HPV Prevalence in the Dutch cervical cancer screening population (DuSC study): HPV testing using automated HC2, cobas and Aptima workflows

Background: Primary high risk (hr)HPV screening will be introduced in The Netherlands in January 2017. Our aim was to determine the hrHPV prevalence in the Dutch cervical cancer screening population (DuSC study). Methods: A total of 12,113 residual PreservCyt cervical samples from the Dutch population based cytology screening program were rendered anonymous, randomized and tested for hrHPV using 3 HPV assays on their respective automated platforms: QIAGEN's digene® HC2 HPV DNA Test® (HC2, signal amplification), Roche Cobas® HPV test (DNA amplification) and Hologic Aptima® HPV Test (RNA amplification). To determine the agreement between results generated using the different assays, pair wise comparison of the systems was performed by determining kappa coefficients. Results: The selected samples were representative for the population based screening program with respect to age distribution and cytology classification. HrHPV prevalences found were: 8.5% for HC2 (n = 959), 8.1% for cobas (n = 919) and 7.5% for Aptima (n = 849), resulting in a mean hrHPV prevalence of 8.0 ± 0.5%. Although the hrHPV prevalences of the different assays are in the range of 8%, there was a significant difference in prevalence for the HC2 vs. Aptima assay (p-value = 0.007). A clear age dependency was found, with an hrHPV prevalence ranging from 18.7 ± 1.2% in women 29-33 years of age to 4.2 ± 0.2% in women 59-63 years of age. Furthermore, a correlation between hrHPV prevalence and severity of cytology was observed, ranging from 5.5 ± 0.4% in normal cytology to 95.2 ± 1.7% in severe dysplasia. Indeed, kappa coefficients of 0.77, 0.71 and 0.72 (HC2 vs cobas, cobas vs Aptima and Aptima vs HC2, respectively) indicated substantial agreement between the results generated by the different systems. However, looking at the hrHPV positive samples, only 48% of the samples tested positive with all 3 assays. Conclusions: A hrHPV prevalence of 8% was found in this unselected population based screening cohort independently of using HC2, Aptima or cobas. This prevalence is higher than the previously reported 4-5% (POBASCAM and VUSA-Screen trials). Furthermore, the complete automated hrHPV detection workflow solutions from QIAGEN, Roche, and Hologic were successfully used and will be valuable for reliably implementing high throughput hrHPV testing in cervical cancer screening