Background: As the result of successful randomized trials, some countries might switch from cytology-based to HPV test-based cervical screening.

Objective: To evaluate the clinical implications for colposcopy if we switch to HPV test-based screening.

Methods: Data are from the Guanacaste cohort study. At enrollment, 8545 women were screened with three cytologic methods (conventional cytology, liquid-based cytology (LBC) and PapNet), and cervicography; a specimen was also collected for HPV DNA PCR-testing. Final case diagnosis (≥CIN3, CIN2 or ≤HSIL) was assigned at the end of the 7-year study. Using cervicography as a surrogate of colposcopy, we examined what colposcopists would see for each cytologic method compared with HPV DNA testing (13 carcinogenic types). We calculated sensitivity, specificity, NPV and PPV using the final case assignment.

Results: Independent of cytologic method, moving to HPV screening would raise the number of referrals (decreased specificity) and the number of ≥CIN2 cases detected (increased sensitivity). Among the enlarged group of referred women, on a per patient basis, colposcopists would not see a major change in the percentage with acetowhite lesions. Although the highest NPV were found among HPV-negative women with no visible lesions, most cases of ≥CIN2 captured by HPV testing did not have acetowhite lesions at baseline. HPV testing predicted risk of high-grade CIN better than the three cytological tests (Youden’s index); however, sensitive reading of LBC performed similarly to HPV. All the methods were more accurate if the screening was restricted to women >30 years old.

Conclusions: The major concern with switching to HPV screening is how to manage the many HPV-positive women. As a group they are at elevated risk, but many will not even have acetowhite lesions. Colposcopic criteria will need adjustment among women known to be HPV-positive, and evidence-based guidelines will also be needed prior to widespread implementation of primary HPV screening.