Background: Human papillomavirus (HPV) is the major cause of cervical neoplasia but estimates of the population attributable fraction (PAR%), of HPV vary. PAR% has not been derived from longitudinal studies although assessment of HPV exposure prior to the neoplasia diagnosis should increase validity of such estimates.
Aims: Systematic review and meta-analysis of longitudinal studies on HPV associated relative risk (RR) for and PAR% of HPV16 in cervical neoplasia.
Methods: Pertinent data from longitudinal studies was made available through Medline and substituted by various hand searches. HPV associated weighted mean RR, with 95% confidence interval (CI) of cervical neoplasia, and the PAR% of HPV16 in cervical carcinoma were estimated both for unselected and low HPV prevalence populations.
Results: HPV associated RR of cervical carcinoma in PCR-based studies was 17 (95% CI 8.2-33). HPV16 associated RRs in seroepidemiological studies were 3.3 (95% CI 2.2-4.9) for the unselected population, HPV16 seroprevalence 11.0%, and 12.5 (95% CI 5.5-29) for a population with low HPV16 seroprevalence of 5.3%. Corresponding PAR% estimates of HPV16 were 27 and 44%, respectively.
Conclusion: Protective vaccination against HPV16 infection would prevent up to 44% of cervical carcinoma.