To clarify the cause of herpes zoster in immunocompetent children, specific humoral and cellular immunity was determined using an ELISA and a lymphoproliferative assay, respectively, in infants < 1 year of age and children > or = 1 year of age who had chickenpox. Thirteen (59.1%) of 22 infants, 17 (81.0%) of 21 children > or = 1 year of age (P < .02), and 13 (86.7%) of 15 children > or = 2 years of age (P < .001) had positive varicella-zoster virus (VZV)-specific cellular immunity. VZV-specific antibodies in infants were significantly lower than those in children > or = 1 year old (P < .01) and > or = 2 years old (P < .001). The possibility of subclinical reactivation was demonstrated by an increase in the specific cellular or humoral immunity (or both) in all of 6 infants who had negative specific cellular immunity. The low response of specific immunity and the immunologic evidence of reactivation in infants after chickenpox provide a possible explanation for the finding that chickenpox in infancy is a risk factor for herpes zoster.