Age-specific and genotype-specific carcinogenic human papillomavirus prevalence in a country with a high cervical cancer burden: results of a cross-sectional study in Estonia

Objectives To describe age-specific and type-specific carcinogenic human papillomavirus (HPV) prevalence prior to large-scale effect of HPV vaccines in Estonia and to analyse the risk factors associated with carcinogenic HPV. Design Cross-sectional study using self-administered questionnaire and self-collected vaginal swabs for detection of HPV infection. Setting Estonian Biobank database. Participants Stratified random sample of women aged 30-33, 57-60 and 67-70 years living in one of the three largest counties in Estonia. Of 3065women approached, 1347 (43.9%) returned questionnaires and specimens for HPV DNA detection. Outcome measures HPV prevalence and fully adjusted ORs with 95% CIs for risk factors. Results HPV prevalence was highest among women aged 30-33 years (18.7%; 95%CI 15.8 to 21.9) followed by those aged 67-70 years (16.7%; 95%CI 12.4 to 22.0) and 57-60 years (10.2%; 95%CI 7.8 to 13.3). HPV16 and HPV56 were the most common among women aged 30-33 years (both 4.0%; 95%CI 2.7 to 5.9), and HPV68 was the most common among women aged 57-60 years (2.8%; 95%CI 1.5 to 4.7) and 67-70 years (6.4%; 95%CI 3.6 to 10.4). Vaccination with nonavalent vaccine would have halved the carcinogenic HPV prevalence among women aged 30-33 years. The odds of infection with carcinogenic HPV were higher among women with six or more sexual partners among younger (OR 2.99; 95%CI 1.54 to 5.81) and older (OR 3.80; 95%CI 1.25 to 11.55)women and lower (OR 0.35; 95%CI 0.17 to 0.72) among younger married women. Conclusions This study demonstrated U-shaped age-specific genotype profile of carcinogenic HPV prevalence, indicating that public health providers should focus on developing exit strategies for the cervical cancer screening programme in Estonia with a possible extension of HPV testing beyond the current screening age of 65 years. Generalisability of the findings of this study may be affected by the low response rate.