Dr P. L. Dwyer, Mercy Hospital for Women, 126 Clarendon Street, East Melbourne, 3002 Victoria, Australia.
Objective To determine the efficacy and safety of a new technique using Atrium polypropylene mesh (Atrium, Hudson, New Hampshire, USA) as an overlay graft for repair of large or recurrent anterior and posterior compartment prolapse.
Design A retrospective review of women who had vaginal prolapse surgery with Atrium mesh reinforcement.
Setting Tertiary referral urogynaecology unit in Australia.
Population Forty-seven women where mesh was placed under the bladder base with lateral extensions onto the pelvic sidewall, 33 women where a Y-shaped mesh was placed from the sacrospinous ligaments to the perineal body and 17 women who had mesh placement in both compartments.
Methods Women were assessed by site-specific vaginal examination pre-operatively and post-operatively at six weeks, six months and two years.
Main outcome measures All complications. Rate of recurrent prolapse assessed by the Baden–Walker halfway classification system.
Results Mean follow up was 29 months (range 6 to 52). Four of 64 women with anterior mesh placement (6%) developed a grade 2 asymptomatic cystocele. Five women (5%) required further surgery for recurrent prolapse at a non-mesh site. Erosion occurred in nine women (9%). Three healed after intravaginal oestrogen cream, five after excision of exposed mesh and vaginal closure and one woman also had surgical closure of a rectovaginal fistula. The risk of mesh erosion decreased over the study period. Urinary, coital and bowel symptoms were significantly improved following surgery.
Conclusions This technique shows promise in correcting pelvic organ prolapse. Vaginal mesh erosion is the most common complication and is related to surgical experience.
Genital prolapse is a common cause of surgery in women, which is frequently unsuccessful. In a recent Australian prevalence survey,1 46% of women aged 15–97 had symptoms of pelvic floor dysfunction and 23.7% of women had undergone a previous pelvic floor repair. A recent North American Study2 showed that the lifetime risk of surgery for prolapse or stress incontinence was 11% with one in three patients requiring more than one surgical repair.
The pelvic floor musculature is considered to be the most important factor in the maintenance of pelvic organ support with the fascia and ligaments providing secondary support.3,4 Weakness of the muscular pelvic floor can be caused by disuse, nerve damage or trauma but is generally not amenable to surgical repair. Dissatisfaction after surgery may be caused by recurrence of prolapse but also by the persistence or development of urinary, anorectal symptoms or dyspareunia caused by excessive scarring or vaginal skin excision.
Synthetic non-absorbable mesh has been used extensively for the abdominal repair of vaginal prolapse by placing mesh over the anterior and posterior vaginal wall and attaching the mesh to the anterior sacrum, providing good vaginal support.5 This success has encouraged us to explore the transvaginal use of mesh in order to improve the anatomical and functional results of pelvic floor surgery. In this study, we assessed the efficiency and complications of Atrium polypropylene mesh (Atrium, Hudson, New Hampshire, USA) in the repair of recurrent or large anterior and posterior compartment vaginal prolapse.
Ninety-seven patients had transvaginal surgery with Atrium mesh reinforcement performed for symptomatic vaginal prolapse between February 1999 and May 2002. Atrium mesh is a monofilament polypropylene woven of 92 g/m2 weight with a pore size (interfibre space) of 800 μm. Seventeen women had mesh placed both in the anterior and posterior vaginal mesh repair (APVMR); 47 women only in the anterior compartment (AVMR) and in 33 women only in the posterior compartment (PVMR). The indication for mesh usage was recurrent vaginal prolapse or where a large fascial defect was unsuitable for standard repair alone. Anterior and posterior Atrium mesh repairs were combined with a number of other pelvic floor surgeries, namely, vaginal hysterectomy (10), sacrospinous fixation (12), bilateral iliococcygeal fixation (4), abdominal sacrocolpopexy (1) and tension-free vaginal (TVT) sling (24).
All patients had a standardised urogynaecological history and examination performed before and after surgery including data on urinary, coital and bowel symptoms. Urodynamic evaluation was performed when indicated for urinary symptoms. All patients were examined vaginally with a Sims speculum in the left lateral and sitting positions with the Valsalva manoeuvre. Vaginal support was graded using the modified Baden–Walker halfway system.6 Prophylactic antibiotics were routinely given intravenously prior to surgery.
Data are presented as mean or median and range and as number (%). Differences are presented with their 95% confidence intervals. Approval was obtained for conducting this study from the hospital Ethics and Research Committee.
For an anterior compartment mesh repair, a midline incision was made along the anterior vaginal wall suburethrally to the vaginal apex and the bladder was separated from the vagina. This dissection was extended bilaterally to the ischial spines and advanced anteriorly along the arcus tendineus. Midline plication of the fascial layer was performed using interrupted 2/0 polydioxanone (PDS) sutures. Atrium mesh (2 × 15 cm) with a widened elliptical mid-portion (5 cm) was placed under the bladder base and each lateral extension was positioned onto the iliococcygeal fascia anterior to the ischial spines (Fig. 1A). The mesh was usually unsutured, although in four cases of complete vaginal eversion, a 1 PDS suture was placed into the iliococcygeal fascia using a Miya hook to attach the mesh and vagina as described previously.7 The mesh overlay was sutured with 2/0 Vicryl sutures at the anterior and posterior margins for stabilisation and to prevent folding. If there was coexistent urodynamic stress incontinence, a TVT sling procedure was performed according to Ulmsten et al.8
For a posterior compartment mesh repair, a midline incision from the perineum to the vaginal apex was made and the vagina detached from the rectum with sharp dissection, which was extended, laterally to the ischiorectal fossa and superiorly onto the sacrospinous ligament. If an enterocele was present, the sac was dissected out and opened; and then closed by high ligation using 2/0 PDS. Any fascial defect in the rectovaginal septum was repaired using 2/0 PDS suture. Atrium mesh (10 × 15 cm) was fashioned in a Y shape, the arms of the Y (2 cm wide), and the body (5 cm wide) as shown in Fig. 1B. The arms of the Y were placed onto the sacrospinous ligaments bilaterally with the main body of mesh overlaying the repaired rectovaginal fascia and the perineal body. The mesh was also stabilised with 2/0 Vicryl sutures placed superiorly, laterally and onto the perineal body. Rectal examination was performed routinely to exclude damage or the inadvertent placement of sutures into the rectum.
For a combined anterior and posterior compartment mesh repair, a midline vertical incision was made suburethrally extending along the full vaginal length to the perineum with wide lateral sharp dissection in the anterior, apical and posterior compartments. The standard repair was performed as outlined above for each compartment. Atrium mesh was fashioned into a cross shape and laid along the repair with the arms of the mesh extending along the iliococcygeal fascia anterior to the ischial spines and the long stem of the cross extending over the vaginal apex, repaired enterocele and rectovaginal fascia to the perineum (Fig. 1C).
In 12 patients with large vaginal eversions or enteroceles, sutures of 1 PDS were placed into the sacrospinous ligament, using a Miya hook, through the arms of the mesh and then through the full thickness of the vaginal epithelium.
Following placement of the mesh overlay, the vagina was closed. A cystourethroscopy and rectal examination were performed to exclude any urinary or rectal injury.
No financial assistance was received from any company in the design or execution of this study.
Atrium polypropylene mesh was used in anterior compartment repairs on 64 women, in the posterior compartment on 50 women and in both compartments on 17 women. The median age at the time of surgery was 61 years (range 30 to 86 years); the median parity was 3 (range 0 to 8) with three women being nulliparous. Many women had previous pelvic surgery including one or more vaginal repairs (48), sacrospinous fixation (15), abdominal sacrocolpopexy (3), vaginal (24) and abdominal hysterectomy (37). Fourteen had a previous Burch colposuspension and 2 a pubovaginal sling.
There were few serious complications. Two women had secondary post-operative haemorrhages (>500 mL) and one woman developed a rectovaginal fistula on the fifth day.
The mean follow up was 29 months (6 to 52 months). Ninety-seven women were reviewed at 6 weeks, 96 women at 6 months, 91 women at 12 months and 77 women had follow up of >24 months. The pre- and post-operative vaginal findings are shown in Table 1.
Table 1. Number of patients with significant vaginal prolapse (≥grade 2 Baden–Walker) pre- and post-operatively.
Repair with mesh
No. with pre-op prolapse ≥2 (%)
Follow up recurrent prolapse ≥2 (cumulative no. of patients)
6 months (94)
12 months (91)
≥24 months (77)
B = bladder, C = cuff, E = enterocele, R = rectum.
Of the 64 women who had polypropylene mesh placed in the anterior compartment, four women (6%) developed an asymptomatic grade 2 cystocele by 12 months and one woman (1.5%) a grade 2 cuff prolapse. Of the 17 women who had mesh reinforcement of both anterior and posterior compartments, one woman (6%) required a further anterior repair after 30 months for recurrent cystocele and repair of posterior mesh erosion. Two patients (4%) who had posterior compartment mesh repairs with standard anterior repair developed an asymptomatic grade 2 cystocele.
There was no recurrence of posterior compartment prolapse in the 50 women who had mesh reinforcement posteriorly, although four women (6%) who had anterior compartment mesh reinforcement developed significant symptomatic posterior compartment prolapse (≥grade 2).
There were nine mesh erosions (9%) overall: three erosions on the anterior vaginal wall and six on the posterior wall. Seven mesh erosions (78%) were diagnosed within six months. Three women were asymptomatic and the vaginal wall re-epithelised after administration of intravaginal oestrogen for between six weeks and three months. Six patients had symptoms of vaginal discharge, bleeding or dyspareunia and were cured after excision of the exposed mesh, undermining of the vaginal edges and then suture closure without tension. One woman had a rectovaginal fistula diagnosed one week post-operatively. On proctoscopy, the Atrium mesh was seen protruding through the rectal mucosa. The rectovaginal fistula was repaired transvaginally seven weeks later with partial excision of the Atrium mesh and a layered closure of the fistula with a Martius graft. Follow up over 30 months showed no recurrence of the fistula. Over the study period, the rate of mesh erosion decreased from 19% in 1999 (3 of 16) to 13% in 2000 (4 of 32) and 4% in 2001/2002 (2 of 49) (differences not significant).
Sixty-seven patients (69%) were sexually active pre-operatively and 66 (68%) were sexually active at the 24-month review. Dyspareunia was reported by 25 patients (25.8%) before surgery, 6 (6.3%) at 6 months, 7 (7.7%) at 12 months and 7 (9.1%) at 24 months (each change significant vs pre-op value, data not shown). In three patients, the dyspareunia occurred de novo following surgery (two AVMR and one PVMR).
Urinary and bowel functional symptoms pre- and post-operatively are shown in Table 2. Urinary and bowel symptoms were common before surgery and were significantly improved following surgery. At 24 months post-operatively, there were three patients with de novo urgency (two had an APVMR and one AVMR). Three patients had de novo urge incontinence (two had APVMR and one had PVMR). One patient had de novo constipation (AVMR).
Table 2. Functional symptoms pre- and post-operatively.
95% confidence intervals of percentage difference.
34.2 (21.7, 46.8)
38.0 (23.6, 52.4)
32.1 (18.9, 45.4)
19.1 (8.9, 29.2)
11.5 (0.7, 22.4)
31.7 (20.0, 43.4)
Twenty-five women had urodynamically confirmed stress incontinence. Twenty-four women had TVT slings and one had a Burch colposuspension. Three women redeveloped symptoms of stress incontinence after surgery (one PVMR and TVT; one PVMR; one APVMR). One patient had voiding difficulty at 24-month review that occurred de novo following AVMR and sling.
In this study, Atrium polypropylene mesh placed under the bladder base with extensions up the pelvic sidewall provided good medium term support for the anterior compartment with a low recurrence rate (6%) but with a small but significant morbidity.
Some pelvic surgeons believe that complete vaginal reconstruction of all compartments is necessary,9 but our philosophy with regard to reparative vaginal surgery has been to repair only areas where defective support was found. Seven patients (7%) in this series developed de novo prolapse during the follow up period in a compartment that clinically had previously appeared well supported. The prophylactic use of synthetic mesh for prolapse women with no anatomical defects requires further evaluation.
The major morbidity was vaginal erosion, which in one case possibly resulted in a rectovaginal fistula. Most cases were diagnosed in the first six months post-operatively and the majority were managed conservatively or with minor surgical intervention. Although fistula formation is a potential risk of mesh, we believe that this is an uncommon complication and we have no further fistulae with subsequent experience in 140 women using Vypro II mesh (Ethicon, Summerville, New Jersey, USA).
Synthetic meshes have been used for abdominal wall hernia surgery since the late 1950s.10,11 Nevertheless, gynaecologists have been reluctant to use synthetic mesh in reconstructive pelvic surgery because of the risk of infection and erosion into the urinary, genital or gastrointestinal tract. The synthetic non-absorbable meshes most commonly used have been polypropylene (Marlex, Prolene, Atrium), polyethylene terephthalate (Mersilene), and expanded polytetrafluoroethylene (PTFE) (Gore-Tex). Polypropylene mesh is monofilament while polyethylene and PTFE are multifilament synthetic meshes. The other important quality of synthetic mesh is their interfibre spaces (pore size), which should be of sufficient size not to harbour bacteria but allow macrophages and leucocytes to penetrate, while also permitting connective tissue ingrowth for permanent support. Bobyn et al.12 found that the best mechanical anchorage occurs when pore size was between 50 and 200 μm with an average of 90 μm. Currently, polypropylene mesh is the most widely used synthetic prosthesis in general surgery and gynaecology.13
In 1996, Julian14 reported a prospective randomised study of synthetic, non-absorbable mesh in 24 women with recurrent anterior wall prolapse. After two years, four women in the control group and none in the mesh group had recurrent anterior wall prolapse. Three women (33%) with mesh repair had erosions. Flood et al.15 placed a strip of Marlex mesh under the vesicourethral junction extending into the Cave of Retzius and found no anterior wall prolapse recurrences but three mesh erosions after three years. Migliari and Usai16 placed mixed fibre mesh (60% polyglactin and 40% polyester) under the urethra and bladder base of 15 women. Two patients developed posterior compartment prolapse after a mean follow up of 23 months.
Synthetic non-absorbable mesh has been used less frequently in the transvaginal repair of posterior compartment prolapse. Watson et al.17 placed a strip of polypropylene mesh (Marlex) between the rectum and the vagina for posterior compartment prolaspe. After a mean of 29 months, eight of nine patients had improved defecation. One patient had a wound infection and one developed dyspareunia post-operatively. There were no mesh erosions and the incidence of prolapse pre- and post-operatively was not stated.
Synthetic absorbable mesh has been used for prolapse in both compartments. Sand et al.18 reported a prospective randomised study of three strips of polyglactin 910 mesh (Vicryl mesh; Ethicon) placed on the anterior endopelvic fascia. At 12 months follow up, there was a significantly lower recurrence rate in the mesh group compared with the controls (25%vs 43%, P= 0.02). Weber et al.19 found no significant difference in recurrence of cystocele in their prospective randomised study using the same mesh.
The positioning of the mesh as much as the type used influences the anatomical and functional outcome. In this study, the mesh was laid over the fascial repair in both compartments. The use of extensions onto the pelvic sidewall or to the sacrospinous ligaments superiorly provided a strong continuous scaffold for fibrous tissue ingrowth extending from the pelvic floor ligaments and muscles superiorly through the rectovaginal septum onto the perineal body, thus mimicking the normal fascial supports of each compartment.
A standard ‘fascial’ repair was performed but no or little vaginal epithelium was removed. We found that the vagina adheres to the underlying mesh and the vaginal wall remodels to the appropriate site once the underlying pulsion pressures have been relieved, providing a better functional result while avoiding vaginal stenosis caused by excessive vaginal epithelial removal.
The functional results of surgery are as important as the anatomical outcome. In our study, the incidence of dyspareunia halved following surgery irrespective of compartment. Salvatore et al.20 found an increased incidence of dyspareunia following the use of Prolene mesh. In our study, the risk of new functional symptoms was low following vaginal mesh surgery. We believe that lateral dissection and wide placement of synthetic mesh overlay is important in preventing excessive folding and scarring in the anterior and posterior compartments.
The risk of mesh erosion varies depending on the type of mesh used and its placement. Other possible factors affecting the incidence of erosion are the use of prophylactic antibiotics, menopausal status and age, the use of pre-operative oestrogens and, post-operatively, pelvic infection. Current evidence would suggest that monofilament polypropylene has the lowest incidence of infection and erosion compared with the other non-absorbable meshes.13 Recently, lighter meshes of polypropylene and vicryl (Vypro I and II; Ethicon) have been introduced for abdominal hernia repair and a monofilament polypropylene mesh (Gynemesh PS; Ethicon) for pelvic reconstructive surgery. These are lighter (area weight 50 g/m2) compared with Atrium (92 g/m2), Prolene (85 g/m2) and Marlex (95 g/m2), more flexible, softer and easier to handle, and have been shown in animal studies to have less contracture and folding during healing. To date, there are no studies in humans to indicate the rate of erosion of these meshes.
This retrospective study described a new approach to the vaginal placement of synthetic mesh with a good anatomical and functional outcome but does not prove that mesh reinforcement provides better long term support than fascial repair alone. This would require a prospective randomised trial similar to a recently published prolapse study21 comparing the abdominal and vaginal approach. However, our experience is that even the most complicated genital prolapse can be successfully treated transvaginally using mesh while maintaining adequate vaginal size and function. We need to clarify when and where synthetic grafts should be used in the anterior and posterior vaginal compartments, how it should be best placed and the most appropriate type of prosthesis.
The authors would like to thank Dr Judith Goh for her contribution with the early collection of data.