BackgroundVaricella-zoster virus (VZV) reactivation is common but difficult to predict in HIV-infected persons.ObjectiveSince qualitative VZV antibodies can determine past VZV disease or vaccination, we evaluated whether quantitative VZV antibody levels over time can predict future zoster.Study designUS Military HIV Natural History (NHS) participants with a zoster diagnosis at least 5 years after HIV diagnosis (n = 100) were included. Zoster-negative controls (n = 200) were matched by age, race, gender, and CD4 count at HIV diagnosis. Repository plasma specimens collected at baseline and prior to zoster diagnosis were evaluated using a quantitative anti-VZV ELISA assay. Differences in quantitative VZV levels were analyzed by Wilcoxon Mann–Whitney and Fisher’s exact tests.ResultsMedian CD4 count at HIV diagnosis was similar for cases and controls (535 [IQR 384–666] vs. 523 [IQR 377–690] cells/μL; p = 0.940), but lower for cases at zoster diagnosis (436 [IQR 277–631] vs. 527 [IQR 367–744] cells/μL; p = 0.007). Antiretroviral therapy (ART) use prior to zoster diagnosis was lower for cases (52.0%) compared to controls (64.5%; p = 0.025). Cases had similar mean VZV antibody levels prior to zoster diagnosis compared to controls [2.25 ± 0.85 vs. 2.44 ± 0.96 index value/optical density (OD) ratio; p = 0.151] with no difference in the change in antibody levels over time (0.08 ± 0.71 vs. 0.01 ± 0.94 index value/OD per year; p = 0.276).ConclusionQuantitative VZV antibody levels are stable in HIV-infected persons and do not predict zoster reactivation. Low CD4 count and lack of ART use appear to be better predictors of future zoster diagnosis.