The association between symptom or microscopy based diagnosis of bacterial vaginosis (BV) and response to treatment in women with recurrent BV

Abstract

Background Bacterial vaginosis (BV) is a common condition, yet accurate diagnosis remains a challenge. We explored symptom and microscopy-based diagnosis of BV and assessed the influence of these diagnostic approaches on treatment response. Methods BV diagnosis based on patient-reported symptoms, and vaginal swab gram stain microscopy at local laboratories and a central laboratory were compared for women recruited into the VITA trial in England. Multivariable analysis was used to assess the association between the method of diagnosis and symptom resolution 2 weeks after metronidazole treatment. Results 517 women presenting with vaginal discharge (470/517 [91%]) and/or malodour (440/517 [85%]) were included. The accuracy of patients’ vaginal symptoms compared to local laboratory microscopy for BV diagnosis were —discharge, sensitivity 90% and specificity 5%; malodour, sensitivity 84% and specificity 12%, and compared to a central laboratory diagnosis were —discharge, sensitivity 91% and specificity 8%; malodour, sensitivity 88% and specificity 18%. 143/204 (70%) participants reported symptom resolution after treatment and this was associated with a positive baseline local laboratory diagnosis (adjusted relative risk—aRR 1.64 [1.02 to 2.64]), but not with a positive central laboratory diagnosis (aRR 1.14 [0.95 to 1.37]). Symptom resolution occurred in 75% (83/111) of women who were symptom positive/central laboratory BV positive compared to 65% (58/89) who were symptom positive/microscopy negative. Conclusion Symptoms correlated poorly with microscopy-based BV diagnosis but two-thirds of women with a symptom positive/microscopy negative diagnosis experienced symptom resolution following metronidazole treatment. Further studies are needed to determine the optimal investigation and treatment for microscopy-negative women with typical BV symptoms.