Comparison of urine, self-collected vaginal swab, and cervical swab samples for detecting human papillomavirus (HPV) with Roche Cobas HPV, Anyplex II HPV, and RealTime HR-S HPV assay

Background Human papillomavirus (HPV) is well established as the main cause of cervical cancer. Non-invasive self-collected urine and vaginal sampling have the potential advantage of increasing patient compliance with cervical cancer screening. Methods Self-collected vaginal and urine samples and clinician-collected cervical samples were collected from 101 patients, including 84 patients with high grade squamous intraepithelial lesion and 17 patients with benign ovarian disease. Each sample was evaluated with RealTime HR-S HPV, Anyplex™ II HPV, and Cobas® HPV assays. The concordance of urine and of self-collected vaginal samples with cervical samples was assessed using the kappa (k) statistic. Results In any high-risk HPV (hrHPV), the concordance of self-collected vaginal and urine samples compared to cervical samples was moderate (k 0.49–0.58) and fair to moderate (k 0.33–0.51), respectively. In HPV 16/18, the concordance of vaginal and urine samples compared to cervical samples was almost perfect (k 0.81–0.86) and moderate to substantial (k 0.59–0.63), respectively. Among the three methods for HPV detection, RealTime HR-S showed the highest concordance with vaginal (k: any hrHPV 0.58, HPV 16/18 0.86) and urine samples (k: any hrHPV 0.51, HPV 16/18 0.63) compared to cervical samples. Conclusion HPV tests using self-collected vaginal samples and urine showed substantial and moderate agreement compared with cervical samples, respectively, although HPV tests using these samples were still inferior to clinician-collected cervical samples. Further research is needed on the clinical performance of HPV testing using urine and self-collected vaginal samples as the screening method.