Human papillomavirus (HPV) testing has been suggested for primary screening of cervical cancer. Prediction of future high-grade cervical lesions is crucial for effectiveness and cost. We performed a case control study in a retrospective cohort of women with at least two cervical smears, all but the last one being negative, from the organized cervical screening programme in Florence, Italy. We searched for high-risk HPV in all previous, archival, smears from cases (new histologically confirmed cervical intraepithelial neoplasia (CIN) grade II or worse) and in one previous smear from each control (last smear cytologically normal, matched by age and interval (latency) from last smear). We applied polymerase chain reaction (PCR), and the b-globin gene was used as a DNA preservation marker. High-risk HPV was identified in 71/92 (77.17%) previous smears from 79 cases and 17/332 controls (5.12%). The odds ratio (OR) was 63.76 (95% CI 30.57-132.96). Among cases the proportion of HPV-positive smears declined slightly with increasing latency. Among cases, HPV was found in 81.24% (95% CI 69.93-88.96%) of smears with latency < 4 years and in 67.80% (95% CI 47.72-82.93%) of those taken at longer intervals, up to 6 years. These findings suggest that testing for high-risk HPV allows predicting 80% of CINII/III 3 years before the cytological diagnosis and two thirds 6 years before. They also suggest that testing women negative for high-risk HPV at longer interval and strictly following-up women who are HPV positive could be an effective strategy for cervical cancer screening.
Copyright 2000 Cancer Research Campaign.