Literature review > Issue_3 > Review Ison et al. 

In 1892 Doderlein (1) described three grades of "vaginal cleanliness" based on microscopic examination of unstained vaginal fluid. Grade I indicates a "clean vagina" where lactobacilli predominate. Grade III is Doderlein's "pathological flora" where the discharge is described as "profuse and purulent or rather scanty and watery," and therefore appears to describe patients with bacterial vaginosis (BV) or trichomoniasis. Grade II is intermediate between Grades I and III and is associated with a mixed flora including fewer lactobacilli than seen in Grade I.

Numerous methods for microscopic diagnosis of BV have been described since that time. Dunkelberg (2) used the Gram stain to diagnose BV by the detection of clue cells. Spiegel et al. (3) described a normal vaginal Gram stain as one containing the Lactobacillus morphotype with or without the Gardnerella morphotype, and BV as the presence of the Gardnerella morphotype plus other organisms with a reduced quantity of lactobacilli. Nugent et al. (4) reported that the intercenter reproducibility of the Spiegel criteria was relatively low (r=0.61), although this suboptimal value may have been due at least in part to use of different Gram stain method: the Kopeloff modification with a basic fuchsin counterstain was not used in the former case. Nugent et al. developed a 10-point scoring system and divided scores into normal, intermediate and BV and they reported improved intercenter reproducibility (r =0.82). Kappa statistics in other studies were 0.62 - 0.77 and 0.6 - 0.72 (5-6). Therefore, the Nugent criteria performed as well as or slightly better than the Spiegel criteria.

In the paper under review, Ison and Hay sought to validate what they described as a simplified grading system for Gram stain diagnosis of BV. Subjects were participants in a treatment trial and all were women with BV as diagnosed by the Amsel clinical criteria (7). Sixteen specimens were from initial visits and the remainder were from post treatment follow-up visits, that is, all but 16 specimens were from women who had received either metronidazole gel or clindamycin cream for treatment of BV within the prior 35 days. Generally women who have been on antibiotic within the prior 30 days are excluded from studies such as this. It would have been better to sample women with a normal exam and no recent antibiotic use.

Similar to Doderlein's grades, Ison and Hay describe the following: Grade I (normal flora) lactobacillus morphotype only, Grade II (intermediate flora) reduced lactobacillus morphotype with mixed bacterial morphotypes; Grade III (BV), mixed bacterial morphotypes with few or absent lactobacillus morphotypes. One person read all of the slides. Therefore, reproducibility of readings by multiple microscopists was not evaluated.

Ison and Hay described two additional Gram stain categories. Grade 0 was used to indicate smears with no bacteria and Grade IV those with only gram-positive cocci attached to epithelial cells. Both of these patterns may be most common in women who had recently received antibiotic therapy. It would have been useful to know at which visit Grade 0 and IV smears were seen.

Thirteen specimens were Grade II, intermediate, and seven of the patients met one or more clinical criteria for BV. Presumably these were all post-treatment samples. Again, the number of days post treatment would be of interest. Ison and Hay stated that most women with Grade II were clinically normal, however these were women who had recently received antibiotics. Hillier et al. (8) showed that most pregnant women with an intermediate Nugent score (4-6) in second trimester became normal (30%) or developed BV (32%) by their third trimester. Ison and Hay presented no follow-up data in their paper.

Because there was no control group, i.e., women without BV who received no antibiotic treatment within 30 days, this is not a standard validation. In addition, Nugent et al. identified the primary issue in Gram stain interpretation as intrareader reproducibility. Because one person read all of the slides in this study, that issue was not addressed in this paper.

Another article published at about the same time as the Ison and Hay article reported intraobserver variation in microscopic diagnosis of BV (9). Smears collected at several international locations were shared among investigators and the local method for diagnosis of BV was used. The kappa scores among the six investigators who used the Hay/Ison protocol range from 0.74 to 0.89.

A major source of variation in Gram stain interpretation pointed out in the Forsum article is consistent identification of the lactobacillus morphotype (9). Dunkelberg had previously noted this issue. In my experience the main confounder in interreader agreement is identification of gram-positive rods that may be small lactobacilli, gardnerella, corynebacteria or bifidobacteria. All of the microscopic methods would benefit from clarification of the identity of these organisms.

References:

1. Doderlein A: Das Scbeidensekret und seine Beddeutung far das Puerperalfieber Leipzid: Durr, 1982.

2. Dunkelberg WE: Diagnosis of Haemophilus vaginalis vaginitis by Gram-stained smears. Am J Obstet Gynecol 91:998-1000, 1965.

3. Spiegel CA, Amssel R, Holmes KK. Diagnosis of bacterial vaginosis by direct gram stain of vaginal fluid. J Clin Microbiol 18:170-177, 1983.

4. Nugent RP, Krohn ma AND hillier SL. Reliability of diagnosing bacterial vaginosis is improved by a standardized method of gram stain interpretation. J Clin Microbiol 29:2970391, 1991.

5. Joesoef MR, Hillier SL, Josodiwondo S and Linnan M. Reproducibility of a scoring system for Gram stain diagnosis of bacterial vaginosis. J Clin Microbiol 29:1730-1731, 1991.

6. Zarakolu P, Hodoglugil NNS, Aydin F, et al. Reliability of interpretation of gram-stained vaginal smears by Nugent's scoring system for diagnosis of bacterial vaginosis. Diagn Microbiol Infect Dis 48:77-80, 2004.

7. Amsel R, Totten PA, Spiegel CA et al. Nonspecific vaginitis. Diagnostic criteria and microbial and epidemiologic association. Am J Med 74:14-22, 1983.

8. Hillier SL, Krohn MA, Nugent RP, Gibbs RS. Characteristics of three vaginal flora patterns assessed by gram stain among pregnant women. Am J Obstet Gynecol 166:938-944, 1992.

9. Forsum U, Jakobsson T, Larsson PG, et al. An international study of the intraobserver variation between interpretations of vaginal smear criteria of bacterial vaginosis. APMIS 110:811-818, 2002.

about SDI | newsletters | grants | publications | literature reviews

WHO Home - WHO Search - TDR Home - SDI Home - SDI Contact us
(c) WHO/OMS 2001