HPV Detection in Clinician-Collected, Self-Collected Vagina, Urine, and Menstrual Blood Samples.

OBJECTIVE This study evaluates the concordance, sensitivity, and specificity of self-collected samples-vaginal, urine, and menstrual blood-for detecting high-risk human papillomavirus (HrHPV) and high-grade squamous intraepithelial lesion (HSIL) compared with clinician-collected specimens for cervical cancer screening.

METHODS This prospective study was carried out in the colposcopy clinic of a tertiary hospital in Hong Kong. A consecutive series of premenopausal women referred to our colposcopy clinic due to abnormal Pap smears were screened and recruited. All participants provided self-collected vaginal, urine, and menstrual blood samples. Clinicians collected cervical samples for cytology and HPV testing. A colposcopy examination was performed if indicated. Testing for HrHPV in specimens was analyzed using a 14-type HPV-DNA test (Cobas 4800). Colposcopy with biopsy was used as the reference standard for histologic confirmation of cervical pathology.

RESULTS Of the 167 subjects included in the analysis, 85 (50.9%) had HrHPV infection. The overall concordance between clinician-collected HrHPV results and self-collected samples was 83.2% (vaginal), 79% (urine), and 82.1% (menstrual blood). Among the 167 cases, 139 underwent colposcopy and 42 cases (30.9%) of HSIL+ were identified. The relative sensitivity for HSIL was 0.89 (95% CI, 0.79-1.02) for vaginal, 0.76 (95% CI, 0.63-0.91) for urine, and 0.83 (95% CI, 0.70-0.98) for menstrual blood samples in comparison to clinician-collected cervical samples.

DISCUSSION Self-collected vaginal, urine, and menstrual blood samples show moderate to good concordance with clinician-collected samples for HrHPV detection. However, only self-collected vaginal samples demonstrate comparable sensitivity and specificity for detecting HSIL.