Human papillomavirus testing versus cytology in primary cervical cancer screening: End-of-study and extended follow-up results from the Canadian cervical cancer screening trial

Abstract

The Canadian Cervical Cancer Screening Trial was a randomized controlled trial comparing the performance of human papillomavirus (HPV) testing and Papanicolaou cytology to detect cervical intraepithelial neoplasia of grades 2 or worse (CIN2+) among women aged 30–69 years attending routine cervical cancer screening in Montreal and St. John's, Canada (n = 10,154). We examined screening and prognostic values of enrollment cytologic and HPV testing results. Extended follow-up data were available for St. John's participants (n = 5,754; 501,682.6 person-months). HPV testing detected more CIN2+ than cytology during protocol-defined (82.9 vs. 44.4%) and extended (54.2 vs. 19.3%) follow-up periods, respectively. Three-year risks ranged from 0.87% (95% CI: 0.37–2.05) for HPV-/Pap- women to 35.77% (95% CI: 25.88–48.04) for HPV+/Pap+ women. Genotype-specific risks ranged from 0.90% (95% CI: 0.40–2.01) to 43.84% (95% CI: 32.42–57.24) among HPV− and HPV16+ women, respectively, exceeding those associated with Pap+ or HPV+ results taken individually or jointly. Ten-year risks ranged from 1.15% (95% CI: 0.60–2.19) for HPV−/Pap− women to 26.05% (95% CI: 15.34–42.13) for HPV+/Pap+ women and genotype-specific risks ranged from 1.13% (95% CI: 0.59–2.14) to 32.78% (95% CI: 21.15–48.51) among women testing HPV− and HPV16+, respectively. Abnormal cytology stratified risks most meaningfully for HPV+ women. Primary HPV testing every 3 years provided a similar or greater level of reassurance against disease risks as currently recommended screening strategies. HPV-based cervical screening may allow for greater disease detection than cytology-based screening and permit safe extensions of screening intervals; genotype-specific testing could provide further improvement in the positive predictive value of such screening.