BACKGROUND The live attenuated varicella vaccine is recommended for HIV-infected children who are not severely immunosuppressed. This study aimed to assess the immunogenicity and safety of varicella vaccination among HIV-infected children who had severe immunosuppression before receiving antiretroviral therapy. METHODS Sixty HIV-infected children with no history of chickenpox or herpes zoster infection with CD4 T lymphocyte counts ≥ 15% or ≥ 200 cell/mm were enrolled. Two doses of varicella vaccine were administered at the time of enrollment and at 3 months. Varicella zoster virus (VZV) antibody was tested at baseline and 3 months after each dose by the enzyme-linked immunosorbent assay technique. An antibody titer >20 HU/mL was regarded as protective. RESULTS The median (interquartile range) of age, CD4 nadir, and current CD4 percentage were 11.2 (8.5-12.8) years, 9.5% (3-14), and 28% (22-32), respectively. Fifty-seven children (95%) received antiretroviral therapy for a median of 27 months. Among 34 children (57%) who were VZV seronegative at baseline, 11.8% (95% CI, 3.3%-27.5%) and 79.4% (95% CI, 62.1%-91.3%) were VZV seroconverted after first and second dose of vaccine, respectively. Children who had VZV seroconversion were more likely to have HIV RNA <1.7 copies/mL (92.6% vs. 71.4%, P = 0.18). Among 26 children who were seropositive at baseline, the geometric mean titers were increased from 56.7 to 107.9 and 134.6 unit/mL, respectively. Local and systemic reactions of grade 1 and 2 were reported in 13% and 4% of children, respectively. There was a trend toward better response among children with younger age, high CD4, and viral suppression. CONCLUSIONS Administration of the 2 doses of varicella vaccine resulted in high seroconversion rates without serious adverse reactions. Varicella vaccination for HIV-infected children should be encouraged.