255P

Return to Contents

Search CAHR Abstracts


COMPARISON OF SELF-OBTAINED AND PHYSICIAN-OBTAINED CERVICAL SWAB SAMPLES FOR DETECTION OF HUMAN PAPILLOMAVIRUS (HPV) DNA IN WOMEN INFECTED WITH HIV

F Coutlée, C Hankins, S Walmsley, D Money, K Pourreaux, P Forest, J-O Shu, Canadian Women’s HIV Study Group
Centre Hospitalier de l’Université de Montréal; Direction de la santé publique de Montréal; Department of Epidemiology and Biostatistics, McGill University; Toronto General Hospital, Toronto; Oak Tree Clinic, Vancouver

Background: HPV causes precancerous and cancerous cervical lesions in HIV-infected women. Standard HPV screening requires a gynaecological examination. Non-invasive sampling methods could facilitate screening for genital HPV infection.
Objectives: To compare self-obtained vaginal swab samples (SOVS) and cervical swab samples obtained by a physician (CSSP) for the detection and typing of HPV DNA in a cohort of HIV-infected women.
Methods: Paired swab samples obtained at the same visit were processed and tested for b-globin for specimen accuracy. b-globin-positive lysates were amplified for HPV DNA with PGMY09-PGMY11 and amplicons were typed for 27 HPV genotypes (18 high-risk and 9 low-risk) using the Line blot assay.
Results: Of 173 paired samples, one sample in each of 8 pairs tested negative for b-globin, leaving 165 pairs from 140 different women for comparison. Paired samples were obtained at 3 visits for 7 women, at 2 visits for 10 and at one visit for 123. The sensitivity and specificity of SOVS compared to CSSP for HPV DNA detection reached 97.8% (87 of 89 HPV-positive samples) and 72.4% (55 of 76 HPV-negative samples), respectively. Results from 142 (86.1%) paired samples were concordant for HPV DNA detection (Kappa value of 0.74). Considering only high-risk types, concordant results were obtained for 142 sample pairs (73 positive, 69 negative) and discordant results were obtained for 23 sample pairs (22 positive SOVS and 1 positive CSS) (Kappa of 0.73). Although a high level of concordance was obtained for HPV-16 and HPV-33 (93.7% and 100%, respectively), results for some important high risk types were less frequently concordant (58% for HPV-18, 20% for HPV-45).
Conclusions: Self-obtained vaginal samples demonstrate a high degree of sensitivity for the detection of HPV infection. The relationship between vaginal HPV detection and progression of cervical dysplasia in HIV infected women needs to be further studied.

NEXT ABSTRACT >