Background: HPV testing is part of primary screening. The optimal screening intervals should be based on the long-term predictive values (absolute risks) of clinically-relevant outcomes after testing positive or negative.

Objectives: To estimate the risk of all grades of abnormal cytology and histology in the years following a screening visit, including HPV typing and baseline cytology.

Methods: At entry into the Portland Kaiser cohort study in 1988-89, >20,000 women were screened by cytology; a second, cervicovaginal specimen was collected and frozen to permit typing for >40 types of HPV by MY09-MY11 PCR. Using the Kaiser cytology and histology database, we generated 16 years of prospective follow-up after the baseline screening visit, and calculated the absolute risks of all grades of subsequent cytologic and histologic abnormality.

Results: The elevated risk of CIN3/cancer following baseline HPV16 positivity lasted for approximately 10 years; for HPV18, risk of CIN3/cancer rose through 16 years. HPV31 also strongly predicted risk of CIN3. A negative HPV test at baseline, especially among women 30 and older, predicted extremely low risk of cervical cancer throughout follow-up. Baseline HPV status strongly predicted low-grade cytologic and histologic abnormalities, but only for a few years after baseline. Overall, HPV testing provided more enduring risk stratification than baseline cytology. At the meeting, we will present full cytologic and histologic results for all baseline HPV types and cytologic interpretations, stratified by age.

Conclusions: In this cohort, CIN3/cancer diagnoses were predicted by HPV18 positivity for at least 16 years, demonstrating the unique carcinogenic quality of this type. HPV16 was the strongest predictor of CIN3/cancer but cases were found within 10 years suggesting a more rapid and/or more overt carcinogenic process. The negative predictive value of a single HPV test among women 30 and older shows that HPV testing will permit lengthened screening intervals.