Ninety-two adult human immunodeficiency virus (HIV)-infected patients with suspected herpes zoster were included in a study. The clinical diagnosis of herpes zoster was verified by examination of blister cell and fluid material or serum samples. Antigen detection by a direct immunofluorescence assay with a fluorescein isothiocyanate-labelled monoclonal antibody, virus isolation, and serologic methods (in-house varicella-zoster virus [VZV] immunoglobulin G [IgG] and IgM enzyme-linked immunosorbent assays and the commercial Enzygnost assay) were compared. The direct immunofluorescence assay was found to be the most sensitive method, diagnosing 85 of 92 infections (92%), while the sensitivity of virus isolation was 65% (60 of 92 patients). Despite the use of two different serological methods, only 60 of 92 patients (65%) had significant VZV IgG titer rises, and only 26 of 92 patients (28%) had detectable VZV IgM. The lack of a VZV IgG antibody titer rise was found to correlate with low CD4 counts in peripheral blood and high VZV IgG titers in the acute-phase serum sample. The frequency of IgM-positive sera was lower than that expected from reports of studies with patients without AIDS. This may be related to early antiviral treatment or deficient antibody production due to the HIV-related immunosuppression. There was no significant difference in CD4 counts between VZV IgM-positive and -negative patients.