Vaginal mesh was initially used to improve anatomic outcomes and decrease reoperation for recurrent prolapse. Preliminary studies found significant short-term benefits with few complications, which led to its widespread adoption.1–4 In fact, several randomized trials have shown improved outcomes for anterior prolapse.5–9 However, initial enthusiasm for mesh has been tempered by the U.S. Food and Drug Administration advisory concerning complications, which include mesh erosion, mesh contraction, de novo pain and dyspareunia, bleeding, and infection.10
In addition, recent studies have drawn into question the definition of successful surgery for pelvic organ prolapse. Barber et al11 found that the symptom of vaginal bulge postoperatively was more strongly correlated with the patient's assessment of overall improvement when compared with anatomic success alone. Although vaginal mesh may improve anatomic outcomes, those outcomes may not correlate with improved patient satisfaction when compared with traditional prolapse surgery without mesh.
Consequently, potential anatomic benefits must be evaluated in the context of patient symptoms and must be weighed against the significant risk of complications. Unfortunately, most studies are limited to 1-year follow-up. The purpose of our study is to report 3-year outcomes of a double-blind, multicenter, randomized controlled trial (RCT) comparing traditional vaginal prolapse repair without mesh and vaginal colpopexy repair with mesh. Our primary study outcome at 1 year was the anatomic cure of prolapse.12,13 For this analysis, we evaluate success based on composite outcomes including symptomatic 3-year quality-of-life questionnaires and at least 2-year anatomic support data. Our secondary aim is to describe stress urinary incontinence and sexual health outcomes at 3 years after the incident surgery.
MATERIALS AND METHODS
This was a planned final analysis of our multicenter, double-blind RCT comparing traditional vaginal prolapse surgery without mesh with vaginal colpopexy repair with mesh for women with symptomatic stage 2–4 prolapse. Six fellowship-trained reconstructive pelvic surgeons conducted the study after institutional review board approval was obtained at each of the three sites. Specific details regarding the study design, methods, surgical interventions, and outcomes at 3 months and 1 year have previously been published.12,13 Women were assigned a treatment group using computer-generated random numbers stratified by the presence or absence of a uterus. After anesthesia was induced in the operating room, the surgeon opened an opaque sealed envelope revealing the surgical intervention: either traditional vaginal prolapse surgery without mesh (combined anterior, posterior colporrhaphy, or both; uterosacral ligament suspension; or sacrospinous ligament fixation) or vaginal colpopexy repair with mesh (synthetic mesh implant kit using Anterior Prolift, Modified Anterior Prolift with apical arms, or Total Prolift). Women with uterovaginal prolapse underwent vaginal hysterectomy per study protocol. The patients and postoperative examiners at each site were blinded to the assigned treatment. Stopping criteria included greater than 15% mesh exposures, greater than 1% mesh infection, greater than 1% fistula formation, and greater than 5% de novo dyspareunia rate.
Patients were seen for postoperative follow-up visits at 3 months, 1 year, 2 years, and 3 years. At each visit, a physician or nurse practitioner not involved in the original surgery performed a blinded examination and collected validated quality-of-life questionnaires. All adverse events were recorded and complications categorized using a modified Clavien-Dindo classification system.14 For the purposes of this analysis, we included those participants with complete 3-year quality-of-life questionnaires and at least 2-year anatomical data. Participants who were lost to follow-up were excluded.
Our primary outcome evaluated anatomic, symptomatic, and composite definitions of cure. Anatomical measurements were made using the Pelvic Organ Prolapse Quantification (POP-Q) examination15 with cure being defined in three different ways: 1) POP-Q stage 1 or less; 2) no prolapse beyond the hymen (Ba, Bp, C 0 or less); or 3) no prolapse beyond the hymen with adequate apical support above the midvagina (Ba, and Bp 0 or less, C to TVL/2 or less). Degree of symptom bother was evaluated using the short forms of the Pelvic Floor Distress Inventory16 and its subscales. The effect of prolapse on the patient’s quality-of-life was assessed using the Pelvic Floor Impact Questionnaire16 and its subscales as well as a Patient Global Impression of Improvement.17 Finally, we used the Pelvic Organ Prolapse/Urinary Incontinence Sexual Questionnaire18 to assess patient sexual function. Symptomatic cure was defined in three different ways: 1) the absence of bulge symptoms (Pelvic Floor Distress Inventory item 3, “Do you usually have a bulge or something falling out that you can see or feel in the vaginal area?” with a response of “no”); 2) Patient Global Impression of Improvement responses of “very much better” or “much better”; or 3) the absence of bulge symptoms and Patient Global Impression of Improvement responses of “very much better” or “much better.” Participants who underwent reoperation for recurrent prolapse or pessary use were excluded from the analysis for further anatomic and subjective outcomes but were considered failures for composite outcome cure rates.
We defined bothersome dyspareunia as a response of “usually” or “always” on Pelvic Organ Prolapse/Urinary Incontinence Sexual Questionnaire question 5 (“Do you feel pain during sexual intercourse?”). Bothersome baseline dyspareunia was recorded and we tracked resolution and persistence of symptoms. De novo dyspareunia was defined as those without baseline bothersome symptoms who developed bothersome dyspareunia at 3 years. Similar analyses were performed for stress urinary incontinence symptoms using question 17 (“Do you usually experience urine leakage related to coughing, sneezing, or laughing?”) of the Pelvic Floor Distress Inventory. Bothersome symptoms required a response of “yes” with bother of “somewhat,” “moderately,” or “quite a bit.” Bothersome symptom resolution and persistence was tracked. De novo stress urinary incontinence was defined as those without baseline bothersome symptoms that developed bothersome stress urinary incontinence symptoms at 3 years.
Sample size calculation conducted for the primary outcome at 1 year required 45 participants in each arm to detect a 20% difference in success (70% no mesh and 90% mesh) with an alpha of .05 and 80% power assuming 15% loss to follow-up. Statistical analysis was performed using SAS 9.1 software. One patient did not receive the assigned mesh treatment because the surgeon felt there was inadequate vaginal caliber, so nonmesh repair was performed. We analyzed this subject in the nonmesh group for the 3-month and 1-year outcomes rather than using an intent-to-treat approach. This subject was lost to follow-up after 12 months and therefore not included in this 3-year analysis. First, we performed the complete case analysis with assumption of missing data completely at random. Baseline characteristics of the study population were calculated according to treatment group. Pelvic Organ Prolapse Quantification individual points, POP-Q stage, and quality-of-life questionnaires (Pelvic Floor Distress Inventory, Pelvic Floor Impact Questionnaire, Pelvic Organ Prolapse/Urinary Incontinence Sexual Questionnaire) at baseline and 3 years were analyzed within groups and between groups. We calculated means and standard deviations or medians and interquartile ranges for continuous variables and frequency (proportion) for the categorical variables. t-tests, Wilcoxon signed rank, and Wilcoxon rank-sum tests were used to compare the means and distributions for continuous variables as appropriate. The exact χ2 was used to compare the proportions for the categorical variables. Anatomic, symptomatic, and combined definitions of cure outcomes, dyspareunia, and stress urinary incontinence symptoms between groups were compared using Fisher’s exact test. In addition, we performed analyses adjusted for censoring using inverse probability weighting to examine whether this analysis will change the conclusion of complete case analysis. The patient who changed treatment groups was analyzed in the mesh group for the sensitivity analysis. Logistic regression model was used to estimate the inverse probability weighting.19 A P value <.05 was considered statistically significant.
From January 2007 to August 2009 a total of 238 patients planning vaginal surgery for stage 2–4 prolapse were approached for inclusion in our study. Sixty-five women were enrolled (33 mesh, 32 no mesh) before the study was prematurely halted as a result of a 15.6% mesh exposure rate. Table 1 lists the baseline characteristics of our study population. There were no significant differences between the two treatment groups. When we compared the data in Table 1 with the original study population baseline data, we discovered that those excluded from the 3-year analysis (n=14) resulting from lack of 3-year quality-of-life questionnaires were more likely to be current smokers (6 [43%] compared with 0, P<.01) with fewer years of college or graduate education (4 [29%] compared with 38 [75%], P<.01) compared with those who were included (n=51). For participants with mesh, those excluded (n=8) were more likely to be current smokers (4 [50%] compared with 0, P<.01), have fewer years of college or graduate education (2 [25%] compared with 21 [84%], P<.01), and have more severe POP-Q stage (stage IV: three [38%] compared with two [8%], P=.04) compared with those included (n=25). For participants without mesh, those excluded (n=6) were more likely to be current smokers (2 [33%] compared with 0, P<.01).
At 3 years, 51 (78%) participants had quality-of-life questionnaires (25 mesh, 26 no mesh) and 41 (63%) had 2- or 3-year POP-Q examinations (20 mesh, 21 no mesh). Length of follow-up was similar for both groups with a median of 3.05 years (interquartile range 2.97–3.15). Three participants died (one mesh, two no mesh) of causes unrelated to the surgical procedure. All three participants with recurrent prolapse who required reoperation were in the mesh group. A total of eight women were lost to follow-up (three mesh, five no mesh) (Fig. 1). The reoperations for recurrent prolapse and mesh complications have been described in detail in the 1-year report.13 Only one prolapse reoperation occurred more than 1 year after surgery in a woman who had previously undergone a prolapse reoperation before 1 year after surgery. This subject initially underwent anterior mesh repair with a transobturator sling and cystotomy repair. She underwent a sling revision and iliococcygeus suspension 4.5 months after the index surgery for voiding dysfunction, urinary retention, vaginal vault prolapse, and a posterior enterocele. She later underwent a second reoperation 17.5 months after the index surgery with a posterior vaginal mesh repair and sacrospinous ligament fixation for recurrent vaginal vault prolapse, enterocele, and rectocele. There was one new mesh exposure since the 1-year outcomes (five total, three that required surgical excision) were presented. The small mesh exposure at the posterior apex was identified at the 3-year study visit in an asymptomatic patient who had previously undergone surgical excision of mesh exposure at the same location 4 months after the index Total Prolift. There were no other study-related adverse events since the 1-year outcomes.
There were no differences between the mesh and no-mesh groups at the final examination with respect to POP-Q stage and individual POP-Q points (Appendix 1, available online at http://links.lww.com/AOG/A421). The majority in each group (90% mesh and 86% no mesh) showed an improvement in the overall stage of prolapse from baseline to 3 years (P<.01). Symptomatic improvement was observed with no differences in scores between groups at 3 years for the Pelvic Floor Distress Inventory, Pelvic Floor Impact Questionnaire, and Pelvic Organ Prolapse/Urinary Incontinence Sexual Questionnaire including the subscale scores (Appendix 2, available online at http://links.lww.com/AOG/A422).
Cure rates did not differ between groups using a variety of definitions (Table 2). Overall anatomic cure rates were higher when the hymen was used as the threshold (85% mesh, 71% no mesh) compared with POP-Q stage 1 or less (45% mesh, 43% no mesh). Anatomic cure was lowest for the anterior compartment and did not differ between groups (POP-Q stage less than 1 [13 (65%) mesh compared with 9 (43%)] no mesh, relative risk [95% confidence interval] 1.52 [0.84–2.73], P=.21) and those with no prolapse beyond the hymen (19 [95%] mesh compared with 15 [71%] no mesh, relative risk [95% confidence interval] 1.33 [0.99–1.77], P=.09). Bulging symptoms were absent in 92% of mesh and 81% of no-mesh participants. The majority of each group (88% mesh, 81% no mesh) was much better or very much better on the Patient Global Impression of Improvement. At the termination of the study, three participants in the mesh arm (13%) required reoperation for recurrent prolapse and nobody used a pessary postoperatively. Although success rates were lower using a combined definition, there was no difference between groups. Thirty-five percent of mesh participants and 38% of nonmesh participants were cured when POP-Q stage 1 or less was used in the combined definition compared with 65% of mesh and 57% of nonmesh participants when the hymen was used in the definition.
We discovered similar results when we analyzed the anatomic and symptomatic outcomes (Table 2; Appendices 1 and 2, http://links.lww.com/AOG/A421 and http://links.lww.com/AOG/A422) adjusting for censoring using inverse probability weighting (data not included).
Of the 14 (56%) mesh participants who were sexually active at baseline, nine had dyspareunia, four had no dyspareunia, and one did not respond. Dyspareunia persisted in five, resolved in two, and two were no longer sexually active at 3 years. Two women developed de novo dyspareunia at 3 years. Of the 11 (42%) no-mesh participants who were sexually active at baseline, seven had dyspareunia and four had no dyspareunia. Dyspareunia persisted in five, resolved in one, and one did not respond at 3 years. One woman developed de novo dyspareunia at 3 years. Four participants (one mesh, three no mesh) who were not sexually active at baseline became sexually active and one had dyspareunia (no mesh). There were no significant differences between groups with respect to sexual activity and dyspareunia rates at baseline or 3 years.
At enrollment, 13 (52%) of mesh and 12 (46%) of no-mesh participants had baseline stress urinary incontinence symptoms. A total of eight concomitant slings were performed in the mesh group and six in the no-mesh group. The majority of those undergoing a sling for symptomatic stress urinary incontinence were cured (five of seven mesh, three of four no mesh), and no new stress urinary incontinence developed when a sling was performed in participants without baseline stress urinary incontinence symptoms (one mesh, two no mesh). Stress urinary incontinence persisted in the majority of those who did not undergo a concomitant sling (six of six mesh, five of eight no mesh). There were three delayed procedures performed in two mesh participants (one collagen injection followed by a sling and one collagen injection alone). There were five cases of de novo stress urinary incontinence (three mesh, two no mesh) and one in each group underwent a sling with symptom resolution. There were no significant differences in stress urinary incontinence outcomes at baseline or 3 years between groups.
We found no significant differences in cure rates at 3 years between the mesh and no-mesh groups regardless of the definition used. Moreover, the mesh group had a greater than 15% risk of mesh exposure. There are several other RCTs with shorter follow-up that showed high mesh exposure rates of 14–21%5,9,20,21 and significant improvement in anatomic outcomes associated with the use of mesh in the anterior compartment.5,7–9,20,21 A summary of recent RCTs on the subject is provided as Appendix 3 (available online at http://links.lww.com/AOG/A423). Appendix 3 (available online at http://links.lww.com/AOG/A423). Although we found the anterior wall cure at 3 years was more likely to occur in the mesh group, the difference was not statically significant. The most likely explanation for this is the fact that we are underpowered to detect a difference as a result of stopping the study early because of mesh exposures. This may also be accounted for by the use of blinded examiners, which decrease surgeon bias. In fact, our recent study found that an unblinded examination might account for as much as a 15% difference in objective cure at 12 months.22
Interestingly, although there was no difference between the mesh and no-mesh groups for the combined outcomes, there was an overall high rate of recurrence in our study (65% mesh compared with 62% no mesh using POP-Q stage 1 or less and 35% mesh compared with 43% no mesh using the hymen). There are several factors that may contribute to this outcome. First, our study analyzed all compartments as compared with most other RCTs that focused on the anterior compartment alone.5,7–9 Second, as previously stated, surgeon bias was minimized by the use of blinded examiners. Finally, we used a robust definition of success that includes both objective and subjective outcomes. Altman et al8 used a similar definition and found the recurrence rate to be elevated despite the use of unblinded examiners (39% mesh compared with 66% no mesh).
When the hymen was used as a threshold for anatomic cure, the success rates increased (85% mesh compared with 71% no mesh). This definition has been shown to be more clinically relevant. Swift et al23 found that pelvic floor symptoms increased from an average of 0.5 symptoms in patients with stage I pelvic support to 2.1 symptoms in women when the leading edge of prolapse extended beyond the hymen. This notion is underscored in our study by the lack of bulging and protrusion symptoms in 92% of mesh and 81% of no-mesh participants.
Our study revealed low rates of de novo dyspareunia (8% mesh compared with 4% no mesh) and sexual function did not differ between groups. Overall rates of dyspareunia range from 2% to 16%7–9,20 in the studies listed in Appendix 3. We had a high rate of baseline dyspareunia (64%) in our study population and only approximately 50% were sexually active at baseline. Stress urinary incontinence symptoms generally improved with a sling and there only five cases of de novo stress urinary incontinence. The few patients without baseline bothersome stress urinary incontinence who underwent a sling did not develop stress urinary incontinence symptoms. There were very few subsequent procedures for bothersome stress urinary incontinence symptoms despite the fact that several patients had persistent or de novo stress urinary incontinence. This may indicate a patient's reluctance to undergo additional surgical procedures for bothersome stress urinary incontinence.
Our study had some limitations. There was a lack of statistical power secondary to early termination and the resultant low number of study participants. There was significant loss to follow-up with 51 (78%) of patients having 3-year quality-of-life questionnaires and 41 (63%) of patients having 2- or 3-year POP-Q examinations. As a result, we only had 6% power to detect a 2% (45% mesh and 43% no mesh) difference in POP-Q stage less than 1 and 19% power to detect a 14% (85% mesh and 71% no mesh) difference for no prolapse beyond the hymen using alpha of .05. However, the length of follow-up was similar between groups. After censoring participants who were lost to follow-up and who had surgery for recurrent prolapse, we found that there were differences in the baseline characteristics of the current 3-year study population and the original study population that may have affected our outcomes. Nevertheless, when additional analysis adjusting for censoring was performed using inverse probability weighting, the conclusions were unchanged.
The major strength of our study is the double-blind, multicenter RCT design and the duration of follow-up. The majority of the studies in the literature are limited to 1-year follow-up. In addition, our study included all vaginal compartments. In our review of the literature, we were able to identify two other multicompartment studies, which had only 1-year follow-up data. In addition, we used a robust definition for successful repair that includes both objective and subjective outcomes.
We are reassured by the fact that overall quality-of-life improvement was high in both groups despite the fact that there was no observed anatomic benefit with the addition of mesh and a greater than 15% mesh exposure rate. Our study and recent studies highlight the need for a balanced approach to the use of vaginal mesh that includes a thorough patient consent process before its use in prolapse repairs.
1. Abdel-Fattah M, Ramsay I; West of Scotland Study Group. Retrospective multicenter study of the new minimally invasive mesh repair devices for pelvic organ prolapsed. BJOG 2008;115:22–30.
2. Fatton B, Amblard J, Debodinance P, Cosson M, Jacquetin B. Transvaginal repair of genital prolapsed: preliminary results of a new tension-free vaginal mesh (Prolift technique)—a case series multicentric study. Int Urogynecol J Pelvic Floor Dysfunct 2007;18:743–52.
3. Aungst MJ, Friedman EB, von Pechmann WS, Horbach NS, Welgoss JA. De novo stress incontinence and pelvic muscle symptoms after transvaginal mesh repair. Am J Obstet Gynecol 2009;201:73.e1–7.
4. Gauruder-Burmester A, Koutouzidou P, Rohne J, Gronewold M, Tunn R. Follow-up after polypropylene mesh repair of anterior and posterior compartments in patients with recurrent prolapse. Int Urogynecol J Pelvic Floor Dysfunt 2007;18:1059–64.
5. Nieminen K, Hiltunen R, Takala T, Heiskanen E, Merikari M, Niemi K, et al.. Outcomes after anterior vaginal wall repair with mesh: a randomized, controlled trial with a 3 year follow-up. Am J Obstet Gynecol 2010;203:235.e1–8.
6. Hiltunen R, Nieminen K, Takala T, Heiskanen E, Merikari M, Niemi K, et al.. Low-weight polypropylene mesh for anterior vaginal wall prolapse: a randomized controlled trial. Obstet Gynecol 2007;110:455–62.
7. Nguyen JN, Burchette RJ. Outcome after anterior vaginal prolapse repair: a randomized controlled trial. Obstet Gynecol 2008;111:891–8.
8. Altman D, Väyrynen T, Engh ME, Axelsen S, Falconer C; Nordic Transvaginal Mesh Group. Anterior colporrhaphy versus transvaginal mesh for pelvic-organ prolapse. N Engl J Med 2011;364:1826–36.
9. Menefee SA, Dyer KY, Lukacz ES, Simsiman AJ, Luber KM, Nguyen JN. Colporrhaphy compared with mesh or graft-reinforced vaginal paravaginal repair for anterior vaginal wall prolapse: a randomized controlled trial. Obstet Gynecol 2011;118:1337–44.
11. Barber MD, Brubaker L, Nygaard I, Wheeler TL, Schaffer J, Chen Z, et al.; Pelvic Floor Disorders Network. Defining success after surgery for pelvic organ prolapse. Obstet Gynecol 2009;114:600–9.
12. Iglesia CB, Sokol AI, Sokol ER, Kudish BI, Gutman RE, Peterson JL, et al.. Vaginal mesh for prolapse: a randomized controlled trial. Obstet Gynecol 2010;116:293–303.
13. Sokol AI, Iglesia CB, Kudish BI, Gutman RE, Shveiky D, Bercik R, et al.. One-year objective and functional outcomes of a randomized clinical trial of vaginal mesh for prolapse. Am J Obstet Gynecol 2012;206:86.e1–9.
14. Dindo D, Demartines N, Clavien PA. Classification of surgical complications: a new proposal with evaluation in a cohort of 6336 patients and results of a survey. Ann Surg 2004;240:205–13.
15. 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.
16. 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;193:103–3.
17. Yalcin I, Bump R. Validation of two global impression questionnaires for incontinence. Am J Obstet Gynecol 2003;189:98–101.
18. Rogers RG, Coates KW, Kammerer-Doak D, Khalsa S, Qualls C. A short form of the Pelvic Organ Prolapse/Urinary Incontinence Sexual Questionnaire (PISQ-12). Int Urogynecol J Pelvic Floor Dysfunct 2003;14:164–8.
19. Hogan JW, Roy J, Korkontzelou C. Handling drop-out in longitudinal studies. Stat Med 2004;23:1455–97.
20. Withagen MI, Milani AL, den Boon J, Vervest HA, Vierhout ME. Trocar-guided mesh compared with conventional vaginal repair in recurrent prolapse: a randomized controlled trial. Obstet Gynecol 2011;117:242–50.
21. Halaska M, Maxova K, Sottner O, Svabik K, Mlcoch M, Kolarik D, et al.. A multicenter, randomized, prospective, controlled study comparing sacrospinous fixation and transvaginal mesh in the treatment of posthysterectomy vaginal vault prolapse. Am J Obstet Gynecol 2012;207:301.e1–7.
22. Antosh DD, Iglesia CB, Vora S, Sokol AI. Outcome assessment with blinded versus unblinded POP-Q exams. Am J Obstet Gynecol 2011;205:489.e1–4.
23. Swift SE, Tate SB, Nicholas J. Correlation of symptoms with degree of pelvic organ support in a general population of women: what is pelvic organ prolapse?. Am J Obstet Gynecol 2003;189:372–7.