Vault prolapse and rectocele: assessment of repair using sacrocolpopexy with mesh interposition


Correspondence: Professor S. L. Stanton, St. George's Hospital, Urogynaecology Unit, Cranmer Terrace, London SW17 ORE, UK.


Objective To assess the sacrocolpopexy with mesh interposition in women with pelvic organ prolapse.

Design A prospective study.

Setting Tertiary referral urogynaecology and pelvic floor reconstruction unit.

Population Twenty-nine consecutive women with symptomatic vault prolapse and rectocele.

Main outcome measures Subjective and objective success rates and complications.

Results The mean age was 57 years. The mean number of past prolapse operations was 2.6 which included two past sacrospinous ligament fixations and 17 past posterior repairs. The mean follow up was 14 months. There was an increase in constipation from 41% to 50%, a decrease in faecal soiling from 21% to 10%, and an increase in incomplete defecation from 24% to 36%. Dyspareunia decreased from 38% to 17%, and there was some improvement in the stress and urge incontinence. There was a significant reduction of vault prolapse and rectocele (P < 0.001). All women with Stage II and Stage III vault prolapse were corrected, with an increase in Stage I prolapse from 20% to 27%. All women with Stage II and Stage III rectocele were corrected with a decrease in Stage I prolapse from 36% to 7%. The only significant interoperative complication was a cystotomy. One mesh became infected post-operatively which required removal.

Conclusions Sacrocolpopexy and mesh interposition is a safe and reliable operation for the correction of vault prolapse and rectocele. A long term follow up is necessary to detect any late complications.


In an ageing population the issue of prolapse, its prevention, and treatment are increasingly important. It is difficult to assess the incidence as these problems are often a source of embarrassment and patients may be reluctant to discus them. However, large epidemiological studies have found that prolapse may account for 10%–20% of all major gynaecological surgeries, and 13%–30% of all hysterectomies1. One study2 found an 11.1% lifetime risk of undergoing a single operation for pelvic organ prolapse. This same study found that 29% of patients will have recurrence requiring at least one re-operation. The true incidence of uterine and post-hysterectomy vault prolapse is unknown. Vault prolapse has been estimated to affect between 1% and 43% of patients following hysterectomy3 and is often accompanied by prolapse of the anterior or posterior vaginal walls.

Numerous surgical techniques have been described for the treatment of post-hysterectomy vaginal vault prolapse, yet none of them has been shown to be superior as a long term treatment which is safe, well-tolerated, and effective. A recent review of transvaginal repair of vault prolapse found over 40 different procedures in the literature4. Techniques include abdominal sacrocolpopexy, sacrospinous fixation, endopelvic fascia fixation, ileo-coccygeal fixation, vault suspension with rectus sheath5, and uterosacral ligament fixation6. All of these procedures need to be combined with a posterior colporrhaphy to address the posterior vaginal wall prolapse.

Sacrospinous fixation has been found to be an effective procedure with success rates ranging from 64%–97%7. This procedure can be complicated by haemorrhage, increased stress incontinence, urge incontinence, cystocele, dyspareunia, and buttock pain. Barks-dale et al.8 found that there were nerve fibres present throughout the length of the sacrospinous ligament, and that these may be related to post-operative pain. Benson et al.9 found evidence of pudendal and perineal nerve damage due to vaginal dissection during sacrospinous fixation, compared with abdominal sacrocolpopexy.

Sacrocolpopexy has also been found to be an effective procedure for the repair of vault prolapse, with success rates ranging from 84%–98%10–12. Complications include bleeding from the pre-sacral vessels, detachment or erosion of the mesh securing the vault to the sacrum and stress incontinence. If the patient has a rectocele than a separate posterior colporrhaphy has to be performed which involves repositioning the patient and an additional vaginal incision. Posterior colporrhaphy is associated with a risk of dyspareunia and of recurrence of the rectocele13.

The sacrocolpopexy with mesh interposition was designed to simplify the procedure and reduce the above complications by performing both operations using a single mesh via an abdominal approach. This was first described by Villet et al.14 in a presentation of 56 cases followed for 13.8 months with a 70% subjective and 86% objective cure, based on a physical examination. Cundiff et al.15 describe the same procedure in 11 women but with a mean follow up of only 11 weeks. Twenty one percent continued to have Stage I prolapse and 16% Stage II prolapse.

The aim of this report was to determine the role of sacrocolpopexy with mesh interposition in the correction of both vault prolapse and rectocele and note the benefits and complications.


This is a prospective and ongoing study of the safety and efficacy of the sacrocolpopexy with mesh interposition. The procedure (Fig. 1) is a sacrocolpopexy, with dissection of the rectum from the posterior vaginal wall down to the perineal body. The patient is positioned in the lithotomy Trendelenberg position, and the abdomen and perineum are sterilely draped. A 14 French gauge Foley catheter is inserted into the bladder, and an obturator is placed into the vagina to bring it to as near a vertical position as possible to facilitate the rectal dissection. Through a Pfannenstiel incision the peritoneum overlying the posterior vaginal wall is incised and the rectum separated from the vagina, down to the perineal body, using a combination of sharp and blunt dissection and diathermy. Once the perineal body is reached, the rectum becomes more difficult to separate from the posterior vaginal wall, particularly when there has been a previous posterior repair, and more bleeding is encountered. A 20 × 4 cm Teflon mesh (Bard Limited) is sutured, to the perineal body using Ethibond No. 1 sutures and then an additional two to three sutures are placed on either side of the mesh to secure it to the posterior vaginal wall up to the vault. If there is a significant cystocele without incontinence, a further piece of mesh is attached to the anterior vaginal wall after the bladder has been dissected distally. The peritoneum is then opened over the anterior longitudinal ligament of S1 or S2 vertebra depending upon ease of access, and the mesh is either tunnelled from the vaginal apex under the peritoneum to the ligament or peritoneal flaps are created which are then closed afterwards across the mesh. Blood vessels are carefully diathermied in front of the anterior longitudinal ligament and two Ethibond sutures are then inserted. The mesh is now laid along the posterior vaginal wall and the appropriate length is determined so that it lies entirely without tension. It is then sutured to the anterior longitudinal ligament and peri-tonealised after removing excess mesh. Patients are given a prophylactic 1 g metronidazole rectally 2 hours prior to surgery and 1 g cephradine intravenously on induction. Prophylactic heparin is started pre-operatively and continued until the patient is fully ambulant. The Foley catheter and a redivac drain are left in place for 24 hours post-operatively. If a colposuspension is performed at the same time, a suprapubic catheter is inserted. Most operations take up to one a half hours.

Figure 1.

Sagittal section of the female pelvis showing sacrocolpopexy and mesh correction of high cystocele, vault prolapse and rectocele. Anteriorly the mesh extends from the upper third of the anterior wall of the vagina to the apex and posteriorally from the perineal body alongside the posterior vaginal wall to the vault and from there to the anterior longitudinal ligament covering the first sacral vertebra.

The patients were seen six weeks post-operatively, and then yearly. A questionnaire was administered addressing the following symptoms: stress incontinence, urge incontinence, frequency, urgency, constipation, incomplete defecation, faecal incontinence, and dyspareunia (if sexually active). The prolapse exam was performed according to International Continence Society (ICS) ordinal stages of pelvic organ prolapse16 as 0-none; I—distal portion of the prolapse is > 1 cm above the hymenal ring; II—prolapse is (1 cm proximal or distal to the hymenal ring; III—prolapse is > 1 cm below the hymen, but no further than 2 cm less than the total vaginal length; and IV—complete or near complete (within 2 cm) vaginal eversion. Data was analysed using the Wilcoxon matched pairs signed rank sum test.


Twenty nine patients with a symptomatic vault prolapse and rectocele, three of whom had a cystocele, were seen over a three year period, and underwent repair with sacrocolpopexy with mesh interposition. The average follow up was 14 months (range 6–32). Demographics are presented in Table 1. Two patients had a past sacrospinous operation and 17 patients had a previous rectocele repair. Pre-operative and post-operative symptoms and signs are compared in Tables 2 and 3, respectively. Nine patients agreed to a colposuspension for stress incontinence. The remaining three patients did not agree to a colposuspension as their stress incontinence was not a problem. A bilateral salpingo-oophorectomy was performed in 18 women. Three patients with a Stage II high cystocele had an additional piece of mesh sutured to the anterior vaginal wall for a distance of nor more than 4 cm and then attached to the main mesh. There were no significant intra-operative complications aside from one cystotomy, which was repaired primarily, and one case of foot-drop, which resolved spontaneously. The mean estimated blood loss was 320 mL (50–1500) and the mean hospital stay was six days (3–8). Post-operative complications included one mesh infection, which required mesh removal. In seven patients, full ICS pelvic organ prolapse grading was available both pre and post-operatively, and in those, the genital hiatus was shortened by a mean value of 1.5 cm (0–3 cm). The perineal body was shortened by 0.4 cm mean (0–2 cm). The 9 patients with persistent stress incontinence were not troubled by it. When incontinence was present pre-operatively, pre and post-operative urodynamics were carried out. No new case of stress incontinence was confirmed. When comparing those patients who have not had previous prolapse surgery with those who had, there was no post-operative difference in subjective awareness of prolapse or urinary or bowel symptoms or objective presentation of cystocele, rectocele or vault prolapse.

Table 1.  Demographics.
Age (years)5744–76
Weight (kg)6751–84
Parity (median)20–6
Previous prolapse surgeries2.61–7
Table 2.  Pre- and post-operative symptoms.
Stress incontinence12 (41)9 (31) 
Urgency12 (41)16 (55) 
Urge incontinence12 (41)11 (38) 
Prolapse29 (100)3 (10)<0.001
Constipation12 (41)15 (50) 
Faecal soiling6 (21)3 (10) 
Incomplete defaecation7 (24)11 (36) 
Dyspareunia11 (38)5 (17) 
Table 3.  Pre- and post-operative signs.
Physical ExamPre-operativePost-operativeP
Vault prolapse   
 Stage I6 (20)8 (27) 
 Stage II12 (40)0 (0) 
 Stage III13 (43)0 (0) 
 TOTAL29 (100)8 (27)<0.001
 Stage I11 (36)2 (7) 
 Stage II17 (57)0 (0) 
 Stage III2 (7)0 (0) 
 TOTAL29 (100)2 (7)<0.001
 Stage II3 (11)0 (0) 


Similar procedures have been described in two other articles. Villet et al.14 reported on a series of 56 women undergoing a similar procedure. They claimed a 70% subjective (symptomatic) and an 86% objective (no prolapse or only Stage 1 prolapse on clinical examination) rate. Their procedure differed from ours in that they carried out the dissection of the rectovaginal septum only down to the level of the levators and used a 6 cm wide mersilene mesh but did not describe its reperitonealisation. Their major complications included one haematoma, one bowel occlusion requiring surgical intervention and bowel resection, six post-operative hernias, one prosthesis detached from the sacrum, resulting in failure of the repair, and one mesh erosion. They felt that the mesh erosion could have been avoided by using an absorbable vaginal suture, improving on the sterile technique and ensuring that the mesh was not under tension. An additional difference was that some patients had a hysterectomy and repair performed at the same time: those who had a simultaneous hysterectomy had a worse subjective outcome (13/34 failure rate). Finally, the procedure improved painful defecation in 10 of the 13 patients but worsened constipation in five. There was no explanation for this.

The main difficulty with this study is the paucity of description of data collection and analysis. It is uncertain if patients were followed prospectively or by retrospective chart review. There is no discussion of pre and post-operative symptoms making it difficult to draw conclusions about the dyschesia and constipation.

In a study by Cundiff et al.15, 19 patients were treated with a mean follow up of only 11 weeks. In their procedure a perineorrhaphy was performed in patients with separation of the superficial perineal muscles, and anterior and posterior vaginal fascia was approximated in appropriate patients. Two trapezoidal pieces of mesh, one placed anteriorly and the second posteriorly, were used and attached separately to the sacrum (to reduce the risk of mesh erosion and tension on the vagina). Finally, a Halban culdoplasty was performed.

Their pre- and post-operative evaluation included a review of symptoms of constipation, splinting, rectal prolapse, and faecal incontinence. The pelvic examination was according to the ICS pelvic organ prolapse standards, and six patients underwent pre-operative defaecography, three of whom had post-operative defaecography. Complications included one cystotomy and one mesh erosion. All patients had improvement in prolapse by at least one stage, although three had Stage II prolapse post-operatively. They found a 66% improvement in bowel symptoms and a significant improvement in measurement of the genital hiatus. Perineal descent was improved on defaecography. There was no comment on sexual function or urinary symptoms and urodynamic tests were only briefly mentioned in the final discussion by the author. The major point of the article, that perineal descent should be addressed in order to prevent further pudendal nerve damage and recurrence of prolapse, is quite important but objective evidence will be needed to confirm that this can be avoided.

As with any procedure, proper patient selection is important in obtaining good surgical results. We believe that this procedure is useful for patients with symptomatic vault prolapse and rectocele who wish to remain sexually active. It may be beneficial for patients with perineal descent as well.

We have not attempted to repair fascial defects in this context, partly because we are doubtful that this is an effective cure and have preferred to use the mesh as our main supporting mechanism. We are also sceptical about actually being able to find and demonstrate these defects. In attaching the mesh to the perineal body we agree with Cundiff et al.15, that it should limit perineal body descent and may prevent pudendal nerve stretch. So far, we have found it an effective method of correcting the anatomy of the rectocele and decreasing the feeling of prolapse but are dissatisfied by the persistence of bowel symptoms, save for a decrease in faecal soiling. We are encouraged by the decrease in dyspareunia - always a potential complication of rectocele surgery. We are concerned about the potential risk of infection involved by passage of suture material after the perineorraphy from the vaginal to the abdominal field as mentioned by Cundiff et al.15. We have not found the need to add a perineorraphy to this procedure and our long term results will determine whether this is correct or not. Three of our patients had an additional piece of mesh sutured to the anterior vaginal wall to correct a high cystocele: this has been effective at the time of this review and needs further long term follow up to determine whether it is a realistic method of controlling the cystocele.

The most important complication of sacrocolpopexy is haemorrhage and the formation of haematomas. Since 1961 over 39 papers have been published involving over 1300 patients. The incidence of haemorrhage and haematoma was 2.9%. Most occurred from the small veins and occasionally the artery on the anterior surface of the sacrum. We minimised this risk by careful dissection and the generous use of diathermy. Alternatively, the use of staples or bone anchors in some situations, may be as safe. When all else fails, arrest of bleeding by thumbtacks is useful. The other sites of bleeding are from the peritoneal edges in the dissection of the return from the vagina and in creating peritoneal flaps to cover the mesh. This is usually venous and can be arrested by diathermy.

The second commonest complication is mesh erosion and infection which is found in about 1.7% of cases. It is sometimes difficult to know whether erosion or infection has come first. Erosion may be minimised by not placing the mesh under any tension, using a wide mesh and suturing it over a large area with multiple sutures to reduce tension at any one area. The avoidance of infection means eschewing the vaginal route for any part of the mesh insert, pre-operative antibiotics, a careful sterile technique and avoidance of passage of permanent sutures through the vaginal mucosa. Amid17 in classifying artificial meshes, notes that Type III or macroporous prosthesis with multifilamentous or microporous components (e.g. Teflon, Mersilene) and braided suture material can harbour bacteria and, therefore consideration should be given to using a Type I or totally macroporous prosthesis (e.g. Prolene, Marlex), containing pores of > 75μ which deter housing and growth of bacteria partly by admitting macrophages but also by promoting rapid fibroplasia and angiogenesis within the larger pores. These also prevent infiltration and growth of bacteria. The uncomfortable handling characteristics of Type I meshes compared with Type III meshes governed the use of Teflon here.

Less common complications include bowel adhesion to the mesh which, we have avoided by peri-tonealising the mesh. Stress incontinence, which occurred in one patient and worsened in four in the previous study on sacrocolpopexy has not been encountered here12, which may be due to a change in technique of mesh insertion whereby the mesh was inserted loosely and without tension. Amid17 mentions that mesh may contract by up to 20% of its length which is another important reason for placing the mesh very loosely.

The lack of improvement in incomplete defecation and constipation was a disappointing aspect of our results. This may be a reflection of underlying bowel motility disorder: Kahn and Stanton13 in a retrospective study on rectocele repair, showed that while the symptom and sign of rectocele improved, the symptom of incomplete emptying and constipation did not. Alternatively, it may be due to a lack of sensitivity of our questionnaire: patients were asked to define symptoms as mile, moderate or severe. We did not define these adequately. Furthermore, there were some patients who had mild pre-operative symptoms which improved but did not resolve, therefore, these patients would have an unchanged classification. To overcome this we will add a visual analogue score to our questionnaire in an ongoing study and use pre and post-operative isotope defaecography in those patients with pre-operative bowel symptoms.


Sacrocolpopexy with mesh interposition appears to be an effective, safe and well tolerated procedure. Persistence and sometimes worsening of bowel symptoms and the applicability of an anterior mesh to correct cystocele and the role of objective bowel investigations need to be determined.