Bacterial vaginosis diagnosis and treatment in postmenopausal women: a survey of clinician practices

Abstract

Objective:

Some diagnostic features of the genitourinary syndrome of menopause (GSM) and bacterial vaginosis (BV) overlap, such as low levels of vaginal Lactobacillus and pH>5. We sought to determine clinicians’ diagnostic and treatment practices for post-menopausal women presenting with BV and GSM scenarios and how commercial molecular screening tests are utilized.

Methods:

Anonymous surveys were sent to practicing women’s health clinicians to evaluate assessment and treatment strategies for post-menopausal women presenting with BV and GSM scenarios.

Results:

When given a scenario of a post-menopausal woman with symptoms overtly positive for BV, a majority of providers (73%) would conduct a wet mount, though only 35% would evaluate full Amsel’s criteria. A majority (89%) recommended treatment with antibiotics, 28.2% recommended vaginal estrogen in addition to antibiotics and 11.8% recommended vaginal estrogen alone. Of providers who would use a molecular swab, 30% would wait for results before treating the patient’s symptoms. When given a scenario of a post-menopausal woman presenting with GSM, a majority (80%) recommended vaginal estrogen, and only 4.6% recommended antibiotics. Few (16%) responders would evaluate with a molecular swab, half of whom would wait for results before prescribing treatment. Clinicians in practice less than 10 years were more likely to rely on molecular swabs than those who had been practicing longer (p<0.0003).

Conclusions:

Methods used to evaluate postmenopausal women with vaginal symptoms vary. Future studies of post-menopausal women that differentiate diagnostic criteria between BV and GSM and validate commercial molecular testing for BV in women over age 50 are needed.

INTRODUCTION

Bacterial vaginosis (BV) is characterized in part by low relative abundance of key Lactobacillus spp and high levels of anaerobic and facultative bacteria.¹ The common presenting symptoms of BV are abnormal vaginal discharge and fishy odor, although half of women with BV are asymptomatic². Among reproductive-age women, BV has a U.S. prevalence of 29% and is a common cause of abnormal vaginal discharge.^{3, 4} Reports of BV’s prevalence in postmenopausal women vary more widely (5.4%−38%), likely because the standard criteria for diagnosis are more relevant for pre-menopausal women.^5–10

Several modalities are available for diagnosis of BV. Although rarely used in the clinical setting, Nugent’s method of Gram stain scoring is a standardized classification system for identifying BV samples based on morphology.⁸ Amsel’s clinical criteria, which can be performed as point-of-care testing, is more commonly used by clinicians due to its ease and immediacy of results that can aid in management decisions. Amsel’s criteria has moderate reproducibility compared to Nugent scoring, with sensitivity ranging from 37–70% and specificity ranging from 94–99%.¹¹ Where Amsel’s criteria relies on symptoms, clinical evaluation, in-office microscopy and pH testing for determination of BV, Gram-stain scoring requires a sample to be sent to a laboratory for microscopy. Nugent scoring is typically only used in research settings and provides only morphological information, which provides limited insight into the composition of the vaginal microbiota.

Nucleic acid amplification tests (NAAT) for vaginitis (vulvovaginal candidiasis and BV) have emerged and are becoming increasingly used in clinical applications. As an infectious correlate has not been identified in BV, these nucleic acid tests quantify loads for a panel of various bacteria, including Gardnerella vaginalis, Atopobium vaginae, BV associated bacteria (BVAB), and others such as Megaspharea phylotype 1 and 2. The tests may also distinguish multiple types of Lactobacillus spp including L. crispatus, L. gasseri, and L. jensennii.¹¹ The tests report either binary positive/negative or a risk category of low/moderate/high risk for BV with certain Lactobacillus species being protective against BV diagnosis and the other bacteria as more indicative of BV.

The diagnosis and treatment of BV in postmenopausal women is not standardized. The vaginal microbiota in postmenopausal women is known to have a different makeup compared to reproductive-aged women. In general, post-menopausal women have lower levels of estrogen and glycogen in the vaginal mucosa as well as thinner vaginal epithelium, lower levels of Lactobacillus and higher microbial diversity and higher vaginal pH.^12–16 However, 20–50% of post-menopausal women do retain Lactobacillus spp.^{17, 18} Given that the natural physiology of aging can mimic some of the same changes seen in BV, the validity of the standard diagnostic criteria in postmenopausal women is unclear. The original studies that developed both the Nugent scoring and Amsel’s criteria, the two most commonly accepted forms of diagnosis for research and clinical purposes, excluded menopausal women. As one of these two methods are used as the standard diagnostic criteria against which all subsequently developed diagnostic tools are compared and validated, the optimal method for diagnosing BV in post-menopausal women is unknown. The utility of the Nugent scoring system in post-menopausal women has recently been called into question given that many women will have an abnormal score with no pathology or symptoms detectable.^{6, 10}In benchmarking of NAAT studies for BV, post-menopausal women were either explicitly excluded or represented a small proportion and were not analyzed separately.^19–24 Consequently, it is unclear how molecular diagnostics tests for BV are utilized in post-menopausal women.

We sought to survey healthcare practitioners who treat menopausal women to determine their practice patterns regarding diagnostic approaches, and specifically, use of molecular assays for detection of BV in post-menopausal women.

MATERIALS AND METHODS

Survey Design

The study survey consisted of nine demographic questions involving provider training, practice setting, experience, and education as well as fourteen content-based questions. These questions were sent to eight practicing OB/GYNs for comment and validation prior to release. The content-based questions were divided into three sections: 1) scenario 1 regarding evaluation and management of a menopausal woman with symptoms of BV, 2) scenario 2 regarding evaluation and management of a post-menopausal woman with genitourinary syndrome of menopause (GSM)²⁵, and 3) questions to assess clinician use of molecular assays (see Survey, Supplemental Digital Content 1, which contains the survey questions sent to potential participans). RedCap was utilized as the method of survey distribution and data collection. The survey was e-mailed to providers with a link that included a consent page followed by the survey. The survey responses were anonymous.

An incentive of a $10 Amazon e-gift card was given to participants who completed the survey. The initial surveys were sent in November, 2017 and data collection concluded in April, 2018.

Recruitment

An email explaining the study was sent to 400 physicians and certified nurse midwives who practice in the fields of Obstetrics and Gynecology, Family Medicine and Internal Medicine. The providers were selected by identifying those whose email addresses were publicly available via their institutions websites. Additionally, information regarding the study was posted on a private, national social media group of female physician members. Program directors of all accredited Obstetrics and Gynecology residency programs were also sent information and encouraged to send the email to the members of their residency programs. Interested clinicians were asked to send an email to the study coordinator using their work email address to verify authenticity of their employment. Those who contacted the study coordinator were sent an individualized link to a REDCap survey.

Ethical Approval

This survey was approved by the University of Maryland Baltimore Human Research Protections Office. No patient data was collected.

Statistical Analysis

Data was analyzed using SAS Studio software. Incomplete surveys were defined as those who did not complete the content-based questions. To determine correlations between swab use for diagnosis and demographic features of survey responders, we used chi squared analysis to calculate p-values. For survey responses, we calculated frequencies and column percentages of individuals choosing a response in each question.

RESULTS

Of the 382 surveys that were distributed, 172 were initiated and 152 (39.8%) were completed. The large majority of respondents practice Obstetrics and Gynecology and 65% reported less than quarter of their practice visits were with menopausal women. Practice settings and number of years of experience was distributed (Table 1).

Table 1:

Demographics of respondents

Characteristics	Total (N=152)	% of population
OB/GYN specialty	146*	96.05
NAMS-Qualified?	10	6.57
Experience
In training	51	33.55
1–5 years out of training	37	24.34
5–10 years out of training	35	23.02
>10 years	29	19.07
Practice Setting
Community Hospital	43	28.28
Teaching Hospital	106	69.73
Hospital-affiliated clinic	36	23.68
Private practice	19	12.5
Post-Menopausal Patient Care:
0–25%	98	64.67
25–50%	31	20.39
50–75%	15	9.86
75–100%	8	5.26

Table 1. Demographics of providers that have completed the survey. The providers that were not OB/GYN trained had a background in Family Medicine (2), Internal medicine (3), and Other- not specified (1).

Scenario 1: Bacterial Vaginosis

The first scenario described a 60-year-old woman with vaginal discharge, itching and dyspareunia and pale vaginal mucosa. When asked the next step in their workup of a woman with this presentation, 73.0% reported that they would do a wet mount with 35.5% indicating that they would evaluate all of Amsel’s criteria, 19.7% indicating that they would do a wet mount and a molecular swab simultaneously, and 11.8% would use a molecular swab as their only test (Table 2).

Table 2:

Diagnostic testing choice for post-menopausal woman with bacterial vaginosis (Scenario 1) and genitourinary syndrome of menopause (Scenario 2)

	Scenario 1: Bacterial vaginosis % (N)	Scenario 2: GUSM %(N)
No diagnostic testing	9.86 (15)	49.34 (75)
Wet Prep (total)	73.02 (111)	36.84 (56)
Wet prep alone	30.26 (46)	7.89 (12)
+ whiff test	42.76 (65)	3.95 (6)
+ whiff test, pH	35.52 (54)	11.18(17)
+ swab	19.73 (30)	7.89 (12)
+ GC/CT testing	12.50 (19)	5.92 (9)
BV Swab alone	11.84 (18)	8.50 (13)
pH alone	1.31 (2)	1.97 (3)
Prefer not to answer	3.95 (6)	3.29 (5)

Breakdown of diagnostic tool combinations. % Column was calculated by the rate a diagnostic tool was chosen by the surveyed providers. GUSM genitourinary syndrome of menopause.

Overall, 89.4% of respondents recommended treatment with either oral or vaginal antibiotics, most commonly oral metronidazole (Table 3). Additionally, 61 respondents (40.1%) recommended treatment with vaginal estrogen. Of these, 18 (29.5% of the subset recommending estrogen, 11.8% of the total cohort) would treat with vaginal estrogen alone and 43 (70.5% of the subset recommending estrogen, 28.2% of the overall cohort) recommended dual treatment with vaginal estrogen plus oral and/or vaginal antibiotics. Only the respondents who chose wet mount (N=111) as part of their initial diagnosis were given the information that the patient was overtly positive for all components of Amsel’s criteria. Of this subset of respondents, 98.2% would treat with oral or vaginal antibiotics, most commonly metronidazole. Additionally, 34 respondents (30.6%) would treat with vaginal estrogen. Of these, none would treat with vaginal estrogen only; all would treat the patient with a combination of vaginal estrogen and oral or vaginal antibiotics. A small amount of those who initially used wet mount microscopy for diagnosis, 1.8%, reported that they would still send a follow-up molecular swab after the results were given that were overtly positive for all of Amsel’s criteria. Of this population (N=2), both chose to treat with oral metronidazole prior to the results of the molecular test being available.

Table 3:

Treatment for post-menopausal woman with bacterial vaginosis (Scenario 1) and genitourinary syndrome of menopause (Scenario 2)

Treatment Choicesa
	Scenario 1 BV % (N)	Scenario 2 GUSM % (N)
No treatment	2.63 (4)	1.31 (2)
No empiric treatment; awaiting results	11.18 (17)	9.86 (15)
Oral Antibiotics	63.81 (97)	2.63 (4)
PO MTZ only	59.21 (90)	1.97 (3)
PO Clindamycin + MTZ	4.60 (7)	0.66 (1)
Vaginal Antibiotics	25.6 (39)	1.31 (2)
V MTZ only	21.05 (32)	1.31 (2)
V Clindamycin only	0.66 (1)	0
V MTZ + Clindamycin	3.95 (6)	0
Boric Acid Vaginal Suppository	2.63 (4)	1.97 (3)
Vaginal Estrogen	40.13 (61)	80.92 (123)
Vaginal estrogen only	11.8 (18)	78.95 (120)
+PO antibiotics	21.05 (32)	1.31 (2)
+ V Antibiotics	8.55 (13)	0.66 (1)
PO Estrogen	1.97 (3)	1.31 (2)
Probiotics	6.57 (10)	1.31 (2)
Vaginal Lubricants	19.73 (30)	36.84 (56)
Other	2.63 (4)	1.31 (2)

The above table shows each treatment tool option and its selection rates by surveyed providers (total = 152) regarding how they would treat scenarios 1 and 2. BV bacterial vaginosis, GUSM genitourinary syndrome of menopause.

^a.Categories are not mutually exclusive

Of the respondents who would order a molecular swab (N=56), 18 of them (32.1%) would use the swab as their only test and 17 (30.3%, 10.8% of overall cohort) would not recommend treatment until the results of the swab were available. Of those who would order/perform other tests in addition to the swab, all of them would recommend treatment prior to receiving the swab results. The most common treatments offered (N=38) were vaginal estrogen alone (18.4%), oral or vaginal antibiotics (57.9%) or a combination of antibiotics and vaginal estrogen (23.6%). Nearly all respondents (94.4%) that would use the swab as their only test reported that they would wait until the results of the swab were available prior to recommending treatment.

Scenario 2: Genitourinary syndrome of menopause

In the second scenario, a 62-year-old woman presents with the complaint of intermittent vaginal irritation and is found to have pale vaginal mucosa on exam. Given these findings, approximately half of respondents (49.3%) reported that they would order no further diagnostic testing (Table 2). For those who would order testing, 36.84% would perform a wet mount and 16.44% indicated that they would send a molecular swab. Of those that chose to send a molecular swab, 43.4% of those would do the swab as their only test. Respondents that chose wet mount as a diagnostic tool were given information that it was unremarkable and the vaginal pH was 6.

For treatment, most respondents (80.9%) reported that they would recommend vaginal estrogen and 36.8% reported that they would recommend vaginal lubrication with or without estrogen (Table 3). A small percentage of respondents (4.6%) would recommend treatment with oral or vaginal antibiotics; of note, all of these respondents would also chose to send a molecular swab. Of the respondents who chose to send a molecular swab, 48% of them would await results before offering treatment, which represented 7.9% of the overall respondents.

Use of molecular swabs

Over half (57.6%) of respondents reported using molecular swabs when evaluating women with vaginal symptoms with 18.4% of respondents reporting using them 75–100% of the time. Only 2.0% reported restricting their use to pre-menopausal women. Type of practitioner, geographic location of practice and practice setting were not statistically significantly related to likelihood of reporting use of molecular swabs. Those respondents who were in training or had been in practice for less than 10 years were significantly more likely (OR 5.58, 95% CI: 2.20–14.18, p<0.0003) to use swabs for diagnosis than those with greater than 10 years of experience.

The reasons given for choosing to use molecular swabs were varied. Almost a third (29.9%) of practitioners believe that molecular swabs are more accurate than clinical diagnosis and 31% reported that they rely on these tests because they do not have access to a microscope. The most common response (43.7%) was that respondents use molecular swabs when they are unable to make a definitive diagnosis clinically.

Over one quarter of respondents (26.4%) believed that the commercial swabs cost less than $100 and one third (33.3%) of respondents had no knowledge of the cost. Only 6.9% of respondents believed that the tests cost $300 or more.

DISCUSSION

Our study found that the practice patterns regarding methods used to evaluate post-menopausal women with vaginal symptoms is variable. Overall frequency of use and reasons for use of molecular tests for diagnosis of vaginal symptoms also varied among practitioners.

The finding in our study that there are very few practitioners that restrict their use of molecular BV screening tests to pre-menopausal women is notable. None of the molecular test manufacturers give specific age ranges for use of the swabs [BD Affirm^TM VPIII Microbial Identification Test; BD MAX^TM Vaginal Panel Package Insert; Aptima Vaginal Swab Specimen Collection Kit Package Insert; NuSwab^SM: Smart Testing Has Arrived; SureSwab®, Vaginosis/Vaginitis Plus]. The studies performed for validation of these molecular assays do not exclude post-menopausal women, but none specify menopausal status of participants or analyze menopausal women separately. As most enroll through sexually transmitted infection screening clinics, it is very likely that menopausal women are underrepresented in these studies. Of the five largest studies, one excludes women over 50²², one reports age range 19 to 67 with median age of 25²³, one reports mean age of 34 with no range given²⁰, and one reported only a range of 17 to 55 with no breakdown or average age reported²¹. Only one study gives specifics regarding the ages of women enrolled and this study reported 4.2% of participants were over the age of 50.¹⁹ Two additional studies are referenced in the BDMax guide; one specifies that it included only reproductive aged women and the other includes only pregnant women.^26–28

There is biological plausibility to the concern that the validity of molecular tests for diagnosis of BV may differ in pre- versus post-menopausal women. Post-menopausal women are significantly less likely to have vaginal microbiota dominated by Lactobacillus spp. and as many as a third of post-menopausal women have no detectable Lactobacillus spp..^{6, 10, 12, 18, 29} Given that molecular assay panels evaluating for BV utilize Lactobacillus in their algorithms, specific validation in post-menopausal women would be prudent.

Given the lack of clarity in how to interpret traditional testing for BV when evaluating a postmenopausal woman, it is possible that practitioners are turning to molecular tests in search of a more sensitive and specific diagnostic tool. This is evidenced by the fact that nearly a third of our respondents report their reason for use of molecular tests to be increased perceived accuracy in diagnosis and almost half use molecular tests when they are unable to make a diagnosis based on clinical findings. However, in post-menopausal women, the molecular assay may be identifying a normal, physiology rather than BV and additional validation and reporting metrics may be needed to avoid misinterpretation of these results by clinicians.

This study also elucidates other potential issues with the use of molecular tests in the evaluation of vaginal symptoms. The reliance of some practitioners on molecular assays rather than point of care testing may lead to a delay in treatment of symptomatic women. Although the sensitivity of clinical diagnosis is shown to be rather low in many studies, the specificity is typically high, particularly with a positive whiff test and elevated pH.^{30, 31} In both clinical scenarios given in this study, nearly one in ten practitioners would not treat at the initial visit but would await results of a molecular swab despite a combination of symptoms and physical exam findings that are classically found in BV and GSM. Given the high specificity of clinical diagnosis, there is likely little clinical utility or cost benefit to performing these molecular assays in these scenarios and awaiting the results of these assays prior to offering treatment may lead to unnecessary delays.

Additionally, few practitioners had an accurate understanding of the potential costs of these molecular studies. Although the rates that are negotiated by insurance companies are not publicly available, the standard costs without insurance for these tests range from $300 to $1,200 (personal communication with representatives).

The optimal treatment for a BV-like presentation in post-menopausal women is also unknown. We found that many practitioners chose to treat with antibiotics as well as local estrogen. Although vaginal estrogen as treatment or prevention of BV has not been specifically studied, there are studies showing that hormone therapy in menopausal women increases the abundance of Lactobacillus spp.^{6, 32}

There are of course limitations to this study. The survey was composed of multiple choice questions that did not allow comment and may not have captured all of the nuances present in clinical care. As with any survey, respondent bias may affect the results. As less than 40% of those who received a survey responded and the demographics of those who did not respond could not be obtained. Many of the respondents were in training and most have a relatively small percentage of postmenopausal women in their practice. Given this, it is unclear how the cohort of respondents compares to the larger population of Women’s Health practitioners. However, it is possible that including a cohort of practitioners who are not sub-specialized in menopausal care actually increases the generalizability of the results.

CONCLUSION

In summary, our study identified variations in practice in both evaluation and treatment of BV and GSM in menopausal women. This highlights the need for future studies that differentiate between changes in vaginal bacteria in menopause and the most effective way to treat a symptomatic and often low-Lactobacillus state in menopause.