Prostate Cancer and Prostatic Diseases (2005) 8, 206–209 & 2005 Nature Publishing Group All rights reserved 1365-7852/05 $30.00
www.nature.com/pcan
Risk factors for disease progression in patients with lower urinary tract symptoms/benign prostatic hyperplasia (LUTS/BPH): a systematic analysis of expert opinion FC Lowe1*, J Batista2, R Berges3, E Chartier-Kastler4, G Conti5, F Desgrandchamps6, K Dreikorn7, M O’Leary8, M Perez9, M Speakman10, J Trachtenberg11, A Tubaro12, B Meesen13, L Smets13 & H Stoevelaar14 1
St Luke’s-Roosevelt Hospital Center, New York, NY, USA; 2Fundacion Puigvert/ CM Teknon, Barcelona, Spain; 3PAN-Klinik am Neumarkt, Cologne, Germany; 4GH Pitie´-Salpe´trie`re, Paris, France; 5Ospedale S. Anna, Como, Italy; 6CHU Saint Louis, Paris, France; 7Zentralkrankenhaus St-Juergenstrasse, Bremen, Germany; 8Brigham & Women’s Hospital, Boston, Massachusettes, USA; 9Medical Sciences University of the Sa˜o Paulo Santa Casa Hospital, Sa˜o Paulo, Brazil; 10Taunton & Somerset Hospital, Taunton, UK; 11University of Toronto & Princess Margaret Hospital, Toronto, Canada; 12 La Sapienza University, Sant’Andrea Hospital, Rome, Italy; 13Ismar Healthcare, Lier, Belgium; and 14Erasmus University Medical Center, Rotterdam, The Netherlands
Disease progression has become an important issue for the management of lower urinary tract symptoms/benign prostatic hyperplasia (LUTS/BPH). Although several risk factors have been identified, no specific patient risk profiles have been established that can be useful in the day-to-day management of LUTS/BPH. In this study, an international panel of urologists developed a risk classification based on the attribution of a risk score to 243 unique patient profiles. From the perspective of clinical decision making, it was concluded that postvoid residual, symptom severity and maximum flow rate are the most relevant determinants of the risk of disease progression. Prostate Cancer and Prostatic Diseases (2005) 8, 206–209. doi:10.1038/sj.pcan.4500806; published online 14 June 2005
Keywords: lower urinary tract symptoms; benign prostatic hyperplasia; disease progression; risk factors
Introduction Disease progression has become an important issue in the ongoing debate on the appropriate management of lower urinary tract symptoms suggestive of benign prostatic hyperplasia (LUTS/BPH). Community-based *Correspondence: FC Lowe, Department of Urology, St Luke’s/ Roosevelt Hospital, 425 West 59th Street, Suite 3A, New York, NY 10019, USA. E-mail:
[email protected] Received 23 January 2005; accepted 27 February 2005; published online 14 June 2005
studies and randomised controlled trials (RCTs) have identified several clinical parameters that are associated with the risk of symptom deterioration, serious complications such as acute urinary retention (AUR) and/or the need for invasive therapy.1–7 These include age, prostate volume, prostate-specific antigen (PSA), symptom severity, maximum flow rate (Qmax), and postvoid residual (PVR).1–7 The relevance of identifying patients at different risk of disease progression has become manifest since recent studies have shown that long-term treatment with a1-adrenoceptor (AR) antagonists, 5a-reductase inhibitors (RIs) and their combination reduce the risk of clinical progression, AUR, and the need for invasive
Risk factors for disease progression in LUTS/BPH FC Lowe et al
therapy.6 However, the a1-AR antagonist doxazosin only reduced the occurrence of AUR compared to placebo in patients with a low-to-intermediate PSA, whereas combination therapy of the a1-AR antagonist with the 5a-RI finasteride provided the largest benefit with regard to reducing the development of AUR and need of invasive therapy, in particular in high-risk patients such as those with a high PSA/large prostate volume.8 In order to be able to determine the most appropriate treatment strategy (watchful waiting, medication, (minimally) invasive therapies) in relation to the probability and magnitude of disease progression, we need to develop a risk classification based on the combination of all (potentially) relevant risk factors. Unfortunately, currently available epidemiological and clinical studies mainly provide information on the impact of single risk factors and are therefore less useful in the day-to-day management of LUTS/BPH patients for whom the total risk is determined by the combination of different clinical variables. As a first step to develop a risk classification that can be used at the patient-specific level, we conducted a panel study in which international experts on LUTS/BPH assessed the risk of disease progression for a large number of hypothetical patients.
Table 1 Overview of clinical variables and categories used for the construction of 243 patient scenarios (risk profiles) Variables
Categories
1. Symptoms (total I-PSS)
(a) 0–7 (b) 8–19 (c) 20–35 (a)o30 (b) 30–59 (c) X60 (a)o1.4 (b) 1.4–3.9 (c) X4.0 (a)o15 (b) 10–15 (c) o10 (a)o50 (b) 50–150 (c)4150
2. Prostate volume (ml) 3. PSA (ng/ml) 4. Qmax (ml/s) 5. PVR (ml)
Number of risk profiles/patient cases: 35 ¼ 243.
Table 2 Agreement (X8 panellists with the same score) on the classification of the risk of disease progression for all patient scenarios (n ¼ 243) Risk score
Agreement Number of scenarios
Percentage of all scenarios
Very low Low Intermediate High Very high
1 5 41 11 2
0.4 2.1 16.9 4.5 0.8
Total
60
24.7
(Very) low Intermediate (Very) high
19 41 41
7.8 16.9 16.9
101
41.6
Methods The assessment of risk profiles was conducted within the framework of a larger study on the appropriateness of LUTS/BPH treatments9 using the RAND Appropriateness Method (RAM).10 In that study, a panel of 12 urologists from eight countries in Europe and North and South America assessed the appropriateness of seven common treatments for LUTS/BPH for 243 different patient scenarios. These scenarios were based on the permutation of the values of several clinical variables, considered by the panel as most relevant risk factors for disease progression. Selection of variables was based on an extensive literature review11 and the opinions of the panellists about their relevance for treatment choice in daily clinical practice. The final set of clinical variables and their cutoff points is shown in Table 1. In an earlier phase of this study, age was also included as a potentially relevant risk factor.12 As this variable proved to be irrelevant to the judgement of risk progression after a first assessment by the panel, it was excluded from further analyses. Panellists were asked to attribute a risk score for disease progression to each of the patient scenarios. The 5-point scale that was used for this purpose ranged from 1 (very low) – 5 (very high). Risk of disease progression was defined as the occurrence of any of the following events in the next 1–4 years: symptom deterioration experienced as significant by the patient and/or, AUR, overflow incontinence or any other serious complications, clearly due to BPH and/or, need for invasive treatment. As there are yet no absolute cutoff points for the risk of disease progression,11 the risk score had to be considered as an ordinal scale in which consecutive marks represent a higher risk of disease progression. Agreement on risk assessment for each patient scenario was calculated
207
5-point scale 1 2 3 4 5
3-point scale 1–2 3 4–5
Total
using the cutoff point such that X8 panellists had the same score. The relation between risk scores and clinical variables was investigated using logistic regression analysis.
Results All panellists (n ¼ 12) returned the appropriateness assessments and risk scores. Few panellists did not complete a small part of the scores because they believed that the associated scenarios represented patients unlikely to have LUTS/BPH. Since no consensus existed on the exclusion of particular scenarios and the quantitative impact was limited (maximum of two out of 12 missing scores for 23% of the scenarios), we decided to include all scenarios, correcting the figures for missing data. Table 2 shows the extent of agreement between panellists. Using the subdivision into five categories, Prostate Cancer and Prostatic Diseases
Risk factors for disease progression in LUTS/BPH FC Lowe et al
208
the agreement was low (on average 25%). After rescaling the risk score into three categories, the percentage of agreement increased to 42%. However, as the risk score represents a ‘relative’ assessment (no absolute end points were defined), differences between panel members may result from their particular scoring behaviour (eg, ‘conservative’ vs ‘extreme’). The most important issue here is the consistency in underlying patterns, that is, the question as to whether particular (combinations of) variables are associated with a higher risk score. Therefore, we studied the relationship between the risk scores and the clinical variables included, using logistic regression analysis. As the mean panel scores showed a fairly normal distribution, we took this measure as the dependent (dichotomous) variable. In all statistical
Table 3 Logistic regression coefficients for the impact of risk factors on the mean risk score (o3.0 vs X3.0). Variable Symptoms (total I-PSS) Prostate volume (ml) PSA (ng/ml) Qmax (ml/s) PVR (ml) Constant value
Value
b
s.e.
P-value
8–19 20–35 30–59 X60 1.4–3.9 X4.0 10–15 o10 50–150 4150
4.16 12.07 5.00 8.42 2.39 6.24 3.17 9.64 3.98 16.39 21.99
1.15 2.42 1.25 1.77 0.94 1.44 1.05 2.06 1.06 3.20 4.32
0.000
models, we used different cutoff points for an ‘elevated’ risk (mean risk score X3.0, X3.5 and X4.0). As explanatory variables, we included all five risk factors and potentially relevant interactions. The regression results for the dichotomy low risk vs intermediate/high risk (cutoff point ¼ 3.0) are summarised in Table 3. For all variables, the coefficients were positive and in the expected direction. Considerable PVR (4150 ml), severe symptoms (total International Prostate Symptom Score (I-PSS) 20–35) and poor Qmax (o10 ml/s) showed the highest contribution to a risk score X3.0. No meaningful interactions were found, indicating that the impact of the various risk factors is merely cumulative. This cumulative effect is illustrated in Figure 1. It can be seen that the impact of moderate ‘unfavourable’ factors on the risk score is very low, but that the combination with severe abnormalities, particularly high PVR, severe symptoms and poor Qmax, considerably increases the probability of an elevated risk (risk score X3.0). Albeit the impact of high PVR and poor Qmax was somewhat higher than that of severe symptoms, additional analyses using higher cutoff points for an elevated risk (risk score X3.5 and X4.0) showed similar patterns.
0.000 0.000 0.000 0.000
The regression results show the impact of each factor in comparison to a reference patient, having no unfavourable conditions (total I-PSS 0–7, prostate volume o30 ml, PSAo1.4 ng/ml, Qmax 415 ml/s, PVR o50 ml).
Discussion Although several epidemiological and clinical studies have substantially increased our knowledge about the risk of disease progression in patients with LUTS/BPH,1–7 their results are of limited usefulness for daily clinical practice. Most studies focused on the impact of single risk factors but practising doctors are faced with patients
Figure 1 Cumulative effect of different risk factors on the risk score, probabilities based on logistic regression analysis, and values calculated against a reference patient having no unfavourable conditions (left column). Prostate Cancer and Prostatic Diseases
Risk factors for disease progression in LUTS/BPH FC Lowe et al
often representing combinations of several ‘unfavourable’ conditions. This study explored the relative contribution of various clinical variables to the risk of disease progression for LUTS/BPH as perceived by a panel of international experts. Although the absolute agreement on risk classification for different patient profiles was modest, the regression analysis made clear that there is good consensus that particular (combinations) of risk factors constitute a higher risk of disease progression than do others. It appears that the impact of considerable PVR (4150 ml), poor Qmax (o10 ml) and severe symptoms (total I-PSS 20–35) are the most dominant factors predicting an elevated risk of disease progression. Their impact seems to be merely cumulative, and no meaningful interaction effects were found that suggest that the simultaneous presence of particular risk factors has an additional effect. The only clinical trial that included all risk factors we took into consideration is the Medical Therapy Of Prostatic Symptoms (MTOPS) trial,4,6 comparing different types of medication with placebo in 3,047 LUTS/BPH patients for an average period of 4.5 y. Patients who received placebo (n ¼ 737) were more likely to experience clinical progression (rise of total I-PSS X4 points or the development of AUR, urinary incontinence, recurrent urinary tract infections/urosepsis or renal insufficiency) when they were older (X62 y), had a low Qmax (o10.6 ml/s), a large prostate volume (X31 ml), increased PSA (X1.6 ng/ml) and/or a high PVR (X39 ml). The development of AUR seemed to be affected by a large prostate volume and high PSA. Baseline symptom severity seemed to be important for symptomatic progression and the need of invasive therapy. Our results are partly in line with these figures. The most important difference is the fact that our study did not reveal age as an important risk factor. This may be due to the fact that age is strongly correlated with symptom severity. In addition, prostate volume and PSA appeared to be of less importance than PVR, symptoms and Qmax. This is probably to be ascribed to differences in the definitions used. The MTOPS study showed that prostate volume and PSA are mainly predictive for the risk of developing AUR, whereas the definition used in our study was much broader (including also symptomatic progression and need of invasive therapy). The results of our study are very much in line with questionnaires assessing the opinion of urologists with regard to risk factors of disease progression. These clearly show that a high PVR, poor Qmax and increased symptom severity, are considered to be of major importance for determining the risk of progression in individual LUTS/BPH patients.13,14 As such, these data can be seen as complimentary to the findings of clinical trials and epidemiological studies. However, the subjective nature of the risk assessments necessitates further research into the predictive value of risk profiles for
disease progression and into their relevance for treatment choice.
209
Acknowledgements This study was supported by an unrestricted educational grant from Boehringer Ingelheim GmbH.
References 1 Jacobsen SJ et al. Natural history of prostatism: risk factors for acute urinary retention. J Urol 1997; 158: 481–487. 2 McConnell JD et al. The effect of finasteride on the risk of acute urinary retention and the need for surgical treatment among men with benign prostatic hyperplasia. Finasteride Long-Term Efficacy and Safety Study Group. N Engl J Med 1998; 338: 557–563. 3 Roehrborn CG et al. Serum prostate-specific antigen concentration is a powerful predictor of acute urinary retention and need for surgery in men with clinical benign prostatic hyperplasia. PLESS Study Group. Urology 1999; 53: 473–480. 4 McConnell JD et al. Baseline measures predict the risk of benign prostatic hyperplasia clinical progression in placebo-treated patients. J Urol 2003; 169 (Suppl): 332 (Abstract 1287). 5 Roehrborn CG, Boyle P, Nickel JC. PSA is a significant predictor of objective parameters in men at risk of BPH progression. J Urol 2003; 169 (Suppl): 364–365 (Abstract 1362). 6 McConnell JD et al. The long-term effect of doxazosin, finasteride, and combination therapy on the clinical progression of benign prostatic hyperplasia. N Engl J Med 2003; 349: 2387–2398. 7 Jimenez-Cruz F. Identifying patients with lower urinary tract symptoms/benign prostatic hyperplasia (LUTS/BPH) at risk for progression. Eur Urol Suppl 2003; 2: 6–12. 8 Kaplan SA et al. Baseline symptoms, uroflow, and post-void residual urine as predictors of BPH clinical progression in the medically treated arms of the MTOPS trial. J Urol 2003; 169 (4 Suppl): 332–333 (Abstract 1289). 9 Speakman MJ et al. Integrating risk profiles in the treatment choice for patients with lower urinary tract symptoms/benign prostatic hyperplasia. A systematic analysis of expert opinion. Eur Urol Suppl 2004; 3: 61 (Abstract 233). 10 Brook RH et al. A method for the detailed assessment of the appropriateness of medical technologies. Int J Technol Assess Health Care 1986; 2: 53–63. 11 Meesen B. LUTS/BPH progression A literature review on risk factors for progression and impact of medical therapy on progression. Lier (Belgium), ISMAR Healthcare. 2003. 12 Lowe F et al. Risk factors for disease progression in patients with lower urinary tract symptoms/benign prostatic hyperplasia; a systematic analysis of expert opinion. Eur Urol Suppl 2004; 3: 39 (Abstract 147). 13 Djavan B, Nickel JC, de la Rosette J, Abrams P. The urologist view of BPH progression: results of an international survey. Eur Urol 2002; 41: 490–496. 14 McNicholas T. LUTS, the case is altered. Eur Urol Suppl 2002; 1: 28–35.
Prostate Cancer and Prostatic Diseases
