- Email: [email protected]
Epidemiologic evaluation of reoperation for surgically treated pelvic organ prolapse and urinary incontinence
Epidemiologic evaluation of reoperation for surgically treated pelvic organ prolapse and urinary incontinence Amanda L. Clark, MD, Thomas Gregory, MD, Virginia J. Smith, MD, and Renee Edwards, MD Portland, Ore OBJECTIVE: The purpose of this study was to measure the risk of reoperation for surgically treated pelvic organ prolapse and urinary incontinence in a community-based population. STUDY DESIGN: A 5-year prospective, observational study was conducted of women who had undergone pelvic organ prolapse and urinary incontinence surgery in 1995. The cohort of 376 women was identified by International Classification of Diseases, 9th revision, and current procedural terminology codes in 149,554 reproductive-aged women within the Kaiser Permanente Northwest membership. RESULTS: Thirty-six women underwent 40 cases of reoperation. By survival analysis, 13% of women underwent reoperation by 71 months. Having undergone previous pelvic organ prolapse and urinary incontinence surgery increased the risk of reoperation to 17% compared with 12% for women who underwent a first procedure (log rank, P = .04). No association was observed with age, body mass index, parity, previous hysterectomy not for prolapse, vaginal versus abdominal approach, severity of prolapse, ethnicity, chronic lung disease, smoking, previous corticosteroid use, and estrogen status. CONCLUSIONS: Future reoperation is a significant risk of morbidity for women who undergo pelvic organ prolapse and urinary incontinence surgery. (Am J Obstet Gynecol 2003;189:1261-7.)
Key words: Prolapse, urinary incontinence, surgery, reoperation
Surgery for pelvic organ prolapse and urinary incontinence (POPUI) in women is common, as shown by an 11.1% lifetime risk of undergoing a first procedure.1 Reoperation is also common, but the risk of reoperation has not been well quantified. The prevalence of reoperation is high (43%-56%)2,3 in the academic referral populations in which most clinical research is conducted. This prevalence is likely to be higher than that in the general population, because the need for reoperation is often the reason for referral to a tertiary care center. Olsen et al1 reported a lower prevalence of reoperation for POPUI (29.2%) in a community-based population. Reoperation for recurrent POPUI implies surgical failure, but the terms are not synonymous. Reoperation is the ‘‘tip of the iceberg’’ of surgical failure, because many women choose conservative therapy after unsatisfactory results. Unfortunately, many women opt to adapt their lives to live with their symptoms. Reoperation results from more than direct surgical failure. Reoperation often From the Division of Urogynecology and Reconstructive Pelvic Surgery, Department of Obstetrics and Gynecology, Oregon Health and Science University and Kaiser Permanente Northwest. Received for publication November 13, 2002; revised March 10, 2003; accepted June 20, 2003. Reprint requests: Amanda L. Clark, MD, Division Chief, Urogynecology and Reconstructive Pelvic Surgery, Associate Professor, Obstetrics and Gynecology, 3181 SW Sam Jackson Park Road, L466, Oregon Health and Science University, Portland, OR 97239. E-mail: [email protected] Ó 2003, Mosby, Inc. All rights reserved. 0002-9378/2003 $30.00 + 0 doi:10.1067/S0002-9378(03)00829-9
occurs because of the phenomenon that successful surgery in 1 anatomic compartment predisposes to worsening of prolapse in another or to new onset of urinary incontinence.2-4 Additionally, risk factors that are associated with the development of POPUI may be prognostic for recurrence and reoperation. Cited risk factors for POPUI include advancing age, white race,5 obesity,6 vaginal delivery,7 estrogen deficiency, activities or conditions that are associated with chronically increased intra-abdominal pressure, smoking,8,9 underlying connective tissue disease, neuropathy, and previous hysterectomy.10 Large sample sizes that are representative of the general population are required to assess these variables. Reported risks of surgical failure for POPUI tend to focus on single procedures or single anatomic sites and are representative of tertiary care referral populations. These studies fail to capture the complexities of reoperation and its burden on women who are affected and the surgeons who care for them. In 1995, we identified a cohort of women who were undergoing POPUI surgery from a community-based population. In this study, we have followed that cohort for 5 years to measure the risk of reoperation for recurrent POPUI and to evaluate risk factors that may predict future reoperation. Material and methods In 1995, a cohort was defined as all women (n = 376) who underwent surgical treatment of POPUI, that was 1261
1262 Clark et al
identified by International Classification of Diseases, 9th revision (ICD9), and current procedural terminology (CPT) codes from hospital discharge summaries at Kaiser Permanente Northwest (KPNW). Eight of the 384 original cases in 1995 underwent reoperation in 1995 and were analyzed as reoperations rather than index cases. Reoperation for POPUI was defined by the same ICD9 and CPT codes in years 1995 through 2000 to define the index cases. Reoperations for surgical complications (such as hemorrhage, bowel obstruction) were not included in this analysis. For all cases identified by ICD9 codes, the operative report of each case was reviewed by the authors to verify the component procedures that comprised each surgery. The dates of membership in the Kaiser health plan were determined for this cohort of 376 women through the years 1995 to 2000. This project was approved for human investigation by the Institutional Review Board. The original population who were at risk for surgically treated POPUI was defined to include all women who were >20 years old and consisted of 149,554 women. Overall, KPNW is a large health maintenance organization with nearly 393,000 members that draws from an estimated service area of >2 million people. The ethnic composition of the membership is 97.3% white, 0.3% African American, 1.3% Asian, and 1.1% Hispanic. KPNW members are well stratified with respect to the socioeconomic indicators of education level and household income. Members of KPNW were chosen as our study group because (1) health histories are recorded in a uniform fashion, (2) medical records are maintained in one centralized location, and (3) membership demographics are representative of the local population. The risk of reoperation was determined by the KaplanMeier procedure for survival analysis. The start date was set at the time of the first surgery (index operation) that occurred in 1995. The event of interest was defined as reoperation. In the original cohort, 70.8% of cases were first procedures, and 29.2% were reoperations. When subjects underwent >1 reoperation during the followup period, each case was re-entered as a new case in the survival analysis to observe further reoperation. The data were censored if the patient dropped out of the health plan or when the study ended in December 2000. In 1995, patient characteristics that were likely to predict the development of POPUI were obtained by a review of both the inpatient and outpatient records. Data accuracy was enhanced by the standardized systems for recording clinical data that are routine in the Kaiser system. Clinical data were obtained from standardized forms that are used for the preanesthetic assessment, the annual health examination, the preoperative history and physical examination, and the operative report. Information on estrogen and corticosteroid use was obtained from the
November 2003 Am J Obstet Gynecol
longitudinal computer database of the centralized Kaiser pharmacy. The variables that were examined included patient age, self-reported race, height, weight, vaginal parity, estrogen status, smoking history, and aspects of the patient medical and surgical history. Chronic lung disease was defined as reactive airway disease that required the use of bronchodilators, chronic bronchitis, or chronic obstructive pulmonary disease. The surgical history that was recorded included the preoperative assessment of pelvic floor support and the operation(s) that had been performed. The descriptive terms that were used by the operating gynecologists and urologists to document pelvic floor support for each vaginal compartment were assigned a prolapse severity grade of 0, 1+, 2+, and 3+. In patients with a history of previous pelvic surgery, the medical records were reviewed for the details of these operations. The surgeon who performed the operation was recorded in 56% of the cases. The operations were classified according to the compartment or compartments that were addressed in the surgery: anterior, posterior, and apical. Operations were also classified according to the condition for which the patient underwent surgery: pelvic organ prolapse alone, urinary incontinence alone, or both. Colposuspension, needle procedures, and slings were classified as operations for stress incontinence. Anterior colporrhaphy and paravaginal repair were designated as operations for prolapse, with the recognition that some surgeons perform these procedures for incontinence. Diagnostic criteria for the condition of stress incontinence were not recorded systematically and were not available for classification in this analysis. The analysis of risk factors was managed and analyzed with the use of computer programs that included Filemaker Pro (Filemaker, Inc, Santa Clara, Calif), Excel (Microsoft Corporation, Bellevue, Wash), and Statistical Package for the Social Sciences (SPSS Inc, Chicago, Ill). Cox regression and the Fisher exact test were used to determine statistical significance of the clinical characteristics in the comparison of women who underwent reoperations to women who did not. A post hoc power analysis was performed with the Web-based SSP statistical calculator for survival analysis from the John Hopkins Center for Clinical Research at: http://iws1.jhsph.edu/ Research/Centers/CCT/javamarc/
Results Fig 1 is a diagram that describes the timeline and outcomes of the study. The original cohort is shown, and for patients who had undergone previous POPUI surgery at the initiation of the cohort, the number of previous operations that they experienced is shown. The number of reoperations undergone by the cohort in the 5-year follow-up period is given. In total, 36 women underwent
Volume 189, Number 5 Am J Obstet Gynecol
Clark et al 1263
Fig 1. The index cases that were performed in 1995 are plotted on a time line to show both the antecedent operations and the need for reoperation as the reoperation relates to previous POPUI surgeries.
40 reoperations. Eight women had reoperation in the index year of 1995, and 32 women had reoperation in the subsequent 5 years. Two women underwent 2 reoperations, and 1 woman underwent 3 reoperations in the follow-up period. The survival of POPUI surgeries that are measured by reoperation is plotted against time to reoperation (months) in Fig 2. The results are stratified according to whether the index surgery was a first POPUI surgery or whether the woman had undergone previous POPUI surgery. At 71 months, the overall survival was 87% ± 2.7%, or conversely, the proportion of reoperation was 13%. When stratified for previous POPUI surgery, 12% ± 4.1% of the women who had their first POPUI procedure (n = 264 women) had undergone reoperation as compared with 17% ± 3.5% of women who had undergone previous POPUI surgery (n = 112 women; P = .04, log rank). At the end of 71 months, 300 women (78%) of the original cohort were still enrolled in the Kaiser health plan. The characteristics of the women who were lost to follow-up were compared with those women who remained in the cohort. Those women who were lost to follow-up were slightly younger (54 vs 58 years, P = .014), had a higher proportion with chronic lung disease (odds ratio, 2.5 [95% CI, 1.4-4.2], P = .001), and were more likely to have a smoking history (odds ratio, 1.67 [95% CI, 1.04-2.68], P = .034). All other characteristics were equivalent. Subject demographics and analysis of potential risk factors for the women in this cohort are shown in Table I. A previous operation for POPUI predicted a 1.9-fold increase in relative hazard (hazard ratio, 1.9, P = .04) of reoperation, but previous hysterectomy not for prolapse was not associated with an increased risk of reoperation. Those variables in the univariate analysis with probability values of < 0.25 were entered into a multiple Cox regression model and were found not to influence the
Fig 2. The survival of a POPUI surgery that was performed in 1995 is measured by the outcome of reoperation, stratified into those women who had a first POPUI surgery (open circles) in 1995 and those women who had a POPUI surgery before 1995 (closed squares, P < .05, log rank test).
relative hazard for previous POPUI surgery. Other variables were not found to be statistically associated with future reoperation. A post hoc power analysis was performed for this cohort. Assuming a power of 0.8, the study has the power to detect a 4-fold increase in risk that is associated with parity, an 8-fold risk with previous hysterectomy, a 6-fold risk reduction in risk with estrogen use, a 5-fold reduction in risk with smoking, and a 14-fold reduction in risk with chronic lung disease. The power to detect no difference varied from 0.07 to 0.63. A sample size of approximately 15,000 women would be required to provide a power of 0.8 to detect no difference for all variables in this study. A complete list of the 40 cases of reoperation and the preceding operations are listed in Table II. Every occurrence of reoperation is unique, except one. In Table II, the cases are grouped according to whether the reoperation occurred in the same anatomic compartments that were addressed in the preceding surgery (n = 24 cases, 60%), in a different compartment than in the preceding surgery (n = 13 cases, 32.5%), or both (n = 3 cases, 7.5%). In most cases, the number of compartments that was addressed is fewer in the reoperation than in the preceding operation. To evaluate patterns of reoperation, the cases have been categorized according to anatomic compartment or the condition that indicated surgery (Table III). The percentage of cases that were followed by any reoperation is shown. By compartment, reoperation was more likely when the preceding operation involved repair of the apex. By condition, the likelihood of reoperation was similar for all three groups.
1264 Clark et al
November 2003 Am J Obstet Gynecol
Table I. Univariate analysis of risk factors for reoperation for POPUI Demographic Analysis by Cox regression Mean age (y) Mean body mass index (kg/m2) Vaginal parity > 1 (%) Previous surgery for POPUI Hysterectomy not for prolapse (%) Worst grade of prolapse (%) 0 1+ 2+ 3+ Current estrogen use (%) Ever smoked (%) History of chronic lung disease (%)
Relative hazard (95% CI) 58.0 27.9 96 29.2 25.2 4.7 18.1 56.7 20.4 80.5 41.1 21.8
0.99 1.02 1.47 1.90 0.73
(0.96-1.02) (0.96-1.08) (0.20-10.7) (1.02-3.25) (0.32-1.70)
.48 .55 .70 .04 .48
1.45 (0.85-2.48)
.17
0.79 (0.37-1.69) 0.33 (0.08-1.38) 0.51 (0.18-1.45)
.21 .13 .54
Analysis by Fisher exact test (when any cell size = 0)
Reoperation status Yes No
Race (n) White Other History of corticosteroid use (n) Yes No
Table IV lists the component procedures within the complete operations of the index. The percentage of cases that resulted in reoperation in the same site and for the same condition were calculated. For example, for the abdominal retropubic urethropexy, the percentage of reoperations that included any incontinence component procedure was calculated. For vaginal hysterectomy for prolapse, the percentage that is shown is of any subsequent operation to repair the apex. The numbers are too small to report as the true risk of reoperation, but they allow a comparison of risk in this population to risks of failure and reoperation in published series. No single component predominates. The global reoperation risk comprises low risks from multiple procedures. The proportion of commonly reported ‘‘other site, other condition’’ reoperations was also calculated. Reoperation for urinary incontinence after a pelvic organ prolapse procedure (implying latent stress incontinence) occurred in 7.5% of cases. After abdominal retropubic urethropexies (RPUs), the following observations were made. Of 132 abdominal RPUs that were performed, 5% of the patients underwent reoperation at the vaginal apex. In 103 cases in which no apical repair was performed, 3% of the patients subsequently required apical suspension. In the 29 cases in which an apical procedure was performed concomitantly with the abdominal RPU, 4% of the cases required resuspension of the apex. The posterior compartment required reoperation in 4% of the 132 abdominal RPU cases; 79 of the 132 cases did not have a concomitant posterior repair. Of those 79 cases, 1% required a subsequent posterior colporrhaphy. Of the 43 cases in which a posterior colporrhaphy was performed at
P value
238 10
25 0
.61
6 241
0 25
.58
the same time as the abdominal RPU, 9% of the patients underwent a repeat posterior colporrhaphy. Of the 214 cases in which the primary surgeon was recorded, 63 surgeons were represented. Most of the surgeons were gynecologists. Urologists and colorectal surgeons performed or assisted with a small number of cases. None of the surgeons were fellowship trained in urogynecology or female pelvic floor medicine. Seventynine procedures (37%) were performed by 5 gynecologists with a special interest in urogynecology who performed at least 10 procedures per year. One hundred thirty-five procedures (63%) were performed by 58 surgeons who performed < 10 procedures per year. Twenty-six surgeons performed only 1 POPUI procedure in 1995. An analysis of reoperation by type of surgeon or surgeon’s experience was precluded by the small sample size. Comment Not only do women have a high lifetime risk of undergoing POPUI surgery, but also they face a high risk of undergoing reoperation for POPUI surgery. We expect that additional cases will come to reoperation as time passes. In the retrospective analysis of this cohort, the time interval between a first POPUI procedure and a second procedure averaged 12.5 years. Thus, our report likely represents only an initial snapshot of the total reoperations that this cohort will experience ultimately. The accuracy of our reported reoperation risk is limited by the 22% of women who were lost to follow-up, but the Kaplan Meier method compensates for this, which results in an SE of 2.7%. The women who were lost to follow-up
Clark et al 1265
Volume 189, Number 5 Am J Obstet Gynecol
Table II. List of reoperations Index operation Reoperation at same anatomic site (n = 24; 60.0%) Vaginal enterocele Abdominal RPU, PR Needle RPU TVH for POP, AR, PR, vaginal enterocele Needle RPU, AR, PR Needle RPU Needle RPU Needle RPU TVH for POP TVH for POP TAH for POP PR, AR Abdominal RPU, paravaginal, PR Abdominal RPU, PR AR, Vaginal enterocele, urethrolysis TAH for POP, abdominal RPU TVH for POP, abdominal RPU, SSLF, vaginal enterocele AR with mesh, vaginal enterocele Vaginal enterocele, PR TVH for POP, AR, PR, vaginal enterocele TVH for POP, AR, PR Abdominal RPU, PR, SSLF TVH for POP, abdominal RPU, PR Subtotal TAH, abdominal RPU, AR, PR, abdominal enterocele Reoperation at different anatomic site n = 13 (32.5%) Abdominal RPU TAH no POP, abdominal RPU Abdominal RPU Needle RPU Needle RPU TVH no POP, PR PR PR TVH for POP TVH for POP TVH for POP, SSLF TVH for POP PR, Vaginal enterocele Reoperation at both same and different site n = 3 (7.5%) Abdominal RPU Needle RPU TAH for POP, AR
Reoperation Abdominal enterocele PR Sling Vaginal enterocele, PR Paravaginal, abdominal RPU Sling Abdominal RPU, ureteroneostomy Needle RPU Sacrocolpopexy Sacrocolpopexy Sacrocolpopexy, abdominal enterocele Sling, AR PR Sling, PR AR with mesh excision, urethrolysis, Martius graft Sacrocolpopexy AR AR, vaginal enterocele, urethrolysis Sacrocolpopexy Abdominal RPU Sacrocolpopexy, abdominal RPU PR, vaginal enterocele, PR Vaginal enterocele Vaginal enterocele, AR PR Vaginal enterocele TAH for POP TAH for POP PR, Vaginal enterocele AR TAH for POP, AR TAH for POP, abdominal RPU Sling Needle RPU, AR AR, PR AR, needle RPU Needle RPU Vaginal enterocele, AR with mesh Sling, PR Vaginal enterocele, AR, PR
PR, Posterior repair; TVH, total vaginal hysterectomy; AR, anterior repair; TAH, total abdominal hysterectomy; POP, pelvic organ prolapse; SSLF, sacrospinous ligament fixation.
differed slightly from the women who remained to the end of the study period. Specifically, chronic lung disease and history of smoking were more prevalent in those women who left the cohort compared with those women who remained. If the prevailing sentiment is that chronic lung disease and smoking predispose to the recurrence of POPUI, then our reported risk may underestimate the true risk of reoperation. The accuracy of this analysis is also dependent on the accuracy of medical coding of procedures and the accuracy of data entry. Error rates for coding are estimated to be < 5% for medical coding and < 0.5% for nonmedical data entry.11 The generalizability of this analysis is influenced by the study population and the surgeons and their criteria for choosing surgical intervention. The population is a community-based managed care population that is well
stratified in terms of socioeconomics. This cohort is predominantly white, which limits our ability to detect differences according to ethnicity. There is no standardized approach for making the decision to perform POPUI surgery at KPNW. A diversity of surgical decision making is represented by at least 63 surgeons, who are primarily general obstetrician/gynecologists of varying backgrounds and experience. The choice of surgical intervention may also be influenced by variables such as access to medical care and financial considerations for both patients and health care providers. These variables are moderated in this mature managed-care environment, in which there are neither financial incentives nor disincentives to receiving surgical care. Having failed a POPUI surgery was the only risk factor that was identified in this population. Other variables that
1266 Clark et al
November 2003 Am J Obstet Gynecol
Table III. Reoperation by condition and compartment 1995 Index cases (n)
Reopertion (%)*
154 137 89
13 8 10
154 28 24 58 33 17 69
8 11 33 7 15 12 9
By condition Pelvic organ prolapse Urinary incontinence Both By compartment Anterior Posterior Apex Anterior and posterior Anterior and apex Posterior and apex All *Any condition or in any compartment.
surgeons commonly believe to be risk factors for surgical failure did not show significance in this study. For the variable of smoking, there was a trend toward decreasing the risk of reoperation, and this is contrary to common expectations. The observational design of this study may have precluded the detection of an effect from smoking and other potential risk factors because surgeons may have selected procedures differently on the basis of a clinical assessment of risk. In addition, more time may be required to see an effect on reoperation from the same risk factors that predict the development of POPUI. Having a failed POPUI surgery is likely a marker for a risk factor that has not yet been measured. The loss of support that occurs after neuromuscular damage in the pelvic floor may allow normal intra-abdominal forces to exceed the capacity of endogenous connective tissues.12,13 Currently, there is no accurate clinical tool for the assessment of neuromuscular function, which may be a very important predictor of failure that has never been well studied. Intrinsic weakness of the connective tissue, from genetic variation, is often posited a risk factor for the development of POPUI, which is evidenced by a differing risk with family history of prolapse and by ethnicity. Information on family history of prolapse was not available in this data set. Reoperation implies failure of the initial procedure but does not measure that failure exactly. Many women who are not treated successfully by surgery opt for conservative therapy. Reoperation represents the most severe occurrences of a failed procedure and includes surgical correction of adverse effects of the original procedure. Pelvic floor disorders result from structural failure of the muscles and connective tissue that comprise the pelvic floor. Defects are presumed to result from the elongation or breakage of the connective tissue and muscle and/or nerve injury that interferes with muscle function. Surgery aims to repair endogenous connective tissue or insert replacement materials to correct abnormal connective tissue and/or to compensate for abnormal neuromuscu-
Table IV. Reoperation specific to condition and compartment
Urinary incontinence Abdominal RPU Needle RPU Sling Anterior prolapse Anterior colporrhaphy Paravaginal repair Posterior prolapse: Posterior colporrhaphy Apical prolapse Total vaginal hysterectomy for pelvic organ prolapse Total abdominal hysterectomy for pelvic organ prolapse Sacrocolpopexy Sacrospinous Ligament Fixation Vaginal enterocele repair Abdominal enterocele repair Colpocleisis
1995 Index cases (n)
Reoperation (%)*
132 80 6
1 8 0
112 13 169
7 0 3
70
7
12
25
13
0
17 50 10 3
18 8 10 0
*Performed for same condition and in same compartment.
lar function. Direct failure occurs when the repaired tissue or grafted material is unable to maintain normal anatomy. Most reoperations (60%) occurred at the same anatomic site, which implies direct failure of the component surgical procedure. The site-specific reoperation risks in this cohort are comparable or lower to that reported in the literature. For sacrocolpopexy, Benson et al4 reported a 2.6% reoperation risk at the apex, compared with 0% in this study. For sacrospinous ligament fixation, Benson et al reported a 12% reoperation risk at the apex, compared with 17.5% in this group. No data were found for sitespecific reoperation risks for incontinence procedures, but the reoperation risk in this study (abdominal RPU, 0.8%; needle RPU, 8.3%; sling, 0%) was much lower compared with the reported 5-year surgical failure rates (abdominal RPU, 6%; needle RPU, 33%; and sling, 7%).14 The lower risk of site-specific reoperation in this community-based sample compared with academic samples supports the previous observation that the overall risk of reoperation is lower in community-based populations than in referral populations. An indirect failure occurs when a corrective procedure in one compartment predisposes to new failure in another compartment (for example, apical prolapse after retropubic urethropexy2 and posterior wall prolapse after retropubic urethropexy3). An alternate cause of apparent indirect failure results from the failure to identify and correct site-specific defects in all anatomic compartments. A smaller but significant proportion of reoperations (32.5%) occurred at different sites than in the preceding surgery. Apical and posterior compartment defects have been reported after abdominal retropubic urethropexy.
Volume 189, Number 5 Am J Obstet Gynecol
Wiskind et al2 reported a 29% risk of reoperation for apical defects after Burch colposuspension, which compares to a 3% risk in this study. Kjolhede et al3 reported a 29% risk of reoperation for rectocele after Burch colposuspension, which compares to a 1% risk in our study. The risks of other site-specific reoperation in community population that was compared are again lower than were reported in referral populations. An ideal surgical treatment for POPUI is a single surgical procedure that restores pelvic floor function with no adverse effects and lasts a lifetime. Clearly, we are far from that ideal. Women and their surgeons justifiably are frustrated by the lack of good clinical outcome data to guide surgeons in selecting operations for POPUI. National support is needed to fund and organize the large and expensive clinical trials that will be needed to address the complexities of outcome in pelvic floor surgery. Fortunately, this work has just begun. It is the responsibility of pelvic floor surgeons to keep this issue as a priority on the national research agenda for women’s health. REFERENCES 1. Olsen AL, Smith VJ, Bergstrom JO, Colling JC, Clark AL. Epidemiology of surgically managed pelvic organ prolapse and urinary incontinence. Obstet Gynecol 1997;89:501-6. 2. Wiskind AK, Creighton SM, Stanton SL. The incidence of genital prolapse after the Burch colposuspension. Am J Obstet Gynecol 1992;167:399-405.
Clark et al 1267
3. Kjolhede P, Noren B, Ryden G. Prediction of genital prolapse after Burch colposuspension. Acta Obstet Gynecol Scand 1996;75:849-54. 4. Benson JT, Lucente V, McClellan E. Vaginal versus abdominal reconstructive surgery for the treatment of pelvic support defects: a prospective randomized study with long-term outcome evaluation. Am J Obstet Gynecol 1996;175:1418-22. 5. Bump RC. Racial comparisons and contrasts in urinary incontinence and pelvic organ prolapse. Obstet Gynecol 1993;81:421-5. 6. Dwyer PL, Lee ET, Hay DM. Obesity and urinary incontinence in women. BJOG 1988;95:91-6. 7. Foldspang A, Mommsen S, Lam GW, Elving L. Parity as a correlate of adult female urinary incontinence prevalence. J Epidemiol Commun Health 1992;46:595-600. 8. Bump RC, McClish DK. Cigarette smoking and urinary incontinence in women. Am J Obstet Gynecol 1992;167:1213-8. 9. Bump RC, McClish DM. Cigarette smoking and pure genuine stress incontinence of urine: a comparison of risk factors and determinants between smokers and nonsmokers. Am J Obstet Gynecol 1994; 170:579-82. 10. Mant J, Painter R, Vessey M. Epidemiology of genital prolapse: observations from the Oxford Family Planning Association Study. BJOG 1997;104:579-85. 11. McLemore T, Lawrence L. Plan and operation of the National Survey of Ambulatory Surgery. Vital Health Stat 1 19971-4. 1-124. 12. Smith AR, Hosker GL, Warrell DW. The role of partial denervation of the pelvic floor in the aetiology of genitourinary prolapse and stress incontinence of urine: a neurophysiological study. BJOG 1989; 96:24-8. 13. Smith AR, Hosker GL, Warrell DW. The role of pudendal nerve damage in the aetiology of genuine stress incontinence in women. BJOG 1989;96:29-32. 14. Leach GE, Dmochowski RR, Appell RA, Blaivas JG, Hadley HR, Luber KM, et al. Female Stress Urinary Incontinence Clinical Guidelines Panel summary report on surgical management of female stress urinary incontinence: the American Urological Association. J Urol 1997;158:875-80.
Key words: Prolapse, urinary incontinence, surgery, reoperation
Surgery for pelvic organ prolapse and urinary incontinence (POPUI) in women is common, as shown by an 11.1% lifetime risk of undergoing a first procedure.1 Reoperation is also common, but the risk of reoperation has not been well quantified. The prevalence of reoperation is high (43%-56%)2,3 in the academic referral populations in which most clinical research is conducted. This prevalence is likely to be higher than that in the general population, because the need for reoperation is often the reason for referral to a tertiary care center. Olsen et al1 reported a lower prevalence of reoperation for POPUI (29.2%) in a community-based population. Reoperation for recurrent POPUI implies surgical failure, but the terms are not synonymous. Reoperation is the ‘‘tip of the iceberg’’ of surgical failure, because many women choose conservative therapy after unsatisfactory results. Unfortunately, many women opt to adapt their lives to live with their symptoms. Reoperation results from more than direct surgical failure. Reoperation often From the Division of Urogynecology and Reconstructive Pelvic Surgery, Department of Obstetrics and Gynecology, Oregon Health and Science University and Kaiser Permanente Northwest. Received for publication November 13, 2002; revised March 10, 2003; accepted June 20, 2003. Reprint requests: Amanda L. Clark, MD, Division Chief, Urogynecology and Reconstructive Pelvic Surgery, Associate Professor, Obstetrics and Gynecology, 3181 SW Sam Jackson Park Road, L466, Oregon Health and Science University, Portland, OR 97239. E-mail: [email protected] Ó 2003, Mosby, Inc. All rights reserved. 0002-9378/2003 $30.00 + 0 doi:10.1067/S0002-9378(03)00829-9
occurs because of the phenomenon that successful surgery in 1 anatomic compartment predisposes to worsening of prolapse in another or to new onset of urinary incontinence.2-4 Additionally, risk factors that are associated with the development of POPUI may be prognostic for recurrence and reoperation. Cited risk factors for POPUI include advancing age, white race,5 obesity,6 vaginal delivery,7 estrogen deficiency, activities or conditions that are associated with chronically increased intra-abdominal pressure, smoking,8,9 underlying connective tissue disease, neuropathy, and previous hysterectomy.10 Large sample sizes that are representative of the general population are required to assess these variables. Reported risks of surgical failure for POPUI tend to focus on single procedures or single anatomic sites and are representative of tertiary care referral populations. These studies fail to capture the complexities of reoperation and its burden on women who are affected and the surgeons who care for them. In 1995, we identified a cohort of women who were undergoing POPUI surgery from a community-based population. In this study, we have followed that cohort for 5 years to measure the risk of reoperation for recurrent POPUI and to evaluate risk factors that may predict future reoperation. Material and methods In 1995, a cohort was defined as all women (n = 376) who underwent surgical treatment of POPUI, that was 1261
1262 Clark et al
identified by International Classification of Diseases, 9th revision (ICD9), and current procedural terminology (CPT) codes from hospital discharge summaries at Kaiser Permanente Northwest (KPNW). Eight of the 384 original cases in 1995 underwent reoperation in 1995 and were analyzed as reoperations rather than index cases. Reoperation for POPUI was defined by the same ICD9 and CPT codes in years 1995 through 2000 to define the index cases. Reoperations for surgical complications (such as hemorrhage, bowel obstruction) were not included in this analysis. For all cases identified by ICD9 codes, the operative report of each case was reviewed by the authors to verify the component procedures that comprised each surgery. The dates of membership in the Kaiser health plan were determined for this cohort of 376 women through the years 1995 to 2000. This project was approved for human investigation by the Institutional Review Board. The original population who were at risk for surgically treated POPUI was defined to include all women who were >20 years old and consisted of 149,554 women. Overall, KPNW is a large health maintenance organization with nearly 393,000 members that draws from an estimated service area of >2 million people. The ethnic composition of the membership is 97.3% white, 0.3% African American, 1.3% Asian, and 1.1% Hispanic. KPNW members are well stratified with respect to the socioeconomic indicators of education level and household income. Members of KPNW were chosen as our study group because (1) health histories are recorded in a uniform fashion, (2) medical records are maintained in one centralized location, and (3) membership demographics are representative of the local population. The risk of reoperation was determined by the KaplanMeier procedure for survival analysis. The start date was set at the time of the first surgery (index operation) that occurred in 1995. The event of interest was defined as reoperation. In the original cohort, 70.8% of cases were first procedures, and 29.2% were reoperations. When subjects underwent >1 reoperation during the followup period, each case was re-entered as a new case in the survival analysis to observe further reoperation. The data were censored if the patient dropped out of the health plan or when the study ended in December 2000. In 1995, patient characteristics that were likely to predict the development of POPUI were obtained by a review of both the inpatient and outpatient records. Data accuracy was enhanced by the standardized systems for recording clinical data that are routine in the Kaiser system. Clinical data were obtained from standardized forms that are used for the preanesthetic assessment, the annual health examination, the preoperative history and physical examination, and the operative report. Information on estrogen and corticosteroid use was obtained from the
November 2003 Am J Obstet Gynecol
longitudinal computer database of the centralized Kaiser pharmacy. The variables that were examined included patient age, self-reported race, height, weight, vaginal parity, estrogen status, smoking history, and aspects of the patient medical and surgical history. Chronic lung disease was defined as reactive airway disease that required the use of bronchodilators, chronic bronchitis, or chronic obstructive pulmonary disease. The surgical history that was recorded included the preoperative assessment of pelvic floor support and the operation(s) that had been performed. The descriptive terms that were used by the operating gynecologists and urologists to document pelvic floor support for each vaginal compartment were assigned a prolapse severity grade of 0, 1+, 2+, and 3+. In patients with a history of previous pelvic surgery, the medical records were reviewed for the details of these operations. The surgeon who performed the operation was recorded in 56% of the cases. The operations were classified according to the compartment or compartments that were addressed in the surgery: anterior, posterior, and apical. Operations were also classified according to the condition for which the patient underwent surgery: pelvic organ prolapse alone, urinary incontinence alone, or both. Colposuspension, needle procedures, and slings were classified as operations for stress incontinence. Anterior colporrhaphy and paravaginal repair were designated as operations for prolapse, with the recognition that some surgeons perform these procedures for incontinence. Diagnostic criteria for the condition of stress incontinence were not recorded systematically and were not available for classification in this analysis. The analysis of risk factors was managed and analyzed with the use of computer programs that included Filemaker Pro (Filemaker, Inc, Santa Clara, Calif), Excel (Microsoft Corporation, Bellevue, Wash), and Statistical Package for the Social Sciences (SPSS Inc, Chicago, Ill). Cox regression and the Fisher exact test were used to determine statistical significance of the clinical characteristics in the comparison of women who underwent reoperations to women who did not. A post hoc power analysis was performed with the Web-based SSP statistical calculator for survival analysis from the John Hopkins Center for Clinical Research at: http://iws1.jhsph.edu/ Research/Centers/CCT/javamarc/
Results Fig 1 is a diagram that describes the timeline and outcomes of the study. The original cohort is shown, and for patients who had undergone previous POPUI surgery at the initiation of the cohort, the number of previous operations that they experienced is shown. The number of reoperations undergone by the cohort in the 5-year follow-up period is given. In total, 36 women underwent
Volume 189, Number 5 Am J Obstet Gynecol
Clark et al 1263
Fig 1. The index cases that were performed in 1995 are plotted on a time line to show both the antecedent operations and the need for reoperation as the reoperation relates to previous POPUI surgeries.
40 reoperations. Eight women had reoperation in the index year of 1995, and 32 women had reoperation in the subsequent 5 years. Two women underwent 2 reoperations, and 1 woman underwent 3 reoperations in the follow-up period. The survival of POPUI surgeries that are measured by reoperation is plotted against time to reoperation (months) in Fig 2. The results are stratified according to whether the index surgery was a first POPUI surgery or whether the woman had undergone previous POPUI surgery. At 71 months, the overall survival was 87% ± 2.7%, or conversely, the proportion of reoperation was 13%. When stratified for previous POPUI surgery, 12% ± 4.1% of the women who had their first POPUI procedure (n = 264 women) had undergone reoperation as compared with 17% ± 3.5% of women who had undergone previous POPUI surgery (n = 112 women; P = .04, log rank). At the end of 71 months, 300 women (78%) of the original cohort were still enrolled in the Kaiser health plan. The characteristics of the women who were lost to follow-up were compared with those women who remained in the cohort. Those women who were lost to follow-up were slightly younger (54 vs 58 years, P = .014), had a higher proportion with chronic lung disease (odds ratio, 2.5 [95% CI, 1.4-4.2], P = .001), and were more likely to have a smoking history (odds ratio, 1.67 [95% CI, 1.04-2.68], P = .034). All other characteristics were equivalent. Subject demographics and analysis of potential risk factors for the women in this cohort are shown in Table I. A previous operation for POPUI predicted a 1.9-fold increase in relative hazard (hazard ratio, 1.9, P = .04) of reoperation, but previous hysterectomy not for prolapse was not associated with an increased risk of reoperation. Those variables in the univariate analysis with probability values of < 0.25 were entered into a multiple Cox regression model and were found not to influence the
Fig 2. The survival of a POPUI surgery that was performed in 1995 is measured by the outcome of reoperation, stratified into those women who had a first POPUI surgery (open circles) in 1995 and those women who had a POPUI surgery before 1995 (closed squares, P < .05, log rank test).
relative hazard for previous POPUI surgery. Other variables were not found to be statistically associated with future reoperation. A post hoc power analysis was performed for this cohort. Assuming a power of 0.8, the study has the power to detect a 4-fold increase in risk that is associated with parity, an 8-fold risk with previous hysterectomy, a 6-fold risk reduction in risk with estrogen use, a 5-fold reduction in risk with smoking, and a 14-fold reduction in risk with chronic lung disease. The power to detect no difference varied from 0.07 to 0.63. A sample size of approximately 15,000 women would be required to provide a power of 0.8 to detect no difference for all variables in this study. A complete list of the 40 cases of reoperation and the preceding operations are listed in Table II. Every occurrence of reoperation is unique, except one. In Table II, the cases are grouped according to whether the reoperation occurred in the same anatomic compartments that were addressed in the preceding surgery (n = 24 cases, 60%), in a different compartment than in the preceding surgery (n = 13 cases, 32.5%), or both (n = 3 cases, 7.5%). In most cases, the number of compartments that was addressed is fewer in the reoperation than in the preceding operation. To evaluate patterns of reoperation, the cases have been categorized according to anatomic compartment or the condition that indicated surgery (Table III). The percentage of cases that were followed by any reoperation is shown. By compartment, reoperation was more likely when the preceding operation involved repair of the apex. By condition, the likelihood of reoperation was similar for all three groups.
1264 Clark et al
November 2003 Am J Obstet Gynecol
Table I. Univariate analysis of risk factors for reoperation for POPUI Demographic Analysis by Cox regression Mean age (y) Mean body mass index (kg/m2) Vaginal parity > 1 (%) Previous surgery for POPUI Hysterectomy not for prolapse (%) Worst grade of prolapse (%) 0 1+ 2+ 3+ Current estrogen use (%) Ever smoked (%) History of chronic lung disease (%)
Relative hazard (95% CI) 58.0 27.9 96 29.2 25.2 4.7 18.1 56.7 20.4 80.5 41.1 21.8
0.99 1.02 1.47 1.90 0.73
(0.96-1.02) (0.96-1.08) (0.20-10.7) (1.02-3.25) (0.32-1.70)
.48 .55 .70 .04 .48
1.45 (0.85-2.48)
.17
0.79 (0.37-1.69) 0.33 (0.08-1.38) 0.51 (0.18-1.45)
.21 .13 .54
Analysis by Fisher exact test (when any cell size = 0)
Reoperation status Yes No
Race (n) White Other History of corticosteroid use (n) Yes No
Table IV lists the component procedures within the complete operations of the index. The percentage of cases that resulted in reoperation in the same site and for the same condition were calculated. For example, for the abdominal retropubic urethropexy, the percentage of reoperations that included any incontinence component procedure was calculated. For vaginal hysterectomy for prolapse, the percentage that is shown is of any subsequent operation to repair the apex. The numbers are too small to report as the true risk of reoperation, but they allow a comparison of risk in this population to risks of failure and reoperation in published series. No single component predominates. The global reoperation risk comprises low risks from multiple procedures. The proportion of commonly reported ‘‘other site, other condition’’ reoperations was also calculated. Reoperation for urinary incontinence after a pelvic organ prolapse procedure (implying latent stress incontinence) occurred in 7.5% of cases. After abdominal retropubic urethropexies (RPUs), the following observations were made. Of 132 abdominal RPUs that were performed, 5% of the patients underwent reoperation at the vaginal apex. In 103 cases in which no apical repair was performed, 3% of the patients subsequently required apical suspension. In the 29 cases in which an apical procedure was performed concomitantly with the abdominal RPU, 4% of the cases required resuspension of the apex. The posterior compartment required reoperation in 4% of the 132 abdominal RPU cases; 79 of the 132 cases did not have a concomitant posterior repair. Of those 79 cases, 1% required a subsequent posterior colporrhaphy. Of the 43 cases in which a posterior colporrhaphy was performed at
P value
238 10
25 0
.61
6 241
0 25
.58
the same time as the abdominal RPU, 9% of the patients underwent a repeat posterior colporrhaphy. Of the 214 cases in which the primary surgeon was recorded, 63 surgeons were represented. Most of the surgeons were gynecologists. Urologists and colorectal surgeons performed or assisted with a small number of cases. None of the surgeons were fellowship trained in urogynecology or female pelvic floor medicine. Seventynine procedures (37%) were performed by 5 gynecologists with a special interest in urogynecology who performed at least 10 procedures per year. One hundred thirty-five procedures (63%) were performed by 58 surgeons who performed < 10 procedures per year. Twenty-six surgeons performed only 1 POPUI procedure in 1995. An analysis of reoperation by type of surgeon or surgeon’s experience was precluded by the small sample size. Comment Not only do women have a high lifetime risk of undergoing POPUI surgery, but also they face a high risk of undergoing reoperation for POPUI surgery. We expect that additional cases will come to reoperation as time passes. In the retrospective analysis of this cohort, the time interval between a first POPUI procedure and a second procedure averaged 12.5 years. Thus, our report likely represents only an initial snapshot of the total reoperations that this cohort will experience ultimately. The accuracy of our reported reoperation risk is limited by the 22% of women who were lost to follow-up, but the Kaplan Meier method compensates for this, which results in an SE of 2.7%. The women who were lost to follow-up
Clark et al 1265
Volume 189, Number 5 Am J Obstet Gynecol
Table II. List of reoperations Index operation Reoperation at same anatomic site (n = 24; 60.0%) Vaginal enterocele Abdominal RPU, PR Needle RPU TVH for POP, AR, PR, vaginal enterocele Needle RPU, AR, PR Needle RPU Needle RPU Needle RPU TVH for POP TVH for POP TAH for POP PR, AR Abdominal RPU, paravaginal, PR Abdominal RPU, PR AR, Vaginal enterocele, urethrolysis TAH for POP, abdominal RPU TVH for POP, abdominal RPU, SSLF, vaginal enterocele AR with mesh, vaginal enterocele Vaginal enterocele, PR TVH for POP, AR, PR, vaginal enterocele TVH for POP, AR, PR Abdominal RPU, PR, SSLF TVH for POP, abdominal RPU, PR Subtotal TAH, abdominal RPU, AR, PR, abdominal enterocele Reoperation at different anatomic site n = 13 (32.5%) Abdominal RPU TAH no POP, abdominal RPU Abdominal RPU Needle RPU Needle RPU TVH no POP, PR PR PR TVH for POP TVH for POP TVH for POP, SSLF TVH for POP PR, Vaginal enterocele Reoperation at both same and different site n = 3 (7.5%) Abdominal RPU Needle RPU TAH for POP, AR
Reoperation Abdominal enterocele PR Sling Vaginal enterocele, PR Paravaginal, abdominal RPU Sling Abdominal RPU, ureteroneostomy Needle RPU Sacrocolpopexy Sacrocolpopexy Sacrocolpopexy, abdominal enterocele Sling, AR PR Sling, PR AR with mesh excision, urethrolysis, Martius graft Sacrocolpopexy AR AR, vaginal enterocele, urethrolysis Sacrocolpopexy Abdominal RPU Sacrocolpopexy, abdominal RPU PR, vaginal enterocele, PR Vaginal enterocele Vaginal enterocele, AR PR Vaginal enterocele TAH for POP TAH for POP PR, Vaginal enterocele AR TAH for POP, AR TAH for POP, abdominal RPU Sling Needle RPU, AR AR, PR AR, needle RPU Needle RPU Vaginal enterocele, AR with mesh Sling, PR Vaginal enterocele, AR, PR
PR, Posterior repair; TVH, total vaginal hysterectomy; AR, anterior repair; TAH, total abdominal hysterectomy; POP, pelvic organ prolapse; SSLF, sacrospinous ligament fixation.
differed slightly from the women who remained to the end of the study period. Specifically, chronic lung disease and history of smoking were more prevalent in those women who left the cohort compared with those women who remained. If the prevailing sentiment is that chronic lung disease and smoking predispose to the recurrence of POPUI, then our reported risk may underestimate the true risk of reoperation. The accuracy of this analysis is also dependent on the accuracy of medical coding of procedures and the accuracy of data entry. Error rates for coding are estimated to be < 5% for medical coding and < 0.5% for nonmedical data entry.11 The generalizability of this analysis is influenced by the study population and the surgeons and their criteria for choosing surgical intervention. The population is a community-based managed care population that is well
stratified in terms of socioeconomics. This cohort is predominantly white, which limits our ability to detect differences according to ethnicity. There is no standardized approach for making the decision to perform POPUI surgery at KPNW. A diversity of surgical decision making is represented by at least 63 surgeons, who are primarily general obstetrician/gynecologists of varying backgrounds and experience. The choice of surgical intervention may also be influenced by variables such as access to medical care and financial considerations for both patients and health care providers. These variables are moderated in this mature managed-care environment, in which there are neither financial incentives nor disincentives to receiving surgical care. Having failed a POPUI surgery was the only risk factor that was identified in this population. Other variables that
1266 Clark et al
November 2003 Am J Obstet Gynecol
Table III. Reoperation by condition and compartment 1995 Index cases (n)
Reopertion (%)*
154 137 89
13 8 10
154 28 24 58 33 17 69
8 11 33 7 15 12 9
By condition Pelvic organ prolapse Urinary incontinence Both By compartment Anterior Posterior Apex Anterior and posterior Anterior and apex Posterior and apex All *Any condition or in any compartment.
surgeons commonly believe to be risk factors for surgical failure did not show significance in this study. For the variable of smoking, there was a trend toward decreasing the risk of reoperation, and this is contrary to common expectations. The observational design of this study may have precluded the detection of an effect from smoking and other potential risk factors because surgeons may have selected procedures differently on the basis of a clinical assessment of risk. In addition, more time may be required to see an effect on reoperation from the same risk factors that predict the development of POPUI. Having a failed POPUI surgery is likely a marker for a risk factor that has not yet been measured. The loss of support that occurs after neuromuscular damage in the pelvic floor may allow normal intra-abdominal forces to exceed the capacity of endogenous connective tissues.12,13 Currently, there is no accurate clinical tool for the assessment of neuromuscular function, which may be a very important predictor of failure that has never been well studied. Intrinsic weakness of the connective tissue, from genetic variation, is often posited a risk factor for the development of POPUI, which is evidenced by a differing risk with family history of prolapse and by ethnicity. Information on family history of prolapse was not available in this data set. Reoperation implies failure of the initial procedure but does not measure that failure exactly. Many women who are not treated successfully by surgery opt for conservative therapy. Reoperation represents the most severe occurrences of a failed procedure and includes surgical correction of adverse effects of the original procedure. Pelvic floor disorders result from structural failure of the muscles and connective tissue that comprise the pelvic floor. Defects are presumed to result from the elongation or breakage of the connective tissue and muscle and/or nerve injury that interferes with muscle function. Surgery aims to repair endogenous connective tissue or insert replacement materials to correct abnormal connective tissue and/or to compensate for abnormal neuromuscu-
Table IV. Reoperation specific to condition and compartment
Urinary incontinence Abdominal RPU Needle RPU Sling Anterior prolapse Anterior colporrhaphy Paravaginal repair Posterior prolapse: Posterior colporrhaphy Apical prolapse Total vaginal hysterectomy for pelvic organ prolapse Total abdominal hysterectomy for pelvic organ prolapse Sacrocolpopexy Sacrospinous Ligament Fixation Vaginal enterocele repair Abdominal enterocele repair Colpocleisis
1995 Index cases (n)
Reoperation (%)*
132 80 6
1 8 0
112 13 169
7 0 3
70
7
12
25
13
0
17 50 10 3
18 8 10 0
*Performed for same condition and in same compartment.
lar function. Direct failure occurs when the repaired tissue or grafted material is unable to maintain normal anatomy. Most reoperations (60%) occurred at the same anatomic site, which implies direct failure of the component surgical procedure. The site-specific reoperation risks in this cohort are comparable or lower to that reported in the literature. For sacrocolpopexy, Benson et al4 reported a 2.6% reoperation risk at the apex, compared with 0% in this study. For sacrospinous ligament fixation, Benson et al reported a 12% reoperation risk at the apex, compared with 17.5% in this group. No data were found for sitespecific reoperation risks for incontinence procedures, but the reoperation risk in this study (abdominal RPU, 0.8%; needle RPU, 8.3%; sling, 0%) was much lower compared with the reported 5-year surgical failure rates (abdominal RPU, 6%; needle RPU, 33%; and sling, 7%).14 The lower risk of site-specific reoperation in this community-based sample compared with academic samples supports the previous observation that the overall risk of reoperation is lower in community-based populations than in referral populations. An indirect failure occurs when a corrective procedure in one compartment predisposes to new failure in another compartment (for example, apical prolapse after retropubic urethropexy2 and posterior wall prolapse after retropubic urethropexy3). An alternate cause of apparent indirect failure results from the failure to identify and correct site-specific defects in all anatomic compartments. A smaller but significant proportion of reoperations (32.5%) occurred at different sites than in the preceding surgery. Apical and posterior compartment defects have been reported after abdominal retropubic urethropexy.
Volume 189, Number 5 Am J Obstet Gynecol
Wiskind et al2 reported a 29% risk of reoperation for apical defects after Burch colposuspension, which compares to a 3% risk in this study. Kjolhede et al3 reported a 29% risk of reoperation for rectocele after Burch colposuspension, which compares to a 1% risk in our study. The risks of other site-specific reoperation in community population that was compared are again lower than were reported in referral populations. An ideal surgical treatment for POPUI is a single surgical procedure that restores pelvic floor function with no adverse effects and lasts a lifetime. Clearly, we are far from that ideal. Women and their surgeons justifiably are frustrated by the lack of good clinical outcome data to guide surgeons in selecting operations for POPUI. National support is needed to fund and organize the large and expensive clinical trials that will be needed to address the complexities of outcome in pelvic floor surgery. Fortunately, this work has just begun. It is the responsibility of pelvic floor surgeons to keep this issue as a priority on the national research agenda for women’s health. REFERENCES 1. Olsen AL, Smith VJ, Bergstrom JO, Colling JC, Clark AL. Epidemiology of surgically managed pelvic organ prolapse and urinary incontinence. Obstet Gynecol 1997;89:501-6. 2. Wiskind AK, Creighton SM, Stanton SL. The incidence of genital prolapse after the Burch colposuspension. Am J Obstet Gynecol 1992;167:399-405.
Clark et al 1267
3. Kjolhede P, Noren B, Ryden G. Prediction of genital prolapse after Burch colposuspension. Acta Obstet Gynecol Scand 1996;75:849-54. 4. Benson JT, Lucente V, McClellan E. Vaginal versus abdominal reconstructive surgery for the treatment of pelvic support defects: a prospective randomized study with long-term outcome evaluation. Am J Obstet Gynecol 1996;175:1418-22. 5. Bump RC. Racial comparisons and contrasts in urinary incontinence and pelvic organ prolapse. Obstet Gynecol 1993;81:421-5. 6. Dwyer PL, Lee ET, Hay DM. Obesity and urinary incontinence in women. BJOG 1988;95:91-6. 7. Foldspang A, Mommsen S, Lam GW, Elving L. Parity as a correlate of adult female urinary incontinence prevalence. J Epidemiol Commun Health 1992;46:595-600. 8. Bump RC, McClish DK. Cigarette smoking and urinary incontinence in women. Am J Obstet Gynecol 1992;167:1213-8. 9. Bump RC, McClish DM. Cigarette smoking and pure genuine stress incontinence of urine: a comparison of risk factors and determinants between smokers and nonsmokers. Am J Obstet Gynecol 1994; 170:579-82. 10. Mant J, Painter R, Vessey M. Epidemiology of genital prolapse: observations from the Oxford Family Planning Association Study. BJOG 1997;104:579-85. 11. McLemore T, Lawrence L. Plan and operation of the National Survey of Ambulatory Surgery. Vital Health Stat 1 19971-4. 1-124. 12. Smith AR, Hosker GL, Warrell DW. The role of partial denervation of the pelvic floor in the aetiology of genitourinary prolapse and stress incontinence of urine: a neurophysiological study. BJOG 1989; 96:24-8. 13. Smith AR, Hosker GL, Warrell DW. The role of pudendal nerve damage in the aetiology of genuine stress incontinence in women. BJOG 1989;96:29-32. 14. Leach GE, Dmochowski RR, Appell RA, Blaivas JG, Hadley HR, Luber KM, et al. Female Stress Urinary Incontinence Clinical Guidelines Panel summary report on surgical management of female stress urinary incontinence: the American Urological Association. J Urol 1997;158:875-80.