Hypothesis / aims of study
Introduction:
The use of pelvic mesh in prolapse and incontinence surgery is associated with complications including pain, dyspareunia, vaginal mesh exposure, and infection (1). Controversy exists surrounding the safety of mesh use for vaginal surgery (2).
Pain and mesh exposure through the vaginal epithelium are the most reported complications and a cause of litigation internationally (3). The mainstay of treatment for pelvic mesh complications is surgical removal of the mesh. However, it is not yet clear if this resolves symptoms.
There is paucity of data available to provide an evidence-based understanding of the complex relationship between the vaginal microbiome and foreign implants (2, 3). One potential risk factor for vaginal mesh extrusion is thought to be due to a primary failure of healing. It is postulated that microbial colonization and infection may be related to complications (3). Trillions of microorganisms inhabit the average healthy human body, most with low virulence (2). The incidence of infection related mesh complications is not well established with rates reported from 1-68% (3). The clinical challenge lies in establishing the difference between colonization of bacteria with infective source that leads to mesh complications (3). We seek to improve the understanding of the pathognomonic relationship between pelvic mesh exposure, pain, and complications with bacterial colonization. Thereby, improving the knowledge on the aetiology of pelvic mesh complications.
Primary Outcomes:
1. To report the rates of bacteria isolated in cultured microbiome from explanted pelvic mesh in
a cohort of Australian women referred to a tertiary pelvic mesh centre.
2. Investigate correlation between bacteria identified with rates of mesh exposure and pelvic pain
Secondary outcomes:
1. Using the brief pain inventory questionnaire assess the change in mean pain severity score before
and after pelvic mesh removal.
2. Rates of pre-existing chronic pain conditions including endometriosis, fibromyalgia, arthritis and
other pain conditions in patients undergoing pelvic mesh removal surgery. Rates of obesity, smoking
and compliance with prescribed vaginal estrogen were assessed.
Study design, materials and methods
We conducted a retrospective record review of 190 women who underwent pelvic mesh removal in our Australian tertiary centre. All patients that underwent complete or partial excisions of pelvic mesh implants over a 2.25 year period were included. Participants were identified through their hospital unique record number with their clinical electronic database of explanted pelvic mesh and cross validated by pathology records.
Data was collected from clinical notes, operation reports, pathology results and outcomes from the brief pain inventory questionnaire. The questionnaire was issued to each patient pre-operatively at initial consult and at 24 months (following completion of pelvic mesh removal surgery). Variables assessed were rates of endometriosis, fibromyalgia, pain conditions and diabetes. Baseline characteristics collected included age, body mass index (BMI), parity and smoking.
Analysis was conducted using Stata 17.0 (StataCorp). Descriptive statistics are reported for presentation of pain and the bacterial cultures isolated from explanted mesh.
To assess the relationship between bacterial culture with mesh exposure and bacterial culture with the presentation of pain, a chi square test was conducted. Some of the cultures were only present in one or two patients. In those cases, a Fisher’s exact test was used. P-values less than 0.05 indicate a statistically significant relationship between outcomes.
Pain scores were assessed using the brief pain inventory questionnaire. The response to this standardized questionnaire was analyzed using the mean severity score. The mean severity score at baseline and following mesh removal (at 24 months in the mesh service) were assessed using the Wilcoxon signed-rank test.
Results
In our cohort of women 67% had complete mesh excision and 33% had partial mesh excision. Of those, 63% were midurethral slings and 33% were vaginal prolapse meshes. 62% of patients presented with pain and 45% presented with mesh exposure.
Cultures from the explanted mesh itself found Staphylococcus (23.15 %) was the most frequently isolated bacteria. Followed by Mixed anaerobic bacteria (7.36%), Mixed enteric bacteria (6.3%), Streptococcus milleri group (Strep. milleri) (5.7%), Candida (3.7%), Actinomyces group and Escherichia coli (E.coli) (2.1%), Enterococcus faecalis (1.6%), Streptococcus (strep) agalactiae (1.05%). Only a small proportion (0.5%) of Aerococcus urinae, Enterobacter cloacae, Propionibacterium acnes, Bacteroides fragilis, Finegoldia magna, Mycoplasma, and Eikenella corrodens was cultured (see graph 1).
Pelvic mesh was found to be exposed at time of removal in 45.6% of the cohort. A statistically significant correlation between vaginal mesh exposure and bacteria cultured was identified in 5 bacteria types (see table 1). These included mixed anaerobic bacteria 10.84% (p-value 0.04), Strep milleri group 13.41%, (p-value <0.005), Actinomyces group 4.82% ( p-value 0.042), Candida 8.43% ( p-value 0.004) and Mixed enteric bacteria 10.84%, (p-value 0.04).
Of women presenting with pain, there was a statistically significant correlation with cultures of strep milleri group 2.7% (p-value 0.21) (see table 1). The brief pain questionnaire was completed by 33% (63/190) of patients. The mean severity score from the brief pain inventory (short form) questionnaire was found to be lower at 24 months when compared to the baseline, however the difference is not statistically significant (p=0.1695).
Fibromyalgia was reported in 11.5% of our population, arthritis 16%, endometriosis 10% and ‘other’ pain conditions of 29.4%. 89% of women undergoing pelvic mesh removal were obese (BMI >25), the median BMI of the cohort was 33.2. Over a third of our population were current smokers (32%) and only 32% were using vaginal estrogen.
Interpretation of results
We identified a total of 18 bacteria types in the cultured microbiome of explanted transvaginal mesh in our cohort. Whilst Actinomyoces has been shown to be associated with invasive bacterial disease associated with pelvic mesh complications, our retrospective cohort review identified an additional 4 bacteria groups associated with mesh exposures. Although some of the identified bacteria are found in the normal vaginal microbiome, it is plausible that others may have a pathogenic effect on vaginal mesh and subsequent mesh extrusion. We also found a statistically significant correlation between pelvic pain and the isolated culture of the Strep milleri group.
The authors acknowledge certain limitations to this study including the lack of standardized microbial culture reporting on pelvic mesh in our local laboratory. Unfortunately, only a small number of women completed both the baseline and 24 months questionnaire. This study was not powered to evaluate outcome therefore would have likely contributed to the non-statistically significant findings.
There were higher rates of patients presenting with pain conditions in our cohort of women compared to the general population. It is possible that pre-existing chronic pain conditions may predispose patients to experience pelvic mesh complications.