Comparison of self‐sampling versus clinician‐sampling HPV DNA cervical cancer screening in South African women living with and without HIV

INTRODUCTION: Human papilloma virus (HPV) testing is the preferred primary screening method for cervical cancer, and HPV self-sampling has the potential to improve screening uptake. We compared the test performance and agreement of self- and clinician-sampling HPV in women living with HIV (WLWH) and HIV-negative women in Gauteng and Western Cape Province, South Africa. METHODS: Cervical cancer screening was performed in previously unscreened women using cytology, self- and clinician-collected HPV (Hybrid Capture 2 assay). Screen-positive participants underwent colposcopy and biopsy, and a random sample of negative screens had punch biopsy regardless of colposcopy findings. Performances of screening tests were evaluated against histology results adjusted for verification bias. The kappa statistic for HPV concordance between self- and clinician-collected HPV was also calculated. RESULTS: Included for analysis were 903 women with a mean age of 41.5 years, of which 43.3% were WLWH. In WLWH/HIV-negative women, the sensitivity of self-sampling HPV was 74.7%/67.4% and of clinician-sampling HPV was 77.1%/59.2% to detect cervical intraepithelial neoplasia (CIN) grade3+, whereas the specificity for clinician-sampling HPV was 75.0%/86.8% compared with 70.3%/79.8% for self-sampling HPV to detect CIN2+. In both HIV groups, Cohen's kappa agreement between the two sampling methods was good (0.63 and 0.62, respectively). CONCLUSION: Self-collected HPV sampling is an acceptable and viable option for primary cervical cancer screening in all women given the superior sensitivity to cytology. The statistically significant difference in HPV test performance among HIV-negative women according to sampling method confirms the benefit of a further triage test when self-sampling is used for HPV-base cervical screening, especially in HIV-negative women.