Assessment of self taken swabs versus clinician taken swab cultures for diagnosing gonorrhoea in women: single centre, diagnostic accuracy study

OBJECTIVE To compare gonorrhoea detection by self taken vulvovaginal swabs (tested with nucleic acid amplification tests) with the culture of urethral and endocervical samples taken by clinicians. DESIGN Prospective study of diagnostic accuracy. SETTING 1 sexual health clinic in an urban setting (Leeds Centre for Sexual Health, United Kingdom), between March 2009 and January 2010. PARTICIPANTS Women aged 16 years or older, attending the clinic for sexually transmitted infection (STI) testing and consenting to perform a vulvovaginal swab themselves before routine examination. During examination, clinicians took urethral and endocervical samples for culture and an endocervical swab for nucleic acid amplification testing. INTERVENTIONS Urethra and endocervix samples were analysed by gonococcal culture. Vulvovaginal swabs and endocervical swabs were analysed by the Aptima Combo 2 (AC2) assay; positive results from this assay were confirmed with a second nucleic acid amplification test. MAIN OUTCOME MEASURES Positive confirmation of gonorrhoea. RESULTS Of 3859 women with complete data and test results, 96 (2.5%) were infected with gonorrhoea (overall test sensitivities: culture 81%, endocervical swabs with AC2 96%, vulvovaginal swabs with AC2 99%). The AC2 assays were more sensitive than culture (P<0.001), but the endocervical and vulvovaginal assays did not differ significantly (P=0.375). Specificity of all Aptima Combo 2 tests was 100%. Of 1625 women who had symptoms suggestive of a bacterial STI, 56 (3.4%) had gonorrhoea (culture 84%, endocervical AC2 100%, vulvovaginal AC2 100%). The AC2 assays were more sensitive than culture (P=0.004), and the endocervical and vulvovaginal assays were equivalent to each other. Of 2234 women who did not have symptoms suggesting a bacterial STI, 40 (1.8%) had gonorrhoea (culture 78%, endocervical AC2 90%, vulvovaginal AC2 98%). The vulvovaginal swab was more sensitive than culture (P=0.008), but there was no difference between the endocervical and vulvovaginal AC2 assays (P=0.375) or between the endocervical AC2 assay and culture (P=0.125). The endocervical swab assay performed less well in women without symptoms of a bacterial STI than in those with symptoms (90% v 100%, P=0.028), whereas the vulvovaginal swab assay performed similarly (98% v 100%, P=0.42). CONCLUSION Self taken vulvovaginal swabs analysed by nucleic acid amplification tests are significantly more sensitive at detecting gonorrhoea than culture of clinician taken urethral and endocervical samples, and are equivalent to endocervical swabs analysed by nucleic acid amplification tests. Self taken vulvovaginal swabs are the sample of choice in women without symptoms and have the advantage of being non-invasive. In women who need a clinical examination, either a clinician taken or self taken vulvovaginal swab is recommended.