Introduction. JW Moul 1,2, RM Mooneyhan 2, T-C Kao 3, DG McLeod 1,2 and DF Cruess 3

Prostate Cancer and Prostatic Diseases (1998) 5, 242±249 ß 1998 Stockton Press All rights reserved 1365±7852/98 $12.00 http://www.stockton-press.co.uk/pcan Preoperative and operative factors to predict incontinence, impotence and stricture after radical prostatectomy JW Moul 1,2, RM Mooneyhan 2, T-C Kao 3, DG McLeod 1,2 and DF Cruess 3 1 Urology Service, Department of Surgery & Department of Clinical Investigation, Walter Reed Army Medical Center, Washington, DC 20307-5001; 2 Center for Prostate Disease Research, Department of Surgery; and 3 Department of Preventive Medicine and Biometrics, Division of Epidemiology and Biostatistics, Uniformed Services University of the Health Sciences, 4301 Jones Bridge Road, Bethesda, MD 20814-4799, USA The purpose of this study was to determine the incidence of patient-self reported post prostatectomy incontinence, impotence, bladder neck contracture or stricture, better, same or worse quality of life and willingness for same treatment again in a large group of radical prostatectomy (RP) patients and to determine if these morbidities are predictable with demographic, surgical or prostate cancer (PC) factors. Methods: A patient self-reporting questionnaire was completed and returned by 374 out of 458 eligible (81.7%) RP patients from one center, 267 (72.2%) have been operated since 1990 and all of whom were a minimum six month postoperative (75%>1 y). Questionnaire results were independently analyzed by a third party and correlated to demographic, operative, and tumor factors in an ongoing comprehensive PC database. Results: The patient self-reported incidence of post prostatectomy incontinence (any degree), impotence, and bladder neck contracture or stricture was 72.2, 87.4, and 25.9%, respectively. The reported rate of incontinence requiring protection was 39.0% and only 2.4% had persistent bladder neck contracture/stricture. Pathologic stage (continuous variable) was the only factor to signi cantly predict incontinence and no factor could predict impotence or bladder neck contracture/ stricture in univariate analysis. No factor was predictive of morbidity by multivariate analysis. Despite incontinence and impotence signi cantly affecting QOL self-reporting (P ˆ 0.001, 0.001, respectively) and willingness to undergo RP again (P ˆ 0.001, 0.067, respectively), the majority of patients would choose surgery again. Conclusions: Although radical prostatectomy morbidity is common and affects patient-reported overall QOL, most patients would choose the same treatment again. Demographic, preoperative, operative, and tumor factors did not reliably predict patient-reported morbidity in this series. Keywords: prostate; cancer; incontinence; impotence; stricture; radical prostatectomy Introduction Because of the aging population, PSA-screening, and public awareness of prostate cancer (PC), the use of radical prostatectomy (RP) has increased dramatically Correspondence: Dr JW Moul, Department of Surgery, USUHS, 4301 Jones Bridge Road, Bethesda, MD 20814±4799, USA. Received 3 December 1997; revised 24 February 1998; accepted 25 February 1998 over the last decade. 1 This increase has led to the recognition and reality of postoperative morbidities including impotence, 2 incontinence, 3 and bladder neck contracture and/or urethral stricture and their potential impact on quality of life. 4 A number of recent patient surveys have documented the incidence of these problems in the Medicare, 5 CHAMPUS, 6 managed care, 7 private sectors, 8 and by multicenter pre- and post questionnaire. 9 None of the studies to date have attempted to de ne patient, surgical and/or tumor factors that may be risk factors for the

subsequent occurrence of these patient-reported outcomes. The goal of this study was to conduct a large patient self-reported questionnaire survey of sexual, urinary, and quality of life and to correlate results to an ongoing retrospective and prospective PC database to examine for factors that may predict the likelihood of these morbidities. Materials and methods A questionnaire was designed to assess the presence and degree of patient self-reported post-prostatectomy incontinence (PPI), impotence (IMP), bladder neck contracture and/or urethral stricture (S), quality of life (QOL) and willingness to undergo the same cancer treatment again, this was based on the instrument of Fowler et al 5 and is available upon request, and is also given in Table 2. The questionnaire was eld-tested for comprehension and clarity by 50 members of the Walter Reed US TOO, Inc., prostate cancer support group prior to study use. Study subjects were patients who had been related by radical retropubic prostatectomy at Walter Reed Army Medical Center between 1973±1994. Of the 471 patients treated by RP during this interval, 13 had died at the survey time of March 1995 leaving 458 patients to whom the survey was mailed. A business reply envelope was included with the questionnaire and after 30 d a followup letter was mailed to remind nonresponders. After 60 d, 374 surveys had been returned (81.7%) which formed the study cohort, for these responders, 96 (25.7%) completed the survey between six months and one year from surgery, 134 (35.8%) between 1±3 y, and 144 (38.5%) greater than 3 y from RP. As part of the Department of Defense Center for Prostate Disease Research (CPDR) program, Walter Reed Army Medical Center is one of ten military hospitals designated to collect prospective and retrospective comprehensive clinical data on all prostate cancer patients. As of 1 January 1994, data collection was prospective on all new PC patients at Walter Reed and similar retrospective data has also been collected on all patients treated since 1970 through inpatient and outpatient record review and interviews of living patients. Standardized data collection forms for transrectal ultrasound biopsy, registration, staging, surgery, radiation, hormonal, cryotherapy, brachytherapy, follow-up, and necropsy have been developed and are available upon request. Data has been collected by physicians and research personnel. This data is entered and maintained on a relational database (Oracle Corporation). The questionnaire project and CPDR database have been approved by the Walter Reed Department of Clinical Investigation Human Use Institutional Review Board. Grading conformed to the Gleason system and pathologic staging to the 1992 TNM system. 10; 11 The de nition of incontinence from the patient questionnaire and used for subsequent analysis in Tables 3 and 4 was for any dripping or leaking problems after surgery and the use of protection (pads, diapers, rubber pants or clamps). Impotence was de ned as any patient who could have erections when stimulated before surgery but could not have full erections after surgery. Statistical analysis Questionnaire data on the 374 patients and the corresponding complete diagnostic, demographic, staging and surgical data for these patients were provided to two of us (TK, DFC) for independent tabulation and statistical analysis, a 5% signi cance level was utilized in all statistical tests. The Statistical Analysis System (SAS) statistical computer software package was used to do all statistical calculations. Chi-square tests were used to determine if there is a signi cant association between categorical risk factors and categorical outcome variables [for example, race (black/white) vs impotence (yes/no)]. Analysis of variance (ANOVA) or two-sample t-test was used to test for statistical signi cance of the difference in the means of continuous risk factors variables across the categories of outcome variables (for example, average tumor volume in those incontinent compared to average tumor volume in those not incontinent post-surgery). The multivariate approach of stepwise logistic regression was used to determine the effect of a number of risk factors simultaneously on the dichotomous outcome variables of incontinence (absence vs presence) and quality of life (better or same vs worse). The risk factors considered in the construction of the logistic regression models included age at surgery, race, surgical era, Gleason sum, pathologic T stage, PSA, operation time, and estimated blood loss (EBL). Stepwise logistic regression rst identi es the risk factor that is the strongest predictor of the outcome variable of interest, followed by the next strongest, etc., until there is no statistically signi cant bene t in adding new risk factors to the prediction model. Additionally, at each step, any risk factor already present in the model is examined to see if it still is a statistically important predictor of the outcome variable when other risk factors have been added to the model. To determine the placement and retention of risk factors in the model, signi cance levels for entry of 15% and removal of 20% were used; these more liberal levels are common in stepwise logistic regression model building. Results The demographic, stage, grade and surgical data contained in the CPDR database of the 374 patient who responded to the patient-self-reported questionnaire is given in Table 1 and the data from the questionnaire is given in Table 2. Notably, the vast majority (86.1%) of patients were less than or equal to 70 y of age at the time of RP and the mean age was 63.2 y. There was racial diversity with 21.9% African American patients and the majority (63.6%) of all patients were married. Gleason sum grade was similar to other series with a majority of patients (61.3%) being intermediate (Gleason 5±7) grade. Similarly, pathologic stage was similar to other reports with 162 (43.3%) organ-con ned and 199 (53.2%) nonorgan-con ned patients. Mean pre-treatment PSA for 308 evaluable patients (66 patients treated prior to PSA availability) was 9.80 ng/ml with a median of 6.75 ng/ml and range of 0.1±112.5 ng/ml. A majority (267±71.2%) of patients were treated since 1990 and the operative time (4.15 hours-mean) and operative estimated blood loss 243

244 (2069 cc mean, 1700 cc median) were similar to other reports during this era from teaching hospitals. Most noteworthy regarding the survey results in Table 2 are the rates of post prostatectomy incontinence and impotence. Fully, 270 patients (72.2%) reported some degree of incontinence at the time of survey, however, 204 (54.6%) reported a few drops or less than a tablespoon and 96 (25.6%) noted leakage once or less a day. Furthermore, among the 374 respondents, protection by using pads (32.6%), diapers (9.6%) or clamps (2.7%) was reported. Of 320 preoperative potent patients, 41 patients (12.8%) reported natural full erections postoperatively and a signi cant additional number were using penile self-injection (40 patients 10.7%), vacuum devices (125 patients 33.4%) or had a penile prosthesis (50 patients 13.4%). Although 97 patients (25.9%) reported at least one episode of bladder neck contracture or stricture requiring treatment, only 9 (2.4%) patients reported that the problem was not resolved at the time of the survey. Regarding overall patient-perceived quality of life, 214 Table 1 Demographic, stage, grade and surgical data of 374 racial prostatectomy patients 4.150.07 (3.93) 2069 94 (1700) 5.88 0.63 (3.61) Age at surgery: Mean SE (median) a 63.2 0.4 (63.8) 70 y 322 (86.1%) >70 y 46 (12.3%) Unknown 6 (1.6%) Race Caucasian 285 (76.2%) Black 82 (21.9%) Other 7 (1.9%) Marital status Married 238 (63.6%) Single/divorce/widowed 134 (35.8%) Unknown 2 (0.5%) Gleason sum grade (RP specimen min) 4 45 (12.0%) 5 75 (20.0%) 6 77 (20.5%) 7 78 (20.8%) 8 34 (9.1%) Unknown/not performed 65 (17.4%) Pathologic stage T2a 48 (12.8%) T2b 24 (6.4%) T2c 90 (24.1%) T3a 101 (27.0%) T3b 58 (15.5%) T3c 40 (10.7%) Unknown 13 (3.5%) Pre-treatment PSA mean (median) b 9.80 69 (6.75) Surgical era <1990 103 (27.5%) 1990 268 (71.1%) unknown 5 (1.3%) Operation time (h) mean SE (median) c Estimated blood loss (cc) mean SE (median) d Total tumor volume (cc) mean SE (median) e a Based on 368 patients. b PSA ˆ maximum of pre-biopsy and pre-treatment value, ng/ml, based on 308 patients (82.4% of cohort). c Based on 310 patients. d Based on 315 patients. e Based on 96 patients. (57.2%) patients reported it to be the same or better and 281 (75.1%) patients would choose the same treatment again. Table 3a provides the bivariate analysis of demographic, tumor and surgical variables with incontinence (any self-reported incontinence and using protectionpads, diapers, clamps from Table 2, II2 and 3), impotence (no full erections after RP but having full erection before from Table 2, III2) and stricture (self-reported bladder neck contracture or stricture requiring treatment from Table 2, IV1). In the bivariate analysis, there was no demographic, preoperative or operative variable that was signi cantly associated with these three patient self-reported morbidities. Potent patients (mean 60.4 y) were younger than impotent patients (63.1 y) which tended to signi cance (P ˆ 0.083). Similar non-signi cant trends were seen for incontinence and Gleason sum>6 (P ˆ 0.074), and bladder neck contracture/stricture and higher estimated blood loss (P ˆ 0.089). Without dichotomizing pathologic stage (that is, using all categories from Table 1), we found that incontinence was signi cantly associated with higher pathologic stage (P ˆ 0.007) and a trend for higher Gleason sum grade (P ˆ 0.073). Similarly, with continuous Gleason sum grade, impotence was slightly associated with advancing grade (P ˆ 0.067). Higher tumor volume (available for 94 patients) as a continuous variable was marginally associated with impotence (P ˆ 0.083). For patient self-reported quality of life and choosing the same treatment again, Table 3b reveals three signi cant associations. Patients reporting worse quality of life were younger (P ˆ 0.023), blacks reported better QOL (P ˆ 0.006), and patients treated prior to 1990 reported a better QOL (P ˆ 0.017). None of the other bivariate preoperative or operative factors were related to QOL or reported willingness to undergo surgery again. Higher Gleason sum grade as a continuous variable was associated with not choosing RP again (P ˆ 0.004). Table 4 provides data on the correlation of patient selfreported quality of life and willingness to undergo radical prostatectomy again with self reported incontinence, impotence and bladder neck contracture/stricture. For patients who reported incontinence, 57.2% reported a worse QOL than before treatment compared to 29.0% of the continent men (P<0.001). Similarly, 47.2% of impotent patients reported worse QOL compared to 18.4% who regained potency (P ˆ 0.001). There was no correlation of self reported bladder neck contracture/stricture and QOL. Regarding self-reported willingness to undergo surgery again, 28.1% of incontinent men would not choose surgery again vs 13.3% of those who reported continence (P ˆ 0.001). Less signi cant trends to not choosing RP again were seen for impotence and bladder neck contracture/stricture. Overall, 71.9±79.2% of incontinent or impotent men or those with bladder neck contracture/stricture report that they would choose the same treatment again and 42.8±56.4% had the same or better self-reported QOL despite reporting these morbidities. Multivariate stepwise logistic regression analyses were performed for a number of outcome variables. For the outcome variable incontinence (absence vs presence), stepwise logistic regression selected only one surgical variable, estimated blood loss (odds ratio (OR) ˆ 0.9998,

Table 2 Patient self-reported questionnaire results in 374 radical prostatectomy patients 245 Yes (%) No (%) Unknown/not applicable not answered (%) I. Cancer recurrence questions 1. As far as you know, do you have prostate cancer anywhere now? 43 (11.5) 319 (85.3) 12 (3.2) 2. Since your prostate surgery, have you had: a. Any pills or injections of estrogens or hormones for PC? 26 (7.0) 342 (91.4) 6 (1.6) b. Surgery to remove testicles? 18 (4.8) 336 (89.8) 20 (5.3) c. Radiation treatments? 74 (19.8) 286 (76.5) 14 (3.7) II. Incontinence questions 1. After prostate surgery some men nd that they have a problem with dripping or leaking urine. a. Did you have that problem to any degree immediately after your 300 (80.2) 69 (18.4) 5 (1.3) prostate surgery? b. Did you have any additional surgery to help stop dripping or 39 (10.4) 327 (87.4) 8 (2.1) leaking? 2. a. Do you currently have any problem at all with dripping or leaking? 270 (72.2) 101 (27.0) 3 (0.8) b. If yes, have you experienced any of the following in the past month? dripped or leaked urine when you coughed or sneezed? 186 (49.7) 75 (20.1) 113 (30.2) c. Dripped or leaked urine when bladder was full before getting to 192 (51.3) 70 (18.7) 112 (29.9) bathroom? d. When you drip about how much usually comes out? a few drops 126 (33.7) less than a tablespoon 78 (20.9) a tablespoon or more 52 (13.9) unknown/not answered 118 (31.6) e. How often do you drip or leak urine? less than once a day 48 (12.8) once a day 48 (12.8) more than once a day 150 (40.1) 3. Do you use any of the following to help with wetness at this time? pads 122 (32.6) 148 (71.7) 104 (27.8) adult diapers 36 (9.6) 234 (62.6) 104 (27.8) rubber pants 2 (0.5) 268 (71.7) 104 (27.8) clamps 10 (2.7) 260 (69.5) 104 (27.8) 4. Have you tried any of the following for leakage? biofeedback 3 (10.8) 267 (71.4) 104 (27.8) behavioral training 30 (8.0) 240 (64.2) 104 (27.8) pelvic exercise training 141 (37.7) 128 (34.2) 105 (28.1) III. Sexual function questions 1. The year prior to your prostate surgery, could you have erections 320 (85.6) 52 (13.9) 2 (0.5) when you were stimulated? 2. Since your prostate surgery, are you capable of having full erections? 41 (11.0) 327 (87.4) 6 (1.6) 3. Since your prostate surgery, have you tried any of the following? penile injections (shots) 40 (10.7) 286 (76.5) 48 (12.8) vacuum suction device 125 (33.4) 230 (61.5) 19 (5.1) penile implant surgery 50 (13.4) 305 (81.6) 19 (5.1) 4. Are you happy with your degree of sexual function in general? very satis ed 41 (11.0) somewhat satis ed 63 (16.8) somewhat unsatis ed 87 (23.3) very unsatis ed 132 (35.3) IV. Post surgical stricture questions 1. One problem some patients have after prostate surgery is called strictures. Strictures are scar tissue that can form in the urinary tract that can make it dif cult to urinate. a. Since your prostate surgery, has a doctor had to dilate your urinary 97 (25.9) 266 (71.1) 11 (2.9) tract or perform any surgery to treat strictures? b. Is this stricture problem now resolved? 81 (21.7) 9 (2.4) 284 (75.9) V. Comments 1. After your prostate cancer treatment, do you feel your overall quality of life is: better than before 63 (16.8) same as before 151 (40.4) worse than before 142 (38.0) unknown/not answered 18 (4.8) 2. If you had the choice of cancer treatment today, would you take the same course of treatment? 281 (75.1) 66 (17.6) 27 (7.2)

246 Table 3a Bivariate analysis of demographic, tumor and surgical variables related to outcome variables: incontinence, impotence and stricture Incontinence Impotence Surgical stricture? Yes No Yes No Yes No Total 147 (40%) 224 (60%) 277 (88%) 38 (12%) 97 (27%) 266 (73%) Age at surgery.(y) Mean.(SE) 63.1.(0.5) 63.2.(0.5) 63.1.(0.4) 60.4.(1.4) 62.7.(0.6) 63.4.(0.4) N 146. 220. 274. 37. 96. 262. Race Black 32.(40%) 49.(60%) 62.(87%) 9.(13%) 16.(20%) 64.(80%) White 113.(40%) 170.(60%) 212.(89%) 26.(11%) 81.(29%) 195.(71%) Surgical era year 1990 39.(38%) 64.(62%) 74.(85%) 13.(15%) 31.(32%) 67.(68%) year 1990 107.(41%) 156.(59%) 201.(89%) 24.(11%) 65.(25%) 195.(75%) PSA.(maximum level of prebiopsy and pre-treatment PSAs) Mean.(SE) 10.0.(1.2) 9.7.(0.9) 10.3.(0.8) 8.1.(1.5) 11.3.(1.5) 9.3.(0.8) N 121. 184. 229. 29. 77. 224. Gleason sum 6 72.(37%) 123.(63%) 148.(87%) 23.(13%) 45.(23%) 148.(77%) 6 53.(47%) 59.(53%) 85.(91%) 8.(9%) 27.(25%) 81.(75%) Pathologic T stage T2a,b,c 59.(36%) 104.(64%) 128.(89%) 16.(11%) 45.(28%) 114.(72%) T3a,b,c 86.(42%) 118.(58%) 146.(87%) 22.(13%) 51.(26%) 149.(74%) Operation time.(h) Mean.(SE) 4.2.(0.1) 4.1.(0.1) 4.1.(0.1) 4.2.(0.2) 4.3.(0.1) 4.1.(0.1) N 125. 183. 228. 30. 78. 224. Estimated blood loss.(cc) Mean.(SE) 2238.(184) 1939.(99) 2035.(116) 2063.(298) 2448.(291) 1930.(79) N 124. 188. 232. 34. 79. 230. Tumor total volume.(cc) Mean.(SE) 5.25.(0.7) 6.3.(1.0) 6.1.(0.7) 3.8.(1.1) 5.6.(1.1) 6.0.(0.8) N 39. 56. 82. 12. 23. 71. Chi-square or two-sample t-test was used to decide if a risk factor was related to each outcome variable. No signi cance was found (P>5%). Table 3b Bivariate analysis of demographic, tumor and surgical variables related to outcome variables: quality of life, treatment again Quality of life Treatment again Better Same Worse Yes No Total 63 (18%) 151 (42%) 142 (40%) 281 (81%) 66 (19%) Age at surgery.(y) Mean.(SE) 64.8.(0.8) 63.6.(0.6) 62.1.(0.6) a 63.0.(0.4) 63.4.(0.9) N 61 149. 140. 276. 66. Race Black 22.(28%) 24.(31%) 32.(41%) a 58.(76%) 18.(24%) White 38.(14%) 124.(46%) 109.(40%) 216.(82%) 48.(18%) Surgical era year 1990 21.(22%) 49.(50%) 27.(28%) a 79.(85%) 14.(15%) year 1990 41.(16%) 100.(39%) 113.(45%) 198.(79%) 52.(21%) PSA. (maximum level of prebiopsy and pre-treatment PSAs) Mean.(SE) 9.4.(1.2) 8.6.(0.7) 11.2.(1.5) 9.6.(0.8) 11.4.(1.9) N 48. 124. 123. 232. 56. Gleason sum 6 34.(18%) 78.(41%) 77.(41%) 151.(80%) 37.(20%) 6 17.(16%) 45.(42%) 45.(42%) 83.(82%) 18.(18%) Pathologic T stage T2a,b,c 23.(15%) 75.(47%) 61.(38%) 130.(84%) 24.(16%) T3a,b,c 39.(20%) 76.(39%) 79.(41%) 150.(79%) 39.(21%) Operation time.(h) Mean.(SE) 4.2.(0.1) 4.2.(0.1) 4.1.(0.1) 4.1.(0.1) 4.2.(0.2) N 54. 119. 123. 234. 56. Estimated blood loss.(cc) Mean.(SE) 2290.(253) 2071.(172) 1930.(119) 2010.(85) 2266.(372) N 54. 124. 124. 240. 56. Tumor total volume.(cc) Mean.(SE) 5.47.(1.1) 5.2.(1.0) 6.5.(1.2) 6.0.(0.7) 5.5.(1.4) N 19. 34. 36. 79. 11. a Denotes signi cant.(p<5%), where Chi-square or two-sample t-test was used to decide if a risk factor was related to each outcome variable.

Table 4 Comparison of patient self-reported incontinence, impotence, and stricture and self-reported quality of life and willingness to undergo radical prostatectomy again 247 Quality of life Treatment again Better Same Worse P-value b Yes No P-value Incontinence Yes 16 (11.6) 43 (31.2) 79 (57.2) 0.001 a 97 (71.9) 38 (28.1) 0.001 a No 46 (21.2) 108 (49.8) 63 (29.0) 183 (86.7) 28 (13.3) Impotence Yes 35 (13.2) 105 (39.6) 125 (47.2) 206 (79.2) 54 (20.8) No 11 (29.0) 20 (52.6) 7 (18.4) 0.003 a 34 (91.9) 3 (8.1) 0.083 Stricture Yes 15 (16.0) 38 (40.4) 41 (43.6) 0.434 68 (73.9) 24 (26.1) 0.033 a No 47 (18.3) 110 (43.0) 99 (38.7) 209 (83.6) 41 (16.4) a Indicates signi cant association between the two corresponding variables. b P-value of Mantel-Haenszel Chi-square test adjusted or the surgical era (before or after 1990). its 95% con dence interval (CI) ˆ [0.9996, 1.0000], P ˆ 0.0651, N ˆ 230) in the nal model. This implied that a cubic centerimeter incremental increase in blood loss would slightly decrease the odds of having continence. For another outcome variable, quality of life (dichotomized as `better or same' vs `worse'), the nal logistic model selected age (OR ˆ 1.047, 95% CI ˆ [1.004, 1.092], P ˆ 0.309), surgical era (year >1990 vs <1990: OR ˆ 0.458, 95% CI ˆ [0.183, 1.147], P ˆ 0.0956) and PSA (OR ˆ 0.981, 95% CI ˆ [0.958, 1.005], P ˆ 0.1185; N ˆ 221). By using the same set of potential risk factors, without dichotomizing pathologic stage and Gleason sum grade, similar results were found except for quality of life in which this nal logistic model selected age (OR ˆ 1.050, 95% CI ˆ [1.006, 1.095], P ˆ 0.0238, and PSA (OR ˆ 0.977, 95% CI ˆ [0.951, 1.004], P ˆ 0.0949; N ˆ 216). Discussion This is the rst patient-self-reported outcome survey study that has been correlated to a database of patient, tumor and surgical variables in an attempt to discover potential predictors of common radical prostatectomy morbidities of impotence, incontinence and urethral/ bladder neck contracture/strictures. To our surprise, few factors were correlated to these three morbidity outcomes. Pathologic stage (continuous variable) was the only statistically signi cant factor associated with post prostatectomy incontinence. High Gleason sum (>6) was marginally (P ˆ 0.07) associated with incontinence as well. Not surprisingly, younger men were more likely to regain potency (P ˆ 0.083). Furthermore, higher Gleason sum (P ˆ 0.067) and tumor volume (P ˆ 0.083) as continuous variables were marginally associated with impotence. A higher estimated blood loss was the only marginal (P ˆ 0.09) risk actor for bladder neck contracture/stricture. In multiple logistic regression analysis using both binary data and continuous data for stage and grade, no factor remained signi cant to predict incontinence and only estimated blood loss trended to signi cance (P ˆ 0.065). With a larger study, it is possible that one or more of these factors may emerge that convincingly predict these three common morbidities. Alternatively, these morbidities may be truly based on individual patient or surgical factors. Self-reported worse quality of life was associated with incontinence and impotence, however, the majority of men reported that they would choose radical prostatectomy again. Despite our careful attempt to link patient self-reported outcome to database variables, there are a number of limitations to our study. Our quality of life assessment was a crude, simple question about QOL and was not based on a validated QOL questionnaire. Our question about choosing the same treatment again is dubious since most studies that ask this question whether it be radiation or surgical procedures report positive replies by patients. We were unable to assess the impact of type of nerve sparing surgery or impact of primary surgeon, we were not con dent in reliably assigning nerve sparing status from operation reports from the retrospective database patients prior to 1994. This is unfortunate since two recent reports came to different conclusions regarding nervesparing and continence. 9,12 Finally, even though we asked about pretreatment potency we did not control for preoperative continence. Fortunately, a recent survey found that pretreatment incontinence was reported by less than 5% of similar patients. 9 Despite these limitations, we believe our study adds to the growing body of literature in this area. Our self-reporting survey results are similar to recent reports even though our patients were somewhat younger (mean±63.2 y) on average than a recent similar survey of Medicare bene ciaries. 5 Fowler et al 5 found a 40% incidence of incontinence (any degree), a 20% incidence of stricture requiring treatment, and a 11% potency rate in 757 Medicare patients who underwent RP. In a smaller survey of similarly aged men (mean 65 y) from one private practice, Lim et al 8 noted a 43% incontinence rate (any degree) and only 4% of patients had spontaneous erections suf cient for vaginal penetration. Jonler et al 13 noted a 74% rate of incontinence after RP with 34% noting that dripping was a signi cant problem. Lim et al 8 found that 82 out of 89 (92%) surveyed patients would choose RP again, in line with our rate of 81.1%. Most recently, Talcott et al 9 conducted a preoperative and postoperative prospective survey of RP patients from multiple hospitals. Even with bilateral nerve sparing RP, only 11% of their cohort were fully potent. Unilateral nerve sparing was not signi cantly better than non-nerve-sparing for preserving potency in their study. Interestingly, regarding

248 post prostatectomy incontinence, 14% of Talcott et al 9 nonnerve sparing patients were wearing pads but 50% of nerve sparing patients required pads. In contrast, Eastham et al 12 found that nerve-sparing RP was an independent predictor of improved continence. Our quality of life assessment was a simple question regarding the patients self reported perceived QOL in relation to before treatment. We did not employ a standardized QOL instrument as other investigators 7,8 and this was a limitation, however, this was not the primary focus of our study. Despite this, our simple QOL assessment did correlate to morbidity. Similar to Lim et al 8 and Litwin et al 7 we found that incontinence and impotence did impact patient self-reported QOL. Overall, approximately 40% of this cohort of 374 RP patients reported a worse QOL than prior to treatment and it was highly related to incontinence and impotence. Post-prostatectomy incontinence is a common problem in the era of increased use of radical prostatectomy as treatment for prostate cancer. 14,15 Although initially felt to be related to bladder and sphincteric dysfunction with similar frequency, recent study suggests that sphincteric etiology predominates. 16 With this in mind, it may be surprising to some that operation time, blood loss, pretreatment PSA and tumor volume did not predict incontinence. Either these factors do not uniformly in uence risk of incontinence or our study was too small to detect a signi cant difference. Regarding age, Catalona and Bigg 20 also did not nd an association that older men were more likely to experience post prostatectomy incontinence. However, our study as well as that of Catalona and Bigg is skewed toward younger carefully-selected otherwise healthy men. It is possible that a study with a higher proportion of older men with comorbidities or more advanced disease may show that age is associated with incontinence. Similarly, had we had more men with locally advanced (stage C or T3B,C) disease there may have been an independent association of post prostatectomy incontinence and stage. Tumor volume assessed by whole-mount three dimension reconstruction was only available for 94 patients; this factor may have predicted morbidity in a larger cohort. It was reassuring that despite more locally advanced disease, 17 black men treated by RP did not have a higher risk of these three morbidities. Furthermore, the percentage of black and white men reporting worse quality of life was similar (40 vs 41%) and blacks were actually more likely to report better post-operative QOL than whites (28 vs 14%). The rate of return of potency in the era of nerve sparing RP is a debated issue. Although potency rates in reports from centers of excellence have ranged form 54±71%, 18±20 these are selected populations and the potency is based on clinic visits. The patient self reported survey results from less selected groups by us and others 5,8,9 is probably more re ective of the scope of morbidity in general community practice. Only younger age and low Gleason ( 4) were found to predict return of potency. Our results are very similar to the recent report of Talcott et al 9 in which 11% of bilateral nerve sparing patients who had pre-operative potency were fully potent at one year post-operatively. Furthermore, 100% of unilateral nerve sparing and non-nerve sparing patients reported inadequate erections. Although this current study focused on one treatment, radical prostatectomy, there is an urgent need to conduct similar studies of the variety of treatments now used for localized prostate cancer including external beam radiation, brachytherapy and cryotherapy. European investigators have embarked on such studies, 21 however, much more work is needed. Conclusions In a patient self-reported questionnaire and database study of 374 radical prostatectomy patients of multiple surgeons from one centre predominantly treated in the late 1980s and early 1990s, impotence and incontinence were higher than generally recognized. Any degree of incontinence was reported by 72.2%, 39.0% were wearing protection, but the majority had mild stress incontinence. Impotence was reported by 87.2% of the 320 men who were potent preoperatively. Despite these side effects, a majority of men report the same or better quality of life after operation and the majority would choose the same treatment again. Similar studies of mid-to-late 1990s era patients should be conducted to determine if improvements in surgical technique result in higher patient satisfaction. Acknowledgements This research was supported by the following grants: Department of Clinical Investigation, Walter Reed Army Medical Center, Washington, DC, Grant No. 2898 and The Center for Prostate Disease Research, a program of Henry M. Jackson Foundation for the Advancement of Military Medicine, 1401 Rockville Pike, Suite 600, Rockville, MD 20852-3007, Grant No. 0998-901-1022. References 1 Lu-Yao GL, McLerran D, Wasson J, Wennberg JE. An assessment of radical prostatectomy. Time trends, geographic variation and outcomes. JAMA 1993; 265: 2633±2636. 2 Ofman US. Sexual quality of life in men with prostate cancer. Cancer 1995; 75: 1949±1953. 3 Blaivas JG. Urinary incontinence after radical prostatectomy. Cancer 1995; 75: 1978±1982. 4 Litwin MS. Health-related quality of life after treatment for localized prostate cancer. Cancer 1995; 75: 2000±2003. 5 Fowler FJ et al. Patient-reported complications and follow-up treatment after radical prostatectomy. Urology 1993; 42: 622±629. 6 Optenberg SA, Wojcik BE, Thompson IM. Morbidity and mortality following radical prostatectomy: A national analysis of civilian health and medical program of the Uniformed Services (CHAMPUS) bene ciaries. J Urol 1995; 153: 1870±1872. 7 Litwin MS et al. Quality of life outcomes in men treated for localized prostate cancer. JAMA 1995; 273: 129±135. 8 Lim AJ et al. Quality of life: radical prostatectomy versus radiation therapy for prostate cancer. J Urol 1995; 154: 1420±1425. 9 Talcott JA et al. Patient-reported impotence and incontinence after nerve-sparing radical prostatectomy. J Nat Cancer Inst 1997; 89: 1117±1123.

10 Gleason DF and the Veterans Administration Cooperative Urological Research Group. Histological grading and clinical staging of prostate carcinoma. In: Tannenbaum M (ed). Urologic Pathology: The Prostate. Lea and Febiger: Philadelphia, 1977, pp 171±197. 11 Ohori M, Wheeler TM, Scardino PT. The new American joint committee on cancer and international union against cancer TNM classi cation of prostate cancer. Cancer 1994; 73: 104±114. 12 Eastham JA et al. Risk factors for urinary incontinence after radical prostatectomy. J Urol 1996; 156: 1707±1713. 13 Jonler M, Messing EM, Rhodes PR, Bruskewitz RC. Sequelae of radical prostatectomy. Br J Urol 1994; 74: 352. 14 Moul JW. An important goal of prostatectomy: minimizing incontinence. Contemp Urol 1994; 6: 15±28. 15 Moul JW. For incontinence after prostatectomy, tap a diversity of treatments. Contemp Urol 1994; 6: 78±88. 16 Chao R, Mayo ME. Incontinence after radical prostatectomy: detrusor or sphincter causes. J Urol 1995; 154: 16±18. 17 Moul JW et al. Prostate-speci c antigen values at the time of prostate cancer diagnosis are higher in African-American men. J Am Med Assoc 1995; 274: 1277±1281. 18 Walsh PC, Epstein JI, Lowe FC. Potency following radical prostatectomy with wide unilateral excision of the neurovascular bundle. J Urol 1987; 138: 823. 19 Leandri P, Rossignol G, Gautier JR, Ramon J. Radical retropubic prostatectomy: morbidity of quality of life. Experience with 620 consecutive cases. J Urol 1992; 147: 883. 20 Catalona WJ, Bigg SW. Nerve-sparing radical prostatectomy: evaluation of results after 250 patients. J Urol 1990; 143: 538. 21 Fossa SD et al. In uence of urological morbidity on quality of life in patients with prostate cancer. Eur Urol 1997; 31 (Suppl 3): 3±8. 249