Background Understanding patterns of human papillomavirus (HPV) infection in partnerships is essential in exploring transmission of HPV in sexual networks. Risk factors for HPV infection have yet to be explored at the dyad level. We studied features that predict presence of HPV in a new sexual partnership.

Methods We analysed data from the HITCH Cohort Study of recently-formed couples. Women aged 18–24 attending university/college in Montreal, Canada and their male partners were recruited in 2006–10. Self-collected vaginal swabs and clinician-obtained swabs from the penis and scrotum were tested for DNA of 36 HPV types. We analysed baseline data from 479 couples. HPV in a partnership was defined as the presence of 1 or more HPV types in either or both partners. We used Poisson regression to calculate prevalence ratios with 95% confidence intervals for candidate risk factors.

Results Most women were unvaccinated (88%). 67% of partnerships harboured HPV. For 49% both partners were HPV+. Detection was associated with the combined total of the male’s and female’s lifetime partners; from 27.5% among couples who jointly had no more than 4 partners to 94.2% among couples with > 20. Couples reporting concurrent partners were 2.8 times (95% CI 1.7–4.5) more likely to have HPV compared to those with a 12-month gap since the last extra-dyadic partner but this effect disappeared after adjustment for number of partners. Couples who always or frequently used condoms with their previous partner(s) were 29% (95% CI 9–45%) less likely to have HPV after accounting for number of partners and gap length/concurrency.

Conclusions Number of extra-dyadic partners (past or concurrent) predicts the likelihood of HPV in a partnership. Condoms may have some impact on limiting spread of HPV, although protection was incomplete. Epidemiologic monitoring of HPV in sexual networks is needed, particularly in populations with suboptimal HPV vaccine coverage.