Objective
Pessaries are important options for women with pelvic floor disorders, but many pessary users experience bacterial vaginosis (BV). The aim of this study was to evaluate the effect of TrimoSan gel (Milex Pessaries, Cooper Surgical, Trumbull, CT) on BV prevalence among pessary users.
Study Design
Women presenting for a pessary fitting completed questionnaires on vaginal symptoms and hormone therapy use and underwent a BV BLUE test and slide collection for BV analysis by Nugent’s criteria. Following pessary fitting, women were randomized to either standard pessary care with the use of TrimoSan placed vaginally twice weekly or to standard pessary care without TrimoSan gel. Women returned 2 weeks and 3 months later for a repeat slide collection for Gram stain, BV BLUE testing, and completion of questionnaires on vaginal symptoms and desire to continue the pessary.
Results
There were 184 women randomized after successful fitting (92 to the TrimoSan group), and 147 (79%) presented for 3-month follow up. Mean age was 56 ± 16 years; patients were mostly white (57%) or Hispanic (23%), and 36% were using hormone therapy. The groups did not differ in the prevalence of BV by Nugent’s criteria at 2 weeks (20% TrimoSan vs 26% no gel, P = .46) or 3 months (24% TrimoSan vs 23% no gel, P = .82), nor did they differ in BV by BV BLUE testing at 2 weeks (0% TrimoSan vs 4% no gel, P = .12) or 3 months (3% TrimoSan vs 0% no gel, P = .15). The prevalence of at least one vaginal symptom did not differ between groups at 2 weeks (44% TrimoSan vs 45% no gel, P = .98) or 3 months (42% TrimoSan vs 32% no gel, P = .30). The TrimoSan group was equally likely to want to continue their pessary use compared with the standard care group at 2 weeks (90% vs 86%, P = .64) and 3 months (63% vs 60%, P = .76).
Conclusion
TrimoSan gel in the first 3 months of pessary use does not decrease the prevalence of BV or vaginal symptoms and does not alter the likelihood of a woman desiring to continue pessary use.
Pelvic floor support disorders are common and debilitating to women, resulting in a significant deterioration in quality of life and overall health. Pelvic floor disorders are prevalent, with up to 25% of women having at least one symptomatic pelvic floor disorder. Although surgery for these disorders is increasingly common, many women cannot undergo surgery or prefer nonsurgical treatment options.
Many women can be successfully fitted with a pessary, a silicone device placed in the vagina to provide pelvic support for the treatment of pelvic organ prolapse and stress urinary incontinence. Although pessaries relieve symptoms in the majority of users, nearly half of women discontinue pessary use within 1 year because of a variety of troublesome side effects, including vaginal discharge and odor. Patients who meet established goals of pessary use are more likely to continue the pessary, so clinicians are in need of evidence-based means to maintain pessary satisfaction after fitting.
Bacterial vaginosis (BV), a shift in the ecological balance of the vagina leading to malodorous vaginal discharge, is common among pessary users and may negatively affect their experience with a pessary. Prior authors have published that up to 30% of women with a pessary experience BV, in contrast to 10% in the general population. Many providers attempt to avoid pessary-related BV by counseling patients on hygiene measures and care, but no randomized trials investigate the care of a pessary after fitting. TrimoSan gel (Milex Pessaries, Cooper Surgical, Trumbull, CT), a mildly acidic vaginal lubricant that is dispensed with some pessaries, may lower the pH of the vagina and keep the vaginal ecosystem in balance during pessary use, but no clinical trials examine its effect on BV or on other outcomes.
We conducted a randomized, controlled trial to determine whether TrimoSan gel decreases the prevalence of BV or bothersome vaginal symptoms during the first 3 months of pessary use. We hypothesized that the TrimoSan gel would decrease BV and vaginal symptoms compared with women not using the TrimoSan gel.
Materials and Methods
This was a multicenter, randomized, single-blind, controlled trial of patients who were fitted for a pessary for any indication at 2 tertiary care centers, from July 1, 2010, through Dec. 31, 2011, at MedStar Washington Hospital Center (Washington, DC) and from July 1, 2012, through May 5, 2014, at the University of New Mexico Health Sciences Center (Albuquerque, NM).
The primary outcome was the determination of BV by Nugent’s criteria on Gram stain 3 months after pessary fitting. The study was approved by the institutional review boards at both clinical sites after ethics committee approval and was registered at ClinicalTrials.gov (identifier number NCT01471457 ).
Potential patients were identified prior to or at the time of a pessary fitting visit, and they were enrolled after providing written, informed consent. We approached all patients who were older than 18 years of age, suitable for follow-up, and undergoing a pessary fitting for any indication.
Patients were excluded if they were younger than 18 years of age, had a known allergic or suspected adverse reaction to TrimoSan gel or any of its components, had an inability to use a pessary or TrimoSan as indicated, already had used a pessary within the last 12 months, had a history of frequent or chronic BV (>2 episodes per year or symptoms longer than 6 months of the last 12 months), had a history of active vaginal ulcerative disease (active ulcers from atrophy, herpes, or mesh erosion with more than 2 episodes of ulcers per year or last ulcer <1 month ago), were using long-term antibiotics for indications not listed previously, were unable to speak English, were unable to provide informed consent, or were unable to be fitted with a pessary.
If patients were randomized at the time of the initial pessary fitting but unable to retain the pessary or be successfully fitted before leaving the office examination and declined a second attempt at pessary fitting, they were excluded after randomization for failure to be fitted with a pessary. Patients who were randomized and fit with a pessary who were planning on wearing their pessary continuously at home and coming to the office for cleaning/removal were still included in the study.
Patients completed baseline questionnaires on their medical history and health, use of hormone therapy (HT) orally and/or vaginally, any vaginal products or medications being used in the vagina, and a questionnaire on vaginal symptoms including vaginal discharge, fishy odor, clear or gray vaginal discharge, too much vaginal discharge, vaginal discharge interfering with sex, vaginal itching, vaginal pain, vaginal pain interfering with sex, and vaginal sores.
Any HT use (any HT) was defined as the use of systemic (oral, transdermal, or vaginal HT), and vaginal hormone therapy use (vaginal HT) was defined as any hormone therapy given via a vaginal route (cream, ring, etc). Therefore, the group using HT included the women using vaginal HT plus women using other forms of HT. The vaginal HT group included women already using vaginal HT upon entry into the study and women who initiated vaginal HT at the time of pessary initiation.
Prior to pessary fitting, study patients had a clean, plastic speculum inserted, and vaginal secretion swabs were collected on 2 sterile cotton tip swabs, with care taken to sample from the middle third of the vagina. One swab was used to perform BV BLUE testing (OSOM Rapid Tests; Sekisui Diagnostics, Lexington, MA) at the time of the visit. If the patient was found positive for BV by BV BLUE at this baseline visit, she was treated with 500 mg of oral metronidazole twice daily for 7 days. If the patient had an allergy to this medication, the patient was prescribed 300 mg of oral clindamycin twice daily for 7 days.
The second vaginal secretion swab was rolled onto a dry, clean glass microscope slide, air dried without a coverslip, and transported to the laboratory and stored at room temperature until the time of Gram staining. After Gram staining, Nugent’s criteria was used to define the presence of BV and allowed the laboratory assessor of the Gram stain to be remote from the clinical study site and unaware of the patient’s treatment allocation.
Following the completion of vaginal specimen collection and questionnaires, patients were fitted with a pessary that was comfortable and allowed them to adequately void. At the time of pessary fitting, patients were randomized to either proceed with standard pessary hygiene plus the use of TrimoSan gel (half applicator of gel vaginally twice a week) or to standard pessary care without the use of TrimoSan gel.
The randomization scheme was a blocked randomization sequence with randomly alternating block size (block size ranging from 5 to 10) with a 1:1 ratio of TrimoSan gel to no TrimoSan gel, stratified by women who intended to remove their pessary at least daily and women who intended to remove their pessary less often than daily. The randomization sequence was prepared by a statistician uninvolved in patient recruitment or allocation and was placed in sequentially numbered opaque sealed envelopes for the 2 different strata by an investigator who also had no involvement in patient recruitment or treatment allocation.
Women stated their intention for pessary removal frequency to the research staff performing the randomization during their pessary fitting, and the investigator opened the next sequentially numbered opaque sealed envelope in the appropriate strata to assign the patient’s randomization group. Patients randomized to the TrimoSan group were then given TrimoSan gel and instructions for its use, and patients allocated to the control group were specifically instructed not to use TrimoSan gel. Patients were not blinded to their treatment allocation.
Women were seen at 2 weeks and 3 months later for examination and completed repeat questionnaires on pessary use, hygiene practices, and vaginal symptoms, including additional questions regarding increase in the 4 vaginal symptoms of discharge, itching, pain, and sores since pessary fitting. Women were asked how often they removed their pessary (daily or more often, less often than daily but at least once per week, less often than once per week) and how often they wore their pessary (never wear, wear 1-7 days per month, >7 days per month, daily). They were also asked to rank their desire to continue their pessary by a 5 point Likert scale. At each examination, a midvaginal secretion swab was done to perform the BV BLUE test in the clinic and to prepare the microscopy slide for Gram staining.
For the analysis of how frequently women removed their pessary, frequent pessary removal was defined as daily or more often, whereas less often pessary removal was defined as less often than daily. Pessary satisfaction was defined as a reply to the question, “How much do you want to keep wearing your pessary in the future?” Replies included quite a bit or moderately. Excellent pessary satisfaction was defined as a reply of quite a bit to this question.
We analyzed the prevalence of reported vaginal symptoms individually and also for the presence of at least 1 reported vaginal symptom. We also analyzed the prevalence of women reporting an increase in 1 of 4 vaginal symptoms since pessary fitting (vaginal discharge, itching, pain, sores) and the prevalence of reporting an increase in at least 1 of these 4 symptoms since pessary fitting.
The primary outcome was the rate of BV by Gram stain (Nugent’s criteria) at 3 months after the pessary fitting. The Nugent’s score was determined by following a standardized Gram staining of the microscope slide. BV was defined as a Nugent’s score of 7 or greater, in accordance with Nugent’s criteria.
Secondary outcomes included the rate of BV at 2 weeks after pessary fitting by Nugent’s criteria, the rate of BV at 2 weeks and 3 months after pessary fitting by BV BLUE test, the prevalence of individual vaginal symptoms or at least 1 reported vaginal symptom at 2 weeks and 3 months following pessary fitting, and the prevalence of the patient reporting an increase in at least 1 of the 4 key vaginal symptoms (discharge, itching, pain, sores) since pessary fitting at 2 weeks or 3 months.
We performed analyses both by intention-to-treat (primary analysis) and by per-protocol analysis. Women were asked to report their frequency of TrimoSan use at each follow-up visit. Women who used TrimoSan gel at least once weekly were considered compliant with TrimoSan use. The per-protocol analysis was performed by comparing those women who were compliant with TrimoSan gel at least once a week (half of the recommended dosing) to those women who used the gel less frequently or did not use the gel. Power calculation was performed based on a 30% reported prevalence of BV in pessary wearers, in contrast to a prevalence of 10% in the general population.
For the primary outcome of BV as detected by Gram stain 3 months after pessary fitting, the study sought to achieve a power of 88% at an α of 0.05 to detect a lowering of the prevalence of BV at 3 months in the TrimoSan-using group from 30% back to the baseline rate of 10%. This translated to a sample size of 62 in each group. To account for at least a 15% dropout rate after randomization, we planned to enroll 75 women in each group. After initial recruitment in the first study site yielded a lower follow-up rate than expected, the sample size was expanded to 180 patients so that a 70% rate of follow-up at 3 months would achieve the predetermined power.
The effect of dichotomous variables (such as randomization to TrimoSan, any HT use, vaginal HT use, and frequent pessary removal) on dichotomous outcomes was determined by χ 2 tests (with Fisher exact test used in which there were significantly less than the expected number of data points), and the effect of dichotomous variables on continuous variables was determined by t tests. For the nonparametric variables (such as parity and number of vaginal deliveries), the Wilcoxon rank-sum test was used to compare between 2 groups.
We utilized the Mantel-Haenzel method to test the interaction of different frequencies of pessary wearing and pessary removal (as stated by the patient at follow-up) with the relationship between the use of TrimoSan gel and dichotomous outcomes. We also used the Mantel-Haenzel method to test the effect of HT use or vaginal HT use on the relationship between the use of TrimoSan gel and dichotomous outcomes. We used general linear models to examine the effect of frequency of pessary removal and HT on the relationship between TrimoSan gel use and continuous outcomes (such as the total Nugent’s score).
We used a linear regression model to determine what other patient characteristics had an impact on the relationship between frequent pessary removal or HT and outcomes when significant findings were present in the uncorrected analysis. If certain patient characteristics were found to have an impact on the relationship, we repeated the analysis with correction for this variable.
We utilized a χ 2 test to analyze whether there were significant differences in BV prevalence or vaginal symptoms at 2 weeks or 3 months between different reported frequencies pessary use or pessary removal, again utilizing the Fisher exact test where more appropriate. As noted above, we defined frequent pessary removal as once-daily removal or more often and analyzed the effect of frequent removal by a χ 2 test on the BV and vaginal symptom outcomes.
Results
There were 466 women screened for eligibility during the study period ( Figure ), with 86 women excluded because of study criteria, 85 women declining to participate, and 108 women not presenting for pessary fitting after being screened. The remaining 187 women who consented for participation were randomized, of which 3 additional women were excluded after randomization for the inability to be fitted with a pessary at their initial visit and patient declination to attempt another fitting. This left 184 active eligible participants, 92 randomized to TrimoSan use and 92 randomized to no TrimoSan use.
Patient characteristics are presented in Table 1 and did not vary between the 2 groups at baseline. The mean age of the patients was 58.6 ± 16.4 years, the mean body mass index (BMI) was 28.3 ± 7.5 kg/m 2 , the median parity was 3, and the majority of the women were white (57%), Hispanic (23%), or African American (10%). The prevalence of BV (by Nugent’s criteria or by BV BLUE testing) and the total Nugent’s score were similar between the groups at baseline. There were no differences between the groups in the number of women requiring pessary refitting at 2 weeks or 3 months. The 2 groups also did not differ in the frequency of wearing or removing the pessary reported by the patients at 2 weeks or 3 months and did not differ in the prevalence of any HT use or vaginal HT use.
| Variable | TrimoSan, n (%) or mean ± SD a or median (variance) b (n = 92) | No TrimoSan, n (%) or mean ± SD a or median (variance) b (n = 92) | Total patients (n = 184) |
|---|---|---|---|
| Age, y a | 58 ± 15 | 60 ± 17 | 59 ± 16 |
| BMI, kg/m 2 a | 29 ± 7.6 | 28 ± 7.4 | 28 ± 7.5 |
| Race | |||
| White | 56 (61) | 48 (52) | 105 (57) |
| African American | 6 (7) | 13 (14) | 19 (10) |
| Hispanic | 23 (25) | 20 (22) | 43 (23) |
| Asian | 1 (1) | 3 (3) | 4 (2) |
| Other | 5 (5) | 8 (9) | 13 (7) |
| Parity b | 3 (2.6) | ||
| Vaginal deliveries b | 2 (2.4) | ||
| Indication for pessary use | |||
| Prolapse | 41 (44) | 45 (49) | 86 (47) |
| Urinary incontinence | 35 (38) | 30 (33) | 65 (35) |
| Both prolapse and urinary incontinence | 13 (14) | 10 (11) | 23 (13) |
| Smoking | |||
| Current | 8 (9) | 6 (7) | 14 (8) |
| Past | 26 (28) | 28 (30) | 54 (29) |
| Charlson Comorbidity Index a | 0.77 ± 1.1 | 0.89 ± 1.8 | 0.82 ± 1.5 |
| Insurance type | |||
| Private | 36 (39) | 38 (41) | 74 (40) |
| Public | 27 (29) | 34 (37) | 61 (33) |
| None | 7 (8) | 6 (7) | 13 (7) |
| Education level | |||
| Graduate | 25 (27) | 28 (30) | 53 (29) |
| College | 37 (40) | 29 (32) | 66 (36) |
| High school | 20 (22) | 28 (30) | 48 (26) |
| Less than high school diploma | 3 (3) | 4 (4) | 7 (4) |
| Job activity level | |||
| Sedentary work | 16 (17) | 16 (17) | 32 (17) |
| Homemaker | 18 (20) | 18 (20) | 36 (20) |
| Light labor | 20 (22) | 26 (28) | 46 (25) |
| Heavy labor | 8 (9) | 5 (5) | 13 (7) |
| Unemployed or retired | 26 (28) | 19 (21) | 45 (24) |
| Prior pelvic surgery | 41 (49) | 39 (42) | 80 (43) |
| Prior pessary use | 9 (10) | 4 (4) | 13 (7) |
| Vaginal infection history | |||
| Vaginal infection last 12 mo | 10 (11) | 7 (8) | 17 (9) |
| HT | |||
| Using HT at time of recruitment | 22 (24) | 24 (26) | 46 (25) |
| Vaginal HT at time of recruitment | 19 (21) | 19 (21) | 38 (21) |
| HT prescribed at time of recruitment | 13 (14) | 14 (15) | 27 (15) |
| Pessary refitting | |||
| Required refitting at 2 wks | 13 (14) | 12 (13) | 25 (14) |
| Required refitting at 3 mo | 15 (16) | 12 (13) | 27 (15) |
| BV measures | |||
| BV by Nugent’s criteria at baseline | 23 (25) | 15 (16) | 38 (21) |
| Total Nugent’s score at baseline a | 4.5 ± 2.0 | 4.8 ± 2.4 | 4.6 ± 2.2 |
| BV by BV BLUE at baseline | 1 (1) | 2 (2) | 3 (2) |
a Designates that the mean +/− standard deviation is reported for that value
b Designates that the median (variance) is reported for that value.
There were 160 women of 184 eligible study patients (87%) who presented for either 2-week or 3-month follow-up, with 147 (79%) women presenting at 3 months for follow-up and 133 of them (72%) qualifying for the primary outcome measure with an interpretable 3 month Gram stain. Women who presented for some type of follow up (n = 160) were more likely to have had a vaginal infection within the last year (16% vs 6%, P = .04) but were otherwise similar in characteristics to the women who never presented for a follow-up after the initial visit (n = 24).
The prevalence of BV did not differ significantly between groups at 2 weeks or 3 months ( Table 2 ). Pessary satisfaction was not different between the groups at 2 weeks or 3 months, and excellent pessary satisfaction was also similar between the groups at both time points ( P > .05). Pessary satisfaction was substantially lower at 3 months than at 2 weeks in both groups and overall (88% at 2 weeks vs 62% at 3 months, P < .01). Individual and composite vaginal symptoms also did not differ between the groups at 2 weeks or 3 months, although the prevalence of at least 1 reported vaginal symptom was high (>30% in both groups at 2 weeks and >40% in both groups at 3 months; Table 2 ). Also, the prevalence of an increase in at least 1 vaginal symptom since pessary fitting was high (>30% in both groups at 2 weeks and 3 months; Table 2 ) but did not differ between the groups.
