Systematic Review of the Efficacy and Safety of Using Mesh Or Grafts in Surgery for Anterior And/Or Posterior Vaginal Wall Prolapse

Systematic review of the efficacy and safety of using mesh or grafts in surgery for anterior and/or posterior vaginal wall prolapse

Xueli Jia, Cathryn Glazener, Graham Mowatt, Graeme MacLennan, Cynthia Fraser, Jennifer Burr

October 2007

INTERVENTIONAL PROCEDURES PROGRAMME NATIONAL INSTITUTE FOR HEALTH AND CLINICAL EXCELLENCE

REVIEW BODY REPORT

Title Systematic review of the efficacy and safety of using mesh or grafts in surgery for anterior and/or posterior vaginal wall prolapse

Produced by Health Services Research Unit University of Aberdeen 3rd Floor, Health Sciences Building Foresterhill Aberdeen AB25 2ZD

Authors Xueli Jia, Cathryn Glazener, Graham Mowatt, Graeme MacLennan, Cynthia Fraser, Jennifer Burr

Correspondence to Xueli Jia Research Fellow Health Services Research Unit University of Aberdeen 3rd Floor, Health Sciences Building Foresterhill AB25 2ZD Tel: (01224) 559801 Fax: (01224) 554580 Email: [email protected]

Date completed October 2007

‘Home unit’ details

The Health Services Research Unit (HSRU) is a multidisciplinary research group of about 50 people based in the University of Aberdeen. The Unit is core-funded by the Chief Scientist Office of the Scottish Government Health Directorates, and has responsibility for the following general remit:

1. To study or evaluate clinical activities with a view to improving effectiveness and efficiency in health care; 2. To work for the implementation of proven changes in clinical activities; 3. To encourage and support similar work thoughout Scotland; 4. To train NHS staff in Scotland, and others, in the principles and practice of health services research in general, and health care evaluation in particular.

The Unit has an established portfolio of health services research focusing on two main programmes – health care assessment and delivery of care. The Unit is one of the two research groups that make up the ‘Review Body’ for the National Institute for Health and Clinical Excellence Interventional Procedures Programme. The other is based at Health Services Research, University of Sheffield.

Contributions of authors

Xueli Jia screened the search results, contacted manufactures, assessed studies for inclusion, undertook data abstraction and quality assessment, conducted meta-analysis, and drafted the review. Cathryn Glazener drafted the scope, determined outcome categories, provided advice on assessing studies for inclusion, conducting meta-analysis, and on drafting of the review, drafted the discussion part, and commented on drafts of the review. Graham Mowatt commented on the scope of the review, drafted letters for contacting mesh/graft manufacturers for additional information, supervised the conduct of the review, and commented on drafts of the review. Graeme MacLennan conducted the statistical analysis, drafted the data analysis section of the review, and commented on drafts of the review. Cynthia Fraser developed and ran the literature search strategies, obtained papers, formatted the references, and drafted sections concerning search strategies and search results. Jennifer Burr supervised the conduct of the review, and commented on drafts of the review.

ii Conflict of interest

Cathryn Glazener was an author on one conference abstract that was included in the review.

Acknowledgements

We thank Chris Bain (Consultant Obstetrician and Gynaecologist, Aberdeen Royal Infirmary) for providing specialist advice and commenting on the draft of the report. The Health Services Research Unit receives a core grant from the Scottish Government Health Directorates. The views expressed are those of the authors and not necessarily those of the funding bodies.

Special thanks also go to James Browning (Mpathy Medical Devices Ltd.), Christine Clarke (Bard Ltd.), Hazel Edwards (American Medical Systems (UK), Ltd.), Adrian Griffin (Johnson & Johnson Medical Ltd.), William Hynes (WL Gore & Associations (UK) Ltd.), and Coloplast Ltd. for providing properties of mesh/graft produced by the manufacturers and published studies relating to mesh/graft. We also thank Paul Moran for providing additional information for a study identified from the National Research Register database.

iii TABLE OF CONTENTS

Executive summary xi List of abbreviations xxiv

1 OBJECTIVE OF THE REVIEW 1

2 BACKGROUND 2 2.1 Description of the underlying health problem 2 2.1.1 Epidemiology 2 2.1.2 Burden of disease 2 2.1.3 Aetiology, pathology and prognosis 3 2.2 Current management and alternative procedures 3 2.3 The interventional procedure under review 4 2.3.1 Description of the interventional procedure 4 2.3.2 Suitable candidates and relevant subgroups 7 2.3.3 Proposed clinical indications/contraindications and putative impact of the 7 procedure 2.3.4 Personnel involved and skill/experience required 8 2.3.5 Current use in the UK 8 2.3.6 Equipment or devices required 9

3 METHODS FOR REVIEWING SAFETY AND EFFICACY 10 3.1 Search strategy 10 3.2 Inclusion and exclusion criteria 11 3.2.1 Types of studies 11 3.2.2 Types of participants 11 3.2.3 Types of interventions 12 3.2.4 Types of outcomes 12 3.3 Quality assessment strategy 14 3.4 Data extraction strategy 14 3.5 Data analysis 14

4 RESULTS 17 4.1 Number and type of included studies 17 4.1.1 Included studies 18 4.1.2 Ongoing studies 22

iv 4.2 Number and type of excluded studies; reason for exclusion 24 4.3 Quality of available evidence 24 4.4 Anterior repair 26 4.4.1 Efficacy 26 4.4.1.1 Subjective failure: persistent prolapse symptoms 29 4.4.1.2 Objective failure: recurrent prolapse at original site 30 4.4.1.3 De novo prolapse (i.e. new prolapse at other sites) 34 4.4.1.4 Need for further surgery for prolapse 35 4.4.1.5 Persistent urinary incontinence 36 4.4.1.6 Persistent bowel symptoms 36 4.4.1.7 Persistent dyspareunia 36 4.4.1.8 Improvement of disease-specific quality of life (RCTs and non-randomised 36 comparative studies only 4.4.2 Safety 37 4.4.2.1 Blood loss 40 4.4.2.2 Damage to surrounding organs during operation 41 4.4.2.3 Mesh/graft erosion 41 4.4.2.4 Further operation needed for mesh/graft erosion 42 4.4.2.5 De novo urinary incontinence 43 4.4.2.6 De novo bowel symptoms 43 4.4.2.7 De novo dyspareunia 44 4.4.2.8 Infection 44 4.4.2.9 Other serious adverse effects 44 4.4.3 Other outcomes (operation time and surgeon experience) 45 4.4.3.1 Operation time 45 4.4.3.2 Surgeon experience 45 4.4.4 Conclusions of the evidence for and against the use of mesh/graft in anterior 46 vaginal wall prolapse repair 4.5 Posterior repair 49 4.5.1 Efficacy 49 4.5.1.1 Subjective failure: persistent prolapse symptoms 52 4.5.1.2 Objective failure: recurrent prolapse at original site 52 4.5.1.3 De novo prolapse (i.e. at other sites) 54 4.5.1.4 Need for further surgery for prolapse 54 4.5.1.5 Persistent urinary incontinence 54 4.5.1.6 Persistent bowel symptoms 54 4.5.1.7 Persistent dyspareunia 54

v 4.5.1.8 Improvement of disease-specific quality of life (RCTs and non-randomised 55 comparative studies only) 4.5.2 Safety 55 4.5.2.1 Blood loss 58 4.5.2.2 Damage to surrounding organs during operation 58 4.5.2.3 Mesh/graft erosion 59 4.5.2.4 Further operation needed for mesh/graft erosion 59 4.5.2.5 De novo urinary incontinence 59 4.5.2.6 De novo bowel symptoms 59 4.5.2.7 De novo dyspareunia 60 4.5.2.8 Infection 60 4.5.2.9 Other serious adverse effects 60 4.5.3 Other outcomes (operation time and surgeon experience) 61 4.5.3.1 Operation time 61 4.5.3.2 Surgeon experience 61 4.5.4 Conclusions of the evidence for and against the use of mesh/graft in posterior 61 vaginal wall prolapse repair 4.6 Anterior and/or posterior repair 64 4.6.1 Efficacy 64 4.6.1.1 Subjective failure: persistent prolapse symptoms 66 4.6.1.2 Objective failure: recurrent prolapse at original site 66 4.6.1.3 De novo prolapse (i.e. at other sites) 68 4.6.1.4 Further operation needed for prolapse 68 4.6.1.5 Persistent urinary incontinence 68 4.6.1.6 Persistent bowel symptoms 68 4.6.1.7 Persistent dyspareunia 68 4.6.1.8 Improvement of disease-specific quality of life (RCTs and non-randomised 69 comparative studies only) 4.6.2 Safety 69 4.6.2.1 Blood loss 72 4.6.2.2 Damage to surrounding organs during operation 72 4.6.2.3 Mesh/graft erosion 72 4.6.2.4 Further operation needed for mesh/graft erosion 73 4.6.2.5 De novo urinary incontinence 74 4.6.2.6 De novo bowel symptoms 74 4.6.2.7 De novo dyspareunia 74 4.6.2.8 Infection 74

vi 4.6.2.9 Other serious adverse effects 75 4.6.3 Other outcomes (operation time and surgeon experience) 75 4.6.3.1 Operation time 75 4.6.3.2 Surgeon experience 75 4.6.4 Conclusions of the evidence for and against the use of mesh/graft in anterior 75 and/or posterior vaginal wall prolapse repair

5 DISCUSSION 78 5.1 Discussion of main results 78 5.1.1 Anterior vaginal wall prolapse repair 78 5.1.2 Posterior vaginal wall prolapse repair 79 5.1.3 Anterior and/or posterior vaginal wall prolapse repair 80 5.1.4 Generis issues related to mesh 81 5.2 Assumptions, limitations, and uncertainties 82

6 CONCLUSIONS 85 6.1 Implications for the NHS 85 6.2 Implications for further research 86

7 REFERENCE LIST 88

vii LIST OF APPENDICES

Appendix 1 Search strategies 96 Appendix 2 Checklist of quality assessment of randomised controlled trials 102 Appendix 3 Checklist of quality assessment of non-randomised studies 103 Appendix 4 Reference list of included studies 104 Appendix 5 Characteristics of the included studies by study design 110 Appendix 6 Summary of properties of mesh/graft used by included studies 152 Appendix 7 Summary of operation techniques used by included studies 158 Appendix 8 Number and type of excluded studies; reasons for exclusion 163 Appendix 9 Detailed quality assessment results for included studies: randomised 170 controlled trials (full-text) Appendix 10 Detailed quality assessment results for non-randomised comparative 171 studies Appendix 11 Detailed quality assessment results for registry and case series 172 Appendix 12 Summary of outcomes: anterior prolapse, efficacy 173 Appendix 13 Summary of outcomes: anterior prolapse, safety 178 Appendix 14 Summary of outcomes: anterior prolapse, operation time 184 Appendix 15 Summary of outcomes: posterior prolapse, efficacy 185 Appendix 16 Summary of outcomes: posterior prolapse, safety 187 Appendix 17 Summary of outcomes: posterior prolapse, operation time 191 Appendix 18 Summary of outcomes: anterior and/or posterior prolapse, efficacy 192 Appendix 19 Summary of outcomes: anterior and/or posterior prolapse, safety 195 Appendix 20 Summary of outcomes: anterior and/or posterior prolapse, operation time 199

viii LIST OF TABLES

Table 1 Classification of vaginal wall prolapse stages 3 Table 2 The different mesh and grafts in vaginal wall prolapse repair 6 Table 3 Relative indications and contraindications to the use of mesh or graft 7 Table 4 Surgeons’ choice of procedure and use of mesh or graft in anterior and 8 posterior vaginal wall prolapse repair Table 5 Search results 17 Table 6 Papers selected for full assessment 18 Table 7 Number of studies and reports included 19 Table 8 Characteristics of women treated with mesh or graft and treatment details 21 Table 9 Characteristics of ongong studies 23 Table 10 Anterior repair, efficacy: summary of crude event rates, by type of 27 mesh/graft, by study design Table 11a Comparison of odds ratios of objective failure between different study 34 designs Table 11b Bayesian meta-analysis models, anterior repair: objective failure, recurrent 34 prolapse at original site Table 11c Indirect comparison, anterior repair: objective failure, recurrent prolapse at 34 original site Table 12 Anterior repair, safety: summary of crude event rates, by type of mesh/graft, 38 by study design Table 13 Posterior repair, efficacy: summary of crude event rates, by type of 50 mesh/graft, by study design Table 14 Bayesian meta-analysis models, posterior repair: objective failure, recurrent 53 prolapse at original site Table 15 Posterior repair, safety: summary of crude event rates, by type of 56 mesh/graft, by study design Table 16 Anterior and/or posterior repair, efficacy: summary of crude event rates, by 65 type of mesh/graft, by study design Table 17 Anterior and/or posterior repair, safety: summary of crude event rates, by 70 type of mesh/graft, by study design

ix LIST OF FIGURES

Figure 1 Summary of quality assessment of the RCTs (full-text, n=6) 24 Figure 2 Summary of quality assessment of the non-randomised comparative studies 25 (n=7) Figure 3 Summary of quality assessment of the registry (n=1) and the case series 25 (n=24) Figure 4 Anterior repair, efficacy, subjective failure, mesh or graft versus procedures 29 without mesh/graft: evidence from RCTs Figure 5 Anterior repair, efficacy, objective failure, mesh or graft versus procedures 30 without mesh/graft: evidence from RCTs Figure 6 Anterior repair, efficacy, objective failure, different types of mesh or graft: 32 evidence from RCTs Figure 7 Anterior repair, safety, blood loss (ml), mesh or graft versus procedures without 40 mesh/graft: evidence from RCTs Figure 8 Efficacy of anterior repair, summary of crude event rates (95% CI, any study 46 design), by type of mesh/graft Figure 9 Safety of anterior repair, summary of crude event rates (95% CI, any study 47 design), by type of mesh/graft Figure 10 Posterior repair, efficacy, objective failure, mesh or graft versus procedures 53 without mesh/graft: evidence from RCTs Figure 11 Efficacy, posterior repair, summary of crude event rates (95% CI, any study 62 design), by type of mesh/graft Figure 12 Safety, posterior repair, summary of crude event rates (95% CI, any study 62 design), by type of mesh/graft Figure 13 Anterior and/or posterior repair, efficacy, objective failure, mesh or graft versus 67 procedures without mesh/graft: evidence from RCTs Figure 14 Efficacy of anterior and/or posterior repair, summary of crude event rates (95% 77 CI, any study design), by type of mesh/graft Figure 15 Safety anterior and/or posterior repair, summary of crude event rates (95% CI, 77 any study design), by type of mesh/graft

x EXECUTIVE SUMMARY Background

Pelvic organ prolapse (POP) is a protrusion of one or more pelvic organs (bladder, rectum, uterus, vaginal vault, bowel) through vaginal fascia into the vagina and the downward displacement (‘prolapse’) of the associated vaginal wall from its normal location to or outside the vaginal opening. POP affects a woman’s quality of life by its local physical effects (pressure, bulging, heaviness or discomfort) or its effect on urinary, bowel or sexual function. POP can be classified according to the compartment affected as: anterior vaginal wall prolapse (urethrocele, cystocele); posterior vaginal wall prolapse (rectocele, enterocele); prolapse of the cervix or uterus; and prolapse of the vaginal vault (which can only occur after prior hysterectomy). A woman can present with prolapse of one or more of these sites. The present review focuses on anterior and posterior vaginal wall prolapse.

Current treatment options for anterior and posterior vaginal wall prolapse include pelvic floor muscle training (PFMT), use of pessaries (mechanical devices such as rings or shelves), and surgery including anterior or posterior colporrhaphy and site-specific defect repair. Surgery can be augmented with implantation of mesh or graft materials, with the aim of reducing the risk of failure.

There are numerous types of mesh and graft materials available, which vary according to type of material, structure, and physical properties such as absorbability. There are no existing classification systems for mesh and grafts. We have used the term ‘mesh’ for synthetic material and ‘graft’ for biological material. For the purposes of this review we have defined four classes of mesh/graft: absorbable synthetic mesh; absorbable biological graft; combined absorbable/non-absorbable mesh/graft (termed ‘combined’ hereafter); and non- absorbable synthetic mesh.

Mesh or graft repair is theoretically suitable for any degree of symptomatic anterior and/or posterior vaginal wall prolapse. It was first introduced in response to the high failure rate in both primary and secondary procedures: about 30% of women need an operation for recurrent prolapse. In the UK, it has been most often used for women with recurrent prolapse or with congenital connective tissue disorders (such as Ehlers-Danlos or Marfan’s syndrome).

The technique for implanting mesh or graft varies widely between gynaecologists. It can be positioned and sutured over the fascial defect as an ‘inlay’, or the whole vagina can be

xi surrounded by mesh/graft (‘total mesh’). In addition, it is now possible to insert mesh by using ‘introducer kits’. These kits involve blind penetration of pelvic spaces by trocars in order to thread mesh into supportive positions.

Objective

To systematically review the evidence for the efficacy and safety of mesh or grafts in surgery for anterior and/or posterior vaginal wall prolapse. The specific objectives were to compare: (a) efficacy and safety between procedures using mesh/graft (any type) and no mesh/graft, and (b) efficacy and safety between different types of mesh/graft.

Methods

We searched electronic databases, conference proceedings and relevant websites, contacted experts in the field and scanned bibliographies of retrieved papers. Extensive highly sensitive electronic searches were conducted to identify reports (both full text papers and conference abstracts) of published and ongoing studies on the safety and efficacy of mesh/graft used in the repair of pelvic organ prolapse. Searches were restricted to publications from 1980 onwards and to those published in the English language. The search strategies were designed to retrieve all relevant publications for the repair of pelvic organ prolapse: both anterior and posterior for the present review, and cervical, uterine and vaginal vault prolapse repair for a subsequent review. Conference proceedings from five major urogynaecological organisations were searched from 2005 onwards for additional reports of randomised controlled trials (RCTs). Reference lists of all included studies were scanned to identify additional potentially relevant studies. Eleven manufacturers were identified and contacted for information on properties of mesh/graft produced and for any studies related to mesh/graft.

We considered the following types of studies: • RCTs (full text and conference abstracts); • Non-randomised comparative studies (full text only); • Case series with a minimum sample size of 50 women (full text only). Studies with a mean follow-up time of one year or over were included for both efficacy and safety. One year was considered a minimum adequate period of time in which to assess the efficacy of prolapse repair. Studies with a mean follow up of less than one year were included for safety outcomes only.

xii • Population-based registry reports that prospectively collected audit data on the use of mesh or graft for anterior and posterior vaginal wall prolapse.

We did not search for studies that only reported procedures without mesh/graft.

One reviewer screened titles/abstracts and undertook data extraction. The same reviewer assessed the quality of the included studies using one of two separate checklists according to study design. Both checklists were developed through the Review Body for Interventional Procedures.

Data analysis was conducted for three subgroups of women according to the type of prolapse being repaired: • Anterior vaginal wall prolapse repair; • Posterior vaginal wall prolapse repair; and • Anterior and/or posterior vaginal wall repair (not reported separately).

A meta-analysis of RCTs, using Cochrane Collaboration Review Manager (RevMan 4.2) software, was conducted to directly compare the efficacy and safety of procedures using mesh/graft versus procedures without mesh/graft, and different types of mesh or graft with each other.

Crude event rates (and 95% confidence intervals) for each of the intervention categories were tabulated by summing across studies for all outcomes, and also according to study design (RCT, non-randomised comparative studies, case series/registry). A binominal distribution approximation was used to calculate 95% confidence intervals.

Additionally, Bayesian meta-analysis models were used to model the objective failure rates for the different interventions. This approach to evidence synthesis enabled data from RCTs and non-randomised comparative studies to be combined in analyses. The specific type of model used was a (Bayesian) binomial random effects model. Differences between interventions were assessed by the corresponding odds ratio and 95% credible interval (Crl). Crls are the Bayesian equivalent of confidence intervals. The results from these meta- analysis models are referred to as ‘unadjusted’ odds ratios.

Meta-analysis models that adjusted for study type were also used. The models were extended by including one study type effect to allow for the differences in reliability between

xiii non-randomised comparative studies and RCTs. The results from these models are referred to as ‘adjusted’ odds ratios.

We carried out, where data permitted, ‘head to head’ indirect comparisons of objective failure rates of the different mesh types. These are reported as odds ratios and 95% CrIs adjusted for study design.

WinBUGS software was used to produce the Bayesian meta-analysis models.

Results

Number and quality of included studies

From the initial 1633 publications identified by the search strategy, 49 studies (reported in 67 publications) were included, of which six were full-text RCTs, 11 were RCTs available as conference abstracts, seven were non-randomised comparative studies, one was a prospective registry, and 24 were case series with a minimum sample size of 50 women. Seven case reports and 25 case series with a sample size of less than 50 were identified but excluded in accordance with our inclusion criteria. Six manufacturers provided data on mesh/graft properties and related studies, all of which had already been identified by our searches. The table below shows the 49 included studies by type of prolapse repair and by study design. For the 17 RCTs, 14 compared mesh/graft with no mesh/graft, and three compared different types of mesh/graft.

Study design Studies Anterior Posterior Anterior and/or repair repair posterior repair1 Efficacy and safety RCTs (full text) 6 5 2 0 RCTs (conference abstracts) 11 7 1 3 Non-randomised comparative studies 7 4 2 1 Case series with sample size ≥50 and mean follow up ≥1y 162 10 2 5 Subtotal 40 27 7 9 Safety only Registry, mean follow up <1y 13 1 1 1 Case series with sample size ≥50 but mean follow up <1y 8 2 1 4 Total 49 30 9 14 1Anterior and posterior repair not reported separately. 2One case series reported anterior and posterior repair separately. 3The registry report reported anterior, posterior, anterior and posterior, and anterior or posterior separately.

xiv Eight ongoing RCTs and one other ongoing registry were also identified.

The included studies took place during the period 1996 – 2007 and in 12 countries. The median follow up was 13 months (range 1 to 51 months). Thirty-one studies had a median follow up of one year or more. In total, 4569 women were treated with mesh or graft. In studies providing this information, the mean age was 64 years (range 24 to 96 years). Seventy-two percent of repairs were primary procedures. The most common use of mesh or graft was for anterior repair (54%, 2472/4569). Overall, just over half of the studies used non-absorbable synthetic mesh (51%, 2320/4569) but for anterior repair alone and for posterior repair alone, absorbable biological graft was the most common alternative (46% (1124/2472) and 29% (121/417) respectively). Fifty-six percent (1404/2497) of women had a concomitant procedure for urinary incontinence and 37% (953/2583) had a hysterectomy.

Only the methodological quality of the full text studies was assessed. The methodological quality of the six RCTs was generally high. Adequate approaches to sequence generation for randomisation were reported in all studies except one. Concealment of treatment allocation was adequate in all RCTs except two. All follow-up periods were one year or more. All studies used intention-to-treat analysis in that women were analysed in the groups to which they were randomised. For the seven included non-randomised comparative studies, mean follow up was less than one year in two studies. For the registry and case series, mean follow up was one year or more in 17 studies. The remaining eight studies were included only for the safety outcomes. None of the studies reported how women were followed up (e.g. clinical review, case note survey, questionnaires) but most studies reported objective prolapse assessment, implying observer measurement. The drop-out rates ranged from 0 to 30%.

Summary of evidence: anterior repair

Thirty studies involving 2472 women provided data on the use of mesh/graft for anterior repair (12 RCTs, 4 non-randomised comparative studies, 1 registry, and 13 case series). Four studies used absorbable synthetic mesh, 14 studies used absorbable biological graft, one study used combined mesh/graft, and 14 studies used non-absorbable synthetic mesh. The median follow-up time was 14 months (range 1 to 38 months). Of the 12 RCTs, 7 were reported only as conference abstracts. Two RCTs and one non-randomised comparative study compared different types of mesh/graft and the others compared mesh/graft with no mesh/graft.

xv Efficacy

Figure A Efficacy of anterior repair, summary of crude event rates (95% CI, any study design), by type of mesh/graft

9/10, 5/49, 13/14, 17/44 100 90 80 70 60 50 40 184/640, 63/273, 186/1041, 48/548 NA, NA, 8/58, 8/45 30 20 19/179, 5/112, 36/484, 1/55 2/85, 16/174, 9/280, 3/234 Crude rate (%) rate Crude 10 0 Subjective Objective failure de novo Further Persistent Persistent Persistent failure prolapse operation urinary bow el symptom dyspareunia (rectocele) needed for incontinence (no data) (no data) prolapse (stress urinary (recurrent or de incontinence) novo)

No mesh/graft Absorbable synthetic mesh Absorbable biological graft Non-absorbable synthetic mesh

There were too few data reported for most outcomes (the primary efficacy outcome, subjective failure; and the secondary outcomes: de novo prolapse; further operation for recurrent or de novo prolapse; persistent urinary or bowel problems; and sexual (dyspareunia) problems) to draw reliable conclusions (Figure A).

However, in 10 RCTs involving 1148 women, there was some evidence that any mesh was better than no mesh for preventing objectively determined recurrence of the anterior prolapse (RR 0.48, 95% CI 0.32 to 0.72). When evidence from other study types was also considered, there was a trend in the crude objective failure rates with procedures not using mesh/graft having the highest failure rate (29%, 95% CI 25 to 32% [184/640]), followed by procedures with absorbable synthetic mesh (23%, 19 to 28% [63/273]), absorbable biological graft (18%, 16 to 20% [186/1041]), and non-absorbable synthetic mesh (9%, 7 to 11% [48/548]) (Figure A).

Furthermore, there was corroboratory evidence from the Bayesian meta-analysis of different mesh types and different study designs that objective failure decreased with increasing mesh strength and durability, from absorbable synthetic to absorbable biological to non-absorbable synthetic. Further evidence from indirect comparisons of different types of mesh or graft indicated that non-absorbable synthetic mesh (reference technique) had a statistically significantly lower objective failure rate than both absorbable synthetic mesh (OR 0.23, 95% Crl 0.12 to 0.44) and absorbable biological graft (OR 0.37, 95% Crl 0.23 to 0.59).

xvi This trend appeared to be supported by the finding of an increased need for re-operation for recurrent and de novo prolapse which was highest in women treated with absorbable synthetic mesh: 9%, (16/174), compared with 3% (9/280) for biological grafts and 1% (3/234) for non-absorbable synthetic mesh (Figure A) within an average follow up of 1.5 years. However, counter-intuitively, the re-operation rate for women with no mesh was lower, at 2% (2/85); but this estimate is based on one small study with short follow up (one year) and as such should be interpreted with caution.

Safety

Figure B Safety of anterior repair, summary of crude event rates (95% CI, any study design), by type of mesh/graft

70 NA, NA, NA, 4/11 60

) 50 40

30 1/88, 0/147, 3/198, 4/161 NA, 1/147, 35/581, 68/666 NA, 0/63, 3/42, 3/44 1/93, 0/35, 2/212, 2/248 0/19, 0/112, 0/94, 6/251 NA, 1/35, 4/154, 23/347 4/142, 0/112, 5/477, 11/558

Crude rate (% 20

10 0 Blood loss Damage to Mesh/graft Further De novo urinary De novo bowel De novo Infection Other serious requiring surrounding erosion operation incontinence symptoms (no dyspareunia adverse events transfusion organs needed for (stress urinary data) mesh erosion incontinence)

No mesh/graft Absorbable synthetic mesh Absorbable biological graft Non-absorbable synthetic mesh

For anterior repair, there were too few data on safety outcomes (blood loss; damage to surrounding organs; new urinary or bowel symptoms; new dyspareunia; or infection) to identify or rule out important adverse effects related to the use of mesh/graft either because the studies were not sufficiently large or the adverse effects were rare (Figure B).

There was some evidence to support the trends in mesh strength mentioned above (in objective failure rates and re-operation rates). Mesh/graft erosion increased from 1%, 95% CI 0.1 to 4% (1/147, absorbable synthetic) to 6%, 95% CI 4 to 8% (35/581, absorbable biological) to 10%, 95% CI 8 to 13% (68/666, non-absorbable synthetic). Women with non- absorbable synthetic mesh were also most likely to require an operation to remove it partially or completely because of mesh/graft erosion (7%, 95% CI 5 to 10% [23/347]), versus 3%, 95% CI 0 to 3.3% (1/350) for absorbable synthetic mesh, and 3%, 95% CI 1 to 7% (4/154) for absorbable biological graft.

xvii Summary of evidence: posterior repair

Only 9 studies involving 417 women treated with mesh/graft reported data on the use of mesh in posterior repair (3 RCTs, 2 non-randomised comparative studies, 1 registry report, and 3 case series). Three studies used absorbable synthetic mesh, three used absorbable biological graft, two used combined mesh/graft, and two studies used non-absorbable synthetic mesh. We did not find any RCTs or non-randomised comparative studies that compared different types of mesh/graft for posterior repair. The median follow up was 12 months (range 1 to 17 months).

Efficacy

Figure C Efficacy of posterior repair, summary of crude event rates (95% CI, any study design)

70 NA, NA, 5/14, NA, NA 60 19/58, NA, 14/82, 5/43, NA 50 40 18/142, 6/70, 19/93, NA, 2/31 9/60, NA, 9/78, NA, NA 3/70, NA, 2/29, NA, NA 30 20 Crude rate (%) rate Crude 10 0 Subjective failure Objective failure De novo prolapse Further operation Persistent urinary Persistent bow el Per s is tent (no data) needed for prolapse incontinence (no symptom dyspareunia (recurrent or de data) novo)

No mesh/graft Absorbable synthetic mesh Absorbable biological graft Combined mesh/graft Non-absorbable synthetic mesh

For posterior repair, there were too few data reported for any of the efficacy outcomes (subjective and objective failure rate; de novo prolapse; further operation for recurrent or de novo prolapse; persistent urinary or bowel problems; and sexual (dyspareunia) problems) to draw reliable conclusions or to carry out further statistical analyses (Figure C).

xviii Safety

Figure D Safety of posterior repair, summary of crude event rates (95% CI, any study design)

13/94, 0/5, 7/48, NA, 4/106 50 3/79, 0/5, 1/31, 0/90, 1/71 2/79, 0/5, 1/31, 0/90, 3/71 40 NA, 4/25, NA, 2/36, NA )

30 NA, NA, 0/28, 16/115, 2/31 NA, NA, NA, 11/90, 20 NA, NA, NA, 2/45, 1/29

Crude rate (% 10

0 Blood loss Damage to Mes h/graft Further operation De novo urinary De novo bowel De novo Infection requiring surrounding erosion needed for mesh incontinence (no symptoms dyspareunia transfusion organs erosion data)

No mesh/graft Absorbable synthetic mesh Absorbable biological graft Combined mesh/graft Non-absorbable synthetic mesh

For posterior repair, there were too few data on safety outcomes (blood loss; damage to surrounding organs; mesh erosion; further operations for mesh erosion; new urinary or bowel symptoms; new dyspareunia; infection) to identify or rule out important adverse effects related to the use of mesh/graft either because the studies were not sufficiently large or the adverse effects were rare (Figure D).

Summary of evidence: anterior and/or posterior repair

Fourteen studies involving 1680 women treated with mesh/graft reported data on the use of mesh in anterior and/or posterior repair (3 RCTs, one non-randomised comparative study, one registry report, and 9 case series). One study used absorbable synthetic mesh, none of the studies used absorbable biological graft, one study used a combined mesh/graft, 10 studies used non-absorbable synthetic mesh, and two studies used more than one of the above types of mesh/graft. None of the RCTs or non-randomised comparative studies compared different types of mesh or grafts. The median follow up was 13 months (range 1 to 51 months).

xix Efficacy

Figure E Efficacy of anterior and/or posterior repair, summary of crude event rates (95% CI, any study design), by type of mesh/graft

70 14/34, 14/32, NA, 0/148 60 50 NA, NA, 1/10, NA 40 27/109, 2/26, 11/143, 41/645 NA, NA, NA, 46/203 NA, NA, NA, 1/21 30

20 NA, NA, NA, 7/161

Crude rate (%) rate Crude 10 0 Subjective Objective De novo Further Per s is tent Persistent Per s is tent failure failure prolapse (no operation urinary bow el symptom dyspareunia data) needed for incontinence prolapse (any type) (recurrent or de novo)

No mesh/graft Absorbable synthetic mesh Combined mes h/gr af t Non-absorbable synthetic mesh

For anterior and/or posterior repair, there were too few data reported for any of the efficacy outcomes (subjective and objective failure rate; de novo prolapse; further operation for recurrent or de novo prolapse; persistent urinary or bowel problems; and sexual (dyspareunia) problems) to draw reliable conclusions (Figure E).

For objective failure, there was a trend in the crude rates with procedures not using mesh/graft having the highest failure rate (25%, 95% CI 18 to 34% [27/109], followed by procedures with absorbable synthetic mesh (8%, 2 to 24% [2/26], combined mesh/graft (8%, 4 to 13% [11/143], and non-absorbable synthetic mesh (6%, 5 to 9% [41/645] (Figure E). There were too few data available to carry out Bayesian meta-analysis. Therefore, we could draw no firm conclusions about the effectiveness of any of the mesh/graft types compared with no mesh in anterior and/or posterior repair.

xx Safety

Figure F Safety of anterior and/or posterior repair, summary of crude event rates (95% CI, any study design), by type of mesh/graft

25 NA, 10/78, 3/42 20

15 1/35, NA, 11/810 NA, 9/143, 62/1119 NA, NA, 34/355 NA, NA, 1/47

10 NA, 4/143, 12/541 NA, 6/143, 45/1098 NA, NA, 33/661 NA, NA, 3/278 5 Crude rate (%) 0 Blood loss Damage to Mesh/graft Further De novo De novo De novo Infection Other serious requiring surrounding erosion operation urinary bowel dyspareunia adverse transfusion organs needed for incontinence symptoms events mesh erosion

No mesh/graft Combined mesh/graft Non-absorbable synthetic mesh

For anterior and/or posterior repair, there were too few data on safety outcomes (blood loss; damage to surrounding organs; mesh erosion; further operations for mesh erosion; new urinary or bowel symptoms; new dyspareunia; or infection) to identify or rule out important adverse effects related to the use of mesh/graft either because the studies were not sufficiently large or the adverse effects were rare (Figure F).

Summary of evidence: operation time and surgeon experience

The mean operation times were all less than 170 minutes in the studies which reported this information. The operation time, however, may be affected by concomitant procedures conducted, such as operations for urinary incontinence. Only the registry report provided information on surgeon experience: all surgeons were experienced in pelvic floor surgery and had received pretrial hands-on training in the use of the commercially available device which was used for mesh insertion.

Conclusions

Implications for the NHS

In general, the evidence for most efficacy and safety outcomes was too sparse to provide meaningful conclusions about the use of mesh/graft in anterior and/or posterior vaginal wall prolapse surgery.

In anterior vaginal wall prolapse repair, there was some short-term evidence suggesting that mesh/graft (any type) could reduce objective prolapse recurrence rates compared with no

xxi mesh/graft. In the comparison between different types of mesh/graft, there was a trend for the objective failure rate to decrease as mesh strength and durability increased. Non- absorbable synthetic mesh had statistically significantly lower objective failure rates (9%) than absorbable synthetic mesh (23%) and absorbable biological graft (18%). However, there was no information about efficacy in the long term. There was insufficient evidence to compare objective prolapse recurrence rates in women having posterior or anterior and/or posterior repairs.

While there might be some evidence of differences in objective efficacy related to the use of mesh, these must be considered alongside any safety concerns. There was some evidence to suggest that mesh/graft may cause problems with erosion and a subsequent need for operations to remove the foreign material. For anterior repair, mesh/graft erosion increased from 0.7% using absorbable synthetic mesh to 6.0% using absorbable biological grafts and to 10.2% using non-absorbable synthetic mesh. The surgical removal rate due to erosion (complete or partial) was 2.9% for absorbable synthetic mesh, 2.6% for absorbable biological graft, and 6.6% for non-absorbable synthetic mesh. The results for mesh erosion were similar in women having posterior repair and anterior and/or posterior repair. However, the numbers were too few to conduct statistical analyses to compare the erosion rates between different types of mesh or graft.

For all prolapse types, there was not enough information to compare any of the other efficacy outcomes (subjective failure, de novo prolapse, further operation needed for prolapse, persistent urinary and bowel symptoms, and sexual problems) nor any of the other safety outcomes (blood loss, damage to surrounding organs, new urinary and bowel symptoms, and new sexual problems) between mesh/graft (any type) and no mesh/graft or between different types of mesh/graft.

Implications for further audit or research

The RCTs currently available only as conference abstracts may become available as full text papers in the future. The expected completion time for the eight on-going studies (seven

xxii RCTs and one registry) ranges from 2008 to 2011. A further review incorporating these studies may be informative.

This review highlights the fact that there is very little evidence to assess whether the use of mesh/graft for prolapse repair improves the health and quality of life of women undergoing surgery.

One year outcomes are too soon to assess whether prolapse surgery is successful in the long-term. The mean time to first re-operation is reported in the literature as 12 years, and therefore failure at one year represents a serious underestimation of efficacy. Prospective studies will require extended follow up to assess meaningful mesh/graft failure.

It is increasingly recognised that in prolapse surgery, subjective failure is a more appropriate outcome measure of efficacy than objective failure. The need for a repeat operation should be determined, not by the presence of prolapse, but by whether that prolapse is affecting a woman’s quality of life. However, criteria for measuring such subjective prolapse outcomes are difficult to quantify and the most appropriate methods are still being evaluated.

Therefore, rigorous RCTs are required to determine the comparative efficacy of using mesh/graft and its optimal place in clinical practice. The RCTs should: • Primarily compare the subjective failure rate in procedures using mesh/graft versus those without mesh/graft, and between different types of mesh/graft; • Use validated patient reported outcome measures, if available; • Have sufficient power to detect clinically meaningful differences in both efficacy and safety; and • Have the capacity to assess outcomes in the long term (at least 5 years), including cost- effectiveness.

In addition, prospective data collection should be considered in which the operative and clinical details of women undergoing prolapse surgery with mesh can be recorded so that sufficient efficacy and safety data can be gathered to guide the use of mesh or grafts in the future. In particular, long term outcomes need to be considered as it may be many years before a prolapse recurrence becomes symptomatic.

xxiii LIST OF ABBREVIATIONS

BMI Body Mass index CENTRAL Cochrane Controlled Trials Register CI Confidence interval CrI Credible Interval GATE Generic Appraisal Tool for Epidemiology IUGA International Urogynaecological Association NHS National Health Services HTA Health Technology Assessment OR Odds Ratio PFMT Pelvic Floor Muscle Training POP Pelvic Organ Prolapse POP-Q Pelvic Organ Prolapse Quantification System ReBIP Review Body of Interventional Procedures RCTs Randomised Controlled Trials RR Relative risk SCI Science Citation Index STARR Stapled Transanal Rectal Resection

xxiv 1 OBJECTIVE OF THE REVIEW

The aim of this study was to systematically review the evidence for the efficacy and safety of mesh or grafts in surgery for anterior and/or posterior vaginal wall prolapse.

The specific objectives were to:

(a) Compare efficacy and safety between procedures using mesh/graft (any type) and no mesh/graft; (b) Compare efficacy and safety between different types of mesh/graft.

1 2 BACKGROUND

2.1 Description of the underlying health problem

2.1.1 Epidemiology

A woman can present with prolapse of one or more of these sites. The present review will focus on anterior and posterior vaginal wall prolapse. We will focus on upper vaginal prolapse (cervix, uterus, or vault) in a second review.

POP is common and is seen in 50% of parous women.2 Data from the Oxford Family Planning Association indicate an annual incidence of hospital admission with a diagnosis of prolapse as 20.4 per 10,000. There is a lifetime risk of 7% of having surgery for it.3 In England and Wales in 2005-2006,4 over 28,000 prolapse operations were performed. Around 90% of these are for anterior and/or posterior POP. Since 2001, there has been an increase of about 1000 prolapse operations in England and Wales each year. Government figures for population predictions show that there will be a 1.2 million increase in the number of women between 2004 (27.0 million) and 2014 (28.2 million). This increase is almost entirely due to more women in the 50 to 85 age group, the population most likely to need prolapse surgery.5 This suggests that the increase in prolapse surgery is likely to continue.

2.1.2 Burden of disease

POP affects a woman’s quality of life by its local physical effects (pressure, bulging, heaviness or discomfort) or its effect on urinary, bowel or sexual function. Urinary symptoms include both incontinence and retention (incomplete emptying), bowel symptoms may include constipation or faecal incontinence and sexual symptoms such as difficulty or inability to have intercourse due to pain (dyspareunia) or embarrassment.

2 2.1.3 Aetiology, pathology and prognosis

Vaginal delivery is the greatest risk factor for POP. Neuromuscular injury6,7 and mechanical injury (for example damage to the fascial connective tissue or pelvic floor muscles,8 occurring during vaginal delivery has been identified as the main reason for POP. Nulliparous women with connective tissue disorders (e.g., Ehlers-Danlos or Marfan’s syndrome) or women with spina bifida and poorly functioning pelvic floor muscles are susceptible to POP and can present with advanced POP at a young age.

A number of objective clinical classification or grading systems are in use. The simplest include the ordinal grading of prolapse as first, second and third degree such as the Baden- Walker9 and Beecham10 classification system. In an attempt to improve the accuracy of prolapse assessment, the International Continence Society has devised the POP quantification (POP-Q) System with proven interobserver and intraobserver reliability.11,12 It allows for a standardized, accurate objective description of physical findings, and provides a means of quantification of change in individual patients, and comparison between series of patients. The Baden-Walker and POP-Q classifications are summarised in Table 1.

Table 1 Classification of vaginal wall prolapse stages Baden-Walker classification system POP-Q system - - Stage 0 No descent of pelvic structures during straining

First The anterior vaginal wall, from the urethral meatus to Stage I The leading surface of the prolapse is >1 degree the anterior fornix, or the saccular protrusion of the cm above the hymenal ring. rectovaginal wall, descends halfway to the hymen.

Second The anterior vaginal wall and underlying bladder, or the Stage II The leading edge of the prolapse extends degree sacculation, extend to the hymen. from 1 cm above the hymen to 1 cm below the hymenal ring.

Third The anterior vaginal wall and underlying urethra and Stage III The prolapse extends to between 2 and 3 degree bladder, or the sacculation protrudes, are outside the cm below the hymenal ring, but there is hymen. not complete vaginal eversion.

- - Stage IV The vagina is completely everted.

2.2 Current management and alternative procedures

Current treatment options for anterior and posterior vaginal wall prolapse include pelvic floor muscle training (PFMT), use of pessaries (mechanical devices such as rings or shelves), and surgery.13

3 Surgery is considered usually because the prolapse is more severe and is causing significant symptoms. A recent development has been the use of mesh or grafts in the surgical repair of POP. Despite the high service use, the evidence base for rational management is scanty.14

2.3 The interventional procedure under review

2.3.1 Description of the interventional procedure

Surgical repair includes anterior or posterior colporrhaphy and site-specific defect repair.15 It can be augmented with mesh or graft materials, to reduce the risk of failure, analogous to the use of similar materials in hernia repair.16 The procedures, with or without mesh/graft, is usually conducted as an inpatient procedure under a general or spinal anaesthetic.

Some women may have other gynaecological problems which require surgical correction at the same time. These concomitant procedures include cervical amputation, hysterectomy, or surgery for urinary incontinence. These are usually conducted under the same anaesthetic.

Colporrhaphy

Anterior colporrhaphy is for the repair of anterior vaginal wall prolapse. It involves dissection of the vaginal mucosa through a midline incision of the anterior vagina in order to expose the underlying bladder and pubocervical fascia. The fascia is then repaired by plication, the redundant vaginal tissue is excised, and closed in the midline.

Posterior colporrhaphy, for posterior vaginal wall prolapse repair, involves a vaginal incision and plication of the levator ani. Posterior vaginal wall prolapse can also be repaired transanally. This involves plication of the rectal serosa to buttress the posterior vaginal wall. Stapled Transanal Rectal Resection (STARR) is a newer technique for posterior vaginal wall prolapse repair.

Site-specific repair

With improved understanding of anatomical supports of the vagina, some surgeons choose site-specific fascial defect repair rather than colporrhaphy as colporrhaphy addresses only the issue of a central fascial defect and was considered to contribute to dyspareunia. In site- specific repair, fascia plication is avoided to reduce the risk of dyspareunia. The defects or

4 tears of the fascia are identified and sutured. For instance, a paravaginal repair, where the vagina is sutured to the arcus tendineus fascia bilaterally, is used if lateral fascial defects are identified.

Mesh or graft repair

Mesh or graft repair differs from traditional repair in that a piece of mesh or graft is used to support and strengthen the vaginal fascia while it heals. More than one piece of mesh/graft may be used in each procedure. The technique of implanting mesh or graft varies widely between gynaecologists. It can be positioned and sutured over the fascial defect as an ‘inlay’16, or the whole vagina can be surrounded by mesh/graft (‘total mesh’). ‘Total mesh’ techniques involve insertion of large pieces of mesh, often using a blind introducer or trocar, which aim to suspend rather than reinforce the prolapsing organs.

There are numerous types of mesh and grafts available in terms of absorbability by the host tissue, type of material, structure, and physical properties such as tensile strength and elasticity. There are no existing classification systems for mesh and grafts. We have used the term ‘mesh’ for synthetic material and ‘graft’ for biological material and have devised our own classification system based on type of material and reported properties. Some synthetic material (e.g. based on polyglactin) and all biological materials are absorbable, i.e. they are eventually absorbed by the host tissue. Therefore, in theory, absorbable mesh or graft materials are less likely to cause problems with erosion, but may have higher failure rates because they are eventually absorbed. In contrast, non-absorbable mesh, which is always synthetic, is thought to be more likely to cause erosion but with lower failure rates. Combined absorbable/non-absorbable mesh/graft (termed ‘combined’ hereafter) is a composite of both absorbable and non-absorbable material.

Monofilament or multifilament, and porosity in mesh structure

Non-absorbable synthetic mesh can be made from knitted single-fiber filaments (monofilament materials) or they can be braided with monofilament yarns, further woven as multifilament fibers in different ways and with different pore sizes. Knitted fabrics have a more open structure than woven. Non-absorbable synthetic mesh are usually further classified according to their pore size. It can be divided into four types:16-18 totally macroporous with a pore size of >75μm (type I); totally microporous (type II); mixture of macro and microporous (type III), and submicronic (type IV).

5 Table 2 summarises the most common types according to their absorbability, material, and structure.

Table 2 The different mesh and grafts in vaginal wall prolapse repair Type of mesh and grafts Material Examples (commercial or trade name) Absorbable synthetic mesh Polyglactin Vicryl (multifilament)

Absorbable biological graft Autograft Fascia lata NA (patient’s own tissue used) Rectus sheath NA (patient’s own tissue used)

Allograft Fascia lata Tutoplast Dura mata Duraderm Dermis AlloDerm

Xenografts Porcine dermis Pelvicol Porcine small-intestine submucosa SIS

Combined mesh/graft Polyglactin and polypropylene Vypro II (mono-multifilament) Porcine collagen and Avaulta (monofilament) polypropylene Non-absorbable synthetic mesh Type I (totally macroporous) Polypropylene Marlex (monofilament) Prolene (monofilament) Type II (totally microporous) Polytetrafluoroethylene Goretex (multifilament) Type III (macro-microporous) Polyethylene tetraphalate Mersilene (multifilament) Type IV (submicronic pore size) Silicone or silicone-coated mesh Polyatex

Commercially available devices for insertion of ‘total mesh’ systems

There are also commercially available devices/kits to faciliate mesh/graft insertion via in-to- out or out-to-in introducers, such as Perigee, Prolift, and Avaulta (commercial names). After vaginal incision and dissection of vaginal wall fascia, incisions for mesh/graft introducer points are made in the skin. The surgeon palpates and controls the introducer movement within the pelvic cavity using landmarks (anatomical) as it perforates. Mesh/graft is connected to the perforating end of the introducer. The mesh is stabilised by the mesh tapes that are pulled through the introducer points.

6 2.3.2 Suitable candidates and relevant subgroups

Mesh or graft repair is theoretically suitable for any degree of symptomatic anterior and/or posterior vaginal wall prolapse. It may be more suitable for women with recurrent prolapse, who have a higher risk of failure compared with women undergoing their first repair, or for women with congenital connective tissue disorders (such as Ehlers-Danlos or Marfan’s). Older women may also have a higher risk of failure. However, there is as yet no evidence to determine whether these or other confounding factors such as menopausal status, body mass index (BMI), or concomitant operations affect prognosis, and hence choice of mesh use or not.

2.3.3 Proposed clinical indications/contraindications and putative impact of the procedure

In a roundtable forum of the International Urogynaecological Association (IUGA) in 2005 on the clinical implications of mesh and grafts,19 it was recognised that the current expansion of the use of mesh and graft materials in POP surgery is not based on reliable evidence. There was a consensus that the indications and contraindications for the use of mesh and graft were unclear. However, based on clinical experience, the committee produced a provisional list of suggested contraindications for the use of mesh and graft (Table 3).

Table 3 Relative contraindications to the use of mesh or graft Contraindications • History of previous pelvic radiation • Severe urogenital atrophy • Immunosuppressed patient • Presence of active pelvic or vaginal infection • Patient currently on systemic steroids • Host factors including: a. Poorly controlled diabetes b. Morbid obesity c. Heavy smokers

7 2.3.4 Personnel involved and skill/experience required

The majority of prolapse surgery is carried out by gynaecologists in the UK. Gynaecologists with a special interest in urogynaecology may also undertake this procedure.

Knowledge regarding indications for mesh and graft usage, specific biologic and chemical properties of the available types of mesh or graft, short- and long- term risks and benefits of graft usage, as well as the optimal implantation technique are required to optimise mesh or graft use.19 Surgeons need to be aware of and have the expertise to manage any complications associated with the use of mesh or graft.

2.3.5 Current use in the UK

Approximately 28,000 anterior and posterior prolapse repair procedures were performed in the NHS in England and Wales in 2005/2006, but no data specific to mesh or graft use were available.4 Jha & Moran20 conducted a survey on the use of mesh and grafts amongst 398 gynaecologists and urogynaecologists in the UK. They found that some gynaecologists were already using mesh or grafts for colporraphy and did so more often in women with recurrent prolapse (Table 4). Usage of mesh/graft is also affected by the financial situation as not all gynaecologists have access to mesh/graft in the UK.

Table 4 Surgeons’ choice of procedure and use of mesh or graft in anterior and posterior vaginal wall prolapse repair Colporrhaphy Fascial plication + mesh or graft Paravaginal repair Others Anterior vaginal wall prolapse Primary 77% 10% (24% synthetic mesh, 76% biological graft) 6% 7% Recurrent 45% 34% (28% synthetic mesh, 72% biological graft) 15% 6%

Posterior vaginal wall prolapse Primary 75% 9% (40% synthetic mesh, 60% biological graft) 11% 5% Recurrent 38% 49% (44% synthetic mesh, 56% biological graft) 6% 7%

Data source: National survey on the management of prolapse in the UK20

Apart from the consensus statement, IUGA 2005, concerning contraindications for treatment and the experience required by the practitioner to undertake the procedure,19 to our best knowledge, there are no established guidelines in the UK or elsewhere on the use of mesh or graft for pelvic organ prolapse.

8 2.3.6 Equipment or devices required

Compared with traditional repair (colporrhaphy or site-specific defect repair), the only extra requirement is a piece of mesh or graft large enough to cover the defect, or a commercially available device. Most of the commercially available devices are disposable. If the gynaecologist chooses to use solvent-dehydrated cadaveric fascia lata, saline solution must be used to soak the graft before use.

9 3 METHODS FOR REVIEWING SAFETY AND EFFICACY

3.1 Search strategy

We searched electronic databases, conference proceedings and relevant websites, contacted experts in the field and scrutinised bibliographies of retrieved papers. Extensive highly sensitive electronic searches were conducted to identify reports (both full text papers and conference abstracts) of published and ongoing studies on the safety and efficacy of mesh/graft used in the repair of pelvic organ prolapse. Searches were restricted to publications from 1980 onwards and to those published in the English language. The search strategies were designed to retrieve all relevant publications for the repair of pelvic organ prolapse: both anterior and posterior for the present review, and cervical, uterine and vaginal vault prolapse repair for a following review. This was felt to be the most efficient approach.

The databases searched were: Medline (1980-June Wk 3 2007), Medline In-Process (3rd July 2007), Embase (1980 – 2007 Wk 26), Biosis (1985- 5th July 2007), Science Citation Index (1980 – 2nd Ist July 2007), Cochrane Controlled Trials Register (The Cochrane Library, Issue 2 2007) ISI Conference Proceedings (1990 – 27th June 2007) as well as current research registers (National Research Register (Issue 2, 2007), Current Controlled Trials (April 2007) and Clinical Trials (April 2007)). Additional databases searched for systematic reviews and other background information included the Cochrane Database of Systematic Reviews (The Cochrane Library, Issue 2, 2007), Database of Abstracts of Reviews of Effectiveness (June 2007) and the HTA Database (June 2007). Conference proceedings of major urogynaecological organisations (American Urogynecologic Society, American Urological Association, European Association of Urology, European Society of Gynecological Endoscopy, Incontinence Society and International Urogynecological Association) for 2005 onwards were scrutinised for additional reports of randomised controlled trials (RCTs). Reference lists of all included studies were scanned to identify additional potentially relevant studies. Full details of the search strategies used and websites consulted are documented in Appendix 1.

Eleven manufacturers were identified and contacted for properties of mesh/graft produced and for any studies related to mesh/graft. The eleven manufacturers were: American Medical Systems UK Ltd., Atrium Mecical Corp., Bard Ltd., Boston Scientific Ltd., Johnson & Johnson Medical Ltd., WL Gore & Associations (UK) Ltd., Coloplast Ltd., Mpathy Medical Devices Ltd., Tyrx Pharma Inc., Cook Biotech Inc., and Organogenesis Inc.

10 3.2 Inclusion and exclusion criteria

3.2.1 Types of studies

We considered the following types of studies: • Randomised controlled trials (RCTs) (full text and conference abstracts); • Non-randomised comparative studies (full text only); • Case series with a minimum sample size of 50 women (full text only). Studies with a mean follow-up time of one year or over were included for both efficacy and safety. One year was considered a minimum adequate period of time in which to assess the efficacy of prolapse repair. Studies with a mean follow up of less than one year were included for safety outcomes only. • Population-based registry reports that prospectively collected audit data on the use of mesh or graft for anterior and posterior vaginal wall prolapse.

We did not search for studies that only reported procedures without mesh/graft, hence data related to such procedures are derived from control groups in RCTs and non-randomised comparative studies only.

We excluded case reports as we considered that there were sufficient data to capture rare adverse events. For conference abstracts, only those published from 2005 onwards were screened and only RCTs were considered. Other conference abstracts were excluded in order to ensure that the number of included studies remained manageable within the timeframe for the review. We excluded earlier reports of studies reproduced in later publications, where the results related to the same outcomes. The review was restricted to English language studies due to the complexity of the terminology in this topic area. Animal studies were not considered.

Studies reporting anterior and/or posterior prolapse repair, plus cervix, uterus, or vaginal vault prolapse repair, but not reporting results separately, were excluded.

3.2.2 Types of participants

The participants considered were women undergoing anterior and/or posterior vaginal wall prolapse surgery. Studies of women with prolapse caused by pelvic trauma, congenital disease, or prolapse after creation of neovagina were excluded. Women could be

11 undergoing other concomitant operations, such as hysterectomy or a continence procedure as long as the main indication for surgery was anterior or posterior prolapse.

3.2.3 Types of interventions

The interventions considered were anterior and/or posterior vaginal wall prolapse repair with mesh or graft. There were no restrictions on type of mesh or graft or technique of using mesh or graft. For RCTs and non-randomised comparative studies, there were no restrictions on the comparator. The comparator could be another operation using mesh or graft, or a type of surgery which did not involve mesh.

The types of mesh or grafts were classified as follows: • Absorbable synthetic mesh (eg. polyglactin) • Absorbable biological graft (eg. porcine dermis, fascia lata) • Combined mesh/graft (eg. polypropylene and polyglactin composite, polypropylene mesh coated with absorbable porcine collagen) • Non-absorbable synthetic mesh (eg polypropylene)

3.2.4 Types of outcomes

Efficacy

Primary outcomes for efficacy included: • Persistent prolapse symptoms (subjective) • Recurrent prolapse at original site (objective)

Secondary outcomes for efficacy include: • De novo prolapse, i.e. new prolapse at other sites that was free of prolapse at baseline • Need for further surgery for prolapse (both recurrent and de novo) • Persistent urinary symptoms • Persistent bowel symptoms • Persistent dyspareunia • Disease specific quality of life (RCTs and non-randomised comparative studies only).

For objective failure, outcomes measured by different systems, such as POP-Q system and Baden-Walker system, were combined. For de novo prolapse (i.e. at other sites), only women free of prolapse at other sites at baseline were considered. Persistent urinary

12 symptoms were included as efficacy outcome although surgeries aiming at curing urinary symptoms may be conducted at the same time. For persistent urinary symptoms, bowel symptoms, and dyspareunia, only women having these symptoms at baseline were considered. For disease specific quality of life, only RCTs and non-randomised comparative studies were considered because the baseline quality of life in registry or case series could not be adjusted for.

Safety

Safety outcomes included: • Blood loss • Damage to surrounding organs during operation • Mesh/graft erosion • Further operation needed for mesh/graft erosion • De novo urinary incontinence • De novo bowel symptoms • De novo dyspareunia • Infection • Other potentially serious adverse effects

We were aware that the blood loss may be not only due to the repair of vaginal wall prolapse, but also the concomitant procedures such as procedures for urinary incontinence or hysterectomy. Another factor that may affect blood loss is the use of commercially available devices, i.e. the devices with introducers, where more dissection is needed; hence more blood loss (personal communication with Dr. Chris Bain, 28th September 2007).

For de novo urinary incontinence, bowel symptoms, and dyspareunia, only women free of these symptoms at baseline were considered.

Other outcomes

Other outcomes considered were operation time and surgeon experience. We were aware that the difference in operation time across studies is also affected by the conduct of concomitant procedures, such as surgery for incontinence and hysterectomy.

13 3.3 Quality assessment strategy

One reviewer assessed the quality of the studies, using one of two separate checklists according to study design. Both checklists were developed through the Review Body for Interventional Procedures (ReBIP): • A 13-question checklist was used to assess the quality of RCTs (see Appendix 2). This was adapted from Verhagen and colleagues.21 We removed the question on blinding of surgeon because it is not applicable. • An 18-question checklist was used to assess the quality of non-randomised comparative studies with the same checklist minus four questions used to assess the quality of case series (see Appendix 3). The checklist was adapted from several sources, including the NHS Centre for Reviews and Dissemination’s guidance for those carrying out or commissioning reviews,22 Verhagen and colleagues,21 Downs and Black23 and the Generic Appraisal Tool for Epidemiology (GATE).

Each question was worded so that a rating of ‘Yes’ was always optimal in terms of methodological quality.

3.4 Data extraction strategy

One reviewer screened the titles (and abstracts where available) of all papers identified by the search strategy. Full text copies of all reports deemed to be potentially relevant were obtained and assessed by the same reviewer for inclusion. Any areas of uncertainty (around 50 queries) were resolved by consultation with a second reviewer or clinical advisor.

We developed and piloted a data extraction form. One reviewer extracted details of study design, methods, participants, interventions and outcomes.

3.5 Data analysis

Data analysis was conducted for three subgroups of women according to the type of prolapse being repaired: • Anterior vaginal wall prolapse repair • Posterior vaginal wall prolapse repair, and • Anterior and/or posterior vaginal wall repair in which o All women received both anterior and posterior prolapse repair, or

14 o Some of the women received anterior repair and others received posterior repair, but their data were not reported separately, or o Some of the women received both anterior and posterior repair and others received either anterior or posterior repair, but their data were not reported separately.

Direct comparisons (RCTs only)

A meta-analysis of RCTs, using Cochrane Collaboration Review Manager (RevMan 4.2) software, was conducted to directly compare the efficacy and safety of procedures using mesh/graft versus procedures without mesh/graft. Where possible, we compared the efficacy and safety of different types of mesh or graft where these were compared directly.

For categorical outcomes we related the numbers reporting an outcome to the numbers at risk in each group to derive a relative risk (RR). For continuous variables we used means and standard deviations to derive a weighted mean difference. As a general rule, a fixed effects model was used for calculations of summary estimates and their 95% confidence intervals (CI).

Trials were only combined if the interventions and participant groups were similar enough on clinical criteria. When heterogeneity was suspected from the X2 test for heterogeneity (at the 10% level) or the I2 statistic (over 40%24), we used a random effects model to derive summary estimates and their associated 95% CI.

Crude event rates (all study types)

Crude event rates (and 95% confidence intervals calculated by using binominal distribution approximation) for each of the intervention categories were tabulated by summing across studies for all outcomes, and also according to study design (RCT, non-randomised comparative studies, case series/registry) to facilitate qualitative assessment of potential heterogeneity of event rates across different study designs.

Bayesian meta-analysis models and indirect comparisons

Additionally, Bayesian meta-analysis models were used to model the objective failure rates for the different interventions. A ‘cross-design’ approach was adopted to allow non- randomised evidence to be included while avoiding the strong assumption of the equivalence

15 of studies implicit in the crude event rates. This approach to evidence synthesis enabled data 25 from RCTs and non-randomised comparative studies studies to be included. The specific type of model used was a (Bayesian) binomial random effects model. Differences between interventions were assessed by the corresponding odds ratio and 95% credible interval (Crl). Crls are the Bayesian equivalent of confidence intervals. A 95% Crl for an odds ratio has a 95% probability of containing the population odds ratio. The results from these meta-analysis models are referred to as ‘unadjusted’ odds ratios.

Meta-analysis models that adjusted for study type were also used. The models were extended by including one study type effect to allow for the differences between non- randomised comparative studies and RCTs. The results from these models are referred to as ‘adjusted’ odds ratios.

Included, where data permitted, are ‘head to head’ indirect comparisons of the different mesh types reported as odds ratios and 95% CrIs adjusted for study design.

26 WinBUGS software was used to produce Bayesian meta-analysis models.

Subgroup analysis

Pre-specified subgroup analysis was planned for • Different mesh types within non-absorbable mesh according to the porosity of mesh structure, i.e. type I to IV (see Table 2); • ‘Mesh inlay’ versus ‘total mesh’ inserted using commercially available introducers; • Primary versus secondary repairs

However, these could not be conducted due to the lack of data.

16 4 RESULTS

4.1 Number and type of included studies

The results of the searches are summarised in Table 5. The numbers retrieved from the searches in Science Citation Index, Biosis and CENTRAL include only the additional reports found after excluding those identified from the Medline/Embase multifile search; the results of the searches of ISI conference proceedings and of conference abstracts include additional reports identified after deduplication against SCI and Biosis. A total of 1633 publications were identified. In addition the details of 18 potentially relevant ongoing studies were noted. The titles and abstracts (where available) were screened and 446 reports (317 full text papers; 129 conference abstracts) were selected for full assessment.

Table 5 Search results Database Number retrieved Number selected for assessment Primary reports: Medline/Embase/Medline In Process multi file 862 261 search(after deduplication in Ovid) SCI 223 42 Biosis 276 53 CENTRAL 5 5 ISI Conference proceedings 27 10 Conference abstracts 226 71 Total 1619 442 Background: CDSR 4 1 DARE 0 0 HTA database 10 3 Total 14 4 Total assessed for review 1633 446 Ongoing studies: NRR 3 0 CCT 3 1 Clinical Trial 12 7 Total 18 8

17 Table 6 details the outcome of assessment of the selected reports with 67 reports (42 full text papers; 25 abstracts) included in the review. Appendix 4 lists the bibliographic details for all included studies. Seven case reports and 25 case series with a sample size of less than 50 were identified but excluded according to our inclusion criteria. Six of the 11 contacted manufacturers responded (American Medical Systems (UK) Ltd., Bard Ltd., Johnson & Johnson Medical Ltd., WL Gore & Associations (UK) Ltd., Coloplast Ltd., Mpathy Medical Devices Ltd.) and provided data on mesh/graft properties and related studies, all of which had already been identified.

Table 6 Papers selected for full assessment Full text Conference papers abstracts Included in review 42 25 Retained for background information 22 0 Retained for companion review on vaginal vault and uterine prolapse 144 54 repair Excluded 100 59 Total 308 138

4.1.1 Included studies

We identified 49 studies (6 full-text RCTs, 11 RCTs available as conference abstracts, 7 non-randomised comparative studies, one prospective registry of cases, and 24 case series) (Table 7). One RCT27 and one case series28 reported anterior and posterior repair separately. The registry report29 reported anterior repair, posterior repair, and anterior and posterior repair separately. The registration was prospective and was during a six month period from 25 centres in Denmark, Norway, Finland and Sweden. The frequency and type of complications occurring during all surgical procedures using a commercial available device (Prolift, non-absorbable synthetic ‘total mesh’) and its associated hospital stay were recorded.

18 Table 7 Number of studies and reports included Study design Studies Anterior repair Posterior Anterior and/or (reports) repair posterior repair Efficacy and Safety RCTs (full text) 6 (8) 527,30-35 227,36 0

RCTs (conference abstracts) 11 (23) 737-54 155 356-59

Non-randomised comparative 7 (7) 460-63 264,65 166 studies Case series with sample size ≥50 16 (20) 10 28,67-79 228,80 581-85 and mean follow-up ≥1y Subtotal 40(58) 26 7 9 Safety only Registry, mean follow-up <1y 1 (1) 129 129 129

Case series with sample size ≥50 8 (8) 386-88 189 490-93 but mean follow-up <1y Total 49 (67) 30 9 14

Seven studies were reported in more than one report. There were two reports of the study by De Ridder and colleagues30,31 with the primary reference being De Ridder 2002;30 two reports of the study by Gandhi and colleagues32,33 with the primary reference being Gandhi 2005;32 four reports of the study by De Tayrac and colleagues68-71 with the primary reference being De Tayrac 2006;71 two reports of the study by Kobashi and colleagues75,76 with the primary reference being Kobashi 2002;76 five reports of the study by Cervigni39-41,51,52 with the primary reference being Cervigni 2007;41 eight reports of the study by Kocjancic and colleagues44-50,54 with the primary reference being Kocjancic 2007;47 and two reports of the study by Lim and colleagues57,58 with the primary one being Lim 2007.58

Appendix 5 provides details of the characteristics of studies by study design.

The studies took place during the period 1996 – 2007. Eighteen studies took place in the USA, seven studies in France, six each in Australia and Italy, two in the UK, one each in Belgium, Canada, Denmark, Egypt, Finland, South Africa, Spain and Sweden, one was a multi-centre study (the registry), and one did not report this information.

Table 8 summarises the characteristics of patients treated with mesh or graft and the treatment details for the 49 included studies. The median follow up was 13 months (range 1 to 51). The mean follow up in 31 studies was one year or over. 4569 women were treated

19 with mesh or graft in total. In studies providing details, the mean age was 64 years (range 24 to 96 years). 72% repairs were primary procedures. Mesh or graft was most often used for anterior repair (54%, 2472/4569). Overall, just over half of the studies used non-absorbable synthetic mesh (51%, 2320/4569) but for anterior repair alone and for posterior repair alone, absorbable biological graft was the most common alternative (46% (1124/2472) and 29% (121/417) respectively). 56% (1404/2497) of women had a concomitant procedure for urinary incontinence and 37% (953/2583) had a hysterectomy.

Appendix 6 provides a full list of all the mesh/graft types used in the included studies, their generic and trade (commercial) names, manufacturers and reported material properties. Appendix 7 provides details of surgical techniques used in each study. The surgical techniques for implanting mesh/graft varied considerably across studies.

Table 8 Characteristics of women treated with mesh or graft and treatment details Anterior Posterior Anterior and/or posterior Total N studies N women N studies N women N studies N women N studies N women Total number of women 301 2472 91 417 141 1680 491 4569

Age (years), medians, range 23 65 (24 – 96) 7 60 (31 – 85) 12 63 (28 – 91) 39 64 (24 – 96)

Primary/secondary repair Primary repair 10 762 (71%) 1 48 (68%) 4 166 (77%) 12 976 (72%) Secondary repair 10 309 (29%) 1 23 (32%) 4 51 (23%) 12 383 (28%)

Type of mesh or graft Absorbable synthetic mesh 4 289 (12%) 3 79 (19%) 1 32 (2%) 8 400 (9%) Absorbable biological graft 14 1124 (46%) 3 121 (29%) 0 0 17 1245 (27%) Combined mesh/graft 1 50 (2%) 2 115 (28%) 1 143 (9%) 4 308 (7%) 21 Non-absorbable synthetic mesh 14 1009 (40%) 2 102 (24%) 11 1209 (72%) 27 2320 (51%) ≥1 type mesh or graft listed above 0 0 0 0 2 296 (18%) 2 296 (6%)

Concomitant procedure Procedure for Incontinence 12 786/1111 (71%) 3 85/126 (68%) 8 533/1260 (42%) 23 1404/2497 (56%) Hysterectomy 11 579/1194 (49%) 3 16/53 (30%) 9 358/1336 (27%) 23 953/2583 (37%)

Follow-up In total, median (range) of mean follow up 30 14m (1 – 38) 9 12m (1 – 17) 14 10m (1 – 51) 49 13m (1 – 51) ≥1y, median (range) of mean follow up 23 22m (12 – 38) 6 14m (12 – 17) 6 22m (12 – 51) 31 17m (12 – 51) <1y, median (range) of mean follow up 7 6m (1 – 8) 3 6m (1 – 6) 8 6m (1 – 11) 18 6m (1 – 11)

Notes: 1Total number of studies was less than the sum of numbers of studies in anterior repair, posterior repair, and anterior and/or posterior repair, because three studies27-29 reported different types of prolapse repair separately.

4.1.2 Ongoing studies

Eight ongoing studies were identified (Table 9), seven of which were RCTs (Bombieri; Brandao; Nieminen; Sung; Tayrac; Verleyen; Withagen), and one was a registry (Zylstra) of efficacy and safety results of using products of American Medical Systems (USA, see Appendix 7 for mesh details). Three studies were of anterior vaginal wall prolapse repair, one study of posterior prolapse repair, two of anterior and/or posterior prolapse repair, and one of prolapse repair in any compartment. For studies reporting sample size, all sample sizes were over 70. The registry study aims to recruit 1000 women.

Three studies used absorbable biological graft, one study used absorbable synthetic mesh, one study used combined mesh/graft, and three studies used non-absorbable synthetic mesh. Six of the seven RCTs compared mesh or graft with traditional repair, and one compared absorbable biological graft with absorbable synthetic mesh.

For studies reporting a follow-up period, all follow-up periods were one year or more. These studies are due to completed in 2008 to 2011.

Table 9 Characteristics of ongoing studies ID Compartment mesh repaired N Mesh/graft Follow-up Time to be completed RCT Bombieri, UK Posterior 120 A: absorbable biological graft 5 years 1 Apr. 2009 B: traditional repair

Brandao, Brazil (personal communication, Anterior and/or posterior NR A: non-absorbable synthetic mesh NR 2008 Adrian Griffin, 2 Aug. 2007) B: traditional repair

Nieminen, Finland Anterior 202 A: non-absorbable synthetic mesh NR May 2008 B: traditional repair

Sung, US Posterior 160 A: absorbable biological graft NR NR (started in Jan. 2004) B: traditional repair

23 Tayrac, France Anterior 194 A: combined mesh/graft NR Jun. 2011 B: traditional repair

Verleyen, Belgium Anterior 79 A: absorbable biological graft NR NR B: absorbable synthetic mesh

Withagen, the Netherland Anterior and/or posterior 194 A: non-absorbable synthetic mesh 1y NR (started in Aug. 2006) B: traditional repair

Registry Zylstra, US Any compartment 1000 Non-absorbable synthetic mesh 2y Sep. 2008

Notes:

NR: not recorded

4.2 Number and type of excluded studies; reason for exclusion

A list of potentially relevant studies identified by the search strategy, for which full text papers were obtained, but which subsequently failed to meet the inclusion criteria given in Section 3.2, and were therefore excluded, is given in Appendix 8.

4.3 Quality of available evidence

Only the methodological quality of the full text studies was assessed. Appendices 9 to 12 provide details of the quality assessment of the included RCTs, non-randomised comparative studies, and registry and case series respectively. The results are summarised in Figures 1, 2 and 3 respectively.

For RCTs, approaches to sequence generation for randomisation were reported in all studies except the study by De Ridder and colleagues.30 Treatment allocation was adequately concealed in all studies except the studies by De Ridder and colleagues30 and by Sand and colleagues (both unclear).27 All follow-up periods were one year or more. All studies used intention-to-treat analysis in that women remained in the groups to which they were allocated, but some reported loss to follow up: in these cases results were abstracted as reported. No studies reported information on surgeon experience.

Figure 1 Summary of quality assessment of the RCTs (full-text, n=6)

Q13. Experienced practitioner

Q12. ITT or no w ithdraw s/drop-outs after recruitment

Q11. Withdraw als/drop-outs unlikely to cause bias

Q10. For continuous data, variability reported

Q9. Patients blinded

Q8. Outcome assessor blinded Yes No Q7. Mean follow -up ≥1y Unclear Q6. Treatments w ere similar betw een groups Not applicable Q5. Intervention cleary defined

Q4. Inclusion criteria clearly specified

Q3. Similar baseline characteristics betw een groups

Q2. Concealment of randomisation

Q1. Randomisation

0% 20% 40% 60% 80% 100%

For non-randomised comparative studies, no studies reported surgeon experience. Mean follow-up was less than one year in two studies.64,66

Figure 2 Summary of quality assessment of the non-randomised comparative studies (n=7)

Q18. Analyses adjusted for confounding factors Q17. Important prognostic factors identified

Q16. Follow -up similar betw een groups Q15. Withdraw als/drop-outs unlikely to cause bias Q14. Information provided on non-responsents/drop-outs

Q13. Mean follow -up ≥1y Q12. Outcome assessor blinded

Q11.Objective outcome measures used Yes Q10. Important outcomes considered No Q9. Appropriate treatment place Unclear Q8. Experienced practitioner Not applicable Q7. Intervention clearly defined Q6. Similar baseline characteristics betw een groups Q5. Data colloection w as undertaken prospectively Q4. Selection of patients w as consecutive Q3. Similar participant baseline characteristics

Q2. Participant inclusion/exclusion criteria described Q1. Participant representativeness

0% 20% 40% 60% 80% 100%

For the registry and case series, mean follow-up was one year or more in 17 studies. The remaining eight studies were included for the safety review only as stated in Section 3.2.1 (Inclusion and exclusion criteria for type of studies). No studies reported methods of follow- up. The drop-out rates varied from 0 to 30%.

Figure 3 Summary of quality assessment of the registry (n=1) and the case series (n=24)

Q17. Important prognostic factors identified Q15. Withdraw als/drop-outs unlikely to cause bias

Q14. Information provided on non-responsents/drop-outs Q13. Mean follow -up ≥1y

Q11.Objective outcome measures used

Q10. Important outcomes considered Yes Q9. Appropriate treatment place No Q8. Experienced practitioner Unc lear

Q7. Intervention clearly defined Not applicable Q5. Data colloection w as undertaken prospectively

Q4. Selection of patients w as consecutive Q3. Similar participant baseline characteristics

Q2. Participant inclusion/exclusion criteria described Q1. Participant representativeness

0% 20% 40% 60% 80% 100%

4.4 Anterior repair

30 studies involving 2472 women provided data on the use of mesh/graft in anterior repair (12 RCTs, 4 non-randomised comparative studies, 1 registry, and 13 case series, see Table 7). The median follow-up time was 14 months (range 1 to 38). Of the 12 RCTs, 7 were reported only as conference abstracts. Two of the RCTs compared different types of mesh/graft and the others compared mesh/graft with no mesh/graft.

Four studies27,30,35,78 used absorbable synthetic mesh, 14 studies30,32,34,41,43,48, 60,61,63,73,74,76,77,79 used absorbable biological graft, one study67 used combined mesh/graft, and 14 studies28,29,37,38,41,42, 53,61,62,71,72,86,87,88 used non-absorbable synthetic mesh. One study34 (n 206) included only women with primary prolapse, one study62 (n 24) only included secondary prolapse, nine studies27,29,60, 61,71,72,73,78, 87 included both primary (n 740) and secondary (n 310) prolapse but did not report the results separately, and the remaining 19 studies did not report whether the prolapses treated were primary or secondary.

Sacrocolpopexy is a surgical technique normally used for vault prolapse repair. However, one study67 reported that it was used for stage IV cystocele repair. This was the only study that used combined mesh (polyglactin and polypropylene) in anterior repair. We were aware that clinically, it can be difficult to identify vault and anterior vaginal wall prolapse, so it is possible that many of the women in this study might have vault prolapse or have both vault prolapse and anterior vaginal wall prolapse. Therefore, the results of this study were presented separately.

4.4.1 Efficacy

Individual studies are summarised by type of outcome (Appendix 12). The crude event rates, by study design and by type of mesh or graft, are summarised in Table 10.

Table 10 Anterior repair, efficacy: summary of crude event rates, by type of mesh/graft, by study design

Study No mesh Absorbable synthetic Absorbable biological Non-absorbable synthetic N n N % n N % n N % n N % A. Subjective failure: persistent prolapse symptoms RCT 2 19 160 11.9% 15 153 9.8% Non-rand 1 0 19 0.0% 1 24 4.2% Case series 4 5 112 4.5% 20 309 6.5% 1 55 1.8% Total 19 179 10.6% 5 112 4.5% 36 486 7.4% 1 55 1.8% 95% CI 6.9% to 16.0% 1.9% to 10.0% 5.4% to 10.1% 0.0% to 6.5% B. Objective failure: recurrent prolapse at original site RCT 12 179 591 30.0% 52 161 32.3% 77 419 18.4% 39 282 13.8% Non-rand 3 5 49 10.2% 23 80 28.8% 1 37 2.7% Case series 8 11 112 9.8% 86 542 15.9% 8 229 3.5% Total 184 640 28.8% 63 273 23.1% 186 1041 17.9% 48 548 8.8% 27 95% CI 25.4% to 32.4% 18.5% to 28.4% 15.7% to 20.3% 6.7% to 11.4% C. De novo prolapse at other sites (rectocele) RCT 0 Non-rand 0 Case series 2 8 58 13.8% 8 45 17.8% Total 8 58 13.8% 8 45 17.8% 95% CI 7.2% to 24.9% 9.3% to 31.3% D. Further operation needed for prolapse (recurrent or de novo) RCT 2 2 85 2.4% 14 62 22.6% 3 63 4.8% 0 92 0.0% Non-rand 0 Case series 5 2 112 1.8% 6 217 4.8% 3 142 2.1% Total 2 85 2.4% 16 174 9.2% 9 280 3.2% 3 234 1.3% 95% CI 0.6% to 8.2% 5.7% to 14.4% 1.7% to 6.0% 0.4% to 3.7%

Study No mesh Absorbable synthetic Absorbable biological Non-absorbable synthetic N n N % n N % n N % n N % E. Persistent urinary incontinence (stress urinary incontinence) RCT Non-rand 1 9 10 90.0% 13 14 92.9% Case series 3 5 49 10.2% 17 44 38.6% Total 9 90 90.0% 5 49 10.2% 13 14 92.9% 17 44 38.6% 95% CI 59.6% to 98.2% 4.4% to 21.8% 68.5% to 98.7% 25.8% to 53.4% F. Persistent bowel symptoms (no studies reported data on this outcome) G. Persistent dyspareunia (no studies reported data on this outcome)

4.4.1.1 Subjective failure: persistent prolapse symptoms

Two RCTs,32,34 one non-randomised comparative study63 and four case series71,73,77,78 involving 832 women reported data on subjective failure after anterior repair (Table 10 A) (see Appendix 12 for the results for individual studies).

Direct comparisons from randomised evidence (mesh/graft vs. no mesh/graft)

Both RCTs32,34 used an absorbable biological graft. Two different types of graft were used. Solvent-derived fascia lata was used by Gandhi and colleagues32 and porcine dermis was used by Meschia and colleagues.34 The follow-up time was a mean of 13 months (range 1.4 to 50) in one RCT32 and one year for all women in the other.34 Meta-analysis shows no statistically significant difference in subjective failure rates between repair with an absorbable biological graft and without (15/153, 10% vs. 19/160, 12%; RR 0.83, 95% CI 0.44 to 1.57, Figure 4).

Figure 4 Anterior repair, efficacy, subjective failure, mesh or graft versus procedures without mesh/graft: evidence from RCTs

Review: Efficacy and safety of using mesh or grafts in surgery for vaginal wall prolapse (RCTs) Comparison: 01 Mesh or graft versus none for anterior vaginal wall prolapse Outcome: 01 Subjective failure: persistent prolapse symptoms

Study Mesh/graft No mesh/graft RR (fixed) Weight RR (fixed) or sub-category n/N n/N 95% CI % 95% CI

01 Absorbable biological graft versus none Gandhi 2005 6/55 6/57 31.73 1.04 [0.36, 3.02] Meschia 2007 9/98 13/103 68.27 0.73 [0.33, 1.63] Subtotal (95% CI) 153 160 100.00 0.83 [0.44, 1.57] Total events: 15 (Mesh/graft), 19 (No mesh/graft) Test for heterogeneity: Chi² = 0.27, df = 1 (P = 0.60), I² = 0% Test for overall effect: Z = 0.59 (P = 0.56)

0.1 0.2 0.5 1 2 5 10 Favours mesh/graft Favours no mesh/graf

Crude event rates (all study types)

When evidence from other study types was also considered (Table 10 A), the subjective failure rate was 10.6% (19/179, 95% CI 6.9% to 16.0) for procedures without mesh/graft, 4.5% (5/112, 95%CI 1.9% to 10.0%) with absorbable synthetic mesh, 7.4% (36/486, 95% CI 5.4 to 10.1%) with absorbable biological graft, and 1.8% (1/55, 95% CI 0 to 5.5%) with non- absorbable synthetic mesh.

Time to failure

No studies in this section reported time to failure.

4.4.1.2 Objective failure: recurrent prolapse at original site

We included any stage of recurrent prolapse at the original site. In the studies reporting this outcome, if more than one stage was reported, the lowest (best) one was used, for instance, if a study reported failure rate at stage II or more and failure rate at stage III or more, the result based on stage II was used.

Twelve RCTs,27,30,32,34,35,37,38,41-43,47,53 three non-randomised comparative studies61-63 and eight case series28,71-74,76,78,79 involving 2470 women reported data on objective failure (Appendix 12).

Direct comparisons from randomised evidence (mesh/graft vs. no mesh/graft)

Figure 5 shows the meta-analysis results of RCTs by type of mesh/graft.

Figure 5 Anterior repair, efficacy, objective failure, mesh or graft versus procedures without mesh/graft: evidence from RCTs

Study Mesh/graft No mesh/graft RR (random) Weight RR (random) or sub-category n/N n/N 95% CI % 95% CI

01 Absorbable synthetic mesh versus none Sand 2001 18/73 30/70 16.69 0.58 [0.35, 0.93] Weber 2001 15/26 36/57 18.32 0.91 [0.62, 1.34] Subtotal (95% CI) 99 127 35.01 0.74 [0.46, 1.18] Total events: 33 (Mesh/graft), 66 (No mesh/graft) Test for heterogeneity: Chi² = 2.31, df = 1 (P = 0.13), I² = 56.7% Test for overall effect: Z = 1.25 (P = 0.21)

02 Absorbable biological graft versus none Gandhi 2005 16/76 23/78 15.52 0.71 [0.41, 1.24] Meschia 2007 7/98 20/103 11.56 0.37 [0.16, 0.83] Hviid 2005 abs. 0/10 0/12 Not estimable Kocjancic 2007 abs 9/85 20/91 12.77 0.48 [0.23, 1.00] Subtotal (95% CI) 269 284 39.85 0.55 [0.37, 0.81] Total events: 32 (Mesh/graft), 63 (No mesh/graft) Test for heterogeneity: Chi² = 1.94, df = 2 (P = 0.38), I² = 0% Test for overall effect: Z = 3.02 (P = 0.002)

03 Non-absorbable synthetic mesh versus none Al-Nazer 2007 abs. 1/20 6/20 3.35 0.17 [0.02, 1.26] Ali 2006 abs. 3/46 5/43 6.20 0.56 [0.14, 2.21] Hiltunen 2006 abs. 7/92 32/85 12.28 0.20 [0.09, 0.43] Nguyen 2007 abs. 1/31 7/32 3.32 0.15 [0.02, 1.13] Subtotal (95% CI) 189 180 25.15 0.24 [0.13, 0.43] Total events: 12 (Mesh/graft), 50 (No mesh/graft) Test for heterogeneity: Chi² = 2.03, df = 3 (P = 0.57), I² = 0% Test for overall effect: Z = 4.69 (P < 0.00001)

Total (95% CI) 557 591 100.00 0.48 [0.32, 0.72] Total events: 77 (Mesh/graft), 179 (No mesh/graft) Test for heterogeneity: Chi² = 20.01, df = 8 (P = 0.01), I² = 60.0% Test for overall effect: Z = 3.54 (P = 0.0004)

0.1 0.2 0.5 1 2 5 10 Favours mesh/graft Favours no mesh/graf

Two RCTs27,35 used absorbable synthetic mesh (polyglactin). The difference in objective failure rate between using absorbable synthetic mesh and procedures without mesh/graft (33/99, 33% vs. 66/127, 52%; RR 0.74, 95% CI 0.46 to 1.18) was not statistically significant.

The follow up in both RCTs was at one year. The heterogeneity, however, was relatively high (P=0.13, I2 = 56.7%).

In the four RCTs which used absorbable biological grafts, one used solvent-dried fascia lata,32 the rest used porcine dermis.34,43,47 The objective failure rate with absorbable biological graft was significantly lower than without mesh/graft (32/269, 12% vs. 63/284, 23%; RR 0.55, 95%CI 0.37 to 0.81). There was no evidence of statistical heterogeneity (I2 = 0%). The mean follow-up times were all one year or more (13 months, 14 months, 1 year) except in the small study by Hviid and colleagues43 (3 months).

Four RCTs37,38,42,53 used non-absorbable synthetic mesh (polypropylene). One of these studies53 used a commercially available device (i.e. a kit with introducer). The objective failure when using non-absorbable synthetic mesh was significantly lower than without mesh/graft (12/189, 6.4% vs. 50/180, 28.8%; RR 0.24, 95% CI 0.13 to 0.43). There was no statistically significant heterogeneity. The follow-up time in two RCTs was one year37,42 and 6 months in the others.38,53

Overall, comparing mesh/graft (any type) with procedures without mesh/graft, the objective failure rate of using mesh/graft was significantly lower (77/557, 13.8% vs. 179/597, 30.0%; RR 0.48, 95% CI 0.32 to 0.72). There was statistically significant heterogeneity (I2 60.0%, P=0.01) which might be explained by the two studies using absorbable synthetic mesh.27,35

Direct comparisons from randomised evidence (different types of mesh/graft)

Figure 6 shows the meta-analysis results of the two RCTs which compared different types of mesh/graft.

Figure 6 Anterior repair, efficacy, objective failure, different types of mesh or graft: evidence from RCTs

Review: Efficacy and safety of using mesh or grafts in surgery for vaginal wall prolapse (RCTs) Comparison: 02 Synthetic mesh versus biological graft for anterior vaginal wall prolapse Outcome: 01 Objective failure: recurrent prolapse at original site

Study Mesh Graft RR (random) Weight RR (random) or sub-category n/N n/N 95% CI % 95% CI

01 Absorbable synthetic mesh versus absorbable biological graft De Ridder 2002 19/62 6/63 47.26 3.22 [1.38, 7.52] Subtotal (95% CI) 62 63 47.26 3.22 [1.38, 7.52] Total events: 19 (Mesh), 6 (Graft) Test for heterogeneity: not applicable Test for overall effect: Z = 2.70 (P = 0.007)

02 Non-absorbable synthetic mesh versus absorbable biological graft Cervigni 2007 abs. 27/93 39/87 52.74 0.65 [0.44, 0.96] Subtotal (95% CI) 93 87 52.74 0.65 [0.44, 0.96] Total events: 27 (Mesh), 39 (Graft) Test for heterogeneity: not applicable Test for overall effect: Z = 2.16 (P = 0.03)

Total (95% CI) 155 150 100.00 1.38 [0.28, 6.86] Total events: 46 (Mesh), 45 (Graft) Test for heterogeneity: Chi² = 11.78, df = 1 (P = 0.0006), I² = 91.5% Test for overall effect: Z = 0.40 (P = 0.69)

0.1 0.2 0.5 1 2 5 10 Favours mesh Favours graft

One RCT30 reported the objective failure rate between absorbable synthetic mesh (polyglactin) and absorbable biological graft (porcine dermis): the result significantly favoured absorbable biological graft (19/62, 31% vs. 6/63, 10%; RR 3.22, 95% CI 1.38 to 7.52). Mean follow-up time was 26 months. Another RCT41 compared a non-absorbable synthetic mesh (polypropylene) with an absorbable biological graft (porcine dermis) with results significantly favouring non-absorbable synthetic mesh (27/93, 29% vs. 39/87, 45%; RR 1.38, 95% CI 0.28 to 6.86). Follow-up time ranged from 6 to 28 months.

Crude event rates (all study types)

When evidence from other study types was also considered (Table 10 B), the objective failure rate was 28.8% (184/640, 95% CI 25.4 to 32.4%) for procedures without mesh/graft, 23.1% (63/273, 95% CI 18.5 to 28.4%) for absorbable synthetic mesh, 17.9% (186/1041, 95% CI 15.7 to 20.3%) for absorbable biological graft, and 8.8% (48/548, 95% CI 6.7 to 11.4%) for non-absorbable synthetic mesh.

Bayesian meta-analysis suggested that non-randomised comparative studies had lower failure rates for mesh, which may indicate a bias when including data from these sources (Table 11a). Results adjusted for the study design showed that both absorbable biological graft and non-absorbable synthetic mesh appeared to be superior to no mesh (Table 11b). The data also favoured absorbable synthetic mesh but the difference was not statistically significant.

Indirect comparisons (Table 11c) suggested that non-absorbable synthetic mesh had a significantly lower objective failure rate than both absorbable synthetic mesh and absorbable biological graft but the difference between absorbable synthetic mesh and absorbable biological graft was not statistically significant.

Table 11a Comparison of odds ratios of objective failure between different study designs OR 95% CrI RCTs Reference study design - Non-randomised comparative studies 0.72 0.18 o 3.25

Table 11b Bayesian meta-analysis models, anterior repair: objective failure, recurrent prolapse at original site Categories n N OR (adjusted for study design) 95% CrI No mesh/graft 184 640 Reference technique - Absorbable synthetic mesh 52 161 0.82 0.50 to 1.32 Absorbable biological graft 120 555 0.51* 0.36 to 0.72 Non-absorbable synthetic mesh 41 344 0.19* 0.12 to 0.30

Table 11c Indirect comparison, anterior repair: objective failure, recurrent prolapse at original site Comparisons OR (adjusted for study design) 95% CrIc Absorbable synthetic mesh versus absorbable biological graft 0.64 0.36 to 1.06 Non-absorbable synthetic mesh versus absorbable synthetic mesh 0.23* 0.12 to 0.44 Non-absorbable synthetic mesh versus absorbable biological graft 0.37* 0.23 to 0.59

Notes:

*Statistically significant 1. Analyses were based on RCTs and non-randomised comparative studies. 2. n = cumulative number of patients experiencing the event, N = cumulative number of patients analysed by the studies. 3. CrI. Credible interval with 95% probability of containing the true OR.

Time to failure

Handel and colleagues61 reported a mean time to failure of 4.9 months (range 0.5–20) but did not report timing for each group separately. In this non-randomised comparative study, absorbable biological graft (porcine dermis) and non-absorbable synthetic mesh (polypropylene) were compared with procedures without mesh/graft. Kobashi and colleagues76 reported that 10 out of 17 failures occurred within 3 months. No other studies reported time to failure.

Evidence from another non-randomised comparative study (original site prolapse and de novo site prolapse not reported separately)

Powell and colleagues77 reported prolapse after surgery in any compartment (stage ≥2) during follow-up (median 24.7 months, range 12-57), but did not report rates for each compartment separately, therefore we could not distinguish recurrent from de novo prolapse. The rates were 23.1% (9/39) in women treated with donor fascia lata and 15.8% (3/19) in women treated with autologous fascia lata (both are biological absorbable grafts). Mean time to failure was 14.4 months (2-40).

Evidence from another case series (cystocele treated with sacrocolpopexy)

In this study67, at a median of 12 months (range 1-50) follow-up, 8/50 (16.0%) patients were found to have a recurrent stage III or IV cystocele after procedures using combined mesh (polyglactin and polypropylene). The authors also stated that recurrences occurred only when the mesh extended from the mid-bladder base to the sacrum, i.e. partial support; but no recurrences occurred when the mesh supporting the bladder extended from the urethra to the sacrum, i.e. complete support.

4.4.1.3 De novo prolapse (i.e. new prolapse at another site)

Three case series71,73,77 involving 296 women reported data on this outcome (Appendix 12): two studies used absorbable biological graft73,77 and one study used non-absorbable synthetic mesh.71

Direct comparison was not possible as there were no RCTs reporting this outcome. One study73 reported de novo prolapse at different sites separately (rectocele, vault prolapse and enterocele). One study71 reported de novo rectocele. Crude event rates were calculated for de novo rectocele. The rate was 13.8% (95%CI 7.2 to 24.9%) for the procedures with absorbable biological graft (mean follow up 17 months, range 6 to 60) and 17.8% (95%CI 9.3 to 31.3%) with a non-absorbable synthetic mesh (mean follow up 37 months, range 24 to 60). In the study reporting de novo vault prolapse and enterocele73, the rate was 4.2% (7/165, absorbable biological graft).

Powell and colleagues77 reported that 17.9% (5/28) of women had de novo prolapse at a mean follow up of 25 months (12 to 57) but did not report the sites of prolapse.

No studies reported when the new prolapse occurred.

Evidence from another case series (cystocele treated with sacrocolpopexy)

Cronje and colleagues67 reported that 2.0% (1/50) of patients treated with sacrocolpopexy for stage IV cystocele developed stage III or IV rectocele; and 2.0% (1/50) developed stage III or IV vaginal vault prolapse.

4.4.1.4 Need for further surgery for prolapse

Both recurrent prolapse and de novo prolapse requiring further operation were considered in this section.

Two RCTs30,42 and five case series72,74,77-79 involving 773 women reported data for this outcome (See Appendix 12).

Direct comparisons (mesh/graft vs. no mesh/graft)

One RCT42 used non-absorbable synthetic mesh (polypropylene). There was no statistically significant difference in the numbers of women initially treated with mesh who needed a further operation for recurrent prolapse compared with procedures without mesh/graft (0/92, 0% vs. 2/85, 2%; RR 0.18, 95% CI 0.01 to 3.80).

Direct comparison (different types of mesh/graft)

One RCT30 compared absorbable synthetic mesh (polyglactin) and absorbable biological graft (porcine dermis). More women using absorbable synthetic mesh required reoperation compared with those using absorbable biological grafts (14/62, 23% vs. 3/63, 5%; RR 4.74, 95% CI 1.43 to 15.69).

Crude event rates (all study types)

When evidence from other study types was also considered, a further operation was needed for prolapse in 2.4% (2/85, 95% CI 0.6 to 8.2%) for procedures without mesh/graft, 9.2% (16/174, 95% CI 5.7 to 14.4%) for absorbable synthetic mesh, 3.2% (9/280, 95% CI 1.7 to 6.0%) for absorbable biological graft, and 1.3% (3/234, 95% CI 0.4 to 3.7%) for non- absorbable synthetic mesh (see Table 10 D).

4.4.1.5 Persistent urinary incontinence

One non-randomised comparative study63 and seven case series28,72,73,76-79 involving 639 women reported persistent stress urinary incontinence, urge incontinence, or any type of urinary incontinence (see Appendix 12). The mean follow-up time ranged from 12 months to three years. For most studies, it was impossible to estimate the number of women having a concomitant urinary incontinence operation at the time of surgery.

Crude event rates (all study types)

Crude event rates were calculated for persistent stress urinary incontinence. The rate was 90.0% (9/10, 95% CI 59.6 to 98.2%) for the procedure without a graft, 10.2% (5/49, 95%CI 4.4 to 21.8%) with absorbable synthetic mesh, 92.9% (13/14, 95% CI 68.5 to 98.7%) with absorbable biological graft, and 38.6% (17/44, 95% CI 25.7 to 53.4%) with non-absorbable synthetic mesh (see Table 10 E).

In the studies28,63,72,78 reporting persistent urge incontinence (n 122 in total, any types of mesh), the rates ranged from 13.3 to 100%. In the studies reporting persistence of all types of incontinence (n 400 in total, all used absorbable biological graft), the rates ranged from 22.2 to 54.6%.

4.4.1.6 Persistent bowel symptoms

No studies reported persistent bowel symptoms in women having the symptoms at baseline.

4.4.1.7 Persistent dyspareunia

No studies reported persistent dyspareunia in women having the symptoms at baseline.

4.4.1.8 Improvement of disease-specific quality of life (RCTs and non-randomised comparative studies only)

One non-randomised comparative study60 involving 28 women reported this outcome using a validated prolapse associated quality of life questionnaire (PQOL, Appendix 12). After a two- year follow up, there was no statistically significant difference in the mean quality of life scores between absorbable biological graft group (small intestine submucosa) and no mesh/graft group (p = 0.137).

4.4.2 Safety

The outcomes from individual studies are summarised by type of outcomes in Appendix 13. The crude event rates are summarised in Table 12.

Studies reporting no adverse events

Gomelsky and colleagues74 using absorbable biological graft (porcine dermis, n=70, mean follow up 24 months), Kocjancic and colleagues47 using absorbable biological graft (porcine dermis, n=85, follow up 2 years), Ali and colleagues38 using non-absorbable synthetic mesh (polypropylene, n=54, follow up 6 months), and Julian and colleagues62 using non- absorbable synthetic mesh (polypropylene, n=12, follow up 2 years), reported that no significant complications occurred during or around the time of surgery. Sand and colleagues27 using absorbable synthetic mesh (polypropylene, n=73, follow up 1 year) and Hviil and colleagues43 using absorbable biological graft (porcine dermis, n=19, follow up 3 months) reported that no complications occurred during follow up.

Study which used sacrocolpopexy to treat cystocele

Cronje and colleagues67 reported that no major perioperative complications (no details reported) occurred in 50 patients with stage IV cystocele who were treated with sacrocolpopexy (where a mesh has to be used). A combined synthetic mesh (polyglactin and polypropylene) was used in this study.

Table 12 Anterior repair, safety: summary of crude event rates, by type of mesh/graft, by study design Study No mesh Absorbable synthetic Absorbable biological Non-absorbable synthetic N n N % n N % n N % n N % A. Blood loss (number of women requiring blood transfusion) RCT 1 1 74 1.4% 0 35 0.0% Non-rand 2 0 14 0.0% 1 70 1.4% Case series 5 0 0 0.0% 0 112 0.0% 2 128 1.6% 4 161 2.5% Total 1 88 1.1% 0 147 0.0% 3 198 1.5% 4 161 2.5% 95% CI 0.2% to 6.2% 0.0% to 2.5% 0.5% to 4.4% 1.0% to 6.2% B. Damage to surrounding organs during operation RCT 0 Non-rand 1 0 19 0.0% 0 24 0.0% Case series 5 0 0 0.0% 0 112 0.0% 0 70 0.0% 6 251 2.4% Total 0 19 0.0% 0 112 0.0% 0 94 0.0% 6 251 2.4% 95% CI 0.0% to 16.8% 0.0% to 3.3% 0.0% to 3.9% 1.1% to 5.1% C. Mesh/graft erosion RCT 7 1 35 2.9% 2 270 0.7% 27 262 10.3% Non-rand 4 12 94 12.8% 2 37 5.4% Case series 9 0 112 0.0% 21 217 9.7% 39 367 10.6% Total 1 147 0.7% 35 581 6.0% 68 666 10.2% 95% CI 0.1% to 3.8% 4.4% to 8.3% 8.1% to 12.7% D. Further operation needed for mesh/graft erosion RCT 2 1 35 2.9% 1 98 1.0% Non-rand 2 3 56 5.4% 1 12 8.3% Case series 3 22 335 6.6% Total 1 35 2.9% 4 154 2.6% 23 347 6.6% 95% CI 0.0% to 3.3% 1.0% to 6.5% 4.5% to 9.7% E. De novo urinary incontinence (stress urinary incontinence) RCT 0 Non-rand 0 Case series 4 0 63 0.0% 3 42 7.1% 3 44 6.8% Total 0 63 0.0% 3 42 7.1% 3 44 6.8% 95% CI 0.0% to 5.7% 2.5% to 19.0% 2.3% to 18.2% F. De novo bowel symptoms (no studies reported this outocme) G. De novo dyspareunia RCT 0 Non-rand 0 Case series 1 4 11 36.4% Total 4 11 36.4% 95% CI 15.2% to 64.6% H. Infection RCT 1 4 97 4.1% 1 105 1.0% Non-rand 4 0 45 0.0% 1 94 1.1% 0 12 0.0% Case series 7 0 112 0.0% 4 383 1.0% 10 441 2.3% Total 4 142 2.8% 0 112 0.0% 5 477 1.0% 11 558 2.0%

Study No mesh Absorbable synthetic Absorbable biological Non-absorbable synthetic N n N % n N % n N % n N % 95% CI 0.0% to 0.0% 0.0% to 3.3% 0.4% to 2.4% 1.1% to 3.5% I. Other serious adverse events RCT 1 1 74 1.4% 0 35 0.0% Non-rand 2 0 19 0.0% 1 80 1.3% Case series 3 1 132 0.8% 4 248 1.6% Total 1 93 1.1% 0 35 0.0% 2 212 0.9% 4 248 1.6% 95% CI 0.2% to 5.8% 0.0% to 9.9% 0.3% to 3.4% 0.6% to 4.1%

4.4.2.1 Blood loss

Blood loss, ml

Four RCTs,34,38,43,47 the registry,29 and two case series72,77 involving 843 women reported this outcome (see Appendix 13).

In direct comparisons (mesh/graft vs. no mesh/graft), three RCTs34,38,47 reported a mean (SD) blood loss (Figure 7). Two RCTs34,47 used absorbable biological graft (porcine dermis). There was no statistically significant difference in blood loss between using absorbable biological graft and procedures without mesh/graft (RR -16 ml, 95% CI -36 to 4). One RCT38 used non-absorbable synthetic mesh (polypropylene). There was no statistically significant difference between using non-absorbable synthetic mesh and procedures without mesh/graft (RR 14, 95% CI -16 to 44). There was no overall statistically significant difference in blood loss between using mesh/graft (any type) and procedures without mesh/graft (WMD -7ml, 95% CI -23 to 10).

Figure 7 Anterior repair, safety, blood loss (ml), mesh or graft versus procedures without mesh/graft: evidence from RCTs

Study Mesh/graft No mesh/graft WMD (fixed) Weight WMD (fixed) or sub-category N Mean (SD) N Mean (SD) 95% CI % 95% CI

01 Absorbable biological graft versus none Meschia 2007 98 151.00(112.00) 103 167.00(96.00) 34.06 -16.00 [-44.90, 12.90] Kocjancic 2007 abs 100 151.00(112.00) 106 167.00(96.00) 34.88 -16.00 [-44.56, 12.56] Subtotal (95% CI) 198 209 68.94 -16.00 [-36.32, 4.32] Test for heterogeneity: Chi² = 0.00, df = 1 (P = 1.00), I² = 0% Test for overall effect: Z = 1.54 (P = 0.12)

02 Non-absorbable synthetic mesh versus none Ali 2006 abs. 54 64.50(70.40) 54 50.30(89.00) 31.06 14.20 [-16.07, 44.47] Subtotal (95% CI) 54 54 31.06 14.20 [-16.07, 44.47] Test for heterogeneity: not applicable Test for overall effect: Z = 0.92 (P = 0.36)

Total (95% CI) 252 263 100.00 -6.62 [-23.49, 10.25] Test for heterogeneity: Chi² = 2.64, df = 2 (P = 0.27), I² = 24.1% Test for overall effect: Z = 0.77 (P = 0.44)

-100 -50 0 50 100 Favours mesh/graft Favours no mesh/graf

In the other studies which reported this outcome, the RCT by Hviid and colleagues43 reported a median (range) for blood loss. The blood loss was 75ml (range 25 to 150ml) in 10 women treated with absorbable synthetic graft and 57ml (range 25 to 150ml) in 12 women treated with procedures without mesh/graft. The mean blood losses in the registry (n=106, non-absorbable synthetic mesh, commercially available, 190ml) and two case series (n=58, absorbable biological graft, 329ml; n=142, non-absorbable synthetic mesh, 442ml) were all less than 450ml.

Blood loss requiring blood transfusion

One RCT,35 two non-randomised comparative studies,60,61 the registry,29 and four case series71,74,77,78 involving 594 women reported this outcome (see Appendix 13).

In a direct comparison, the RCT used absorbable synthetic mesh (polyglactin).35 There was no significant difference in the number of women requiring blood transfusion compared with procedures with no mesh/graft (0/35, 0% vs. 1/74, 1.4%; RR 0.69, 95% CI 0.03 to 16.63).

When evidence from other study types was also considered, 1.1% (1/88, 95% CI 0.2 to 6.2%) of women treated with procedures without mesh/graft required blood transfusion, 0% (0/147, 95% CI 0 to 2.5%) for absorbable synthetic mesh, 1.5% (3/198, 95% CI 0.5% to 4.4%) for absorbable biological graft, and 2.5% (4/161, 95% CI 1.0% to 6.2%) for non- absorbable synthetic mesh (see Table 12 A).

4.4.2.2 Damage to surrounding organs during operation

One non-randomised comparative study,63 the registry,29 and four case series71,74,78,87 involving 476 women reported this outcome (see Appendix 13). Six cases of damage occurred in total with all associated with the use of non-absorbable synthetic mesh.

Four cases were reported as bladder injury or perforation with two cases occurring in the registry29 where commercially available insertion devices were used, and two cases, also occurred in the registry, reported as urethral perforation.

Crude event rates (all study types)

When evidence from all study types was considered, damage to surrounding organs occurred in 0% (0/19, 95% CI 0 to 16.8%) of procedures without mesh/graft, 0% (0/112, 95% CI 0 to 3.3%) for absorbable synthetic mesh, 0% (0/94, 95% CI 0 to 3.9%) for absorbable biological graft, and 2.4% (6/251, 95% CI 1.1 to 5.1%) for non-absorbable synthetic mesh (see Table 12 B).

4.4.2.3 Mesh/graft erosion

Outcomes reported as mesh/graft erosion, vaginal erosion, vaginal mesh extrusion, and minor mesh exposure were considered in this section.

Seven RCTs,34,35,38,41,42,47,53 four non-randomised comparative studies,60-63 and nine case series28,71,72,74,77-79,86 (two case series were reported separately in one paper86) involving 1394 women reported this outcome (see Appendix 13).

Direct comparison (different types of mesh/graft)

One RCT41 compared the occurrence of mesh/graft erosion between procedures using absorbable biological graft (porcine dermis) and non-absorbable synthetic mesh (polypropylene) during a follow-up time of 6 to 28 months. No statistically significant difference was found in the occurrence of mesh/graft erosion (0/87, 0% vs. 6/93, 6.5%; RR 0.08, 95% CI 0 to 1.44).

Crude event rates (all study types)

When evidence from other study types was also considered (Table 12 C), mesh/graft erosion occurred in 0.7% (1/147, 95% CI 0.1 to 3.8%) of procedures using absorbable synthetic mesh (with a mean follow-up time ranging from 21 to 23 months); 6.0% (35/581, 95% CI 4.4 to 8.3%) for absorbable biological graft (with a mean follow-up time ranging from 14 to 25 months except one study which was 1 month); and 10.2% (68/666, 95% CI 8.1 to 12.7%) for procedures with non-absorbable synthetic mesh (with a mean follow-up time ranging from 6 to 37 months).

Time to occurrence of erosion

De Tayrac and colleagues71 (at a mean follow up of 37 months, range 24 to 60) reported that all erosions (5/55) occurred between 6 weeks and 6 months. Deffieux and colleagues86 reported that 16 of 27 erosions occurred within 12 weeks, and the remaining 11 occurred between 12 weeks and 21 months. Flood and colleagues72 reported that 3/143 erosions occurred at 3 months, 4 years and 7 years respectively. Julian and colleagues62 reported that all mesh/graft erosions (3/12) occurred within 6 months.

4.4.2.4 Further operation needed for mesh/graft erosion

Outcomes reported as mesh removal and partial removal/excision were considered in this section. Two RCTs,34,35 two non-randomised comparative studies61,62 and three case series71,72,86 involving 536 women reported this outcome (see Appendix 13).

Crude event rates (all study types)

Mesh removal was required by 2.9% (1/35, 95% CI 0.5 to 14.5%) of women treated with absorbable synthetic mesh, 2.6% (4/154, 95% CI 1.0 to 6.5%) with absorbable biological graft, and 6.6% (22/335, 95% CI 4.5 to 9.7%) with non-absorbable synthetic mesh (see Table 12 D).

Erosion managed by conservative treatment

Powell and colleagues77 reported mesh/graft erosions (6/58) resolved with debridement and vaginal oestrogen. Simsiman and colleagues79 reported that mesh/graft erosions (15/89) were resolved by wound debridement and vaginal oestrogen.

4.4.2.5 De novo urinary incontinence

Six case series72,73,77-79,88 involving 248 women with no urge incontinence at baseline, 149 women with no stress urinary incontinence at baseline, and 30 women with no incontinence (of any type) reported the occurrence of de novo urinary incontinence (see Appendix 13). The mean follow-up time ranged from 1 to 3.2 years except for one study88 (3 months).

Direct comparison was not possible because no RCT reported this outcome. Crude event rates were calculated for de novo stress urinary incontinence which were reported by four case series77-79,88. The rate was 0% (0/63, 95% CI 0 to 5.7%) in women treated with absorbable synthetic mesh, 7.1% (3/42, 2.4 to 19.0%) with absorbable biological graft and 6.8 % (3/44, 95% CI 2.3 to 18.2%) with non-absorbable synthetic mesh (Table 12 E).

Three studies reported de novo urge incontinence.73,78,79 One study78 used absorbable synthetic mesh with occurrence of 7.5% (5/67). Two studies73,79 used absorbable biological graft (fascial lata or porcine dermis) with occurrence of 8.5% (13/153) and 28.6% (8/28) respectively.

One study72 did not report type of incontinence. Non-absorbable synthetic mesh was used in this study. De novo incontinence occurred at 3.3% (1/30).

4.4.2.6 De novo bowel symptoms No studies reported data for this outcome in women free of bowel symptoms at baseline.

4.4.2.7 De novo dyspareunia

One case series using non-absorbable synthetic mesh28 involving 11 women reported this outcome (see Appendix 13). The crude event rate was 36.4% (4/11, 95% CI 15.2 to 64.6%) after a mean of 17 month follow up (Table 12 G).

4.4.2.8 Infection

Outcomes reported as infection, mesh infection, urinary tract infection, wound infection, and retropublic abscess were considered in this section.

One RCT,42 four non-randomised comparative studies,60-63 the registry29 and six case series71-73,76,78,86 involving 1289 women reported this outcome (see Appendix 13).

Direct comparison (mesh/graft vs. no mesh/graft)

The RCT42 used non-absorbable synthetic mesh (polypropylene). The occurrence of infection (reported as postoperative infection) was not statistically significantly different from procedures without mesh/graft (1/105, 1.0% vs. 4/97, 4.1%; RR 0.23, 95% CI 0.03 to 2.03)

Crude event rates (all study types)

When evidence from other study types was also considered, infection occurred in 2.8% (4/142, 95% CI 1.1 to 7.0%) of procedures without mesh/graft, 0% (0/112, 95% CI 0 to 3.3%) for absorbable synthetic mesh, 1.0% (5/477, 95% CI 0.4 to 2.4%) for absorbable biological graft, and 2.0% (10/558, 95% CI 1.1 to 3.5%) for non-absorbable synthetic mesh (see Table 12 H).

4.4.2.9 Other serious adverse effects

One RCT,35 two non-randomised comparative studies,61,63 the registry,29 and two case series involving 588 women reported outcomes on other serious adverse effects including pulmonary embolism, return to theatre, fistula, osteitis pubis, osteomyelitis, anaemia, postoperative stroke, and postoperative myocardial ischaemia (Appendix 13).

Direct comparison (mesh/graft vs. no mesh/graft)

The RCT35 used absorbable synthetic mesh. The adverse event reported was pulmonary embolism. No case occurred in the mesh group (0/35) and one occurred in the group without mesh/graft (1/74). The difference was not statistically significant.

Crude event rates (any study design)

In the other studies, in total, six cases of serious adverse events occurred (Table 12 I). One (1/56, 1.8%) was return to theatre61 (reason unclear) after using absorbable biological graft (porcine dermis). One (1/132, 0.8%) was osteitis pubis after using porcine dermis.76 One (1/106, 1.0%) was anaemia during hospital stay after using non-absorbable synthetic mesh (polypropylene, commercial available device).29 One (1/142, 0.7%) postoperative stroke occurred in a woman with a history of heart disease after using non-absorbable synthetic mesh (polypropylene) (she was lost to follow-up six weeks after surgery). The other two cases (2/142, 1.4%) were postoperative myocardial ischaemia after using non-absorbable synthetic mesh (polypropylene)72 (no details of the myocardial ischaemia were reported).

4.4.3 Other outcomes (operation time and surgeon experience)

The outcomes from individual studies are summarised in Appendix 14.

4.4.3.1 Operation time

One RCT,43 the registry29 and one case series71 involving 183 women reported this outcome. The RCT compared the median (range) operation time between procedures using absorbable synthetic mesh and those without mesh /graft. The median operation time was less than 40 minutes for the mesh group and for the no mesh group. The cases reported in the registry used a commercially available kit device (Prolift) with non-absorbable synthetic mesh (polypropylene). The mean operation time was 73 minutes (SD 32). The case series used non-absorbable synthetic mesh, with a median operation time of 102 minutes (range 50 to 150).

4.4.3.2 Surgeon experience

Only the registry report29 reported information on surgeon experience. All surgeons were experienced in pelvic floor surgery and had received pretrial hands-on training in the use of the commercially available device. No information on learning curve and minimum number of patients treated in the training was reported.

4.4.4 Conclusions of the evidence for and against the use of mesh/graft in anterior vaginal wall prolapse repair

The median follow-up time in the 30 studies reporting data on anterior repair was 14 months (range 1 to 38).

Efficacy

The crude event rates (95% CI) (any study designs) in Table 10 (efficacy) are presented graphically in Figure 8.

Figure 8 Efficacy of anterior repair, summary of crude event rates (95% CI, any study design), by type of mesh/graft

9/10, 5/49, 13/14, 17/44 100 90 80 70 60 50 40 184/640, 63/273, 186/1041, 48/548 NA, NA, 8/58, 8/45 30 20 19/179, 5/112, 36/484, 1/55 Crude rate (%) rate Crude 10 0 Subjective Objective failure de novo Further Persistent Persistent Persistent failure prolapse operation urinary bow el symptom dyspareunia (rectocele) needed for incontinence (no data) (no data) prolapse (stress urinary (recurrent or de incontinence) novo)

No mesh/graft Absorbable synthetic mesh Absorbable biological graft Non-absorbable synthetic mesh

Sufficient evidence was only available for objective failure rate. There was a trend in objective failure rate in that procedures without mesh/graft had the highest failure rate (28.8%), followed by procedures with absorbable synthetic mesh (23.1%), absorbable biological graft (17.9%), and finally non-absorbable synthetic mesh had the lowest failure rate (8.8%) (Table 10 B). In the comparison between different types of mesh/graft, Bayesian meta-analysis and indirect comparison suggested that absorbable synthetic mesh and absorbable biological graft had statistically significantly higher objective failure rate than non- absorbable synthetic mesh (OR 4.12, 95% Crl 2.27 to 7.70, and OR 2.97, 95% Crl 1.93 to 4.61 respectively) (Table 11c). There was insufficient evidence to compare other efficacy outcomes between mesh/graft (any type) and no mesh/graft, nor for comparisons between different types of mesh/graft.

Safety

The crude event rates (95% CI) (any study designs) in Table 12 (safety) are presented in Figure 9.

Figure 9 Safety of anterior repair, summary of crude event rates (95% CI, any study design), by type of mesh/graft

70 NA, NA, NA, 4/11 60

) 50 40

30 1/88, 0/147, 3/198, 4/161 NA, 1/147, 35/581, 68/666 NA, 0/63, 3/42, 3/44 0/19, 0/112, 0/94, 6/251 NA, 1/35, 4/154, 23/347 4/142, 0/112, 5/477, 11/558

Crude rate (% 20

No mesh/graft Absorbable synthetic mesh Absorbable biological graft Non-absorbable synthetic mesh

For most safety outcomes, the event rates were <5%. For mesh/graft erosion, there appears to be a trend that absorbable synthetic mesh had the lowest erosion rate, followed by absorbable biological graft, and non-absorbable synthetic mesh had the highest erosion rate. However, there was insufficient evidence to compare mesh/graft erosion and any other safety outcomes between mesh/graft (any type) and no mesh/graft, nor for comparisons between different types of mesh/graft.

For other outcomes (operation time and surgeon experience), the mean operation times were all less than 120 minutes in the studies which reported this outcome. Only the registry report29 reported information on surgeon experience: all surgeons were experienced in pelvic floor surgery and had received pretrial hands-on training in the use of the commercially available device.

A balance sheet is presented in the following page to summarise the evidence for and against the use of mesh/graft in anterior vaginal wall prolapse repair.

Balance sheet: conclusion of the evidence for and against the use of mesh/graft comparing with no mesh/graft in anterior vaginal wall prolapse repair Absorbable synthetic mesh Absorbable biological graft Non-absorbable synthetic mesh N. participant N. participant N. participant For Unclear Against For Unclear Against For Unclear Against (N. study) (N. study) (N. study) Efficacy Subjective failure rate √ 112 (1) √ 665 (5) √ 55 (1) Objective failure rate √ 400 (4) √ 1418 (13) √ 783 (11) De novo prolapse (i.e. at other sites) √ 0 √ 238 (3) √ 45 (1) Further operation needed for prolapse √ 174 (2) √ 280 (4) √ 319 (2) Persistent urinary symptoms √ 94(2) √ 484 (5) √ 44 (2) Persistent bowel symptoms √ 0 √ 0 √ 0 Persistent dyspareunia √ 0 √ 0 √ 0 Improvement of disease-specific QoL √ 0 √ 28 (1) √ 0 Safety Blood loss √ 147 (2) √ 641 (7) √ 411 (4) Damage to surrounding organs √ 112 (1) √ 113 (2) √ 251 (3) Mesh/graft erosion √ 147 (2) √ 581 (7) √ 666 (9)

48 Further operation needed for mesh/graft erosion √ 35 (1) √ 154 (2) √ 347 (4) De novo urinary incontinence √ 63 (1) √ 198 (3) √ 74 (2) De novo bowel symptoms √ 0 √ 0 √ 0 De novo dyspareunia √ 0 √ 0 √ 11 (1) Infection √ 112 (1) √ 510 (5) √ 667(5) Other serious adverse effects √ 109 (1) √ 231 (3) √ 248 (2) Other outcomes Operation time √ 0 √ 22 (1) √ 161 (2) Surgeon experience √ 0 √ 0 √ 106 (1)

Note:

QoL: quality of life For: statistically significant difference, favouring mesh/graft in a direct comparison (RevMan meta-analysis) where only RCTs were considered and in the Bayesian meta-analysis where all types of study designs were considered. Against: statistically significant difference, favouring no mesh/graft in RevMan meta-analysis and Bayesian meta-analysis. Unclear: insufficient evidence or no statistically significant difference in RevMan meta-analysis or Bayesian meta-analysis.

4.5 Posterior repair

Only 9 studies involving 417 women treated with mesh/graft reported data on the use of mesh in posterior repair (3 RCTs, 2 non-randomised comparative studies, one registry report, and 3 case series, see Table 7). We did not find any RCTs which compared different types of mesh/graft for posterior repair. The median follow-up time was 12 months (range 1 to 17).

Three studies27,55,65 used absorbable synthetic mesh, three studies36,64,80 used absorbable biological graft, two studies55,89 used combined mesh/graft, and two studies28, 29 used non- absorbable synthetic mesh. One study29 included both primary (n 48) and secondary (n 23) prolapse but did not report the results separately, and the other eight studies did not report whether the prolapses treated were primary or secondary.

4.5.1 Efficacy

The outcomes from individual studies are summarised, by types of outcome, in Appendix 15. The crude event rates are summarised in Table 13.

Table 13 Posterior repair, efficacy: summary of crude event rates, by type of mesh/graft, by study design Study No mesh Absorbable synthetic Absorbable biological Combined Non-absorbable synthetic N n N % n N % n N % n N % n N % A. Subjective failure: persistent prolapse symptoms RCT 1 9 60 15.0% 6 28 21.4% Non-rand 0 Case series 1 3 50 6.0% Total 9 60 15.0% 9 78 11.5% 95% CI 8.1% to 26.1% 6.2% to 20.5% B. Objective failure: recurrent prolapse at original site RCT 2 17 122 13.9% 6 65 9.2% 12 26 46.2% Non-rand 2 1 20 5.0% 0 5 0.0% 2 17 11.7% Case series 2 5 50 10.0% 2 31 6.5% Total 18 142 12.7% 6 70 8.6% 19 93 20.4% 2 31 6.5% 95% CI 8.2% to 19.1% 4.0% to 17.5% 13.5% to 29.7% 1.8% to 20.7% 50 C. De novo prolapse at other sites (no studies reported data for this outcome) D. Further operation needed for prolapse RCT 1 3 70 4.3% 2 29 6.9% Non-rand 0 Case series 0 Total 3 70 4.3% 2 29 6.9% 95% CI 1.5% to 11.9% 1.9% to 22.0% E. Persistent urinary incontinence (no studies reported data for this outcome) F. Persistent bowel symptoms RCT 1 19 58 32.8% 5 29 17.2% Non-rand 0 Case series 2 9 53 17.0% 5 43 11.6% Total 19 58 32.8% 14 82 17.1% 5 43 11.6% 95% CI 22.1% to 45.6% 10.5% to 26.6% 5.2% to 24.6%

Study No mesh Absorbable synthetic Absorbable biological Combined Non-absorbable synthetic N n N % n N % n N % n N % n N % G. Persistent dyspareunia RCT 0 Non-rand 0 Case series 1 5 14 35.7% Total 5 14 35.7% 95% CI 16.3% to 61.2%

4.5.1.1 Subjective failure: persistent prolapse symptoms

One RCT36 and one case series80 involving 138 women reported this outcome (see Appendix 15). The RCT compared absorbable biological graft with procedures without mesh/graft. There was no significant difference found (6/28, 21.4% vs. 9/60, 15.0%; RR 1.43, 95% CI 0.56, 3.62). The case series used another type of absorbable biological graft, solvent-dried fascia lata. The subjective failure rate after a mean of 13.7 months (range 6 to 23) follow-up was 6.0% (3/50). When the evidence from both study types was considered, the subjective failure rate was 15.0% (9/60, 95% CI 8.1 to 26.1%) for procedures without mesh/graft and 11.5% (9/78, 95% CI 6.2 to 20.5%) for absorbable biological graft (see Table 13 A).

4.5.1.2 Objective failure: recurrent prolapse at original site

Two RCTs,27,36 two non-randomised comparative studies64,65 and two case series28,80 involving 336 women reported objective failure (see Appendix 15). Mean follow-up time ranged from 12 to 14 months except for the study by Altman and colleagues64(6 months). No studies reported time to failure.

Direct comparison (mesh/graft vs. no mesh/graft)

One RCT27 used absorbable synthetic mesh (polyglactin). No significant difference in objective failure was found between absorbable synthetic mesh and procedures without mesh/graft (6/65, 9.2% vs. 7/67, 10.4%; RR 0.88, 95% CI 0.31 to 2.49, Figure 10). The other RCT36 used absorbable biological graft (porcine dermis). The failure rate with absorbable biological graft was significantly higher than after procedures without mesh/graft (12/26, 46.2% vs. 10/55, 18.2%; RR 2.54, 95% CI 1.26 to 5.10). In total, the objective failure rate with mesh/graft (any type) was not statistically significantly different for procedures without mesh/graft (18/91, 19.8% vs. 17/122, 13.9%; RR 1.61, 95% CI 0.57 to 4.54), but the statistical heterogeneity was high (I2 = 64.6%).

Figure 10 Posterior repair, efficacy, objective failure, mesh or graft versus procedures without mesh/graft: evidence from RCTs

Review: Efficacy and safety of using mesh or grafts in surgery for vaginal wall prolapse (RCTs) Comparison: 03 Mesh or graft versus none for posterior vaginal wall prolapse Outcome: 02 Objective failure: recurrent prolapse at original site

Study Mesh/graft No mesh/graft RR (random) Weight RR (random) or sub-category n/N n/N 95% CI % 95% CI

01 Absorbable synthetic mesh versus none Sand 2001 6/65 7/67 43.34 0.88 [0.31, 2.49] Subtotal (95% CI) 65 67 43.34 0.88 [0.31, 2.49] Total events: 6 (Mesh/graft), 7 (No mesh/graft) Test for heterogeneity: not applicable Test for overall effect: Z = 0.23 (P = 0.81)

02 Absorbable biological graft versus none Paraiso 2006 12/26 10/55 56.66 2.54 [1.26, 5.10] Subtotal (95% CI) 26 55 56.66 2.54 [1.26, 5.10] Total events: 12 (Mesh/graft), 10 (No mesh/graft) Test for heterogeneity: not applicable Test for overall effect: Z = 2.62 (P = 0.009)

Total (95% CI) 91 122 100.00 1.61 [0.57, 4.54] Total events: 18 (Mesh/graft), 17 (No mesh/graft) Test for heterogeneity: Chi² = 2.82, df = 1 (P = 0.09), I² = 64.6% Test for overall effect: Z = 0.89 (P = 0.37)

0.1 0.2 0.5 1 2 5 10 Favours mesh/graft Favours no mesh/graf

Crude event rates (all study types)

When the evidence from other study types was also considered, the objective failure rate was 12.7% (18/142, 95% CI 8.2 to 19.1%) for procedures without mesh/graft, 8.6% (6/70, 95% CI 4.0 to 17.5%) for absorbable synthetic mesh, 20.4% (19/93, 95% CI 13.5 to 29.7%) for absorbable biological graft, and 6.5% (2/31, 95% CI 1.8 to 20.7%) for non-absorbable synthetic mesh (Table 13 B).

A Bayesian meta-analysis adjusting for study design, and indirect comparisons could not be conducted because there was very little evidence available (only two studies of each design). From the evidence from unadjusted Bayesian meta-analysis, there were no significant differences in objective failure between any type of mesh/graft and no mesh/graft and the confidence intervals were very wide, using a random effect at the study level (Table 14).

Table 14 Bayesian meta-analysis models, posterior repair: objective failure, recurrent prolapse at original site Categories n N OR (unadjusted) 95% CrI No mesh 18 142 Reference technique - Absorbable synthetic mesh 6 70 0.59 0.01 to 15.06 Absorbable biological graft 19 93 2.10 0.14 to 40.03 Non-absorbable synthetic mesh 2 31 0.57 0.01 to 43.84

Notes: 1. n = cumulative number of patients experiencing the event, N = cumulative number of patients analysed by the studies. 2. CrI. Credible interval with 95% probability of containing the true OR. 3. It was not possible to provide an OR adjusted for study design due to the small number of studies

4.5.1.3 De novo prolapse (i.e. at other sites)

No studies reported de novo prolapse in women free of prolapse at other sites at baseline.

4.5.1.4 Need for further surgery for prolapse

Only one study, a RCT36 using absorbable biological graft (porcine dermis) reported this outcome (Appendix 15, Table 13 D). All types of prolapse, i.e. recurrent and de novo, were considered for further opeartion. There was no significant differences between using absorbable biological graft and procedures without mesh/graft (3/29, 10.3% vs. 3/70, 4.3%; RR 2.41, 95% CI 0.52 to 11.27).

4.5.1.5 Persistent urinary incontinence

No studies reported persistent urinary incontinence in women having this symptom at baseline.

4.5.1.6 Persistent bowel symptoms

One RCT36 using absorbable biological graft (porcine dermis), one case series80 using absorbable biological graft (fascia lata), and another case series89 using combined mesh/graft (polyglactin-polypropylene), involving 183 women reported this outcome (Appendix 15, Table 13 F). In the RCT36, no significant difference was found between using absorbable biological graft and procedures without mesh/graft during a follow-up of one year (5/29, 17.2% vs. 19/58, 32.8%; RR 0.53, 95% CI 0.22 to 1.27). The crude event rates for persistent bowel symptoms were 32.8% (19/58, 95% CI 22.1 to 45.6%) for procedures without mesh/graft, 17.1% (14/82, 95% CI 10.5 to 26.4%) with absorbable biological graft, and 11.6% (5/43, 95% CI 5.2 to 24.6%) with non-absorbable synthetic mesh.

4.5.1.7 Persistent dyspareunia

Only one case series80 using absorbable biological graft (fascia lata) reported persistent dyspareunia in women having this symptom at baseline. At a mean follow up of 13.7 months (range 6 to 23), dyspareunia was persistent in 35.7% (5/14, 95% CI 16.3 to 61.2%) of women.

4.5.1.8 Improvement of disease-specific quality of life (RCTs and non-randomised comparative studies only)

One RCT36 involving 45 women reported this outcome using a disease-specific questionnaire (PFIQ-7, the Pelvic Floor Impact Questionnaire-7, ranged between 0-300, 300 indicates the greatest adverse impact on quality of life) (see Appendix 13). Disease-specific quality of life after two-year follow-up was not statistically significantly different between procedures using absorbable biological graft (porcine dermis) and those without mesh/graft (RR -11.00, 95% CI -24.15 to 2.15)

4.5.2 Safety

The outcomes from individual studies are summarised, by type of outcomes, in Appendix 16. The crude event rates are summarised in Table 15.

Studies reporting no adverse events occurred

Castelo-Branco and colleagues65 reported that no complications occurred during surgery using absorbable synthetic mesh (polyglactin) (n=10). Sand and colleagues,27 using absorbable synthetic mesh, reported that no adverse effects occurred during the one-year follow up (n=132). Altman and colleagues,64 using absorbable biological graft (porcine dermis), reported that there were no perioperative mortalities, major complications or readmissions within 30 days of surgery (n=32).

Table 15 Posterior repair, safety: summary of crude event rates, by type of mesh/graft, by study design Study No mesh Absorbable synthetic Absorbable biological Combined Non-absorbable synthetic N n N % n N % n N % n N % n N % A. Blood loss (number of women requiring blood transfusion) RCT 1 3 74 4.1% 1 31 3.2% Non-rand 1 0 5 0.0% 0 5 0.0% Case series 2 0 90 0.0% 1 71 1.4% Total 3 79 3.8% 0 5 0.0% 1 31 3.2% 0 90 0.0% 1 71 1.4% 95% CI 1.3% to 10.6% 0.0% to 43.4% 0.6% to 16.2% 0 to 4.1% 0.2% to 7.6% B. Damage to surrounding organs during operation RCT 1 2 74 2.7% 1 31 3.2% Non-rand 1 0 5 0.0% 0 5 0.0% Case series 2 0 90 0.0% 3 71 4.2% Total 2 79 2.5% 0 5 0.0% 1 31 3.2% 0 90 0.0% 3 71 4.2% 95% CI 0.7% to 8.8% 0.0% to 43.4% 0.6% to 16.2% 0 to 4.1% 1.4% to 11.7%

56 C. Mesh/graft erosion RCT 2 0 28 0.0% 5 25 20.0% Non-rand 0 Case series 2 11 90 12.2% 2 31 6.5% Total 0 28 0.0% 16 115 13.9% 2 31 6.5% 95% CI 0.0% to 12.1% 8.7% to 21.4% 1.8% to 20.7% D. Further operation needed for mesh/graft erosion RCT 0 Non-rand 0 Case series 1 11 90 12.2% Total 11 90 12.2% 95% CI 7.0% to 20.6% E. De novo urinary incontinence (no studies reported this outcome) F. De novo bowel symptoms

Study No mesh Absorbable synthetic Absorbable biological Combined Non-absorbable synthetic N n N % n N % n N % n N % n N % RCT 0 Non-rand 0 Case series 2 2 45 4.4% 1 29 3.4% Total 2 45 4.4% 1 29 3.4% 95% CI 1.2% to 14.8% 0.6% to 17.2% G. De novo dyspareunia RCT 0 Non-rand 0 Case series 2 4 25 16.0% 2 36 5.6% Total 4 25 16.0% 2 36 5.6% 95% CI 6.4% to 34.7% 1.5% to 18.1% H. Infection RCT 1 11 74 14.9% 6 31 19.4%

57 Non-rand 2 2 20 10.0% 0 5 0.0% 1 17 5.9% Case series 1 4 106 3.8% Total 13 94 13.8% 0 5 0.0% 7 48 14.6% 4 106 3.8% 95% CI 8.3% to 22.2% 0.0% to 43.4% 7.2% to 27.2% 1.5% to 9.3% I. Other serious adverse effects (inappropriate to calculate crude event rates)

4.5.2.1 Blood loss

Blood loss, ml

One RCT36 using absorbable biological graft and the registry29 using non-absorbable synthetic mesh (polypropylene, commercially available device) involving 176 women reported this outcome (Appendix 16). The mean blood loss in both studies was 200ml or less. In the RCT, the mean blood loss with absorbable biological graft (200ml) was similar to the procedures without mesh/graft (150ml), but no SDs were reported. The study reported that there was no statistically significant difference.

Blood loss requiring blood transfusion

One RCT,36 one non-randomised comparative study,65 the registry,29 and one case series89 involving 276 women reported the number of women requiring blood transfusion (Appendix 16). In the RCT, absorbable biological graft (porcine dermis) was used. There was no significant difference in the need for blood transfusion between using absorbable biological graft and procedures without mesh/graft (1/31, 3.2% vs. 3/74, 4.1%; RR 0.80, 95% CI 0.09 to 7.36).

When the evidence from other study types was also considered, blood transfusion was required by 3.8% (3/79, 95% CI 1.3 to 10.6%) of women treated with procedures without mesh/graft, 0% (0/5, 95% CI 0 to 43.4%) with absorbable synthetic mesh, 3.2% (1/31, 95% CI 0.6 to 16.2%) with absorbable biological graft, 0% (0/90, 95% CI 0 to 4.1%) with combined synthetic mesh, and 1.4% (1/71, 95% CI 0.2 to 1.4%) with non-absorbable synthetic mesh (Table 15 A).

4.5.2.2 Damage to surrounding organs during operation

One RCT,36 one non-randomised comparative study,65 the registry29 and one case series89 involving 276 women reported this outcome (Appendix 16).

In the RCT, absorbable biological graft (porcine dermis) was used. There was no significant difference in damage to bladder, bowel or urethra between using graft and procedures without mesh/graft (1/31, 3.2% vs. 2/74, 2.7%; RR 1.19, 95% CI 0.11 to 12.69). When evidence from other study types was also considered, damage to surrounding organs occurred in 2.5% (2/79, 95% CI 0.7 to 8.8%) of women treated with procedures without

mesh/graft, 0% (0/5, 95% CI 0 to 43.4%) with absorbable synthetic mesh, 3.2% (1/31, 95% CI 0.6 to 16.2%) with absorbable biological graft, 0% (0/90, 95% CI 0 to 4.1%) with combined synthetic mesh, and 4.2% (3/71, 95% CI 1.4 to 11.7%) with non-absorbable synthetic mesh (Table 15 B). The regitry was the only study that used commercially available insertion kits. Rectal perforation was the reported complication with non- absorbable synthetic mesh (polypropylene, all commercially available devices).

4.5.2.3 Mesh/graft erosion

Two RCTs36,55 and two case series28,89 involving 174 women reported this outcome (Appendix 16). The mean follow-up time ranged from 6 to 18 months. When evidence from all study types was considered, mesh/graft erosion was 0% (0/28, 95% CI 0 to 12.1%) for absorbable biological graft, 13.9% (16/115, 95% CI 8.7 to 21.4%) for combined synthetic mesh, and 6.5% (2/31, 95% CI 1.8 to 20.7%) for non-absorbable synthetic mesh (Table 15 C).

Lim and colleagues89 reported 7 of the 11 minor mesh protrusions occurred between 6 and 12 weeks.

4.5.2.4 Further operation needed for mesh/graft erosion

In one case series, Lim and colleagues89 reported that all 11 women who had minor mesh protrusions after using combined synthetic mesh (polypropylene and polyglactin) were treated by surgery (Appendix 16, Table 15 D).

4.5.2.5 De novo urinary incontinence

No studies reported this outcome.

4.5.2.6 De novo bowel symptoms

Two case series28,89 reported this outcome (Appendix 16). One study89 used combined synthetic mesh (polypropylene and polyglactin). During 6 to 12 week follow up, 2 of 45 (4.4%, 95% CI 1.2 to 14.8%, Table 15 F) women free of constipation at baseline had de novo constipation after surgery and 1 of 35 (2.9%, 95% CI 0.5 to 14.5%) women free of defaecation difficulties developed de novo defaecation difficulties. In the other case series,28 during a mean follow-up of 17 months, 3.4% (1/29, 95% CI 0.6 to 17.2%, Table 15 F) of

women had de novo faecal incontinence after surgery with non-absorbable synthetic mesh (polypropylene).

4.5.2.7 De novo dyspareunia

Two case series80,89 involving 61 women free of dyspareunia at baseline reported this outcome (see Appendix 16). De novo dyspareunia occurred in 16.0% (4/25, 95% CI 6.4 to 34.7%) of women treated with absorbable biological graft (solvent-dried fascia lata) during a mean follow up of 14 months,80 and occurred at 5.6% (2/36, 95% CI 1.5 to 18.1%) with combined synthetic mesh (polypropylene and polyglactin) during a follow up of 6 months89 (Table 15 G). Neither study reported the timing of the onset of dyspareunia or duration.

4.5.2.8 Infection

One RCT,36 two non-randomised comparative studies64,65 and the registry29 involving 253 women reported this outcome (see Appendix 16). In the RCT,36 19.4% (6/31) women treated with absorbable biological graft and 14.9% (11/74) women treated with procedure without mesh/graft had pelvic abscess (RR 1.30, 95% CI 0.53 to 3.21). The trialists did not report the timing of the events. When evidence from other study types was also considered, infection occurred in 13.8% (13/94, 95% CI 8.3 to 22.2%) of women treated with procedures without mesh/graft, 0% (0/5, 95% CI 0 to 43.4%) with absorbable synthetic mesh, 14.6% (7/48, 95% CI 7.2 to 27.2%) with absorbable biological graft, and 3.8% (4/106, 95% CI 1.5 to 9.3%) with non-absorbable synthetic mesh (Table 15 H).

4.5.2.9 Other serious adverse effects

One RCT36 using absorbable biological graft (porcine dermis) and the registry29 using non- absorbable synthetic mesh (polypropylene, commercially available device) involving 176 women reported other serious adverse effects (Appendix 16). In the RCT, one woman (1/31) returned to theatre, one (1/31) had a pulmonary-related event, and four women (4/31) required emergency room evaluation; in the group without mesh/graft, one woman (1/74) had a myocardial infarction, one (1/74) had a pulmonary-related event, three women (3/74) were readmitted, and 11 cases (11/74) required emergency room evaluations. No further details or outcomes for these events were reported. Direct comparison (RevMan meta- analysis) could not be conducted because some events might have occurred in the same women, and this also made calculating the crude event rates difficult. In the registry, Altman29 reported that no women had anaemia.

4.5.3 Other outcomes (operation time and surgeon experience)

The outcomes from individual studies are summarised in Appendix 17.

4.5.3.1 Operation time

One RCT36 using absorbable biological graft (porcine dermis) and the registry29 using non- absorbable synthetic mesh (polypropylene, commercially available device) involving 176 women reported this outcome. There was no statistically significant difference in operation time between using absorbable biological graft and procedures without mesh/graft in the RCT (168min vs. 150min; 95% CI -7.77 to 45.77). The mean operation time was 57.6 minutes (SD 19.6) in the registry.29

4.5.3.2 Surgeon experience

Only the registry29 report reported information on surgeon experience. All surgeons were experienced in pelvic floor surgery and had received pretrial hands-on training in the use of a commercially available device.

4.5.4 Conclusions of the evidence for and against the use of mesh/graft in posterior vaginal wall prolapse repair

Only nine studies reported data on posterior vaginal wall prolapse repair. The median follow- up time was 12 months (range 1 to 17). The crude event rates in Table 13 (efficacy) are presented graphically in Figure 11. The crude event rates in Table 15 (safety) are presented graphically in Figure 12. In the studies which reported data on duration of surgery, the operation times were all less than 170 minutes. Only the registry report reported surgeon experience.

Figure 11 Efficacy of posterior repair, summary of crude event rates (95% CI, any study design)

No mesh/graft Absorbable synthetic mesh Absorbable biological graft Combined mesh/graft Non-absorbable synthetic mesh

Figure 12 Safety of posterior repair, summary of crude event rates (95% CI, any study design)

50 3/79, 0/5, 1/31, 0/90, 1/71 2/79, 0/5, 1/31, 0/90, 3/71 13/94, 0/5, 7/48, NA, 40 NA, 4/25, NA, 2/36, NA )

30 NA, NA, 0/28, 16/115, 2/31 NA, NA, NA, 11/90, 20 NA, NA, NA, 2/45, 1/29

Crude rate (% 10

No mesh/graft Absorbable synthetic mesh Absorbable biological graft Combined mesh/graft Non-absorbable synthetic mesh

Balance sheet: evidence for and against the use of mesh/graft comparing with no mesh/graft in posterior vaginal wall prolapse repair Absorbable synthetic Absorbable biological graft Combined mesh/graft Non-absorbable synthetic N. women N. women N. women N. women For Unclear Against For Unclear Against For Unclear Against For Unclear Against (N. study) (N. study) (N. study) (N. study) Efficacy Subjective failure rate √ 0 √ 138 (2) √ 0 √ 0 Objective failure rate √ 142 (2) √ 163 (3) √ 0 √ 31 (1) De novo prolapse (i.e. at other sites) √ 0 √ 0 √ 0 √ 0 Further operation needed for prolapse √ 0 √ 99 (1) √ 0 √ 0 Persistent urinary symptoms √ 0 √ 0 √ 0 √ 0 Persistent bowel symptoms √ 0 √ 140 (2) √ 43 (1) √ 0 Persistent dyspareunia √ 0 √ 14 (1) √ 0 √ 0 Improvement of disease-specific QoL √ 0 √ 45 (1) √ 0 √ 0 Safety Blood loss √ 10 (1) √ 105 (1) √ 90 (1) √ 71 (1) Damage to surrounding organs √ 10 (1) √ 105 (1) √ 90 (1) √ 71 (1) 63 Mesh/graft erosion √ 0 √ 28 (1) √ 115 (2) √ 31 (1) Further operation needed for mesh/graft erosion √ 0 √ 0 √ 90 (1) √ 0 De novo urinary incontinence √ 0 √ 0 √ 0 √ 0 De novo bowel symptoms √ 0 √ 0 √ 45 (1) √ 29 (1) De novo dyspareunia √ 0 √ 25 (1) √ 36 (1) √ 0 Infection √ 10 (1) √ 168 (2) √ 0 √ 106 (1) Other serious adverse effects √ 0 √ 105 (1) √ 0 √ 71 (1) Other outcomes Operation time √ 0 √ 31 (1) √ 0 √ 71 (1) Surgeon experience √ 0 √ 0 √ 0 √ 71 (1)

Note:

4.6 Anterior and/or posterior repair

Fourteen studies involving 1680 women treated with mesh/graft reported data on the use of mesh in anterior and/or posterior repair (3 RCTs, one non-randomised comparative study, one registry report, and 9 case series, see Table 7). None of the RCTs compared different types of mesh or grafts. The median follow-up time was 13 months (range 1 to 51).

One study56 used absorbable synthetic mesh, none of the studies used absorbable biological graft, one study83 used a combined mesh/graft, 10 studies29,58,59,81,82,84,85,91,92,93 used non- absorbable synthetic mesh, and two studies66,90 used more than one of the above types of mesh/graft. One study34 only included women with primary prolapse (n 71), two studies29,93 included both primary (n 130) and secondary (n 51) prolapse but did not report the results separately, and the other 11 studies did not report whether the prolapses treated were primary or secondary.

4.6.1 Efficacy

In this section, studies are included which presented data from women having anterior or posterior repair or both, but which failed to report separately outcomes for the different prolapse sites. The outcomes from individual studies are summarised, by type of outcomes, in Appendix 18. The crude event rates are summarised in Table 16.

Table 16 Anterior and/or posterior repair, efficacy: summary of crude event rates, by type of mesh/graft, by study design Study No mesh Absorbable synthetic Combined Non-absorbable synthetic N n N % n N % n N % n N % A. Subjective failure: persistent prolapse symptoms RCT 1 14 34 41.2% 14 32 43.8% Non-rand 0 Case series 2 0 148 0.0% Total 14 34 41.2% 14 32 43.8% 0 148 0.0% 95% CI 26.4% to 57.8% 28.2% to 60.7% 0.0% to 2.5% B. Objective failure: recurrent prolapse at original site RCT 3 27 109 24.8% 2 26 7.7% 18 83 21.7% Non-rand 0 Case series 4 11 143 7.7% 23 562 4.1% Total 27 109 24.8% 2 26 7.7% 11 143 7.7% 41 645 6.4% 95% CI 17.6% to 33.6% 2.1% to 24.1% 4.3% to 13.2% 4.7% to 8.5% C. De novo prolapse at other sites (no studies reported data on this outcome) D. Further operation needed for prolapse RCT 0 Non-rand 0 Case series 2 7 161 4.3% Total 7 161 4.3% 95% CI 2.1% to 8.7% E. Persistent urinary incontinence (any types) RCT 0 Non-rand 0 Case series 4 46 203 22.7% Total 46 203 22.7% 95% CI 17.4% to 28.9% F. Persistent bowel symptoms RCT 0 Non-rand 0 Case series 1 1 21 4.8% Total 1 21 4.8% 95% CI 0.8% to 22.7% G. Persistent dyspareunia RCT 0 Non-rand 0 Case series 1 1 10 10.0% Total 1 10 10.0% 95% CI 1.8% to 40.4%

4.6.1.1 Subjective failure: persistent prolapse symptoms

One RCT56 and two case series81,82 involving 214 women reported this outcome (see Appendix 18). The RCT used absorbable synthetic mesh (polyglactin) with a follow-up of 6 months. There was no significant difference in subjective failure rates between using absorbable synthetic mesh and procedures without mesh/graft (14/32, 43.8% vs. 14/34, 41.2%; RR 1.11, 95% CI 0.42 to 2.95%). The two case series both used non-absorbable synthetic mesh (polypropylene) and no cases (0/148) of subjective failure occurred (Table 16 A). The mean follow up time in one case series81 was 30.7 months (SD 1.7) and 51 months (range 12 to 115) in the other.82

No studies reported time to failure.

4.6.1.2 Objective failure: recurrent prolapse at original site

Three RCTs34,56,58 and five case series81-85 involving 923 women reported this outcome (Appendix 18).

Direct comparison (mesh/graft vs. no mesh/graft)

One RCT56 compared objective failure of procedures using absorbable synthetic mesh with those without mesh/graft after 3-month follow up. No significant difference was found (2/26, 7.7%, 4/32, 12.5%; RR 0.62, 95% CI 0.12 to 3.10, Figure 13). Two RCTs34,58 used non- absorbable synthetic mesh (polypropylene), with one34 using a commercially available device. The results favoured no mesh in the study using the commercially available device (7/21, 33.3% vs. 3/17, 17.7%; RR 1.89, 95% CI 0.57 to 6.22) after 3-month follow-up and favoured mesh in the other study (11/62, 17.7% vs. 20/60, 33.3%; RR 0.53, 95% CI 0.28 to 1.01) after one year follow up but neither was statistically significant and the heterogeneity was high (I2 70.3%).

Figure 13 Anterior and/or posterior repair, efficacy, objective failure, mesh or graft versus procedures without mesh/graft: evidence from RCTs

Review: Efficacy and safety of using mesh or grafts in surgery for vaginal wall prolapse (RCTs) Comparison: 04 Mesh or graft versus none for anterior and/or posterior vaginal wall prolapse Outcome: 02 Objective failure: recurrent prolapse at original site

Study Mesh/graft No mesh/graft RR (random) Weight RR (random) or sub-category n/N n/N 95% CI % 95% CI

01 Absorbable synthetic mesh versus none Alahdin 2006 abs. 2/26 4/32 18.94 0.62 [0.12, 3.10] Subtotal (95% CI) 26 32 18.94 0.62 [0.12, 3.10] Total events: 2 (Mesh/graft), 4 (No mesh/graft) Test for heterogeneity: not applicable Test for overall effect: Z = 0.59 (P = 0.56)

02 Non-absorbable synthetic mesh versus none Lim 2007 abs. 11/62 20/60 52.12 0.53 [0.28, 1.01] Meschia 2007 abs. 7/21 3/17 28.94 1.89 [0.57, 6.22] Subtotal (95% CI) 83 77 81.06 0.90 [0.27, 3.08] Total events: 18 (Mesh/graft), 23 (No mesh/graft) Test for heterogeneity: Chi² = 3.36, df = 1 (P = 0.07), I² = 70.3% Test for overall effect: Z = 0.16 (P = 0.87)

Total (95% CI) 109 109 100.00 0.79 [0.35, 1.77] Total events: 20 (Mesh/graft), 27 (No mesh/graft) Test for heterogeneity: Chi² = 3.39, df = 2 (P = 0.18), I² = 40.9% Test for overall effect: Z = 0.57 (P = 0.57)

0.1 0.2 0.5 1 2 5 10 Favours mesh/graft Favours no mesh/graf

Crude event rates (all study types)

When evidence from other study types was also considered, the objective failure rate was 24.8% (27/109, 95% CI 17.6 to 33.6%) for procedures without mesh/graft, 7.7% (2/26, 95% CI 2.1 to 24.1%) for absorbable synthetic mesh, 7.7% (11/143, 4.3 to 13.2%) with combined mesh/graft, and 6.4% (41/645, 4.7 to 8.5%) for non-absorbable synthetic mesh (Table 16 B). The mean follow-up time for the case series ranged from 13 to 51 months.

There were insufficient data (only three RCTs) to conduct Bayesian meta-analysis. Therefore we could draw no conclusions about which mesh option was most effective.

Time to failure

Rozet and colleagues85 reported that 7 of 13 failures occurred within 6 months. No other studies reported time to failure.

Evidence from another study

Vakilli and colleagues66 reported development of prolapse in any compartment of any degree during follow-up (mean 9 months, range 3-67), but did not specify which compartment, i.e. whether the prolapse occurred at the original site or a de novo site. 34.7% (34/98) occurred in the mesh/graft group and 42.5% (91/214) occurred in the no-mesh group.

4.6.1.3 De novo prolapse (i.e. at other sites)

No studies reported de novo prolapse in women without prolapse at other sites at baseline.

4.6.1.4 Further operation needed for prolapse

Two case series82,84 using non-absorbable synthetic mesh (polypropylene, Table 16 D) and one non-randomised comparative study66 using more than one type of mesh (absorbable biological graft or non-absorbable synthetic mesh) involving 473 women reported this outcome (see Appendix 18). One case series82 reported that no women needed an operation for recurrent (same-site) prolapse (0/64). The other84 reported that 7.2% (7/97) women needed a new operation for de novo prolapse. In the non-randomised comparative study, 8.2% (8/98) of women treated with mesh/graft needed a further operation for prolapse (unclear whether recurrent or de novo) and 9.4% (20/214) of women treated with procedures without mesh/graft needed a further operation.

4.6.1.5 Persistent urinary incontinence

Four case series82,84,92,93, all using non-absorbable synthetic mesh, involving 303 women reported persistent urinary incontinence. The mean follow up ranged from 25 weeks to 51 months. The crude event rate for persistent urinary incontinence was 22.7% (46/203, 95% CI 17.4 to 28.9%, Table 16 E).

4.6.1.6 Persistent bowel symptoms

Only one case series84 reported persistent bowel symptoms in women having the symptoms at baseline. Non-absorbable synthetic mesh was used. After a mean 29-month follow up (range 6 to 52), bowel symptoms were persistent in 4.8% (1/21, 95% CI 0.8 to 22.7%, Table 16 F) of women.

4.6.1.7 Persistent dyspareunia

Only one case series83 with a mean follow up of 13 months (range 10 to 19) reported persistent dyspareunia in women having the symptoms at baseline. Combined mesh/graft (polypropylene coated with atelocollagen) was used. Dyspareunia was persistent in 10.0% (1/10, 95% CI 1.8 to 40.4%) of women.

4.6.1.8 Improvement of disease-specific quality of life (RCTs and non-randomised comparative studies only)

Only one study, a RCT,56 involving 60 women reported this outcome (see Appendix 18). There was no significant difference in disease-specific quality of life after 6-month follow up between using absorbable synthetic mesh (polyglactin) and procedures without mesh/graft (RR -0.10, 95% CI -1.44 to 1.24).

4.6.2 Safety

The outcomes from individual studies are summarised, by type of outcomes, in Appendix 19. The crude event rates are summarised in Table 17.

All studies in this section used either combined mesh/graft or non-absorbable synthetic mesh.

Table 17 Anterior and/or posterior repair, safety: summary of crude event rates, by type of mesh/graft, by study design Study No mesh Combined Non-absorbable synthetic N n N % n N % n N % A. Blood loss (number of women requiring blood transfusion) RCT 1 1 35 2.9% 0 36 0.0% Non-rand 0 Case series 7 0 0 0.0% 11 774 1.4% Total 1 35 2.9% 11 810 1.4% 95% CI 0.5% to 14.5% 0.8% to 2.4% B. Damage to surrounding organs during operation RCT 0 Non-rand 0 Case series 6 4 143 2.8% 12 541 2.2% Total 4 143 2.8% 12 541 2.2% 95% CI 1.1% to 7.0% 1.3% to 3.8% C. Mesh/graft erosion RCT 2 5 83 6.0% Non-rand 0 Case series 8 9 143 6.3% 57 1036 5.5% Total 9 143 6.3% 62 1119 5.5% 95% CI 3.3% to 11.5% 4.3% to 7.0% D. Further operation needed for mesh/graft erosion RCT 1 3 62 4.8% Non-rand 0 Case series 8 6 143 4.2% 42 1036 4.1% Total 6 143 4.2% 45 1098 4.1% 95% CI 1.9% to 8.9% 3.1% to 5.4% E. De novo urinary incontinence RCT 0 Non-rand 0 Case series 5 34 355 9.6% Total 34 355 9.6% 95% CI 6.9% to 13.1% F. De novo bowel symptoms RCT 1 1 47 2.1% Non-rand 0 Case series 0 Total 1 47 2.1% 95% CI 0.4% to 11.1% G. De novo dyspareunia RCT 0 Non-rand 0 Case series 2 10 78 12.8% 3 42 7.1% Total 10 78 12.8% 3 42 7.1%

Study No mesh Combined Non-absorbable synthetic N n N % n N % n N % 95% CI 7.1% to 22.0% 2.5% to 19.0% H. Infection RCT 0 Non-rand 0 Case series 6 33 661 5.0% Total 33 661 5.0% 95% CI 3.6% to 6.9% I. Other serious adverse events RCT 0 Non-rand 0 Case series 4 3 278 1.1% Total 3 278 1.1% 95% CI 0.4% to 3.1%

4.6.2.1 Blood loss

Blood loss, ml

The registry29 and three case series82,85,93 involving 578 women reported this outcome (Appendix 19). All studies used non-absorbable synthetic mesh (polypropylene or polyester covered with silicone). Two studies29,93 used commercially available devices. All studies reported a mean blood loss less than 350 ml.

Blood loss requiring blood transfusion

One RCT,34 the registry,29 and five case series82,84,85,92,93 involving 845 women reported this outcome (Appendix 19). All used non-absorbable synthetic mesh. In the RCT, the difference in blood transfusion rate between mesh and no mesh groups was not significantly different (0/36, 0% vs. 1/35, 2.9%; RR 0.32, 95% CI 0.01 to 14.5%). When the evidence from other study types was also considered, 2.9% (1/35, 95% CI 0.5 to 14.5%) of women required blood transfusion after procedures without mesh/graft and 1.4% (11/810, 95% CI 0.8 to 2.4%) when non-absorbable synthetic mesh was used (Table 17 A).

4.6.2.2 Damage to surrounding organs during operation

The registry29 and four case series82,83,92,93 involving 684 women reported this outcome (Appendix 19). The damage included bladder, rectal, and urethral injury. Damage occurred in 2.8% (4/143, 955 CI 1.1 to 7.0%) of women treated with combined mesh/graft (polypropylene coated with atelocollagen) and in 2.2% (12/451, 95% CI 1.3 to 3.8%) of women treated with non-absorbable synthetic mesh (Table 17 B).

The registry and one case series93 used a commercially available insertion device. There was insufficient evidence to compare commercially available insertion devices with procedures that did not use such devices.

4.6.2.3 Mesh/graft erosion

Outcomes reported as mesh/graft erosion, mesh rejection, and mesh exposure were considered in this section.

Two RCTs,34,58 one non-randomised comparative study66 and nine case series81-85,90-93 involving 1558 women reported this outcome (Appendix 19). The mean follow-up time ranged from 2 to 41 months. In total, mesh/graft erosion occurred in 6.3% (9/143, 95% CI 3.3 to 11.5%) of women treated with combined mesh/graft and 5.5% (62/1119, 95% CI 4.3 to 7.0%) of women treated with non-absorbable synthetic mesh (Table 17 C).

Two studies66,90 used more than one type of mesh/graft. The erosion rate was 25.5% (25/98) using absorbable biological graft or non-absorbable synthetic mesh after a mean follow-up time of 9 months.66 The erosion rate (any visible mesh exposure) was 6.6% (13/198) using combined mesh/graft or non-absorbable synthetic mesh during 6-weeks to 6-month follow up.

Time that erosion occurred

Amrute and colleagues81 reported that the two erosions (2/76) occurred at 7 months and 12 months respectively.

Achtari and colleagues90 reported that 5 of 13 mesh exposures occurred within 6 weeks, and the other 8 occurred between 6 weeks and 6 months (combined or non-absorbable synthetic mesh used). De Tayrac and colleagues83 reported that all of the 9 erosions occurred within 3 months (combined mesh/graft used). Dwyer and colleagues84 reported that 7 of 9 erosions occurred within 6 months (non-absorbable synthetic mesh used).

4.6.2.4 Further operation needed for mesh/graft erosion

One RCT58 and eight case series81-85,91-93 involving 1241 women reported this outcome (see Appendix 19). In total, 4.2% (6/143, 95% CI 1.9 to 8.9%) of women treated with combined mesh/graft) needed a further operation compared with 4.1% (45/1098, 95% CI 3.1 to 5.4%) of women treated with non-absorbable synthetic mesh (see Table 17 D).

Conservative treatment for mesh/graft erosion

Collinet and colleagues91 reported that 9/34 mesh/graft erosions were completely healed by antiseptics and antibiotics. Fatton and colleagues93 reported that 2 of 5 mesh/graft erosions were successfully treated by vaginal hormone therapy. Vakili and colleagues66 reported that all 25 mesh/graft erosions were successfully treated by debridement or topical oestrogen alone.

4.6.2.5 De novo urinary incontinence

Five case series81,82,84,85,93 involving 355 women free of urinary incontinence at baseline reported this outcome (Appendix 19). All used non-absorbable synthetic mesh and the follow-up time ranged from 6 to 51 months. 9.6% (34/355, 95% CI 6.9 to 13.1%) of women in total developed de novo urinary incontinence (Table 17 E).

4.6.2.6 De novo bowel symptoms

Only one study, a case series84 reported this outcome (Table 17 F). One of 47 women free of bowel symptoms at baseline developed de novo constipation after a mean of 29 months (range 6 to 52) follow up. Non-absorbable synthetic mesh (polypropylene) was used.

4.6.2.7 De novo dyspareunia

Two case series83,84 involving 120 women free of dyspareunia at baseline reported this otucome (Appendix 19). 12.8% (10/78, 95% CI 7.1 to 22.0%) of women treated with combined mesh/graft developed de novo dyspareunia after a mean follow up of 13 months (range 10 to 19) compared with 7.1% (3/42, 95% CI 2.5 to 19.0%) of women treated with non-absorbable synthetic mesh after a mean follow up of 29 months (range 6 to 52) (Table 17 G).

4.6.2.8 Infection

One non-randomised comparative study,66 the registry29 and four case series82,85,92,93 involving 973 women reported this outcome (see Appendix 19). The non-randomised comparative study66 used more than one type of mesh/graft (absorbable biological graft or non-absorbable synthetic mesh). 18.4% (18/98) of women treated with mesh/graft had an infection (vaginal/graft infection) during a mean follow up of 9 months (range 3 to 67) compared with 4.7% (10/214) of those treated without mesh/graft. For non-absorbable synthetic mesh, when evidence from all study types was considered, 5.0% (33/661, 95% CI 3.6 to 6.9%) of women had an infection (reported as urinary tract infection, chronic inflammation of the cervix, postoperative spondylitis) (see Table 17 H). When the infection occurred was not reported.

4.6.2.9 Other serious adverse effects

The registry29 and two case series84,93 involving 278 women reported other serious adverse effects (see Appendix 19). All used non-absorbable synthetic mesh. One (1/97) rectovaginal fistula occurred at the fifth day after operation.84 Two (2/110) haematomas requiring secondary surgical treatment were reported in the other case series.93 In total, other serious adverse effects occurred in 1.1% (3/278, 95% CI 0.4 to 3.1%, Table 17 I).

4.6.3 Other outcomes (operation time and surgeon experience)

The outcomes from individual studies are summarised by type of outcomes in Appendix 20.

4.6.3.1 Operation time

The registry29 and four case series82,85,92,93 involving 661 women reported this outcome. All used non-absorbable synthetic mesh, with the registry29 (n 71, mean 86 minutes) and one case series93 (n 111, mean 90 minutes) using a commercially available device. The mean operation time in studies that did not use such a device ranged between 97 to 276 minutes.

4.6.3.2 Surgeon experience

The registry29 report and a case series83 reported information on surgeon experience. In the registry,29 all surgeons were experienced in pelvic floor surgery and had received pretrial hands-on training on the use of a commercially available device. In the case series,83 all 18 surgeons were experienced in POP repair.

4.6.4 Conclusions of the evidence for and against the use of mesh/graft in anterior and/or posterior vaginal wall prolapse repair

Fourteen studies reported anterior and/or posterior vaginal wall prolapse repair. For most outcomes, there was only one study available. The median follow-up time was 13 months (range 1 to 51). Figure 14 presents the crude event rates in Table 16 (efficacy) graphically. Figure 15 presents the crude event rates in Table 17 (safety) graphically. Operation times were all less than 120 minutes (n=661).

Figure 14 Efficacy of anterior and/or posterior repair, summary of crude event rates (95% CI, any study design), by type of mesh/graft

70 14/34, 14/32, NA, 0/148 60 50 NA, NA, 1/10, NA 40 27/109, 2/26, 11/143, 41/645 NA, NA, NA, 46/203 NA, NA, NA, 1/21 30

20 NA, NA, NA, 7/161

No mesh/graft Absorbable synthetic mesh Combined mes h/gr af t Non-absorbable synthetic mesh

Figure 15 Safety of Anterior and/or posterior repair, summary of crude event rates (95% CI, any study design), by type of mesh/graft

25 NA, 10/78, 3/42 20

15 1/35, NA, 11/810 NA, 9/143, 62/1119 NA, NA, 34/355 NA, NA, 1/47

No mesh/graft Combined mesh/graft Non-absorbable synthetic mesh

Balance sheet: evidence for and against the use of mesh/graft comparing with no mesh/graft in anterior and/or posterior vaginal wall prolapse repair Absorbable synthetic Absorbable biological graft Combined mesh/graft Non-absorbable synthetic N. women N. women N. women N. women For Unclear Against For Unclear Against For Unclear Against For Unclear Against (N. study) (N. study) (N. study) (N. study) Efficacy Subjective failure rate √ 66 (1) √ 0 √ 0 √ 148 (2) Objective failure rate √ 135 (2) √ 0 √ 143 (1) √ 645 (5) De novo prolapse (i.e. at other sites) √ 0 √ 0 √ 0 √ 0 Further operation needed for prolapse √ 0 √ 0 √ 0 √ 161 (2) Persistent urinary symptoms √ 0 √ 0 √ 0 √ 203 (4) Persistent bowel symptoms √ 0 √ 0 √ 0 √ 21 (1) Persistent dyspareunia √ 0 √ 0 √ 10 (1) √ 0 Improvement of disease-specific QoL √ 60 (1) √ 0 √ 0 √ 0 Safety Blood loss √ 0 √ 0 √ 0 √ 845 (7) Damage to surrounding organs √ 0 √ 0 √ 143 (1) √ 541 (4) Mesh/graft erosion √ 0 √ 0 √ 143 (1) √ 1119 (9) 77 Further operation needed for mesh/graft erosion √ 0 √ 0 √ 143 (1) √ 1098 (8) De novo urinary incontinence √ 0 √ 0 √ 0 √ 355 (5) De novo bowel symptoms √ 0 √ 0 √ 0 √ 47 (1) De novo dyspareunia √ 0 √ 0 √ 78 (1) √ 42 (1) Infection √ 0 √ 0 √ 0 √ 661 (5) Other serious adverse effects √ 0 √ 0 √ 0 √ 278 (3) Other outcomes Operation time √ 0 √ 0 √ 0 √ 661 (6) Surgeon experience √ 0 √ 0 √ 0 √ 71 (1)

Note:

5 DISCUSSION

This systematic review was prompted by concerns regarding the efficacy and safety of mesh/graft in surgery for anterior and/or posterior prolapse repair compared with procedures without mesh/graft. As more than one type of mesh/graft is available, we also sought to compare the efficacy and safety between different types of mesh/graft.

5.1 Discussion of main results

In this review, we identified 49 studies (6 full-text RCTs, 11 RCTs available as conference abstracts, 7 non-randomised comparative studies, one prospective registry report, and 24 case series) involving 4569 women treated with mesh/graft in total.

5.1.1 Anterior vaginal wall prolapse repair

Thirty studies involving 2472 women provided data on the use of mesh/graft in anterior repair (12 RCTs, 4 non-randomised comparative studies, 1 registry, and 13 case series, see Table 7). Four studies used absorbable synthetic mesh, 14 studies used absorbable biological graft, one study used combined mesh/graft, and 14 studies used non-absorbable synthetic mesh. The median follow-up time was 14 months (range 1 to 38). Of the 12 RCTs, 7 were reported only as conference abstracts. Two of the RCTs compared different types of mesh/graft and the others compared mesh/graft with no mesh/graft.

There were too few data for most of the outcomes, including the primary outcome, cure of symptoms, to draw reliable conclusions. However, there was some evidence that any mesh was better than no mesh for preventing objective recurrence of anterior repair (RR 0.48, 95% CI 0.32 to 0.72). Furthermore, there was corroboratory evidence from the indirect comparisons of different mesh types and different study designs that the efficacy increased with increasing mesh strength and durability, from absorbable synthetic (whose strength is reduced to 25% after 21 days, Appendix 6) to absorbable biological (which theoretically remains as an inert collagen matrix resistant to breakdown, Appendix 6) to non-absorbable synthetic (which is theoretically permanent). The evidence from the indirect comparisons indicated that non-absorbable synthetic mesh was statistically significantly less likely to fail than no mesh/graft or any other type of mesh/graft.

This trend appeared to be supported by the finding of an increased need for re-operation for recurrent or new prolapse in women treated with absorbable synthetic mesh (9.2%) versus

3.2% with biological graft versus 1.3% in women in whom non-absorbable synthetic mesh was used (Table 10 D). Although the re-operation rate for women undergoing surgery without mesh/graft appeared to be lower, at 2.4%, this estimate was based on a small number of trials with few participants, short follow up time and poor quality study designs and as such should be interpreted with caution. Long term (12 years) reoperation rates for women undergoing prolapse surgery without mesh have been reported as 30% (based on a retrospective cohort study of 400 women derived from a population of 150,000 women aged 20 or older).3

In terms of safety, compared with surgery without mesh/graft, there was not enough evidence to identify or rule out important adverse effects related to the use of mesh including blood loss, damage to surrounding organs, infection, new urinary or bowel symptoms or new dyspareunia (Figure 8). This is because the studies were not large enough to identify rare safety effects.

However, there was some evidence to support a trend that more mesh/graft erosions were associated with increasing mesh strength. The rate of mesh/graft erosion increased from 0.7% (1/147, absorbable synthetic) to 6.0% (35/581, absorbable biological) to 10.2% (68/666, non-absorbable synthetic, Table 12 C). Women having a non-absorbable synthetic mesh were also most likely to require an operation to remove it completely or partially because of mesh/graft erosion (6.6%, 23/347, versus 2.9%, 1/35, and 2.6%, 4/154, for the other two types respectively, Table 10 D). No statistical analysis could be conducted to compare mesh/graft erosions or operation for mesh/graft erosions between different types of mesh because of the small number of women in each group and relatively rare events. The reported occurrence of other adverse effects was less than 5%.

5.1.2 Posterior vaginal wall prolapse repair

Only 9 studies involving 417 women treated with mesh/graft reported data on the use of mesh in posterior repair (3 RCTs, 2 non-randomised comparative studies, one registry report, and 3 case series). Three studies used absorbable synthetic mesh, three studies used absorbable biological graft, two studies used combined mesh/graft, and two studies used non-absorbable synthetic mesh. We did not find any RCTs which compared different types of mesh/graft for posterior repair. The median follow-up time was 12 months (range 1 to 17).

In terms of efficacy, there were too few data reported for any of the efficacy outcomes (subjective and objective failure rate; de novo prolapse; further operation for recurrent or de novo prolapse; persistent urinary or bowel problems; and sexual (dyspareunia) problems to draw reliable conclusions or to carry out further statistical analyses (Table 13).

For safety, there were too few data on safety outcomes (blood loss; damage to surrounding organs; mesh erosion; further operations for mesh erosion; new urinary or bowel symptoms; new dyspareunia; infection) to identify or rule out important adverse effects related to the use of mesh/graft either because the studies were not sufficiently large or the adverse effects were rare (Table 15).

5.1.3 Anterior and/or posterior vaginal wall prolapse repair

Fourteen studies involving 1680 women treated with mesh/graft reported data on the use of mesh in anterior and/or posterior repair (3 RCTs, one non-randomised comparative study, one registry report, and 9 case series). One study used absorbable synthetic mesh, none of the studies used absorbable biological graft, one study used a combined mesh/graft, 10 studies used non-absorbable synthetic mesh, and two studies used more than one of the above types of mesh/graft. No RCTs compared different types of mesh or grafts. The median follow-up time was 13 months (range 1 to 51).

For efficacy, there were too few data reported for any of the efficacy outcomes (subjective and objective failure rate; de novo prolapse; further operation for recurrent or de novo prolapse; persistent urinary or bowel problems; and sexual (dyspareunia) problems) to draw reliable conclusions or to carry out further statistical analyses (Table 16).

For objective failure, there was a trend in the crude objective failure rates with procedures not using mesh/graft having the highest failure rate (25%, 95% CI 18 to 34% [27/109], followed by procedures with absorbable synthetic mesh (8%, 2 to 24 [2/26], combined mesh/graft (8%, 4 to 13% [11/143], and non-absorbable synthetic mesh (6%, 5 to 9% [41/645] (Table 16). Bayesian meta-analysis could not be conducted due to lack of data. Therefore, no firm conclusions could be drawn about the effectiveness of any of the mesh/graft types compared to no mesh for anterior and/or posterior repair.

For safety, there were too few data on safety outcomes (blood loss; damage to surrounding organs; mesh erosion; further operations for mesh erosion; new urinary or bowel symptoms; new dyspareunia; or infection) to identify or rule out important adverse effects related to the

use of mesh/graft either because the studies were not sufficiently large or the adverse effects were rare (Table 17).

5.1.4 Generic issues related to mesh

Mesh/graft erosion

The clinical importance of mesh/graft erosion was difficult to assess. The diagnosis was both problematic as different authors used different definitions (mesh erosion, vaginal mesh extrusion, minor mesh exposure), and its clinical impact controversial as some gynaecologists operated on erosions34,35,61,62,71,72,81-86,89,91-93 (eg Tables 12 D, 15 D) whereas others treated erosions with debridement, vaginal oestrogens, antiseptics or antibiotics.66,77,79,91,93

Use of introducers / commercially available kits

Although the numbers were not sufficient to perform meaningful sub-group analyses within mesh types, the use of blind introducer kits has given rise to some concern. These have only been used to date with non-absorbable synthetic mesh. In total, for the safety outcome, ‘damage to surrounding organs’, there were 6/476 (1.3%) events for anterior repair, 6/276 (2.2%) for posterior repair and 16/684 (2.3%) for anterior and/or posterior repair, giving a total of 28/1436 (1.9%). Of the 28 events, 2.8% (14/503) were associated with an introducer kit and 1.5% (14/933) were not. The rate associated with an introducer kit was similar to that reported by a study100 that was identified after the review was completed in which introducer kits were used (2.7%).

Dyspareunia

One of the anecdotally cited contra-indications for the use of mesh is the likelihood of dyspareunia. This outcome is more problematic to measure because some women are not sexually active, but not all studies take this factor into account when reporting their sexual function data. Secondly, some women may be sexually inactive because of their prolapse surgery (especially when the outcome is measured within 6 months of operation). Thirdly, many studies do not measure or report this outcome at all. We made the best estimates by using two outcomes: persistent dyspareunia in women having dyspareunia at baseline (efficacy), and de novo dyspareunia in women without dyspareunia at baseline (safety). However, few studies reported data in this way.

While it is possible that dyspareunia may be caused or aggravated by mesh, its occurrence after POP surgery emphasises that all gynaecologists should be aware of the potential need for sexual function after prolapse operations. Research should be directed at identifying the reasons for the pain, its duration and effect on quality of life, surgical approaches aimed at minimising it, and interventions to treat it when it does occur.

Relationship between objective and subjective cure

The conundrum in prolapse surgery is that objective prolapse recurrence is not necessarily related to continuation of prolapse symptoms (subjective failure) and indeed can cause other symptoms or de novo prolapse, due to altered directional forces resulting from anatomical changes. It was therefore disappointing to find veryl little evidence relating to our primary outcome (subjective failure) and other genitourinary symptoms of importance to women (urinary, bowel and sexual function) in any of the trials or studies reviewed. These subjective outcomes should be considered to be of primary importance both clinically and in future research.

5.2 Assumptions, limitations, and uncertainties

A substantial number of studies currently only available as conference abstracts were excluded, including RCTs published before 2005 and all non-randomised comparative studies and case series, in order to make the task manageable within the available timeframe.

We did not search for studies which reported procedures without mesh/graft because these procedures were not the main focus of this review and also because of the limited time available for this review (initial searches of MEDLINE and EMBASE from 1996 onwards identified 2008 titles/abstracts of such studies). The data on procedures without mesh/graft in this report were based on the control groups of RCTs and non-randomised comparative studies only. Therefore, the results for ‘no mesh/graft’ were not derived from a comprehensive literature review and hence should be interpreted with caution. However, since the evidence for most outcomes for procedures with mesh/graft was insufficient, including studies which reported procedures without mesh/graft would only increase the accuracy of estimates in the ‘no mesh/graft’ group, but would not have changed the conclusions of this review (i.e. there was insufficient evidence to draw conclusions for most comparisons between procedures with and without mesh/graft or between different types of mesh/graft).

For this review, one year was considered as an adequate minimum period of time to assess the efficacy of prolapse repair. In this review, only 31 studies reported efficacy outcomes at one year or more. Nine studies with follow up of less than one year were included only for safety outcomes. The remaining nine studies (two non-randomised comparative studies and seven RCTs in conference abstracts) were included despite their short follow up because of their study design. However, even one-year outcomes are too soon to judge whether prolapse surgery is successful in the long-term. The mean time to first re-operation is reported in the literature as 12 years,3 and therefore failure at one year represents a serious underestimation of efficacy. Prospective studies would require extended follow up to assess meaningful mesh/graft failure.

In terms of methodological quality of the included full text RCTs (n=6), most RCTs reported adequate approaches to sequence generation for randomisation (5/6) and the randomisation was adequately concealed (4/6). Outcome assessors were blinded in two studies. All studies used intention-to-treat analysis (i.e. participants remained in the group to which they were allocated). The completeness of follow up in the included studies ranged between 0 to 30%. No studies reported how the authors followed up their participants.

A substantial amount of RCT evidence was available only from conference abstracts (11 of 17 RCTs). They were not amenable to quality assessment and their evidence should be interpreted with caution, as should the results of any analyses combining them with full text studies. On the other hand, they were deemed to be of higher quality by virtue of their use of randomisation as opposed to non-randomised comparative studies or case series.

Apart from using meta-analysis in RevMan for the RCTs to compare the efficacy and safety between different types of mesh/graft, we carried out crude event rate analysis of the RCTs and non-randomised comparative studies by treating each arm as a case series. We then combined the results from each arm with those from the matching case series. We were aware that non-randomised comparative studies and case series are more prone to systematic biases than other study designs such as RCTs. In particular, selection bias in the types of participants, not adjusting for confounding factors (e.g. severity of prolapse, age, co- morbidities such as diabetes or other illness), and withdrawal rates may affect the reliability and size of the treatment effects. However, we consider this as an alternative way to compare all available mesh/graft types and were able to perform analysis which adjusted for this bias to a certain extent.

Some adverse effects, particularly serious adverse events, occurred infrequently. As these events are so infrequent their estimated event rates may be prone to random error.

Categorising some of the reported outcomes was problematic. For instance, cut-off points used to determine objective failure rates varied between studies as to the degree of prolapse when failure was deemed to have occurred. Outcomes reported as mesh exposure were grouped with those reported as mesh/graft erosion. All types of infections such as urinary tract infection, wound infection and pelvic abscess were grouped together. Dyspareunia was particularly poorly reported.

Pre-specified subgroup analysis by different mesh types within non-absorbable mesh, i.e. type I to IV (see Table 6), could not be conducted because the data were too scarce. Pre- specified subgroup analysis by ‘total mesh’ and ‘mesh inlay’ could also not be conducted because of lack of data.

It is important to determine whether safety and efficacy of mesh differs between primary repair and recurrent prolapse repair. However, only one62 of the included studies reported exclusively on women having recurrent repairs (a small comparative study of only 12 women in each of 2 arms). Of the 49 included studies, only 12 reported the case mix of primary (72%) and secondary (28%) operations in 1359 women (Table 8, section 4.1.1) but no study reported the outcome data separately for the two groups. These data, however, suggest that many gynaecologists are already using mesh in women for primary repairs. They are in contrast with a study of UK gynaecologists’ self reported use of mesh, in which the majority claimed that they would only use mesh for secondary repairs.20

The results were considered generalisable as the majority of studies recruited participants from routine practice without restriction on the severity of prolapse or other patient characteristics.

6 CONCLUSIONS

6.1 Implications for the NHS

In general, the evidence for most efficacy and safety outcomes was too sparse to provide meaningful conclusions about the use of mesh/graft in anterior and/or posterior vaginal wall prolapse surgery.

There was some short-term evidence that suggested that objective prolapse recurrence rates in anterior vaginal wall prolapse repair were less if mesh or grafts were used. In the comparison between different types of mesh/graft, there was a trend for the objective failure rate to decrease as mesh strength and durability increased. Absorbable synthetic mesh (23%) and absorbable biological graft (18%) had statistically significantly higher objective failure rates than non-absorbable synthetic mesh (9%). However, no long term efficacy data were available.

There was insufficient evidence to compare objective recurrence rates in women having posterior or anterior and/or posterior repairs. For all prolapse types, there was not enough information to compare any other efficacy outcomes (subjective failure, de novo prolapse, further operation needed for prolapse, persistent urinary and bowel symptoms, and sexual problems) between mesh/graft (any type) and no mesh/graft or between different types of mesh/graft.

While there might be some evidence of differences in objective efficacy related to the use of mesh, these must be considered alongside any safety concerns. There was insufficient evidence to compare any of the pre-specified safety outcomes (blood loss, damage to surrounding organs, new urinary and bowel symptoms, and new sexual problems) between mesh/graft (any type) and no mesh/graft, or between different types of mesh/graft.

However, there was some evidence to suggest that mesh/graft may cause problems with erosion and a subsequent need for operations to remove the foreign material. For anterior repair, mesh/graft erosion increased from 0.7% using absorbable synthetic mesh to 6.0% using absorbable biological grafts and to 10.2% using non-absorbable synthetic mesh. More minor cases of mesh erosion can be treated locally with oestrogen creams or debridement. However, some women will require surgical removal of mesh under general anaesthetic. The surgical removal rate due to erosion (complete or partial) was 2.9% for absorbable synthetic mesh, 2.6% for absorbable biological graft, and 6.6% for non-absorbable synthetic

mesh. The results for mesh erosion were similar in women having posterior repair and anterior and/or posterior repair. However, the numbers were too few to conduct statistical analyses to compare the erosion rates between different types of mesh or graft.

We were unable to measure potential differences in efficacy and safety of mesh/graft between primary repairs and recurrent prolapse repairs. Of the 49 included studies, only one study reported exclusively on women having recurrent repairs, and the remainder did not report these subgroups separately. We were also unable to measure the efficacy and safety of prolapse repairs with and without introducer kits because there were insufficient data for such devices.

6.2 Implications for future research

The RCTs currently available only as conference abstracts may be available as full text studies in the future. Full text papers of the conference abstracts (RCTs published before 2005 and all non-randomised comparative studies and case series) that were excluded may also be available in the future. The expected completion time of the eight on-going studies (seven RCTs and one registry) ranges from 2008 to 2011. A further review incorporating these studies may be informative.

This review highlights the fact that there is very little evidence to assess whether the use of mesh/graft for prolapse repair improves the health and quality of life of women undergoing prolapse surgery. Therefore, high quality RCTs, with validated patient-reported outcome measures, primarily comparing the failure rate between using mesh/graft and procedures without mesh/graft, and between different types of mesh/graft, are required to determine the comparative efficacy of using mesh/graft and its optimal place in clinical practice. However, there is not yet a validated and generally accepted measure of subjective prolapse symptoms, although several have been published.94-99

It is also important to know how often further surgery is required for recurrent prolapse. To judge this, studies with long term follow up over many years would be required. In addition, the need for supplementary treatment such as pelvic floor muscle training or the use of pessaries to treat residual, recurrent or de novo prolapse, or new urogenital symptoms such as de novo urinary incontinence, must be taken into account and quantified.

Prolapse research has benefited from a general agreement that a reliable, reproducible and objective outcome measure, the POP-Q, should be used both in research and clinical

practice. Any future research should consider the use of this standardised objective prolapse measurement system while acknowledging that subjective symptoms are equally important.

Rigorous assessment of the use of mesh or grafts in prolapse surgery should be limited to RCTs with adequate power to detect clinically meaningful differences which take account of the outcomes most important to women. In addition, prospective data collection should be considered in which the operative and clinical details of women undergoing prolapse surgery with mesh can be recorded so that sufficient efficacy and safety data can be gathered to guide the use of mesh or grafts in the future. In particular, long term outcomes need to be considered as it may be many years before a prolapse recurrence becomes symptomatic.

7 REFERENCE LIST

1 Kohli N, Goldstein,DP. An overview of the clinical manifestations, diagnosis, and classification of pelvic organ prolapse [webpage on the Internet]. Waltham, MA: UpToDate; 2007 [accessed September 2007]. Available from: URL: http://patients.uptodate.com/topic.asp?file=pelvic_s/2164.

2 Brubaker L, Bump R, Jacquetin B, Schuessler B, Weidner A, Zimmern P et al. Pelvic organ prolapse. Incontinence: 2nd International Consultation on Incontinence. Plymouth: Health Publication Ltd; 2002. p. 243-266.

3 Olsen AL, Smith VJ, Bergstom JO, Colling JC, Clark AL. Epidemiology of surgically managed pelvic organ prolapse and urinary incontinence. Obstet Gynecol 1997;89:501-6.

4 Hospital Episode Statistics: Main operations 2005-2006 [database on the Internet]. Hospital Episode Statistics; 2007 [accessed September 2007]. Available from: URL: http://www.hesonline.org.uk/Ease/servlet/ContentServer?siteID=1937&categoryID=19 2.

5 Population Projections Database [database on the Internet]. Goverment Actuary's Department; 2007 [accessed September 2007]. Available from: URL: http://www.gad.gov.uk/Population/index.asp.

6 DeLancey JO, Kearney R, Chou Q, Speights S, Binno S. The appearance of levator ani muscle abnormalities in magnetic resonance images after vaginal delivery. Obstet Gynecol 2003;101:46-53.

7 Snooks SJ, Swash M, Mathers SE, Henry MM. Effect of vaginal delivery on the pelvic floor: a 5 year follow-up. Br J Surg 1990;77:1358-60.

8 Tulikangas PK. Defect theory of pelvic organ prolapse. Clin Obstet Gynecol 2005;48:662-7.

9 Baden WF, Walker TA. Genesis of the vaginal profile: a correlated classification of vaginal relaxation. Clin Obstet Gynecol 1972;15:1048-54.

10 Beecham CT. Classification of vaginal relaxation. Am J Obstet Gynecol 1980;136:957-8.

11 Bump RC, Mattiasson A, Bo K, Brubaker LP, DeLancey JO, Klarskov P et al. The standardization of terminology of female pelvic organ prolapse and pelvic floor dysfunction. Am J Obstet Gynecol 1996;175:10-7.

12 Hall AF, Theofrastous JP, Cundiff GW, Harris RL, Hamilton LF, Swift SE et al. Interobserver and intraobserver reliability of the proposed International Continence Society, Society of Gynecologic Surgeons, and American Urogynecologic Society pelvic organ prolapse classification system. Am J Obstet Gynecol 1996;175:1467-70.

13 Hagen S, Stark D, Maher C, Adams EJ. Conservative management of pelvic organ prolapse in women. Cochrane Database of Systematic Reviews 2006;Issue 4. Art. No.: CD003882. DOI: 10.1002/14651858.CD003882.pub3.

14 Maher C, Baessler K, Glazener CM, Adams EJ, Hagen S. Surgical management of pelvic organ prolapse in women. Cochrane Database of Systematic Reviews 2007;Issue 3. Art. No.: CD004014. DOI: 10.1002/14651858.CD004014.pub3.:

15 Vimplis S, Hooper P. Assessment and management of pelvic organ prolapse. Curr Obstet Gynaecol 2005;15:387-93.

16 Deprest J, Zheng F, Konstantinovic M, Spelzini F, Claerhout F, Steensma A et al. The biology behind fascial defects and the use of implants in pelvic organ prolapse repair. Int Urogynecol J Pelvic Floor Dysfunct 2006;17:S16-S25.

17 Amid PK. Classification of biomaterials and their related complication in abdominal wall hernia surgery. Hernia 1997;1:15-21.

18 Baessler K, Maher CF. Mesh augmentation during pelvic-floor reconstructive surgery: Risks and benefits. Curr Opin Obstet Gynecol 2006;18:560-6.

19 Davila GW, Drutz H, Deprest J. Clinical implications of the biology of grafts: Conclusions of the 2005 IUGA Grafts Roundtable. Int Urogynecol J Pelvic Floor Dysfunct 2006;17:S51-S55.

20 Jha S, Moran PA. National survey on the management of prolapse in the UK. Neurourol Urodyn 2007;26:325-31.

21 Verhagen AP, de Vet HC, de Bie RA, Kessels AG, Boers M, Bouter LM et al. The Delphi list: a criteria list for quality assessment of randomized clinical trials for conducting systematic reviews developed by Delphi consensus. J Clin Epidemiol 1998;51:1235-41.

22 Undertaking systematic reviews of research on effectiveness. CRD's guidance for those carrying out or commissioning reviews. CRD Report No 4. University of York: Centre for Reviews and Dissemination; 2001.

23 Downs SH, Black N. The feasibility of creating a checklist for the assessment of the methodological quality both of randomised and non-randomised studies of health care interventions. J Epidemiol Community Health 1998;52:377-84.

24 Higgins JPT, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta- analyses. Br Med J 2003;327:557-60.

25 Sutton AJ, Abrams KR. Bayesian methods in meta-analysis and evidence synthesis. Stat Methods Med Res 2001;10:277-303.

26 Spiegelhalter D, Thomas A, Best N. WinBUGS: Bayesian inference using Gibbs sampling. User manual, version 1.4. Cambridge: MRC Biostatistics Unit; 2003.

27 Sand PK, Koduri S, Lobel RW, Winkler HA, Tomezsko J, Culligan PJ et al. Prospective randomized trial of polyglactin 910 mesh to prevent recurrence of cystoceles and rectoceles. Am J Obstet Gynecol 2001;184:1357-62.

28 Milani R, Salvatore S, Soligo M, Pifarotti P, Meschia M, Cortese M. Functional and anatomical outcome of anterior and posterior vaginal prolapse repair with prolene mesh. BJOG 2005;112:107-11.

29 Altman D, Falconer C. Perioperative morbidity using transvaginal mesh in pelvic organ prolapse repair. Obstet Gynecol 2007;109:303-8.

30 De Ridder D. The use of biomaterials in reconstructive urology. Eur Urol Suppl 2002;1:7-11.

31 De Ridder D, Claerhout F, Verleyen P, Boulanger S, Deprest J. Porcine dermis xenograft as reinforcement for cystocoele stage III repair: A prospective randomized controlled trial. Neurourol Urodyn 2004;23:435-6.

32 Gandhi S, Goldberg RP, Kwon C, Koduri S, Beaumont JL, Abramov Y et al. A prospective randomized trial using solvent dehydrated fascia lata for the prevention of recurrent anterior vaginal wall prolapse. Am J Obstet Gynecol 2005;192:1649-54.

33 Gandhi S, Abramov Y, Goldberg RP, Botros SM, Miller JR, Beaumont J et al. Fascia lata for the prevention of recurrent posterior vaginal wall prolapse: subanalysis of a randomized controlled trial. J Pelvic Med Surg 2005;11:S53.

34 Meschia M, Pifarotti P, Bernasconi F, Magatti F, Riva D, Kocjancic E. Porcine skin collagen implants to prevent anterior vaginal wall prolapse recurrence: a multicenter, randomized study. J Urol 2007;177:192-5.

35 Weber AM, Walters MD, Piedmonte MR, Ballard LA. Anterior colporrhaphy: a randomized trial of three surgical techniques. Am J Obstet Gynecol 2001;185:1299- 304.

36 Paraiso MFR, Barber MD, Muir TW, Walters MD. Rectocele repair: A randomized trial of three surgical techniques including graft augmentation. Am J Obstet Gynecol 2006;195:1762-71.

37 Al-Nazer MA, Ismail WA, Gomma IA. Comparative study between anterior colporraphy versus vaginal wall repair with mesh for management of anterior vaginal wall prolapse. Int Urogynecol J Pelvic Floor Dysfunct 2007;18:S49-S50.

38 Ali S, Han HC, Lee LC. A prospective randomized trial using Gynemesh PS for the repair of anterior vaginal wall prolapse. Int Urogynecol J Pelvic Floor Dysfunct 2006;17:S221.

39 Cervigni M, Natale F, Weir J, Galante L, Agostini M, Pajoncini C. Propsective randomized trial of two new materials for the correction of anterior compartment prolapse: pelvicol and prolene soft. Annual Meeting, International Continence Society, Montreal, Canada, 22nd Aug-2nd Sep 2005. Abstract no 118.

40 Cervigni M, Natale F, Galante L, La Penna C, Weir J, Panei M et al. Prospective randomized trial of two new materials for the correction of anterior compartment prolapse: pelvicol and prolene soft. Int Urogynecol J Pelvic Floor Dysfunct 2005;16:S84.

41 Cervigni M, Natale F, Weir J, Antomarchi F. Prospective randomized controlled study of the use of a synthetic mesh (Gynemesho (R)) versus a biological mesh (Pelvicol (R)) in recurrent cystocele. J Urol 2007;177:423.

42 Hiltunen R, Nieminen K, Takala T, Hesikanen E, Niemi K. Transvaginal mesh repair of the anterior compartment prolapse; a randomized comparing prospective study. Int Urogynecol J Pelvic Floor Dysfunct 2006;17:S142.

43 Hviid UH, Rudnicki M. A randomised controlled study of biomesh (Pelvicol) for vaginal anterior wall prolapse. Int Urogynecol J Pelvic Floor Dysfunct 2005;16:S58.

44 Kocjancic E, Meschia M, Pifarotti P, Gattei U, Magatti F, Crivellaro S et al. Multicentre randomized trial of Pelvicol TM implant in the treatment of cystocele. J Urol 2005;173:233.

45 Kocjancic E, Pifarotti R, Magatti F, Bernasconi F, Riva W, Frea B et al. Evaluation of pelvicol implants in preventing recurrence of anterior vaginal wall prolapse: Randomised multicentre study. Eur Urol Suppl 2006;5:191.

46 Kocjancic E, Pifarotti P, Magatti F, Bernasconi F, Riva D, Frea B et al. Pelvicol implants for prevention of anterior vaginal wall prolapse: Randomized multicenter study. J Urol 2006;175:293.

47 Kocjancic E, Crivellaro S, Bernasconi F, Magatti F, Frea B, Meschia M. A two years follow-up, prospective randomized study on cystocele repair with or without pelvicol implant. J Urol 2007;177:453.

48 Kocjancic E, Crivellaro S, Bernasconi F, Magatti F, Frea B, Meschia M. Cystocele repair with or without pelvicol implant: A two years follow-up. Eur Urol Suppl 2007;6:238.

49 Meschia M, Pifarotti P, Magatti F, Bernasconi F, Riva D, Kojancic E. Porcine skin collagen implants (Pelvicol (TM)) to prevent anterior vaginal wall prolapse recurrence: A randomized study. Neurourol Urodyn 2005;24:587-8.

50 Meschia M, Pifarotti P, Riva D, Bernasconi F, Magatti F, Kocjancic E. Pelvicol implant for anterior vaginal wallprolapse repair: a two years randomized follow-up study. Int Urogynecol J Pelvic Floor Dysfunct 2006;17:S64-S65.

51 Natale F, Antornarchi F, La Penna C, Lo Voi R, Agostini M, Signore M et al. Prospective randomized controlled study of the use of a synthetic mesh (Gynemesh (R)) versus a biological mesh (Pelvicol (R)) in recurrent cystocele. Neurourol Urodyn 2006;25:555-7.

52 Natale F, Antomarchi F, La Penna C, Lo Voi R, Agostini M, Signore M et al. Perspective randomized controlled study between the use of a synthetic mesh (Gynemesh) and a biological mesh (Pelvicol). Int Urogynecol J Pelvic Floor Dysfunct 2006;17:S138.

53 Nguyen JN, Burchette RJ. Anatomic support and visceral function following anterior colporrhaphy versus polypropylene mash transobturator prolapse repair: Avia trial. Int Urogynecol J Pelvic Floor Dysfunct 2007;18:S51.

54 Pifarotti P, Magatti F, Bernasconi F, Riva D, Kocjancic E, Meschia M. Multicentrer randomized trial of pelvicol implant to prevent recurrence of anterior vaginal wall prolapse in women undergoing primary surgery for genital prolpase. Int Urogynecol J Pelvic Floor Dysfunct 2005;16:S52.

55 Lim YN, Muller R, Hitchins S, Tang A, Naidu A, Greenland H. Is it worth incorporating mesh in pasterior colporrhaphy: a randomized study. Int Urogynecol J Pelvic Floor Dysfunct 2006;17:S178-S179.

56 Allahdin S, Bain C, Glazener C. Feasibility study for a randomized controlled trial evaluating the use of absorbable mesh, polydioxanone and polyglactin sutures for anterior and posterior vaginal wall prolapse repairs. Annual Meeting, International Continence Society, Churchill, New Zealand, 27th Nov-1st Dec 2006. Abstract no 412.

57 Higgs PJ, Carey MP, Goh J, Krause H, Leong A, Cornish A. Randomised controlled trial comparing vaginal prolapse repair with mesh augmentation to traditional vaginal repair: a 6-month follow-up. Int Urogynecol J Pelvic Floor Dysfunct 2006;17:S64.

58 Lim YN, Carey MP, Higgs PJ, Goh J, Krause H, Leong A. Vaginal colporrhaphy versus vaginal repair with mesh for pelvic organ prolapse: a randomized controlled trial. Int Urogynecol J Pelvic Floor Dysfunct 2007;18:S38-S39.

59 Meschia M, Baccichet R, Cervigni M, Guercio E, Maglioni Q, Narducci P. A multicenter randomized trial on transvaginal mesh repair of severe genital prolapse with the Perigee-Apogee system: the Perapo study. Int Urogynecol J Pelvic Floor Dysfunct 2007;18:S10.

60 Chaliha C, Khalid U, Campagna L, Digesu GA, Ajay B, Khullar V. SIS graft for anterior vaginal wall prolapse repair--a case-controlled study. Int Urogynecol J Pelvic Floor Dysfunct 2006;17:492-7.

61 Handel LN, Frenkl TL, Kim YH. Results of cystocele repair: a comparison of traditional anterior colporrhaphy, polypropylene mesh and porcine dermis. J Urol 2007;178:153-6.

62 Julian TM. The efficacy of Marlex mesh in the repair of severe, recurrent vaginal prolapse of the anterior midvaginal wall. Am J Obstet Gynecol 1996;175:1472-5.

63 Leboeuf L, Miles RA, Kim SS, Gousse AE. Grade 4 cystocele repair using four-defect repair and porcine xenograft acellular matrix (Pelvicol): Outcome measures using SEAPI. Urology 2004;64:282-6.

64 Altman D, Mellgren A, Blomgren B, Lopez A, Zetterstrom J, Nordenstam J et al. Clinical and histological safety assessment of rectocele repair using collagen mesh. Acta Obstet Gynecol Scand 2004;83:995-1000.

65 Castelo-Branco C, Garrido JF, Ribas C, Iglesias X. Posterior vaginal wall repair with synthetic absorbable mesh: A new technique for an old procedure. J Gynecol Surg 1998;14:111-7.

66 Vakili B, Huynh T, Loesch H, Franco N, Chesson RR. Outcomes of vaginal reconstructive surgery with and without graft material. Am J Obstet Gynecol 2005;193:2126-32.

67 Cronje HS, Prollius A, de Beer JAA. Stage IV cystocele treated by sacrocolpopexy. Int J Gynecol Obstet 2006;92:153-4.

68 De Tayrac R, Fernandez H, Sand PK. Surgical repair of cystocele with mesh by the vaginal route. Am J Obstet Gynecol 2002;186:852-3.

69 De Tayrac R, Gervaise A, Chauveaud-Lambling A, Fernandez H. Combined genital prolapse repair reinforced with a polypropylene mesh and tension-free vaginal tape in women with genital prolapse and stress urinary incontinence: a retrospective case- control study with short-term follow-up. Acta Obstet Gynecol Scand 2004;83:950-4.

70 De Tayrac R, Gervaise A, Chauveaud A, Fernandez H. Tension-free polypropylene mesh for vaginal repair of anterior vaginal wall prolapse. J Reprod Med 2005;50:75- 80.

71 De Tayrac R, Deffieux X, Gervaise A, Chauveaud-Lambling A, Fernandez H. Long- term anatomical and functional assessment of trans-vaginal cystocele repair using a tension-free polypropylene mesh. Int Urogynecol J Pelvic Floor Dysfunct 2006;17:483-8.

72 Flood CG, Drutz HP, Waja L. Anterior colporrhaphy reinforced with Marlex mesh for the treatment of cystoceles. Int Urogynecol J Pelvic Floor Dysfunct 1998;9:200-4.

73 Frederick RW, Leach GE. Cadaveric prolapse repair with sling: intermediate outcomes with 6 months to 5 years of followup. J Urol 2005;173:1229-33.

74 Gomelsky A, Rudy DC, Dmochowski RR. Porcine dermis interposition graft for repair of high grade anterior compartment defects with or without concomitant pelvic organ prolapse procedures. J Urol 2004;171:1581-4.

75 Kobashi KC, Mee SL, Leach GE. A new technique for cystocele repair and transvaginal sling: the cadaveric prolapse repair and sling (CAPS). Urology 2000;56:9-14.

76 Kobashi KC, Leach GE, Chon J, Govier FE. Continued multicenter followup of cadaveric prolapse repair with sling. J Urol 2002;168:2063-8.

77 Powell CR, Simsiman AJ, Menefee SA. Anterior vaginal wall hammock with fascia lata for the correction of stage 2 or greater anterior vaginal compartment relaxation. J Urol 2004;171:264-7.

78 Safir MH, Gousse AE, Rovner ES, Ginsberg DA, Raz S. 4-Defect repair of grade 4 cystocele. J Urol 1999;161:587-94.

79 Simsiman AJ, Luber KM, Menefee SA. Vaginal paravaginal repair with porcine dermal reinforcement: Correction of advanced anterior vaginal prolapse. Am J Obstet Gynecol 2006;195:1832-6.

80 Kobashi KC, Leach GE, Frederick R, Kuznetsov DD, Hsiao KC. Initial experience with rectocele repair using nonfrozen cadaveric fascia lata interposition. Urology 2005;66:1203-7.

81 Amrute KV, Eisenberg ER, Rastinehad AR, Kushner L, Badlani GH. Analysis of outcomes of single polypropylene mesh in total pelvic floor reconstruction. Neurourol Urodyn 2007;26:53-8.

82 Costantini E, Mearini L, Bini V, Zucchi A, Mearini E, Porena M. Uterus preservation in surgical correction of urogenital prolapse. Eur Urol 2005;48:642-9.

83 De Tayrac R, Devoldere G, Renaudie J, Villard P, Guilbaud O, Eglin G et al. Prolapse repair by vaginal route using a new protected low-weight polypropylene mesh: 1-year functional and anatomical outcome in a prospective multicentre study. Int Urogynecol J Pelvic Floor Dysfunct 2007;18:251-6.

84 Dwyer PL, O'Reilly BA. Transvaginal repair of anterior and posterior compartment prolapse with Atrium polypropylene mesh. BJOG 2004;111:831-6.

85 Rozet F, Mandron E, Arroyo C, Andrews H, Cathelineau X, Mombet A et al. Laparoscopic sacral colpopexy approach for genito-urinary prolapse: experience with 363 cases. Eur Urol 2005;47:230-6.

86 Deffieux X, De Tayrac R, Huel C, Bottero J, Gervaise A, Bonnet K et al. Vaginal mesh erosion after transvaginal repair of cystocele using Gynemesh or Gynemesh-Soft in 138 women: a comparative study. Int Urogynecol J Pelvic Floor Dysfunct 2007;18:73- 9.

87 Petros PEP, Richardson PA, Goeschen K, Abendstein B. The tissue fixation system provides a new structural method for cystocoele repair: A preliminary report. Aust N Z J Obstet Gynaecol 2006;46:474-8.

88 Rodriguez LV, Bukkapatnam R, Shah SM, Raz S. Transvaginal paravaginal repair of high-grade cystocele central and lateral defects with concomitant suburethral sling: report of early results, outcomes, and patient satisfaction with a new technique. Urology 2005;66:57-64.

89 Lim YN, Rane A, Muller R. An ambispective observational study in the safety and efficacy of posterior colporrhaphy with composite Vicryl-Prolene mesh. Int Urogynecol J Pelvic Floor Dysfunct 2005;16:126-31.

90 Achtari C, Hiscock R, O'Reilly BA, Schierlitz L, Dwyer PL. Risk factors for mesh erosion after transvaginal surgery using polypropylene (Atrium) or composite polypropylene/polyglactin 910 (Vypro II) mesh. Int Urogynecol J Pelvic Floor Dysfunct 2005;16:389-94.

91 Collinet P, Belot F, Debodinance P, Ha DE, Lucot JP, Cosson M. Transvaginal mesh technique for pelvic organ prolapse repair: mesh exposure management and risk factors. Int Urogynecol J Pelvic Floor Dysfunct 2006;17:315-20.

92 Cosson M, Rajabally R, Bogaert E, Querleu D, Crepin G. Laparoscopic sacrocolpopexy, hysterectomy, and burch colposuspension: feasibility and short-term complications of 77 procedures. J Soc Laparoendosc Surg 2002;6:115-9.

93 Fatton B, Amblard J, Debodinance P, Cosson M, Jacquetin B. Transvaginal repair of genital prolapse: preliminary results of a new tension-free vaginal mesh (Prolifttrade mark technique)-a case series multicentric study. Int Urogynecol J Pelvic Floor Dysfunct 2007;18:743-52.

94 Price N, Jackson SR, Avery K, Brookes ST, Abrams P. Development and psychometric evaluation of the ICIQ Vaginal Symptoms Questionnaire: the ICIQ-VS. BJOG 2006;113:700-12.

95 Lukacz ES, Lawrence JM, Buckwalter JG, Burchette RJ, Nager CW, Luber KM. Epidemiology of prolapse and incontinence questionnaire: validation of a new epidemiologic survey. Int Urogynecol J Pelvic Floor Dysfunct 2005;16:272-84.

96 Barber MD, Walters MD, Bump RC. Short forms of two condition-specific quality-of- life questionnaires for women with pelvic floor disorders (PFDI-20 and PFIQ-7). Am J Obstet Gynecol 2005;193:103-13.

97 Tan JS, Lukacz ES, Menefee SA, Powell CR, Nager CW, San Diego Pelvic Floor Consortium. Predictive value of prolapse symptoms: a large database study. Int Urogynecol J Pelvic Floor Dysfunct 2005;16:203-9.

98 Digesu GA, Khullar V, Cardozo L, Robinson D, Salvatore S. P-QOL: a validated questionnaire to assess the symptoms and quality of life of women with urogenital prolapse. Int Urogynecol J Pelvic Floor Dysfunct 2005;16:176-81.

99 Barber MD, Kuchibhatla MN, Pieper CF, Bump RC. Psychometric evaluation of 2 comprehensive condition-specific quality of life instruments for women with pelvic floor disorders. Am J Obstet Gynecol 2001;185:1388-95.

100 Abdel-fattah M, Ramsay I on behalf of the West of Scotland Study Group. Retrospective multicentre study of the new minimally invasive mesh repair devices for pelvic organ prolapse. BJOG 2008;115:22-30.

APPENDIX 1 Search strategies

MEDLINE ( 1966-June Wk 3 2007) EMBASE (1980-2007 Wk 26) (Medline In Process 3rd July 2007) ) Ovid Multifile Search URL: http://gateway.ovid.com/athens

1 pelvic organ prolapse/ use emez 2 vaginal vault prolapse/ use emez 3 uterine prolapse/ 4 prolapse/ use mesz 5 rectocele/ 6 cystocele/ 7 enterocele/ use emez 8 (prolapse adj3 pelvic).tw. 9 (cystoc?ele$ or urethroc?ele$ or cystourethroc?ele$).tw. 10 (enteroc?ele$ or rectoc?ele$ or sigmoidoc?ele$).tw. 11 (prolapse adj2 (genital or vagina? or uter$ or cervic$ or apical)).tw. 12 or/1-11 13 urogenital surgical procedures/ or gynecologic surgical procedures/ or urologic surgical procedures/ use mesz 14 gynecologic surgery/ or urologic surgery/ use emez 15 (surgery or surgical or repair$).tw. 16 12 and (13 or 14 or 15) 17 pelvic floor/su 18 vagina/su 19 ((anterior or posterior) adj1 repair).tw. 20 (colporrhaphy or colpoperineorraphy or cystopex$).tw. 21 (sacro$ hysteropex$ or sacral hysteropex$ or sacrohysteropex$).tw. 22 (sacro$ colpopex$ or sacral colpopex$ or sacrocolpopex$).tw. 23 (sacropex$ or intravaginal slingplast$).tw. 24 (suspension adj2 (uter$ or vault or sacro$)).tw. 25 12 and su.fs. 26 or/16-25 27 surgical mesh/ 28 exp fascia/ 29 (mesh or graft$ or allograft$ or fascia).tw. 30 polypropylenes/ or polyglactin 910/ 31 (prolene or trelex or avaulta or pelvitex or prolift or polyform or marlex or gynemesh or gore tex).tw,dv. 32 (apogee or perigee or straight in or prolite or ultrapro or mersilene or vypro or novasilk or minimesh or pivit or surgsis).tw,dv. 33 (axis or suspend).dv 34 (faslata or fortagen).tw,dv 35 (porcine dermis or pelvicol or pelvisoft).tw,dv 36 or/27-35 37 26 and 36 38 limit 37 to yr="1980 - 2007" 39 eng.la. 40 38 and 39 41 animals/ not humans/ 42 (animal/ or nonhuman/) not human/ use emez 43 40 not (41 or 42) 44 clinical trial.pt. 45 exp controlled study/ use emez 46 randomi?ed.ab.

47 randomly.ab. 48 trial.ab. 49 groups.ab. 50 comparative study.pt. 51 follow-up studies/ use mesz 52 time factors/ use mesz 53 Treatment outcome/ use emez 54 major clinical study/ use emez 55 clinical trial/ use emez 56 (preoperat$ or pre operat$).mp. use mesz 57 (chang$ or evaluat$ or reviewed or baseline).tw. 58 (prospective$ or retrospective$).tw. use mesz 59 (cohort$ or case series).tw. use mesz 60 (compare$ or compara$).tw. use emez 61 case report/ use emez 62 case reports.pt. 63 case?.tw. 64 or/44-63 65 ae.fs 66 co.fs 67 postoperative complications/ 68 intraoperative complications/ 69 or/65-68 70 43 and 64 71 43 and 69 72 (letter or editorial or note or comment).pt. 73 ((70 or 71) not 72 ) 74 remove duplicates from 73

Science Citation Index (1980- 1st July 2007) Biosis (1985-5th July 2007) ISI Proceedings (1990-27th June 2007) Web of Knowledge URL: http://wok.mimas.ac.uk/

#1 TS=(prolapse SAME (pelvic OR genital OR vagina*)) #2 TS=(prolapse SAME (uter* OR cervic* OR apical)) #3 TS=(cystocele* OR cystocoele* or cystourethroc* or urethrocele* or urethrocoele*) #4 TS=(eneterocele* or enterocoele* or rectocele* or rectocoele* or sigmoidocele* or sigmoidocoele*) #5 #1 or #2 or #3 or #4 #6 TS=(repair SAME (anterior or posterior)) #7 TS=(colporrhaphy OR colpoperineorraphy OR cystopex*) #8 TS=(sacro* hysteropex* OR sacral hysteropex* OR sacrohysteropex*) #9 TS=(sacro* colpopex* or sacral colpopex* OR sacrocolpopex*) #10 TS=(sacropex* OR intravaginal slingplast*) #11 TS=(suspension SAME (uter* OR vault OR sacro*)) #12 #6 OR #7 OR #8 OR #9 OR #10 OR #11 #13 #5 OR #12 #14 TS=(mesh OR fascia) #15 TS=graft #16 TS=allograft #17 TS=(polypropylene* OR prolene OR trelex OR avualta OR pelvitex OR prolift OR polyform OR marlex OR gynemesh OR gore tex) #18 TS=(polyglactin OR vicryl OR tutoplast OR faslata OR fortagen) #19 TS=(porcine dermis OR pelvicol OR pelvisoft)

#20 #13 AND (#14 OR #15 OR #16 OR #17 OR #18 OR #19) DocType=All document types; Language=English; Timespan=1980-2007

Cochrane Library ( Issue 2, 2007) URL: http://www3.interscience.wiley.com/cgi-bin/mrwhome/106568753/HOME

#1 MeSH descriptor Uterine Prolapse explode all trees #2 MeSH descriptor Prolapse explode all trees #3 MeSH descriptor Rectocele explode all trees #4 (prolapse near/3 pelvic):ti,ab,kw or (prolapse near/2 (genital or vagina* or uter*)):ti,ab,kw or (prolapse near/2 (cervic* or apical)):ti,ab,kw #5 (cystoc?ele* or urethroc?ele* or cystourethroc?ele*):ti,ab,kw or (enteroc?ele* or rectoc?ele* or sigmoidoc?ele*):ti,ab,kw #6 (#1 OR #2 OR #3 OR #4 OR #5) #7 MeSH descriptor Urogenital Surgical Procedures explode all trees #8 MeSH descriptor Gynecologic Surgical Procedures explode all trees #9 MeSH descriptor Urologic Surgical Procedures explode all trees #10 (surgery or surgical or repair*):ti,ab,kw #11 (#7 OR #8 OR #9 OR #10) #12 (#6 AND #11) #13 MeSH descriptor Pelvic Floor explode all trees with qualifier: SU #14 MeSH descriptor Vagina explode all trees with qualifier: SU #15 ((anterior or posterior) near/1 repair):ti,ab,kw #16 (colporrhaphy or colpoperineorraphy or cystopex*):ti,ab,kw or (sacro* hysteropex*):ti,ab,kw or (sacral hysteropex*):ti,ab,kw or (sacrohysteropex*):ti,ab,kw #17 (sacro* colpopex*):ti,ab,kw or (sacral colpopex*):ti,ab,kw or (sacrocolpopex*):ti,ab,kw #18 (sacropex* or intravaginal slingplast*):ti,ab,kw #19 (suspension near/2 (uter* or vault or sacro*)):ti,ab,kw #20 (#12 OR #13 OR #14 OR #15 OR #16 OR #17 OR #18 OR #19) #21 MeSH descriptor Surgical Mesh explode all trees #22 MeSH descriptor Fascia explode all trees #23 MeSH descriptor Polypropylenes, this term only #24 MeSH descriptor Polyglactin 910, this term only #25 (mesh or graft* or allograft* or fascia):ti,ab,kw #26 (prolene or trelex or avaulta) or (pelvitex or prolift or polyform) or (marlex or gynemesh or gore tex) #27 (vicryl or tutoplast or faslata or fortagen) #28 (porcine dermis or pelvicol or pelvisoft) #29 (#21 OR #22 OR #23 OR #24 OR #25 OR #26 OR #27 OR #28) #30 (#20 AND #29)

DARE and HTA Databases (June 2007) NHS Centre for Reviews & Dissemination URL: http://nhscrd.york.ac.uk/welcome.htm

10. enteroc* OR rectoc* OR sigmoidoc* 11. enterocele* OR enterocoele* 12. rectocele* OR rectocoele* 13. sigmoidocele* OR rectocoele* 14. sigmoidocele* OR sigmoidocoele* 15. #1 or #2 or #3 or #4 or #5 or #6 or #7 or #8 or #9 or #10 or #11 or #12 or #13 or #14 16. MeSH Pelvic Floor QUALIFIERS SU EXPLODE 1 2 17. MeSH Vagina QUALIFIERS SU EXPLODE 1 18. ( ( anterior OR posterior ) AND repair ) 19. colporrhaphy OR colpoperineorraphy OR cystopex* 20. sacro* AND hysteropex* OR sacral AND hysteropex* OR sacrocrohysteropex* 21. sacro* AND colpopex* OR sacral AND colpopex* OR sacrocolpopex* 22. sacropex* OR intravaginal AND slingplast* 23. suspension AND ( uter* OR vault OR sacro* ) 24. #16 or #17 or #18 or #19 or #20 or #21 or #22 or #23 25. #15 or #24 26. MeSH Surgical Mesh 27. MeSH Fascia EXPLODE 1 28. MeSH Polypropylenes EXPLODE 1 2 3 29. MeSH Polyglactin 910 EXPLODE 1 2 30. prolene OR trelex OR avaulta OR pelvitex 31. prolift OR polyform OR marlex OR gynemesh OR gore 32. vicryl 33. tutoplast OR faslata OR fortagen 34. porcine OR dermis OR pelvicol OR pelvisoft 35. #26 or #27 or #28 or #29 or #30 or #31 or #32 or #33 or #34 36. #25 and #35

National Research Register (Issue 2, 2007) URL: http://www.update-software.com/National/

1. MeSH Uterine Prolapse 2. MeSH Prolapse 3. MeSH Rectocele 4. MeSH Bladder Diseases 5. ( prolapse AND ( pelvic OR genital OR vagina* ) ) 6. ( prolapse AND ( uter* OR cervic* OR apical ) ) 7. cystocele* OR cystocoele* 8. urethrocele* OR urethrocoele* 9. cystourethroc* 10. enteroc* OR rectoc* OR sigmoidoc* 11. enterocele* OR enterocoele* 12. rectocele* OR rectocoele* 13. sigmoidocele* OR rectocoele* 14. sigmoidocele* OR sigmoidocoele* 15. #1 or #2 or #3 or #4 or #5 or #6 or #7 or #8 or #9 or #10 or #11 or #12 or #13 or #14 16. MeSH Pelvic Floor QUALIFIERS SU EXPLODE 1 2 17. MeSH Vagina QUALIFIERS SU EXPLODE 1 18. ( ( anterior OR posterior ) AND repair ) 19. colporrhaphy OR colpoperineorraphy OR cystopex* 20. sacro* AND hysteropex* OR sacral AND hysteropex* OR sacrohysteropex* 21. sacro* AND colpopex* OR sacral AND colpopex* OR sacrocolpopex* 22. sacropex* OR intravaginal AND slingplast* 23. suspension AND ( uter* OR vault OR sacro* ) 24. #16 or #17 or #18 or #19 or #20 or #21 or #22 or #23 25. #15 or #24

26. MeSH Surgical Mesh 27. MeSH Fascia EXPLODE 1 28. MeSH Polypropylenes EXPLODE 1 2 3 29. MeSH Polyglactin 910 EXPLODE 1 2 30. prolene OR trelex OR avaulta OR pelvitex 31. prolift OR polyform OR marlex OR gynemesh OR gore 32. vicryl 33. tutoplast OR faslata OR fortagen 34. porcine OR dermis OR pelvicol OR pelvisoft 35. #26 or #27 or #28 or #29 or #30 or #31 or #32 or #33 or #34 36. #25 and #35

Clinical Trials (April 2007) URL: http://clinicaltrials.gov/ct/gui/c/r

(prolapse and (vaginal or pelvic or uterine or cervical or genital or apical)

Current Controlled Trials (April 2007) URL: http://www.controlled-trials.com/

(prolapse and (vagina% or pelvic or uter% or cervic% or genital or apical)

Conference Proceedings Abstracts:

American Urogynecologic Society 2006 Annual Scientific Meeting, Palm Springs,CA., 19-21 October. 2005 Annual Scientific Meeting, Atlanta, GA., October.

American Urological Association Annual Meetings 2005-7

Available form: URL: HTUH http://www.abstracts2view.com/aua/sessionindex.phpUTH

European Association of Urology 2007 Annual Congress, Berlin,Germany, 21-24 March. 2006 Annual Congress, Paris, France, 5-8 April. 2005 Annual Congress, Istanbul, Turkey, 16-19 March.

European Society of Gynecological Endoscopy 15th Annual Congress, Strasbourg, France, 5-7 October. 14th Annual Congress, Athens, Greece, 6-8 October.

Incontinence Society 2006 Annual Meeting,Christchurch,New Zealand, 27 November-1 December. 2005 Annual Meeting,Montreal, Canada, 22 August-2 September.

International Urogynecological Association 2007 Annual Meeting, Cacun,Mexico, 12-16 June. 2006 Annual Meeting, Athens,Greece, 6-9 September. 2005 Annual Meeting, Copenhagen, Denmark, 9-12 August.

Websites searched for other evidence-based reports and background information:

American Medical Systems URL: http://www.americanmedicalsystems.com/ American Urogynecologic Society URL: http://www.augs.org/

100

Americal Urological Association URL: http://www.auanet.org/ Atrium Medical URL: http://www.atriummed.com/atrium.asp Bard Urological URL: http://www.bardurological.com/ Boston Scientific URL: http://www.bostonscientific.co.uk/ Colplast URL: http://www.urology.coloplast.com/ Ethicon Women’s Health & Urology URL: http://www.womenshealthsolutions.co.uk/ European Society of Gynecological Endoscopy URL: http://www.esge.org/ European Association of Urology URL: http://www.uroweb.org/ Gore Medical URL: http://www.goremedical.com/ Incontinence Continence Society: URL https://www.icsoffice.org/ International Urogynecological Association URL: http://www.iuga.org/ Johnson & Johnson URL: http://www.jnjgateway.com/ Mpathy URL: http://www.mpathy.org/ Organogenesis URL: http://www.organogenesis.com/ Tutogen Medical URL: http://www.tutogen.com/

101

APPENDIX 2 Checklist of quality assessment of randomised controlled trials

Criteria Yes No Unclear Comment 1. Was the assignment to the treatment groups really random?

2. Was the treatment allocation concealed from those responsible for entering patients into trials, i.e. not knowing upcoming assignments in advance?

3. Were the groups similar at baseline in terms of prognostic factors, e.g. age, duration of disease, disease severity?1 4. Were the eligibility criteria specified? 5. Was the intervention (and comparison) clearly defined? 6. Were the groups treated in the same way apart from the intervention received? 7. Was there a follow-up period ≥ 1 year? 8. Was the outcome assessor blinded to the treatment allocation? 9. If patient blind is possible, were the patients blinded?2 10. If having primary outcome measures as continuous data, were the point estimates and measures of variability presented?3 11. Were the withdrawals/drop-outs having similar characteristics as those completed the study and therefore unlikely to cause bias?4 12. Did the analyses include all women according to randomised groups, i.e. intention-to-treat analysis?5 13. Was the operation undertaken by somebody experienced in performing the procedure?6

Note:

1. ‘Yes’ if two or more than two factors were similar. 2. If patient blinding is impossible, note ‘impossible’ in comment area and leave other cells blank. 3. If having no primary outcome measures as continuous data, note ‘no continuous data’ in comment area and leave other cells blank. 4. ‘Yes’ if no withdrawal/drop out; ‘No’ if drop-out rate ≥30% or differential drop-out. 5. ‘Yes’ if no withdrawals/drop out after enrol 6. ‘Yes’ if the practitioner received training on conducting the procedure before or conducted same kind of procedure before, i.e. no learning curve.

102

APPENDIX 3 Checklist of quality assessment of non-randomised studies

Criteria Yes No Unclear Comments 1. Were participants a representative sample selected from a relevant patient population, e.g. randomly selected from those seeking for treatment despite of age, duration of disease, primary or secondary disease, and severity of disease? 2. Were the inclusion/exclusion criteria of participants clearly described? 3. Were participants entering the study at a similar point in their disease progression, i.e. severity of disease? 4. Was selection of patients consecutive? 5. Was data collection undertaken prospectively? 6. Were the groups comparable on demographic characteristics and clinical features? 7. Was the intervention (and comparison) clearly defined? 8. Was the intervention undertaken by someone experienced at performing the procedure?1 9. Were the staff, place, and facilities where the patients were treated appropriate for performing the procedure? (E.g. access to back-up facilities in hospital or special clinic) 10. Were all the important outcomes considered? 11. Were objective (valid and reliable) outcome measures used, including satisfaction scale? 12. Was the assessment of main outcomes blind? 13. Was follow-up long enough (≥1y) to detect important effects on outcomes of interest? 14. Was information provided on non-respondents, dropouts?2 15. Were the withdrawals/drop-outs having similar characteristics as those completed the study and therefore unlikely to cause bias?3 16. Was length of follow-up similar between comparison groups 17. Were all the important prognostic factors identified, e.g. age, duration of disease, disease severity?4 18. Were the analyses adjusted for confounding factors? The same form was adapted to assess the quality of case series after taking out question 6, 12, 16 and 18.

Note: 1. ‘Yes’ if the practitioner received training on conducting the procedure before or conducted same kind of procedure before, i.e. no learning curve. 2. ‘No’ if participants were from those whose follow up records were available (retrospective) 3. ‘Yes’ if no withdrawal/drop out; ‘No’ if drop-out rate ≥30% or differential drop-out, e.g. those having most severe disease died during follow up but the death was not due to treatment; no description of those lost. 4. ‘Yes’ if two or more than two factors were similar.

103

APPENDIX 4 Reference list of included studies

Achtari 2005 Achtari C, Hiscock R, O'Reilly BA, Schierlitz L, Dwyer PL. Risk factors for mesh erosion after transvaginal surgery using polypropylene (Atrium) or composite polypropylene/polyglactin 910 (Vypro II) mesh. Int Urogynecol J Pelvic Floor Dysfunct 2005;16(5):389-94.

Al-Nazer 2007 Al-Nazer MA, Ismail WA, Gomma IA. Comparative study between anterior colporraphy versus vaginal wall repair with mesh for management of anterior vaginal wall prolpase. Int Urogynecol J Pelvic Floor Dysfunct 2007;18(Suppl 1):S49-S50.

Ali 2006 Ali S, Han HC, Lee LC. A prospective randomized trial using Gynemesh PS for the repair of anterior vaginal wall prolapse. Int Urogynecol J Pelvic Floor Dysfunct 2006;17(Suppl 2):S221.

Allahdin 2006 Allahdin S, Bain C, Glazener C. Feasibility study for a randomized controlled trial evaluating the use of absorbable mesh, polydioxanone and polyglactin sutures for anterior and posterior vaginal wall prolapse repairs. Annual Meeting, International Continence Society, Churchill, New Zealand, 27th Nov-1st Dec 2006.Abstract no 412.

Altman 2004 Altman D, Mellgren A, Blomgren B, Lopez A, Zetterstrom J, Nordenstam J et al. Clinical and histological safety assessment of rectocele repair using collagen mesh. Acta Obstet Gynecol Scand 2004;83(10):995-1000.

Altman 2007 Altman D, Falconer C. Perioperative morbidity using transvaginal mesh in pelvic organ prolapse repair. Obstet Gynecol 2007;109(2 Pt 1):303-8.

Amrute 2007 Amrute KV, Eisenberg ER, Rastinehad AR, Kushner L, Badlani GH. Analysis of outcomes of single polypropylene mesh in total pelvic floor reconstruction. Neurourol Urodyn 2007;26(1):53-8.

Castelo-Branco 1998 Castelo-Branco C, Garrido JF, Ribas C, Iglesias X. Posterior vaginal wall repair with synthetic absorbable mesh: A new technique for an old procedure. J Gynecol Surg 1998;14(3):111-7.

Cervigni 2007 Primary reference Cervigni M, Natale F, Weir J, Antomarchi F. Prospective randomized controlled study of the use of a synthetic mesh (Gynemesho (R)) versus a biological mesh (Pelvicol (R)) in recurrent cystocele. J Urol 2007;177(4):423.

Secondary references Cervigni M, Natale F, Galante L, La Penna C, Weir J, Panei M et al. Prospective randomized trial of two new materials for the correction of anterior compartment prolapse: pelvicol and prolene soft. Int Urogynecol J Pelvic Floor Dysfunct 2005;16(Suppl 2):S84.

Cervigni M, Natale F, Weir J, Galante L, Agostini M, Pajoncini C. Propsective randomized trial of two new materials for the correction of anterior compartment prolapse: pelvicol and

104

prolene soft. Annual Meeting, International Continence Society, Montreal, Canada, 22nd Aug-2nd Sep 2005.Abstract no 118.

Natale F, Antomarchi F, La Penna C, Lo Voi R, Agostini M, Signore M et al. Perspective randomized controlled study between the use of a synthetic mesh (Gynemesh) and a biological mesh (Pelvicol). Int Urogynecol J Pelvic Floor Dysfunct 2006;17(Suppl 2):S138.

Natale F, Antornarchi F, La Penna C, Lo Voi R, Agostini M, Signore M et al. Prospective randomized controlled study of the use of a synthetic mesh (Gynemesh (R)) versus a biological mesh (Pelvicol (R)) in recurrent cystocele. Neurourol Urodyn 2006;25(6):555-7.

Chaliha 2006 Chaliha C, Khalid U, Campagna L, Digesu GA, Ajay B, Khullar V. SIS graft for anterior vaginal wall prolapse repair--a case-controlled study. Int Urogynecol J Pelvic Floor Dysfunct 2006;17(5):492-7.

Collinet 2006 Collinet P, Belot F, Debodinance P, Ha DE, Lucot JP, Cosson M. Transvaginal mesh technique for pelvic organ prolapse repair: mesh exposure management and risk factors. Int Urogynecol J Pelvic Floor Dysfunct 2006;17(4):315-20.

Cosson 2002 Cosson M, Rajabally R, Bogaert E, Querleu D, Crepin G. Laparoscopic sacrocolpopexy, hysterectomy, and burch colposuspension: feasibility and short-term complications of 77 procedures. J Soc Laparoendosc Surg 2002;6(2):115-9.

Costantini 2005 Costantini E, Mearini L, Bini V, Zucchi A, Mearini E, Porena M. Uterus preservation in surgical correction of urogenital prolapse. Eur Urol 2005;48(4):642-9.

Cronje 2006 Cronje HS, Prollius A, de Beer JAA. Stage IV cystocele treated by sacrocolpopexy. Int J Gynecol Obstet 2006;92(2):153-4.

De Ridder 2002 Primary reference De Ridder D. The Use of biomaterials in reconstructive urology. Eur Urol Suppl 2002;1(10):7- 11.

Secondary reference De Ridder D, Claerhout F, Verleyen P, Boulanger S, Deprest J. Porcine dermis xenograft as reinforcement for cystocoele stage III repair: A prospective randomized controlled trial. Neurourol Urodyn 2004;23(5-6):435-6.

De Tayrac 2007 De Tayrac R, Devoldere G, Renaudie J, Villard P, Guilbaud O, Eglin G et al. Prolapse repair by vaginal route using a new protected low-weight polypropylene mesh: 1-year functional and anatomical outcome in a prospective multicentre study. Int Urogynecol J Pelvic Floor Dysfunct 2007;18(3):251-6.

De Tayrac 2006 Primary reference De Tayrac R, Deffieux X, Gervaise A, Chauveaud-Lambling A, Fernandez H. Long-term anatomical and functional assessment of trans-vaginal cystocele repair using a tension-free polypropylene mesh. Int Urogynecol J Pelvic Floor Dysfunct 2006;17(5):483-8.

105

Secondary references De Tayrac R, Gervaise A, Chauveaud A, Fernandez H. Tension-free polypropylene mesh for vaginal repair of anterior vaginal wall prolapse. J Reprod Med 2005;50(2):75-80.

De Tayrac R, Gervaise A, Chauveaud-Lambling A, Fernandez H. Combined genital prolapse repair reinforced with a polypropylene mesh and tension-free vaginal tape in women with genital prolapse and stress urinary incontinence: a retrospective case-control study with short-term follow-up. Acta Obstet Gynecol Scand 2004;83(10):950-4.

De Tayrac R, Fernandez H, Sand PK. Surgical repair of cystocele with mesh by the vaginal route. Am J Obstet Gynecol 2002;186(4):852-3.

Deffieux 2007 Deffieux X, De Tayrac R, Huel C, Bottero J, Gervaise A, Bonnet K et al. Vaginal mesh erosion after transvaginal repair of cystocele using Gynemesh or Gynemesh-Soft in 138 women: a comparative study. Int Urogynecol J Pelvic Floor Dysfunct 2007;18(1):73-9.

Dwyer 2004 Dwyer PL, O'Reilly BA. Transvaginal repair of anterior and posterior compartment prolapse with Atrium polypropylene mesh. BJOG 2004;111(8):831-6.

Fatton 2007 Fatton B, Amblard J, Debodinance P, Cosson M, Jacquetin B. Transvaginal repair of genital prolapse: preliminary results of a new tension-free vaginal mesh (Prolifttrade mark technique)-a case series multicentric study. Int Urogynecol J Pelvic Floor Dysfunct 2007 18(7):743-52.

Flood 1998 Flood CG, Drutz HP, Waja L. Anterior colporrhaphy reinforced with Marlex mesh for the treatment of cystoceles. Int Urogynecol J Pelvic Floor Dysfunct 1998;9(4):200-4.

Frederick 2005 Frederick RW, Leach GE. Cadaveric prolapse repair with sling: intermediate outcomes with 6 months to 5 years of followup. J Urol 2005;173(4):1229-33.

Gandhi 2005 Primary reference Gandhi S, Goldberg RP, Kwon C, Koduri S, Beaumont JL, Abramov Y et al. A prospective randomized trial using solvent dehydrated fascia lata for the prevention of recurrent anterior vaginal wall prolapse. Am J Obstet Gynecol 2005;192(5):1649-54.

Secondary reference Gandhi S, Abramov Y, Goldberg RP, Botros SM, Miller JR, Beaumont J et al. Fascia lata for the prevention of recurrent posterior vaginal wall prolapse: subanalysis of a randomized controlled trial. J Pelvic Med Surg 2005;11(Suppl):S53.

Gomelsky 2004 Gomelsky A, Rudy DC, Dmochowski RR. Porcine dermis interposition graft for repair of high grade anterior compartment defects with or without concomitant pelvic organ prolapse procedures. J Urol 2004;171(4):1581-4.

Handel 2007 Handel LN, Frenkl TL, Kim YH. Results of cystocele repair: a comparison of traditional anterior colporrhaphy, polypropylene mesh and porcine dermis. J Urol 2007;178(1):153-6.

106

Hiltunen 2006 Hiltunen R, Nieminen K, Takala T, Hesikanen E, Niemi K. Transvaginal mesh repair of the anterior compartment prolapse; a randomized comparing prospective study. Int Urogynecol J Pelvic Floor Dysfunct 2006;17(Suppl 2):S142.

Hviid 2005 Hviid UH, Rudnicki M. A randomised controlled study of biomesh (Pelvicol) for vaginal anterior wall prolapse. Int Urogynecol J Pelvic Floor Dysfunct 2005;16(Suppl 2):S58.

Julian 1996 Julian TM. The efficacy of Marlex mesh in the repair of severe, recurrent vaginal prolapse of the anterior midvaginal wall. Am J Obstet Gynecol 1996;175(6):1472-5.

Kobashi 2005 Kobashi KC, Leach GE, Frederick R, Kuznetsov DD, Hsiao KC. Initial experience with rectocele repair using nonfrozen cadaveric fascia lata interposition. Urology 2005;66(6):1203-7.

Kobashi 2002 Primary reference Kobashi KC, Leach GE, Chon J, Govier FE. Continued multicenter followup of cadaveric prolapse repair with sling. J Urol 2002;168(5):2063-8.

Secondary reference Kobashi KC, Mee SL, Leach GE. A new technique for cystocele repair and transvaginal sling: the cadaveric prolapse repair and sling (CAPS). Urology 2000; 56(6 Suppl 1):9-14.

Kocjancic 2007 Primary reference Kocjancic E, Crivellaro S, Bernasconi F, Magatti F, Frea B, Meschia M. A two years follow- up, prospective randomized study on cystocele repair with or without pelvicol implant. J Urol 2007;177(4):453.

Secondary references Kocjancic E, Crivellaro S, Bernasconi F, Magatti F, Frea B, Meschia M. Cystocele repair with or without pelvicol implant: A two years follow-up. Eur Urol Suppl 2007;6(2):238.

Kocjancic E, Pifarotti R, Magatti F, Bernasconi F, Riva W, Frea B et al. Evaluation of pelvicol implants in preventing recurrence of anterior vaginal wall prolapse: Randomised multicentre study. Eur Urol Suppl 2006;5(2):191.

Kocjancic E, Pifarotti P, Magatti F, Bernasconi F, Riva D, Frea B et al. Pelvicol implants for prevention of anterior vaginal wall prolapse: Randomized multicenter study. J Urol 2006;175(4):293.

Kocjancic E, Meschia M, Pifarotti P, Gattei U, Magatti F, Crivellaro S et al. Multicentre randomized trial of Pelvicol TM implant in the treatment of cystocele. J Urol 2005;173(4):233.

Meschia M, Pifarotti P, Riva D, Bernasconi F, Magatti F, Kocjancic E. Pelvicol implant for anterior vaginal wallprolapse repair: a two years randomized follow-up study. Int Urogynecol J Pelvic Floor Dysfunct 2006;17(Suppl 2):S64-S65.

Meschia M, Pifarotti P, Magatti F, Bernasconi F, Riva D, Kojancic E. Porcine skin collagen implants (Pelvicol (TM)) to prevent anterior vaginal wall prolapse recurrence: A randomized study. Neurourol Urodyn 2005;24(5-6):587-8.

107

Pifarotti P, Magatti F, Bernasconi F, Riva D, Kocjancic E, Meschia M. Multicentrer randomized trial of pelvicol implant to prevent recurrence of anterior vaginal wall prolapse in women undergoing primary surgery for genital prolpase. Int Urogynecol J Pelvic Floor Dysfunct 2005;16(Suppl 2):S52.

Leboeuf 2004 Leboeuf L, Miles RA, Kim SS, Gousse AE. Grade 4 cystocele repair using four-defect repair and porcine xenograft acellular matrix (Pelvicol): Outcome measures using SEAPI. Urology 2004;64(2):282-6.

Lim 2005 Lim YN, Rane A, Muller R. An ambispective observational study in the safety and efficacy of posterior colporrhaphy with composite Vicryl-Prolene mesh. Int Urogynecol J Pelvic Floor Dysfunct 2005;16(2):126-31.

Lim 2006 Lim YN, Muller R, Hitchins S, Tang A, Naidu A, Greenland H. Is it worth incorporating mesh in pasterior colporrhaphy: a randomized study. Int Urogynecol J Pelvic Floor Dysfunct 2006;17(Suppl 2):S178-S179.

Lim 2007 Primary reference Lim YN, Carey MP, Higgs PJ, Goh J, Krause H, Leong A. Vaginal colporrhaphy versus vaginal repair with mesh for pelvic organ prolapse: a randomized controlled trial. Int Urogynecol J Pelvic Floor Dysfunct 2007;18(Suppl 1):S38-S39.

Secondary reference Higgs PJ, Carey MP, Goh J, Krause H, Leong A, Cornish A. Randomised controlled trial comparing vaginal prolapse repair with mesh augmentation to traditional vaginal repair: a 6- month follow-up. Int Urogynecol J Pelvic Floor Dysfunct 2006;17(Suppl 2):S64.

Meschia 2007a Meschia M, Pifarotti P, Bernasconi F, Magatti F, Riva D, Kocjancic E. Porcine skin collagen implants to prevent anterior vaginal wall prolapse recurrence: a multicenter, randomized study. J Urol 2007;177(1):192-5.

Meschia 2007b Meschia M, Baccichet R, Cervigni M, Guercio E, Maglioni Q, Narducci P. A multicenter randomized trial on transvaginal mesh repair of severe genital prolapse with the Perigee- Apogee system: the Perapo study. Int Urogynecol J Pelvic Floor Dysfunct 2007;18(Suppl 1):S10.

Milani 2005 Milani R, Salvatore S, Soligo M, Pifarotti P, Meschia M, Cortese M. Functional and anatomical outcome of anterior and posterior vaginal prolapse repair with prolene mesh. BJOG 2005;112(1):107-11.

Nguyen 2007 Nguyen JN, Burchette RJ. Anatomic support and visceral function following anterior colporrhaphy versus polypropylene mash transobturator prolapse repair: Avia trial. Int Urogynecol J Pelvic Floor Dysfunct 2007;18(Suppl 1):S51.

Paraiso 2006 Paraiso MFR, Barber MD, Muir TW, Walters MD. Rectocele repair: A randomized trial of three surgical techniques including graft augmentation. Am J Obstet Gynecol 2006;195(6):1762-71.

108

Petros 2006 Petros PEP, Richardson PA, Goeschen K, Abendstein B. The tissue fixation system provides a new structural method for cystocoele repair: A preliminary report. Aust N Z J Obstetr Gynaecol 2006;46(6):474-8.

Powell 2004 Powell CR, Simsiman AJ, Menefee SA. Anterior vaginal wall hammock with fascia lata for the correction of stage 2 or greater anterior vaginal compartment relaxation. J Urol 2004;171(1):264-7.

Rodriguez 2005 Rodriguez LV, Bukkapatnam R, Shah SM, Raz S. Transvaginal paravaginal repair of high- grade cystocele central and lateral defects with concomitant suburethral sling: report of early results, outcomes, and patient satisfaction with a new technique. Urology 2005;66(5 Suppl):57-64.

Rozet 2004 Rozet F, Mandron E, Arroyo C, Andrews H, Cathelineau X, Mombet A et al. Laparoscopic sacral colpopexy approach for genito-urinary prolapse: experience with 363 cases. Eur Urol 2005;47(2):230-6.

Safir 1999 Safir MH, Gousse AE, Rovner ES, Ginsberg DA, Raz S. 4-Defect repair of grade 4 cystocele. J Urol 1999;161(2):587-94.

Sand 2001 Sand PK, Koduri S, Lobel RW, Winkler HA, Tomezsko J, Culligan PJ et al. Prospective randomized trial of polyglactin 910 mesh to prevent recurrence of cystoceles and rectoceles. Am J Obstet Gynecol 2001;184(7):1357-62.

Simsiman 2006 Simsiman AJ, Luber KM, Menefee SA. Vaginal paravaginal repair with porcine dermal reinforcement: Correction of advanced anterior vaginal prolapse. Am J Obstet Gynecol 2006;195(6):1832-6.

Vakili 2005 Vakili B, Huynh T, Loesch H, Franco N, Chesson RR. Outcomes of vaginal reconstructive surgery with and without graft material. Am J Obstet Gynecol 2005;193(6):2126-32.

Weber 2001 Weber AM, Walters MD, Piedmonte MR, Ballard LA. Anterior colporrhaphy: a randomized trial of three surgical techniques. Am J Obstet Gynecol 2001;185(6):1299-304.

109

APPENDIX 5 Characteristics of the included studies by study design

Studies included for both safety and efficacy

RCT

Study details Participant characteristics Intervention characteristics Results Author, year Inclusion criteria: patients for Intervention A (n 20): anterior EFFICACY, 12m Al-Nazer 200737 treatment of ≥stage II (POP-Q) vaginal wall colporrhaphy + Objective Failure anterior genital prolapse mesh POP: failure rate (no definition), cystocele, A 1/20; B 6/20 Type of prolapse: anterior Type of mesh: soft Publication type: conference popypropylene mesh, abstract Gynemesh PS, Gynecare, Study design: RCT Ethicon, France Setting: hospital Country: Egypt Intervention B (n 20): anterior Recruitment/Treatment dates: vaginal wall colporrhaphy 110 2003-2005 Length of follow-up: 12m results available Source of funding: NR, no support from industry

Author, year Inclusion criteria: women Intervention A (n 32): anterior, EFFICACY Allahdin 200656 scheduled for anterior or posterior posterior or paravaginal repair + Objective Failure repair for pelvic organ prolapse mesh POP: 3m, residual stage 2 prolapse, A 2/26, B 4/32 Type of prolapse: anterior and/or posterior Number of patients lost to Type of mesh: absorbable, Subjective Failure follow-up: 73 were randomised, 7 polyglactin POP: 6m, symptom score, 7 no symptoms – 35, N, mean (SD), A Publication type: conference were excluded after as they were 27, 10.9 (3.9); B 33, 11.3 (5.4) abstract found to be unfit for surgery or did Intervention B (n 34): anterior, not have a prolapse operation within posterior, or paravaginal repair Quality of life Study design: RCT the study period. Disease-specific: 6m, quality of life not been affected by this Concomitant operation operation, A 14/32, B 14/34 Setting: university hospital Hysterectomy, 14; Recruitment/Treatment dates: Cervical amputation, 18; May 2005-Aug2005 Incontinence: 13 TVT Length of follow up: 6 months Country: UK

Study details Participant characteristics Intervention characteristics Results Additional information: Funding: none Number of surgeon: 10, were trained in the use of mesh if not already familiar with it Author, year Inclusion criteria: patients with a Intervention A (n 54): standard SAFETY Ali 200638 grade 3 or 4 cystourethrocoele anterior colporrhaphy + mesh Mesh erosion: 6m, A 3/46; B not applicable Blood loss: mean, SD, A 64.5 (70.4); B 50.3 (89.0) ml. Type of prolapse: anterior Number of patients lost to Type of mesh: polypropylene All complications: no intraoperative complications in either group. follow-up: 108 were randomly mesh, Gynemesh PS Publication type: conference assigned, 6m, A 46, B 43 available abstract Intervention B (n 54): standard EFFICACY, 6m Study design: RCT anterior colporrhaphy Objective Failure Setting: hospital POP: Baden-Walker, cystocele ≥grade 2, A 3/46; B 5/43 Country: NR Recruitment/Treatment dates: Feb 2004 – Jul 2005 Length of follow-up: 6m results available Source of funding: NR, not supported by industry

111 Author, year Inclusion criteria: patients with Intervention A (n 93): anterior SAFETY, 6-28m Cervigni 200739-41,51,52 recrrent POP stage ≥2 repair + Gynemesh Mesh erosion: erosions, A 6/93, B 0/87

Type of prolapse: anterior Number of patients lost to Type of mesh: Gynemesh follow-up: NR, seems 0 EFFICACY, 6-28m Publication type: conference Age (mean): 64.35 Intervention B (n 87): anterior Objective Failure abstract Parity (mean): 2 repair + Pelvicol POP: cystocele ≥2, A 27/93; B 39/87 Study design: RCT Postmenopausal: 168 Country: Italy Type of grafts: human dermis Recruitment/Treatment dates: Jun 2003-June 2004 Length of follow-up: 6-28m Source of funding: none

Author, year Inclusion criteria: patients with a Intervention A (n 65): anterior EFFICACY, 24m De Ridder 200230,31 stage III cystocele repair (Raz 4 defect repair) with Objective Failure Pelvicol POP: recurrence, A, 6/63; B 19/62 Type of prolapse: anterior Number of patients lost to follow-up: 134 were recruited; Type of grafts: porcine dermis Further need for treatment Publication type: full text (data 24m, A 2/65; B 7/69 Repeat POP surgery: recurred cystocele, A 3; B 14 in De Ridder 2004 [abstract]

Study details Participant characteristics Intervention characteristics Results was used) Age (mean, range): A, 70.09 (24- Intervention B (n 69): anterior 86); B 69.47 (36-83) repair (Raz 4 defect repair) with Study design: RCT Vycril Anterior only / posterior only / Setting: university hospital both: A 55/0/10; B 56/0/13 Type of mesh: polyglactin

Country: Belgium Concomitant operation Hysterectomy, A 38; B 41 Recruitment/Treatment dates: Jan 2001 – Dec 2003 Additional information: Number of surgeon: 3 Length of follow-up: mean, 25m; A, 24.8 (5.3); B, 25.9 (6.4)

Author, year Inclusion criteria: anterior vaginal Intervention A (n 76): standard SAFETY Gandhi 200532,33 wall prolapse to the hymen or anterior coporrhaphy with NR beyond while straining, ≥18y cardaveric fascia lata Type of prolapse: anterior EFFICACY (median of 13m follow up, range 1.4 to 50m) Exclusion criteria: plans for Type of graft: Tutoplast (Mentor Objective Failure Publication type: Full-text pregnancy within 12m Corporation, Santa Barbara, POP (according to Baden-Walker and POP-Q):

112 Calif) cardaveric fascia, saline Stage II anterior wall prolapse or worse: A, 16/76; B, 23/78 Study design: RCT Number of patients lost to follow- washed, inorganic solvents Prolapse to the hymen or beyound: A, 8/76; B, 13/78 up: no return: A, 0; B, 1 used, gamma-irradiated. Setting: University hospital No 1y return: A, 10; B, 10 Subjective Failure Intervention B (n 78): standard A B Country: U.S. Age (mean, SD): anterior coporrhaphy Pelvic pressure Pre- 41/68 48/68 A, 64.9 (11.7); B, 65.5 (11.6) Post- 9/68 18/68 Recruitment/Treatment dates: Concomitant operation: Resolved 32/41 33/48 Jul. 1999 – Nov. 2002 Parity (median, range): Incontinence: A 51/76; B, 43/78 Persistent 9/41 15/48 A,3 (1-10); B, 3 (1-7) Hysterectomy: A 37/76; B, 37/78 De novo 0/27 2/20 Length of follow-up, median Pelvic pain Pre- 23/67 23/67 (range): 13m (1.4-50), n 134 Primary/secondary repair: NR Additional information: Post- 5/67 13/67 ≥1y Number of surgeon: NR, but all Resolved 20/23 18/23 Anterior only / posterior only / surgeries were supervised by a Persistent 3/23 5/23 Source of funding: supported both: A, 1/0/75; B, 5/0/73 single physician. De novo 2/44 8/44 by an unrestricted educational Abdominal pain Pre- 30/69 28/70 grant from Mentor Corp. They Vault: A, 33/76; B, 33/78 Post- 12/69 14/70 had no effort to influence the Resolved 21/30 19/28 trial, presentation of its Persistent 7/24 4/18 outcomes, or publication of the De novo 2/44 5/42 data. Vaginal bulging Pre- 55/67 57/66 Post- 6/67 6/66

Study details Participant characteristics Intervention characteristics Results Resolved 49/55 51/57 Persistent 6/55 6/57 De novo 3/29 0/9 Slow urine stream Pre- 48/68 42/65 Post- 16/68 19/65 Resolved 34/48 28/42 Persistent 14/48 14/42 De novo 2/20 5/23 Postvoid fullness Pre- 53/72 52/76 Post- 22/72 28/76 Resolved 34/53 30/52 Persistent 19/53 22/52 De novo 3/19 6/24 Author, year Inclusion criteria: patients with Intervention A (n 105): anterior SAFETY Hiltunen 200642 symptomatic cystocele of stage ≥II colporraphy + mesh Mesh erosion: minor exposure of mesh, 12m, 17/92 (POP-Q) Infection: postoperative infection: A 1/105; B 4/97 Type of prolapse: anterior Type of mesh: light, self- tailored, monofilament Publication type: conference polypropylene, Parietene light, EFFICACY, 12m abstracts Sofradim, Trevoux, France Objective Failure

113 POP: POP-Q ≥stage II or any symptomatic cystocele, A 7/92; B Study design: RCT Intervention B (n 97): traditional 32/85; anterior colporraphy Mean (SD), Aa, A -2.44 (0.78); B -1.47 (1.43); Ba, A -2.52 (0.89), Setting: hospital B -1.56 (1.46)

Country: Finland Other symptoms not cured Urinary: mean residual urine volumes, A 24.7 (26.7), B 41.5 (55.3) Recruitment/Treatment dates: ml Apr 2003 – May 2005 Sexual: Dyspareunia: A 0/49; B 2/42

Length of follow-up: 12m Further need for treatment results available Repeat POP surgery: A 0/92; B 2/85

Source of funding: NR, not supported by industry

Author, year Inclusion criteria: patients with Intervention A (n 19): anterior SAFETY Hviid 200543 grade ≥2 prolapse (ICS) repair + Pelvicol Blood loss: N, mean, range, A 19, 75 (25-150) ml; B 20, 57 (25- 150) ml Type of prolapse: anterior Number of patients lost to Type of grafts: porcine dermis All complications: no complications were observed. follow-up: 22 had been to the 3m Publication type: conference follow-up, A 10; B 12; still recruiting Intervention B (n 20): traditional

Study details Participant characteristics Intervention characteristics Results abstract Age (mean): 59 (40-84) antetior repair EFFICACY, 3m Objective Failure Study design: RCT POP: no definition, A 0/10; B 0/12; Difference of Aa pre- and post- operation within each group, A 5.5 Setting: University hospital (3-7)cm; B 4 (2-9) (P<0.02)

Country: Denmark Other outcomes Operating time (N, mean, range): A 19, 38 (23-55) min; B 20, 28 Length of follow-up: 3m (17-50) min results available

Source of funding: NR, not supported by industry

Author, year Inclusion criteria: women with Intervention A (n 100): anterior SAFETY Kocjancic 200744-50,54 primary anterior vaginal wall repair + Pelvicol Mesh erosion: in 2y, graft rejection, 1/85 prolapse >stage II Blood loss: mean, SD, A 151 (112); B 167 (96) Type of prolapse: anterior Type of grafts: porcine dermis All complications: no intraoperative complications Number of patients lost to Publication type: conference follow-up: 206 were enrolled, 176 Intervention B (n 106): anterior

114 abstracts were available at 2y (A 85; B 91) repair EFFICACY, 2y Study design: RCT Objective Failure Setting: hospital Primary/Secondary repair: 206/0 POP: unsatisfactory anatomic outcomes at point Ba, A 9/85; B Country: Italy 20/91 Recruitment/Treatment dates: Mar 2003 – Jun 2004 Length of follow-up: mean, 14m; 2y results available Source of funding: None

Author, year Inclusion criteria: women with Intervention A (n 62): traditional SAFETY, 12m Lim 200757,58 symptomatic prolapse >=stage 2 on vaginal repair + Gynemesh PS Mesh erosion: mesh exposure 4 POP-Q examination requiring mesh Type of prolapse: anterior and anterior and posterior vaginal repair posterior Type of mesh: polypropylene EFFICACY Exclusion criteria: cervical or vault Objective Failure Publication type: conference descent beyond the introitus Intervention B (n 60): traditional POP: 6m, stage 0-1 prolapse, A 81%; B 67%; 12m, A 82.5%; B abstracts vaginal repair 66.0% (assume no patients lost follow-up) Number of patients lost to Study design: RCT follow-up: 128 were recruited, 6 Other symptoms not cured were excluded due to not meeting Sexual: 12m, de novo dyspareunia, A 28%; B 25% Setting: hospital criteria.

Study details Participant characteristics Intervention characteristics Results Other symptoms de novo Country: Australia Sexual: 12m, dyspareunia, A 18%; B 21% (not reported number of women free of dyspareunia at baseline) Length of follow-up: 12m results available Satisfaction Visual analogue score >=80/100, 6m, A 89%; B 68%; 12m, A Source of funding: Johnson 88%, B 83% and Johnson, Ethicon (Somerville, NJ) Further need for treatment Mesh removal surgery: 3 Author, year Inclusion criteria: women Intervention A (n 25): posterior SAFETY Lim 200655 requiring posterior vaginal repair site-specific repair + Vypro 2 Mesh erosion: mesh vaginal erosion, A 19%; B N/R

Type of prolapse: posterior Number of patients lost to Type of grafts: combined follow-up: 80 women were EFFICACY Publication type: conference randomised, 65 were available Intervention B (n 9): posterior Objective Failure abstracts site-specific repair + Vicryl POP: POP-Q Ap or Bp >=1cm, NR Study design: RCT Age (mean): A 57.6 (9.9); B 66.8 Setting: hospital (8.8); C 55.4 (12.5) Type of mesh: synthetic Time till failure: median ‘prolapse-free’ duration: A 24.0 (12.8- Country: Australia absorbable 35.2)m, B 3.0 (1.7-4.3) m, C 26 (23.7 – 28.3) m

115 Length of follow-up (median, SD): A 13.6 (9.0); B 12.1 (12.1); Intervention C (n 31): site- Additional information C 12.2 (9.5) specific repair Due to an unacceptable high prolapse recurrence rate at interim analysis, the Vicryl mesh arm was discontinued 6m after Source of funding: NR, not commencement of study. supported by industry

Author, year Inclusion criteria: women with Intervention A (n 98): anterior SAFETY Meschia 200734 stage II or greater anterior vaginal vaginal repair with Pelvicol Vaginal extrusion of the porcine implant: 1m, 1, was removed, wall prolapse (point Ba -1 or implant reinforcement graft culture showed bacterial colonisation Type of prolapse: anterior greater) according to POP-Q Type of graft: Pelvicol (porcine Blood loss, ml (N, mean, SD): A, 151 (112); B, 167 (96) Publication type: Full-text Exclusion criteria: age >80y, any dermis) Other measurements: previous pelvic surgery, diabetes Vaginal bleeding/hematoma, early postoperative, A, 3; B, 0 Study design: RCT mellitus, and collagen disease. Intervention B (n 103): anterior vaginal repair EFFICACY, 1y Setting: University hospital Number of patients lost to follow- Objective Failure up: A, 2; B, 3 Concomitant operation: POP: Country: Italy Incontinence: A, 4; B, 3 Point Ba >-1: A, 7/98; B, 20/103 Age (mean, SD): A, 65 (8); B, 65 Hysterectomy with McCall Stage III: A, 2/98; B, 2/103 Recruitment/Treatment dates: (9) culdoplasty: A+B, 188 Mar. 2003 – Jun. 2004 Parity (median, range): A, 2 (0-6); Subjective Failure

Study details Participant characteristics Intervention characteristics Results B, 2 (0-5) Additional information: POP: Length of follow-up: 1 year BMI (mean, SD): A, 25.8 (4); B, Number of surgeon: NR A B 25.1 (3) Prolase sensation 9/98 13/103 Source of funding: NR, but no Incontinent: A, 22; B, 18 Bladder symptom financial support from the mesh Stress incontinence 10/98 14/103 manufacture Primary/secondary repair: A, Overactive bladder 15/98 18/103 100/0; B, 106/0 Sexual problems Dyspareunia/total sexually active 7/47 5/48 Anterior only / posterior only / Visual analog scale 1-10, mean, SD 1.5 (1.7)1.5 (1.6) both: A, 33/0/67; B, 39/0/67 Further need for treatment Mesh removal surgery: 1m, A 1/98, B 0/103.

Other outcomes Length of hospital stay (mean, SD, range): A,4.4 (1.5) (3-13); B, 4.7 (1.3), (3-10)

Author, year Inclusion criteria: women with Intervention A (n 36): SAFETY, postoperative Meschia 200759 severe genital prolapse (>=stage III transvaginal anterior and Mesh erosion: 1/21 mesh exposure anterior vaginal wall prolapse and posterior mesh repair without

116 Type of prolapse: anterior and >=stage II uterine prolapse) hysterectomy Other: A, 1 persistent voiding difficulty, 3 complaining of perineal posterior planning to go primary surgery pain; B, 1 recurrent urinary tract infection (UIT) And 2 difficulty in Type of mesh: Perigee-Apogee bowel emptying Publication type: conference Number of patients lost to system abstract follow-up: 71 were enrolled, A 36, All complications: there were no intraoperative complications B 35, 3m follow-up visit were Intervention B (n 35): fascial except for 1 patient in group 2 in which a blood loss of 500ml was Study design: RCT available from 38, A 21, B 17 reconstructive surgery with recorded. concomitant hysterectomy Setting: hospital Primary/Secondary repair: all/0 EFFICACY, 3m Country: Italy Other information: Objective Failure A, mean, point Ba, +3.1 (0.9); C +2 POP: no definition, A, 3/21 anterior, 3/21 apical, and 1/21 Recruitment/Treatment dates: (1.3); posterior recurrences; B 2/17 anterior, 1/17 posterior recurrences Jul 2006 – to date B, mean, point Ba, +2.0 (0.9); C +2.2 (1.5) Length of follow-up: at 3m follow-up visit, 38 were available for clinical assessment

Source of funding: NR, not supported by industry

Study details Participant characteristics Intervention characteristics Results Author, year Inclusion criteria: women with Intervention A (n 31): anterior SAFETY, 6m Nguyen 200753 stage >=II anterior vaginal prolapse prolapse repair + polypropylene Mesh erosion: vaginal mesh exposure, 1/31 requiring surgical correction mesh Type of prolapse: anterior Dyspareunia: A 2/21; B 4/25 Number of patients lost to Type of mesh: Perigee Publication type: conference follow-up: A 31, B 32 available at Transobturator Prolapse Repair EFFICACY abstract 6m System Objective Failure Study design: RCT POP: optimal (both points As and Ba were at stage 0), A 17/31, B Setting: hospital Intervention B (n 32): anterior 10/32; Country: US colporhaphy Satisfactory (both points Aa or Ba were at stage I and improved Recruitment/Treatment dates: from preoperative staging), A 13/31; B 15/32; Length of follow-up: 6m Unsatisfactory (either point Aa or Ba was >=stage II, or results available unchanged or worse than preoperative staging): A 1/31; B 7/32 Source of funding: American Medical Systems

Author, year Inclusion criteria: patients Intervention A (n 31): site- SAFETY Paraiso 200636 undergoing surgery for stage II or specific posterior repair with Mesh erosion: no graft exposure or graft complications during greater posterior vaginal wall graft implant follow-up. Type of prolapse: posterior prolapse, >=21y, did not desire

117 future vaginal delivery. Type of graft: porcine-derived, Blood loss: A 200ml (50-3500); B 150ml (50-950); C 150 (50-600) Publication type: Full text acellular collagen matrix graft Exclusion criteria: underwent (Fortagen, Organogenesis, Inc., Intra- or postoperative (short-term) complications: Study design: RCT concomitant colorectal procedures, Canton, MA). A, n 31 B, n 37 C, n 37 allergic to pork products, or Blood loss requiring transfusion 1 3 0 Setting: Clinic unwilling to accept porcine product Intervention B (n 37): posterior Return to theatre 1 0 0 implantation. colporrhaphy Damage surrounding tissue Country: U.S. Bladder 0 0 2 Number of patients lost to follow- Intervention C (n 37): site- Urethra 1 0 0 Recruitment/Treatment dates: up: A, 3; B, 4; C, 0 specific posterior repair Bowel 0 0 0 Jun. 2002 – Dec. 2004 Infection requiring antibiotics 3 4 0 Age (mean, SD): Concomitant operation: Pulmonary 1 0 1 Length of follow-up (mean, A, 60 (11); B, 61 (12); C, 62 (9) Continence: Other measurements SD, range): 17.5 (7)m, 4.4 – A, 15/31; B, 17/37; C, 17/37 Cardiac/myocardial infarction 0 1 0 33.7m Parity (mean, range): Hysterectomy: Ileus 0 1 0 A, 3 (1-6); B, 3 (1-6); C, 3 (1-8) A, 13/31; B, 12/37; C, 14/37 Pelvic abscess 6 4 7 Source of funding: An Urinary tract infection 1 0 0 unrestricted research grant from Postmenopausal: A, 26/31; B, Additional information: Readmissions 0 2 1 Organogenesis Inc. (Canton, 32/37; C, 35/37 Number of surgeon: NR Emergency room evaluations 4 6 5 MA). Organogenesis had no involvement in the design, Primary/secondary repair: NR EFFICACY implementation, analysis or Objective Failure, 1y

Study details Participant characteristics Intervention characteristics Results writing of the manuscript. Anterior only / posterior only / POP-Q point >-2: A, 12/26; B, 4/28; C, 6/27 both: POP-Q point >=0 (to or beyond the hymen): A, 0/12/19; B, 0/17/20; C, 0/11/26 A, 5/25; B 2/28; C, 2/27 POP-Q point, median (range): Vault repair: A, -2 (-3, 1); B, -3 (-3, 0); C, -3 (-3, 0) A, 19/31; B, 22/37; C, 26/37 Subjective Failure, 1y POP (worsening of prolaps symptoms and/or colorectal symptoms using The Pelvic Floor Distress Inventory-20 (PFDI-20)): A, 6/28; B, 5/31; C, 4/29

Other symptoms not cured or de novo Bowel: A B C Defecatory dysfunction Baseline 29 (97%) 28 (80%) 30 (85%) 1y 5 (21%) 9 (32%) 10 (35%) Splinting to defecate Baseline 12 (51%) 15 (43%) 21 (58%) 1y 2 (7%) 6 (19%) 7 (21%)

118 Hard straining Baseline 21 (68%) 24 (67%) 27 (77%) 1y 12 (41%) 11 (35%) 14 (42%) Felling of incomplete emptying Baseline 27 (93%) 22 (62%) 24 (69%) 1y 8 (28%) 14 (45%) 17 (51%)

Sexual: A B C Sexually active Baseline 17 (53%) 17 (46%) 18 (46%) 1y 16 (57%) 19 (61%) 21 (63%) PISQ-121 Baseline, mean (SD) 33 (8) 29 (8) 31 (8) 1y, mean (SD) 37 (5) 36 (5) 36 (7) Dyspareunia, never/seldom Baseline 12 (50%) 10 (45%) 12 (52%) 1y 16 (84%) 11 (55%) 16 (72%) Dyspareunia, sometimes Baseline 12 (50%) 6 (26%) 9 (39%) 1y 2 (11%) 5 (25%) 3 (14%)

Study details Participant characteristics Intervention characteristics Results Dyspareunia, usually/always Baseline 0 2 (9%) 7 (30%) 1y 1 (6%) 3 (14%) 4 (20%) 1 POP/Urinary Incontinence Sexual Questionnaire short form, range 0-48, higher score, better sexual function

Quality of Life Disease specific: A B C PFIQ-71 Baseline, N, mean (SD) 31, 63 (64) 35, 65 (69) 36, 87 (66) 1y, N, mean (SD) 24, 10 (23) 28, 10 (18) 29, 22 (38) 2y, N, mean (SD) 13, 5 (13) 15, 16 (32) 17, 16 (31) PFDI-202 Baseline, N, mean (SD) 31, 116 (55) 35, 114 (56) 36, 146 (66) 1y, N, mean (SD) 24, 34 (37) 28, 39 (30) 29, 46 (53) 2y, N, mean (SD) 13, 32 (33) 15, 44 (32) 17, 53 (46) 1The Pelvic Floor Impact Questionnaire-7, range 0-300, higher scores indicating greater adverse impact on QoL. 2The Pelvic Floor Distress Inventory-20, range 0-300, higher

119 scores indicating greater distress.

Satisfaction Global index of improvement at 1y compare with before treatment, scale 1-7: A B C Much better or very much better 90% 74% 88% Somewhat better 7% 19% 12%

Further need for treatment The re-operation rate for pelvic organ prolaps at 2y (any segment): A, 3/29; B, 1/33; C, 2/37

Other outcomes Operating time, min (N, mean, SD): A, 31, 169 (62); B, 37, 150 (68); C, 37, 151 (69)

Length of hospital stay(N, median, range): A, 31, 2 (1-6); B, 37, 2 (1-19), C, 2 (1-7)

Author, year Inclusion criteria: cystocele Intervention A (n 73): anterior SAFETY Sand 200127 (Grade 2 to hymen), age >18y, colporraphy with or without 1 year, no adverse effects

Study details Participant characteristics Intervention characteristics Results ambulatory, willing to comply with posterior colporraphy + Type of prolapse: anterior or return visits. polyglactin mesh reinforcement posterior, reported separately EFFICACY, 1y Exclusion criteria: pregnancy or Type of mesh/grafts: Objective Failure Publication type: full-text contemplating pregnancy in the next polyglactin 910 mesh (Vicryl POP II or greater (Baden Walker) 12m, anterior enterocele or mesh, Ethicon, Somerville, NJ), Anterior POP: A, 18/73, B, 30/70 Study design: RCT paravaginal defect alone absorbable Posterior POP: A, 6/65; B, 7/67

Setting: University hospital Number of patients lost to follow-up: A, 6; B, 11 Intervention B (n 70): Country: US Age (mean): A, 65.0; B, 62.5 Standard anterior colporrhaphy Parity (mean): A 2.8; B, 2.8 with or without posterior Recruitment/Treatment dates: BMI: NR colporrhaphy Sep. 1995 – Apr. 1999 Postmenopausal: A, 60; B, 54

Length of follow-up: 1 year Primary/Secondary repair: A, Concomitant operation 55/18; B, 49/21 Hysterectomy: A, 36/73; B, Source of funding: NR 39/70 Anterior only / posterior only / Continence procedure: A, 58/73; both: A, 8/0/65; B, 3/0/67 B, 52/70

120 Additional information: Number of surgeon: NR Author, year Inclusion criteria: patients Intervention A (n 35): standard SAFETY Weber 200135 underwent operation for anterior anterior coporrhaphy + mesh Mesh erosion: timing not stated, 1/35 vaginal prolapse Blood loss requiring transfusion: A, 0/35; B, 1/39; C, 0/35 Type of prolapse: anterior Type of mesh: polyglactin 910 PE: postoperative, A, 0/35; B, 0/39; C, 1/35 Exclusion criteria: an anti- mesh, Vicryl, Ethicon, Publication type: full-text incontinence procedure other than Somerville, NJ EFFICACY, 1y suburethral placation (i.e. Burch Objective Failure Study design: RCT colposuspension, sling, or needle Intervention B (n 39): Standard POP: suspension) was planed as part of colporrhaphy Either point Aa or Ba stage II or greater: A, 15/26; B, 23/33; C, Setting: clinic their surgical procedures. 13/24 Intervention C (n 35): Country: U.S. Number of patients lost to follow- ultralateral anterior colporrhaphy Further need for treatment up: 1y, A, 9/35; B, 6/39; C, 11/35 Mesh removal surgery: timing not stated, 1/35 Recruitment/Treatment dates: Concomitant operation: Jun. 1996 – May 1999 Age (mean, SD): A, 66 (11.2); B, NR 65.6 (11.2); C, 62.4 (13.3) Length of follow-up (mean, Additional information: range): 23.3m (4.5-44.4), 1y Postmenopausal: A, 31/35; B, Number of surgeon: 5 results available for n 83 34/39; C, 28/35

Study details Participant characteristics Intervention characteristics Results

Source of funding: The American College of Obstricians and Gynecologists/Ethicon Research Award for Innovations in Gynecologic Surgery, and by the Department of Gynecology and Obstetrics at the Cleveland Clinic Foundation

121

Non-randomised comparative studies

Study details Participant characteristics Intervention characteristics Results Author, year Inclusion criteria: patients Intervention A (n 17): rectocele SAFETY Altman 200464 referred for the treatment of repair + porcine collagen mesh Mesh erosion: NR symptomatic posterior vaginal wall Infection: urinary tract infection, A 1/17; B 2/15 Type of prolapse: posterior prolapse, stage >= grade III Type of grafts: porcine dermis, Other: conservative managed bleeding from the wound edge, A 2; Pelvicol (C R Bard, Murray Hill, B 0; rectal emptying difficulties, A 0; B 1 Publication type: full-text Exclusion criteria: recurrent NJ, USA) rectocele, systemic inflammatory All complications: there were no perioperative mortalities, major Study design: non-randomised disease, connective tissue Intervention B (n 15, age- complications or readmissions in either of the two groups within comparative study disorders, diabetes, impaired matched): posterior colporraphy 30d of surgery. immune system disease, and Setting: hospital systemic steroid treatment. Type of mesh/grafts: traditional rectocele repair EFFICACY, 6m Country: Sweden Number of patients lost to Objective Failure follow-up: NR, retrospective study Concomitant operation POP: A, stage I 15, stage II 2; B, stage I 14, stage II 1. Recruitment/Treatment dates: Age (mean, range): A, 60.3 (42- Hysterectomy: A, 2; B, 2 NR 75); B, 58.6 (43-68) Cervical amputation: A, 2; B, 1 Length of follow-up: NR, 6m Parity (mean, range): A, 2.1 (0-4); Other outcomes results available B, 2.0 (0-3) Additional information: Length of hospital stay(N, mean, range): A, 17, 3d (2-6); B, 15,

122 Number of surgeon: NR 4.1d (2-6) Source of funding: NR, the Postmenopausal: A, 15; B, 11 mesh manufacture company had no influence over the Anterior only / posterior only / Additional information design or execution of this study both: A, 0/16/1; B, 0/8/3 There were no significant changes in macrophage count, nor did they provide any inflammatory cell count, total mesh-tissue interface cell count, or in financial support Other information: histopathological inflammatory grading when comparing HRT: A 12, B NR. postoperative with preoperative values (data obtained by biopsy). Additional information on study design: participants in the control group was randomly selected women undergoing traditional posterior colporraphy at the same clinic 1y prior to the study, i.e. historical study

Author, year Inclusion criteria: NR Intervention A (n 5): posterior SAFETY Castelo-Branco 199865 wall repair (‘usual procedure’) + No complications occurred during surgery; no febrile morbidity, Exclusion criteria: NR mesh cuff or posterior vaginal wall infections, thrombophlebitis, rectal Type of prolapse: posterior injury, or hemorrhagic complications Number of patients lost to follow- Type of mesh: polyglacolic acid Publication type: full-text up: A, 0; B, 0 mesh, absorbable, no other Other: 1m

Study details Participant characteristics Intervention characteristics Results information available Posterior wall granulomata: A, 2/5; B, 1/5 Study design: case-control Age (mean, SD, range): A, 57 Mild vulvar hematoma: A, 0/5; B, 1/5 (7.1), 46-64; B, 55.6 (8.2), 43-65 Intervention B (n 5): posterior Setting: hospital clinic wall repair (‘usual procedure’) EFFICACY, 1y Parity: Objective Failure Country: Spain 1: A, 1; B 0 POP: A, 0/5; B, 0/5, all rectum remained in a normal position, 2: A, 2; B, 2 Concomitant operation: confirmed by Urethro-Cysto-Vagino-Rectography Recruitment/Treatment dates: >=3: A 2; B, 3 Incontinence: A, 1; B, 1 Jan. – Sep. 1996 Hysterectomy: A, 1; B, 3 Other symptoms not cured Primary/secondary repair: NR Sexual: Length of follow-up: data Additional information: Unpleasant sexual intercourse: A, 0/5; B, 2/5 available for 1m and 1y Anterior only / posterior only / Number of surgeon: NR both: A, 2/3/2; B, 4/1/4 Source of funding: NR

Author, year Inclusion criteria: NR Intervention A (n 14): anterior SAFETY, 2y Chaliha, 200660 repair with small intestine Mesh erosion: A, 0/14; B, 0/14 Exclusion criteria: NR submucosa graft Blood loss: at or post-operation, there were no significant Type of prolapse: anterior complications such as excessive bleeding (assume equivalent to Number of patients lost to follow- Type of grafts: small intestine no blood transfusion).

123 Publication type: full-text up: NR submucosa (no other Infection: A, 0/14; B, 0/14 information available) Study design: non-randomised Age (mean, range): A, 70 (51-86); EFFICACY comparative study B, 60 (47-79) Intervention B (n 14): Objective Failure traditional anterior repair POP: POP-Q, mean score Setting: teaching hospital Primary/secondary repair: A, 12/2; A B B, 12/2 Concomitant operation: Pre- 6m 2y Pre- 6m 2y Country: UK Incontinence: A, 0; B 0 AA 1.14 -2.07 -1.04 1.61 -1.32 -0.75 Anterior only / posterior only / BA 1.64 -2.07 -1.07 2.25 -1.14 -0.61 Recruitment/Treatment dates: both: A, 14/0/0; B, 14/0/0 Additional information: C -2.93 -5.57 -4.86 1.71 -3.93 -5.21 2001 – 2003 Number of surgeon: 1 AP -0.64 -2.82 -2.54 0.21 -2.32 -1.86 BP -0.04 -2.82 -2.40 1.14 -1.82 -1.79 Length of follow-up: 6m and D -3.64 -0.43 -0.43 0.32 -0.14 0 2y results available TVL 6.93 5.57 5.36 4.29 6.57 6.68 GH 4.18 2.86 2.57 1.46 3.29 3.04 Source of funding: NR PB 3.14 4.07 3.57 3.21 3.64 3.21 TVL: total vaginal length; GH: genital hiatus; PB: perineal body

Quality of Life Disease specific: Validated prolapse QoL questionnaire, mean of score A B

Study details Participant characteristics Intervention characteristics Results Pre- 6m 2y Pre- 6m 2y General health perception 32 16 20 41 34 36 Prolapse impact 81 5 14 62 14 14 Role limitations 65 4 10 35 17 8 Physical limitations 58 5 1 38 14 1 Social limitations 36 3 -2 20 9 2 Personal relationships 65 7 -1 52 4 0 Emotions 64 5 10 42 13 -1 Sleep/energy 58 17 15 36 23 5 Severity measures 42 5 4 36 11 3 Author, year Inclusion criteria: patients Intervention A (n 56): cystocele SAFETY Handel 200761 underwent cystocele repair repair + porcine dermal graft Mesh erosion: mesh extrusion, A 12/56; B 1/25 Blood loss requiring blood transfusion: A 1/56, B NR, C NR Type of prolapse: anterior Exclusion criteria: NR Type of grafts: NR trade name Return to theatre (short-term): A 1/56, B NR, C NR. Infection: mesh infection, A 1/56, B NR Publication type: full-text Number of patients lost to Intervention B (n 25): cystocele follow-up: 20/119 follow-up not repair + polypropylene mesh Other: vaginal incision dehiscence, A 6/56 (1 also had graft Study design: non-randomised available extrusion, recurrent cystocele, blood loss requiring transfusion; 1 comparative study Type of mesh: NR trade name return to theatre for removal of remaining porcine mesh; 1 mesh Primary/Secondary repair: A, infection, the vaginal epithelium was debrided and any remaining th 124 Setting: University hospital 36/20; B, 24/1; C, 17/1 Intervention C (n 18): porcine graft was exicised; the 4 one had the remaining graft traditional repair removed in the office, 5th and 6th required no further surgical Country: US Anterior only / posterior only / treatment), B NR, C NR. both: A,18/0/38; B, 7/0/18; C, Concomitant operation Recruitment/Treatment dates: 6/0/12 Incontinence: A 48, B 20, C 9, Jan 1999 – Aug 2005 pubovaginal sline (mersilene EFFICACY mesh, polypropylene mesh, or Objective Failure Length of follow-up (mean, fascia lata) POP: Baden-Walker, cystocele >=stage 2: A 20/56; B 1/25; C 1/18 range): all, 13.5m (2-46); A, 17m; B, 13m; C 9m Hysterectomy: A 46, B 25, C 18 Time till failure: all, mean 4.9m (0.5-20)

Source of funding: NR Additional information: Further need for treatment Number of surgeon: NR Mesh removal surgery: A 3/56 partial removal Additional study on study design: in the early series traditional anterior colporrhaphy was performed but it was gradually abandoned in favour of a porcine dermal graft. The most recent cases involved the use of polypropylene only for cystocele repair, i.e. historical

Study details Participant characteristics Intervention characteristics Results study

Author, year Inclusion criteria: patients with 2 Intervention A (n 12): support SAFETY Julian 199662 or more postsurgical recurrences of defects repair with mesh Mesh erosion: graft related complications, <=6m, 3/12 (1, 0.5cm severe anterior vaginal wall opening below the graft; 2, 3mm area of granulation tissue over Type of prolapse: anterior prolapse and had III or IV degree Type of mesh: Marlex the graft; 3, two tiny perforating mesh fibers) defect of the anterior vaginal polypropylene mesh (Bard Publication type: full-text segment Vascular System Division, CR Infection: inhospital infection, A, 0/12; B, 0/12 Bard, Billerica, Mass.) Study design: case-control Exclusion criteria: NR Any complications: there were no significant intraoperative Intervention B (n 12): support complications in either group. Setting: University hospital Number of patients lost to follow- defect repair: urethropexy, up: NR anterior colporrhaphy, Country: US paravaginal repair. EFFICACY Age (mean, range): A, 63 (37-82); Objective Failure, 2y Recruitment/Treatment dates: B, 66 (46-78) POP: recurrent anteior vaginal wall prolapse (>= II degree), A, Jan 1989 – Dec 1992 Parity (mean, SD): A, 3.2 (0-8); B, Concomitant operation: 0/12; B, 4/12 (2 II degree, 2 III degree). 3.3 (0-11) Length of follow-up: 2y results Further need for treatment available Primary/secondary repair: A, 0/12; Additional information: Mesh removal surgery: 1/3 erosions B, 0/12 Number of surgeon: NR 125 Source of funding: NR

Author, year Inclusion criteria: women Intervention A (n 24): 4-defect SAFETY Leboeuf 200463 underwent grade 4 cystocele repair repair + Pelvicol Mesh erosion: erosion, A, 0; B, 0 Other organ damage: bladder perforation, A, 0; B 0 Type of prolapse: anterior Exclusion criteria: NR Type of grafts: xenograft Infection: A, 0; B, 0 collagen matrix, Pelvicol, Bard Other: fistula, vaginal stenosis, A, 0; B, 0; Publication type: full-text Number of patients lost to Postoperative urinary retention, A, 1/24; B, 1/19 follow-up: 2/45 not available for Intervention B (n 19): 4-defect Study design: case-control follow-up repair Age (mean, range): 65 (33-91) EFFICACY Setting: university hospital Additional information: Objective Failure Anterior only / posterior only / Number of surgeon: 1 POP: A, 3 (2 asymptomatic grade 2 and 1 symptomatic grade 3); Country: US both: 8/0/35 B, 0;

Recruitment/Treatment dates: Vault only / uterine only / both: Subjective Failure Oct 1998 – Oct 2002 10/43 had modified McCall POP: A, 1 who was symptomatic; B, 0 Length of follow-up (mean): procedure 15m Other symptoms not cured Other information: Urinary: SUI, A, 13/14 who had SUI; B, 9/10 who had SUI; Source of funding: NR SUI, 24/43; Urge incontinence, A 10/14 who had urge incontinence; B, 3/12

Study details Participant characteristics Intervention characteristics Results Urge incontinence, 26/43 who had urge incontinence; SEAPI scores of zero (stress, emptying, anatomy, protection, instability): A B Pre- Post- Pre- Post- Stress 14 22 5 17 Emptying 8 21 8 18 Anatomy 0 22 0 18 Protection 16 20 7 13 Instability 12 15 5 13

Other symptoms de novo Urinary: SUI, total, 2/19 who had no SUI; urge incontinence, 2/17 who had no urge incontinence

Author, year Inclusion criteria: all patients Intervention A (n 98): anterior SAFETY, 3-67m Vakili 200566 underwent vaginal reconstruction or posterior graft-augmented Mesh erosion: 25 (28.4%) with 18 requiring debridement, the surgery for prolapse during the remaining 7 erosions were treated successfully with topical Type of prolapse: anterior indicated time interval Type of mesh/grafts: estrogen alone. and/or posterior 1, Freeze-dried cadeveric fasica

126 Exclusion criteria: procedure was lata (banked); 2, solvent-dried Infection: vaginal/graft infection, A 18/98 (18.4%); B 10/214 (4.7%) Publication type: full-text obliterative (9.e. colpocleisis), cadeveric fascia lata (Tutoplast, operative report not available, Mentor, Santa Barbara, CA); 3, Other: granulation tissue, A 38/98 (38.8%); B 37/214 (17.3%) Study design: case-control follow-up <3m Solvent-drived cadaveric fascia lata (Intaxen, American Medical EFFICACY, 3-67m Setting: University hospital Number of patients lost to follow- Systems, Inc. Minnetonka, MN); Objective Failure up: NR, retrospective, follow up 4, porcine small subintestinal POP/ de novo POP: Country: US <3m were excluded already submucosa (SIS, Cook Any descent of any compartment of the vagina below the normal Biotechnology Inc. West anatomic position, i.e. >0 by the Baden- Walker halfway system), Recruitment/Treatment dates: Age (mean): A 65.4; B 60.7 Lafayette, IN); 5, Acellular A 34/98 (34.7%); B 91/214 (42.5%) Feb 1998 – Jun 2004 Parity (mean): A 3.3; B 2.9 porcine dermis (Pelvicol, Bard, BMI (mean): A 26.3; B 26.1 Covington, GA); 6, autologus Stage III prolapse: A 2/98 (2.0%); B 6/214 (2.8%) Length of follow-up (mean, Postmenopausal: A 89 (92.7%); B fascia lata; 7, polypropylene range): 9m (3-67) 173 (83.2%) (Gynemesh, Gynecare Other symptoms not cured worldwide, New Brunswick, NJ) Urinary: subjective incontinence, A 25/98 (25.5%); B 59/214 Source of funding: NR Anterior only / posterior only / (27.6%) both: A, 74/22/0; B, NR Intervention B (n 214): restorative repairs Further need for treatment, only if the additional surgery was Vault only / uterine only / both: performed by one of the two primary surgeons vault or uterine, A 62 (63.2%); B Concomitant operation: 124 (57.9%) Incontinence, A 66 (67.3%); B Repeat POP surgery:, A 8 (8.2%); B 20 (9.3%) (unclear de novo 142 (66.4%) or not)

Study details Participant characteristics Intervention characteristics Results Additional information: Hysterectomy, A 7 (7.1%); B 23 Hormone replacement therapy: A (10.7%) Incontinence operation: A 5/98 (5.1%); B 9/214 (4.3%) 61 (70.1%); B 131 (76.6%) Additional information: POP-Q measurements (cm), mean: Number of surgeon: 2 A B Aa 1.0 0.6 Ba 3.2 2.7 C 1.6 0.2 GH 6.0 5.2 PB 3.3 3.9 TVL 7.5 8.3 Ap -0.4 -0.6 Bp 1.2 0.7

127

Case series (follow-up >=1y, n>=50)

Study details Participant characteristics Intervention characteristics Results Author, year Inclusion criteria: patients Intervention A (n 76): 52 SAFETY Amrute 200781 underwent tension-free 4-point underwent tension-free placement Mesh erosion: vaginal erosion, 2/76, occurred at 7m and 12m, fixation for PoP with the BioArc device, 24 symptoms were consisted of vaginal discharge, sizes ranged Type of prolapse: anterior and underwent bone anchor from 1-1.5vm posterior Exclusion criteria: NR placement Other: obstruction, 15m, 1/76 (1.1%); palpable retained vaginal Publication type: full-text Number of patients lost to Type of mesh: polypropylene, suture, 1/76 (1.1%) follow-up: 20/96 not available for BioArc device, American Medical Study design: retrospective follow up Systems, Minnetonka, Minnetonka case series Age (mean, SD): 69.3 (11.3) EFFICACY, mean 30.7 (1.7) Concomitant operation Objective Failure Setting: medical centre Primary/Secondary repair: NR Hysterectomy, 36/76 (47.4%) POP: no defination, but all asymptomatic, 4/76 (5.2%)

Country: US Anterior only / posterior only / Additional information: Other symptoms not cured both: NR Number of surgeon: NR Urinary: recurrent SUI, 2 (2.1%); Recruitment/Treatment dates: For those with preoperative incontinence (n 36), average pad Jan 2000 – Jun 2005 Vault only / uterine only / both: used/d: pre- 2.1 (0.4), post- 0.8 (0.2) Length of follow-up (mean, 0/0/76

128 SD): 30.7m (1.7) Sexual: dyspareunia, 2/21 (9.6%) who sexually active Other information: Source of funding: NR Incontinence at baseline: 36/76 Other symptoms de novo Urinary: urgency, 12 (15.7%) of whom 6/40 were incontinent

Satisfaction Scale 1-10 (most satisfied), mean (SD): 7.9 (0.3)

Further need for treatment Mesh removal surgery: for 2 mesh erosion, excision of the exposed mesh; for the retained suture, simple excision. Other: for 2 recurrent SUI

Additional information: Baseline PoP-Q examination was not recorded, so unclear whether all women had cystocele and vault prolapse

Author, year Inclusion criteria: patients with Intervention A: A, SAFETY Costantini 200582 symptomatic uterovaginal colposacropexy with uterus Mesh erosion: mesh erosion, A 0; B 3 prolapse grade III-IV conservation (n 34); B, Blood loss: median, intraoperative, A 200ml, B 325ml; requiring Type of prolapse: anterior and hysterectomy followed by transfusion, A 2; B 2 posterior Exclusion criteria: for uterine sacropexy (n 38)

Study details Participant characteristics Intervention characteristics Results conservation group, no post- Infection: would infection, A 0; B 2 Publication type: full-text menopausal bleeding, no previous Type of mesh: Marlex, no further Other: fever, A 1; B 1; perivesical haematoma, A 2; B 4; voiding cervical intraepithelial neoplasia, details dysfunction, A 4; B 1; incisional hernia, A 2; B 1 Study design: prospective case no abnormal cervical smears or series uterine disease including uterine Concomitant operation enlargement or cervical ulceration. Incontinence: A, 28/34; B 30/38 EFFICACY Setting: university hospital anterior colposuspension; Objective Failure Number of patients lost to Hysterectomy: A, 0; B 38 POP: cystocele, A 5/34 asymptomatic grade 2 (in the first 8 Country: Italy follow-up: A, 2 did not return after cases of this series, only the posterior rectangular mesh was 1 and 2y, but were included in the Additional information: used and central cystocele recurred in 5, 4 of them are in A); B Recruitment/Treatment dates: analysis; B, 1 died 3y from cancer Number of surgeon: NR, all 1/38; Jun 1995 – Dec 2003 unrelated to surgery, but was surgery was performed by or Cervical/vault prolapse (<6m above the hymen): A 0; B0 Length of follow-up (mean, included in the analysis. under the suspension of the A B range): 51m (12-115) Age (mean, SD): A, 61 (12); B 62 senior author. Pre- Post- Pre- Post- (8) C, cm -2 (-6, 8) -6 (-8.5, -3) -2 (-6.5, 7) -5.5 (-7, -4) Source of funding: NR Parity (median, range): A, 2 (1- D -2.7 (-7, 10) -8 (-9, -6) -3 (-7.5, 10) - 8); B 2 (0-3) Total vaginal 8 (6, 11) 8 (6, 11) 7 (6, 10) 6.5 (6, 10) BMI (median, range): A, 24.4 length (18.6-31.2); B 25.4 (16-35.1) Aa 2.2 (-2, 3) -2 (-3, -1) 2.5 (-2.5, 3) -2.2 (-3, 0) Postmenopausal: A, 26; B, 31 Ba 3 (-1, 8) -4 (-6.5, -1) 3 (-1, 7) -3 (-5, -1)

129 Anterior only / posterior only / Other symptoms not cured both: A, 0/0/34 all; B, 0/0/38 all Urinary: A, 4/22; B 7/20 who had incontinence

Vault only / uterine only / both: Other symptoms de novo A, 34; B 38 Urinary: A 1/22; B 0/18 who had no incontinence

Other information: Satisfaction A B Satisfied and would repeat surgery again: A 31/34; B 33/38 Uterine prolapse Grade 1-2 20 27 Further need for treatment Grade 3 14 11 Repeat POP surgery: A 0; B0 Cystocele Mesh removal surgery: A 0; B 3 vaginal revision for mesh Grade 1-2 3 1 erosion Grade 3 31 37 Other outcomes Operating time (N, mean, range): A, 34, 89min (60-110); B, 38, 115min (80-135)

Author, year Inclusion criteria: patients with Intervention A (n 50): SAFETY Cronje 200667 stage IV cystocele sacrocolpopexy All complications: no major perioperative complications occurred.

Study details Participant characteristics Intervention characteristics Results Type of prolapse: anterior Exclusion criteria: NR Type of mesh: combined, Vypro, Johnson & Johnson, Brussels, EFFICACY Publication type: full-text Number of patients lost to Belgium Objective Failure follow-up: NR POP: grade 3 and 4 cystoceles, 8/50 Study design: retrospective Age (median): 65 Additional information: de novo POP: grade 3 or 4 rectocele, 1/50; grade 3 or 4 vaginal case series Parity (median): 3 Number of surgeon: NR, but vault prolapse, 1/50 patients were consecutively Setting: university hospital Other information: recruited Additional information Urinary symptoms: 15/50 No cystocele reoccurred in the patients in whom the mesh Country: South Africa supporting the bladder extended from the urethra to the sacrum, i.e. complete support. Recurrences occurred only when the Recruitment/Treatment dates: mesh extended from the mid-bladder base to the sacrum, i.e. NR partial support. Length of follow-up (median, range): 12m (1-50)

Source of funding: University

Author, year Inclusion criteria: at least a Intervention (n 143): SAFETY De Tayrac 200783 symptomatic vaginal wall prolapse Anterior and/or posterior repair Mesh erosion: 9/143 (5/9 within 3m, 8/9 anterior, 5/9

130 at stage 2 to 4 in the International with mesh symptomatic) Type of prolapse: anterior POP Staging System (Ba or Bp and/or posterior >=-1) and an impairment on her Type of mesh: Ugytex, Sofradim, Other organ damage: intraoperative, 3/143 bladder injuries quality of life. France, distributed by Bard as during cyctocele dissection, 1 rectal injury during the rectocele Publication type: full-text Pelvitex, low weight (38g/m2), dissection, 1 uterine artery haemorrhage during hysterectomy, 1 Exclusion criteria: NR highly porous (89% of average vaginal sulcus perforation during a transobturator tape procedure Study design: prospective case porosity, pores >1.5mm), series Number of patients lost to polypropylene monofilament mesh Other: postoperative (<=1m), 1 urinary retention after a follow-up: 0 coated with a hydrophilic film transobturator tape procedure, 2 vaginal haematomas occurred Setting: 3 public and 6 private composed of atelocollagen, at 2d and 7d after operation respectively. hospitals, and 4 university Age (mean, SD, range): 63 polyethylene glycol and glycerol. hospitals (10.7), 37-91 Concomitant operation: EFFICACY, mean 13m (10-19) Country: France Parity (mean, SD, range): 2.6 Hysterectomy/cervical amputation: Objective Failure (1.4), 1-11 52/3 POP-Q >=II stage: cyctocele 9/132; rectocele 2/76 Recruitment/Treatment dates: Mar 2003 – June 2004 Anterior only / posterior only / Additional information: Subjective Failure & other symptoms not cured both: 67/11/65 Number of surgeon: 18 surgeons Pelvic Floor Distress Inventory (PFDI): Length of follow-up (mean, experienced in vaginal prolapse Pre- Post- SD, range: 13 (2)m (10-19) Vault or uterine repair: 41/143 repair Urinary Distress Inventory (UDI) (/300), mean 84.9 22.6 Colo-Recto-Anal Distress Inventory (CRADI) 80.4 35.7 Source of funding: NR (/300), mean

Study details Participant characteristics Intervention characteristics Results POP Distress Inventory (POPDI) (/300), mean 110.9 32.9

Sexual: Dyspareunia/sexually active: preoperative 10/88, postoperative 1/77

Other symptoms de novo Sexual: dyspareunia/sexually active no dyspareunia preoperation 10/78

Quality of Life Disease specific: Pelvic Floor Impact Questionnaire (PFIQ): Pre- Post- Urinary Impact Questionnaire (UIQ) (/300), mean 77.1 19.2 Colo-Recto-Anal Impact Questionnaire (CRAIQ) 32.2 9.6 (/300), mean POP Impact Questionnaire (POPIQ) (/300), mean 48.2 7.8

Satisfaction

131 ‘Satisfied’ or ‘very satisfied’ with the functional results: 138/143

Further need for treatment Mesh removal surgery: 6/9 erosions

Other outcomes Operating time (N, mean, range): 143, 81min (20-180) Length of hospital stay(N, mean, range): 143, 4.5d (1-23)

Author, year Inclusion criteria: women with Intervention A (n 55): trans- SAFETY De Tayrac 200668-71 symptomatic stage 2 to 4 vaginal cystocele repair using a Mesh erosion: vaginal erosions, all between 6w-6m, 5/55 (9.1%), cystocele who underwent the tension-free polypropylene mesh 4 out of 5 were symptomatic including bloody vaginal discharge Type of prolapse: anterior same trans-vaginal procedure and dyspareunia Type of mesh: Gynemesh, Publication type: full-text Exclusion criteria: NR GyneCare, Ethicon, France Blood loss: requiring transfusion, postoperative, 0/55 Other organ damage: bladder injury, intraoperative, 1/55 Study design: retrospective Number of patients lost to Concomitant operation: case series follow-up: 8, not returned for Incontinence: 22/63 (34.9%) Infection: urinary tract infection, postoperative 4/55; mesh follow-up Hysterectomy: 52/63 (82.5%) infection, 0/55 Setting: hospital Age (mean, SD, range): Additional information: Other: local pain around a mesh shrinkage, 3/55 (5.5%); one day Country: France 62.7 (10.8), 39-83 Number of surgeon: 2 fever, 2/55

Study details Participant characteristics Intervention characteristics Results

Recruitment/Treatment dates: Parity (mean, SD, range): Oct 1999 – Oct 2002 2.8 (1.4), 1-6 EFFICACY, mean 37m (24-60) Objective Failure Length of follow-up (mean, Postmenopausal: 54/63 (85.7%) POP: stage 2 anterior wall prolapse, asymptomatic, 5/55 (9.1%); SD, range): 37 (9.9), 24-60m Incontinent: SUI 25/63 (39.7%) stage 3, symptomtaic, 1/55 (1.8%)

Source of funding: NR Primary/secondary repair: 59/4 de novo POP: rectocele, 8/45 (6.6%) Subjective Failure Anterior only / posterior only / POP: symptomatic, 1/55 (1.8%) who had stage 3 failure both: 45/0/10 (18.1%) Other symptoms de novo Vault only / uterine only / both: Urinary: urgencies, 3/55 vault or uterine 51/63 (80.9%) Sexual: dyspareunia, 4/24 (16.7%) who were sexually active postoperation Additional information: Hormone replacement therapy: Quality of Life 19/54 (35.2%) who were General: significant improvement in QoL (satisfaction), 53/55 postmenopausal (96.4%)

132 Stage 2 anterior prolapse, 9/63 Further need for treatment (14.3%); stage 3, 42/63 (66.7%); Mesh removal surgery: 6w-6m, 4/5 erosions (3 partial excision, stage 4, 12/63 (19%) 1 complete excision)

Other outcomes Operating time (N, mean, range): 55, 102min (50-150) Length of hospital stay(N, mean, range): 55, 3.8d (2-7)

Author, year Inclusion criteria: women had Intervention A (n 97): SAFETY Dwyer 200484 vaginal prolapse surgery with Mesh erosion: 9 (9%), of whom 7 occurred <=6m; 3 were Atrium mesh reinforcement for Type of mesh: polypropylene, asymptomatic and healed after intravaginal oestrogen cream, the Type of prolapse: anterior recurrent or large anterior and Atrium, Hudson, New Hampshire, other 6 had symptoms of vaginal discharge, bleeding or and/or posterior posterior compartment vaginal US dyspareunia; 3/9 were on the anterior wall and the other 6 on prolapse (Baden-Walker >=2); posterior wall. Publication type: full-text indication: recurrent vaginal Concomitant operation prolapse or a large fascia defect Incontinence: 24 tension-free Other: postoperative haemorrhages (>500ml), 2/97; rectovaginal Study design: retrospective unsuitable for standard repair vaginal sling fistula, 5th d, 1/97 case series alone Hysterectomy: 10

Setting: hospital Exclusion criteria: NR Additional information: EFFICACY, >24m Number of surgeon: NR Objective Failure

Study details Participant characteristics Intervention characteristics Results Country: Australia Number of patients lost to Mesh was placed under the POP: grade 2 asymptomatic cystocele, 4/64 who had anterior follow-up: 97 were reviewed at bladder base with lateral mesh placement; cystocele >=grade 2, 2/17 who had both Recruitment/Treatment dates: 6w, 96 at 6m, 91 at 12m, 77 >24m extensions onto the pelvic anterior and posterior repair; posterior prolapse >=2, 0/50 who Feb 1999 – May 2002 Age (mean, range): 61 (30-86) sidewall: 47; had posterior repair Length of follow-up (mean, Parity (mean, range): 3 (0-8) Y-shaped mesh was placed from range): 29m (6-52) the sacrospinous ligaments to the de novo POP: 2/33 anterior prolapse in those had posterior Anterior only / posterior only / perineal body: 33; repair; 1/64 cuff prolase in those had anterior repair; 4/64 Source of funding: NR, but no both: 47/33/17 Mesh placement in both posterior prolapse in those had anterior repair. financial assistance received compartments: 17 from any company in the design Vault only / uterine only / both: Other symptoms not cured or execution of this study vault or uterine 17 (12 Urinary: urge, 9/52 who had urge; urge incontinence, 4/40 who sacrospinous fixation, 4 bilateral had urge incontinence iliococcygeal fixation, 1 abdominal Bowel: constipation, 1/21 who had constipation sacrocolpopexy) Sexual: dyspareunia, pre- 25/67 active; 24m 4/66 active Other:

Other symptoms de novo Urinary: urge, 3/45 (2 had anterior and posterior repair, 1 had anterior repair); urge incontinence, 3/57 (2 had anterior and posterior repair, 1 had posterior repair)

133 Bowel: constipation, 1 who had anterior repair Sexual: dyspareunia, 3/42 who had no dyspareunia

Further need for treatment Repeat POP surgery: 5 (5%) who had recurrent proplase at a non-mesh site (de novo); Mesh removal surgery: 6/9 erosions, excision of exposed mesh and vaginal closure; Other: fistula, 1w, 1/9 erosions, surgical closure of the rectovaginal fistula.

Additional information The risk of mesh erosion decreased over the study period, i.e. vaginal mesh erosion is related to surgical experience

Author, year Inclusion criteria: women Intervention A (n 142): anterior SAFETY Flood 199872 underwent an extended anterior colporrhaphy reinforced with Mesh erosion: mesh erosion into vagina, 3/142, 3m, 4y, 7y colporrhaphy reinforced with Marlex mesh respectively. Type of prolapse: anterior Marlex mesh Blood loss: mean, 442ml Type of mesh: Marlex mesh Infection: reaction or persistent infection because of mesh, 0 Publication type: full-text Exclusion criteria: NR

Study details Participant characteristics Intervention characteristics Results Concomitant operation Other: 1 postoperative stroke, with history of heart disease and Study design: retrospective Number of patients lost to Hysterectomy: 94 was lost to follow-up after the 6w check (no further details); 2 case series follow-up: NR, unclear postoperative myocardial ischaemia, one of them was lost follow- consecutive or not Additional information: up (no further details). Setting: University hospital Age (mean, range): 65 (37-87) Number of surgeon: unclear (‘’the Parity (mean, range): 2.6 (0-9) operation was designed by the Country: Canada senior author, and performed with EFFICACY, mean 3.2y (6w-12y) Primary/Secondary repair: the residents and fellows on the Objective Failure Recruitment/Treatment dates: 120/22 urogynecology service). POP: stage >=II cystocele (descent < halfway to the hymen): Jan 1977 – May 1992 0/142 Length of follow-up (mean, Vault only / uterine only / both: range): 3.2y (6w-12y), less than 12 sacrospinous vault suspension Subjective Failure 20 patients had follow-up <1y and 1 amputation of cervic POP: time to develop was mean 2.9y (6m-8y), 8/142 (5.7%) of whom 6 had recurrent vault and posterior prolase and 2 had Source of funding: NR Other information: posterior prolapse only (unclear de novo or due to posterior or Symptoms of incontinence: vault repair failure) 112/142 (79%); SUI 81%; Other symptoms not cured urge incontinence: 86%; Urinary: incontinence, 8/30 who had SUI mixed stress and urge

134 incontinence: 81/112 (72%) Other symptoms de novo Urinary: SUI, 1/30 who had no incontinence I no prolapse: 0 II descent

Other outcomes Length of hospital stay(N, mean, range): 142, 10.1d (5-28)

Author, year Inclusion criteria: patients had Intervention A (n 251): cadaveric SAFETY, 6-61m Frederick 200573 objectively demonstrated stress prolapse repair with sling (CaPS) Infection: serious infection 0/251 urinary incontinence and grad 2-4 Other: Type of prolapse: anterior cystocele before surgery Type of grafts: Tuto-Plast, prolonged urinary retention (>1m), 3/251 Mentor Corporation, Santa intravaginal granulation tissue, 22/251 (9%) Publication type: full-text Exclusion criteria: NR Barbara, California, solvent dehydrated cadaveric fascia lata Study design: prospective case Number of patients lost to EFFICACY, 6-61m series follow-up: NR, 44/295 not Concomitant operation: Objective Failure

Study details Participant characteristics Intervention characteristics Results analysed because follow-up <6m Incontinence: all 251 POP: any degree of postoperative anterior wall prolapse, 39/251 Setting: hospital Hysterectomy: 28/248 (11%) (16%), of which 21 (54%) had grade I asymptomatic prolapse, Age (mean, range): 66 (31-90) the other 18 had symptomatic cystocele/recurrence II-IV grade. Country: U.S. Additional information: Primary/secondary repair: Number of surgeon: NR de novo POP: II-IV grade rectocele 8/158 (5%), apical (vault & Recruitment/Treatment dates: 226/25 enterocele) 7/165 (4%). 1998 – Jan 2003. Anterior only / posterior only / Other symptoms not cured Length of follow-up (mean, both (n 248): 158/90/0 Urinary: all types of incontinence, i.e. stress urinary incontinence, median, range): 6-61m; follow- urge incontinence, and mixed, 114/251 (45%) cured, 78/251 up for physical examination, Additional information: (31%) >50% improvement, 59/251 (24%) <50% improvement 20m, 17m, 6-60m; follow-up for Baden-Walker grade cystocele: II questionnaire, 24m, 20m 67/251 (27%), III 112/251, IV Other: SEAPI (stress incontinence, emptying, anatomy, 72/251 (29%) protection, inhibition) scores, mean, pre-operation 7.38, post- Source of funding: the second operation 2.94. author (Gary Leach) had financial interest and/or other Other symptoms de novo relationship with Mentor, Boston Urinary: urge incontinence, 13/153 (8.5%) Scientific and AMS Quality of Life

135 Disease specific: prolapse QoL score, range 4-20, mean: preoperation 9.6, post-operation 5.5

Satisfaction Visual analogue scale (0-100%): >=50% satisfied 200/251 (80%)

Author, year Inclusion criteria: women Intervention A (n 70): cystocele SAFETY Gomelsky 200474 underwent surgical correction of repair + porcine dermis Mesh erosion: hematoma, persistent vaginal discharge or high grade cystocele (Baden- interposition graft erosion, 0/70 Type of prolapse: anterior Walker and POP-Q grading system), i.e. grade III: Aa+1 and Type of grafts: Intexen, AMS, Blood loss: requiring transfusion, 0/70 Publication type: full-text Ba+2, or at rest, grade IV: Aa+3 Minnetonka, Minnesota, animal and Ba+4 porcine dermis Other organ damage: injuries to bladder, urethra, ureters or Study design: retrospective small bowel, intraoperative, 0/70 case series Exclusion criteria: NR Concomitant operation: Incontinence, 65/70, pubovaginal Other: superficial vaginal wound separation, 1/70 and was Setting: medical centre Number of patients lost to sling for SUI successfully treated conservatively. Country: US follow-up: NR Recruitment/Treatment dates: Additional information: Any complications: there were no intraoperative complications NR Vault only / uterine only / both: Number of surgeon: NR vault repair, 50/70, bilateral Length of follow-up (mean, vaginal vault suspension to the EFFICACY, mean 24m (12-NR)

Study details Participant characteristics Intervention characteristics Results range): 24m (12-NR) iliococcygeus fascia Objective Failure POP: grade II (Ba 0) cystocele asymptomatic, 6/70 (8.6%); grade Source of funding: NR, but III (Ba +2) and elected not to undergo reoperation, 3/70 (4.3%) one author of the 3 had financial interest and/or other relationship Other outcomes with Bioform, Genyx, Ortho Length of hospital stay: 1-2d McNeil, Surx, Lilly and Watson.

Author, year Inclusion criteria: patients with Intervention A (n 73): site- SAFETY, mean 13.7m (6-23) Kobashi 200580 symptomatic (stool trapping specific repair reinforced with Other: fecal urgency, 1; and/or vaginal/perineal splinting or cadaveric fascia Granulation on tissue from an exposed suture, 8; Type of prolapse: posterior postural modifications to facilitate Vaginal wound separation, 3; stool evacuation) rectoceles Type of grafts: solvent-dried Prolonged postoperative pain, 2; Publication type: full-text underwent a site-specific repair fascia lata, Tutoplast, Mentor Bloody vaginal discharge, 1. reinforced with cadaveric fascia Corporation, Santa Barbara, Study design: retrospective California case series Exclusion criteria: NR EFFICACY, mean 13.7m (6-23) Additional information: Objective Failure Setting: medical centre Number of patients lost to Number of surgeon: NR POP: Baden-Walker, follow-up: 11 not responded to Pre-, n 73 Post-, n 50

136 Country: US the postoperative questionnaire, 0 0 45 23 not underwent physical 1 3 2 Recruitment/Treatment dates: examination. 2 22 3 NR, 2 centres Age (range): 31-86 3 38 0 Length of follow-up (mean, 4 10 0 range): 13.7m (6-23) de novo POP: 1 symptomatic enterocele Source of funding: NR, but the first 2 authors are paid Subjective Failure consultant to Mentor POP: 3/50 were symptomatic Corporation and American Medical System Other symptoms not cured Bowel: stool trapping, pre- 53/62, pos- 9/53, improved 44/53; Vaginal/perineal splinting, pre- 38/62, post- 4/38, improved 34/38. Sexual: dyspareunia, pre- 14/39 sexually active; post- 5/14

Other symptoms de novo Sexual: dyspareunia, 4/25 who had no dyspareunia

Satisfaction 80% satisfied: 41/62;

Study details Participant characteristics Intervention characteristics Results Would repeat procedure: 48/62 Would recommend procedure: 48/62

Author, year Inclusion criteria: patients Intervention A (n 132): cadaveric SAFETY, 6-28m Kobashi 200275,76 undergone cadaveric prolapse prolase repair with sling technique Infection: urinary tract infection, 4 (3.0%) repair with sling for grade 2-4 (CAPS) Other: Type of prolapse: anterior cystoceles with a minimum of 6m Osteitis pubis: 1 (0.7%); follow-up Type of graft: cadaveric fascia Would separation (no details): 3 (2.3%); Publication type: full-text lata, solvent dehydrated Urinary retention: 1 (0.7%); Exclusion criteria: NR (Tutoplast, Mentor Corp., Santa Vaginal pain: 1 (0.7%); Study design: prospective case Barbara, Califonia) Suprapubic pain: 1 (0.7%); series Number of patients lost to Osteomyelitis: 0 follow-up: NR, 40 having no Concomitant operation: NR Setting: hospital completed follow-up parameters were excluded. Additional information: EFFICACY, 6-28m Country: US Number of surgeon: NR Objective Failure Age (mean, range): 62.1 (35-90) POP: Baden-Walker, grade 1 cystocele, 15 (11.4), grade 2, 2 Recruitment/Treatment dates: Incontinent: pure SUI 52 (1.5%), with 10 occurred <3m and 7 occurred 3-12m. NR, could be in a previous (39.4%), occult SUI 7/132, mixed report (Kobashi KC, Mee SL, 54/132 (40.9%) de novo POP: apical prolapse, 13 (9.8%) with 2 small

137 Leach GE. A new technique for enteroceles and 11 moderate enteroceles; rectocele, 2 (1.5%) cystocele repair and Primary/secondary repair: NR transvaginal sling: the cadaveric Other symptoms not cured prolapse repair and sling Additional information: Urinary: 24/108 (22.2%) who had either pure SUI or mixed (CAPS), Urology, 56:9, 2000.) Grade 2 cystocele, 73 (55.3%); incontinence; 1/7 who had occult SUI. grade 3, 43 (32.6%); Length of follow-up (mean, grade 4, 16 (12.1%) Other: SEAPI (Stress-related leakage, emptying ability, anatomy, median, range): 12.4m, 15, 6- protection, inhibition), mean, pre- 5.58, post- 0.90. 28 Other symptoms de novo Source of funding: NR, but 2 Sexual: dyspareunia, 4/NR number of women free of of 4 authors having financial dyspareunia at baseline interest and/or other relationship with Mentor Corp. and American Satisfaction Medical Systems. One of them >=70% satisfied, 99 (75.0%); would under go the surgery again, is 1st author. 105 (79.9%); would recommend the surgery to a friend, 102 (77.4%).

Author, year Inclusion criteria: women Intervention A (n 63): anterior or SAFETY Milani 200528 requiring prolapse repair for posterior repair + mesh Mesh erosion: A, mean 12m (5), 13%; B, mean 14m (6), 6.5% symptomatic anterior or posterior Type of prolapse: anterior or vaginal prolapse, each patient had Type of mesh: Prolene, Ethicon,

Study details Participant characteristics Intervention characteristics Results posterior, reported seperately at least one of the following risk Somervill, New Jersey EFFICACY factors: previous prolapse Objective Failure Publication type: full-text surgery, obesity (BMI>29), chronic Concomitant operation POP: any degree, A, 2/32; B 2/31; bronchopneumonia, chronic Hysterectomy, A 18; B 10 A B Study design: prospective case staining at defecation, history of Pre- Post- Pre- Post- series hernia or collagen disease. Additional information: 0 - 30 - 29 Number of surgeon: NR I - 2 - 2 Setting: university hospital Exclusion criteria: had not II 9 - 19 - completed their family, diabetic or III 16 - 10 - Country: Italy with a recognised or suspected IV 7 - 2 - immunodepression. Recruitment/Treatment dates: Other symptoms not cured NR Number of patients lost to Urinary: Length of follow-up (mean, follow-up: NR A range): 17m (3-48) Age (mean, range): anterior Pre- Post- repair (A) 63 (49-82); posterior Urgency 16 13 Source of funding: NR repair (B) 63 (50-80) Urge incontinence 5 5 SUI 8 7 Anterior only / posterior only / both: 32/31/0 Bowel: B, constipation, pre- 14, post- 9

138 Sexual : dyspareunia, A B Pre- Post- Pre- Post- 6/17 8/17 2/17 21/13?

Other symptoms de novo Bowel: B, anal incontinence, 1/29 who had no anal incontinence Sexual: dyspareunia, A, 4/11 had no dyspareunia

Further need for treatment Mesh removal surgery: 1 mesh removal for pelvic abscess.

Author, year Inclusion criteria: patients with Intervention A (n 39): anterior SAFETY Powell 200477 stage 2 or greater anterior vaginal vaginal wall hammock (AVWH) Mesh erosion: erosion of the vaginal mucosa overlying the compartment relaxation with donor fascia lata AVWH, postoperative, 6/58 (10%) Type of prolapse: anterior underwent anterior vaginal wall hammock (AVWH) procedure with Type of grafts: donor fascia lata, Blood loss: requiring transfusion, 2/58 (3%), intraoperative, mean Publication type: full-text autologous or allograft fascia lata. no further details 329cc (50-1000)

Study design: case series Exclusion criteria: 11 patients Intervention B (n 19): anterior (unclear prospective or who had relocated and had good vaginal wall hammock (AVWH) EFFICACY retrospective) vault support at the last follow-up, with autologous fascia lata Objective Failure

Study details Participant characteristics Intervention characteristics Results but had not been seen within the POP: Setting: hospital last 12m were excluded. POP-Q stage >=2 (prolapse of any vaginal segment to <1cm of Type of grafts: autologous fascia the hymenal ring or greater), 12/58 (19%); A 9/39; B 3/19; Country: US Number of patients lost to lata median time to failure 14.4m (2-40) follow-up: NR, 11 patients met Recruitment/Treatment dates: exclusion criteria were excluded Concomitant operation: De novo POP: 5/28 who had no posterior repair preoperation Jun 1998 – Mar 2001 as stated above Incontinence: 41/58 (71%); A 29/39; B 12/19 Subjective Failure Length of follow-up (median, Age (mean): 61.9; A 63.7; B 56.1 Hysterectomy: 14/58; A 9/39; B POP: symptoms of pelvic pressure or a vaginal bulge requiring range): 24.7m, 12-57 Parity (mean): A 3.4; B 2.8 5/19 further treatment, 2/58 (2%); A 1/39; B 1/19 BMI (mean): A 29.1; B 26.6 Source of funding: Incontinent: SUI, 41/58 (71%) Additional information: Other symptoms not cured government (the Chief, Bureau Number of surgeon: >1, all Urinary: incontinence, 11/41 (27%) had persistent incontinence of Medicine and Surgery, Navy Anterior only / posterior only / patients underwent AVWH after sling procedure Department, Washington DC, both: 28/0/30; A 17/0/22; B 11/0/8 performed by resident physicians Clinical Investigation Program) under the supervision of one of Other symptoms de novo Vault only / uterine only / both: the first 2 authors Urinary: SUI, 1/17 who did not have a sling procedure vault or uterine, 18/58; A 12/39; B 6/19 Further need for treatment Repeat POP surgery: 1/58 failure

139 Mesh removal surgery: 0/6 erosions, all resolved with would Additional information: debridement and vaginal estrogen All patients had apparent bilateral paravaginal defects on Additional information: examination; Age and parity were not predicative but increasing BMI was associated with a significantly higher failure rate in the anterior POP-Q, median, cm: compartment. A B Anterior 2 1 Middle compartment -3 -4 Posterior -1 -2 Author, year Inclusion criteria: women with Intervention A (n 325): anterior SAFETY Rozet 200485 genito-urinary prolapse underwent and posterior mesh repair with Mesh erosion: 6m-5y, vaginal erosions, 3/325 (0.9%); mesh a transperitoneal placement of a both meshes sutured to the infection, 2/325 (0.6%) Type of prolapse: anterior and 100% polyester mesh on the promontory posterior anterior vaginal wall and a Blood loss: requiring transfusion, intra-operative, 3/328, but they posterior mesh on the levator ani Type of mesh: polyatex A4 were converted to open surgery, so not included in the analysis Publication type: full-text muscle 100x150mm, Cousin Biotech (100% polyester single covered Infection: post-operative, spondylitis, 1/325 (0.3%) Study design: retrospective Exclusion criteria: NR silicone mesh) case series (consecutive) Other: Number of patients lost to Concomitant operation: Blood loss, intra-operative, 10-100cc;

Study details Participant characteristics Intervention characteristics Results Setting: hospitals follow-up: NR, 38/363 patients Incontinence: 163/325 (49%) Bowel incarcerations/intestinal obstruction, post-operative, 1/325 were excluded because of having Hysterectomy: 15/325 (4%) (0.3%); Country: France no complete medical chart Persistent constipation, <=6m and resolved within this period, Additional information: 19/325 (6%); Recruitment/Treatment dates: Age (mean, range): 63 (35-78) Number of surgeon: 5 senior Urinary retentions, post-operative, 2/325 (0.5%); Jun. 1996 – Mar. 2003 Incontinent: 177/325 (54%) surgeons in 2 medical centres Port hernia, post-operative, 1/325 (0.3%)

Length of follow-up (mean, Primary/secondary repair: NR range): 14.6m (6m-5y), 67% EFFICACY, 5m-5y were followed up 5y Anterior only / posterior only / Objective Failure both: 0/0/325 POP: (no definition), 13/325 (4%), 10 of which were cystocele Source of funding: NR and 3 rectocele, 7/13 occurred within 6m Additional information: Baden-Walker grading: II 82/325 Other symptoms not cured (25%), III 217/325 (67%), IV Sexual: complain of sexual dysfunction, 0/325 26/325 (8%) Other symptoms de novo Urinary: incontinence, 19/162 (12%)

Satisfaction

140 Satisfy the results of operation: 96% (312/325)

Further need for treatment Mesh removal surgery: 3 partial vaginal removals

Other outcomes Operating time (N, mean, range): 325, 97min (45-156) Length of hospital stay(N, mean, range): 325, 3.7d (2-7)

Author, year Inclusion criteria: patients Intervention A (n 112): 4-defect SAFETY, mean 21m (6-42) Safir 199978 underwent repair of grade 4 repair with mesh Mesh erosion: extrusion of mesh or pelvic abscess related to the cystocele using the 4-defect repair mesh, 0/112 Type of prolapse: anterior technique (grade 4 = extrusion of Type of mesh: polyglycolic acid the bladder base beyond the mesh, NR manufacture or trade Blood loss: requiring transfusion, 0/112 Publication type: full-text vaginal introitus at rest). name Other organ damage: bladder perforation, vesicovaginal fistula, 0/112 Study design: retrospective Exclusion criteria: NR Concomitant operation: case series Hysterectomy: 22/112 Infection: retropubic abscess, 0/112 Number of patients lost to Setting: University hospital follow-up: NR, 18 follow-up <6m Additional information: were excluded Number of surgeon: NR EFFICACY, mean 21m (6-42) Country: US Objective Failure

Study details Participant characteristics Intervention characteristics Results Age (mean, range): 65 (35-96) POP: grade 2 cystocele, 6/112 (5%); grade 4 cystocele, 5/112 Recruitment/Treatment dates: Parity (mean): 2.8 (4%) 1994-1997 Primary/secondary repair: 70/60 Subjective Failure Length of follow-up (mean, (46%) POP: symptomatic: 5/112 (4%) who had recurrent grade 4 range): 21m (6-42) cystocele. Anterior only / posterior only / Source of funding: NR both: 31/0/81 Other symptoms not cured Urinary: urge incontinence, 6/45 (13%) who had urge incontinence; SUI, 5/49 (10%) who had SUI, defined as using pads daily

Other symptoms de novo Urinary: urge incontinence, 5/67 (7%) who had no urge incontinence; SUI, 0/63 who had no SUI.

Further need for treatment Repeat POP surgery: 2/5 recurrent grade 4 cystoceles. Author, year Inclusion criteria: advanced Intervention A (n 89): vaginal SAFETY Simsiman 200679 >=stage II anterior vaginal wall paravaginal repair (VPVR) with Mesh erosion: erosion of the vaginal mucosa overlying the graft,

141 prolapse, underwent reinforced Pelvicol 6-44m, 15 (6.7%), all erosions resolved with office management Type of prolapse: anterior vaginal paravaginal repair (VPVR) including wound debridement and vaginal estrogen with Pelvicol Type of grafts: porcine dermis Publication type: full-tex (Pelvicol, Bard, Covington, GA) Exclusion criteria: NR EFFICACY Study design: retrospective Concomitant operation: Objective Failure case series Number of patients lost to Incontinence: 41/89 (45%) POP: point Ba>-1, 1y, 21/89 follow-up: NR, 22 follow-up <1y Hysterectomy: 54% Ba Pre- Post- Post- mean (SD) Setting: hospital and last visit showed no sign of -3 36 (38%) failure were excluded. Their age, Additional information: >1 -2 25 (26%) Country: US BMI, parity, ethnicity and severity Number of surgeon: performed by -1 7 (7.4%) of POP are similar to included resident or fellow physicians with 0 34 (37.2%) 11 (11.6%) -2.0 (1.3) Recruitment/Treatment dates: women. 1 of the 2 senior authors +1 7 (8.1%) 3 (3.2%) -2.0 (1.1) Sep 2001 – Jan 2004 +2 12 (14%) 4 (4.2%) -2.2 (1.4) Age (mean, SD, range): 59.5 +3 16 (17%) 2 (2.2%) -1.4 (1.9) Length of follow-up (mean, (11.6), 26-82 +4 20 (22.4%) 1 (1.1%) -0.6 (2.1) SD, range): 24 (10.1)m, 6-44, 1y outcome available Parity (median, range): 3 (0-15) Other symptoms not cured BMI (mean, SD, range): 29 (5.5), Urinary: urge incontinence, 1y, 24/58 had urge incontinence Source of funding: NR 20-44 Other symptoms de novo Vault repair: 41/89 (45%) Urinary: 1y, urge incontinence, 8/28 who had no incontinence;

Study details Participant characteristics Intervention characteristics Results SUI, 2/25 (8%) who had no SUI.

Further need for treatment Repeat POP surgery: 6-44m, 5/21 failures

142

Studies included for safety review only

Registry

Study details Participant characteristics Intervention characteristics Results Author, year Inclusion criteria: women Intervention A (n 248): SAFETY, during the surgical procedure and its associated Altman 200729 underwent transvaginal mesh transvaginal mesh surgery hospital stay surgery Type of prolapse: anterior Type of mesh: polypropylene, Mesh erosion: NR and/or posterior, reported Exclusion criteria: NR Prolift, Gynecare, Somerville, NJ Blood loss: requiring transfusion, A 1, B 0, C, 0, D, 0 seperately Bleeding Total A B C D Number of patients lost to Concomitant operation >500ml 4 3 1 0 3 Publication type: full-text follow-up: NR Incontinence: 7 TVT or >1000ml 1 1 0 0 0 Age (mean, SD): overall 67.8 transobturator tension-free tape Mean, SD 154 190 85 153 169 Study design: registry report (10.3); A, anterior repair, 68.1 procedures; (212) (295) (97) (100) (133) (10.2); B, posterior repair, 67.5 Hysterectomy/cervical Setting: 25 centres: 12 (10.1); C, combined anterior and amputation: 21/2 Other organ damage: Sweden, 2 Denmark, 2 Norway, posterior repair, 69.5 (11.2); D, total A B C D 9 Finland mesh, 73.5 (9.4) Additional information: Bladder perforation 2 0 1 2 Parity (median, range): overall, 2 Number of surgeon: all Urethra perforation 2 0 0 0

143 Country: Sweden (0-6); A, 2 (0-6); B, 2 (1-6); C, 2 (0- participating gynaecologists Rectal perforation 0 3 1 0 5); D, 2 (0-4) were experienced pelvic floor Recruitment/Treatment dates: surgeons and had received Infection: urinary tract infection, A, 6; B 3, C 0, D 7; wound Initiated from Oct 2005 and the Primary/Secondary repair: total pretrial hands-on training on the infection, A 0, B 1, C 0, D 0; registration period lasted for 6m 108/40; A, 18/88; B, 48/23; C, 15/5; standardised transvaginal mesh Other: Length of follow-up: NR D, 27/24 techniques, however, this series A B C D represents the first experience Urinary retension 0 1 1 2 Source of funding: The study Vault only / uterine only / both: 5 with a new surgical technique. Postoperative fever (>37.8 C) 2 0 0 2 was entirely funded by sacrospinal fixation Anemia (Hb <100g/L) 1 0 0 0 University-administered Groin pain 1 1 0 0 research funds but pretrial Other information: Buttock pain 0 0 0 0 scientific meetings were Vaginal repair using other meshes: Vaginal hematoma 2 0 0 0 sponsored by Gynecare, 3, no further details; Defecation pain 0 0 0 1 Sweden AB. The transvaginal Defecation difficulties 0 0 0 1 mesh manufacture company A, Anterior repair: 106/248; Inguinal hematoma 0 0 0 0 had no influence over study aim, B, posterior repair: 71/248 design, execution, or analysis C, combined anterior and posterior Other outcomes and interpretation of data. The repair: 20/248; Total A B C D 1st author has an educational D, total mesh (no description): Operation time, 71.9 72.5 57.6 86.9 85.9 advisory position for Gynecare 51/248 nd min, (mean, SD) (29.0) (32.1) (19.6) (30.2) (22.9) Sweden AB. The 2 author is a Hospital stay, d, 3 (1-10) 3 (1-10) 2 (1-7) 3 (2-9) 4 (1-10) member of the Johnson &

Study details Participant characteristics Intervention characteristics Results Johnson advisory board. (median, range)

Additional information Concurrent pelvic floor surgery was associated with an increased risk for minor complications. There were no other predictors of outcomes when assessing the association between age, parity, weight, previous pelvic floor surgery, previous hysterectomy, or concurrent hysterectomy.

144

Case series (n >50, F/U <1y or only reported safety outcomes)

Study details Participant characteristics Intervention characteristics Results Author, year Inclusion criteria: patients underwent Intervention A (198: Atrium SAFETY Achtari 200590 vaginal reconstruction surgery mesh, n 97; Vypro II mesh, n Mesh erosion (defined as any visible mesh exposure identified reinforced by synthetic mesh 101): anterior and/or posterior on vaginal examination): 14/198 (7.1%) (7 Atrium mesh, 7 Vypro Type of prolapse: anterior repair with mesh II mesh; 5 occurred <=6w, 9 between 6w and 6m; 6 meshes and/or posterior Exclusion criteria: NR were used for anterior repair, 7 for posterior repair, and 1 for Type of mesh: Atrium, Vypro II vault repair; 9 were used by consultant surgeons, 5 by fellows) Publication type: full-text Number of patients lost to follow-up: NR Concomitant operation: Symptoms of mesh erosion included vaginal discharge (5), pain Study design: retrospective Incontinence: 67/198 (5) and dyspareunia (2), 5 were asymptomatic. case series Age (mean, SD): 63 (11.6) Hysterectomy: 13/198 Parity (median, range): 3 (2-4) EFFICACY Setting: hospital Postmenopausal: 176/198 Additional information: Further need for treatment Number of surgeon: >1, initially Mesh removal surgery: day 5, 10/14 erosions Country: Australia Primary/secondary repair: NR the consultant surgeon operated on all patients, using Atrium Additional information: Recruitment/Treatment dates: Anterior only / posterior only / both: mesh only, then changing to 1. Surgeon experienced in the technique has less erosions than Feb. 1999 – Jul. 2003 (Atrium Atrium mesh: 49/32/16; Vypro II mesh: composite mesh (Vypro II). less experienced surgeons. mesh was used in 97 patients 41/44/16 Fellow surgeons commenced 2. Patient age was also associated with mesh erosion.

145 until Dec. 2001, and from Jan. using Vypro II only, in Jan. 2002. 3. Hormon replacement treatment (HRT) was not significantly 2002, 101 patients had a Vypro Vault only: 2 associated with erosion but slightly increased risk of erosion II mesh) (OR: 3.0, 95% CI 0.8-10.3, p=0.08)

Length of follow-up (range): 6w-6m

Source of funding: NR

Author, year Inclusion criteria: patients underwent Intervention A (n 277): SAFETY Collinet 200691 surgery due to pelvic prolapse with transvaginal mesh repair Mesh erosion: <=2m, 34/277 (12.27%), 9 of whom were transvaginal mesh technique completely healed by antiseptics and antibiotics, the other 25 Type of prolapse: anterior Type of mesh: polypropylene, required partial mesh exeresis. In 33/34, the exposure site was and/or posterior Exclusion criteria: NR Prolene Soft (169, 61%) or located on the anterior colpotomy scar. Prolene (108, 39%), Gynecare Publication type: full-text Number of patients lost to Other organ damage: rectal injury, 1; bladder injury, 4 follow-up: NR, retrospective, all Concomitant operation Other: haematoma, immediate postoperative, 1 Study design: retrospective analysed SUI, 136 (49.10%), suburethral case series Age (mean, range): 64 (37-81) sling with retropubic or BMI (mean, range): 26 (17-47) transobturator methods EFFICACY Setting: University hospital Postmenopausal: 83.75% Hysterectomy, 164 (59.21%) Further need for treatment Mesh removal surgery: 25/34 mesh exposure required partial

Study details Participant characteristics Intervention characteristics Results Country: France Primary/Secondary repair: NR Additional information: mesh exeresis; one patient presented with vesicovaginal fistula Number of surgeon: NR after resection underwent surgery to repair the fistula Recruitment/Treatment dates: Anterior only / posterior only / both: Jan 2002 - Dec 2003 63/46/166 Additional information Length of follow-up: >=2y, Anatomical and functional results will be analysed in another only reported safety results, and Other information: study; were for 2m HRT: 34/142 (24.65%) SUI: 163 (58.84%) Risk factors of erosion were concomitant hysterectomy and Source of funding: NR Each patient presented with grade 3 inverted T colpotomy, no significant correlation from the prolapse (ICS classification) based on menopause, RHT, or the patient’ medical history; at least one of three stages with functional impairment: No significant differences for prosthetic material used (Prolene Cystoclele: 242 (87.36%) of whom mesh vs. Prolene Soft), number of prostheses (1 or 2) and 61.37% >= grade 3; concomitant management of SUI (suburethral sling or no sling). Hysterocele or colpocele: 167 (60.21%) of whom 33.85% >= grade 3; Rectocele: 236 (85.20%) of whom 23.47% >= grade 3 Author, year Inclusion criteria: patients with Intervention A (n 83): SAFETY Cosson 200292 symptomatic prolapse of the uterus laproscopic sacropexy with Mesh erosion: rejection of the posterior mesh, occurred at 6m,

146 who underwent laproscopic sacral anterior and posterior mesh 1/83; Type of prolapse: anterior and colpopexy reinforcement Return to theatre (short-term): 1 for haemorrhage > 500ml posterior (assume required blood transfusion) Exclusion criteria: NR Type of mesh: Mersilene, Publication type: full-text Ethicon Other organ damage: rectal injuries, 1, during rectovaginal Number of patients lost to dissection; bladder injuries, 1, during the dissection before Study design: retrospective follow-up: 12 not available at 1y, no Concomitant operation placing anterior strip; case series further details. Incontinence: 74 laproscopic Age (mean, range): 47 (28-66) Burch colposuspension Infection: chronic inflammation of the cervix, 1; lower urinary tract Setting: university hospital Parity (mean, range): 3.5 (0-6) Hysterectomy: 60 subtotal infection, 9; Postmenopausal: 5 hysterectomy Country: France Other: unexplained fever for over 24h, 9; postoperative phlebitis, Anterior only / posterior only / both: Additional information: 1; urinary retension, 0 Recruitment/Treatment dates: 0/0/all (83) Number of surgeon: NR, Jun 1996 – May 1998 operation time decreased from Length of follow-up (mean): Vault only / uterine only / both: all 292min to 180min as experience EFFICACY 343d (83) had sacral colpopexy was gained. Objective Failure POP: cystocele, 1st degree: 1/65, occurred 3m after surgery, Source of funding: NR Other information: became symptomatic after 1y with a 3rd degree recurrence; HRT: 4 hysterocele, 1st degree: 1/65; degree 1st 2nd 3rd Total Cystocele 9 34 40 83 Subjective Failure

Study details Participant characteristics Intervention characteristics Results Hysterocele 5 27 31 63 POP: symptomatic cystocele, 1/65 occurred at 1y Enterocele 0 8 5 13 Rectocele 12 29 22 63 Other symptoms not cured Urinary: incontinence, 20/67 who had incontinence

Further need for treatment Repeat POP surgery: 1/65 for the 3rd degree recurrence of cystocele; Mesh removal surgery: 1 who had mesh rejection Other: incontinence, 2

Other outcomes Operating time (N, mean, range): 83, 276 (120-360) min Length of hospital stay(N, mean, range): 83, 3.5 (2-7) d

Author, year Inclusion criteria: patients underwent Intervention A (n 138: 89 SAFETY Deffieux 200786 vaginal reconstruction surgery Gynemesh, 49 Gynemesh- Mesh erosion (all vaginal erosions were reported even when only reinforced either by Gynemesh or Soft): transvaginal repair of a few milimeters): 27/138 (20%), 15/89 with Gynesh (310mm2, Type of prolapse: anterior Gynemesh-Soft beneath the bladder by cystocele using synthetic mesh 75-600) and 12/49 with Gynemesh-Soft mesh (200mm2, 10-100). vaginal approach. 16/27 occurred <12w with 12 asymptomatic, 2 vaginal discharge,

147 Publication type: full-text Type of mesh: Gynemesh, 1 vaginal pain, and 1 vaginal discharge with pain. 11/27 occurred Exclusion criteria: NR Gynemesh-Soft, Gynecare, >12w-21m with 8 asymptomatic, 2 dyspareunia, and 1 Study design: retrospective Ethicon, France dyspareunia with vaginal discharge. case series (patients recruited Number of patients lost to follow-up: consecutively) 4 Concomitant operation: Mesh infection: 0/138 Incontinence: 87/138 (63%) Infection: 0/138 Setting: University hospital Age (mean, range): 62 (30-83) Hysterectomy: 103/138 (75%) Parity (mean, range): 2.7 (0-9) EFFICACY, 6m Country: France Postmenopausal: 116/138 Additional information: Objective Failure Incontinent: Number of surgeon: NR POP (Baden & Walker >= II stage): 7/138 (5%): 3 (3%) with Recruitment/Treatment dates: Gynemesh, 4 (8%) with Gynemesh-Soft Oct. 1999 – Oct. 2004 Primary/secondary repair: NR (Gynemesh was available Oct. Other symptoms de novo 1999 – Dec. 2003; Gynemesh- Anterior only / posterior only / both: Sexual: dyspareunia, 9% in patients with vaginal erosion, 11% in Soft was available from Oct. 118/0/20 those without erosion 2003) Vault repair: 83/138 (60%) Further need for treatment Length of follow-up: minimum Mesh removal surgery: 15/27 (55%) erosions 6m follow-up outcomes Additional information: available. Gynemesh 32.1m Substitutive hormone treatment: 34/138 Additional information: (7.5m – 59.9m); Gynemesh-Soft (25%) Age increases risk of erosion significantly, Baden-Walker stage 7.1m (1-21.9m) Cystocele stage (Baden & Walker): I 3, >2 cystocele protect from erosion significantly.

Study details Participant characteristics Intervention characteristics Results II 37, III 76, IV 23 Source of funding: NR, but not sponsored by manufactures.

Author, year Inclusion criteria: initially, the mesh Intervention A (n 110): mesh SAFETY Fatton 200793 indication was reserved for recurrent repair of genital prolapse Mesh erosion: mesh exposure, 3m, 5/106, 3 of them successfully vaginal prolapse. Primary cases with managed with vaginal hormone therapy; Type of prolapse: anterior significant prolapse were then allowed Type of mesh: soft prolene Blood loss: requiring transfusion, 0/110, all <=300ml and/or posterior into the study after the positive results mesh, Prolift, Gynecare, Other organ damage: bladder injury, 1/110 due to dissection; in the early procedures. All had a Ethicon, Sommerville, NJ, USA rectal injury, 0/110 Publication type: full-text prolapse with at least a >=3 component Infection: urinary tract infection, 13/110 Concomitant operation Other: febrile morbidity (>38.5 C), 2/110; Study design: retrospective Exclusion criteria: NR Incontinence: 45, TVT haematomas requiring secondary surgical management, 2/110; case series Hysterectomy: 15 shrinkage of mesh, 3m, 18/106; Number of patients lost to pelvic abscess, 0/110; Setting: hospital follow-up: 4/110 lost follow-up at 3m Additional information: granuloma without mesh exposure, 3m, 3/106; Age (mean, range): 63.2 (29-90) Number of surgeon: NR bladder or rectal fistula, 0/110; Country: France Parity (mean, range): 3.13 (1-11) persistent perineal pain, 0/110; BMI (mean, range): 26.7 (19-43) vaginal syneshia, 3m, 1/106 Recruitment/Treatment dates: Postmenopausal: 93/110

148 Feb – Sep 2005 Length of follow-up (mean, Primary/Secondary repair: 88/22 EFFICACY, 3m range): 25w (12-42) Objective Failure Anterior only / posterior only / both: Stage 2 Stage 3 Stage 4 Source of funding: NR, no 22/29/59 (total mesh) C V/U E R C V/U E R C V/U E R financial assistance was A+P or total (58)8 4 0 4 1 0 0 0 0 0 0 0 received from any company for Vault only / uterine only / both: 5, A (20) 0 0 1 5 0 0 0 0 0 0 0 0 the execution of this study transvaginal sacrospinous uterine P (28) 6 2 0 2 1 0 0 0 0 2 0 0 fixation A, anterior; P, posterior; C, cystocele; V, vaginal vault prolapse, U, uterine prolapse; E enterocele; R, rectocele. Other information: HRT: 14 (therefore, POP: >=stage 3, A+P or total, 1/58; A, 0/20; P, 0/28; de novo POP: >=stage 3, A, 0/76; P, 3/78; stage 2 can not be Stage 0-1 2 3 4 used as cut-off, because it is unclear whether the same patients Cystocele 24 9 19 58 had prolapse of more than one compartments) Uterine/vault prolapse 35 21 26 28 Other symptoms not cured Enterocele 88 10 5 7 Urinary: incontinence, 7/54 persistent Rectocele 26 26 33 25 Other symptoms de novo Urinary: incontinence, 5/56 had no incontinence

Further need for treatment

Study details Participant characteristics Intervention characteristics Results Repeat POP surgery: 1 symptomatic low grade cystocele; 1 persistent stage 3 cystocele

Mesh removal surgery: 2/5 mesh exposure required mesh resection

Other outcomes Operating time (N, mean, range): 110, 89.9 (30-180) min Length of hospital stay(N, mean, range): 110, 3.75 (2-11) d

Author, year Inclusion criteria: patients underwent Intervention A (n 90): rectocele SAFETY Lim 200589 posterior vaginal wall repair with Vypro repair + mesh Mesh erosion: there was no mesh infection but minor vaginal II mesh mesh protrusion, 6-12w, 7/90 (7.8%), all treated easily with Type of prolapse: posterior Type of mesh: composite trimming without mesh removal; >=6m, 4/31 new onset (12.9%), Exclusion criteria: NR Vicryl-prolene mesh, Vypro II, two were trimmed in the outpatient clinic Publication type: full-text Ethicon, Hamburg, Germany Number of patients lost to Blood loss: requiring transfusion, intraoperative, 0/90 Study design: ambispective follow-up: 12 (13%) could not be Concomitant operation Other organ damage: bowel injury, 0/90 case series (retrospective chart contacted after 6-12w, 39/78 follow-up Incontinence, 69 (76.6%), of Other: hematoma, perioperative, 2/90 (2.2%); vaginal pain, 6- review and prospective <6m at phase II, 8/39 (20.5%) did not which 67 had suburethral sling, 12w, 1/90 (1.1%); pain with defecation, 6-12w, 1/90 (1.1%)

149 assessment) comply data collection at phase II 2 had periurethral collagen injection Setting: hospital Age (N, mean, range): 90, 59.2, 31-83 EFFICACY Additional information: Objective Failure Country: Australia Anterior only / posterior only / both: Number of surgeon: NR POP: rectocele at any degree, >=6m, 5/31 (17%) (6-9m, 2/16; 9- 0/75/15 (16.7%) of which 8 had anterior 12m, 2/11; 12-14m, 1/4) Recruitment/Treatment dates: fascia repair, 7 had anterior mesh Oct 2001 - Oct 2002 reapair Subjective Failure POP: Length of follow-up: NR, 6m Vault only / uterine only / both: vault Symptomatic, >=6m, 3/31 results available or uterine, 9/90 (10%), of which 8 had Lump sensation, 6-12w, 11/54 still present sacrospinous fixation, 1 had posterior Source of funding: NR infracoccygeal sacropexy Other symptoms not cured, 6-12w Bowel: constipation, 12/33 still present; defecation difficulties, Other information: 15/43 still present Rectoceles grade I: 9/90 (10%); Sexual: dyspareunia, 4/23 still present II: 12 (13.3%) III 65 (72.2%) Other symptoms de novo, 6-12w Not specified: 4 (4.5%) Bowel: constipation, 2/45 absent preoperation; defecation difficulties, 1/35 absent preoperation Sexual: dyspareunia, 2/36 absent preoperation Other: lump sensation, 1/24 absent preoperation

Study details Participant characteristics Intervention characteristics Results Author, year Inclusion criteria: patients underwent Intervention A (n 90): cystocele SAFETY Petros 200687 cystocele repair using the Tissue repair using the Tissue Fixation Other organ damage: bladder perforation, 1/90 Fixation System (TFS) System (TFS) Other: haematoma, 7th d, 1; fever, 2nd d, 1 and lasted 3d. Type of prolapse: anterior Exclusion criteria: NR Type of mesh: Tissue Fixation Publication type: full-text System, a 8mm wide mesh tape EFFICACY Number of patients lost to attached superiorly by a soft Objective Failure Study design: case series, follow-up: NR tissue anchor, the tape is POP: >=1 degree, 4/90 (3/45 in transverse TFS group, 1/29 in unclear prospective or Age (mean, range): 63 (29-83) multifilament, trade name or single TFS U-sling group) retrospective manufacture NR. Primary/Secondary repair: 42/48 Other symptoms not cured Setting: university hospitals (1 Additional information: Sexual: dyspareunia, 0/90 Australia, 1 European) Other information: Number of surgeon: NR Cystocele II degree, 27 Other symptoms de novo Country: Australia III degree, 63 Tape as a retro-obturator U- Sexual: dyspareunia, 0/90 sling: 29; transversely between Recruitment/Treatment dates: both arcus tendineus fascia Other outcomes NR pelvic ligaments: 45; along the Length of hospital stay(mean, range): Australian group, 1.5d (1- Length of follow-up (mean, patch of the cardinal ligament in 7); European group, 5d (4-8) range): 8m (3-15) patients with cervical ring

150 defects: 12; Source of funding: NR, but longitudinally along the arcus the first author has a financial tendineus fascia pelvic ligament: interest in the TFA instrument 4; >1 tape was used if needed.

Author, year Inclusion criteria: patients witht high- Intervention A (n 98): SAFETY Rodriguez 200588 grade cystocele who underwent repair transvaginal paravaginal repair + Mesh erosion: NR with the transvaginal paravaginal repair mesh Other: perioperative bladder wall hematoma, 1 Type of prolapse: anterior Exclusion criteria: NR Type of mesh: soft Publication type: full-text polypropylene mesh, has more EFFICACY Number of patients lost to porosity (60% more open) and Objective Failure Study design: prospective case follow-up: NR flexibility (70% more flexible) POP: 0-I, 85%; II, 13%; III, 2% series Age (mean, range): 65 (40-86) than regular polypropylene Post- PoP-Q: Parity (mean): 2.4 mesh Aa Ba Setting: university hospital Postmenopausal: 89 -3 cm 94 82 Concomitant operation -2 cm 3 13 Country: US Anterior only / posterior only / both: Incontinence: 100% had distal -1 cm 1 3 6/0/92 urethra polypropylene sling >=0 0 0 Recruitment/Treatment dates: (DUPS) regardless of symptoms Mean -2.8cm -2.84cm May 2002 – Aug 2003 Other information: of SUI Length of follow-up: NR, HRT: 42 Other symptoms not cured

Study details Participant characteristics Intervention characteristics Results efficacy data were for last SUI: 54 Additional information: Urinary: UDI-6 improvement (Urogenital Distress Inventory): pre- follow-up, the first follow-up time Number of surgeon: NR 2.3; post 0.94; was 3m, so assume >=3m SUI in 54 who had SUI preoperation: Pre- Post- Source of funding: NR No incontinence 0 38 <1 time/w 13 4 >=1 time/w 13 6 >=1 time/d 28 6

Other symptoms de novo Urinary: SUI, 3/44 who had no SUI; Urge incontinence, 2/NR who had urge incontinence

Quality of life Disease-specific: QoL due to genitourinary symptoms questionnaire, 0-6 terrible, pre- 4.7; post 1

Satisfaction Overall satisfaction with the surgery (0-100% satisfaction): average improvement of 89%?

151 Further need for treatment Repeat POP surgery: 3/98 for recurrent cystoceles

APPENDIX 6 Summary of properties of mesh/graft used by included studies

Material Commercial/trade name Manufacture, Other properties, e.g. porous, absorbability (if Studies used the country applicable) mesh/grafts Absorbable synthetic mesh Polyglactin/ Polyglacolic Vicryl mesh Ethicon, Somerville, 90% galactoside and 10% L-lactoside, 7.5ml thick, Allahdin 200656 (abs.) acid NJ, US 1.5 ounces/yard2, Mullen burst strength 60 Castelo-Branco 199865 pounds/inch2, retains at least 25% of its strength beyond 21 days in vivo. The mesh dissolves by De Ridder 200431 (abs.) hydrolysis and acts as a lattice for the formation of De Ridder 200230 dense granulation tissue. Lim 200655 (abs.) Safir 199978 Sand 200127 Verleyen on-going, Weber 200135 Absorbable biological graft Heterologous Porcine small intestinal SIS/SurgiSIS Cook Biotechnology - Chaliha 200660 Submucosa Inc., West Lafayette, Sung on-going, IN Vakili 200566

152 Dura mater Duraderm CR Bard, Covington, - - GA

Dermis Human dermis Repliform Boston Scientific - - Corp. Natick, MA

Human dermis PelviSoft BioMesh C.R. Bard, Inc., US Acellular collagen, provides benefits of nature tissue - together with the porosity of synthetic mesh.

Human dermis AlloDerm LifeCell Corp. Cadaveric, accellular, immunologically inert dermal - Branchburg, NJ matrix

Porcine dermis Pelvicol® Acellular Collagen C.R. Bard, Inc., US Derived from porcine dermis. The cellular Altman 200464 Matrix components of ski are removed, leaves the Bombieri on-going, architecture of dermal collagen bundles and elastin Cervigni 200741 (abs.) fibers intacts. The mature collagen is stabilised by De Ridder 200431 (abs.) cross-linking with diisocyanate which makes it Hviid 200543 (abs.) resistant to breakdown by naturally occurring Kocjancic 2007 (abs.), collagenases. Has been Food and Drug Leboeuf 200463 Administration (FDA)- and comite Europeen (CE)- Meschia 200734

Material Commercial/trade name Manufacture, Other properties, e.g. porous, absorbability (if Studies used the country applicable) mesh/grafts approved for permanent implantation in humans. Simsiman 200679 Vakili 200566 Verleyen on-going

Porcine dermis Fortagen Organogenesis, Inc., Porcine-derived, acellular collagen matrix graft, Paraiso 200636 Canton, MA cross-linked with 1-ethyl-3 (3-dimethylaminopropyl) carbodiimide hydrochloride (EDC).

Porcine dermis - - - Handel 200761

Dermis (unclear human or AxisTM Tutoplast® Coloplast Group, For pelvic floor reconstruction procedures - porcine) Processed Dermis Mentor Corporation

Cadaveric fascia lata Freeze-dried fascia lata - - Banked Vakili 200566

Solvent-dried fascia lata Intaxen American Medical - Gomelsky 200474 Systems, Inc. Vakili 200566 Minnetonka, MN

153 Solvent-dried fascia lata Tutoplast Mentor Corporation, Solvent dehydrated human fascia lata. FDA Gandhi 200532 Santa Barbara, serologic testing protocols are followed to screen Frederick 200573 Califonia donor tissue. Saline washes result in the osmotic Kobashi 200580 destruction of cells, reducing harvested tissues to Kobashi 200276 their fiber and mineral components, destrying Vakili 200566 bacteria, and removing tissue antigenicity. Inorganic solvents are then used to inactive viruses and prions and remove antigens. The implant is finally gamma- irradiated to ensure sterility.

Fascia lata (unclear freeze- Faslata® Allograft Tissue C.R. Bard, Inc., US - - dried or solvent-drived)

Donor fasica lata - - - Powell 200477

Autologous Fascia lata - - - Powell 200477 Vakili 200566 Combined mesh/graft (contain non-absorbable part) Device includes AvaultaTM Anterior C.R. Bard, Inc., US AvaultaTM BioSynthetic Mesh is a monofilament, - polypropylene mesh coated Biosynthetic Support System polypropylene mesh coated in the central section

Material Commercial/trade name Manufacture, Other properties, e.g. porous, absorbability (if Studies used the country applicable) mesh/grafts with absorbable porcine (for transobturator or with an absorbable, hydrophilic film of porcine collagan transischiorectal repair) collagen. The hydrophilic film minimises visceral attachment to the mesh which may occur during the healing process.

Device includes AvaultaTM Posterior C.R. Bard, Inc., US Same as above. For transobturator or - polypropylene mesh coated Biosynthetic Support System transischiorectal repair with absorbable porcine collagen

- Avaulta PlusTM Biosynthetic C.R. Bard, US - - Support System

Polypropylene mesh coated PelvitexTM polypropylene C.R. Bard, Inc., US Comprised of a monofilament, polypropylene mesh - with absorbable porcine mesh coated with a hydrophilic porcine collagen. collagen

Polypropylene mesh coated Ugytex (Pelvitex) Ugytex, softradim, Low weight (38g/m2), highly porous (89% of average De Tayrac 200783 with atelocollagen, France, distributed by porosity, pores >1.5mm), polypropylene De Tayrac on-going

154 polyethylene glycol and Bard as Pelvitex monofilament mesh. The mesh is coated with a glycerol hydrophilic film composed of atecolollagen, polyethylene glycol and glycerol.

Polypropylene and VYPRO II Ethicon, Summerville, A composite mono- and multifilamentous mesh Achtari 200590 polyglactin mesh NJ, US/Hamburg, containing 50% polypropylene and 50% absorbable Lim 200655 (abs.), Germany polyglactin 910. After absorption of polyglactin, the Lim 200589 remaining polypropylene mesh has an area weight of 35g/m2 (lightweight). It is stronger than Vypro I as it has twice the amount of polyglactin 910, Vypro II also has an additional reinforcing monofilament to give a greater overall strength once all the polyglactin 910 yarn has been absorbed. Vypro I is also lightweight.

- Vypro Johnson & Johnson, - Cronje 200667 Brussels, Belgium Non-absorbable synthetic mesh Device includes ApogeeTM Vault Suspension American Medical Sterile, single use product consisting of two stainless Meschia 200759 (abs., polypropylene mesh System (for vaginal vault Systems, US (AMS) steel curved needles and a polypropylene mesh Perigee-Apogee system), repair) implanted. The same mesh is available in three Zylstra on-going configurations: TAPE, CAPE, and BIO-CAPE.

Material Commercial/trade name Manufacture, Other properties, e.g. porous, absorbability (if Studies used the country applicable) mesh/grafts

Device includes PerigeeTM System with American Medical Sterile, single use system consisting of four stainless Nguyen 200753 (abs.) polypropylene mesh InteProTM (for anterior repair) Systems, US steel helical needles and an assembly including unclear this mesh or the AMS InteProTM, a large-pore monofilament one below, Zylstra on- polypropylene mesh. going

Device includes PerigeeTM System for American Medical Sterile, single use system consisting of four stainless Zylstra on-going polypropylene mesh biologic InteGraftTM (for Systems, US steel helical needles, a precut biological graft anterior repair) material to which four polypropylene monofilament self-fixing mesh appendages could attach to.

Device includes Straight-InTM sacral American Medical This system includes Straight-In bone screw Zylstra on-going polypropylene mesh colpopexy system with Systems, US inserter. The Y-sling is constructed of polypropylene InteProTM large pore suture knitted together to form a mesh. The ‘wings’ polypropylene Y-sling (for of the Y-shaped sling are secured to the anterior and vault prolapse repair) posterior sides of the vaginal cuff. The ‘stem’ of the Y-sling is attached via sutures to bone screws located in the sacrum.

155 Polypropylene mesh BioArc device American Medical - Amrute 2006 Systems, US

Polypropylene mesh Atrium Atrium Medical Type I, non-absorbable, monofilamentous, Achtari 200590 Corporation, Hudson, macroporous mesh with an area weight of 92g/m2, Dwyer 200484 N.H. U.S. pore size (interfibre space) 800μm

Polypropylene mesh Polyform™ Synthetic Mesh Boston Scientific, US Made from uncoated monofilament macroporous - Polyproplene, stretchable.

Polypropylene mesh Marlex CR Bard, US - Costantini 200582 Flood 199872 Julian 199662

Polypropylene mesh NovasilkTM Polypropylene Coloplast Group, A soft, supple mesh, weighing 21g/m2 and provides - Mesh (for a variety of pelvic Mentor Corporation multi-directional elasticity. Macroporous, floor procedures, including monofilament polypropylene. cystocele, rectocele, and vaginal vault prolapse)

Device includes Gynecare Prolift* Ethicon, Inc., US This system consists of precut Gynecare Altman 200729 polypropylene mesh Total/Anterior/Posterior Gynemesh* PS implant, guide, retrieval devices, and Fatton 200793

Material Commercial/trade name Manufacture, Other properties, e.g. porous, absorbability (if Studies used the country applicable) mesh/grafts Pelvic Floor Repair System cannulas. The Total mesh implant is shaped for Withagen on-going performing a total vaginal repair which has 6 straps; the Anterior mesh implant is shaped for repair of anterior vaginal defects which has 4 straps; the Posterior mesh implant is shaped for repair of posterior and/or apical vaginal vault defects.

Polypropylene mesh Gynemesh, Gynecare Ethicon, France/ New Monofilament mesh, 90g/m2 Cervigni 200741 (abs.), (Prolene mesh) Brunswick, Collinet 200691 NJ/Edinburgh/Belgiu De Tayrac 200671 m Deffieux 200786 Milani 200528 Vakili 200566

Polypropylene mesh Gynemesh-Soft, Gynecare Ethicon, France Monofilament mesh, 42.7g/m2 Al-Nazer 200737 (abs.) (Prolene soft mesh) Ali 200638 (abs.) Collinet 200691 Deffieux 200786 Lim 200758 (abs.)

156 Polypropylene mesh MINIMESH® (for anterior Mpathy Medical Light (19g/m2), large pore knit construction (2mm), - and posterior repair) Devices Ltd., UK 100 micron interstitial spaces for tissue integration and defence against infection

Polypropylene mesh Parietene light Sofradim, Trevoux, Light monofilament, 6x11cm Hiltunen 200642 (abs.) France

Polypropylene mesh - - - Handel 200761

Soft polypropylene mesh - - More porosity (60% more open) and flexibility (70% Rodgriguez 200588 more flexible) than regular polypropylene mesh.

100% polyester single side Polyatex A4 100x150mm Cousin Biotech, - Rozet 200585 covered silicone mesh country unkown

Polyethylene tetraphalate Mersilene Ethicon Type III Cosson 200292

Polytetraflourethylene Gore-Tex - A microporous multifilament mesh (Type II) -

Popytetrafluoroethylene Teflon Bard, MurrayHill, NJ - -

Material Commercial/trade name Manufacture, Other properties, e.g. porous, absorbability (if Studies used the country applicable) mesh/grafts

- Tissue Fixation System - Was designed as a universal instrument for repair of Petros 200687 (TFS) damaged ligaments and fascia, a 8mm wide mesh tape attached superiorly by a soft tissue anchor, the tape is multifilament

157

APPENDIX 7 Summary of operation techniques used by included studies

Name (other names) Description Studies used the operation Anterior repair Anterior colporrhaphy + NR Al-Nazer 200737 mesh (abs.) Transvaginal mesh repair The uniformly sized and shaped polypropylene mesh is passed through the arcus tendineus fascia pelvis using 4 Altman 200729 using Prolift (device) extension arms in the anterior vaginal compartment. Fatton 200793 Anterior repair + mesh NR Cervigni 200741 (abs.) Traditional anterior repair The vesical fascia and bladder were reflected off the vaginal skin, and then lateral transverse support sutures Chaliha 200660 were placed at the junction of the anterior wall of the vagina and fascia edge. The skin was then trimmed, closed with a continuous suture. Traditional anterior repair + After placement of the transverse sutures, the fascia was dissected further laterally extending under the subpubic Chaliha 200660 small intestine submucosa arch to the pelvic side wall. The 4-layer graft was pre-soaked for 5min in a saline solution, then folded to produce graft a double thickness, and then cut into an octagonal shape which fitted well while being under tension over the anterior wall fascia. The graft was fixed using sutures placed circumferentially around the graft to the anterior wall fascia. The vaginal skin was trimmed and closed as with the traditional anterior repair. Transvaginal + mesh Begins with anterior colpotomy. Cystocele correction warrants bilateral trans-obturator passage of the mesh in Collinet 200691 order to suspend it. The anterior, sub-vesical strap is inserted in arcus tendineus fascia pelvis and the posterior, 158 sub-vesical stap into the arcus tendineus 1cm from the siatic spine. Sacrocolpopexy for anterior Sacrocolpopexy was performed by placing a mesh both anteriorly and posteriorly to the vagina and extending it to Cronje 200667 prolapse the sacrum Anterior repair (Raz 4 defect Cystocoele reduction was done by a plug of mesh and the reapproximation of the perivesical fascia De Ridder 200431 repair) with mesh (abs.) Trans-vaginal cystocele Mesh was placed from the retropubic space to the inferior part of the bladder in a tension-free fashion. The tip of De Tayrac 200671 repair + mesh the mesh was inserted on both sides into the retropubic space and left in a tension-free fashion, i.e. the mesh is not fixed to any structures. The lower part of the mesh was placed under the bladder and left in a tension-free fashion. Transvaginal repair of Mesh (precut measuring 6cm x 15cm) was used (no further details) after the paraversical fossa was opened on Deffieux 200786 cystocele using mesh each side. Anterior repair + mesh After incision, dissection and midline placation of the fascial layer, mesh (2x15cm) with a widened elliptical mid- Dwyer 200484 portion (5cm) was placed under the bladder base and each lateral extension was positioned onto the iliococcygeal fascia anterior to the ischial spines. The mesh was usually unsutured. Anterior colporrhaphy After incision and dissection, a piece of Marlex mesh approximately 4cm long and 1cm wide is placed under the Flood 199872 reinforced with Marlex mesh vesicourethral junction, extending up into the periurethral tunnels, and trimmed so that there is no bunching. It is secured at the level of the bladder neck with sutures. Excess vagina is trimmed, if necessary, taking great care to leave adequate vagina to cover the mesh, and then closed with sutures. The cadeveric prolapse After incision and dissection, bone anchors with sutures are placed into the pubic bone. 4 sutures are placed into Frederick 200573 repair with sling (CaPS) the levator complex on either side and tagged. A fifth suture is placed into the inside of the cervix or vaginal cuff Kobashi 200276 for superior support of the fascia. A 4x8 cm piece of cadeveric fascia lata is cut into T-configuration. The bone

Name (other names) Description Studies used the operation anchor sutures are placed through the sling portion of the fascia. The proximal suture is placed medially in relation to the distal suture to ensure that the lower edge of the sling is snug against the bladder neck. One side of the sling is secured in place up to the bone, and the distal edge of the sling is tacked in place. A small right angle clamp is placed between the urethra and the sling. The anchor sutures are passed through the fascia on this contralateral side and tied with the right angle in place. The edge of the remaining portion of the fascia used to repair the cystocele defect is folded over and the previously placed levator sutures are passed through the edge of the fascia patch bilaterally to reduce the cystocele. Finally, the apical sutures are passed through the proximal aspect of the fascia and secured in place. Standard anterior Wide placation of the endopelvic connective tissue at the midline Gandhi 200532 colporrhaphy Standard anterior A 2x4cm piece of solvent dehydrated cadaveric fascia was used and sutured under the anterior repair as a barrier Gandhi 200532 colporrhaphy + cadaveric fascia lata graft Cystocele repair + porcine After dissection, 3 pair of sutures are inserted into the arcus tendineus fascia pelvis (ATFP) and a 6x8cm Gomelsky 200474 dermis interposition graft rehydrated porcine dermis is tailored to fit in the defect. The graft is then sutured to each ATFP, making sure to leave a slight degree of redundancy in the middle. The proximal end of the graft is sutured to the vaginal apex, incorporating the cardinal ligament complex. The vaginal wall is then closed. Traditional colporrhaphy The uterosacral-cardinal ligmant complex and pubocervical fascia were plicated in the midline Handel 200761 Cystocele repair + porcine The lateral edges of a 3x7cm graft segment were anchored to the pelvic sidewalls at the level of the levator fascia Handel 200761

159 dermal graft/ polypropylene bilaterally. The uterosacral-cardinal ligament complex was plicated in the midline if present and the apical edge of mesh the graft was then fixed in the midline to this area of placation. Additional placation of the pubocervical fascia was not performed. The distal edge of the graft was affixed in the midline to the bladder at least 1cm proximal to the bladder neck. The graft was trimmed as needed. Traditional anterior NR Hiltunen 200642 colporraphy + mesh (abs.) Anterior repair + mesh NR Hviid 200543 (abs.) Support defect repair Urethropexy, anterior colporrhaphy and paravaginal repair Julian 199662 Support defect repair with Similar with support defect repair except that anterior vaginal segment was reinforced by sewing the synthetic Julian 199662 mesh nonabosorbable mesh from the urethrovesicle junction anteiorly to the vaginal apex posteiorly and to the junction of the obturator and levator fascia at the lateral margins of this space. Anterior repair + mesh NR Kocjancic 200747 (abs.) 4-defect repair After incision and dissection, lateral defect of the cystocele was repaired by the first sutures. The second sutures Leboeuf 200463 included the pubourethral ligaments, entrance of the levator to the mid urethra, and periurethral fascia to provide a vaginal wall sling for SUI. A small suprapubic incision was made, and sutures were transformed from the vaginal wall to the suprapubic area. Sutures were firmly applied to the cardinal ligaments approximating them to the midline and providing support to the vaginal vault. Finally, sutures were applied to approximate the perivesical fascia, correcting central cystocele defect. 4-defect repair + mesh Consisted of the same corrections of the lateral and central defects described above, except for the omission of Leboeuf 200463 elevating the urethra. The support mechanism of the mid-urethra complex was provided by either tension-free Safir 199978 tape or a Pelvicol sling. Pelvicol was added at the end of the modified 4-defect repair, by placing sutures through

Name (other names) Description Studies used the operation the urethropelvic and cardinal ligaments. It was then held in place firmly over the plicated perivesical fascia. The vaginal wass was closed over the Pelvicol mesh. This technique was previously described by Safir 1999. Standard anterior Plicate the pubocervical fascia Meschia 200734 colporraphy Standard anterior After standard placation, the 4x7cm re-hydrated porcine dermis is trimmed and positioned over the pubocervical Meschia 200734 colporrhaphy + porcine fascial and anchored with 2 pair of sutures proximal to the endopelvic fascia and distal to the uterosacral-cardinal dermis stumps or to the cervical ring when the uterus was retained. Anterior repair + mesh A conventional anterior repair plus the placement of a mesh: after incision and dissection, the fascial tissue was Milani 200528 plicated anteriorly with sutures. The mesh was inserted and fixed using sutures. Finally, the excess vaginal tissue was resected. Cystocele repair using the Tape as a retro-obturator U-sling in 29 patients, transversely between both arcus tendineus fascia pelvic Petros 200687 Tissue Fixation System ligaments in 45 patients, along the patch of the cardinal ligament in patients with cervical ring defects in 12 (TFS) patients, and longitudinally along the arcus tendineus fascia pelvic ligament in 4; >1 tape was used if needed. Anterior vaginal wall Previously described (Chesson 1999). Fascia was anchored to the arcus and underlying support tissue using Powell 200477 hammock (AVWH) + fascia sutures. The necessary width of the hammock depend on the bi-ischial diameter of each patient, although most lata required a graft width of approximately 12cm to span the distance between the arcus on each side. Transvaginal paravaginal After repair of the central defect of the cystocele, a paravaginal repair of the lateral defect was performed by using Rodgriguez 200588 repair + mesh a circular 5x5cm soft polypropylene mesh attached proximately to the sacrouterine/cardinal ligament, distally to the bladder neck, and laterally to the infralevator obturator fascia. 79 160 Vaginal paravaginal repair A 4x12cm graft was bilaterally attached to the arcus tendious fascial pelvis. The graft is modified into a trapezoid Simsiman 2006 (VPVR) with Pelvicol pattern and typically has 12cm width at the level of the ischial spine and 6-8cm at the most distal portion Restorative repairs (no graft) NR Vakili 200566 Graft-augmented repair The technique for anterior vaginal wall fixation involved fashioning a trapezoid shape from the graft to cover the Vakili 200566 entire vaginal wall. The graft was fixed bilaterally to the arcus tendineus fascia pelvis. The graft was fixed apically either to the vaginal apex or bilaterally to the uterosacral or sacrospinous ligaments if a concurrent vault suspension was performed. The technique was similar for posterior graft with lateral fixation to the arcus tendineus fascia rectovaginalis. Standard anterior Suburethral placation and muscularis plication Weber 200135 colporrhaphy Ultralateral anterior After a midline anterior vaginal incision was made, dissection was performed laterally to the limits of the pubic Weber 200135 colporrhaphy rami on each side. After the suburethral placation, the paravaginal connective tissue was plicated. Standard anterior After midline placation of the vaginal muscularis, a piece of polyglactin 910 mesh (Vicryl) was cut to fit the space Weber 200135 coporrhaphy + mesh over the placation. The mesh was anchored at the lateral limits of the dissection, and the vaginal epithelium was closed over the mesh. Posterior repair Transvaginal mesh repair In the posterior compartment, the uniformly sized and shaped polypropylene mesh is placed through a Altman 200729 using Prolift transgluteal approach, and the two extension arms are fixed to the sacrospinous ligment. Fatton 200793 Rectocele repair + porcine After incision and dissection, the implant was placed in between the vaginal mucosa and the rectovaginal fascia, Altman 200464 collagen mesh minimising direct contact between the collagen mesh and rectum. Neither site-specific repair nor levator sutures were performed. The 4x7 implant was secured to the levator muscle, rectovaginal fascia, and perineal body. The

Name (other names) Description Studies used the operation implant was covered by the remaining mucosa after excision of redundant mucosa. Posterior wall repair (‘usual The rectocele was pushed into the rectum, the levato ani muscle was approximated and sutured, posterior vaginal Castelo-Branco procedure’) mucosa was trimmed away and continuous suture was placed at the pex of the vaginal mucosa and tied down. 199865 Posterior wall repair (‘usual After the rectocele was pushed into the rectum, the mesh was suited to the size of the rectocele and placed over Castelo-Branco procedure’) + mesh and fixed with suture to the margins of both levator an muscles from the apex down to the fourchette. The closure 199865 of the posterior wall was completed without the removal of any vaginal mucosa. Transvaginal + mesh Begins with posterior colpotomy. The mesh is then transperineal, on either side, one strap of the mesh is passed Collinet 200691 into the pararectal space through the sacrosciatic ligament to become exteriorised in incisions located outside and below the arms. Posterior repair + mesh After incision and dissection, any fascial defect in the rectovaginal septum was repaired using sutures. Mesh Dwyer 200484 (10x15cm) was fashioned in a Y shape. The arms of the Y were placed onto the sacrospinous ligaments bilaterally with the main body overlaying the repaired rectovaginal fascia and the perineal body. The mesh was sutured superiorly, laterally and onto the perineal body. Posterior repair + mesh The defect specific repair was performed, and the meshes were implanted with an onlay method with their four Lim 200655 (abs.) corners secured onto fascia tissue. Rectocele repair + mesh After dissection, a piece of mesh was fashioned and laid over the fascial defect, and sutured in a tension-free Lim 200589 manner. The rectovaginal defect itself was not closed. Posterior site-specific repair The surgical technique involved removal of a triangular segment of the posterior vaginal epithelium overlying the Kobashi 200580 + cadaveric fascia lata rectocele, undermining of the posterolateral vaginal flaps to the level of the levator ani, perineoplasty,

161 preplacement of six sutures into the levator ani (arcus complex), the apex of the rectocele, and the ‘inside’ of the perineoplasty, interposition of a 4x7cm graft to transverse the rectovaginal fascial defect, and close of the vaginal epithelium and perineal skin. Posterior repair + mesh A conventional posterior repair plus the placement of a mesh: after incision and dissection, the fascial tissue was Milani 200528 plicated anteriorly with sutures. The mesh was inserted and fixed using sutures. Finally, the excess vaginal tissue was resected. Anterior and/or posterior prolapse repair Anterior and/or posterior Involves using the mesh as overlay to reinforce fascia repair. For anterior compartment repair, the mesh, a Achtari 200590 mesh overlay widened elliptical midportion and two lateral arms, is placed under the bladder base with each lateral extension positioned onto the iliococcygeal fascia. The mesh overlay is sutured at the anterior and posterior margins for stabilisation and to prevent folding. For posterior compartment repair, any fascia defect in the rectovaginal septum is repaired using sutures. The mesh is fashioned in a Y shape and placed onto the sacrospinous ligaments bilaterally with the main body of mesh overlaying the repaired rectovaginal fascia and the perineal body. Combined anterior and posterior compartment repair involves placement of the mesh fashioned into a cross shape and along the arms of the mesh extending along the iliococcygeal fascia anterior to the ischial spines and the long stem of the cross extending to the perineum. Total repair NR Altman 200729 4-point fixation After vaginal dissection, a special fashioned ‘H’ shaped polypropylene mesh is positioned and fixed at 4-points, Amrute 200781 with a single piece of mesh, the anterior arms provide mid-urethral and bladder neck support, the mid-portion of the mesh connects anterior compartment defects, and the posterior arms aid in vaginal vault suspension. A tension-free method is used. Anterior, posterior or NR Allahdin 200656

Name (other names) Description Studies used the operation paravaginal repair + mesh (abs.) Anterior and/or posterior or The mesh has three distinct parts. The anterior part is inserted between the bladder and the vagina and secured Altman 200729 total mesh with Prolift bilaterally by two arms through each obturator foramen. The posterior part is placed between the rectum and the Fatton 200793 vagina and is secured bilaterally by one arm passing through each ischiorectal fossa and sacrospinous ligament. The intermediate section corresponding to the vaginal apex separates the anterior and posterior parts and can be cut if needed. Instruments were designed to facilitate proper implant placement. Sacropexy with anterior and Anterior and posterior lengths of Mersilene mesh were sutured to the vaginal wall. The strips of mesh were Cosson 200292 posterior mesh reinforcement approximately 2-3cm wide and 4-5cm long. When conservation of the uterus was required, the anterior strip was Y-shaped, its 2 arms going through the broad ligament before reaching the sacral promontory. The mesh was suspended to the anterior longitudinal sacral ligament and buried by closing the peritoneum. Hysterectomy followed by Two meshes were attached to anterior and posterior vaginal wall after dissection. Meshes are cut into different Costantini 200582 sacropexy or colposacropexy shapes: both rectangular in hysterectomy followed by sacropexy, one rectangular and one Y-shaped in with uterine conservation, colposacropexy with uterine conservation. both with anterior and posterior reinforcement Anteior and/or posterior Anterior or posterior fascia was not plicated before mesh placement. The mesh was implanted anteriorly either De Tayrac 200671 repair + mesh with two arms into the retropublic space or with 2-4 arms into the obturator foramen. The mesh was implanted posteriorly either with two arms sutured to the sacrospinous ligaments or with two arms via the infracoccygeal sacropexy. The excess vaginal skin was not excised to avoid direct contact between the vaginal scar and the

162 mesh. Anterior and posterior repair After incision and dissection, anteriorly, midline placation of the fascial layer was performed; posteriorly, any Dwyer 200484 + mesh fascial defect in the rectovaginal septum was repaired. Mesh was then fashioned into a cross shape and laid along the repair with the arms 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. Traditional anterior and Performed with fascia plication Lim 200758 (abs.) posterior repair Traditional anterior and NR Lim 200758 (abs.) posterior repair + mesh anterior and posterior mesh NR Meschia 200759 repair without hysterectomy: (abs.) Perigee-Apogee system Anterior and posterior repair Laparoscopic route. Dissection of the posterior portion of the vaginal in order to fix the posterior mesh on the Rozet 200485 with mesh with both meshes levator ani muscles. Adequate dissection of inter vesico-vaginal space through the broad ligament in order to fix sutured to the promontory the anterior prosthesis on the anterior wall. The surgery ends with both meshes sutured to the promontory. Standard anterior +/- A standard mattress suture repair of the anterior endopelvic connective tissue, excess vaginal epithelium was Sand 200127 posterior colporrhaphy excised; posterior colporrhaphy was similarly performed. Standard anterior +/- One piece of mesh was folded into the imbricated endopelvic connective tissue underneath the trigone, a second Sand 200127 posterior colporrhaphy with piece was folded into the imbricated endopelvic connective tissue anterior to the vaginal cuff, and the third piece mesh was used during posterior colporrhaphy, in a similar fashion, cephalad to the deep transverse perineal muscles.

APPENDIX 8 Number and type of excluded studies; reasons for exclusion

Case series with less than 50 subjects (N= 43)

Abbott PD, McDonald TM. Polyethylene terephlatate grafts for repair of enteroceles and rectoceles. Journal of Pelvic Medicine & Surgery 2004;10(1):27-9.

Altman D, Mellgren A, Zetterstrom J. Rectocele repair using biomaterial augmentation: current documentation and clinical experience. Obstetrical & Gynecological Survey 2005;60(11):753-60.

Altman D, Zetterstrom J, Lopez A, Anzen B, Falconer C, Hjern F et al. Functional and anatomic outcome after transvaginal rectocele repair using collagen mesh: a prospective study. Diseases of the Colon & Rectum 2005;48(6):1233-41.

Altman D, Lopez A, Gustafsson C, Falconer C, Nordenstam J, Zetterstrom J. Anatomical outcome and quality of life following posterior vaginal wall prolapse repair using collagen xenograft. International Urogynecology Journal 2005;16(4):298-303.

Athanasiou S, Matthaiou DK, Falagas ME. Vaginal mesh infection due to Bacteroides melaninogenicus: a case report of another emerging foreign body related infection. Scandinavian Journal of Infectious Diseases 2006;38(11-12):1108-10.

Azanjac B, Jorovic M. Transperineal repair of recurrent rectocele using polypropylene mesh. Techniques in Coloproctology 1999;(1):39-41.

Baessler K, Schuessler B. Abdominal sacrocolpopexy and anatomy and function of the posterior compartment. Obstetrics & Gynecology 2001;97(5 Pt 1):678-84.

Baessler K, Stanton SL. Sacrocolpopexy for vault prolapse and rectocele: do concomitant Burch colposuspension and perineal mesh detachment affect the outcome? American Journal of Obstetrics & Gynecology 2005;192(4):1067-72.

Biertho I, Dallemagne B, Dewandre JM, Markiewicz S, Monami B, Wahlen C et al. Intravaginal slingplasty: short term results. Acta Chirurgica Belgica 2004;104(6):700-4.

Canepa G, Ricciotti G, Introini C, Vigliercio G, Puppo P. Horseshoe-shaped marlex mesh for the treatment of pelvic floor prolapse. European Urology 2001;39 (Suppl 2):23-6.

Chung SY, Franks M, Smith CP, Lee JY, Lu SH, Chancellor M. Technique of combined pubovaginal sling and cystocele repair using a single piece of cadaveric dermal graft. Urology 2002;59(4):538-41.

Clemons JL, Myers DL, Aguilar VC, Arya LA. Vaginal paravaginal repair with an AlloDerm graft. American Journal of Obstetrics & Gynecology 2003;189(6):1612-8.

Costantini E, Lombi R, Micheli C, Parziani S, Porena M. Colposacropexy with Gore-tex mesh in marked vaginal and uterovaginal prolapse. European Urology 1998;34(2):111-7.

Costantini E, Del Zingaro M, Pajoncini C, Fornetti P, Zucchi A, Porena M. Sacropexy in marked vaginal and utero-vaginal prolapse. European Urology Supplements 2002;1(1):156- 7.

Cutner A, Spiteri M. The use of submucosal small intestinal mesh in laparoscopic sacrocolpopexy and posterior vaginal repair. Gynecological Surgery 2005;(3):187-9.

163

David-Montefiore E, Barranger E, Dubernard G, Detchev R, Nizard V, Darai E. Treatment of genital prolapse by hammock using porcine skin collagen implant (Pelvicol). Urology 2005;(6):1314-8.

De Tayrac R, Picone O, Chauveaud-Lambling A, Fernandez H. A 2-year anatomical and functional assessment of transvaginal rectocele repair using a polypropylene mesh. International Urogynecology Journal 2006;17(2):100-5.

Dell JR, O'Kelley KR. PelviSoft BioMesh augmentation of rectocele repair: The initial clinical experience in 35 patients. International Urogynecology Journal 2005;(1):44-7.

Fox SD, Stanton SL. Vault prolapse and rectocele: assessment of repair using sacrocolpopexy with mesh interposition. BJOG: An International Journal of Obstetrics & Gynaecology 2000;107(11):1371-5.

Gadonneix P, Ercoli A, Salet-Lizee D, Cotelle O, Bolner B, Van Den AM et al. Laparoscopic sacrocolpopexy with two separate meshes along the anterior and posterior vaginal walls for multicompartment pelvic organ prolapse. Journal of the American Association of Gynecologic Laparoscopists 2004;11(1):29-35.

Groutz A, Chaikin DC, Theusen E, Blaivas JG. Use of cadaveric solvent-dehydrated fascia lata for cystocele repair--preliminary results. Urology 2001;58(2):179-83.

Hilger WS, Cornella JL. Rectovaginal fistula after Posterior Intravaginal Slingplasty and polypropylene mesh augmented rectocele repair. International Urogynecology Journal 2006;17(1):89-92.

Hung MJ, Liu FS, Shen PS, Chen GD, Lin LY, Ho ES. Factors that affect recurrence after anterior colporrhaphy procedure reinforced with four-corner anchored polypropylene mesh. International Urogynecology Journal 2004;15(6):399-406.

Ismail SIMF, Pugh DHO. Mesh erosion and prolapse recurrence following repeat anterior repair with Prolene mesh. Journal of Obstetrics & Gynaecology 2007;(1):94.

Kohli N, Miklos JR. Dermal graft-augmented rectocele repair. International Urogynecology Journal 2003;14(2):146-9.

LaSala CA, Schimpf MO. Occurrence of postoperative hematomas after prolapse repair using a mesh augmentation system. Obstetrics & Gynecology 2007;109(2 Pt2):569-72.

Lazarou G, Powers K, Pena C, Bruck L, Mikhail MS. Inflammatory reaction following bovine pericardium graft augmentation for posterior vaginal wall defect repair. International Urogynecology Journal 2005;(3):242-4.

Lyons TL, Winer WK. Laparoscopic rectocele repair using polyglactin mesh. Journal of the American Association of Gynecologic Laparoscopists 1997;4(3):381-4.

Marinkovic SP, Stanton SL. Triple compartment prolapse: sacrocolpopexy with anterior and posterior mesh extensions. BJOG: An International Journal of Obstetrics & Gynaecology 2003;110(3):323-6.

Mercer-Jones MA, Sprowson A, Varma JS. Outcome after transperineal mesh repair of rectocele: a case series. Diseases of the Colon & Rectum 2004;47(6):864-8.

Migliari R, Usai E. Treatment results using a mixed fiber mesh in patients with grade IV cystocele. Journal of Urology 1999;161(4):1255-8.

164

Migliari R, De Angelis M, Madeddu G, Verdacchi T. Tension-free vaginal mesh repair for anterior vaginal wall prolapse. European Urology 2000;38(2):151-5.

Ng CC, Chong CY. The effectiveness of transvaginal anterior colporrhaphy reinforced with polypropylene mesh in the treatment of severe cystoceles. Annals of the Academy of Medicine, Singapore 2006;35(12):875-81.

Nicita G. Transvaginal pelvic floor reconstruction with a polypropylene mesh in the treatment of incontinent genito-urinary prolapse. Acta Urologica Italica 1997;(4):275-9.

Nicita G. A new operation for genitourinary prolapse. Journal of Urology 1998;160(3 Pt 1):741-5.

Oster S, Astrup A. A new vaginal operation for recurrent and large rectocele using dermis transplant. Acta Obstetricia et Gynecologica Scandinavica 1981;60(5):493-5.

Salomon LJ, Detchev R, Barranger E, Cortez A, Callard P, Darai E. Treatment of anterior vaginal wall prolapse with porcine skin collagen implant by the transobturator route: preliminary results. European Urology 2004;(2):219-25.

Sentilhes L, Sergent F, Resch B, Verspyck E, Descamps P, Marpeau L. Midterm follow-up of high-grade genital prolapse repair by the trans-obturator and infracoccygeal hammock procedure after hysterectomy. European Urology 2007;(4):1065-72.

Sola V, Pardo J, Ricci P, Guiloff E. Tension free monofilament macropore polypropylene mesh (Gynemesh PS) in female genital prolapse repair. International Braz J Urol 2006;32(4):410-4.

Washington JL, Somers KO. Laparoscopic paravaginal repair: a new technique using mesh and staples. Journal of the Society of Laparoendoscopic Surgeons 2003;7(4):301-3.

Watson SJ, Loder PB, Halligan S, Bartram CI, Kamm MA, Phillips RK. Transperineal repair of symptomatic rectocele with Marlex mesh: a clinical, physiological and radiologic assessment of treatment. Journal of the American College of Surgeons 1996;183(3):257-61.

Wheeler II TL, Richter HE, Duke AG, Burgio KL, Redden DT, Varner RE. Outcomes with porcine graft placement in the anterior vaginal compartment in patients who undergo high vaginal uterosacral suspension and cystocele repair. American Journal of Obstetrics & Gynecology 2006;(5):1486-91.

Yamada BS, Govier FE, Stefanovic KB, Kobashi KC. Vesicovaginal fistula and mesh erosion after Perigee (transobturator polypropylene mesh anterior repair). Urology 2006;68(5):1121 e5-7.

Results for different operations not reported separately (N=13)

Antiphon P, Elard S, Benyoussef A, Fofana M, Yiou R, Gettman M et al. Laparoscopic promontory sacral colpopexy: is the posterior, recto-vaginal, mesh mandatory? European Urology 2004;45(5):655-61.

Baessler K, Hewson AD, Tunn R, Schuessler B, Maher CF. Severe mesh complications following intravaginal slingplasty. Obstetrics & Gynecology 2005;106(4):713-6.

Brubaker L, Cundiff G, Fine P, Nygaard I, Richter H, Visco A et al. A randomized trial of colpopexy and urinary reduction efforts (CARE): Design and methods. Control Clin Trials 2003;24(5):629-42.

165

Collopy BT, Barham KA. Abdominal colporectopexy with pelvic cul-de-sac closure. Diseases of the Colon & Rectum 2002;45(4):522-6.

Cosson M, Occelli B, Narducci F, Ego A, Querleu D, Crepin G. Causes of failure of abdominal colposacropexy for the treatment of genital and vaginal vault prolapse. Journal of Gynecologic Surgery 2000;(4):141-8.

Cronje HS. Colposacrosuspension for severe genital prolapse. International Journal of Gynaecology & Obstetrics 2004;85(1):30-5.

Moretti M, Cichero A, Malcangi D, Pittaluga P, Varaldo M. Tension-free prothesic surgery for stress incontinence and cystorectocele: Preliminary results. Acta Urologica Italica 1998;(6):297-300.

Seracchioli R, Hourcabie JA, Vianello F, Govoni F, Pollastri P, Venturoli S. Laparoscopic treatment of pelvic floor defects in women of reproductive age. Journal of the American Association of Gynecologic Laparoscopists 2004;11(3):332-5.

Su KCH, Mutone MF, Terry CL, Hale DS. Abdominovaginal sacral colpoperineopexy: patient perceptions, anatomical outcomes, and graft erosions. International Urogynecology Journal 2007;18(5):503-11.

Sullivan ES, Longaker CJ, Lee PYH. Total pelvic mesh repair: A ten-year experience. Diseases of the Colon & Rectum 2001;(6):857-63.

Thompson PK, Pugmire JE, Sangi-Haghpeykar H. Abdominal sacrocolpopexy utilizing Gore- Tex in genital prolapse: Unresolved issues. Journal of Pelvic Medicine & Surgery 2004;(6):311-7.

Vancaillie TG. MycroMesh is not a suitable soft tissue prosthesis for repair of the defective vaginal wall. Journal of the American Association of Gynecologic Laparoscopists 2003;10(3):424-5.

Verdaasdonk EG, Bueno de Mesquita JM, Stassen LP. Laparoscopic rectovaginopexy for rectal prolapse. Techniques in Coloproctology 2006;10(4):318-22.

Case report without safety data (N= 7)

Addison WA, Timmons MC, Wall LL, Livengood CH, III. Failed abdominal sacral colpopexy: observations and recommendations. Obstetrics & Gynecology 1989;(3 Pt 2):480-3.

Arunkalaivanan AS, Kaur H, Devarajan R. Management of complex vault prolapse with hydronephrosis by using porcine mesh (SIS) sacrocolpopexy and colposuspension. Journal of Obstetrics & Gynaecology 2006;(2):179-81.

Balaloski SP, Richards SR, Singh E. Conservative management of delayed suprapubic abscess after laparoscopic Burch colposuspension using nonabsorbable polypropylene mesh. Journal of the American Association of Gynecologic Laparoscopists 1999;6(2):225-8.

Koscinski T. Tension-free technique in the management of rectocele - Preliminary report. Proktologia 2006;(2):118-24.

Lilford RJ, Johnson N, Batchelor A. repair of prolapse by means of a living ligament. Br J Obstet Gynaecol 1993;100(9):859-60.

166

Parker MC, Phillips RKS. Repair of rectocoele using Marlex mesh. Annals of the Royal College of Surgeons of England 1993;(3):193-4.

Yamada BS, Kobashi KC. Urogenital prolapse and occult stress urinary incontinence. Nature Clinical Practice Urology 2007;4(1):55-8.

Not a report of primary research (N= 6)

Altman D, Zetterstrom J, Mellgren A, Gustafsson C, Anzen B, Lopez A. A three-year prospective assessment of rectocele repair using porcine xenograft. Obstetrics & Gynecology 2006;(1):59-65.

Beer M, Kuhn A. Surgical techniques for vault prolapse: a review of the literature. European Journal of Obstetrics, Gynecology, & Reproductive Biology 2005;119(2):144-55.

Maher C, Baessler K. Surgical management of anterior vaginal wall prolapse: an evidencebased literature review. International Urogynecology Journal 2006;17(2):195-201.

Maher C, Baessler K. Surgical management of posterior vaginal wall prolapse: an evidencebased literature review. International Urogynecology Journal 2006;17(1):84-8.

Novara G, Artibani W. Surgery for pelvic organ prolapse: current status and future perspectives. Current Opinion in Urology 2005;15(4):256-62.

Novara G, Galfano A, Mancini M, Ficarra V, Artibani W. Critical Assessment of Pelvic Floor Surgical Reconstruction Outcome. Eau-Ebu Update Series 2006;(5):202-13.

No relevant outcomes reported or no data (N=5)

Brubaker L. Abdominal sacrocolpopexy with Burch colposuspension to reduce urinary stress incontinence. N Engl J Med 2006;354(15):1557-66.

D'Elia A, Grossi FS, Larocca L, Barnaba D, Sallustio G, Raguso G. Colposuspension: Burch technique in extraperitoneal laparoscopy. Acta Urologica Italica 1998;(5):255-8.

Fischer A. Prolapse Surgery Using Biomaterials. European Urology Supplements 2002;(10):29-32.

Hendee AE, Berry CM. Abdominal sacropexy for vaginal vault prolapse. Clinical Obstetrics & Gynecology 1981;24(4):1217-26.

Steinberg AC, Oyama IA, Sastry D, Vakili B, Caraballo R, Maccarone JL. Vaginal paravaginal repair with the use of porcine dermis. Journal of Pelvic Medicine & Surgery 2006;(3):141-4.

Prolapse repair did not use mesh (N =14)

Bai SW, Jeon JD, Chung KA, Kim JY, Kim SK, Park KH. The effectiveness of modified six- corner suspension in patients with paravaginal defect and stress urinary incontinence. International Urogynecology Journal 2002;13(5):303-7.

167

Bruce RG, El Galley RE, Galloway NT. Paravaginal defect repair in the treatment of female stress urinary incontinence and cystocele. Urology 1999;54(4):647-51.

Cosson M, Collinet P, Occelli B, Narducci F, Crepin G. The vaginal patch plastron for vaginal cure of cystocele. Preliminary results for 47 patients. European Journal of Obstetrics, Gynecology, & Reproductive Biology 2001;95(1):73-80.

Glavind K, Madsen H. A prospective study of the discrete fascial defect rectocele repair. Acta Obstetricia et Gynecologica Scandinavica 2000;79(2):145-7.

Jeng CJ, Yang YC, Tzeng CR, Shen J, Wang LR. Sexual functioning after vaginal hysterectomy or transvaginal sacrospinous uterine suspension for uterine prolapse - A comparison. J Reprod Med 2005;50(9):669-74.

Karram M, Goldwasser S, Kleeman S, Steele A, Vassallo B, Walsh P. High uterosacral vaginal vault suspension with fascial reconstruction for vaginal repair of enterocele and vaginal vault prolapse. American Journal of Obstetrics & Gynecology 2001;185(6):1339-42.

Koyama M, Yoshida S, Koyama S, Ogita K, Kimura T, Shimoya K et al. Surgical reinforcement of support for the vagina in pelvic organ prolapse: concurrent iliococcygeus fascia colpopexy (Inmon technique). International Urogynecology Journal 2005;16(3):197- 202.

Lo TS. Tension-free vaginal tape procedures in women with stress urinary incontinence with and without co-existing genital prolapse. Current Opinion in Obstetrics & Gynecology 2004;16(5):399-404.

Long CY, Liu CM, Wu TP, Hsu SC, Chang Y, Tsai EM. A randomized comparison of vesicourethral function after laparoscopic hysterectomy with and without vaginal cuff suspension. Journal of Minimally Invasive Gynecology 2005;12(2):137-45.

Maher CF, Murray CJ, Carey MP, Dwyer PL, Ugoni AM. Iliococcygeus or sacrospinous fixation for vaginal vault prolapse. Obstet Gynecol 2001;98(1):40-4.

Shull BL, Capen CV, Riggs MW, Kuehl TJ. Bilateral attachment of the vaginal cuff to iliococcygeus fascia: an effective method of cuff suspension. American Journal of Obstetrics & Gynecology 1993;168(6 Pt 1):1669-74.

Singh K, Cortes E, Reid WM. Evaluation of the fascial technique for surgical repair of isolated posterior vaginal wall prolapse. Obstetrics & Gynecology 2003;101(2):320-4.

Thornton WN, Jr., Peters WA, III. Repair of vaginal prolapse after hysterectomy. American Journal of Obstetrics & Gynecology 1983;147(2):140-8.

Zacharin RF. Free Full-Thickness Vaginal Epithelium Graft in Correction of Recurrent Genital Prolapse. Australian & New Zealand Journal of Obstetrics & Gynaecology 1992;32(2):146-8.

Repair not for genital prolapse (N= 12)

Almeida SH, Gregorio EP, Rodrigues MA, Grando JP, Moreira HA, Fraga FC. Banked cadaveric fascia lata: 3-year follow-up. Transplantation Proceedings 2004;36(4):993-4.

Choe JM. The use of synthetic materials in pubovaginal sling. Advances in Experimental Medicine & Biology 2003;539(Pt A):481-92.

168

Clemons JL, Aguilar VC, Sokol ER, Sung VW, Myers DL. Suburethral sling treatment of occult stress incontinence and intrinsic sphincter deficiency in women with severe vaginal prolapse of the anterior vs posterior/apical compartment. American Journal of Obstetrics & Gynecology 2005;192(5):1566-72.

Crain J, Curole D, Hill G, Hurst B, Metzger D, Murphy A et al. Laparoscopic implant of Gore- Tex surgical membrane. Journal of the American Association of Gynecologic Laparoscopists 1995; 2(4):417-20.

Drake NL, Weidner AC, Webster GD, Amundsen CL. Patient characteristics and management of dermal allograft extrusions. International Urogynecology Journal 2005;16(5):375-7.

Ghosh T, Banfield PJ, Klazinga DA. An unusual complication of tension-free vaginal tape procedure: recurrent anterior vaginal wall abscess and sinus formation. Journal of Obstetrics & Gynaecology 2004;24(5):590-1.

Kuah SE, Lee KW, Houghton CR, Korda AR. The management of pulsion enterocoele with the Zacharin abdominoperineal technique (and mesh sacrocolpopexy). Australian & New Zealand Journal of Obstetrics & Gynaecology 2000;40(3):303-7.

Leanza V, Gasbarro N, Caschetto S. New technique for correcting both incontinence and cystocele: T.I.C.T. (Tension-free incontinence cystocele treatment). Urogynaecologia International Journal 2001;15(3):133-40.

Mattox TF, Stanford EJ, Varner E. Infected abdominal sacrocolpopexies: diagnosis and treatment. International Urogynecology Journal 2004;15(5):319-23.

Meschia M, Pifarotti P, Spennacchio M, Buonaguidi A, Gattei U, Somigliana E. A randomized comparison of tension-free vaginal tape and endopelvic fascia plication in women with genital prolapse and occult stress urinary incontinence. American Journal of Obstetrics & Gynecology 2004;190(3):609-13.

Pifarotti P, Spennacchio M, Gattei U, Ronchetti A, Stoppelli S, Meschia M. A randomized prospective comparison of TVT and endopelvic fascia plication in the treatment of occult stress urinary incontinence in patients with genital prolapse: Preliminary data. Urogynaecologia International Journal 2001;(1 Suppl.):55-7.

Smith HO, Genesen MC, Runowicz CD, Goldberg GL. The rectus abdominis myocutaneous flap: modifications, complications, and sexual function. Cancer 1998;83(3):510-20.

169

APPENDIX 9 Detailed quality assessment results for included studies: randomised controlled trials (full-text)

Study ID Q1 Q2 Q3 Q4 Q5 Q6 Q7 Q8 Q9 Q10 Q11 Q12 Q13 Q14 De Ridder, 200230 ? ? ? - + ? + ? ? ? NA ? ? ? Gandhi, 200532 + + + + + + + - ? + NA ? - ? Meschia, 200734 + + + + + + + - ? ? + - - ? Paraiso, 200636 + + + + + + + + ? + + ? - ? Sand, 200127 + - + + + + + ? ? ? NA + - ? Weber, 200135 + + + + + + + + ? - + ? - ?

Note:

170 1. See Appendix 2 for checklist for quality assessment of randomised controlled trials 2. + Yes; - No; ? Unclear; NA, not applicable, i.e. not reported any continuous data

APPENDIX 10 Detailed quality assessment results for non-randomised comparative studies

Study ID Q1 Q2 Q3 Q4 Q5 Q6 Q7 Q8 Q9 Q10 Q11 Q12 Q13 Q14 Q15 Q16 Q17 Q18 Altman, 200464 + + + ? - + + ? + + + ? - - ? + + - Castelo-Branco, 199865 + - + ? ? + + ? + + + ? + + + + + - Chaliha, 200660 + - + + ? + - ? + + + + + - ? + + - Handel, 200761 + + + ? - + + ? + + + ? + - ? + + + Julian, 199662 + + + ? + + + ? + + + ? + + NA + + - Leboeuf, 200363 + + + + ? ? + ? + + + ? + - ? + + - Vakili, 200566 + + + - - - - ? + + + ? - - NA + + +

Note: 171 1. See Appendix 3 for checklist for quality assessment of non-randomised comparative studies 2. + Yes; - No; ? Unclear; NA, not applicable, i.e. no drop-outs

APPENDIX 11 Detailed quality assessment results for registry and case series

Study ID Q1 Q2 Q3 Q4 Q5 Q7 Q8 Q9 Q10 Q11 Q13 Q14 Q15 Q17 Achtari, 200590 + + + + - + + + + + - + NA + Altman, 200729 + - + ? + + + + + + - - ? + Amrute, 200781 + - + - - + ? + + + + - ? + Collinet, 200691 + + + + - + ? + + + - - ? + Cosson, 200292 + + + ? - + ? + + + - - ? - Costantini, 200582 + + + + + + ? + + + + + ? + Cronje, 200667 + + + + - + ? + + + + - ? + De Tayrac, 200783 + + + + + + + + + + + + NA + De Tayrac, 200671 + + + + - + ? + + + + - ? + Deffeux, 200786 + + + + - + ? + + + - - ? + Dwyer, 200484 + + + ? - + ? + + + + - ? + Fatton, 200793 + + + + - + ? + + + - - ? + 72 172 Flood, 1998 + + + ? - + ? + + + + - ? + Frederick, 200573 + + + + + + ? + + + + + - + Gomelsky, 200474 + + + ? - + ? + + + + - ? + Kobashi, 200580 + + + - - + ? + + + + - ? + Kobashi, 200276 + + + ? + + ? + + + + - ? + Lim, 200589 + + + + + + ? + + + - - ? + Milani, 200528 + + + + + + ? + + + + - ? + Petros, 200687 + + + ? ? + ? + + + - - ? + Powell, 200477 + + + ? ? + ? + + + + - ? + Rodriguez, 200588 + + + ? + + ? + + + - - ? + Rozet, 200485 + + + ? - + ? + + + + - ? + Safir, 199978 + + + - - + ? + + + + - ? + Simsiman, 200679 + + + + - + + + + + + + + +

Note: 1. See Appendix 3 for checklist for quality assessment of non-randomised comparative studies; the same checklist was adapted for case series by taking out question 6, 12, 16 and 18; 2. + Yes; - No; ? Unclear; NA, not applicable, i.e. no drop-outs

APPENDIX 12 Summary of outcomes: anterior prolapse, efficacy

Table Anterior vaginal wall prolapse repair: subjective failure rate

Study ID Type of mesh Mesh, No mesh, Outcome reported as Follow-up n/N n/N (time assessed / mean, SD / median, range) Absorbable synthetic mesh Case-series Safir 199978 Polyglycolic acid 5/112 - Symptomatic 21m (6-42) Absorbable biological graft RCTs Gandhi 200532 Solvent-dried fascia lata 6/55 6/57 Vaginal bulging 13m (1.4-50) Meschia 200734 Porcine dermis 9/98 13/103 Prolapse sensation 1y Non-randomised comparative studies Leboeuf 200363 Porcine dermis 1/24 0/19 Symptomatic prolapse Mean 15m Case-series 73

173 Frederick 2005 Solvent-dehydrated fascia 18/251 - Symptomatic 20m (6-60) lata Powell 200477 Donor or autologous 2/58 - Pelvic pressure or vaginal bulging 24.7m (12-57) fascia lata requiring further treatment Non-absorbable synthetic mesh Case-series De Tayrac 200671 Polypropylene 1/55 - Symptomatic 37m (24-60) (Gynemesh)

Table Anterior vaginal wall prolapse repair: objective failure rate (mesh versus no mesh)

Study ID Type of mesh Mesh, No mesh, Outcome reported as Follow-up n/N n/N (time assessed / mean, SD / median, range) Absorbable synthetic mesh RCTs Sand 200127 Polyglactin 18/73 30/70 >=stage II 1y Weber 200135 Polyglactin 15/26 36/57 >= stage II 1y Case-series Safir 199978 Polyglycolic acid 11/112 - >=stage II 21m (6-42) Absorbable biological graft RCTs Gandhi 200532 Solvent-dried fascia lata 16/76 23/78 >=stage II 13m (1.4-50) Hviid 200543 (abs.) Porcine dermis 0/10 0/12 No definition 3m Kojancic 200748 (abs.) Porcine dermis 9/85 20/91 Unsatisfactory anatomic Mean 14m outcomes Meschia 200734 Porcine dermis 7/98 20/103 Ba>-1 1y Non-randomised comparative studies 174 Handel 200761 Porcine dermis 20/56 1/18 >=stage II Mean 17m Leboeuf 200363 Porcine dermis 3/24 0/19 >=stage II Mean: 15m Case-series Frederick 200573 Solvent dehydrated fascia lata 39/251 - Any degree 20m (6-60) Gomelsky 200474 Porcine dermis 9/70 - >=grade II 24m (12-NR) Kobashi 200276 Solvent-dehydrated fascia lata 17/132 - Any grade 3-12m Simsiman 200679 Porcine dermis 21/89 - Any grade 1y Non-absorbable synthetic mesh RCTs Al-Nazer 200737 (abs.) Soft polypropylene (Gynemesh PS) 1/20 6/20 No definition 1y Ali 200638 (abs.) Soft polypropylene (Gynemesh PS) 3/46 5/43 >=stage II 6m Hiltunen 200642 (abs.) Polypropylene (Parietene light) 7/92 32/85 >=stage II or any 1y symptomatic cystocele Ngnyen 200753 (abs.) Polypropylene (Perigee) 1/31 7/32 >=stage II, or unchanged or 6m worse than preoperative stage Non-randomised comparative studies Handel 200761 Polypropylene (trade name NR) 1/25 1/18 >=stage II Mean 13m Julian 199662 Polypropylene 0/12 4/12 >=stage II 2y Case-series De Tayrac 200671 Polypropylene (Gynemesh) 6/55 - >=stage II Mean 37m (24-60) Flood 199872 Polypropylene (Marlex) 0/142 - >=stage II Mean 3.2y (6w-12y) Milani 200528 Polypropylene (Prolene) 2/32 - Any degree 17m (3-48)

Table Anterior vaginal wall prolapse repair: objective failure rate (one mesh versus another)

Study ID Type of mesh 1 n/N Type of mesh 2 n/N Outcome Follow-up reported as (time assessed / mean, SD / median, range) RCTs Cervigni 200741 (abs) Non-absorbable synthetic 27/93 Absorbable biological graft (porcine dermis) 39/87 >=stage II 6-28m mesh (polypropylene (Gynemesh)) De Ridder 200230 Absorbable synthetic mesh 19/62 Absorbable biological graft (porcine dermis) 6/63 No definition 25.9m (6.4) (polyglactin) Non-randomised comparative studies Handel 200761 Non-absorbable synthetic 1/25 Absorbable biological graft (porcine dermis) 20/56 >=stage II Mean 17m mesh polypropylene, trade name NR)

Table Anterior vaginal wall prolapse repair: de novo prolapse 175 Study ID Type of mesh n/N Outcome reported as Follow-up (time assessed / mean, SD / median, range) Absorbable biological graft Case-series Frederick 200573 Solvent dehydrated fascia lata 8/58 >=grade II rectocele 17m (6-60) 7/165 >=grade II vault & enterocele Powell 200477 Donor fascia lata or autologous fascia lata 5/28 NR 24.7m (12-57) Non-absorbable synthetic mesh Case-series De Tayrac 200671 Polypropylene (Gynemesh) 8/45 Rectocele 37m (24-60)

Table Anterior vaginal wall prolapse repair: further operation needed for prolapse (mesh versus no mesh)

Study ID Type of mesh Mesh, n/N No mesh, n/N Recurrent / de novo Absorbable synthetic mesh Case-series Safir 199978 Polyglycolic acid 2/112 - Recurrent Absorbable biological graft Case-series Gomelsky 200474 Porcine dermis 0/70 - Recurrent Powell 200477 Donor fascia lata or autologous fascia lata 1/58 - Unclear Simsiman 200679 Porcine dermis 5/89 - Recurrent Non-absorbable synthetic mesh RCTs Hiltunen 200642 (abs.) Polypropylene (Parietene light) 0/92 2/85 Recurrent Case-series Flood 199872 Polypropylene (Marlex) 3/142 - De novo

Table Anterior vaginal wall prolapse repair: further operation needed for prolapse mesh (one mesh versus another) 176

Study ID Type of mesh 1 n/N Type of mesh 2 n/N Recurrent / de novo RCTs De Ridder 200431 (abs) Absorbable synthetic mesh 14/62 Absorbable biological graft (porcine 3/63 Recurrent (polyglactin) dermis)

Table Anterior vaginal wall prolapse repair: persistent urinary incontinence

Study ID Type of mesh Mesh, n/N No nesh, n/N Measures Follow-up Absorbable synthetic mesh Case series Safir 199978 polyglactin 6/45 - Urge incontinence 21m (6-42) 5/49 - SUI Absorbable biological graft Non-randomised comparative studies Leboeuf 200463 Porcine dermis 13/14 9/10 SUI Mean 15m 10/14 3/12 Urge incontinence Case series Frederick 200573 Solvent dehydrated cadaveric fascia lata 137/251 - All types of incontinence 20m (6-60) Kobashi 200276 Solvent dehydrated cadaveric fascia lata 24/108 - Pure SUI or mixed incontinence 12.4m (6-28) Powell 200477 Donor or autologous fascia lata 11/41 - Incontinence 24.7m (12-57) Simsiman 200679 Porcine dermis 24/58 - Urge incontinence 1y Non-absorbable synthetic mesh Case series Flood 199872 Polypropylene (Marlex) 8/30 - SUI 3.2y (6w-12y) Milani 200528 Polypropylene (Prolene) 5/5 - Urge incontinence 17m (3-48) 177 9/14 - SUI

No studies reported outcomes on persistent bowel symptoms in women having this symptom at baseline.

No studies reported outcomes on persistent dyspareunia in women having this symptom at baseline.

Table Anterior vaginal wall prolapse repair: improvement of disease-specific quality of life (RCTs and non-randomised comparative studies only)

Study ID Type of mesh Mesh, n, No mesh, n, Outcome reported as Follow-up mean (SD) mean (SD) Absorbable biological graft Non-randomised comparative studies Chaliha 200660 Small intestine submucosa 14, mean 14 14, mean 14 Prolapse impact from a validated 2y prolapse OoL questionnaire (PQOL, no further details)

APPENDIX 13 Summary of outcomes: anterior prolapse, safety

Table Anterior vaginal wall prolapse repair: blood loss, ml

Study ID Type of mesh Mesh, n, mean (SD) / median (range) No mesh, n, mean (SD) / median (range) Absorbable biological graft RCTs Hviid 200543 (abs.) Porcine dermis 10, 75 (25-150) 12, 57 (25-150) Kocjancic 200748 (abs.) Porcine dermis 100, 151 (112) 106, 167 (96) Meschia 200734 Porcine dermis 98, 151 (112) 103, 167 (96) Case-series Powell 200477 Donor or autologous fascia lata 58, 329 (50-1000) - Non-absorbable synthetic mesh RCTs Ali 200638 (abs.) Polypropylene (Gynemesh PS) 54, 64.5 (70.4) 54, 50.3 (89.0) Registry Altman 200729 Polypropylene (Prolift) 106, 190 (295) - Case-series Flood 199872 Polypropylene (Marlex) 142, mean 442 - 178

Table Anterior vaginal wall prolapse repair: number of women with blood loss requiring blood transfusion

Study ID Type of mesh Mesh, n/N No mesh, n/N Outcome reported as Absorbable synthetic mesh RCTs Weber 200135 Polyglactin 0/35 1/74 Blood transfusion Case-series Safir 199978 Polyglycolic acid 0/112 - Blood transfusion Absorbable biological graft Non-randomised comparative studies Chaliha 200660 Small intestine submucosa 0/14 0/14 Excessive bleeding Handel 200761 (NR for the other 2 groups) Porcine dermis 1/56 - Blood transfusion Case-series Gomelsky 200474 Porcine dermis 0/70 - Powell 200477 Donor or autologous fascia lata 2/58 - Non-absorbable synthetic mesh Registry Altman 200729 Polypropylene (Prolift) 4/106 - >500ml Case-series De Tayrac 200671 Polypropylene (Gynemesh) 0/55 - Blood transfusion

Table Anterior vaginal wall prolapse repair: damage to surrounding organs during operation

Study ID Type of mesh Mesh, n/N No mesh, n/N Damaged organs Absorbable synthetic mesh Case-series Safir 199978 Polyglacolic acid 0/112 - Bladder perforation, vesicovaginal fistula Absorbable biological graft Non-randomised comparative studies Leboeuf 200463 Porcine dermis 0/24 0/19 Bladder perforation Case-series Gomelsky 200474 Porcine dermis 0/70 - Bladder urethra, ureters or small bowl Non-absorbable synthetic mesh Registry Altman 200729 Polypropylene (Prolift) 4/106 - 2 bladder perforation, 2 urethra perforation, 0 rectal perforation Case-series De Tayrac 200671 Polypropylene (Gynemesh) 1/55 - Bladder injury Petros 200687 NR (Tissue Fixation System, multifilament) 1/90 - Bladder perforation 179

Table Anterior vaginal wall prolapse repair: mesh/graft erosion

Study ID Type of mesh n/N Outcome reported as Follow-up Absorbable synthetic mesh RCTs Weber 200135 Polyglactin 1/35 Mesh erosion 23.3m (4.5-44.4) Case-series Safir 199978 Polyglacolic acid 0/112 Mesh erosion 21m (6-42) Absorbable biological graft RCTs Kocjancic 200748 (abs.) Porcine dermis 1/85 Mesh erosion Mean 14m Meschia 200734 Porcine dermis 1/98 Vaginal extrusion with bacterial colonization 1m Non-randomised comparative studies Chaliha 200660 Small intestine submucosa 0/14 Mesh erosion 2y Leboeuf 200463 Porcine dermis 0/24 Mesh erosion 15m Case-series Gomelsky 200474 Porcine dermis 0/70 Mesh erosion 14m (12-NR) Powell 200477 Donor or autologous fascia lata 6/58 Mesh erosion 24.7m (12-57) Simsiman 200679 Porcine dermis 15/89 Mesh erosion 24m (6-44) Non-absorbable synthetic mesh 180 RCTs Ali 200638 (abs.) Polypropylene (Gynemesh PS) 3/46 Mesh erosion 6m Hiltunen 200642 (abs.) Polypropylene (Parietene light) 17/92 Minor mesh exposure 1y Nguyen 200753 (abs.) Polypropylene (Perigee) 1/31 Mesh exposure 6m Non-randomised comparative studies Julian 199662 Polypropylene (Marlex)) 1/12 Two tiny perforating mesh fibers 2y Case-series De Tayrac 200671 Polypropylene (Gynemesh) 5/55 Vaginal erosions 37m (24-60) Deffieux 200786 Polypropylene (Gynemesh) 15/89 Vaginal erosions 32m (8-60) Deffieux 200786 Polypropylene (Gynemesh Soft) 12/49 Vaginal erosions 7m (1-22) Flood 199872 Polypropylene (Marlex) 3/142 Mesh erosion 3.2y (6w-12y) Milani 200528 Polypropylene (Prolene) 4/32 Mesh erosion 12m (5)

Table Anterior vaginal wall prolapse repair: mesh/graft erosion (one mesh versus another)

Study ID Type of mesh 1 n/N Type of mesh 2 n/N Outcome reported as Follow-up RCTs Cervigni 200741 (abs) Non-absorbable synthetic mesh 6/93 Absorbable biological 0/87 Mesh erosion 6-28m (polypropylene (Gynemesh)) graft (porcine dermis) Non-randomised comparative studies Handel 200761 Non-absorbable synthetic 1/25 Absorbable biological 12/56 Mesh extrusion 17m (2-46) (polypropylene (trade name NR)) graft (porcine dermis)

Table Anterior vaginal wall prolapse repair: further operation need for mesh/graft erosion

Study ID Type of mesh n/N Surgery Absorbable synthetic mesh RCTs Weber 200135 Polyglactin 1/35 Mesh removal Absorbable biological graft RCTs 181 Meschia 200734 Porcine dermis 1/98 Mesh removal Non-randomised comparative studies Handel 200761 (not reported for the other group) Porcine dermis 3/56 Partial removal Non-absorbable synthetic mesh Non-randomised comparative studies Julian 199662 Polypropylene 1/12 Mesh removal Case-series De Tayrac 200671 Polypropylene (Gynemesh) 4/55 3 partial excision, 1 complete excision Deffieux 200786 Polypropylene (Gynemesh or Soft) 15/138 Mesh removal Flood 199872 Polypropylene (Marlex) 3/142 Partial excision

Table Anterior vaginal wall prolapse repair: de novo urinary incontinence

Study ID Type of mesh n/N Outcomes Follow-up Absorbable synthetic mesh Case-series Safir 199978 Polyglycolic acid 5/67 Urge incontinence 21m (6-42) 0/63 Stress urinary incontinence Absorbable biological graft Case-series Frederick 200573 Solvent-dehydrated fascia lata 13/153 Urge incontinence 17m (6-60)

Study ID Type of mesh n/N Outcomes Follow-up Powell 200477 Donor or autologous fascia lata 1/17 Stress urinary incontinence 24.7m (12-57) Simsiman 200679 Porcine dermis 8/28 Urge incontinence 1y 2/25 Stress urinary incontinence Non-absorbable synthetic mesh Case-series Flood 199872 Polypropylene (Marlex) 1/30 Incontinence 3.2y (16w-12y) Rodriguez 200588 Soft polypropylene (60% more porosity, trade name NR) 3/44 Stress urinary incontinence 3m

No studies reported outcome on de novo bowel symptoms.

Table Anterior vaginal wall prolapse repair: de novo dyspareunia

Study ID Type of mesh n/N Follow-up Non-absorbable synthetic mesh Case-series Malani 200528 Polypropylene (Prolene) 4/11 17m (3-48) 182

Table Anterior vaginal wall prolapse repair: infection

Study ID Type of mesh Mesh, No mesh, Outcome reported as Follow-up n/N n/N Absorbable synthetic mesh Case-series Safir 199978 Polyglycolic acid 0/112 - Retropubic abscess 21m (6-42) Absorbable biological graft Non-randomised comparative studies Chaliha 200660 Small intestine submucosa 0/14 0/14 Infection 2y Handel 200761 (NR for the other 2 groups) Porcine dermis 1/56 - Mesh infection Mean 17m Leboeuf 200463 Porcine dermis 0/24 0/19 Infection Mean 15m Case-series Frederick 200573 Solvent dehydrated fascia lata 0/251 - Serious infection 17m (6-61) Kobashi 200276 Solvent dehydrated fascia lata 4/132 - Urinary tract infection 12.4m (6-28) Non-absorbable synthetic mesh RCT Hiltunen 200642 (abs.) Polypropylene (Parietene light) 1/105 4/97 Postoperative infection 1y Non-randomised comparative studies Julian 199662 Polypropylene (Marlex) 0/12 0/12 Inhospital infection 2y

Study ID Type of mesh Mesh, No mesh, Outcome reported as Follow-up n/N n/N Registry Altman 200729 Polypropylene (Prolift) 6/106 - Urinary tract infection Within hospital stay 0/106 - Wound infection Case-series De Tayrac 200671 Polypropylene (Gynemesh) 4/55 - Urinary tract infection 37m (24-60) 0/55 - Mesh infection Deffieux 200786 Polypropylene (Gynemesh or Soft) 0/138 - Mesh infection 32.1m (7.5-60), 0/138 - Infection Soft 7.1m (1-22) Flood 199872 Polypropylene (Marlex) 0/142 - Reaction or persistent 3.2y (6w-12y) infection due to mesh

Table Anterior vaginal wall prolapse repair: other serious adverse effects

Study ID Type of mesh Mesh, No mesh, Outcome reported as Time that occurred n/N n/N Absorbable synthetic mesh RCTs 183 Weber 200135 Polyglactin 0/35 1/74 Pulmonary embolism - Absorbable biological graft Non-randomised comparative studies Handel 200761 (NR for the other 2 groups) Porcine dermis 1/56 - Return to theatre Leboeuf 200463 Porcine dermis 0/24 0/19 Fistula - Case-series Kobashi 200276 Solvent-dehydrated fascia lata 1/132 - Osteitis pubis - 0/132 - Osteomyelitis - Non-absorbable synthetic mesh Registry Altman 200729 Polypropylene (Prolift) 1/106 - Anaemia During hospital stay Case-series Flood 199872 Polypropylene (Marlex) 1/142 - Postoperative stroke With history of heart disease and was lost to follow-up after the 6w check 2/142 - Postoperative myocardial - ischaemia

APPENDIX 14 Summary of outcomes: anterior prolapse, operation time

Table Anterior vaginal wall prolapse repair: operation time

Study ID Type of mesh Mesh, n, mean (SD) / median (range) No mesh, n, mean (SD) / median (range) Absorbable biological graft RCTs Hviid 200543(abs.) Porcine dermis 10, 38min (23-55) 12, 28min (17-50) Non-absorbable synthetic mesh Registry Altman 200729 Polypropylene (Prolift) 106, 72.5 (32.1) - Case-series De Tayrac 200671 Polypropylene (Gynemesh) 55, 102 min (50-150) -

184

APPENDIX 15 Summary of outcomes: posterior prolapse, efficacy

Table Posterior vaginal wall prolapse repair: subjective failure rate

Study ID Type of mesh Mesh, No mesh, Outcome reported as Follow-up (time assessed / n/N n/N mean, SD / median, range) Absorbable biological graft RCTs Paraiso 200636 Porcine dermis 6/28 9/60 Prolapse symptoms 1y Case-series Kobashi 200580 Solvent-dried fascia lata 3/50 - Symptomatic 13.7m (6-23)

Table Posterior vaginal wall prolapse repair: objective failure rate Study ID Type of mesh Mesh, No mesh, Outcome reported as Follow-up (time assessed / n/N n/N mean, SD / median, range) Absorbable synthetic mesh RCTs Sand 200127 Polyglactin 6/65 7/67 >=stage II 1y 185 Non-randomised comparative studies Castelo-Branco 199865 Polyglacolic acid 0/5 0/5 NR 1y Absorbable biological graft RCTs Paraiso 200636 Porcine dermis 12/26 10/55 PoP-Q >-2 1y Non-randomised comparative studies Altman 200464 Porcine dermis 2/17 1/15 >=stage II 6m Case-series Kobashi 200580 Sovent-drived fascia lata 5/50 - Any degree 13.7 (6-23) Non-absorbable synthetic mesh Case-series Milani 200528 Polypropylene (Prolene) 2/31 - Any degree 14m (6)

No studies reported outcomes on de novo prolapse in women having no prolapse at other sites at baseline.

Table Posterior vaginal wall prolapse repair: further operation needed for prolapse

Study ID Type of mesh Mesh, n/N No mesh, n/N Recurrent / de novo Absorbable biological graft RCTs Paraiso 200636 Porcine dermis 3/29 3/70 Any compartment

No studies reported outcome on persistent urinary incontinence.

Table Posterior vaginal wall prolapse repair: persistent bowel symptoms

Study ID Type of mesh Mesh, n/N No mesh, n/N Outcome reported as Follow-up Absorbable biological graft RCTs Paraiso 200636 Porcine dermis 5/29 18/58 Defecatory dysfunction 1y Case series Kobashi 200580 Solvent-dried fascia lata 9/53 - Stool trapping 13.7m (6-23) Combined mesh/graft 186 Case series Lim 200589 Polyglactin-polypropylene 2/23 - constipation 6m 5/43 - Defecation difficulties

Table Posterior vaginal wall prolapse repair: persistent dyspareunia

Study ID Type of mesh Mesh, n/N Outcome reported as Follow-up Absorbable biological graft Case series Kobashi 200580 Solvent-dried fascia lata 5/14 Dyspareunia 13.7m (6-23)

Table Posterior vaginal wall prolapse repair: improvement of disease-specific quality of life (RCTs and non-randomised comparative studies only)

Study ID Type of mesh Mesh, n, No mesh, n, Outcome reported as Follow-up mean (SD) mean (SD) Absorbable biological graft RCTs Paraiso 200636 Porcine dermis 13, 5 (13) 32, 16 (32) PFIQ-7 (the Pelvic Floor Impact Questionnaire-7, 0- 2y 300 greatest adverse impact on OoL)

APPENDIX 16 Summary of outcomes: posterior prolapse, safety

Table Posterior vaginal wall prolapse repair: blood loss, ml

Study ID Type of mesh Mesh, n, mean (SD) / median No mesh, n, mean (SD) / (range) median (range) Absorbable biological graft RCTs Paraiso 200636 Porcine dermis 31, 200 (50-3500) 74, 150 (50-950) Non-absorbable synthetic mesh Registry Altman 200729 Polypropylene (Prolift) 71, 85 (97) -

Table Posterior vaginal wall prolapse repair: blood loss requiring blood transfusion

Study ID Type of mesh Mesh, n/N No mesh, n/N Outcome reeported as

187 Absorbable synthetic mesh Non-randomised comparative studies Castelo-Branco 199865 Polyglacolic acid 0/5 0/5 Hemorrhagic complications Absorbable biological graft RCTs Paraiso 200636 Porcine dermis 1/31 3/74 Blood transfusion Combined mesh/graft Case-series Lim 200589 Polypropylene and 0/90 - Blood transfusion polyglactin Non-absorbable synthetic mesh Registry Altman 200729 Polypropylene (Prolift) 1/71 - >1000ml

Table Posterior vaginal wall prolapse repair: damage to surrounding organs during operation

Study ID Type of mesh Mesh, n/N No mesh, n/N Damaged organs Absorbable synthetic mesh Non-randomised comparative studies Castelo-Branco 199865 Polyglacolic acid 0/5 0/5 Rectum Absorbable biological graft RCTs Paraiso 200636 Porcine dermis 1/31 2/74 Bladder, bowel, or uretheral Combined mesh/graft Case-series Lim 200589 Polypropylene and polyglactin 0/90 - Bowel injury Non-absorbable synthetic mesh Registry Altman 200729 Polypropylene (Prolift) 3/71 - 0 bladder or urethra perforation, 3 rectal perforation

Table Posterior vaginal wall prolapse repair: mesh/graft erosion 188

Study ID Type of mesh n/N Outcome reported as Follow-up Absorbable biological graft RCTs Paraiso 200636 Porcine demis 0/28 Mesh exposure or graft complications 17.5m (4.4-33.7) Combined mesh/graft RCTs Lim 200655 (abs.) Polypropylene and polyglactin 5/25 Mesh erosion 13.6 (9.0) Case-series Lim 200589 Polypropylene and polyglactin 11/90 Minor mesh protrusion 6m Non-absorbable synthetic mesh Case-series Milani 200528 Polypropylene (Prolene) 2/31 Mesh erosion 14m (6)

Table Posterior vaginal wall prolapse repair: further operation need for mesh/graft erosion

Study ID Type of mesh n/N Surgery Combined mesh/graft Case-series Lim 200589 Polypropylene and polyglactin 11/90 Surgery (no details)

No studies reported outcome on de novo urinary incontinence.

Table Posterior vaginal wall prolapse repair: de novo bowel symtpoms

Study ID Type of mesh n/N Outcome reported as Follow-up Combined mesh/graft Case-series Lim 200589 Polypropylene- glycolic acid 2/45 Constipation 6-12w 1/35 Defecation difficultis Non-absorbable synthetic mesh 189 Case-series Milani 200528 Polypropylene (Prolene) 1/29 Fetal Incontinence 17m (3-48)

Table Posterior vaginal wall prolapse repair: de novo dyspareunia

Study ID Type of mesh n/N Follow-up Absorbable biological graft Case-series Kobashi 200580 Solvent-dried fascia lata 4/25 13.7 (6-23) Combined mesh/graft Case-series Lim 200589 Polypropylene and polyglactin 2/36 6m

Table Posterior vaginal wall prolapse repair: infection

Study ID Type of mesh Mesh, No mesh, Outcome reported as Follow-up n/N n/N Absorbable synthetic mesh Non-randomised comparative studies Castelo-Branco 199865 Polyglacolic acid 0/5 0/5 Cuff or posterior vaginal wall infections 1y 0/5 0/5 Thrombophlebitis Absorbable biological graft RCTs Paraiso 200636 Porcine dermis 3/31 4/74 Infection requiring antibiotics 17.5m (4.4-33.7) 6/31 11/74 Pelvic abscess 1/31 0/74 Urinary tract infection Non-randomised comparative studies Altman 200464 Porcine dermis 1/17 2/15 Urinary tract infection 6m Non-absorbable synthetic mesh Registry Altman 200729 Polypropylene (Prolift) 3/106 Urinary tract infection During hospital stay 1/106 Wound infection

190

Table Posterior vaginal wall prolapse repair: other serious adverse effects

Study ID Type of mesh Outcome reported as Mesh, n/N No mesh, n/N Time that occurred Absorbable biological graft RCTs Paraiso 200636 Porcine dermis Return to threatre 1/31 0/74 - Pulmonary 1/31 1/74 Myocardial infarction 0/31 1/74 Readmissions 0/31 3/74 Emergency room evaluations 4/31 11/74 Non-absorbable synthetic mesh Registry Altman 200729 Polypropylene (Prolift) Anaemia 0/71 During hospital stay

APPENDIX 17 Summary of outcomes: posterior prolapse, operation time

Table Posterior vaginal wall prolapse repair: operation time (min)

Study ID Type of mesh Mesh, n, Mean (SD) /median (range) No mesh, n, Mean (SD) /median (range) Absorbable biological graft RCTs Paraiso 200636 Porcine dermis 31, 169 (62) 74, 150 (68) Non-absorbable synthetic mesh Registry Altman 200729 Polypropylene (Prolift) 71, 57.6min (19.6)

191

APPENDIX 18 Summary of outcomes: anterior and/or posterior prolapse, efficacy

Table Anterior and/or posterior vaginal wall prolapse repair: subjective failure rate

Study ID Type of mesh Mesh, n/N No mesh, n/N Outcome reported as Follow-up (time assessed / mean, SD / median, range) Absorbable synthetic mesh RCTs Alladhin 200656 (abs.) Polyglactin 14/32 14/34 No improvement in quality of life 6m (scored 0 using a scale 0-10) Non-absorbable synthetic mesh Case-series Amrute 200781 Polypropylene (BioArc device) 0/76 - Symptomatic 30.7m (1.7) Costantini 200582 Polypropylene (Marlex) 0/72 - Symptomatic 51m (12-115)

Table Anterior and/or posterior vaginal wall prolapse repair: objective failure rate

192 Study ID Type of mesh Mesh, n/N No mesh, n/N Outcome Follow-up (time assessed / reported as mean, SD / median, range) Absorbable synthetic mesh RCTs Allahdin 200656 (abs.) Polyglactin 2/26 4/32 >=stage 2 3m Combined mesh/graft Case-series De Tayrac 200783 Polypropylene coated with atelocollagen 11/143 - >=stage 2 13m (10-19) Non-absorbable synthetic mesh RCTs Lim 200758 (abs.) Polypropylene (Gynemesh PS) 11/62 20/60 >=stage 2 1y Meschia 200759 (abs.) Polypropylene (Perigee-Apogee system) 7/21 3/17 No definition 3m Case-series Amrute 200781 Polypropylene (BioArc device) 4/76 - No definition 30.7m (1.7) Costantini 200582 Polypropylene (Marlex) 2/64 - >=grade 2 51m (12-115) Dwyer 200484 Polypropylene (Atrium) 4/97 - >=stage 2 29m (6-52) Rozet 200485 Polyester covered with silicone 13/325 - No definition 14.6m (6m-5y)

No studies reported outcomes on de novo prolapse.

Table Anterior and/or posterior vaginal wall prolapse repair: further operation needed for prolapse

Study ID Type of mesh Mesh, n/N No mesh, Recurrent / de n/N novo Non-absorbable synthetic mesh Case-series Costantini 200582 Polypropylene (Marlex) 0/64 - Recurrent Dwyer 200484 Polypropylene (Atrium) 7/97 - De novo >=1 type of mesh/graft Non-randomised comparative studies Vakilli 200566 Absorbable biological graft or non- 8/98 (only those performed by one of 20/214 Unclear absorbable synthetic mesh the primary surgeons)

Table Anterior and/or posterior vaginal wall prolapse repair: persistent urinary incontinence

Study ID Type of mesh Mesh, n/N Outcome reported as Follow-up Non-absorbable synthetic mesh Case-series Cosson 200292 Polypropylene (Mersilene) 24/67 Incontinence Mean 343d 193 Costantini 200582 Polypropylene (Marlex) 11/42 Urinary incontinence 51m (12-115) Dwyer 200484 Polypropylene (Atrium) 4/40 Urge incontinence 29m (6-52) Fatton 200793 Polypropylene (Prolift) 7/54 Incontinence 25w (12-42)

Table Anterior and/or posterior vaginal wall prolapse repair: persistent bowel symptoms

Study ID Type of mesh Mesh, n/N Outcome reported as Follow-up Non-absorbable synthetic mesh Case-series Dwyer 200484 Polypropylene (Atrium) 1/21 Constipation 29m (6-52)

Table Anterior and/or posterior vaginal wall prolapse repair: persistent dyspareunia

Study ID Type of mesh Mesh, n/N Outcome reported as Follow-up Combined mesh/graft Case-series De Tayrac 200783 Polypropylene coated with atelocollagen 1/10 Dyspareunia 13m (10-19)

Table Anterior and/or posterior vaginal wall prolapse repair: improvement of disease-specific quality of life (RCTs and non-randomised comparative studies only)

Study ID Type of mesh Mesh, n, No mesh, n, Outcome reported as Follow-up mean (SD) mean (SD) Absorbable synthetic mesh RCTs Allahdin 200656 (abs.) Polyglactin 27, 1.4 (2.5) 33, 1.5 (2.8) 0 symptoms do not interfere with everyday life at all - 10 6m

194

APPENDIX 19 Summary of outcomes: anterior and/or posterior prolapse, safety

Table Anterior and/or posterior vaginal wall prolapse repair: blood loss, ml

Study ID Type of mesh N, mean (SD) / median (range) Non-absorbable synthetic mesh Registry Altman 200729 (total mesh) Polypropylene (Prolift) 51, 169 (133) Altman 200729 (ant. and post.) Polypropylene (Prolift) 20, 153 (100) Case-series Costantini 200582 Polypropylene (Marlex) 72, mean 263 Fatton 200793 Polypropylene (Prolift) 110, all <300ml Rozet 200485 Polyester covered with silicone 325, 10-100

Table Anterior and/or posterior mesh repair: number of women with blood loss requiring blood transfusion

Study ID Type of mesh Mesh, n/N No mesh, n/N Outcome reported as 195 Non-absorbable synthetic mesh RCTs Meschia 200759 (abs.) Polypropylene (Perigee-Apogee system) 0/36 1/35 >500ml Registry Altman 200729 (total mesh) Polypropylene (Prolift) 3/51 - >500ml Altman 200729 (ant. and post.) Polypropylene (Prolift) 0/20 - >500ml Case-series Cosson 200292 Polyethylene tetraphalate (Mersilene) 1/83 - return to theatre due to haemorrhage Costantini 200582 Polypropylene (type # (Marlex) 4/72 - Blood transfusion Dwyer 200484 Polypropylene (Atrium) 0/110 - >500ml Fatton 200793 Polypropylene (Prolift) 0/110 - Rozet 200485 Polyester covered with silicone 3/328 - Blood transfusion

Table Anterior and/or posterior mesh repair: damage to surrounding organs during operation

Study ID Type of mesh n/N Damaged organs Combined mesh/graft Case-series De Tayrac 200783 Polypropylene coated with atelocollagen 4/143 3 bladder injury, 1 rectal injury Non-absorbable synthetic mesh Registry Altman 200729 (total mesh) Polypropylene (Prolift) 2/51 2 bladder perforation, 0 urethra or rectal perforation Altman 200729 (ant. and post.) Polypropylene (Prolift) 2/20 1 bladder perforation, 0 urethra perforation, 1 rectal perforation Case-series Collinet 200691 Polypropylene (Prolene soft 169, Prolene 5/277 1 rectal injury, 4 bladder injury 108, Gynecare) Cosson 200292 Polyethylene tetraphalate (Mersilene) 2/83 1 rectal injury, 1 bladder injury Fatton 200793 Polypropylene (Prolift) 1/110 1 bladder injury, 0 rectal injury

Table Anterior and/or posterior vaginal wall prolapse repair: mesh/graft erosion

Study ID Type of mesh n/N Outcome reported as Follow-up 196 Combined mesh/graft Case-series De Tayrac 200783 Polypropylene coated with atelocollagen 9/143 Mesh erosion 13m (10-19) Non-absorbable synthetic mesh RCTs Lim 200758 (abs.) Polypropylene (Gynemesh PS) 4/62 Mesh exposure 1y Meschia 200759 (abs.) Polypropylene (Perigee-Apogee system) 1/21 Mesh exposure 3m Case-series Amrute 200781 Polypropylene (BioArc device) 2/76 Erosion 30.7m (1.7) Collinet 2006 Polypropylene (Prolene Soft, 169, Prolene 108, Gynecare) 34/277 Mesh erosion 2m Cosson 200292 Polyethylene tetraphalate (Mersilene) 1/83 Mesh rejection Mean 343d Costantini 200582 Polypropylene ((Marlex) 3/72 Erosion 41m (12-115) Dwyer 200484 Polypropylene (Atrium) 9/97 Mesh erosion 6-52m Fatton 200793 Polypropylene (Prolift) 5/106 Mesh exposure 3m Rozet 200485 Polyester covered with silicone (Polyatex) 3/325 Erosion 14.6 (6m-5y) >=1 type of mesh Non-randomised comparative studies Vakili 200566 Absorbable biological or non-absorbable synthetic 25/98 Erosion 9m (3-67) Case series Achtari 200590 Combined or non-absorbable synthetic 13/198 Any visible mesh exposure 6w – 6m

Table Anterior and/or posterior vaginal wall prolapse repair: further operation need for mesh/graft erosion

Study ID Type of mesh n/N Operation Combined mesh/graft Case-series De Tayrac 200783 Polypropylene coated with atelocollagen 6/143 Mesh removal Non-absorbable synthetic mesh RCTs Lim 200758 (abs.) Polypropylene (Gynemesh PS) 3/62 Mesh removal Case-series Amrute 200781 Polypropylene (BioArc device) 2/76 Partial excision Collinet 200691 Polypropylene (Prolene Soft 169, Prolene 108, Gynecare) 25/277 Partial excision Cosson 200292 Polyethylene tetraphalate (Mersilene) 1/83 Mesh removal Costantini 200582 Polypropylene (Marlex) 3/72 vaginal revision for mesh erosion, no further details Dwyer 200484 Polypropylene (Atrium) 6/97 Partial excision Fatton 200793 Polypropylene (Prolift) 2/106 Mesh resection Rozet 200485 Polyester covered with silicone 3/325 Partial removals

197 Table Anterior and/or posterior vaginal wall prolapse repair: de novo urinary incontinence

Study ID Type of mesh n/N Outcome reported as Follow-up Non-absorbable synthetic mesh Case-series Armrute 200781 Polypropylene (BioArc device) 6/40 Urge incontinence 30.7m (1.7) Costantini 200582 Polypropylene (Marlex) 1/40 Incontinence 51m (12-115) Dwyer 200484 Polypropylene (Atrium) 3/57 Urge incontinence 29m (6-52) Fatton 200793 Polypropylene (Prolift) 5/56 Incontinence 25w (12-42) Rozet 200485 Polyester covered with silicone 19/162 Incontinence 14.6m (6m-5y)

Table Anterior and/or posterior vaginal wall prolapse repair: de novo bowel symptoms

Study ID Type of mesh n/N Outcome reported as Follow-up Non-absorbable synthetic mesh Case-series Dwyer 200484 Polypropylene (Atrium) 1/47 Constipation 29m (6-52)

Table Anterior and/or posterior vaginal wall prolapse repair: de novo dyspareunia

Study ID Type of mesh n/N Follow-up Combined mesh/graft Case-series De Tayrac 200783 Polypropylene coated with atelocollagen 10/78 13m (10-19) Non-absorbable synthetic mesh Case-series Dwyer 200484 Polypropylene (Atrium) 3/42 29m (6-52)

Table Anterior and/or posterior vaginal wall prolapse repair: infection

Study ID Type of mesh Mesh, No mesh, Outcome reported as Follow-up n/N n/N Non-absorbable synthetic mesh Registry Altman 200729 (total mesh) Polypropylene (Prolift) 7/51 Urinary tract infection During hospital stay Altman 200729 (ant. and post.) Polypropylene (Prolift) 0/20 Urinary tract infection During hospital stay 0/20 Wound infection 198 Case-series Cosson 200292 Polypropylene (Mersilene) 1/83 Chronic inflammation of the cervix Mean 343d 9/83 Low urinary tract infection Costantini 200582 Polypropylene (Marlex) 2/72 Wound infection 51m (12-115) Fatton 200793 Polypropylene (Prolift) 13/110 Urinary tract infection 252 (12-42) 0/110 Pelvic abscess Rozet 200485 Polyester covered with silicone 1/325 Postoperative spondylitis 14.6m (6m-5y) >=1 type of mesh Non-randomised comparative studies Vakili 200566 Biological graft or polypropylene 18/98 10/214 Vaginal/graft infection 9m (3-67)

Table Anterior and/or posterior vaginal wall prolapse repair: other serious adverse events

Study ID Type of mesh n/N Events Time that occurred Non-absorbable synthetic mesh Registry Altman 200729 (total mesh) Polypropylene (Prolift) 0/51 Anaemia During hospital stay Altman 200729 (ant. and post.) Polypropylene (Prolift) 0/20 Anaemia During hospital stay Case-series Dwyer 200484 Polypropylene (Atrium) 1/97 Rectovaginal fistula Occurred at 5th d after operation Fatton 200793 Polypropylene (Prolift) 2/110 Haemotomas requiring secondary surgical treatment - 0/110 Bladder or rectal fistula

APPENDIX 20 Summary of outcomes: anterior and/or posterior prolapse, operation time

Table Anterior and/or posterior mesh repair: operation time (min)

Study ID Type of mesh N, mean (SD)/median (range) Non-absorbable synthetic mesh Registry Altman 200729 (total) Polypropylene (Prolift) 51, 85.9 (22.9) Altman 200729 (ant. and post.) Polypropylene (Prolift) 20, 86.9 (30.2) Case-series Cosson 200292 Polypropylene (Mersilene) 83, 276 (120-360) Costantini 200582 Polypropylene (Marlex) 72, 102 (60-135) Fatton 200793 Polypropylene (Prolift) 110, 89.9 (30-180) Rozet 200485 Polyester covered with silicone 325, 97 (45-156)

199