Prevalence of Amsel-defined bacterial vaginosis before and after transvaginal ultrasound with lubricant application

Abstract

Emerging evidence suggests a link between hyperosmolal vaginal lubricants and bacterial vaginosis (BV). In 80 patients undergoing transvaginal ultrasound (TVUS) with a standardized lubricant, we observed a significant increase in Amsel-defined BV 1–13 days post-TVUS compared to baseline, likely driven by heightened amine odor (OR: 5.88; 95% CI: 1.22–28.23).

Summary:

A higher prevalence of Amsel-defined bacterial vaginosis and positive whiff tests was observed 1–13 days after transvaginal ultrasounds using hyperosmolal vaginal lubricant, compared to one week before the procedure.

Introduction

Bacterial vaginosis (BV) affects approximately one-third of U.S. reproductive-age women,¹ and is associated with vulvovaginal symptoms¹ and preterm delivery² as well as susceptibility,^{3, 4} transmission,⁵ and persistence⁶ of sexually transmitted infections (STIs). The etiology of BV remains largely unknown; however, it is characterized by a low Lactobacillus spp. abundance and diverse anaerobic bacteria. Observational studies report intravaginal practices and multiple or new sexual partners are associated with increased risk for BV.⁷ Vaginal lubricants are emerging as another potential risk factor.^{8, 9}

Water-based vaginal lubricants are commonly used during sex. They increase personal satisfaction, mitigate vaginal dryness and discomfort,¹⁰ and reduce condom failure.¹¹ Lubricants are also routinely used in clinical procedures including pelvic exams, transvaginal ultrasound (TVUS), and during pregnancy, labor, and delivery.

The World Health Organization recommends lubricants ideally be iso-osmolal (<380 mOsm/kg) with an upper limit of 1,200 mOsm/kg.¹² Many commercial lubricants exceed those limits. Personal and clinical water-based lubricants are compositionally similar; common ingredients include humectants like glycerin, propylene glycol, and polyethylene glycol which improve the feeling and lasting power of the lubricant but also increase its osmolality.^{13, 14}

These osmolality recommendations are derived in part from in vitro studies finding damage to vaginal and colorectal epithelial tissue cultures following exposure to hyperosmolal lubricants.^{14, 15} Some observational studies have reported a higher risk of BV among lubricant users;^{8, 9} however, these studies did not specifically evaluate hyperosmolal lubricants and were limited by variations in frequency, volume and patterns of use.

We sought to evaluate shifts in the prevalence of BV and Amsel clinical criteria following a single exposure to hyperosmolal lubricant in reproductive-age participants referred for TVUS, a procedure that includes a standardized lubricant application.

Methods

We utilized data from the Gynecology and Lubricant Effects (GALE) study.^{16, 17} This study sought to evaluate the association between use of a hyperosmolal lubricant during TVUS (GLIDE, formerly known as E-Z lubricating jelly, Athena Medical Products, ~2,400 mOsm/kg, pH:5.5) and vaginal health over 10 weeks in non-pregnant women referred for TVUS at the University of Maryland Medical Center. This medical lubricant was selected as it was being used by the Department of Radiology and Nuclear Medicine and was representative of commercial lubricant products.¹⁸ It is also manufactured in single-use foil packets (2.7g) and sonographers were instructed to use one packet, though additional packets could be applied if necessary. The study methods have been previously described.^{17, 19, 20} Participants were seen by a clinician and had a pelvic exam at baseline, approximately one week before the TVUS appointment, and at a follow-up appointment, approximately 2–5 days after TVUS (range: 1–13). Data from a third clinical visit in week 10 was not included in this analysis. Participants completed daily health and behavior diaries, and Amsel-BV was diagnosed by study clinicians at each clinical visit by at least three of four clinical criteria (thin homogenous discharge, vaginal pH>4.5, positive whiff test, and clue cells).²¹ Peri/post-menopausal status was assessed by menstrual cycle regularity and taking into account hormonal contraceptive use, adapted from the Stages of Reproductive Aging Workshop system.²⁰ We excluded participants who were peri/post-menopausal, missing clinical data, or reported use of antibiotics, antifungals, condoms (a potential source of additional lubricant), or commercial vaginal lubricants between baseline and the post-TVUS follow-up visit. Of the 121 participants enrolled, 80 were eligible for this analysis.

Exact McNemar’s test for matched pairs was used to evaluate whether participants were more likely to acquire versus resolve BV and individual Amsel criteria after TVUS. Generalized linear mixed effects models with a random intercept were used to evaluate within-participant odds of having each Amsel outcome after versus before TVUS. Time-varying factors reported on daily diaries in the 24 hours preceding each visit, including menses, douching, smoking, and any receptive sex (vaginal, anal, or oral), were evaluated as potential confounders in bivariate analyses.

Results

At baseline, a majority of participants self-reported Black race (N=49, 62%), use of hormonal contraception (N=42, 54%) and having at least one sex partner in the prior two months (N=61, 80%). Seven percent disclosed a BV diagnosis in the two months preceding enrollment, and no participants were excluded due to antibiotic treatment for symptomatic BV at enrollment. Most participants noted on daily diaries that they engaged in sexual activity (59%) and experienced menses (61%) between baseline and the post-TVUS visit. The most common findings from the TVUS radiology reports were uterine fibroids (26%) and ovarian cysts (14%), although many participants had no clinical findings (37%).

Amsel-BV was diagnosed at 20% (32/160) of visits, including 15% (12/80) of baseline and 25% (20/80) of post-TVUS visits (Table 1). McNemar’s exact test for paired proportions indicated that among those with fluctuations in BV status between baseline and post-TVUS (i.e., BV present at baseline but absent post-TVUS follow-up, or absent at baseline but present post-TVUS follow-up), participants were more likely to have BV (p=0.02, N=10) and a positive whiff test (p=0.04, N=9) after TVUS compared to baseline (Table 1), and modeling including all participants found the within-participant odds of having Amsel-BV and positive whiff test were 5.9-fold (95% CI:1.22–28.23) and 5.0-fold (95% CI:1.06–23.62) higher after versus before TVUS, respectively. Although no statistically significant differences were noted for clinician observation of clue cells (p=0.07, N=8), discharge (p=0.34, N=10) or high vaginal pH (p=0.39, N=12) thwith McNemar’s exact test, modeling found the odds of each were higher after TVUS versus before (Table 1). Except for menses, which was associated with both lubricant exposure and pH>4.5, none of the factors recorded in daily diaries were identified as confounders. The timing of the follow-up clinical appointment (four or fewer days after TVUS [N=49, 61%] versus 5–13 days [N=31, 39%]) did not affect results (p-values for interaction=0.53–0.75).

Table 1:

Odds of any bacterial vaginosis and individual Amsel criteria after versus before exposure to lubricant during TVUS in 80 reproductive-age women.

		Present at post-TVUS follow-up
	Present at baseline	No	Yes	OR	95% CI	p-value††
Amsel-BV	No (%)	59 (87%)	9 (13%)	Ref	-	-
	Yes (%)	1 (8%)	11 (92%)	5.88	1.22 – 28.23	0.03
Discharge	No (%)	58 (89%)	7 (11%)	Ref	-	-
	Yes (%)	3 (20%)	12 (80%)	2.16	0.60 – 7.78	0.23
pH > 4.5 †	No (%)	47 (85%)	8 (15%)	Ref	-	-
	Yes (%)	4 (16%)	21 (84%)	2.09	0.46 – 9.43	0.33
Whiff positive	No (%)	60 (88%)	8 (12%)	Ref	-	-
	Yes (%)	1 (8%)	11 (92%)	5.01	1.06 – 23.62	0.04
Clue cells	No (%)	59 (89%)	7 (11%)	Ref	-	-
	Yes (%)	1 (7%)	13 (93%)	4.13	0.89 – 19.07	0.07

^†Adjusted for same-day menses

^††From generalized linear mixed effects modeling

Black participants made up 62% of the cohort, but contributed 100% of Amsel-BV, positive whiff test and clue cell observations, 97% of abnormal discharge, and 87% of pH>4.5. The above findings were similar when the analysis was restricted to Black participants.

Discussion

This analysis investigated the impact of TVUS with hyperosmolal lubricant on BV diagnosis by assessing Amsel-BV and its individual diagnostic criteria before and after the procedure. Among the subset of participants whose Amsel-BV status changed between visits, both Amsel-BV and a positive whiff test were more frequently diagnosed after TVUS versus before, as indicated by McNemar’s exact test. Findings were consistent when modeling the odds of Amsel-BV and its individual criteria across all 80 participants, though the limited number of BV events likely contributed to the wide confidence intervals observed.

Several longitudinal observational studies have explored the effects of lubricant use on the vagina. In this cohort, we previously observed evidence suggesting that the vaginal epithelial barrier may be compromised for up to two weeks after TVUS, as indicated by a decrease in the maturity of shed vaginal epithelial cells,¹⁹ along with an increased risk of BV over 10 weeks among Black participants,¹⁶ and an increased risk of a low-Lactobacillus microbiota after TVUS among peri/post-menopausal participants and those with prior BV history.¹⁷ In a separate cohort, we reported a reduced relative abundance of Lactobacillus crispatus²² and elevated lipids associated with cellular damage²³ after sex with lubricant compared to sex without lubricant. Other studies have noted decreases in vaginal moisture 24 hours after lubricant application²⁴ but reported no differences in vaginal microbiome between moderately (~600–1200 mOsm/kg) and highly (~5100 mOsm/kg) hyperosmolal lubricant users over four weeks,²⁵ or in inflammatory markers in cervicovaginal lavage samples following up to two weeks of daily diaphragm use with a personal lubricant.²⁶ In this study, we aimed to examine Amsel-BV and its individual diagnostic criteria – outcomes of particular interest to clinicians diagnosing BV.

There are some limitations to the analysis. Sample sizes, including the number of Amsel-BV diagnoses and clinical criteria instances, were limited. In the single crossover study design, some observed changes in Amsel-BV or its criteria could reflect typical fluctuations over time⁸ rather than the effects of lubricant use. The median time between TVUS and the first follow-up visit was four days (range:1–13 days), but the length of follow-up did not appear to influence the results. Additionally, the GALE study recruited participants referred for TVUS, which may not represent the broader population’s personal lubricant use practices. Lastly, this study relied on Amsel clinical criteria and did not incorporate other diagnostic approaches, such as microbiologic tests, which could provide further insights.

Studying vaginal lubricant use in an observational setting is challenging due to variations in product composition, volume and frequency of use. This cohort allowed for evaluating the effects of a single use of a hyperosmolal lubricant similar to products commonly used in personal practice. Our findings suggest participants are more likely to develop Amsel-BV and a positive whiff test following exposure to a hyperosmolal lubricant. Future research should prioritize larger cohorts, randomized designs, and examining more frequent lubricant use. Additionally, comparative studies on hyperosmolal versus iso-osmolar lubricants are needed, focusing on their impact on the vaginal microenvironment, BV, and STI risk.