Predictive value of the clinical diagnosis of lower genital tract infection in women

Objective: We hypothesized that diagnostic approaches to lower genital tract infections are inaccurate and proposed this study to evaluate typical approaches. Study design: Clinical diagnoses were made with symptoms, direct observation, wet mount, vaginal pH, and amines in 598 women with genital complaints. Laboratory testing for N gonorrhoeae, yeast, T vaginalis, C trachomatis, and bacterial vaginosis by Gram stain. Results: The most frequent symptoms were vaginal discharge (64%), change in discharge (53%), malodor (48%), and pruritis (32%). The infection rates were 46% bacterial vaginosis, 29% yeast, 12% trichomoniasis, 11% chlamydia or gonorrhea; 21% of the patients had no infection. The symptoms did not predict laboratory diagnosis. Clinical signs and symptoms with office-based tests and microscopy improved the accuracy of diagnoses. Amsel's clinical diagnosis of bacterial vaginosis was the most sensitive at 92%. The sensitivity of wet mount diagnosis of trichomoniasis was 62%, of yeast by microscopy was 22%, and of mucopus for the prediction of gonorrhea and/or chlamydia was 30%. Conclusion: Symptoms alone should not be used to direct treatment in instances in which resources permit more complete evaluation with office-based testing that includes microscopy. Treatment failures or diagnostic uncertainty should prompt specific laboratory testing. (C) 2004 Elsevier Inc. All rights reserved.