Abstract
Point-of-care microscopy is the gold standard for the diagnosis of vaginal discharge in genitourinary (GU) medicine clinics but not used in primary care settings and reproductive health clinics to which many patients present. In our GU medicine clinic setting, we conducted an audit to assess the utility of microscopy of vaginal secretions versus clinical diagnosis alone for the differential diagnosis of uncomplicated lower vaginal infections. Clinical diagnosis (including pH) of bacterial vaginosis had a sensitivity between 85% and 88% at two clinic sites. Our results suggest that it may be safe and more cost-effective to restrict vaginal microscopy to a subgroup of women presenting with vaginal discharge.
Introduction
Vulvo-vaginal candidiasis (VVC) and bacterial vaginosis (BV) alone represent 35% of diagnoses made in women attending genitourinary (GU) medicine clinics. 1 Given the current recommendations to provide an integrated contraception and GU medicine service 2 alongside pressure to increase capacity and tackle the increasing rates of sexually transmitted infections (STIs), we investigated the usefulness of point-of-care microscopy in the management of women presenting with uncomplicated vaginal discharge.
Methods
We conducted a prospective audit in two central London GU medicine clinics (Clinic 1 and Clinic 2) over a three-month period. All women presenting with uncomplicated vaginal discharge who were offered investigations according to clinical protocol were enrolled. This includes microscopy of vaginal secretions, Gram stain and wet-mount microscopy, pH measurement, culture for Candida species and Trichomonas vaginalis (TV).
During the study period, the attending clinician (doctor or specialist nurse) recorded a ‘best judgement’ clinical diagnosis based on history, examination and vaginal pH measurement before receiving microscopy results.
For BV, we defined as gold standard, the presence of three out of four Amsel criteria 3 (vaginal discharge, pH > 4.5 and clue cells on Gram stain) and as ‘best judgement’ clinical diagnosis the presence of abnormal vaginal discharge together with pH > 4.5. For VVC, we defined as microscopy gold standard the presence of hyphae or spores on Gram stain of vaginal secretions and as best judgement clinical diagnosis the presence of thick white discharge and vulvo-vaginitis. These data were used to calculate the sensitivity and specificity of clinical diagnosis alone versus combined clinical diagnosis and same-day microscopy for each condition in both the clinics. A more detailed analysis of all BV clinical diagnoses including notes review was carried out in Clinic 1. Owing to time constraints this was not possible in Clinic 2.
Results
We analysed 391 patients attending the clinic with vaginal discharge (280 in Clinic 1 and 111 in Clinic 2). Patients were excluded from analysis for BV if pH was not recorded, leaving 349 (89%) patients, with 241 (86%) in Clinic 1 and 108 (97%) in Clinic 2. The prevalence of BV by gold standard criteria was 21% (50/241) in Clinic 1 and 16% (17/108) in Clinic 2 and VVC prevalence was 15% in Clinic 1 (41/280) and 13% in Clinic 2 (14/111). TV prevalence was 1.4% (4/280) in Clinic 1 and 6.3% (7/111) in Clinic 2.
Table 1 shows sensitivities and specificities of best clinical diagnosis for diagnosing BV and candida together with the rates of missed diagnosis in both clinics.
Bacterial vaginosis (BV) and candida sensitivity, specificity and missed diagnosis data
Relative to gold standard (see text) for BV and to microscopy for candida
Further analysis of 72 women with a clinical diagnosis of BV in Clinic 1 showed that microscopy was negative for clue cells in 28 patients, but 19 of them had mixed flora (of which 11 patients were treated as BV) and three of 28 were TV-positive.
Discussion
There is good evidence that vaginal pH is the most sensitive single criterion for diagnosing BV 4 and some data have suggested that using two clinical criteria (pH and amine test) does not lead to loss of sensitivity or specificity for BV diagnosis. 5 Our data shows reasonable rates of sensitivity and specificity for best judgement clinical diagnosis of BV in two clinical settings with similar BV prevalence. We also found that negative microscopy results may not change the clinician's treatment decision (data not shown). Candida results were more difficult to interpret, given that VVC may represent simple carriage or disease and often clinical symptoms/signs are in fact due to other conditions, 6 but clinical diagnosis can always be compared with culture results.
Of the four TV cases in Clinic 1, three were given a BV clinical diagnosis and treated with metronidazole.
Conclusion
Our findings suggest that microscopy may add little to the clinical diagnosis of vaginal discharge, especially in a setting with low TV prevalence where pH is measured consistently. We do not wish to suggest that microscopy be replaced by clinical judgement alone. However, considerable time and money could be saved by restricting microscopy to certain groups, such as patients presenting with unclear history, or with recurrent or un-resolving symptoms.