We evaluated a study setting for assessment of the long-term vaccine efficacy (VE) of human papillomavirus (HPV) virus-like-particle (VLP) vaccine against cervical carcinoma. A total of 22,412 16- to 17-year old adolescent women from seven cities in Finland were invited by letter to participate in a phase III study of a quadrivalent HPV (types 6, 11, 16, 18) VLP vaccine, between September 2002 and March 2003. A total of 30,947 18-year old women were invited to participate as unvaccinated controls. These women were asked about their willingness to participate in an HPV vaccination trial and to fill a health questionnaire. These three population-based cohorts of adolescent women, including women vaccinated with HPV vaccine or placebo vaccine and unvaccinated control women, are systematically followed over time. The study cohort database will be linked with the Finnish Cancer Registry using cervical carcinoma in situ (CIS) and invasive cervical carcinoma (ICC) as endpoints. Assuming that the cumulative incidence of CIS and ICC over 15 years is 0.45%, and that there is no loss to follow-up, and power of 80%, the determination of 70% total VE will require 3357 HPV vaccine recipients, 3357 placebo vaccine recipients, and 6714 unvaccinated controls. At the baseline, 2632 (12%) of the invited adolescents volunteered to the phase III vaccination trial, and 6790 (22%) responded to the questionnaire study. During a recruitment period of 10 months, 874 HPV vaccine recipients, 875 placebo recipients and 1919 unvaccinated controls were enrolled. Population-based enrollment of large cohorts of vaccinated and unvaccinated adolescents for passive registry-based follow-up with cervical carcinoma as the end-point is feasible and currently going on in Finland.

Human papillomaviruses (HPVs, most notably types 16 and 18) cause cervical carcinoma, the second most common cancer among women. Vaccination of adolescents against HPV16/18 might prevent large proportion of cervical and other anogenital cancers. However, because of ethical reasons this cannot be proven by clinical studies. To determine the long-term vaccine efficacy (VE) of HPV16/18 virus-like-particle (VLP) vaccine against cervical carcinoma in situ (CIS+) and invasive cervical carcinoma, the following three population-based cohorts of adolescent women have been enrolled: (1) women vaccinated with the HPV vaccine; (2) women vaccinated with hepatitis A control vaccine; and (3) unvaccinated control women. These cohorts will be passively followed for cumulative incidence of CIS+ endpoints by population-based cancer registry. Overall 24,046 16- to 17-year-old adolescent women from 18 cities in Finland were invited between May 2004 and June 2005 to participate in a phase III trial with bivalent HPV16/18 VLP vaccine. A total of 58,996 18- to 19-year-old women were invited in May 2005 to participate as unvaccinated controls. Women who reported their willingness to participate in an HPV vaccination trial had they been 1-2 years younger were eligible. Cumulative incidence (CI) of CIS+ in our cohorts over 15 years is approximately 0.45%. VE of 70% against CIS+ with 80% power requires 3357-3189 HPV16/18 vaccine recipients, 3357-3189 other vaccine recipients, and 6714-9567 unvaccinated controls. We have now enrolled 2404 HPV16/18 vaccine recipients, 2404 hepatitis A-vaccine recipients, and 5130 unvaccinated controls. This enrolment in addition to our earlier enrolment in another phase III trial guarantees enough power so that by 2020 we can ultimately provide data on the efficacy of HPV16/18 vaccination against CIS+.