Objective
The objective of the study was to evaluate the differences in vaginal culture, microscopy, and Gram stain between postmenopausal women who wear pessaries and those who do not to explain pessary-related, bothersome vaginal discharge.
Study Design
Postmenopausal women not using exogenous estrogen who had either been wearing a pessary for at least 3 months or who were undergoing their first pessary fittings were approached for enrollment. Symptoms were assessed, and vaginal fluid was collected for culture, microscopy, and Gram stain. A cross-sectional analysis was performed, comparing the new and return pessary wearers. The new pessary users were also sampled at 2 weeks, 3 months, and 6 months after fitting.
Results
Women who wore pessaries were more likely to be bothered by discharge (30.0% vs 2.1%, P < .001). They were also more likely to show microscopic evidence of vaginal inflammation and vaginitis. Prospective data showed that these changes developed during the first 2 weeks of pessary use. Aerobic and anaerobic organisms were nearly identical in women with and without bothersome vaginal discharge in the cross-sectional analysis and at all time points in the prospective analysis.
Conclusion
Pessary-related, bothersome vaginal discharge develops early and may be due to an inflammatory process in the vagina.
Pessaries have long been utilized for the treatment of urinary incontinence and pelvic organ prolapse. Studies show that 65-86% of pessary candidates are successfully fitted, and 65-89% with prolapse as well as 47-94% with stress incontinence experience symptom improvement with this therapy. There is a paucity of data on the use of pessaries or their side effects, but complaints of vaginal discharge and foul odor are common, with recent studies noting such complaints in 13.5-17.3% of pessary wearers.
Even less is known about the differences in microorganisms present in women who wear pessaries compared with those who do not. The assumption that increased vaginal discharge could represent a vaginal infection is entrenched in clinical ideology and is off-putting for patients and their physicians alike.
One study conducted in 2000 comprehensively evaluated vaginal discharge in pessary users using well-established, objective measures. By applying Nugent’s criteria to vaginal samples taken from pessary wearers and controls, the authors concluded that 32% of pessary wearers developed bacterial vaginosis (BV) compared with 10% of the control group. They also reported that most patients who used pessaries and developed BV did so within the first 6 months of treatment, although some patients did not develop it until 3 years after the pessary fitting.
Despite this study’s many strengths, considering BV in this patient population may be problematic. The disruption of vaginal flora, characteristic of BV, refers to a change from what is normally seen in reproductive-aged women. Prepubertal and postmenopausal women have decreased amounts of lactobacilli and higher vaginal pH values physiologically.
Little is known about the process by which discharge develops in pessary wearers; some assume it is infectious, but it also may be created by inflammatory mediators from a reactive process. No studies can be found that control for menopausal status among pessary users with vaginal discharge, and this is important because circulating and exogenous estrogen is known to alter the microbiologic milieu of the vagina.
The primary objective of this study was to identify, using microscopy, aerobic and anaerobic culture, and 2 well-established vaginitis scoring systems, the differences in the vaginal microenvironment between women who wear pessaries and those who do not over the first 6 months of pessary use. We hypothesized that postmenopausal women undergoing treatment with pessaries would be colonized with different vaginal microorganisms than their pessary-naïve counterparts and anticipated that pessary users would report increasing bothersome vaginal discharge with longer duration of use.
Materials and Methods
Recruitment began after approval was obtained from the Hartford Hospital Institutional Review Board. Subjects were postmenopausal women who had either worn pessaries for at least 3 months (established users, hereafter termed ESTs) or who never had worn pessaries and were presenting for pessary fitting (new users, hereafter termed NEW). Women were approached for enrollment at outpatient visits to the urogynecology practice at Hartford Hospital, a tertiary care center.
Subjects met inclusion criteria if they were postmenopausal (amenorrhea for at least 12 months), able to give consent and answer treatment-related questions, and had either worn their pessaries for at least 3 months or had never worn a pessary but were presenting for pessary fitting. Patients were excluded if they were pre- or perimenopausal (amenorrhea for <1 year), had significant cognitive deficits challenging participation, used exogenous systemic or vaginal estrogen therapy, had used a pessary for 0-3 months, or performed pessary self-maintenance. Systemic or vaginal antibiotic use within the last 2 weeks was also an exclusion criterion.
In women composing the EST group, data were collected only once. In contrast, data were collected at the first visit and 2 weeks, 3 months, and 6 months later in the NEW group. At each study visit, subjects underwent a speculum examination with vaginal sample collection. In the first visit for women in the NEW group, this examination occurred before pessary fitting. Otherwise, it occurred after the pessary was removed for cleaning. Speculums were moistened with tap water only before placement to avoid contamination of culture, Gram stain, or microscopy specimens with water-based lubricant. Next, swabs were placed in the vaginal vault, and specimens were placed immediately in anaerobic transport vials for culture (aerobic and anaerobic) and Gram stain. PH was measured using pH paper (Hydrion S/R Insta-chek; Micro Essential Laboratory, Inc, Brooklyn, NY), and a wet mount was prepared. A brief questionnaire was administered to confirm patients’ avoidance of antibiotics and estrogen use and to evaluate for the subjective complaint of bothersome vaginal discharge.
Demographic and medical data including age, height, and weight (from which body mass index was calculated), and number of years since menopause were collected for each subject. Women who had undergone a hysterectomy prior to the cessation of menstruation were asked the age at which they experienced menopausal symptoms. For pessary wearers, the duration of pessary use was recorded.
Microscopy was performed as follows: after collection, 1 swab was placed in a glass tube containing 5 drops of normal saline. Slides were prepared and examined using an office light microscope at ×40 magnification by 1 of 2 investigators (S.C. and C.M.). Five different fields were examined on each slide. Numbers of white blood cells (WBCs), squamous cells, clue cells, and parabasal cells in each field were recorded. Slides were scored for percent clue cells ([number clue cells/number squamous cells] ×100), number of WBCs (<10 or ≥10), and the presence or absence of parabasal cells. Percentage of clue cells and number of WBCs from the 5 fields examined were averaged for overall scores.
After microscopy findings were recorded, modified Amsel’s criteria (characteristic vaginal discharge; vaginal fluid pH of greater than 4.5; positive amine odor, or whiff test; and 20% or greater prevalence of clue cells on microscopy) were used to determine whether a clinical diagnosis (3 of 4 criteria present) of nonspecific vaginitis could be made. The vaginal discharge criterion was considered present in subjects who answered yes to the question, “Do you feel you have more vaginal discharge than is normal for you?”
Specimens were Gram stained and analyzed for bacterial vaginosis using Nugent’s criteria. The Nugent’s score is the sum of 3 weighted subscores, each of which quantifies a different category of organism (large Gram-positive rods, small Gram-negative to Gram-variable rods, and curved Gram-negative rods). The presence of large Gram-positive rods, such as Lactobacilli , decreases the score. Small Gram-negative and Gram-variable rods ( Gardnerella , Bacteroides , Prevotella , and Porphyromonas species) and curved Gram-negative rods ( Mobiluncus species) increase the score. A Nugent’s score of 0-3 is considered normal, 4-6 is intermediate, and 7-10 is consistent with bacterial vaginosis. Amsel’s and Nugent’s scores were compared between the EST group and the first-visit data from the NEW group.
Aerobic cultures were incubated in a 5-10% CO 2 incubator at 33-37°C for 48 hours. All isolates were identified to the genus and species level using either basic biochemical tests or, if these was unsuccessful, an automated bacterial identification system (Microscan; Dade MicroScan, Inc., Sacramento, CA). Yeast isolates were subcultured to CHROM Candida agar. Anaerobic cultures were plated in anaerobic conditions at 33-37°C for 7 days. Aerotolerance testing was performed on all isolates to determine whether the isolates were truly anaerobes. Isolates were Gram stained. Identification was performed according to the Anaerobic Culture and Identification Procedure. Semiquantitation was not performed, and the cultures were limited to those processed in a clinical, not academic, microbiology laboratory.
Microorganisms were compared between the EST group and the first-visit data of the NEW group using the aerobic and anaerobic organisms most frequently cultivated. In the NEW group, microorganisms were analyzed at each study visit according to whether they were physiological (ie, Lactobacillus ) or pathological (ie, Gardnerella , Bacteroides , Prevotella , Porphyromonas , and Mobiluncus species) vaginal species. The designation of pathophysiological is based on the association of these organisms with BV.
Baseline demographic and medical characteristics were compared using univariate analyses. The Gram stain Nugent’s score (0-10), microscopy data, pH, and questionnaire items also were compared using univariate analyses based on data type and distribution. Given that women in the EST group had been wearing their pessaries for varying periods of time, a Spearman’s rank correlation coefficient was calculated to evaluate whether there was a relationship between duration of pessary use (in months) and BV by Amsel’s or Nugent’s criteria.
The data gathered from the NEW group were analyzed by time point at which they were collected (baseline, 2 weeks, 3 months, 6 months). Culture and microscopy findings were compared between time points using χ 2 and analysis of variance calculations.
An a priori alpha level of α = 0.05 was established such that all results and/or between-group differences yielding P < .05 were deemed statistically significant. SPSS version 19.0 (2010; IBM/SPSS Inc, Chicago, IL) and JMP version 7.0.1 (2007; SAS Institute, Cary, NC) were used for all analyses.
Results
Cross-sectional analysis
Each group (EST and NEW) consisted of 50 women ( Table 1 presents baseline demographic and medical characteristics of both groups). The EST group was older and had been menopausal longer than the NEW group. Body mass index (EST, 27.0, SD, 4.6; NEW, 26.4, SD, 4.9; P = .57) and racial composition (white in 96% EST and 94% NEW subjects, P = 1.00) were similar between groups. Only 1 subject (in the EST group) was a smoker. Women in the EST group were more likely to be bothered by vaginal discharge than women in the NEW group ( Table 1 ).
| Characteristic | EST (n = 50) | NEW (n = 50) | P value |
|---|---|---|---|
| Age (mean ± SD) a (t test) | 80.2 ± 8.5 | 75.6 ± 7.9 | .005 b |
| Years since menopause, all (mean ± SD) a (t test) | 31.8 ± 9.9 | 26.2 ± 8.9 | .004 b |
| Bothered by discharge, n, % c | 15/50, 30.0% | 1/48, 2.1% | < .001 b |
| BV by Amsel’s criteria, n, % c | 4/49, 8.2% | 0/47, 0% | .05 b |
| BV by Nugent’s criteria, n, % c | 12/50, 24.0% | 4/50, 8.0% | .03 b |
| WBCs (≥10 per hpf), n, % d | 19/48, 39.6% | 2/49, 4.1% | < .001 b |
| Parabasal cells present, n, % c | 32/49, 65.3% | 18/49, 36.7% | .005 b |
| Clue cells (mean ± SD), % e | 11.2 ± 24. 6 | 1.6 ± 5.2 | .01 b |
| pH (mean ± SD) a | 6.6 ± 1.3 | 6.3 ± 0.8 | .20 |
b Statistically significant ( P < .05)
d Mann-Whitney U Fisher exact test
Because age was significantly different between the EST and NEW groups, a multivariate analysis was performed to evaluate for confounding for the presence of bothersome vaginal discharge. Variables tested included age, parity, smoking status, and group (EST vs NEW). Only the group was a significant factor, with the EST group being 10.82 more likely to experience this symptom (95% confidence interval, 1.27–92.28, P = .029).
Women in the EST group were more likely to meet criteria for BV by both Amsel’s and Nugent’s criteria than the NEW group. Microscopy showed greater than 10 WBCs/high power field and the presence of parabasal cells more often in women in the EST group than women in the NEW group.
When women in the EST group who were bothered by vaginal discharge were compared with those who were not bothered ( Table 2 ), women who were bothered by discharge were more likely to meet the criteria for BV by Amsel’s but not by Nugent’s criteria. Vaginal pH, the presence of parabasal cells and clue cells, and the finding of 10 or more WBCs/high power field were similar between women bothered and those who were not bothered by vaginal discharge.
| Variable | Bothered | Not bothered | Missing | P value |
|---|---|---|---|---|
| BV by Amsel’s criteria, n, % a | 4/15, 26.7% | 0/34, 0.0% | 1 | .006 b |
| BV by Nugent’s criteria, n, % a | 3/15, 20.0% | 9/35, 25/7% | 0 | .70 |
| pH (mean, SD) c | 7.0 ± 1.3 | 6.47 ± 1.3 | 0 | .19 |
| WBCs (≥10 per hpf), n, % a | 5/14, 35.7% | 14/34, 41.2% | 2 | 1.00 |
| Clue cells (mean, SD), % d | 10.5 ± 22.9 | 11.5 ± 25.6 | 1 | 1.00 |
| Clue cells, n, % a | 5/15, 33.3% | 11/34, 33.4% | 1 | 1.00 |
| Parabasal cells present, n, % a | 10/15, 66.7% | 22/34, 64.7% | 1 | .90 |
b Statistically significant ( P < .05)
The ancillary analysis performed to establish the effect of varying amounts of time on Amsel’s and Nugent’s criteria revealed a Spearman’s rho value for correlation between months of pessary use and BV by Amsel’s criteria of 0.005 ( P = .971). The Spearman’s rho for the relationship between the duration of pessary use and BV by Nugent’s criteria is 0.032 ( P = .823). The mean duration of use in women who met BV by Amsel’s criteria was 47 months compared with 42.4 months in women who did not meet Amsel’s criteria for BV ( P = .788). The mean duration of pessary use in women who met Nugent’s criteria BV was 42.4 months vs 42.5 months in women who did not meet Nugent’s criteria ( P = .996).
Culture results revealed the presence of multiple aerobic and anaerobic organisms in these subjects. Aerobic organisms included Staphylococus , Streptococcus , Corynebacterium , Citrobacter , Klebsiella , Enterococcus , Enterobacter , Lactobacillus , Gardnerella , Serratia , and Hemophilus and Proteus species and Escherichia coli . Anaerobic organisms included Prevotella , Peptostreptococcus , Porhyromonas , Propionibacterium , Fusobacterium , Bacteroides , and Clostridium species. Tables 3 and 4 list the most prevalent organisms in the EST vs NEW women. The most prevalent organisms in the EST women who were bothered vs not bothered by vaginal discharge are included. The organisms are similar, although the relative prevalence of each organism varies slightly between groups.
