The use of urine for cervical cancer screening is gaining international attention, although more data on the relative clinical accuracy of validated human papillomavirus (HPV) DNA tests on urine versus cervical samples are needed. This study primarily seeks to evaluate the clinical performance of Roche cobas 4800 and 6800 HPV Systems in first-void urine, collected at home, compared with clinician-collected cervical samples. Paired first-void urine (index test) and cervical samples (comparator test) from 499 females enrolled at five Belgian colposcopy clinics were analyzed with cobas HPV Systems. Colposcopy and histology of biopsies were used as reference test (trial registration number: NCT03064087). Sample processing protocols and clinical thresholds proposed by the manufacturer for cervical samples were also applied for first-void urine. In the total study population, HPV testing on first-void urine was similarly sensitive [ratio, 0.98; 95% CI, 0.93-1.02] and specific for cobas 4800 HPV (ratio, 1.00; 95% CI, 0.91-1.10) and cobas HPV for use on the cobas 6800 System (ratio, 0.96; 95% CI, 0.91-1.02; ratio, 1.01; 95% CI, 0.93-1.09) compared with cervical samples (P ≥ 0.05). Good to excellent HPV test agreements between paired samples were observed (κ = 0.68 to 0.87). In summary, HPV testing using cobas 4800 and 6800 HPV Systems was as accurate on first-void urine as on cervical samples collected by a clinician.
