The interpretation of Gram-stained slides of vaginal swab specimens is
evaluated for the diagnosis of bacterial vaginosis and correlated wit
h the isolation bacteria and clinical signs : thin homogeneous vaginal
discharge, pH greater-than-or-equal-to 4.7, amine odor and presence o
f clue cells. Gram stained smears were scored following a morphotype c
lassification : (a)- Lactobacillus morphotypes (scored 4 to 0), (b)- s
mall Gram negative (G. vaginalis, scored 0 to 4), (c)- curved Gram var
iable rods (Mobiluncus morphotypes, scored 0 to 2). The scoring system
(0 to 10) was a weighted combination of these morphytypes : a + b + c
. The criterion for bacterial vaginosis was a score greater-than-or-eq
ual-to 7, a score of 0 to 6 was considered as no bacterial vaginosis.
Samples of 709 women, aged 18 to 84 (mean age =39.7) were examined. Th
e prevalence of G. vaginalis culture was 12.5 %, and Mobiluncus infect
ion occured in 1.5 %. The frequency of M hominis was diagnosed in 2.8
% of women. A score of 7 to 10 on Gram stain was observed in 9.7 % in
the population. Compared to G. vaginalis culture, Gram stained smear w
as more sensitive (0.71) than clinical signs of bacterial vaginosis (0
.46). The specificity was similar with the two methods. The assessemen
t of G. vaginalis morphotypes was highly correlated with the G. vagina
lis culture (p <0.001, OR =248). The infection of T. vaginalis or Cand
ida sp. was never associated with a score greater-than-or-equal-to 7.
The Gram-stained smear is a cheap, fast and easy method. It provides g
ood results for diagnosis of bacterial vaginosis.