Introduction:
Midurethral slings (MUS) provided women with an effective treatment option for stress incontinence, with >80% success rate. However, long term mesh complications (including chronic pain, vaginal mesh exposure and lower urinary tract (LUT) perforation) eventually came to light and led to the ban on the use of vaginally implanted mesh in many countries. Explanted vaginal mesh from those with complications show evidence of persistent chronic inflammation. MUS have been shown to be colonised with bacteria which persist as biofilms; the components of biofilms may alter the host implant response.
Changes in the vaginal and urinary microbiome have been related to various disease conditions (1–3). Current published literature regarding the microbiome and mesh complications either only assesses the vaginal microbiome or does not utilise next generation sequencing. Culture techniques are not well adapted to the detection of commensals within the vagina or nearby locations. As each body site carries its own unique microbiome, it is important to consider the mesh with local body sites. Little is understood about the composition of the mesh microbiome, or its relationship with the skin, urinary or vaginal microbiome, and mesh complications.
Hypothesis/ aims:
This study aimed to characterise the microbiome of MUS, and its association with mesh complications. Secondary objectives include assessment of changes in the local microbiome with corresponding complications: vaginal microbiome and vaginal mesh exposure, urinary microbiome and LUT perforation, and skin microbiome and corresponding pain sites.
STUDY DESIGN, MATERIALS AND METHODS
Women (n=74) provided samples of mesh, urine and swabs from the vagina and skin of the groin or suprapubic regions (n=397). These samples were allocated to an associated clinical group: chronic pain, vaginal mesh exposure, LUT perforation or recurrent incontinence (considered the control group). Cases of clinical mesh infection were excluded. Microbiome analysis was performed via next generation 16S amplicon sequencing targeting the V4 region. Data were analysed using QIIME2 and R Studio.
RESULTS
Relative abundances at the genus level by each type of sample are represented in Figure 1. Mesh samples had a higher species diversity compared to vaginal (p <0.0001) and groin (p <0.0001) swabs, but not urine. Although mesh and urinary bacterial communities were both more diverse, the compositions of the mesh and urinary microbiomes were still significantly different, p = 0.001.
The vaginal portion of the MUS had higher abundances of Atopobium, Campylobacter and Moryella than mesh arms.
The vaginal portion of the MUS from those with exposure had greater species diversity (p = 0.0201) than vaginal MUS which was covered, and a distinct microbial community composition (p = 0.0204) with more Actinomyces, Bifidobacterium, Dialister, Fusobacterium, Gardnerella, Oriibacterium and Peptostreococcus.
Vaginal swabs from those with vaginal mesh exposure also corresponded with greater species diversity (p = 0.006) and change in the community (p = 0.014), with greater abundances of Megasphaera, Parvimonas, Peptostreptococcus, Porphyromonas, Prevotella and Snethia. Vaginal swabs from those without a vaginal mesh exposure carried a significantly greater abundance of Lactobacillus.
Mesh arms associated with pain sites carried lower species diversity (p=0.0104) than sites not associated with pain, and a change in community (p=0.003). Actinobaculum and Prevotella were more common in the mesh arms which did not correspond to pain sites, whereas an unknown genus of Proteobacteria was more common in mesh arms corresponding with pain sites.
There was a change in the skin bacterial community in corresponding pain sites (p=0.024), with more Acintobacter, Caponcytophaga, Fusobacterium and Psuedomonas.
LUT perforation resulted in a significantly different urinary microbiome (p=0.011).
Unexpectedly, the microbiome of mesh was most similar to the LUT, not the vagina, with regard to species diversity. It is possible that incontinence at the time of implantation alters the vaginal microbiome by loss of Lactobacillus dominance.
Variations in microbial communities were found in both exposed vaginal portions of MUS and vaginal swabs from sites of exposure; however, these were not of the same bacteria. This suggests that not all bacteria are able to thrive on mesh, and the vagina provides a different environment to mesh itself. Interestingly, the vaginal microbiome in the absence of vaginal mesh exposure contained significantly more Lactobacillus than when there was a mesh exposure. It is not clear whether loss of Lactobacillus dominance occurs due to the exposure, as the implant alters the microbiome, or if the loss of Lactobacillus dominance is a risk factor for vaginal mesh exposure. Loss of Lactobacillus dominance in the vagina is often related to disease conditions.
A change in bacterial community was noted at each body site with each corresponding mesh complication, i.e., the vaginal microbiome with mesh exposure; urinary microbiome with LUT perforation and skin when corresponding to a pain site. Within the vagina and urinary tract, it is difficult to determine the direction of this relationship.
CONCLUDING MESSAGE
This original study defines the mesh microbiome. The mesh microbiome is more similar to the urinary microbiome with regards to bacterial diversity, but the bacterial community of the mesh microbiome remains significantly distinct from the vaginal, urinary and surrounding skin.
Mesh complications are associated with a change in microbial community at the complication sites: vaginal microbiome is significantly different in those with vaginal mesh exposure, the urinary microbiome is significantly different in those with LUT perforation and the skin microbiome is significantly different at a corresponding pain site.
FIGURE 1
