Background: The first vaccine against human papillomavirus (HPV)-related disease is now available. Although it has been designed and tested mainly to protect against cervical lesions, it is also expected to be effective against other anogenital cancers. Associations between HPV and vaginal, vulvar, and anal cancers are well established, but the full extent in terms of age and time since diagnosis of these associations is not well known.
Methods: We established a cohort of all women in Sweden who were aged 18-50 years at some timepoint from 1968 to 2004. Using national registration numbers, we linked this cohort to nationwide population, migration, cancer, and death registers. The incidence rate ratios (IRRs) of vaginal, vulvar, anal, and rectal cancer in women with a history of a cervical intraepithelial neoplasm (CIN), grade 3, compared with women with no such history were estimated by use of multivariate Poisson regression.
Findings: Women with a history of grade 3 CIN had increased risks of cancer of the vagina (6.74 [95% CI 5.24-8.56]), vulva (2.22 [1.79-2.73]), and anus (IRR 4.68 [3.87-5.62]). No excess risk was found for rectal cancer. For all four anatomical sites, the IRRs varied substantially with the amount of time that had elapsed since the date of first diagnosis of grade 3 CIN. Analyses stratified by attained age during follow-up showed that the risk of cancer conferred by a history of diagnosis of grade 3 CIN was highly age dependent. The observed increased risks were not explained by smoking or socioeconomic status.
Interpretation: This study confirms the known association between history of CIN, presumed HPV infection, and increased risk of cancers of the vagina, vulva, and anus by use of large and complete databases, but also shows that this risk varies both by the time from initial diagnosis of grade 3 CIN and by the age of the individual. Further studies are needed to clarify the type of HPV associated with this increase in risk to determine the clinical applicability of the new HPV vaccines.