Effective therapies and reduced AIDS-related morbidity and mortality have shifted the focus in pediatric human immunodeficiency virus (HIV) from minimizing short-term disease progression to maintaining optimal long-term health. We describe the effects of children's age and pre-antiretroviral therapy (ART) CD4 count on long-term CD4 T-cell reconstitution. CD4 counts in perinatally HIV-infected, therapy-naive children in the Paediatric European Network for the Treatment of AIDS 5 trial were monitored following initiation of ART for a median 5.7 years. In a substudy, naive and memory CD4 counts were recorded. Age-standardized measurements were analyzed using monophasic, asymptotic nonlinear mixed-effects models. One hundred twenty-seven children were studied. Older children had lower age-adjusted CD4 counts in the long term and at treatment initiation (P < .001). At all ages, lower counts before treatment were associated with impaired recovery (P < .001). Age-adjusted naive CD4 counts increased on a timescale comparable to overall CD4 T-cell reconstitution, whereas age-adjusted memory CD4 counts increased less, albeit on a faster timescale. It appears the immature immune system can recover well from HIV infection via the naive pool. However, this potential is progressively damaged with age and/or duration of infection. Current guidelines may therefore not optimize long-term immunological health.