Wet smear compared with gram stain diagnosis of bacterial vaginosis in asymptomatic pregnant women.

Abstract

To compare wet smear and Gram stain diagnoses of bacterial vaginosis among asymptomatic pregnant women. Between November 1, 1996 and December 31, 1997, asymptomatic women who initiated prenatal care in our obstetric clinics were invited to participate. Exclusion criteria included antimicrobial use within 2 weeks, cervical cerclage, vaginal bleeding, placenta previa, spermicide use, douching, or intercourse within 8 hours. Clinical diagnosis that required two of three positive criteria for bacterial vaginosis (vaginal pH, whiff test, and clue cells on wet smear) was compared with Gram stain diagnosis (Bacterial vaginosis score 7-10 by Nugent criteria). Population characteristics (n = 69) included an average (+/- standard deviation [SD]) maternal age of 27. 3 +/- 6.6 years, 26 nulliparas (38%), 28 black women (41%), 23 white women (38%), 15 Hispanic women (22%), and three Asian women (4%). The mean (+/-SD) gestational age at entry was 15.6 +/- 7.6 weeks. Twenty-seven percent (18 of 67) of the study population was diagnosed with bacterial vaginosis by definitive Gram stain. Two slides were lost or were of poor quality and not included. Using Gram stain diagnosis of bacterial vaginosis as the standard, clinical diagnosis had sensitivity of 56% (95% confidence interval [CI] 32%, 78%), a specificity of 96% (95% CI 90%, 100%), a positive predictive value of 83%, and a negative predictive value of 85%. In asymptomatic pregnant women, bacterial vaginosis can be diagnosed reliably by Gram stain.