIV-1 infection and the monitoring of pediatric antiretroviral therapy.