Intra-laboratory reproducibility of human papillomavirus identification in cervical specimens by a polymerase chain reaction-based assay

Background: polymerase chain reaction (PCR)-based assays for human papillomavirus (HPV) sequences are in wide use in clinical and epidemiological studies. The reproducibility of these assays is not extensively studied. Objectives: to estimate the intra-laboratory reproducibility of generic and type-specific: HPV diagnoses by the MY09/MY11/HMB01 consensus L1 primer-based PCR assay. Study design: systematically collected specimens (n = 207) were masked and retested. Results: when specimens negative in both initial and repeat assays were excluded from analysis, the diagnostic reproducibility was 98.6% for beta -globin, 90.7% for generic HPV (any HPV type), and 76.9% for type-specific HPVs. The reproducibility of type-specific diagnosis increased with increase in signal strength in the hybridization reaction of the initial assay. When a specimen contained five or more HPV types in the initial assay, it was rare to identify all of the HPV types ill the repeat assay. Conclusions: the degree of reproducibility of the PCR diagnosis should be taken into account in the interpretation of HPV data in clinical and epidemiological studies. (C) 2000 Elsevier Science B.V, All rights reserved.