Published at www.nejm.org May 21, 2007 (10.1056/NEJMoa070416) |
|
Published at www.nejm.org May 21, 2007 (10.1056/NEJMoa070416) |
| |||||||||||||||||||||||||||||||||
Background Many surgical procedures are available for women with urinary stress incontinence, yet few randomized clinical trials have been conducted to provide a basis for treatment recommendations.
Methods We performed a multicenter, randomized clinical trial comparing two procedures — the pubovaginal sling, using autologous rectus fascia, and the Burch colposuspension — among women with stress incontinence. Women were eligible for the study if they had predominant symptoms associated with the condition, a positive stress test, and urethral hypermobility. The primary outcomes were success in terms of overall urinary-incontinence measures, which required a negative pad test, no urinary incontinence (as recorded in a 3-day diary), a negative cough and Valsalva stress test, no self-reported symptoms, and no retreatment for the condition, and success in terms of measures of stress incontinence specifically, which required only the latter three criteria. We also assessed postoperative urge incontinence, voiding dysfunction, and adverse events.
Results A total of 655 women were randomly assigned to study groups: 326 to undergo the sling procedure and 329 to undergo the Burch procedure; 520 women (79%) completed the outcome assessment. At 24 months, success rates were higher for women who underwent the sling procedure than for those who underwent the Burch procedure, for both the overall category of success (47% vs. 38%, P=0.01) and the category specific to stress incontinence (66% vs. 49%, P<0.001). However, more women who underwent the sling procedure had urinary tract infections, difficulty voiding, and postoperative urge incontinence.
Conclusions The autologous fascial sling results in a higher rate of successful treatment of stress incontinence but also greater morbidity than the Burch colposuspension. (ClinicalTrials.gov number, NCT00064662
[ClinicalTrials.gov]
.)
Many surgical procedures have been described for women with stress incontinence, yet few randomized clinical trials have been conducted to provide a basis for treatment recommendations. The fascial-sling procedure and Burch colposuspension are two well-established procedures with reported cure rates of 70 to 85% at 5 to 8 years.9,10 In the Burch modified colposuspension,11 the anterior vaginal wall is suspended (at the level of the bladder neck) with permanent sutures tied to the iliopectineal ligament (Figure 1A). In the autologous sling procedure,12 a harvested strip of rectus fascia is placed transvaginally at the level of the proximal urethra. The fascial strip is secured superiorly to the rectus fascia with permanent sutures (Figure 1B). Although it has been suggested that the sling procedure may result in higher cure rates, this advantage may be offset by increased obstructive complications, such as voiding dysfunction and urge incontinence.13,14 We conducted a multicenter, randomized surgical trial, the Stress Incontinence Surgical Treatment Efficacy Trial, to compare the efficacy and safety of the sling and Burch procedures 24 months after surgery.
|
Patients and Study Design
Women who were planning to undergo stress-incontinence surgery were invited to participate in the trial. Eligibility requirements included documented pure or predominant symptoms of stress incontinence for at least 3 months and a positive standardized urinary stress test.
Details of the study methods have been published previously.15 All study procedures were approved by the institutional review board at each participating clinical center, and written consent was obtained from all women before enrollment. Randomization was performed in the operating room after anesthesia induction with the use of a permuted-block randomization schedule with stratification according to clinical site. The patients were aware of study-group assignments postoperatively. An independent data and safety monitoring board oversaw the progress, interim results, and safety of the study.
Formal interim time-to-event analyses of the primary outcome of overall success were planned for three time points, with the use of an O'Brien–Fleming stopping boundary, and were conducted when 19%, 44%, and 76% of failures had occurred. Although the test statistic at the third analysis crossed the stopping boundary in favor of the sling procedure, the protocol did not require stopping the trial when the boundary was crossed, and the data and safety monitoring board recommended that the study continue. No adjustment was made for these analyses.
Definitions of clinical terms, urodynamic nomenclature, and methods of evaluation of patients were uniform across all sites and in accordance with recommendations from the standardization committees of the International Continence Society.5,16 Key elements of the two surgical procedures were standardized among all participating surgeons and included the use of preoperative antibiotics, skin-incision length, number and type of Burch sutures, fascial-sling length and width, and cystoscopic evaluation of the bladder. Because these procedures are frequently performed in conjunction with surgery for pelvic prolapse, abdominal and vaginal approaches for both pelvic prolapse repair and hysterectomy were permitted. However, surgeons were required to declare before randomization which concomitant procedures would be performed.
The two primary outcomes were composite measures of success in terms of overall urinary-incontinence measures and of stress-incontinence measures specifically. Overall treatment success was defined as no self-reported symptoms of urinary incontinence, an increase of less than 15 g in pad weight during a 24-hour pad test, no incontinence episodes recorded in a 3-day diary, a negative urinary stress test (no leakage noted on examination during cough and Valsalva maneuvers at a standardized bladder volume of 300 ml), and no retreatment for urinary incontinence (including behavioral, pharmacologic, and surgical therapies). Since the study surgeries are intended to correct symptoms of stress incontinence without necessarily improving concomitant urge incontinence and the voiding diary and pad test do not differentiate between urge-incontinence and stress-incontinence events, the definition of success specific to stress incontinence was limited to no self-reported symptoms of stress incontinence, a negative stress test, and no retreatment for stress incontinence.
Data were collected preoperatively and postoperatively at 6 weeks and at 3, 6, 12, 18, and 24 months by means of interview and clinical examination. Baseline measures included sociodemographic characteristics; risk factors for urinary incontinence, including a high body-mass index, a history of vaginal childbirth, and previous surgery for urinary incontinence; quality of life specific to urinary incontinence17; clinical characteristics of urinary incontinence, including current behavioral or pharmacologic therapy, self-reported urinary-incontinence symptoms on a validated questionnaire distinguishing stress leakage from urge leakage,18 quantity of urine leakage on a pad test,19 and the number of incontinence episodes as recorded in a 3-day voiding diary20; findings on physical examination, including urethral hypermobility as measured by the Q-tip test21 and pelvic-organ prolapse22; and urodynamic evaluation, including the presence of urodynamic stress incontinence and detrusor-overactivity incontinence.
The principal investigator at each site reported adverse events to the adverse-events committee, which comprised four investigators who were unaware of site-specific information. In certain cases, the descriptive details of the adverse event may have made it possible to discern the randomized surgical procedure. All adverse events were assigned a severity code according to a modified version of the classification system developed by Dindo and colleagues.23 This system, which has been validated for use among surgical patients, classifies the severity of an event into one of four levels on the basis of the clinical measures taken to treat that event.
Postoperative urge incontinence was defined as treatment of clinically diagnosed new-onset or persistent urge incontinence after the 6-week follow-up visit. Adequacy of voiding was assessed and categorized dichotomously at hospital discharge and again 6 weeks after surgery. Voiding dysfunction was defined by the need for surgical revision to facilitate bladder emptying or the use of any type of catheter after the 6-week visit.
Patient satisfaction was assessed at 24 months with the question "How satisfied or dissatisfied are you with the result of bladder surgery related to urine leakage?" Patients rated their overall satisfaction, choosing one of five options that ranged from "completely satisfied" to "completely dissatisfied." Patients who answered that they were either "completely satisfied" or "mostly satisfied" were classified as being satisfied with the outcome.
Statistical Analysis
We calculated that 260 women per group would provide a power of 80% to detect a 12% difference between study groups (60% vs. 72%) with the use of a two-sided alternative hypothesis at a significance level of 5%. To allow for attrition and missed visits, we recruited a total of 655 women. Treatment success was assessed at follow-up visits every 6 months. If a treatment failed between scheduled visits, it was considered to have failed at the next visit. Data for women whose treatment was not known to have failed but who had not completed all assessments at the 24-month visit were censored at the last visit on which all failure assessments were complete.
For both outcome measures, we compared the success rates in the two groups at 24 months with the use of time-to-event methods for interval censored data.24 We used Kaplan–Meier product-limit analysis to estimate the success rates at 24 months in the two groups and compared the treatment-failure distributions in the two groups, controlling for stratification by clinical site, with the use of the log-rank test. To determine whether concomitant surgery might have had an effect on the results, we tested the interaction between treatment group and concomitant surgery with the use of the Weibull accelerated failure-time model. All analyses were carried out with SAS statistical software, version 9.2 (SAS Institute).
Results
Patients
From February 2002 to June 2004, we screened 2405 women for trial eligibility (Figure 2). Of these women, 556 were ineligible, 1193 declined to participate or withdrew consent, and 1 died before randomization. A total of 655 women were randomly assigned to a study procedure: 326 to undergo the sling procedure and 329 to undergo the Burch procedure. One woman did not undergo the assigned treatment (Burch procedure), and four women were found to be ineligible after randomization (one assigned to the sling procedure and three assigned to the Burch procedure). A total of 520 women (79%) — 265 in the sling group (81%) and 255 in the Burch group (78%) — either were assessed for treatment success at the 24-month visit or were deemed to have had a treatment failure before that visit.
|
|
|
|
|
Treatment-satisfaction rates for the 480 subjects who answered the satisfaction question at 24 months were significantly higher in the sling group than in the Burch group (86% vs. 78%, P=0.02).
Discussion
At 24 months, the pubovaginal fascial sling had significantly higher rates of success — both overall and specific to stress incontinence — than did the Burch colposuspension in women with predominant stress incontinence. These findings were not modified by performance of concomitant surgery for pelvic-organ prolapse. In addition, the frequency of surgical retreatment for stress incontinence was greater in the Burch group than in the sling group. Success rates declined steadily over the 2-year follow-up period, which confirmed previous observations25,26 and underscored the need for long-term follow-up in these patients.
However, the higher success rates in the sling group were offset by higher rates of urinary tract infection, urge incontinence, voiding dysfunction, and the need for surgical revision to improve voiding. The increased efficacy and greater morbidity of the sling procedure confirm and quantify the results of previous systematic reviews27,28,29 and may explain some of the reluctance in the past to use this procedure as a primary surgical treatment for stress incontinence.14
Our large, randomized surgical trial comparing the fascial-sling procedure with the Burch procedure had a robust 24-month follow-up with the use of standardized definitions, procedures, and methods of evaluation to assess a variety of outcome measures and comprehensive postoperative morbidity. The absence of such information to date has precluded rigorous assessment of surgical outcomes for this condition.30,31 Reported success rates of surgery have varied widely.27,28 Factors contributing to this variation have included the lack of standardized outcome measures, differences in the baseline characteristics of the study populations, and the length of follow-up.32,33
Success rates that are based on reporting by patients are consistently lower than those based on physician-reported measures.34,35 Current research guidelines emphasize the importance of evaluating treatment efficacy with composite outcome measures that include both subjective and objective efficacy measures as well as an assessment of morbidity.36,37,38 Success rates in our trial were low, as compared with those in previous studies.9,10 This finding may be related to our use of composite outcome measures, resulting in a stricter definition of success. The substantial variation in failure rates among studies using single-component measures supports the use of composite outcome measures32 and highlights the lack of concordance among several traditional measures.
Our finding that the two procedures had similar success rates as measured by pad tests and voiding diaries may reflect the higher number of patients with symptoms of urge incontinence in the sling group, since these two measures cannot differentiate stress incontinence from urge incontinence. It is likely that we underestimated the rate of postoperative urge incontinence, since our definition was restricted to patients who received treatment for this condition. This factor may explain in part why only 3% of the patients in our trial had new-onset urge incontinence, a rate that is at the low end of the range reported by others.29,39
The higher rate of urinary tract infections reported in the sling group, as compared with the Burch group, may be related to a delayed return to adequate voiding and a prolonged need for catheterization in the sling group. Our definition of urinary tract infection did not require a positive urine culture, and it is possible that some patients received empirical antibiotic therapy for symptoms alone, leading us to overestimate the true incidence of postoperative urinary tract infection in either or both groups. For instance, the higher rate of urge incontinence identified in the sling group may have led to more false diagnoses of urinary tract infection in that group.
All the patients in our study received care in tertiary care centers, and the study population included only women with urethral hypermobility and pure or stress-predominant incontinence. Whether the results can be generalized to other groups of women is uncertain. Both the patients and the health care providers were aware of study-group assignments, and it is possible that bias affected the measurement of some outcomes.
Just over half the women underwent concomitant surgery for pelvic-organ prolapse, a proportion consistent with that in other studies.8 Although we did not find any material differences in success rates on the basis of concomitant surgery, such procedures could potentially influence the number of adverse events identified in both groups.
The sling group also had higher satisfaction rates than did the Burch group, a difference that was consistent with the success rates. However, satisfaction rates were higher in both groups than were success rates, indicating that satisfaction was influenced by factors beyond the resolution of incontinence symptoms. Further analyses are needed to assess the relationships among the satisfaction reported by patients, improvement in the quality of life, and outcome measures described here.
New surgical procedures for stress incontinence continue to be introduced into clinical practice without evaluation of their efficacy and safety in well-designed, randomized clinical trials.27,28 There has been a recent transition from the fascial sling and Burch procedure to the newer midurethral synthetic sling. A previous randomized surgical trial comparing the midurethral sling with the Burch procedure showed similar efficacy of the two procedures,32,40 although that study has been criticized for being underpowered. No randomized trials have compared the midurethral sling with the autologous fascial sling. The relative frequency with which these procedures are performed in the United States is difficult to assess because they have identical procedural codes. Rigorous comparative trials are needed to assess the efficacy and safety of these novel surgical techniques as compared with the efficacy and safety of the procedures studied in our trial.
The number of women undergoing surgical therapy for stress incontinence is increasing, and this trend is likely to continue as the population ages. Our data show that the pubovaginal fascial sling has significantly higher efficacy than the Burch abdominal colposuspension at 24 months in women with predominant stress incontinence, but such success comes at the cost of more complications. Clinicians should discuss such trade-offs when making recommendations to patients regarding the optimal procedure and should emphasize that complete resolution of incontinence symptoms after surgery is unlikely.
Supported by cooperative agreements (U01 DK58225, U01 DK58229, U01 DK58234, U01 DK58231, U01 DK60379, U01 DK60380, U01 DK60393, U01 DK60395, U01 DK60397, and U01 DK60401) with the National Institute of Diabetes and Digestive and Kidney Diseases and by the National Institute of Child Health and Human Development and Office of Research in Women's Health of the National Institutes of Health.
Dr. Albo reports receiving consulting fees from Pfizer and Astellas and grant support from Pfizer; Dr. Richter, receiving lecture fees from Pfizer, Esprit Pharmaceutical, and Indevus Pharmaceuticals and grant support from Pfizer; Dr. Brubaker, receiving consulting fees from Pfizer and Q-Med and grant support from Q-Med, Allergan, and Pfizer; Dr. Norton, receiving consulting fees from Eli Lilly; Dr. Kraus, receiving consulting fees from Lilly ICOS and Pfizer, lecture fees from Novartis, Astellas, Pfizer, and Ortho-McNeil, and grant support from Pfizer and GlaxoSmithKline; Dr. Chai, receiving consulting fees from Astellas and Pfizer and grant support from Pfizer; Dr. Zyczynski, receiving consulting fees from Ethicon and grant support from Novartis and Johnson & Johnson; Dr. Diokno, receiving lecture fees from Astellas, Medtronic, Janssen-Cilag, and Ortho-McNeil, and grant support from GlaxoSmithKline, Allergon, and Allergan, and holding a patent on a modified vaginal speculum marketed by Mentor; Dr. Tennstedt, receiving grant support from Lilly ICOS; Dr. Lloyd, receiving consulting fees from Novartis, GlaxoSmithKline, Pfizer, and AMS, lecture fees from Lilly ICOS, Ortho-McNeil, Bayer, and Boehringer Ingelheim, and grant support from Ortho-McNeil, MediciNova, and Allergan; Dr. FitzGerald, receiving consulting fees from GlaxoSmithKline, lecture fees from Medtronic, and grant support from Pfizer; Dr. Lemack, receiving consulting fees from Pfizer and Allergan and lecture fees from Pfizer and Astellas and having equity interest in Pfizer; Dr. Johnson, receiving lecture fees from Johnson & Johnson; Dr. Stoddard, having equity in Johnson & Johnson, Bristol-Myers Squibb, Elan, and Procter & Gamble; Dr. Menefee, receiving grant support from Boston Scientific and having equity interest in Pfizer; Dr. Varner, receiving lecture fees from Pfizer; Dr. Moalli, receiving lecture fees from Boston Scientific that were contributed to a fellowship research fund; Dr. Steers, receiving consulting fees from Dynogen, Sanofi-Aventis, Astellas, and Allergan, lecture fees from Watson, Astellas, and GlaxoSmithKline Beecham, and grant support from Pfizer and having equity interest in Abbott and Johnson & Johnson. No other potential conflict of interest relevant to this article was reported.
We thank Kathleen Cannon at the New England Research Institutes for her tireless organizational efforts on behalf of the Urinary Incontinence Treatment Network, and Bette Jo Garrett and Patricia Lane at the University of California at San Diego for their editorial assistance.
Source Information
From the University of California, San Diego, San Diego (M.E.A., C.N., S.M.); the University of Alabama at Birmingham, Birmingham (H.E.R., L.K.L., R.E.V.); Loyola University Medical Center, Maywood, IL (L.B., M.F., K.K.); the University of Utah Health Sciences Center, Salt Lake City (P.N.); the University of Texas Health Sciences Center, San Antonio (S.R.K.); the University of Texas Southwestern, Dallas (P.E.Z., G.E.L.); the University of Maryland, Baltimore (T.C.C., H.W.J.); Magee Women's Hospital, University of Pittsburgh, Pittsburgh (H.Z., W.L., P.M.); Beaumont Hospital Medical Center, Royal Oak, MI (A.C.D., L.S.); New England Research Institutes, Watertown, MA (S.T., A.M.S., K.J.D.); Oakwood Hospital, Dearborn, MI (V.M.); the National Institute of Diabetes and Digestive and Kidney Diseases, Bethesda, MD (J.W.K., L.M.N.); and the University of Virginia Health Systems, Charlottesville (W.S.).
This article (10.1056/NEJMoa070416) was published at www.nejm.org on May 21, 2007. It will appear in the May 24 issue of the Journal.
Address reprint requests to Dr. Albo at the Division of Urology, University of California, San Diego Medical Center, 200 W. Arbor Dr., San Diego, CA 92103-8897.
References
The following are members of the Urinary Incontinence Treatment Network: Steering Committee: University of Virginia, Charlottesville — W. Steers (chair); William Beaumont Hospital, Royal Oak, MI, and Oakwood Hospital, Dearborn MI — A. Diokno, S. Khandwala; Loyola University Medical Center, Maywood, IL — L. Brubaker, M. FitzGerald; University of Alabama at Birmingham, Birmingham — H.E. Richter, L.K. Lloyd; University of California, San Diego, San Diego — M. Albo, C. Nager; University of Maryland, Baltimore — T. Chai, H.W. Johnson; University of Pittsburgh, Pittsburgh — H.M. Zyczynski, W. Leng; University of Texas Southwestern, Dallas — P. Zimmern, G. Lemack; University of Texas Health Sciences Center, San Antonio — S. Kraus, T. Rozanski; University of Utah, Salt Lake City — P. Norton, L. Kerr; New England Research Institutes, Watertown, MA — S. Tennstedt, A. Stoddard; National Institute of Diabetes and Digestive and Kidney Diseases, Bethesda, MD — D. Chang, J.W. Kusek, L.M. Nyberg; National Institute of Child Health and Human Development, Bethesda, MD — A.M. Weber. Coinvestigators: R.S. Ashford II, J. Baker, D. Borello-France, K.L. Burgio, S. Chiang, A. Dabbous, P.S. Goode, L.N. Hammontree, K. Kenton, D. Lesser, K. Luber, E. Lukacz, A. Markland, S. Menefee, P. Moalli, K. Peters, E. Sagan, J. Schaffer, A. Simsiman, L. Sirls, R. Starr, R.E. Varner. Study Coordinators: R. Bradt, K. Debes, R. Dinh, J. Gruss, L. Hall, A. Howell, K. Jesse, D.L. Kalinoski, K. Koches, B. Leemon, K. Mislanovich, S. O'Meara, J. Parent, N. Pope, C. Prather, T. Rogers, S. Sluder, M. Tulke. Biostatistical Coordinating Center: K.J. Dandreo, C.J. Leifer, S. McDermott, A. Stoddard (co-principal investigator), S. Tennstedt (principal investigator), L. Tinsley, L. Wruck, Y. Xu. Data and Safety Monitoring Board: Dartmouth-Hitchcock Medical Center, Lebanon, NH — E.A. Gormley (chair); Bristol Urological Institute, Bristol, United Kingdom — P. Abrams; Blue Ridge Medical Associates, Winston-Salem, NC — D. Bland; Northwestern University Medical School, Chicago — J.Q. Clemens; University of Minnesota, Minneapolis — J. Connett; University of Colorado, Aurora — W. Henderson; University of Michigan, Ann Arbor — D. Fenner; University of Pittsburgh, Pittsburgh — S. Kelsey; Brown University School of Medicine, Providence, RI — D. Myers; Johns Hopkins Hospital, Baltimore — J. Mostwin; and Mansoura Urology and Nephrology Center, Mansoura, Egypt — B. Wadie.
| |||||||||||||||||||||||||||||||||
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | TERMS OF USE | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2009 Massachusetts Medical Society. All rights reserved. |
