Improving Gram-stained reproducible result by further adding clue cells in diagnosing bacterial vaginosis.

The reproducibility of interpretation in diagnosing bacterial vaginosis may be enhanced by adding pus cells and clue cells into two different criteria, developed by Spiegel et al. and Nugent et al. The purpose of study was designed to find out which parameter was more reproducible. 100 patients were collected with the diagnosis of bacterial vaginosis as an experimental group, while the other 100 patients who were with routine Papanicolaou smears in gynecologic clinic the collected as a control group. Two slides, including the original and reproducible ones, were obtained from vaginal smears for each patient. Three technicians read the slides randomly by using two different criteria, plus pus cells and clue cells. This showed the agreement for clue cells is the best method regardless of experimental group or control group (Kappa values between 0.708 and 1.000). The intra-observer agreement for the diagnosis of bacterial vaginosis by the method of Nugent et al. is superior to the method of Spiegel et al. Our data show the comparison of Amsel criteria versus Nugent criteria for the diagnosis of bacterial vaginosis with sensitivity of 88.9%, specificity of 55.4%, negative positive value of 62.1%, and positive predictive value of 85.8%. Moreover, our data also demonstrate the comparison of Amsel criteria versus the diagnosis either based on Nugent criteria or the presence of clue cells with sensitivity of 95.7%, specificity of 56.7%, negative positive value of 81.2%, and positive predictive value of 87.1%. The results demonstrate further adding score of the clue cells can enhance the reproducible diagnosis of bacterial vaginosis, which is superior to the methods of Nugent et al. and Spiegel et al.