This pilot study proposes an index to quantify the intensity of high-risk anal sex and presents factors associated with this practice among 65 homosexual men with HIV infection. All of them answered a self-administered questionnaire. Using multiple logistic regression, we first compared those who had had safe sex in the past six months to those who had had unsafe sex. Variables significantly associated with unsafe sex are: drugs and/or alcohol use with sex, and less favorable attitude towards condom use. Finally, multiple linear regression analysis among respondents who had had unsafe sex indicated that the following factors are associated with higher levels of unsafe sex: men between the ages of 32 and 38, satisfied with their social life, non-smokers, with a less favorable attitude towards condom use, and with few social relationships. These results must be interpreted with caution because of the small number of respondents.