Storage Irritative) and Voiding Obstructive) Symptoms as Predictors of Benign Prostatic Hyperplasia Progression and Related Outcomes

Transcription

1 European Urology European Urology ) 1±6 Storage Irritative) and Voiding Obstructive) Symptoms as Predictors of Benign Prostatic Hyperplasia Progression and Related Outcomes Claus G. Roehrborn a,*, John D. McConnell a, Brian Saltzman b, Donald Bergner c, Todd Gray d, Perinchery Narayan e, Thomas J. Cook f, Amy O. Johnson-Levonas f, Wilson A. Quezada f, Joanne Waldstreicher f for the PLESS Study Group) a Department of Urology, The University of Texas Southwestern Medical Center at Dallas, 5323 Harry Hines Blvd., J8-130, Dallas, TX , USA b Harvard Medical School, Beth Israel Hospital, Boston, MA, USA c Tampa Bay Medical, Clearwater, FL, USA d Lovelace Medical Center, Albuquerque, NM, USA e Department of Veterans Affairs, Gainesville, FL, USA f Merck Research Laboratories, Rahway, NJ, USA Accepted 26 April 2002 Abstract Objectives: To assess the utility of voiding and lling symptom subscores in predicting features of benign prostatic hyperplasia BPH) progression, including acute urinary retention AUR) and prostate surgery. Methods: The Proscar Long-termEf cacy and Safety Study PLESS) was a 4-year study designed to evaluate the effects of nasteride versus placebo in men with lower urinary tract symptoms LUTS), clinical evidence of BPH, and no evidence of prostate cancer. A self-administered questionnaire was employed to quantify LUTS at baseline. Receiver operating characteristics ROC) curves were used to assess baseline characteristics frompatients treated with placebo as predictors of outcomes. The characteristics assessed included the overall symptom score Quasi-AUA SI), separate voiding and lling subscores, prostate volume PV) and serum prostate-speci c antigen PSA) levels. Results: PV and PSA were superior to the symptom scores at predicting episodes of spontaneous AUR and all types of AUR. The Quasi-AUA SI and the lling and voiding subscores were effective at predicting progression to surgery; however, PSA was more effective at predicting this outcome. To better evaluate symptoms as predictors of surgery, patients who experienced a preceding episode of AUR were excluded fromthe surgery analysis. In the absence of preceding AUR, the best predictors of future surgery were the Quasi-AUA SI and the lling subscore. Conclusions: Among men with LUTS, clinical BPH and no history of AUR, the overall symptom score and storage subscore are useful parameters to aid clinicians in identifying patients at risk for future prostate surgery. PVand PSA were the best predictors of AUR, while PSA was the best predictor of prostate surgery for all indications). # 2002 Elsevier Science B.V. All rights reserved. Keywords: Benign prostatic hyperplasia; Acute urinary retention; Prostate surgery; BPH progression; Prostatespeci c antigen; AUA symptom score 1. Introduction Benign prostatic hyperplasia BPH) is a highly prevalent histological diagnosis in older men, and may * Corresponding author. Tel ; Fax: address: claus.roehrborn@utsouthwestern.edu C.G. Roehrborn). lead to progressive enlargement of the prostate gland and the development of lower urinary tract symptoms LUTS) [1]. Although many men with LUTS manage well without therapy, others experience a progressive worsening of their condition with the development of bladder outlet obstruction, which may ultimately result in acute urinary retention AUR) or require surgical intervention [2,3] /02/$ ± see front matter # 2002 Elsevier Science B.V. All rights reserved. PII: S )

2 2 C.G. Roehrborn et al. / European Urology ) 1±6 Clinical and epidemiological studies have identi ed numerous risk factors for BPH progression. A study based on a community cohort of men showed that, compared to younger men, the incidence of spontaneous AUR was three times higher in men aged 60 to 69, and eight times higher in men aged 70 to 79 years old [4]. In addition to age, increased incidence of AUR was associated with depressed urinary ow rates and prostate volume PV) >30 ml. Further supportive data come from the Proscar Long-term Ef cacy and Safety Study PLESS), which demonstrated that the incidence of AUR or prostate surgery in placebo-treated men increased from8.9% in patients with the smallest prostate volumes to 22.0% in patients with the largest volumes [5]. Similar data were obtained upon analyzing tertiles of serumprostate-speci c antigen PSA); the incidence of AUR or surgery increased from7.8% in the lowest tertile to 19.9% in the highest tertile. Data fromthis study demonstrated that PSA and PV are powerful predictors of the risk of AUR and surgery in men with LUTS and clinical BPH and no evidence of prostate cancer. The American Urological Association Symptom Index AUA SI) has been categorized into voiding obstructive) symptoms urinary hesitancy, diminished stream, straining, incomplete emptying, interruption of the urinary stream, and dribbling) thought to be associated with obstruction, and lling irritative) symptoms urgency, frequency, and nocturia) believed to arise frominvoluntary detrusor contractions detrusor overactivity). Despite the empiric validity of symptom subscores, there is little evidence supporting their clinical usefulness [6,7]. In this post-hoc analysis, we re-analyzed data frompless to assess whether separate voiding and lling subscores are predictive of BPH progression and related outcomes. 2. Patients and methods 2.1. Study design PLESS was a 4-year, placebo-controlled study evaluating 3040 patients with moderate to severe LUTS and enlarged prostates, with no evidence of prostate cancer. The present analysis includes only placebo-treated patients fromthis trial N ˆ Following a 1-month placebo run-in, patients were randomized to receive either nasteride 5 mg or placebo once daily for 4 years [8]. The primary endpoint, symptom score, was assessed by a self-administered questionnaire that queried subjects regarding the frequency of LUTS. Subsequent to the initiation of PLESS, the AUA SI was adopted as a standard for symptom assessment. The seven symptom components of the AUA SI nocturia, impairment in size and force of urinary stream, urinary frequency, delay, strained, or interrupted urination, and incomplete emptying of the bladder) were represented in the PLESS questionnaire. A pre-de ned linear transformation of the seven symptom components in the PLESS questionnaire was used togeneratethe Quasi-AUASI[8]. TheQuasi-AUASIwasdividedinto voidingandstorage lling)subgroupsaspublishedbywelchetal.[6]. Surgery and the occurrence of AUR during the 4-year period were prede ned secondary endpoints. An Endpoint Committee reviewed all the study documents for each man who underwent prostate surgery or who had AUR requiring catheterization of the bladder. The AUR was classi ed as spontaneous if there was no precipitating factor, or precipitated if there was a factor that may have contributed to the development of retention. PSA and symptom scores were measured in all patients. By design, PV was assessed in a subset of patients 10%) by magnetic resonance imaging Statistical analysis Receiver operating characteristics ROC) curves have been used to evaluate the ability of PV, PSA, and symptom score to predict the AUR and BPH-related surgery [5]. In this report, we used the same type of analysis to examine the ability of lling and voiding symptom subscores to predict BPH-related outcomes. Baseline symptom scores were used to plot ROC curves and the area under the curve AUC) was computed by the method of Hanley and McNeil [9]. The AUC for a particular population denotes the probability that a randomly chosen individual in the affected population has a higher value of a de ned parameter than a randomly chosen individual in the non-affected population. The area under the ROC curve is non-informative when AUC ˆ 0:5; therefore, values furthest from0.5, when statistically signi cant, are usually meaningful. An alpha level of 0.05 was used in the analyses. 3. Results 3.1. Patient accounting A total of 1516 men were randomized to placebo, and ef cacy data were available for 1503 men 99%). By the end of the study, 42% of patients discontinued treatment. Four-year follow-up data was available for 92% of all patients randomized, including those who discontinued treatment. In the remaining 8%, complete information was available until discontinuation of medication or until the 6-month follow-up after discontinuation. Over a 4-year period, AUR developed in 7% of placebo-treated patients, compared to 3% of patients receiving nasteride [8]. Approximately 10% of the patients in the placebo group versus 5% in the nasteride group underwent BPH-related surgery [8]. Among the placebo patients who had surgery, approximately 32% had a preceding episode of AUR either spontaneous or precipitated) Acute urinary retention PV and PSA were previously shown to be the best predictors of spontaneous and all types of AUR in PLESS. In contrast, the Quasi-AUA SI was a poor predictor of AUR [5]. In the present report, ROC curves were used to evaluate the performance of baseline lling and voiding subscores as predictors of all types

3 C.G. Roehrborn et al. / European Urology 42 (2002) 1±6 3 Table 1 Area under the ROC curve for several baseline parameters that may predict acute urinary retention and prostate surgery Outcome Baseline characteristic Filling symptoms Voiding symptoms Quasi-AUA SI score PV Serum PSA All AUR Spontaneous Precipitated * ** *** 0.703*** 0.632*** All surgery Excluding preceding AUR 0.588*** 0.615*** 0.558* 0.568* 0.587*** 0.608*** *** 0.564* * * p < 0:05. ** p < 0:01. *** p < 0:001. of AUR (Table 1; Fig. 1A) and spontaneous AUR (Table 1; Fig. 1B). For comparison, we also show the results from the ROC analyses of baseline PV, PSA and Quasi-AUA SI. The ROC curves for PV were discontinuous in nature due to the limited number of placebo patients who underwent prostate imaging at baseline. Relative to PV and PSA, the lling and voiding subscores were poor predictors of urinary retention (spontaneous, precipitated, and all kinds of AUR). The voiding subscore was numerically better at predicting future episodes of spontaneous AUR compared to the Quasi-AUA SI and the lling subscore. Fig. 1. The ROC curve analyses to measure the ability of baseline parameters, including serum PSA, PV, Quasi-AUA SI, and the separate lling and voiding symptom subscores, to predict (A) all episodes of AUR, and (B) spontaneous episodes of AUR in placebo-treated patients with LUTS and clinical BPH. Fig. 2. The ROC curve analyses to assess the ability of baseline parameters, including serum PSA, Quasi-AUA SI, and the separate lling and voiding symptom subscores, to predict (A) prostate surgery in the entire cohort of placebo-treated patients, or (B) in patients without a preceding episode of AUR.

4 4 C.G. Roehrborn et al. / European Urology ) 1±6 Fig. 3. Percent of patients needing surgery over a 4-year period, excluding all patients who experienced a preceding episode of AUR, strati ed by baseline response to two lling symptoms relating to nocturia and urinary frequency All prostate surgery As previously shown [5], PSA and PV were superior to the Quasi-AUA SI at predicting surgery Fig. 2A; Table 1). Despite the high AUC value for PV, this parameter did not show statistically signi cant predictive power for surgery due to the limited number of patients 10%) with prostate size measurements. The Quasi-AUA SI and the lling subscore were marginally better at predicting progression to surgery in placebotreated patients compared to the voiding subscore BPH-related surgery in the absence of AUR The ability of the Quasi-AUA SI and lling subscore to predict BPH-related surgery substantially improved when patients who had experienced a prior episode of AUR were excluded fromthe analysis Fig. 2B). The AUC values increased from0.587 p < 0:001 to p < 0:001 for the overall symptom score and from0.588 p < 0:001 to p < 0:001 for the lling subscore Table 1). While the voiding subscore remained nearly unchanged by this analysis, the predictive abilities of PV decreased from0.635 p ˆ 0:086 to p ˆ 0:490, and PSA from p<0:001 to p ˆ 0:03. Compared to the Quasi-AUA SI and the lling subscore, PVand PSA were relatively poor predictors of BPH-related surgery in men who did not have a preceding episode of AUR Predictive value of individual lling and voiding questions for surgery In order to explore the relationship between individual symptoms and risk of surgery, we plotted the percentage of patients needing surgery by baseline patient response to individual questions. None of the Quasi-AUA SI questions were predictive of surgery with the exception of two lling questions nocturia and increased urinary frequency) Fig. 3). Men reporting ve or more night time urinations at baseline had more than twice the risk of undergoing surgery than men reporting no night time urination. In addition, those men reporting urinary frequency from ``half the time'' to ``almost always'' at baseline had twice the risk of requiring surgery than patients experiencing less frequent urination. 4. Discussion In a previous analysis of PLESS data, we examined the ability of baseline characteristics to predict important BPH progression and related outcomes [5]. The ROC analyses were used to determine the predictive abilities of PSA, PV and the Quasi-AUA SI for AUR and prostate surgery for all indications including AUR-related prostatectomies). The results of this initial analysis demonstrated that PSA and PV were strong predictors of AUR and prostate surgery; however, the predictive value of PV did not reach statistical signi cance most likely due to the limited number of patients with prostate size measurements [5]. The similarity in the magnitude of the AUC values for PSA and PV most likely re ects the strong correlation between these two parameters [10]. Compared to PSA and PV, symptom severity, as measured by the

5 C.G. Roehrborn et al. / European Urology ) 1±6 5 Quasi-AUA SI, was a relatively poor predictor of AUR and prostate surgery for all indications [5]. In the present analysis, we sought to determine whether subdividing the Quasi-AUA SI score into its lling and voiding components would improve the predictive ability of symptom scores for long-term BPH-related outcomes. The results of this analysis con rm that compared to PV and PSA, symptom severity is a poor predictor of AUR and prostate surgery for all indications [5]. Of the symptom scores, the voiding subscore was the only signi cant predictor of spontaneous AUR p < 0:050, although it did not outperformpv and PSA. To explore whether symptom scores were useful at predicting surgery in the absence of a medical indication, we re-analyzed the symptom score data by excluding patients who developed spontaneous or precipitated AUR prior to undergoing surgery. The AUC values for the Quasi-AUA SI and the lling subscore increased substantially with this analysis, suggesting that lling/irritative and overall symptom scores are useful for assessing future risk of surgery in the absence of other indications. The AUC values for the Quasi-AUA SI and the lling subscore were similar in magnitude and signi cance, con rming that voiding symptoms were not predictive of BPH-related surgery. Two lling questions increased urinary frequency and nocturia) appeared to be driving the predictive ability of the symptom scores for BPH-related surgery. This nding supports the commonly held clinical belief that irritative symptoms are most bothersome to the patient, motivating doctor visits and ultimately leading to the decision to proceed with surgery in the absence of other medical indications. In comparison, the lling subscore was a relatively poor predictor of AUR, suggesting that the factors leading to AUR are clearly different fromthe sensory disturbances of irritative lling. Arrighi et al. [11] analyzed data fromthe Baltimore Longitudinal Study of Aging BLSA) to determine the ability of pre-surgical characteristics to predict subsequent prostatectomy in men with BPH. After controlling for age, ``diminished stream'', incomplete emptying, and increased PV by palpation had a signi cant positive association with the subsequent risk of having a prostatectomy [11]. Voiding symptoms and PV were found to be predictive of future prostatectomy, but nocturia, although prevalent in older men, was not. Unlike our analysis of PLESS, the importance of nocturia was not found in BLSA [11], but differences between the studies may account for such discrepancies. BLSA studied community men, not selected for age or for symptoms suggestive of prostate obstruction, while PLESS studied men of a restricted age range who had moderate-tosevere symptoms and enlarged prostates. Furthermore, BLSA included men with preceding AUR in analyzing symptoms predictive of future prostatectomy, while the PLESS analysis showed that, by excluding these patients, the predictive ability of the scores could be improved. Finally, BLSA examined the ability of presurgical variables assessed just prior to surgery) to predict prostatectomy, whereas PLESS analyzed baseline characteristics in a 4-year prospective clinical trial. Recently, Barry et al. re-examined data from the Veterans Administration Co-operative Study to determine the merit of calculating separate lling and voiding subscores [7]. Although these analyses con- rmed the empiric validity of subdividing the AUA SI into two symptom components, they did not establish the clinical usefulness of symptom subscores in the diagnosis of disease severity, or in identifying patients who may respond to speci c pharmacologic treatments. Our analysis indicates that lling symptoms may be useful in predicting prostate surgery among patients who do not have a speci c indication for surgery, such as AUR. Patients with advanced lling subscores who have not experienced AUR may bene t frommedical therapies proven to reduce the risk of surgery. Nevertheless, prostate volume and PSA are important considerations when weighing options for the treatment of bladder outlet obstruction secondary to BPH, as these measures are the best predictors of surgery for all causes AUR related or not). We recognize that clinical trials do not necessarily predict outcomes observed in clinical practice or in population based studies. General population studies have demonstrated a positive relationship between BPH symptoms and outcomes [4,12,13]; however, the relationships between separate voiding versus lling symptoms and individual outcomes have not been evaluated. This new analysis of PLESS demonstrated that the lling subscore and overall symptom score are useful predictors of BPH-related surgery in men who had not experienced a preceding episode of AUR. It is possible that in a broader sample population of patients with BPH, voiding and storage symptoms may demonstrate more usefulness in predicting BPH progression and related outcomes, especially AUR. It is also likely that baseline PV and PSA would still be the best predictors, since these parameters are more directly linked to the BPH disease process. Acknowledgements This study was sponsored by Merck Research Laboratories, Rahway, NJ. The authors would like to

6 6 C.G. Roehrborn et al. / European Urology ) 1±6 acknowledge the coordinators at the participating sites for their tremendous contribution to this study. The PLESS Study Group includes in alphabetical order): A. Aigen, P. Albertsen, R. Anderson, G. Andriole, S. Auerbach, M. Bamberger, J. Bannow, W. Barzell, D. Bergner, J. Bonilla, R.B. Bracken, W. Brannan, W. Bremner, T. Brown, R. Bruskewitz, R. Castellanos, S. Childs, K.S. Cof eld, T. Cook, C. Cox, E.D. Crawford, B. Dalkin, R.W. devere White, G. Drach, H. Epstein, C. Ercole, D. Falcone, D. Finnerty, W. Fitch, M. Flanagan, J. Fowler, H. Fuselier, D. Garvin, J. Geller, R. Gibbons, P. Gilhooly, M. Gittelman, S. Glickman, J. Gottesman, T. Gray, J. Grayhack, H. Guess, L. Harrison, R. Herlihy, G.B. Hodge Jr., H.L. Holtgrewe, R. Huben, P. Hudson, C.L. Jackson, E. Johnson, D. Kadmon, S. Kandzari, S. Kantor, S. Kaplan, M. Koppel, G. Kornitzer, D. Kozlowski, O. Kurzer, R. Labasky, J. Libertino, M. Lieber, R. Lund, S. Luttge, D. Lynch, G. Malek, N Mangelson, A. Matsumoto, J.D. McConnell, W.S. McDougal, A. Melman, D. Milam, R. Milsten, J. Mitchell, D. Mobley, P. Narayan, J.C. Nickel, L. Oppenheimer, F. Pappas, R. Parra, L. Peterson, J. Rajfer, P. Reddy, M. Resnick, O.F. Rigby, C.G. Roehrborn, N. Romas, S. Rosenberg, S. Rosenblatt, S. Rous, C. Rowe, J. Roy, B. Saltzman, W.P. Sawyer, P. Schellhammer, J. Schmidt, K. Short, T. Shown, D. Siegel, M. Soloway, T. Stanisec, B. Stein, E. Stoner, M. Sullivan, D. Sussman, A.M. Taylor, L. Tenover, J. Waldstreicher, P. Walsh, D. Wang, F. Wei, S. Weiner, G. Wells, H. Wessells, C. White, H. Wise, and G. Zhang. The authors gratefully acknowledge Dr. Michael J. Barry for his insight and comments on the analysis of BPH-related surgery. References [1] Garraway WM, Kirby RS. Benign prostatic hyperplasia: Effects on quality of life and impact on treatment conditions. Urology 1994;44:629±36. [2] Boyle P. Some remarks on the epidemiology of acute urinary retention. Arch Ital Urol 1998;LXX:77±82. [3] Girman CJ, Waldstreicher J. Prevention of the progression and longtermcomplications of benign prostatic hyperplasia. In: Resnick MI, Thompson IM, editors. Advanced Therapy of Prostate Disease. Hamilton BC): Decker Inc., p. 498±507. [4] Jacobsen SJ, Girman CJ, Guess HA, Rhodes T, Oesterling J, Lieber MM. Natural history of prostatism: Longitudinal changes in voiding symptoms in community dwelling men. J Urol 1996;155:595±600. [5] Roehrborn CG, McConnell JD, Lieber M, Kaplan S, Geller J, Malek GH et al. Serumprostate-speci c antigen concentration is a powerful predictor of acute urinary retention and need for surgery in men with clinical benign prostatic hyperplasia. Urology 1999;53:473±80. [6] Welch G, Kawachi I, Barry MJ, Giovannucci E, Colditz GA, Willett WC. Distinction between symptoms of voiding and lling in benign prostatic hyperplasia: Findings fromthe Health Professionals Followup Study. Urology 1998;51:422±7. [7] Barry M, Williford W, Fowler F, Jones KM, Lepor H. Filling and voiding symptoms in the American Urological Association Symptom Index: The value of their distinction in a veterans affairs randomized trial of medical therapy in men with a clinical diagnosis of benign prostatic hyperplasia. J Urol 2000;164:1559±64. [8] McConnell JD, Bruskewitz R, Walsh P, Andriole G, Lieber M, Holtgrewe HL et al. The effect of nasteride on the risk of acute urinary retention and the need for surgical treatment among men with benign prostatic hyperplasia. Finasteride Long-TermEf cacy and Safety Study Group. New Engl J Med 1998;338:557±63. [9] Hanley JA, McNeil BJ. The meaning and use of the area under a receiver operating characteristic ROC) curve. Radiology 1982;143:29±36. [10] Roehrborn CG, Boyle P, Gould AL, Waldstreicher J. Serumprostatespeci c antigen as a predictor of prostate volume in men with benign prostatic hyperplasia. Urology 1999;53:581±9. [11] Arrighi HM, Guess HA, Metter EJ, Fozard JL. Symptoms and signs of prostatismas risk factors for prostatectomy. Prostate 1990;16: 253±61. [12] Girman CJ, Jacobsen SJ, Guess HA, Oesterling JE, Chute CG, Panser LA et al. Natural history of prostatism: Relationship among symptoms, prostate volume, and peak urinary ow rate. J Urol 1995;153:1510±5. [13] Barry MJ, Fowler Jr. FJ, Bin L, Pitts JC, Harris CJ, Mulley AG. The natural history of patients with benign prostatic hyperplasia as diagnosed by North American urologists. J Urol 1997;157:10±5.