PCR based high-risk human papillomavirus (hrHPV) testing on self- and physician-collected specimens have shown similar sensitivity to detect a high-grade cervical lesion (CIN 2+). However, limited evidence is available on the performance and acceptability of different sampling devices. We compared two self-collection methods to detect clinically significant hrHPV infection in Norway.
310 women referred to the treatment for histologically verified CIN 2+ completed self-collection using a dry brush (Evalyn Brush, Rovers, Netherlands) and a dry swab (FLOQSwabs, Coban, Italy), and filled a questionnaire. At the hospital, a physician took an additional cervical specimen (PreservCyt, Hologic, USA). Cytology specimens were blindly analysed by one cytotechnician and one pathologist. Self-specimens and a physician-specimen (reference) were analysed for the presence of 14 hrHPV DNA using three commercially available HPV assays; Anyplex (Seegene, South Korea), Cobas (Roche, USA) and Xpert HPV (Cepheid, USA). Agreement between self- and physician-specimen was assessed with kappa (κ) statistics with bootstrapped 95% confidence intervals (CIs). Absolute and relative sensitivities with 95% confidence intervals were computed using a matched-pair design.
Analyses included 251 women with matched triplets. Overall hrHPV positivity rate was 89% in the reference specimen using Anyplex, and 86% using Cobas and Xpert HPV.
Overall agreement for hrHPV positivity between self- and physician-specimens was highest at 94% using Anyplex on Evalyn (κ = 0.68, 95% CI: 0.53-0.83), and lowest at 82% using Xpert HPV on FLOQSwabs (κ = 0.47, 95% CI: 0.35-0.59). Anyplex detected 95% of the CIN2+ lesions. Corresponding sensitivities for Cobas and Xpert HPV were 93% and 94%. The ability of any hrHPV test to detect CIN2+ from a brush-specimen was similar to the reference, whereas significantly lower sensitivities were demonstrated using a swab. Both devices were well accepted, but women considered a brush easier, less painful and less uncomfortable than a swab. Generally, we did not observe any differences on perceptions on self-collection by sociodemographic status.
We observed significant device effects in detection of the hrHPV DNA and CIN2+ using three validated HPV assays. There were also differences on the acceptability of the sampling devices. If self-collection is considered as an alternative to provider-collection in national screening programmes, both hrHPV assays and sampling devices have to be validated.