We determined the prevalence of type-specific hrHPV infections in the Netherlands on cervical scrapes of 45 362 women aged 18 -65 years. The overall hrHPV prevalence peaked at the age of 22 with peak prevalence of 24%. Each of the 14 hrHPV types decreased significantly with age (P-values between 0.0009 and 0.03). The proportion of HPV16 in hrHPV-positive infections also decreased with age (OR ¼ 0.76 (10-year scale), 95% CI ¼ 0.67 -0.85), and a similar trend was observed for HPV16 when selecting hrHPV-positive women with cervical intraepithelial neoplasia grade 2 or worse (CIN2 þ ) (OR ¼ 0.76, 95% CI ¼ 0.56 -1.01). In women eligible for routine screening (age 29 -61 years) with confirmed CIN2 þ , 65% was infected with HPV16 and/or HPV18. When HPV16/18-positive infections in women eligible for routine screening were discarded, the positive predictive value of cytology for the detection of CIN2 þ decreased from 27 to 15%, the positive predictive value of hrHPV testing decreased from 26 to 15%, and the predictive value of a double-positive test (positive HPV test and a positive cytology) decreased from 54 to 41%. In women vaccinated against HPV16/18, screening remains important to detect cervical lesions caused by non-HPV16/18 types. To maintain a high-positive predictive value, screening algorithms must be carefully re-evaluated with regard to the screening modalities and length of the screening interval. British Journal of Cancer (2008) . Prophylactic vaccines are now available that are effective against incident and persistent HPV16 and 18 infections (Koutsky et al, 2002;Munoz et al, 2003;Harper et al, 2006; Villa et al, 2006). Mass vaccination can have a substantial impact on the cervical cancer incidence, even in developed countries where the incidence is already low because of implementation of organised cervical screening. To make a well-judged decision about the future role of both vaccination and screening in cervical cancer prevention, detailed information about type-specific hrHPV distribution is required. Data on hrHPV prevalence and type distribution in the Netherlands have been reported previously, but only for a relatively small cohort of 3305 women (Jacobs et al, 2000a;Clifford et al, 2005). To obtain reliable estimates of the hrHPV type distribution in relation to age, both for women with normal and for women with abnormal cytology, data from a far larger cohort are needed. In this study, we determined the age-dependent prevalence of 14 hrHPV types and the age-dependent type distribution within hrHPV-positive women from a cohort of 45 362 women in the Netherlands aged 18 -65 years. The results from the current study give an impression of the potential benefits that can be achieved from HPV16/18 vaccination. They can also be used to gain insight into the effects of partial cross-protection against HPV types 31 and 45 (Harper et al, 2006). Furthermore, the figures will serve as inputs for simulation models in which different vaccination and screening strategies will be compared.
MATERIALS AND METHOD...