Literature review > Issue_4 > Review on Taylor-Robinson et al. 

The use of scoring systems for interpreting Gram-stained vaginal smears has allowed recognition of a continuum between normal vaginal flora and bacterial vaginosis, as opposed to the 'all or nothing' diagnosis provided by the standard Amsel criteria. This is useful when studying the relationship between BV and its associated complications. The Nugent system grades the severity of BV between 7 and 10 as well as recognising intermediate grades scoring 4 - 6 [1]. The Hay-Ison criteria that were developed initially in a study of pregnant women just defined normal, BV, and intermediate grades [2]. This system was updated to include a fourth grade of flora, dominated by Gram-positive flora and to recognise the lack of bacteria sometimes seen after broad-spectrum antibiotics (grade 0) [3]. Both scoring systems performed well in an international workshop evaluating diagnostic methods for BV [4].

Whilst the intermediate grade is recognisable on Gram stain, how does it correlate with symptoms and signs of BV? Hillier and colleagues evaluated the intermediate pattern in 7918 pregnant women at 23 to 26 weeks' gestation [5]. They reported that women with intermediate flora were more likely to have some of the Amsel criteria positive than were women with normal flora, but fewer than women with BV. Women with intermediate flora or BV were more likely to have chlamydia or gonorrhoea than were women with normal flora.

This study reports on the relationship between intermediate flora using the Hay-Ison criteria and the symptom of offensive smell, as well as with Amsel criteria. It is part of an RCT of clindamycin cream to treat BV in pregnancy. Vaginal smears were obtained from women booking at the antenatal clinic at a gestational age of 12 - 16 weeks. Data from two follow-up evaluations later in pregnancy were included. Studying pregnant women is one limitation of the study. Abnormal discharge may not be so easily distinguished from physiological discharge, affecting the Amsel criteria. Thus, the specificity of abnormal discharge was poor, with the sign identified in 32% of women who did not have BV on either Amsel criteria or Gram stain. In contrast, an elevated pH (> 4.5) was identified in 9.6%, clue cells in 4.3%, and positive amine test in 0.4% of women with normal flora. I am always concerned about inter-observer error in performing the Amsel criteria and the paper does not state who performed the tests, although the majority were done by two principal researchers (J. Morgan, personal communication), who also read the Gram stains.

The comparison between Gram stain and Amsel criteria is summarised in the table.

The main differences were in intermediate (grade II) flora and Amsel. Thus, 11 (11.7%) of 94 women with grade II flora had four Amsel criteria positive; 24 (25.5%) had three; 18 (19.2%) had two and 28 (29.8%) had one. It is inevitable that diagnostic methods will produce different results close to the cut off between BV and intermediate and such variation should not interfere with the usefulness of Gram staining for either research or clinical care. Since the vaginal flora can be in a dynamic state, changing day by day, symptomatic women should be treated if a Gram stain shows either BV or intermediate flora; asymptomatic women should not. In terms of symptoms, 53% of women with BV on both Gram stain and Amsel criteria reported a malodorous discharge compared to 22% of those positive on Amsel criteria, but intermediate on Gram stain, and 16.6% of those with two Amsel criteria positive and intermediate on Gram stain. Secondary findings of the paper confirm previous findings: BV and intermediate flora can revert to normal in women treated with placebo; and clindamycin cream was less successful in clearing intermediate flora than BV, although the numbers of women studied are small [6].

Gram stain has several advantages over the Amsel criteria for research studies. Of particular value is the ability to store slides and review them subsequently for quality control, as well as speed and simplicity of collecting samples in a clinical environment. If future studies confirm benefit from screening and treating women for BV to improve pregnancy outcome, we will need a simple, reliable, inexpensive near-patient test to allow optimal treatment early in pregnancy.

References:

1.  Nugent RP, Krohn MA, Hillier SL: Reliability of diagnosing bacterial vaginosis is improved by a standardized method of gram stain interpretation. J Clin Microbiol 1991, 29:297-301.

2.  Hay PE, Lamont RF, Taylor-Robinson D, Morgan DJ, Ison C, Pearson J: Abnormal bacterial colonisation of the genital tract and subsequent preterm delivery and late miscarriage. Br Med J 1994, 308:295-298.

3.  Ison CA, Hay PE: Validation of a simplified grading of Gram stained vaginal smears for use in genitourinary medicine clinics. Sex Transm.Infect. 2002, 78:413-415.

4.  Forsum U, Jakobsson T, Larsson PG, Schmidt H, Beverly A, Bjornerem A, Carlsson B, Csango P, Donders G, Hay P, Ison C, Keane F, McDonald H, Moi H, Platz-Christensen JJ, Schwebke J: An international study of the interobserver variation between interpretations of vaginal smear criteria of bacterial vaginosis. APMIS 2002, 110:811-818.

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

6.  Rosenstein IJ, Morgan DJ, Lamont RF, Sheehan M, Dore CJ, Hay PE, Taylor-Robinson D: Effect of intravaginal clindamycin cream on pregnancy outcome and on abnormal vaginal microbial flora of pregnant women. Infect Dis Obstet Gynecol 2000, 8:158-165

about SDI | newsletters | grants | publications | literature reviews

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