Factors affecting the response to hepatitis B vaccination in a primary care setting were examined by means of a review of case notes of patients attending 22 sexually transmissible disease services. Where not available from the notes, presence of antibody to hepatitis B surface antigen (anti-HBs) was determined by testing available stored serum. One hundred and ninety-five patients completed a course of 3 injections and had an anti-HBs assay performed. The highest response rate (anti-HBs > or = 10 IU/L) was found in human immunodeficiency virus (HIV)-negative heterosexual women (16 of 17, 94.1%) followed by HIV-negative heterosexual men (11 of 12, 91.7%); HIV-negative homosexual men (105 of 120, 87.5%); and HIV-positive homosexual men (6 of 14, 42.9%). (For HIV-positive vs HIV-negative homosexual men, P = 0.0003). Eleven of 14 (78.6%) homosexual men of unknown HIV status responded to vaccination. There was a trend to lower CD4+ lymphocyte counts among HIV-infected patients who responded to hepatitis B vaccination (mean 482 cells/cm2) when compared to those that did not respond (632 cells) but this difference was not statistically significant (P = 0.330). Neither the type of vaccine (recombinant, plasma-derived or mixed) nor the length of vaccination course (mean 6.2 months; range 2 to 18) affected response. This study confirmed that vaccination against hepatitis B is much less effective in HIV-infected homosexual men and marginally less effective for HIV-negative homosexual men, though the mechanism for this reduced response is uncertain. Reassuringly vaccine response was not affected by common variables in primary care settings such as vaccine type or delays in the vaccine schedule.