Diagnosis of bacterial vaginosis: Clinical or microscopic? A cross-sectional study.

OBJECTIVE To compare the performance of the Amsel criteria, culture of Gardnerella spp., and wet mount microscopy (WMM) in the diagnosis of bacterial vaginosis. METHODS A cross-sectional cohort study was conducted, consisting of evaluating 749 consecutive women, regardless of symptoms. The Amsel criteria were evaluated, WMM microscopy and Gram staining were performed, and a swab was collected for culture. The gold standard for diagnosis was the Nugent score. RESULTS The sensitivity and specificity for the different approaches were: Amsel criteria 41.3% (95% confidence interval [CI] 33.76-49.18) and 97.8% (95% CI 96.21-98.81); cultures 59.9% (95% CI 51.47-67.85) and 99.0% (95% CI 97.76-99.69); WMM 82.6% (95% CI 76.02-88.05) and 92.4% (95% CI 89.98-94.45), respectively. WMM performed equally well in symptomatic and asymptomatic women. Amsel criteria in scenarios where there is no use of a microscope had very poor sensitivity (22.8% [CI 16.63-29.87%]). CONCLUSION The Amsel criteria have a poor performance for the diagnosis of bacterial vaginosis. WMM performs well but is not routinely used and should be the first approach for the diagnosis of vaginitis.