Evaluation of HPV16PCR detection in self-compared with clinician-collected samples in women referred for colposcopy

Objectives. A clinical prospective evaluation study was conducted to evaluate PCR detection of high-risk human papillomavirus (HPV) type 16 in self-sampled vaginal compared with clinician-collected cervical specimens. Methods. Paired vaginal and cervical specimens were collected from 137 consecutive women referred for colposcopy because of abnormal cervical cytology. In-house and a commercial PCR method for HPV type 16 were used. Self-sampled vaginal HPV 16 detection was compared to histology and physician-collected cervical specimens. Results. Of the 137 patients, 98 had proven abnormal histology and were included in the analysis. Overall, using the cervix HPV detection as reference method, the self-sampled vaginal sample showed sensitivity 91.8%, specificity 96.1% and agreement kappa (kappa) 0.881. Using the histology as reference, all 11 cervical cancer cases were HPV-16-positive in both cervical and vaginal samples, and in 43 high-grade lesions, detection sensitivity in cervix was 72.1% (kappa 0.588) and vagina 67.4% (kappa 0.516). HPV 16 detection did not differ (P = 0.27) between clinician-collected cervical and self-sampled vaginal specimens. Conclusions. The self-collected vaginal sample is highly concordant with the physician-collected cervical sample in HPV 16 detection. (c) 2006 Elsevier Inc. All rights reserved.