Evaluation of two clinical protocols for the management of women with vaginal discharge in southern Thailand

OBJECTIVES: (1) To compare the effectiveness of two clinical protocols for the management of vaginal discharge in the situations where no laboratory facilities are available but speculum examination is possible and where basic laboratory facilities are available. (2) To determine clinical and simple laboratory indicators for diagnosis of patients with vaginal discharge in the local setting. DESIGN: Alternate allocation of subjects to one of two management protocols. SUBJECTS: Women presenting to university gynaecology outpatients department with a complaint of vaginal discharge. METHODS: Subjects were alternately allocated management according to one of two protocols: one without (group A) and one with (group B) immediate access to results of basic laboratory tests. Full clinical assessment including speculum examination and microbiological assessment for infection with gonorrhoea, chlamydia, candida, trichomonas, and bacterial vaginosis was performed on all women. Follow up assessment of clinical and microbiological response was performed 1-2 weeks later. RESULTS: At initial assessment, both groups were similar in all respects except that more group B women had inflammation of the vulva. The prevalences of various conditions were: candidiasis 22%, bacterial vaginosis 38%, trichomoniasis 4%, chlamydia 4%, gonorrhoea 0.4%. There was no association between any demographic characteristic and diagnosis of cause of the discharge. Both protocols resulted in clinically and statistically significant improvements for women with candidiasis, bacterial vaginosis, and trichomoniasis. There were no clinically important differences in outcomes between the two protocols. The sensitivities and specificities of various indicators were: curd-like vaginal discharge for candidiasis, 72% and 100%; homogeneous vaginal discharge for bacterial vaginosis or trichomoniasis, 94% and 88%; absent or scanty lactobacilli for bacterial vaginosis, 99% and 68%; > 20% clue cells for bacterial vaginosis, 81% and 99%; visible endocervical mucopus for chlamydia or gonorrhoea, 36% and 86%; microscopic endocervical mucopus for chlamydia or gonorrhoea, 64% and 69%. CONCLUSIONS: Both protocols were equally effective in managing women with abnormal vaginal discharge. Simple clinical indicators for candidiasis, bacterial vaginosis, or trichomonas as in protocol A are sufficiently sensitive and specific for use in situations with no laboratory support. A modification to protocol A could increase detection of bacterial vaginosis at basic health service level. Further work is needed to identify appropriate indicators for infection with chlamydia or gonorrhoea.