Human papillomavirus (HPV) is a sexually transmitted infection but it is unclear whether differences in transmission efficacy exist between individual HPV types. Information on sexual behavior was collected from 11 areas in four continents among population-based, age-stratified random samples of women of ages > or = 15 years. HPV testing was done using PCR-based enzyme immunoassay. Unconditional logistic regression was used to estimate odds ratios (OR) of being HPV positive and corresponding 95% confidence intervals (95% CI). Variables were analyzed categorically. When more than two groups were compared, floating confidence intervals were estimated by treating ORs as floating absolute risks. A total of 11,337 women (mean age, 41.9 years) were available. We confirmed that lifetime number of sexual partners is associated with HPV positivity (OR for > or = 2 versus 1, 1.86; 95% CI, 1.63-2.11) but the association was not a linear one for HPV18, 31, and 33 (i.e., no clear increase for > or = 3 versus 2 sexual partners). Women who had multiple-type infection and high-risk HPV type infection reported a statistically nonsignificant higher number of sexual partners than women who had single-type and low-risk type infections, respectively. Early age at sexual debut was not significantly related to HPV positivity. Husband's extramarital sexual relationships were associated with an OR of 1.45 (95% CI, 1.24-1.70) for HPV positivity in their wives after adjustment for age and lifetime number of women's sexual partners. We did not observe a significant association with condom use. Our study showed an effect of both women's and their husbands' sexual behavior on HPV positivity. Furthermore, it suggests some differences in the pattern of the association between sexual behavior and different HPV types.