ACCURACY AND INTERLABORATORY RELIABILITY OF HUMAN PAPILLOMAVIRUS DNA TESTING BY HYBRID CAPTURE

Epidemiologists and clinicians wishing to introduce human papillomavir us (HPV) testing into cervical cancer prevention programs need standar dized, reliable, and accurate HPV DNA tests that can detect the full s pectrum of pathogenic HPV types. The Hybrid Capture System assay from Digene (hybrid capture assay) is a nonradioactive kit designed to dete ct 14 HPV types in two groups: a mix of 9 high-risk types associated w ith anogenital cancer (HPV types 16, 18, 31, 33, 35, 45, 51, 52, and 5 6) and another group of 5 low-risk types associated with condyloma acu minatum (HPV types 6, 11, 42, 43, and 44). The assay yields quantitati ve data meant to reflect viral concentration. In a study of 199 cervic al specimens from women with concurrent Pap smears, we assessed the re liability of the new assay by comparing the hybrid capture assay resul ts from three laboratories. We assessed the accuracy of the hybrid cap ture assay in comparison with a reference standard of HPV DNA content (multiple testing by several methods in two reference laboratories). W e also compared the hybrid capture assay results with the concurrent c ytologic diagnoses on the basis of an independent review of each smear by five pathologists. Pairwise interlaboratory agreement rates on HPV positivity for either high-risk or low-risk types ranged from 87 to 9 4%, and kappa values ranged from 0.61 to 0.83. Among specimens positiv e for high-risk types (the most important clinical outcome), the inter laboratory correlations of the quantitative data ranged from 0.60 to 0 .90. Test results from all three laboratories were strongly associated with those of the HPV DNA reference standard and with the concurrent cytopathologic diagnoses. The most common errors were sporadic, appare ntly false-positive results.