Treatment Success of Retropubic and Transobturator Midurethral Slings at 24-months

Abstract

Purpose

Longer-term comparative efficacy information regarding transobturator and retropubic midurethral slings (TMUS and RMUS) is needed. We report 24 month continence rates, complications and symptom outcomes from a randomized equivalence trial.

Material & Methods

Primary outcomes were objective (negative stress test, negative pad test and no retreatment for stress urinary incontinence, SUI) and subjective (no self-report of SUI symptoms, no leakage episodes on 3-day bladder diary and no retreatment for SUI) success at 24 months. Predetermined equivalence margin was ±12%.

Results

516 of 597 (86.4%) randomized participants were assessed. Objective success rates for RMUS and TMUS were 77.3% and 72.3%, respectively (95% CI for difference of 5.1%: −2.0, 12.1%), subjective success rates were 55.7% and 48.3% (CI for difference of 7.4%: −0.7, 15.5%). Neither objective nor subjective success rates met the pre-specified criteria for equivalence.

Patient satisfaction (RMUS 86.3% vs. TMUS 88.1%, p=0.58), frequency of de novo urgency incontinence (RMUS 0% vs. TMUS 0.3%, p= 0.99), and occurrence of mesh exposure (RMUS 4.4% vs. TMUS 2.7%, p=0.26) were not significantly different. RMUS participants had higher rates of voiding dysfunction requiring surgery (3.0% vs. 0%, p=0.002) and urinary tract infections (17.1% vs. 10.7%, p=0.025) while the TMUS group had more neurologic symptoms (9.7% vs. 5.4%, p=0.045).

Conclusions

At 24 months, objective success rates, which met criteria for equivalence at 12 months, no longer met these criteria. Subjective success rates remained inconclusive for equivalence. Patient satisfaction remained high and symptom severity remained markedly improved. Continued surveillance is important in women undergoing MUS.

INTRODUCTION

Midurethral slings with synthetic mesh are the most frequently used surgical procedures for stress urinary incontinence (SUI) in the United States and Europe.^{1, 2} Two common approaches are used to place the sling at the mid-urethra. The sling is passed transvaginally behind the pubic bone with the retropubic approach³ whereas in the transobturator approach it is passed laterally through the obturator foramen in an attempt to avoid the pelvic organs and vasculature in the retropubic space.⁴ Recent meta-analyses of randomized and quasi-randomized controlled trials and prospective studies comparing the retropubic with the transobturator approach suggest they have similar efficacy in the short term. However, the studies considered were of too low quality to permit definitive conclusions to be made about the comparative efficacy and safety of these approaches beyond 12 months. ^5–7

Based on the literature available when TOMUS was designed, we conducted an equivalence trial rather than a superiority or non-inferiority trial in order to more definitively state that the two MUS approaches were equivalent. We previously reported 12 month success rates after surgery in the trial of midurethral slings (TOMUS).⁸ Consistent with the original design of the trial we report success rates, quality of life, patient satisfaction, adverse events and other outcomes of this clinical trial 24 months after surgery.

METHODS

We conducted a multicenter, randomized equivalence trial of retropubic and transobturator midurethral sling surgery in women with SUI. Details of the study design have been previously described.⁹ Women seeking surgery for SUI who were 21 years of age or older, had symptoms of stress predominant urinary incontinence and had a positive stress test at a bladder volume ≤ 300 mL were eligible for the study. Two primary outcomes of surgical success were assessed. Objective success was defined as a negative provocative stress test at bladder volume ≥ 300 cc, a negative 24 hour pad test and no re-treatment for SUI (including behavioral, pharmacological or surgical). Subjective success was defined as the absence of self- reported symptoms of SUI from the Medical, Epidemiological and Social Aspects of Aging (MESA) questionnaire,¹⁰ no leakage of urine on a 3-day voiding diary and no retreatment for SUI. Secondary outcomes included complications/morbidity, self-report of quality of life, symptom bother, satisfaction and global improvement. Quality of Life was assessed by the Incontinence Impact Questionnaire (IIQ) with possible scores of 0 to 400 with higher scores indicating more negative impact on QOL¹¹ and the International Consultation on Incontinence Questionnaire (ICIQ) with possible scores of 0 to 21 with higher scores indicating more negative impact.¹² Symptom bother was assessed by the Urogenital Distress Inventory (UDI)with possible scores of 0 to 300 with higher scores indicating higher distress.¹¹ Additionally, participants completed the Patient Global Impression of Improvement (PGI-I), a single item querying their overall improvement and ranging from very much better to very much worse. ¹³ Adverse events, adjudicated by a group of study investigators blinded to the surgical procedure, were classified with a modification of the Dindo system.¹⁴ The study protocol was approved by an institutional review board at each of the participating sites. Written informed consent was obtained from each study participant. An independent data and safety monitoring board reviewed the progress, interim results and safety of the study.

With 250 women who had available data in each group, it was calculated that this study would have 80% power to demonstrate equivalence between success rates in the two sling approaches (equivalence margin of ±12 percentage points), at a two-sided significance level of 5%. The equivalence margin was chosen for clinical considerations (e.g. even if the two treatments differed by as much as 12 percentage points, we would still feel comfortable deeming the two arms equivalent) and practical considerations (e.g. number of patients it was feasible to enroll in the trial). Generalized linear modeling, assuming a logit link and binomial distribution, was used for calculating the rates of treatment success. Equivalence for the primary outcome was declared if the entire 95% confidence interval for the difference between the two surgical approaches was within the equivalence margin.

Only women who had received their assigned surgery (per protocol) were included in the primary outcome analysis. As in the primary TOMUS paper, per protocol analyses were used here to analyze the primary outcomes because they provide more conservative estimates when evaluating equivalence in contrast to intent to treat analyses that may tend to bias towards concluding that the two arms are similar.^8,15 We performed a secondary analysis of the primary outcome and analyses of secondary outcomes on women based on randomized assignment (intention-to-treat population). The rates of surgical success at 24-months were calculated under the assumption that women who were lost to follow-up were surgical successes since women were assumed to be successes until failures were reported. We performed sensitivity analyses by both excluding women lost to follow-up and considering them to be surgical failures. Fisher’s exact test was used to compare the proportions of participants in each group who had one or more adverse events. Continuous outcomes were analyzed by least-squares modeling methods. Repeated measures modeling was used to assess changes by treatment and/or time (visit) for relevant secondary outcomes.

No formal adjustment for multiple comparisons was made. Analyses were performed with SAS statistical software, version 9.2 (SAS Institute).

RESULTS

The number of subjects screened for eligibility, randomized and assessed at 12 and 24 months post-surgery is shown in Figure 1. Five hundred ninety seven women were randomized and 552 (92.5%) and 516 (86.4%) were assessed per protocol at 12 and 24 months, respectively. The proportion of women who were either assessed per protocol at 24 months or failed surgery within that time was similar in the retropubic and transobturator surgery groups; 253 (84.9%) and 263 (88.0%), respectively.

Figure 1.

Baseline demographic and clinical characteristics were similar in both surgery groups (data not shown) with the exception of valsalva leak point pressure (VLPP), which was lower in the retropubic arm compared to the transobturator arm (mean of 114.4 (SD +/− 43.1) versus 124.2 (+/− 41.4) cm/H₂O respectively, p=0.03)⁸ The mean age of study participants was 53 years (+/− 11 years). Both groups had a median of 2.7 incontinence episodes per day; and a median of 12.5 grams of urine loss measured on a 24 hour pad test. Seventy nine women (13%) reported a prior surgery for urinary incontinence. The frequency of concomitant surgery was similar in the retropubic and transobturator arms (25% and 26%, respectively).

The unadjusted 24 month objective success rates for retropubic and transobturator slings were 77.3% and 72.3% respectively; (95% CI for difference of 5.1%: −2.0, 12.1%) while subjective success rates were 55.7% and 48.3% (CI for difference of 7.4%: −0.7%, 15.5%), respectively. Neither the objective nor subjective success rates met the pre-specified criteria for equivalence. However, each of the confidence intervals included 0% indicating that the success rates cannot be deemed different from one another either. (Figure 2) These rates did not change significantly after statistical adjustment for clinical center, concomitant surgery and the urodynamic measures of VLPP and maximal urethral closure pressure (MUCP) or when study participants who were lost to follow-up were considered either as surgical failures or excluded from the analysis. The proportion of women who were treatment failures at 24 months by the individual components of the primary outcomes is shown in Figure 3. Note that a woman could fail by more than one criterion.

Figure 2.

Figure 3.

Over twenty-four months of follow-up, a total of 75 serious adverse events (SAEs) were identified in 70 women; the rate in the retropubic arm (15.1%) was approximately twice that in the transobturator arm (8.4%), primarily as a result of more intraoperative bladder injuries. Of the 75 SAEs reported, 12 (16.0%) occurred from 12 to 24 months post- surgery (4 in the retropubic arm and 8 in transobturator arm). Three hundred and eight adverse events (AEs) were experienced by 219 women during the 24 months following surgery; the rate was 40.6% (174 AEs in 121 women) in the retropubic and 32.8% (134 AEs in 98 women) in the transobturator arm (Table 1). Of the 308 AEs, 43 (14.0%) occurred from 12 to 24 months. (27 in the retropubic arm and 16 in the transobturator arm)

Table 1.

Adverse Events^a by Treatment Group, Severity^b, and System Up to 2 Years.

Adverse events by treatment group, severity and system up to 2 years
	Retropubic (298)			Transobturator (299)			p Value†
	Total Events	No. Events in 13–24 Mos	No. Unique Pts (%)*	Total Events	No. Events in 13–24 Mos	No. Unique Pts (%)*
Grades III–IV, requiring surgical, endoscopic or radiological intervention
Wound:	11	1	10 (3.4)	11	6	11 (3.7)	>0.99
Mesh exposure‡	10	1	9 (3.0)	6	5	6 (2.0)	0.45
Mesh erosion§	1	0	1 (0.3)	1	0	1 (0.3)
Surgical site infection	0	0	0 (0)	3	1	3 (1.0)
Granulation tissue	0	0	0 (0)	1	0	1 (0.3)
Genitourinary:	25	2	25 (8.4)	13	0	13 (4.3)	0.046
Urethral perforation	1	0	1 (0.3)	0	0	0 (0)
Bladder perforation	15	0	15 (5.0)	0	0	0 (0)
Vaginal epithelial perforation	6	0	6 (2.0)	13	0	13 (4.3)
Recurrent cystitis||	3	2	3 (1.0)	0	0	0 (0)
Vascular/hematological:	1	0	1 (0.3)	1	0	1 (0.3)	>0.99
Pulmonary embolus	0	0	0 (0)	1	0	1 (0.3)
Postop bleeding	1	0	1 (0.3)	0	0	0 (0)
Neurological symptoms¶	1	0	1 (0.3)	0	0	0 (0)	0.50
Voiding dysfunction requiring surgery	9	1	9 (3.0)	0	0	0 (0)	0.002
Pain per pt self-report 6 wks or more**	0	0	0 (0)	0	0	0 (0)	Not available
Other	0	0	0 (0)	3	2	3 (1.0)	0.25
Overall SAEs	47	4	45 (15.1)	28	8	25 (8.4)	0.011
Grades I–II, expectant or pharmacological intervention
Wound:	6	0	6 (2.0)	2	−3	2 (0.7)	0.18
Mesh exposure‡	4	0	4 (1.3)	2	−1**	2 (0.7)
Mesh erosion§	0	0	0 (0)	0	0	0 (0)
Surgical site infection	2	0	2 (0.7)	0	−2††	0 (0)
Genitourinary:	61	15	51(17.1)	35	8	32 (10.7)	0.025
Cystitis - culture proven	27	4	25 (8.4)	16	1	14 (4.7)
Cystitis - empirically treated	16	1	15 (5.0)	9	1	9 (3.0)
Recurrent cystitis||	18	10	16 (5.4)	10	6	10 (3.3)
Vascular/hematological:	20	0	18 (6.0)	7	0	7 (2.3)	0.026
Pulmonary embolus	0	0	0 (0)	0	0	0 (0)
Intraop bleeding	14	0	14 (4.7)	7	0	7 (2.3)
Postop bleeding	6	0	5 (2.0)	0	0	0 (0)
Neurological symptoms:¶	20	5	15 (5.0)	32	1	29 (9.7)	0.041
Numbness	9	1	6 (2.0)	9	1	7 (2.3)
Weakness	11	4	10 (3.4)	23	0	22 (7.4)
Voiding dysfunction:	10	0	10 (3.4)	6	2	6 (2.0)	0.33
Managed with catheter only	6	0	6 (2.0)	4	2	4 (1.3)
Managed with prescription drug only	2	0	2 (0.7)	1	0	1 (0.3)
Managed with behavioral treatment or neuro modulation (± catheter)	2	0	2 (0.7)	1	0	1 (0.3)
Pain per pt self-report 6 wks or more‡‡	7	0	7 (2.3)	7	0	6 (2.0)	0.79
Prescription for persistent urgency incontinence§§	42	6	42 (14.1)	38	8	38 (12.7)	0.63
Other||||	8	1¶¶	7 (2.3)	6	0	6 (2.0)	0.79
Overall AEs‡‡	174	27	121 (40.6)	134	16	98 (32.8)	0.051

Note: One subject randomized to the retropubic procedure received a transobturator sling and had a culture proven cystitis (1 of the 23 in retropubic arm); one subject randomized to the transobturator procedure received the retropubic procedure and had no SAE’s or AE’s.

^aAdverse event (AE) defined as a deviation from the normal intra- or post-operative course (Grades I and II). A single patient can have multiple entries.

^bSeverity grade determined by a slightly modified version of the Dindo¹⁸ classification system, which is based on the level of therapy required to treat an event. Serious Adverse Event (SAE) defined as ≥ Grade III–V; no grade IV or V events occurred in either group.

I – No pharmacologic, surgical, or radiologic intervention (allowed therapeutic regimens include antiemetics, antipyretics, analgesics, diuretics, electrolytes, and physiotherapy).

II – Required pharmacologic treatment with drugs other than such allowed for grade I complications (antibiotics, blood transfusions and total parenteral nutrition are included). III – Required surgical, endoscopic or radiologic intervention.

IV – Life-threatening complication requiring intensive care management. V – Death.

^cP-values for adverse events were calculated from Fisher’s exact tests comparing the number of unique patients with events by treatment group using the total number of patients as denominators.

^dNumber of unique patients is the number of patients having a particular adverse event (less than or equal to the number of events). The percent equals number of unique patients with an adverse event divided by the number of patients in the treatment group.

^eMesh exposure defined as mesh visualized in the vagina through a prior incision area with or without an inflammatory reaction.

^fMesh erosion defined as erosion after primary healing, into an organ or surrounding tissue.

^gRecurrent cystitis defined as presumed UTI with treatment, ≥3 in 1 year after 6 week visit.

^hNeurologic symptoms are defined as patient self report on standardized form of new paresthesias or alteration in motor function that developed between surgery and the 6 week visit. The location of numbness or weakness was ascertained by the patient marking a body map. There were three neurologic SAEs in the Retropubic arm with one report of lower leg numbness, another with upper leg numbness and one with upper leg weakness. The neurologic SAE in the Transobturator arm was upper leg numbness. Neurologic AEs were listed by location.

ⁱThere were 3 reported mesh exposure AEs at the 12 month visit and 2 reported at the 24 month visit, so the difference is −1. Examining more closely, two of the mesh exposures were re-classified as mesh exposure SAEs at the 24 month visit, one was previously reported at 12 months and one was new at the 24 month visit.

^jThere were 2 reported surgical site infection AEs at the 12 month visit and 0 reported at the 24 month visit, so the difference is −2. Examining more closely, both of the surgical site infections were re-classified at the 24 month visit, one as a mesh exposure SAE and the other as a surgical site infection SAE.

^k

Pain defined as self report at or beyond 6 weeks post surgery by the following questions:

1. Patient answers “yes” to the introductory question “Have you had any pain within the last 24 hours as a result of your incontinence operation?” and
2. Patient answers any of the first three McCarthy²³ pain questions at a level 75mm or greater on the visual analog scale (150mm total length). and
3. Patient answers the bother question on the McCarthy visual analog scale at a level 75mm or greater.

^lDe novo urgency urinary incontinence (UUI) defined as baseline pure stress incontinence patient at or beyond 6-week visit now answering any MESA¹⁹ urge question “sometimes” or “often,” and/or, has initiated treatment with anticholinergic medication for UUI.

^mPersistent UUI was defined as a baseline mixed UI patient answering any MESA urge question “sometimes” or “often,” and/or, has initiated treatment with anticholinergics for UUI.

ⁿOther adverse events include: granulation tissue, pyelonephritis, pelvic pain not associated with a UTI, anxiety, thrush, urinary tract colonization, wound edge separation, minor wound, vaginal cuff from hysterectomy/suspension, medication reaction, bi-lobed fluid filled sac, skin irritation

^oThere were two new “other” events that occurred between the 12 and 24 month visits and one event that was re-classified to a neurologic AE at 24 months from an “other” AE at 12 months. Hence, the net change is only 1 additional “other” AE.

When SAE and AEs were combined by complication type, rates of de novo urgency incontinence (0% vs. 0.3%, p> 0.99), persistent urgency incontinence (14.1% vs. 12.7%, p=0.63) and mesh exposures (4.4% vs. 2.7%, p=0.26) were similar in the retropubic and transobturator groups respectively. However, the retropubic group had higher rates of voiding dysfunction requiring surgery (3.0% vs. 0%, p=0.002) and urinary tract infections (17.4% vs. 10.7%, p=0.018) while the transobturator group had higher rates of neurologic symptoms (9.7% vs. 5.4%, p=0.045).

Similarly high rates of patient satisfaction were observed in both the retropubic and transobturator groups (86.3% vs. 88.1%, respectively, p=0.58,) at 24 months. (Figure 4, Panel a) However, the proportion of women who reported being much or very much better on the PGI-I at 24 months was higher in the transobturator group (91.5%) than in the retropubic group (86.8%), p=0.02 (Figure 4, Panel b). Based on repeated measures modeling of the quality of life measures, neither the change from baseline in total UDI scores (retropubic −100 vs. transobturator −107, p =0.13), nor in total IIQ scores, (retropubic − 125 vs. transobturator − 124, p=0.89) differed by treatment group.

Figure 4.

DISCUSSION

This report extends our previous findings from a multi-center equivalence trial of retropubic and transobturator midurethral slings by following study participants up to 24 months after surgery. While objective and subjective success rates for both approaches decreased modestly from 12 to 24 months after surgery, these decreases reflect a continued trend for greater benefit from the retropubic as compared to the transobturator midurethral sling. The decrease in objective success rates over that time period resulted in not meeting our pre-specified criteria for equivalence of the two approaches at 24 months. This is in contrast to our previous report at 12 months of clinical and statistical equivalence of objective success rates for the retropubic and transobturator approaches.⁸ Similar to 12 month findings, the subjective success rates also did not reach the pre-determined criteria for equivalence at 24 months. Despite the decline in objectively and subjectively assessed success rates, study participants from both surgical groups reported a high level of satisfaction as well as improvements in urinary symptom severity and quality of life 24 months after surgery.

As noted above, the decline in success rates also led to a widening difference between the transobturator and retropubic arms between 12 and 24 months. The difference in objective success rates increased from 3% to 5.1% while the difference in subjective success rates increased from 6.4% to 7.1% both in favor of the retropubic approach. A recent Cochrane review comparing the retropubic and transobturator mid-urethral slings demonstrated subjective cure rates of 83% with no statistically significant difference in the subjective cure and improvement rates between the two approaches at 12 months. However longer follow up was not available.⁵ An updated review on surgery for SUI conducted by Novara et al, demonstrated similar subjective efficacy rates and a trend towards higher objective efficacy in the retropubic group similar to our findings.⁷

Use of composite outcomes, such as those in the TOMUS trial to define treatment success, has been shown to result in lower success rates compared to a single outcome measure used in many previous studies comparing the efficacy of midurethral slings.^16–18 This may explain the relatively low success rates we observed at 12 and 24 months after surgery. Our approach of including both objective and subjective outcomes is reasonable given the recent emphasis to consider patient-centered and other domains to assess surgical success for urinary incontinence. When individual components of the primary outcomes were considered, failure by stress testing and stress symptoms on the MESA was higher in the transobturator than the retropubic arm, suggesting the possibility of superior continence rates with the RMUS approach which is consistent with the trends seen in overall treatment success. Similar to several prior reports success rates varied in our study depending on which outcome measure was used.^{17, 18}

Over 24 months the AE profile differed significantly between the two surgical approaches. With respect to specific AEs, intraoperative bladder perforations, postoperative voiding dysfunction requiring surgical intervention and urinary tract infections each occurred more frequently in the retropubic group, whereas groin and lower extremity neurologic symptoms were more commonly observed in the transobturator group. These findings are similar to conclusions reported in a recent systematic review and meta-analysis.⁷ The longer-term follow-up informs clinicians and patients about the continued occurrence of known complications over time.

The TOMUS trial provides important information regarding complications from mesh products used in contemporary surgical procedures for stress urinary incontinence over a two- year period. Our study included a standardized pelvic examination at 24 months to evaluate mesh complications. The overall rates for mesh exposure did not differ significantly by treatment group at twenty-four months. Also, the number of mesh exposures identified in the first year after surgery was higher in the retropubic group whereas in the 13–24 months time period more mesh exposures were seen in the transobturator group highlighting the need for long-term clinical follow-up of patients receiving mid-urethral sling surgery. Others have also reported on a later presentation of mesh complications. ^{6, 19} Our report is especially timely given the recent warnings from the Food and Drug Administration about the use of mesh products in pelvic floor and incontinence surgical procedures.²⁰

There are several strengths of this study. Both objective and subjective measures of surgical success were used to capture a broad spectrum of outcomes. Rates of participant assessment two years after surgery were high and did not vary significantly by surgery group. Complications from surgery were assessed in a standardized manner with standardized definitions across sites. While some may argue that the equivalence margin 12 percentage points is too wide, at the time of the trial design, the investigators concluded that the clinical significance of a smaller margin did not justify a sample size requiring hundreds of additional patients.

CONCLUSION

With an additional 12 months of follow-up to two years post-surgery, the rates of subjective and objective success continued to decline in women who received retropubic and transobturator midurethral slings. The decrease in objective success rates resulted in an inability to conclude that the two approaches were equivalent. The improvement in symptom severity and quality of life measures persists in both groups at 24 months post-operatively which coincides with high rates of participant satisfaction. The cumulative rate of serious adverse events was nearly twice as high in the RMUS group compared to TMUS group at 24 months, but occurred much less often in the second year of follow-up in both groups. Additional problems with mesh and urgency incontinence continued to arise in patients through the 24 month post-operative period, reinforcing the need for continued surveillance in our patients undergoing midurethral slings.

Key of Abbreviations

TMUS

transobturator

RMUS

retropubic midurethral sling

SUI

stress urinary incontinence

CI

confidence interval

MUS

midurethral sling

TOMUS

trial of midurethral

slings mL

milliliter

MESA

Medical, Epidemiological and Social Aspects of Aging

IIQ

Incontinence Impact Questionaire

QOL

quality of life

ICIQ

International Consultation on Incontinence

UDI

Urogenital Distress Inventory

PGI-I

Patient Global Impression of Improvement

VLPP

valsalva leak point pressure

MUCP

maximum urethral closure pressure

SAE

serious adverse event

AE

adverse event

PP

protocol

UI

urinary incontinence