To investigate whether the epidemiologic correlates of cervical cancer are predictors of infection with genital human papillomavirus (HPV), we performed a prevalence survey in two metropolitan areas of Brazil, Recife and São Paulo. The data records of four randomly selected HPV-negative women were matched on the basis of age, clinic, and admission period with those of each of 136 patients with positive HPV DNA hybridizations. Anal intercourse [prevalence rate ratio (PRR) = 1.7] and current pregnancy (PRR = 2.3) were the only variables associated with HPV 6/11 infection (P less than .10). Only the frequency of gynecologic consultations was associated (negatively) with risk of HPV 16/18 infection (P = .0175). Our data failed to provide evidence for the existence of shared risk factors for genital HPV infection and cervical cancer. The frequency of mixed HPV infections was 13 times higher than expected, a finding suggestive of the existence of additional determinants of HPV infection not akin to the general behavioral characteristics of women that are probed in the study.