Waetjen, L. Elaine MD; Subak, Leslee L. MD; Shen, Hui MS; Lin, Feng MS; Wang, Tsung-Hsi MD; Vittinghoff, Eric PhD; Brown, Jeanette S. MD
Urinary incontinence is a distressing, common, and costly condition in women. As many as one third of women older than 40 years report involuntary loss of urine. 1–3 Although not a life-threatening condition, urinary incontinence may cause women to curtail their social and physical activities, which can lead to social isolation and depression. 4–7 In the United States, urinary incontinence costs are estimated at $16 billion per year. 8
Stress urinary incontinence is the most common type of incontinence in women younger than 60 years and accounts for approximately half of incontinence in all women. 2 Behavior modification, physiotherapy, biofeedback, and incontinence devices can be effective in reducing urine loss. 9–12 However, when conservative treatments fail, surgery is the treatment of choice for women with stress incontinence. The prevalence of stress incontinence surgery in the United States appears to be increasing, 13 yet there is limited information as to how the different types of these surgeries vary by region of the country, by age, and by racial category.
To estimate the number of surgeries performed for stress incontinence by type and by region of the United States, we analyzed data from the 1998 National Hospital Discharge Survey. Additionally, we examined the demographic characteristics of the women who underwent these surgeries and described the associated morbidity and mortality.
MATERIALS AND METHODS
We used the 1998 National Hospital Discharge Survey database to estimate the annual number of inpatient stress incontinence surgeries in the United States. The National Hospital Discharge Survey uses a multistage probability sample of nonfederal, short-stay, general hospitals nationwide; this process creates a sample of hospitals with an average length of stay of less than 30 days that represents national and regional distributions. From these selected hospitals, data is collected from hospital discharges, including patient demographics, length of hospital stay, a primary discharge diagnosis with up to six secondary diagnoses, and a primary procedure with up to three other procedures. Within selected hospitals, discharges were chosen for abstraction using systematic sampling. For 1998, data were collected from over 300,000 discharges from 478 (96%) of 495 eligible hospitals.
For our analysis, we selected all International Classification of Disease, 9th Revision (ICD-9) procedure codes representing stress incontinence surgeries. Similarly, we used ICD-9 diagnostic codes to determine discharge diagnoses. Stress incontinence surgeries in our analyses included plication of the urethrovesical junction such as the Kelly procedure (59.3), pubovaginal sling (59.4), Burch and Marshall Marchetti Krantz retropubic suspensions (59.5), paraurethral needle suspensions such as the Pereyra (59.6), levator muscle urethrovesical suspension operations (59.71), and other repairs (injection implants or repairs not specified) for urinary incontinence (59.72, 59.79). Anterior repairs (70.50, 70.51) were only included if the ICD-9 code for incontinence (625.6) was among the listed discharge diagnoses. Finally, we used the following ICD-9 procedure codes to identify major gynecologic procedures commonly done in conjunction with incontinence surgeries. For prolapse surgeries we used rectocele only (70.52), uterine suspension (69.22), vaginal suspension and fixation of the vagina (70.77), and enterocele surgery and operations on the cul-de-sac (70.92). Repair of cystocele and rectocele together (70.50), repair of cystocele only (70.51), vaginal hysterectomy (68.5), subtotal hysterectomy (68.3), total hysterectomy (68.4), vaginectomy (70.44), and obliteration of the vaginal vault (70.8) were also included as prolapse procedures if the ICD-9 codes for pelvic organ prolapse (618.0–618.6, 618.8, and 618.9) were among the discharge diagnoses. Codes for hysterectomy included subtotal (68.3), total (68.4), vaginal (68.5), and other (68.9).
To estimate mortality associated with stress incontinence surgery, we used the National Hospital Discharge Survey disposition code for “dead.” For morbidity, we used multiple ICD-9 codes representing surgical complications listed in the discharge diagnoses. Bleeding complications included codes for hemorrhage (998.11, 623.80), hematoma (998.12), acute anemia (285.10), and transfusion (99.0). Surgical injury included any injury to the bladder, ureter, or urethra (867.00), accidental puncture or laceration during surgery (998.20), and other genitourinary complications (997.50, 996.76). Infections included sepsis (38.90), urinary tract infections (593.89, 595.90, 599.00, 996.64), postoperative fevers (998.89, 780.60), and genital cellulitis (569.49, 597.00, 614.90, 616.10). We used codes for seroma (998.13), dehiscence (998.30), and wound infection (998.59) in our category of wound complications. Pulmonary complications included codes for acute respiratory distress syndrome (518.50) and pneumonia (518.89, 997.30). We categorized codes for pulmonary embolism (415.11, 415.19), congestive heart failure (428.00), myocardial infarction (410.00), and postoperative thromboembolic disease (997.20) as cardiovascular complications. Other complications included cerebrovascular complications (434.91) and anesthetic complications (995.20).
The National Hospital Discharge Survey coded race as white, black, Asian Pacific Islander, American Indian/Eskimo, other, and not stated. Hispanic ethnicity is not specified. We aggregated Asian Pacific Islander and American Indian/Eskimo into “other” because of the small numbers in these categories.
The National Hospital Discharge Survey provides yearly US Census Bureau population estimates nationally, by the Census-designated regions (Northeast, South, Midwest, and West), by race, and by age group. Weights based on probability of selection and adjusted for nonresponse and other factors were used to estimate nationwide numbers of procedures of specific types. In turn, these estimated totals were divided by population totals from the 1998 Census for women aged 20 years and older to calculate population rates for various inpatient surgeries nationally, by region, by race, and by age group. Finally, constants provided in the National Hospital Discharge Survey documentation were used to compute relative standard errors and confidence intervals (CIs) for totals, proportions, and rates among all (but not subgroups of) women. In the CIs for population rates, the population totals, which are based on Census enumeration, were treated as fixed.
In 1998, 134,987 women underwent surgeries for stress incontinence in the United States, at a rate of 13.4 per 10,000 women aged 20 years and older. The mean age of women undergoing incontinence surgery was 54.0 (±13.3) years.
More than half of stress incontinence surgeries were performed in women between the ages of 40 and 59 (Table 1). Table 1 also shows the distribution of incontinence surgeries by age group. Overall, retropubic suspensions were the most commonly performed procedure (46%), followed by anterior repairs (27%). Retropubic suspensions accounted for almost all of the surgeries in women aged 20–29 and almost two thirds of surgeries in women younger than 60. In women aged 60 and older, pubovaginal slings were three times more common than in women younger than 60 (18% compared with 5% of surgeries, respectively).
Figure 1 presents the variation in rates of stress incontinence surgery by age group and by region of the United States. The highest rate of surgery was in 60–69-year-old women (24.4 per 10,000). Women in the South had the highest overall rate of incontinence surgeries (14.8 per 10,000), whereas the Northeast had the lowest rate (9.8 per 10,000). The rate in the Midwest (14.2 per 10,000) and the West (13.4 per 10,000) ranged between these estimates.
There were large racial differences in the rates of stress incontinence surgery (Figure 2). White women had an overall stress incontinence surgery rate that was nearly five times greater than that for black women (11.6 per 10,000 compared with 2.6 per 10,000, respectively). The mean age for stress incontinence surgery in black women was 50.1 (±11.6) years, whereas for white women it was 54.0 (±13.2) years. Race was not stated in 21% of surgeries.
Overall, 68% (CI 63%, 73%) of surgeries for stress incontinence were done in conjunction with another major gynecologic surgery. About one third (36%, CI 31%, 41%) included any procedure for pelvic organ prolapse, and almost half (43%, CI 38%, 48%) were combined with a hysterectomy.
The average length of hospital stay for stress incontinence surgery was 2.8 (±1.5) days. This varied by type of procedure, with average length of stay of 2.5 (±1.5) days for pubovaginal sling and 3.0 (±1.8) days for retropubic suspensions. Hospital stays were longer in older women: 3.3 (±2.5) days for women older than 79 compared with 2.0 (±0.9) days in women aged 20–29 years.
We also analyzed incontinence surgeries by expected source of payment and type of hospital. Almost all expected reimbursement for stress incontinence surgery was from private insurance (eg, Blue Cross/Blue Shield, health maintenance organization/preferred provider organization; 62%) or government funding (eg, Medicare and Medicaid; 30%). For the remaining 8% of surgeries, payment was listed as “other” or was not stated. About 75% of incontinence surgery was performed in non-profit hospitals, and the rest were equally divided between government hospitals (12.5%) and proprietary hospitals (12.5%).
Mortality associated with stress incontinence surgery was 1 per 10,000 surgeries. Overall, 18.3% of the women who underwent incontinence surgery had one or more complications. This rate was similar when stress incontinence surgery was performed alone (17.1%) or in conjunction with other procedures (18.9%). Table 2 shows the frequency and rates of complications by type. Infection accounted for 44% of all complications, whereas surgical injury (24%) and bleeding (16%) were less common. Serious morbidity, such as pulmonary embolism or sepsis, accounted for less than 1% of all complications. There were also racial differences in the complication rates associated with urinary incontinence surgery (overall 20.6% for blacks compared with 9.7% for whites), with infection being almost twice as frequent in black women (14.2% compared with 8.0%).
Approximately 135,000 inpatient surgeries for stress incontinence were performed in the United States in 1998. The frequency of incontinence surgery has increased dramatically over the last ten years—from 78,000 surgeries in 1988. 13 Adjusting for population growth, the rate of incontinence surgery has increased approximately 45%. Stress incontinence surgery is one of the more common procedures performed in women. For comparison, the frequencies of other operations in 1998 include hysterectomy (605,000), tubal ligation (325,000), coronary artery bypass graft (187,000), total knee replacement (169,000), and appendectomy for suspected appendicitis (119,000). 14
Rates of stress incontinence surgery varied by region within the United States. As with other elective gynecologic surgeries, the largest difference was between women in the South and in the Northeast (14.8 compared with 9.8 per 10,000 women). The variation in the regional distribution of surgeries likely reflects regional differences in physician training and patient attitudes regarding surgery.
There was an almost five-fold higher rate of stress incontinence surgeries in white compared with black women in 1998. A number of studies have reported a higher prevalence of stress incontinence in white women compared with black women. 1,15,16 Physiologic research has suggested that levator strength and urethral pressures are greater in black women. 15,17 However, the racial differences in rates of incontinence surgery may be influenced by a number of important factors, including socioeconomic status, insurance status, access to and utilization of care, patient reporting or physician ascertainment of incontinence symptoms, and variations in general attitudes regarding undergoing elective surgery. Our observed two-fold increase in surgical morbidity for black compared with white women, which is similar to the racial differences in complications described for hysterectomy, 18 is concerning and bears further investigation.
Anterior repair was the second most common type of surgery performed for stress incontinence. This frequency is surprising, given solid evidence that anterior repair has poor short- and long-term efficacy compared with all other types of stress incontinence surgery. 19–21 However, this frequency may reflect physician and patient preference for the least invasive surgery for minimal stress incontinence symptoms.
Almost three of four women who underwent incontinence surgeries had concomitant major gynecologic procedures. In this way, the rate of complications (18.3%) reflects the morbidity of combined procedures, as well as of stress incontinence surgeries performed alone. It is, therefore, not surprising that the overall morbidity rate reported here is similar to other common gynecologic surgeries. 18,22
The National Hospital Discharge Survey has several limitations. The accuracy of our estimates depends on the accuracy and completeness of ICD-9 coding for diagnoses and procedures and the collection of demographic characteristics listed on discharge summaries. Only procedures with stress urinary incontinence listed as one of the seven discharge diagnoses recorded in the 1998 National Hospital Discharge Survey were included in this analysis. The accuracy of hospital coding for diagnoses and procedures in the National Hospital Discharge Survey has been reported to exceed the standards of other data from the National Center for Health Statistics. 23 The National Hospital Discharge Survey estimates coding errors to be approximately 0.9% for procedures and diagnoses and 0.2% for demographic data. 24
Another limitation of our estimates is that stratum and primary sampling unit identifiers used by specialized programs for analyzing multistage probability samples are not included in the publicly available National Hospital Discharge Survey data set. Instead, we used approximations recommended in the National Hospital Discharge Survey documentation to compute relative standard errors and CIs. The approximate CIs for population totals may be less reliable than intervals computed using information on stratum and primary sampling unit to account for clustering. In addition, our CIs for population rates do not take account of errors in the denominators and thus may understate the uncertainty of our ratio estimates. However, unlike the numerators, which are weighted sample estimates of population totals, the denominators are based on US Census enumeration, adjusted for population growth and undercounting of some population groups. As a result, the denominator errors are likely to be negligible as compared with the uncertainty of the numerators, or to the uncertainty that would obtain if the denominators were based on sampling rather than enumeration.
This analysis only includes surgeries for incontinence that required hospital admission. Incontinence procedures performed outside the nonfederal, short-stay hospitals, such as tension-free vaginal tape, collagen, or durasphere injections, are not included in the National Hospital Discharge Survey, so we have likely underestimated stress urinary incontinence surgery frequency. In addition, because the National Hospital Discharge Survey does not follow individual patients longitudinally, our morbidity and mortality estimates include only perioperative complications and deaths. They do not reflect posthospitalization complications or readmissions and thus are likely underestimated as well.
Data on race was missing for 21% of discharges for incontinence surgery in 1998, and thus the conclusions regarding racial differences found in our study are limited. However, despite this under-reporting of race, an analysis of missing racial data in the National Hospital Discharge Survey has described an under-representation of white patient discharges, 24 therefore the discrepancy in rates of incontinence surgery is more likely to be greater than what we have reported. In addition, given that the race assignments were made by the admitting hospital, these categories likely reflect appearance rather than self-identification. Finally, the nature of race/ethnicity data collection and the small number of women in the other National Hospital Discharge Survey race/ethnicity categories allowed only the general and simplistic distinctions between black and white.
Stress urinary incontinence surgery is common, especially in white women, and the number of such surgeries performed in the United States has nearly doubled in the last decade. The large regional and racial differences in the rates of incontinence surgery raises important health policy questions regarding physician practices, patient preferences for incontinence treatment, and differential access to and utilization of care.
1. Brown JS, Grady D, Ouslander JG, Herzog AR, Varner RE, Posner SF. Prevalence of urinary incontinence and associated risk factors in postmenopausal women. Heart & Estrogen/Progestin Replacement Study (HERS) Research Group. Obstet Gynecol 1999;94:66–70.
2. Hannestad YS, Rortveit G, Sandvik H, Hunskaar S. A community-based epidemiological survey of female urinary incontinence: The Norwegian EPINCONT study. Epidemiology of Incontinence in the County of Nord-Trondelag. J Clin Epidemiol 2000;53:1150–7.
3. Seim A, Sandvik H, Hermstad R, Hunskaar S. Female urinary incontinence—consultation behaviour and patient experiences: An epidemiological survey in a Norwegian community. Fam Pract 1995;12:18–21.
4. Dugan E, Cohen SJ, Bland DR, Preisser JS, Davis CC, Suggs PK, et al. The association of depressive symptoms and urinary incontinence among older adults. J Am Geriatr Soc 2000;48:413–6.
5. Brown JS, Subak LL, Gras J, Brown BA, Kuppermann M, Posner SF. Urge incontinence: The patient's perspective. J Womens Health 1998;7:1263–9.
6. Grimby A, Milsom I, Molander U, Wiklund I, Ekelund P. The influence of urinary incontinence on the quality of life of elderly women. Age Ageing 1993;22:82–9.
7. Hunskaar S, Vinsnes A. The quality of life in women with urinary incontinence as measured by the sickness impact profile. J Am Geriatr Soc 1991;39:378–82.
8. Wilson L, Brown JS, Shin GP, Luc K, Subak LL. Annual direct cost of urinary incontinence. Obstet Gynecol 2001; 98:398–406.
9. Nygaard I. Prevention of exercise incontinence with mechanical devices. J Reprod Med 1995;40:89–94.
10. Berghmans LC, Hendriks HJ, Bo K, Hay-Smith EJ, de Bie RA, van Waalwijk van Doorn ES. Conservative treatment of stress urinary incontinence in women: A systematic review of randomized clinical trials. Br J Urol 1998;82:181–91.
11. Sirls LT, Foote JE, Kaufman JM, Lightner DJ, Miller JL, Moseley WG, et al. Long-term results of the femsoft urethral insert for the management of female stress urinary incontinence. Int Urogynecol J Pelvic Floor Dysfunct 2002;13:88–95.
12. Miller JM, Ashton-Miller JA, DeLancey JO. A pelvic muscle precontraction can reduce cough-related urine loss in selected women with mild SUI. J Am Geriatr Soc 1998;46:870–4.
13. Korn AP, Learman LA. Operations for stress urinary incontinence in the United States, 1988–1992. Urology 1996;48:609–12.
14. Lawrence L, Hall MJ. 1997 summary: National Hospital Discharge Survey. Adv Data 1999:1–15.
15. Duong TH, Korn AP. A comparison of urinary incontinence among African American, Asian, Hispanic, and white women. Am J Obstet Gynecol 2001;184:1083–6.
16. Graham CA, Mallett VT. Race as a predictor of urinary incontinence and pelvic organ prolapse. Am J Obstet Gynecol 2001;185:116–20.
17. Howard D, Delancey JO, Tunn R, Ashton-Miller JA. Racial differences in the structure and function of the stress urinary continence mechanism. Obstet Gynecol 2000;95:713–7.
18. Kjerulff K, Guzinski G, Langenberg P, Stolley P, Adler Moye N, Kazandijian V. Hysterectomy and race. Obstet Gynecol 1993;82:757–64.
19. Black NA, Downs SH. The effectiveness of surgery for stress incontinence in women: A systematic review. Br J Urol 1996;78:497–510.
20. Glazener CM, Cooper K. Anterior vaginal repair for urinary incontinence in women. Cochrane Database Syst Rev 2001;1:CD001755.
21. 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.
22. Brown JS, Waetjen LE, Subak LL, Thom DH, Van den Eeden S, Vittinghoff E. Pelvic organ prolapse surgery in the United States, 1997. Am J Obstet Gynecol 2002;186:712–6.
23. Dennison C, Pokras R. Design and operation of the National Hospital Discharge Survey: 1988 redesign. Vital Health Stat 1 2000;39:1–42.
24. Kozak LJ. Underreporting of race in the National Hospital Discharge Survey. Adv Data 1995;265:1–12.