Family Practice Vol. 18, No. 3, 253-257
© Oxford University Press 2001
Infectious Diseases |
The effect of introduction of a guideline on the management of vaginal discharge and in particular bacterial vaginosis in primary care
Highlands Health Centre, Fore Street, Ivybridge, Plymouth PL21 0AD, UK,
a Department of Primary Health Care and General Practice, the ITTC Building, Doniford, Plymouth PL6 8BX,
b Genitourinary Medicine Department and
cPlymouth Public Health Laboratory, Doniford Hospital, Plymouth PL6 8DH, UK.
Langsford MJ, Dobbs FF, Morrison GM and Dance DAB. The effect of introduction of a guideline on the management of vaginal discharge and in particular bacterial vaginosis in primary care. Family Practice 2001; 18: 253–257.
Received 22 June 2000; Revised 12 October 2000; Accepted 8 January 2001.
 |
Abstract |
|---|
 Top Abstract IntroductionMethodResultsDiscussionReferences |
|---|
Background. Bacterial vaginosis (BV) is the commonest cause of vaginal discharge, and its association with obstetric and gynaecological complications is being recognized increasingly. It was our impression that BV was poorly understood and underdiagnosed in family practice.
Objective. The aim of this study was to explore the management of patients with vaginal symptoms by family practitioners and to see if the management changed after the assimilation of best practice guidelines.
Method. Family practitioners were invited to complete a baseline questionnaire of their perceived practice, and to record actual practice when consulted about vaginal symptoms, for a minimum of 4 weeks. Consensus best practice guidelines were then provided and practice recorded for a similar period.
Results. Baseline data was received from 34 practitioners and suggested that the symptoms and signs of different vaginal infections were not well known. Most symptomatic patients were only investigated at re-presentation with unresolved symptoms or at recurrence, and 43% of respondents treated with empirical antifungals as a first line approach. Pregnant patients were only occasionally asked about symptoms and only occasionally examined if symptomatic. Pre-guideline practice data from 30 practitioners showed 1.2 patient consultations/week, of which 60% were examined and 55% had a high vaginal swab (HVS) sent. Only 2% had near-patient tests done. Post-guideline data from 23 family practitioners showed a lower recorded consultation rate at 0.7/week, but 90% of these were examined, 77% had an HVS sent and 69% had near-patient tests done. Of the 36 HVS examined by Gram stain, 19 (53%) showed Lactobacillus predominant flora and 10 (28%) suggested BV. Seven (19%) were borderline or ungradable. Only three (8%) showed yeasts, one of which also showed BV.
Conclusions. Baseline data supported our impression that BV was under-recognized. Guidelines appeared to improve the rate of investigation of women consulting with vaginal symptoms.
Keywords. Bacterial vaginosis, family practice, guideline, management.
|
Introduction |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionReferences |
|---|
Bacterial vaginosis (BV) is a polymicrobial clinical syndrome resulting from the replacement of the normal hydrogen peroxide-producing Lactobacillus species in the vagina with a mixture of anaerobic bacteria, Gardnerella vaginalis and Mycoplasma hominis, without signs of vaginal inflammation. It is the commonest cause of vaginal discharge and accounts for 40–50% of cases presenting in family practice.1
BV is being recognized increasingly as a cause of unpleasant symptoms, typically a profuse, offensive smelling discharge which increases after intercourse, and also as being associated with a variety of important gynaecological and obstetric complications, e.g. post-abortion pelvic inflammatory disease,2 post-hysterectomy vaginal cuff infection, mid-trimester miscarriage and pre-term delivery.3
Classically, the diagnosis of BV has been a clinical one at the bedside on the basis of the presence of three out of four of Amsel's criteria.4 These criteria are the presence of a thin, greyish/white homogeneous discharge, a pH of vaginal secretions >4.5, a positive amine whiff test (i.e. the release of a ‘high cheese’ or fishy odour on mixing vaginal secretions with 10% potassium hydroxide) and the presence of microscopically diagnosed clue cells on a saline wet mount.
Increasingly, diagnosis has been by Gram stain using scoring techniques to grade the change from Lactobacillus species dominant flora to replacement with mixed organisms, mostly small Gram-variable rods.5,6 This study was set in South West England, and the local standard laboratory method for investigation of vaginal discharge included the examination of a high vaginal swab (HVS) for clue cells, having recently changed from culture for Gardnerella vaginalis. Plymouth Public Health Laboratory (PHL) had received many telephone calls from family practitioners indicating gaps in knowledge about Amsel's criteria, and the significance of clue cells when reported.
|
Method |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionReferences |
|---|
The study aimed to examine family practitioners' management of vaginal discharge, before and after introduction of a guideline. All 500 family practitioners in the South & West Devon Health Authority area were invited to participate. Fifty-five practitioners expressed an interest, and baseline questionnaires of current perceived practice were obtained from 34 (29 family practitioners and five practice nurses who had been delegated to take part by their family practitioner colleagues).
These practitioners were then asked to record their actual practice for a minimum of 4 weeks when consulted by a woman (over 14 years old) presenting with vaginal symptoms of discharge, itch or soreness. They were asked to record for each consultation whether the patient was examined, any near-patient tests done and laboratory tests requested.
Participants were then provided with some brief facts about BV, a patient advice leaflet comparing BV and ‘thrush’, and a consensus ‘best practice’ guideline flow chart produced jointly by family practitioners, a genito-urinary medicine consultant and a consultant microbiologist after careful examination of the research evidence. They were provided with narrow range (pH 3–6) Whatman pH paper, and a plastic dropper bottle of 10% potassium hydroxide solution.
The guideline (see Figure 1
) suggested that presenting patients should have a history taken including characteristics of the discharge and that they be examined with a speculum unless they declined or had a typical recurrence of a previously fully investigated and diagnosed problem. It was suggested that the appearance of any discharge be noted, that an HVS be taken and rolled on to a glass slide for later Gram stain. The swab was to be touched on to Whatman pH paper and the pH noted, and then the tip placed into a Universal container, 10% potassium hydroxide solution added and any resulting odour noted.
|
Participants were not asked to look for clue cells as very few would have had access to a microscope or the expertise needed. It was suggested that BV be suspected if two out of the three near-patient tests performed were positive.
After assimilating the guidelines, the practitioners were asked to record their actual practice for a further minimum of 4 weeks. They were again asked to record, for each patient consulting, whether they were examined, what investigations were done and also the results of any near-patient tests performed. They were asked to send the glass slides under separate cover to the investigator (ML) who Gram stained them and examined them blind, with a Leitz Dialux microscope at 1300x magnification in the Genito-Urinary Medicine Department. The Gram-stained flora was graded according to Hay's criteria7 and the participants were informed of the result. The slides were then re-examined blind by a senior Medical Laboratory Scientific Officer in Plymouth PHL and subsequently reviewed by both microscopists to attempt to resolve discrepancies.
|
Results |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionReferences |
|---|
Baseline data were received from 34 practitioners—29 family practitioners (15 male and 14 female) and five practice nurses, who were delegated by their family practitioner colleagues as being the member of that team most likely to be consulted about vaginal symptoms. Participating practices included both inner city and rural situations.
Participants were asked how many patients they thought consulted them per month about vaginal symptoms. The replies varied widely from <1/month to >10/month, the median being 5/month. The number of HVS they thought they sent was slightly less, the most common reply being 1–5 per month.
Table 1 shows how often respondents considered they examined (both non-pregnant and antenatal) symptomatic women with a speculum, and at what stage they examined women with recurrent symptoms. The frequency of questioning of antenatal patients about vaginal symptoms is also shown.
|
Family practitioners were asked how they most commonly treated vaginal symptoms of discharge, itch or soreness, and 43% responded that empirical antifungals were their first line approach (see Figure 2
).
|
Participants were asked which of a list of eight symptoms would make them suspicious of a patient's self-diagnosis of ‘thrush’ to try and gauge the level of awareness of the symptoms, signs and associated risk factors for BV. Fifty-nine percent of respondents recognized that a smelly discharge should make them suspicious, but only 50% were suspicious of a watery discharge. Conversely, 56% were suspicious that itch and soreness may not mean thrush! Worryingly, only 41% of practitioners felt that a recent change of partner should make them think twice about accepting the patient's self-diagnosis of thrush.
Pre-guideline actual practice data were obtained from 30 practitioners, and a total of 205 weeks were surveyed. The consultation rate was 4.9 per 4-week period, with a range of 0.4–10.4, i.e. very similar to the perceived consultation level.
Table 2 demonstrates the number of consultations, number examined, HVS sent and near-patient tests performed. Near-patient tests recorded included dipstick urine tests and noting the type of discharge; one participant noted vaginal odour.
|
The rate of examination was very variable between participants, from 0 to 100%, with three family practitioners examining fewer than 10% of patients. It was noted that some patients not examined by the family practitioner had an HVS performed, presumably by the practice nurse. The number of HVS sent per 4-week period (2.7) was less than expected from the baseline questionnaire.
Post-guideline data were only obtained from 23 participants (77% of those recording in the pre-guideline period). Table 2
shows the data in comparison with the pre-guideline figures. It was noted that the number of patients seen was apparently far less than in the pre-guideline period, but a higher percentage were examined. Overall, the number of patients per week who were examined remained very similar.
Near-patient tests were the one item that was changed dramatically: 30 (51%) of patients had all three suggested near-patient tests recorded.
Thirty-six slides were Gram stained and examined by both investigators. Of these, 10 (28%) were suggestive of BV on Hay grading as agreed by both examiners. Nineteen (53%) had Lactobacillus species predominant flora, three (8%) were borderline on Hay grade and four (11%) were ungradable due to very scanty bacteria on the slide, a flora which did not correspond to any of the Hay grades, or lack of agreement between the examiners. Only three slides (8%) showed yeasts, and one of these was also suggestive of BV.
|
Discussion |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionReferences |
|---|
The low participation rate in this study (only 6.8% of those invited) suggests a lack of interest amongst family practitioners in the ‘Cinderella’ symptom of vaginal discharge.
The baseline data support our view that women's complaints of vaginal symptoms are still being managed inadequately, as found by O'Dowd et al. in their family practice study in 1996.8 The answers to the question about the symptoms which make a candidal infection unlikely revealed gaps in the respondents' knowledge of the typical symptoms and signs of common vaginal infections. A high proportion of family practitioners treated women with vaginal symptoms with empirical antifungals with no investigation, and only half of the practitioners investigated women re-presenting with unresolved symptoms. Together with the gaps in knowledge, this suggests that most women with BV are not being treated appropriately. The survey of pre-guideline actual practice tends to support these views.
With the increasing evidence of the importance of BV in pregnancy, it is worrying that only a small minority of practitioners either questioned pregnant women about vaginal discharge or considered that a complaint of a discharge in an antenatal patient warranted investigation. Our guideline did not specifically mention antenatal patients, although it would apply equally to symptomatic pregnant women. Specific advice both on the management of pregnant women with vaginal symptoms and on screening for BV in pregnancy should be made available to family practitioners if the obstetric complications of this condition are to be avoided.
It was disappointing that so many participants dropped out of the study after receiving the guidelines. They were followed up actively but some reported that the work involved with the near-patient tests appeared too onerous and others that life was just too busy. The post-guideline study period fell in the winter, and the ravages of an influenza outbreak may have affected both the participants' time and the consultation pattern.
The apparent difference in consultation rates in the two phases of the study may have been partly seasonal. One practitioner included telephone requests for antifungals in both phases of the study, and noted eight in 12 weeks pre-guideline data but none in 11 weeks post-guideline recording, despite doing a computer search for prescriptions of relevant antifungals. Hurley et al. studied 8700 patients over 5 years and unexpectedly found that yeast vulvo-vaginitis was commoner in the spring than in the winter,9 but Odds et al. found no statistically significant variations in the prevalence of vaginal yeasts with the season of the year.10 The recorded diagnoses from KC60 data (infectious disease returns) for the Plymouth Genito-Urinary Medicine Clinic show no seasonal changes in Candida or bacterial vaginosis levels (personal communication).
The Gram stain results were sent back to the participants and, since 10 cases of probable BV were diagnosed, we hope that awareness of the condition and the level of knowledge have been increased. The participating family practitioners have been empowered to manage women with vaginal symptoms more appropriately and successfully.
|
Acknowledgments |
|---|
This study was approved by the Plymouth Local Research Ethics Committee and carried out with support of members of the SWARM Research Network. Financial support was obtained from 3M Healthcare.
|
References |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionReferences |
|---|
1 Sobel JD. Vaginitis. N Engl J Med 1997; 337: 1896–1903.
2 Blackwell AL, Thomas PD, Wareham K, Emery SJ. Health gains from screening for infection of the lower genital tract in women attending for termination of pregnancy. Lancet 1993; 342: 206–209.[Web of Science][Medline]
3
Hay PE, Lamont RF, Taylor-Robinson D, Morgan DJ, Ison C, Pearson J. Abnormal bacterial colonisation of the lower genital tract and subsequent preterm delivery and late miscarriage. Br Med J 1994; 308: 295–298.
4 Amsel R, Totten PA, Spiegel CA, Chen KC, Eschenbach D, Holmes KK. Non-specific vaginitis. Diagnostic criteria and microbiological and epidemiologic associations. Am J Med 1983; 74: 14–22.[Web of Science][Medline]
5
Nugent RP, Krohn MA, Hillier SL. Reliability of diagnosing bacterial vaginosis is improved by a standardised method of Gram stain interpretation. J Clin Microbiol 1991; 29: 297–301.
6 Schwebke JR, Hillier SL, Sobel JD, McGregor JA, Sweet RL. Validity of the vaginal Gram stain for the diagnosis of bacterial vaginosis. Obstet Gynaecol 1996; 88: 573–576.[Web of Science][Medline]
7 Hay PE, Morgan DJ, Ison CA et al. A longitudinal study of bacterial vaginosis during pregnancy. Br J Obstet Gynaecol 1994; 101: 1048–1053.[Web of Science][Medline]
8 O'Dowd TC, Parker S, Kelly A. Women's experiences of general practitioner management of their vaginal symptoms. Br J Gen Pract 1996; 46: 415–418.[Web of Science][Medline]
9 Hurley R, Leask BGS, Faktor JA, Ivy de Fonseka C. Incidence and distribution of yeast species and of Trichomonas vaginalis in the vagina of pregnant women. J Obstet Gynaecol 1973; 80: 252–257.
10 Odds FC, Webster CE, Riley VC, Fisk PG. Epidemiology of vaginal Candida infection: significance of numbers of vaginal yeasts and their biotypes. Eur J Obstet Gynaecol Reprod Biol 1987; 25: 53–66.[Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  I. Hilmarsdottir, G. S. Hauksdottir, J. D. Johannesdottir, T. Danielsdottir, H. Thorsteinsdottir, and J. H. Olafsson Evaluation of a rapid gram stain interpretation method for diagnosis of bacterial vaginosis. J. Clin. Microbiol., March 1, 2006; 44(3): 1139 - 1140. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||