EUROPEAN UROLOGY 60 (2011)

Transcription

1 EUROPEAN UROLOGY 60 (2011) available at journal homepage: Platinum Priority Kidney Cancer Editorial by Simon P. Kim and R. Houston Thompson on pp of this issue A Stage-for-Stage and Grade-for-Grade Analysis of Cancer-Specific Mortality Rates in Renal Cell Carcinoma According to Age: A Competing-Risks Regression Analysis Maxine Sun a,y, *, Firas Abdollah b,y, Marco Bianchi a,b, Quoc-Dien Trinh a,c, Claudio Jeldres a,d, Zhe Tian a, Shahrokh F. Shariat e, Hugues Widmer d, Kevin Zorn d, Mani Menon c, Francesco Montorsi b, Paul Perrotte d, Pierre I. Karakiewicz a,d a Cancer Prognostics and Health Outcomes Unit, University of Montreal Health Center, Montreal, Canada; b Department of Urology, Vita Salute San Raffaele University, Milan, Italy; c Vattikuti Urology Institute, Henry Ford Health System, Detroit, MI, USA; d Department of Urology, University of Montreal Health Center, Montreal, Canada; e Department of Urology, Weill Medical College, Cornell University, New York, NY, USA Article info Article history: Accepted July 27, 2011 Published online ahead of print on August 5, 2011 Keywords: Renal cell carcinoma Age Cancer-specific mortality Nephrectomy Competing-risks regression Abstract Background: The association of advanced age and cancer control outcomes shows discordant findings. Objective: To evaluate the effect of age on cancer control outcomes in a large population-based cohort of patients diagnosed with renal cell carcinoma (RCC) of all stages. Design, setting, and participants: Using the Surveillance Epidemiology and End Results database, patients with RCC were identified. The population was stratified according to age: < 50, 50 59, 60 69, 70 79, and 80 yr. The effect of age on cancer control outcomes was evaluated using competing-risks regression models. Analyses were repeated stage for stage and grade for grade. Measurements: Cancer-specific mortality (CSM) was measured. Results and limitations: Age categories 50 59, 60 69, 70 79, and 80 yr respectively portended a 1.4-, 1.5-, 1.6-, and 1.9-fold higher risk of CSM than age category <50 yr (all p < 0.001). The effect of advanced age was particularly detrimental in patients with stage I disease: 1.8-, 2.3-, 3.2-, and 3.8-fold higher CSM risk for the same age groups, respectively (all p < 0.001). The effect of age on CSM was at its peak in patients with stage I, low-grade RCC (1.6-, 2.2-, 3.6-, and 4.3-fold, respectively; all p < 0.001) and remained elevated in stage I, high-grade RCC (2.2-, 2.6-, 2.4-, and 3.0-fold higher, respectively; all p < 0.05). Conversely, its effect was virtually absent in patients with stage II IV RCC. Conclusions: Our data suggest that stage I RCC may behave in a more aggressive fashion in elderly patients. Further studies are required to confirm the current findings. Crown Copyright # 2011 Published by Elsevier B.V. on behalf of European Association of Urology. All rights reserved. y Both authors contributed equally to the manuscript. * Corresponding author. Cancer Prognostics and Health Outcomes Unit, CHUM, 1058 rue St-Denis, Montreal, QC, H2X 3J4, Canada. Tel ext 35335; Fax: address: mcw.sun@umontreal.ca (M. Sun). 1. Introduction The incidence of renal cell carcinoma (RCC) continues to increase [1]. In the United States, approximately new cases were diagnosed in 2010 [2]. Despite the innovations in the treatment management of RCC in recent years, mortality rates have continued to rise [1,3]. It is known that the majority of newly diagnosed RCC patients are aged or yr. Given that the average life expectancy has steadily increased in the United States [4], it may be expected that more elderly patients will be diagnosed with RCC in the upcoming years /$ see back matter Crown Copyright # 2011 Published by Elsevier B.V. on behalf of European Association of Urology. All rights reserved. doi: /j.eururo

2 EUROPEAN UROLOGY 60 (2011) As with several other urologic malignancies, age represents an independent prognostic factor of worse survival in patients with RCC [5 10]. In a pertinent study, Hollingsworth et al. [10] examined the Surveillance Epidemiology and End Results (SEER) database to evaluate cancer-specific mortality (CSM) in patients according to age groups. Within that study, only surgically managed localized and regional RCC patients were included. The authors recorded an increasing risk of CSM with more advanced age in a univariable analysis. That being said, the effect of age was not examined with adjustment for other important tumor characteristics, such as stage and grade. The current study has several objectives. First, we examined RCC patients according to different age decades. Second, we sought to perform a stage-for-stage and grade-forgrade analysis of CSM rates according to age groups. Finally, since a large proportion of elderly patients may die of causes other than cancer, we used competing-risks methodology to account for this potential confounder in all our analyses. 2. Materials and methods 2.1. Data source The SEER program database, as reported by the US National Cancer Institute, was used to identify the study population. The SEER program collects patient demographics and publishes cancer incidence and survival data from population-based cancer registries, covering approximately 26% of the US population. Data from 1988 to 2006 from 17 SEER registries were abstracted Study population Individuals with histologically confirmed RCC were identified using the International Classification of Diseases for Oncology [ICD-O] (C67.0, C.67.9). Patients <25 yr were removed from the analysis. All autopsy or death certificate cases were excluded from the current study. Only patients with the main histologic subtypes (clear cell, papillary, chromophobe) were included. Patients with missing tumor (T) or node (N) stage information wereremoved aswellas patientswithnoinformation ondistant metastasis status. This method resulted in a total of RCC patients of all stages Description of covariates Patient age at diagnosis was classified according to decades, similar to a previous study: <50, 50 59, 60 69, 70 79, and 80 yr [10]. Other patient characteristics include race (white, black, other) and year of diagnosis tertiles ( , , ). Treatment type was categorized aspatientswho were surgicallymanaged(partialnephrectomy [PN] vs radical nephrectomy [RN]) or patients for whom surgery was not recommended. The latter categorization was identified via the absence of a concomitant code for RN or PN. Several clinical and pathologic characteristics were available within the SEER database, namely, tumor size (continuously coded), histologic subtype (clear cell, papillary, chromophobe), and Fuhrman grade (low grade [I II], high grade [III IV]). Table 1 Descriptive characteristics of patients with renal cell carcinoma, Surveillance Epidemiology and End Results, Variable <50 yr (n = 6603) yr (n = 8922) yr (n = 9877) yr (n = 8294) >80 yr (n = 2637) p value Gender <0.001 Male 4122 (62.5) 5908 (66.2) 6342 (64.2) 4988 (60.1) 1458 (55.3) Female 2481 (37.6) 3114 (34.9) 3535 (35.8) 3306 (39.9) 1179 (44.7) Race <0.001 White 5407 (81.9) 7398 (82.9) 8398 (85.0) 7231 (87.2) 2370 (89.9) Black 741 (11.2) 980 (11.0) 875 (8.9) 573 (6.9) 139 (5.3) Other 455 (6.9) 544 (6.1) 604 (6.1) 490 (5.9) 128 (4.9) Surgery type <0.001 Radical nephrectomy 5257 (79.6) 7326 (82.1) 8132 (82.3) 7064 (85.2) 2284 (86.6) Partial nephrectomy 1287 (19.5) 1504 (16.9) 1619 (16.4) 1071 (12.9) 238 (9.0) Nonsurgical management 59 (0.9) 92 (1.0) 126 (1.3) 159 (1.9) 115 (4.4) Histologic subtype <0.001 Clear cell 5841 (88.5) 7906 (88.6) 8732 (88.4) 7335 (88.4) 2273 (86.2) Chromophobe 276 (4.2) 256 (2.9) 248 (2.5) 225 (2.7) 102 (3.9) Papillary 486 (7.4) 760 (8.5) 897 (9.1) 734 (8.8) 262 (9.9) Tumor grade <0.001 I II 4943 (74.9) 6338 (71.0) 7051 (71.4) 5895 (71.1) 1825 (69.2) III IV 1660 (25.1) 2584 (29.0) 2826 (28.6) 2399 (28.9) 812 (30.8) T stage <0.001 T (67.6) 5729 (64.2) 6334 (64.1) 5339 (64.4) 1608 (61.0) T (17.2) 1348 (15.1) 1276 (12.9) 963 (11.6) 313 (11.9) T3 943 (14.3) 1753 (19.6) 2148 (21.7) 1896 (22.9) 678 (25.7) T4 57 (0.9) 92 (1.0) 119 (1.2) 96 (1.2) 38 (1.4) N stage N (96.3) 8594 (96.3) 9484 (96.0) 8050 (97.1) 2530 (95.9) N (3.7) 328 (3.7) 393 (4.0) 244 (2.9) 107 (4.1) M stage <0.001 M (97.6) 8572 (96.1) 9543 (96.6) 8100 (97.7) 2578 (97.8) M1 160 (2.4) 350 (3.9) 334 (3.4) 194 (2.3) 59 (2.2) AJCC stage <0.001 I 4408 (66.8) 5639 (63.2) 6247 (63.2) 5258 (63.4) 1574 (59.7) II 1077 (16.3) 1238 (13.9) 1184 (12.0) 902 (10.9) 295 (11.2) III 752 (11.4) 1417 (15.9) 1753 (17.7) 1683 (20.3) 597 (22.6) IV 366 (5.5) 628 (7.0) 693 (7.0) 451 (5.4) 171 (6.5) T = tumor; N = node; M = metastasis; AJCC = American Joint Committee on Cancer.

3 1154 EUROPEAN UROLOGY 60 (2011) The TNM 2002stagingsystem wasusedtoclassifypatientsaccording tothe American Joint Cancer Committee (AJCC) stages: I (T1N0M0), II (T2N0M0), III (T1 3N0 1M0), and IV (T4N0 1M0 1) [11]. The cause of death was defined according to the SEER-specific cause of death code (29010). Deaths from RCC were coded as CSM events. All other deaths were considered as other-cause mortality (OCM) Statistical analysis Frequencies and proportions were generated for categorical variables. We relied on the competing-risks regression methodology to assess CSM, as described by Fine and Gray [12]. The latter accounts for the effect of OCM and provides the most unbiased estimates of CSM. Age-stratified cumulative incidence CSM rates were generated for different groups and compared with the Gray test [13]. Subsequently, univariable and multivariable competing-risks regression models were used to test the effect of age (at <50, 50 59, 60 69, 70 79, and 80 yr) on CSM rates. Covariates included gender, race, year of diagnosis, tumor size, histologic subtype, AJCC stage, and Fuhrman grade. To assess the magnitude of the effect related to age, we repeated all multivariable competing-risks regression models after stratifying according to AJCC stage (stages I, II, III, and IV) and tumor grade (I II vs III IV). Finally, to reduce the bias associated with pathologic staging data, the aforementioned analyses were repeated within patients who underwent RN. Prior to fitting the competing-risks models, the proportional cause-specific hazards assumption, theproportionalityofthe hazards ofthecumulativeincidence function assumption, and the linearity assumption were tested [14]. All statistical tests were performed using the R statistical package (v ) or SPSS (Chicago, IL, USA). Finally, all tests were two-sided, with a significance level set at p < Results Overall, RCC patients were identified between 1988 and 2006 (Table 1). Several baseline characteristics differed according to age groups. For example, more females were aged 80 yr relative to those aged <50 yr (45% vs 38%; odds ratio [OR]: 1.35; p < 0.001). A lower proportion of black race was recorded in persons aged 80 yr (5% vs 11%; OR: 0.44; p < 0.001). Patients of different age groups also differed according to clinical and pathologic characteristics. In general, patients aged 80 yr had a more advanced T stage than patients aged <50 yr (T4: 1.4% vs 0.9%; OR: 1.68; p=0.01), a higher proportion of node-positive patients (4.1% vs 3.7%; OR: 1.1; p = 0.5), and more advanced Fuhrman grade (III IV: 31% vs 25%; OR: 1.33; p < 0.001). Patients aged yr had a higher rate of distant metastasis than patients aged <50 yr (3.4% vs 2.9%; OR: 1.6; p < 0.001). The number of deaths from CSM and OCM for the entire cohort, stratified according to age group and disease stage, is illustrated in Table 2. The overall respective 5- and 10-yr CSM rates, after accounting for OCM, were 8.7% and 12.5% for patients aged <50, 12.4% and 18.7% for patients aged 50 59, 13.2% and 20.7% for patients aged 60 69, 14.8% and 23.9% for patients aged 70 79, and 18.0% and 26.2% for patients aged 80 yr (all p < 0.001, Fig. 1). When cumulative incidence CSM rates were stratified according to AJCC stage, older age continuously portended to higher CSM rates than their younger counterparts (data not shown). The multivariable age-specific hazard ratios (HRs) predicting CSM in the entire cohort and stratified according to AJCC stage are depicted in Figure 2A 2E. Patients aged 50 59, 60 69, 70 79, and 80 yr demonstrated a higher rate of CSM than their counterparts aged <50 yr (HR: 1.4, 1.5, 1.6, and 1.9, respectively; all p < 0.001), even after accounting for OCM (Fig. 2A). The risk of CSM was 1.8-, 2.3-, 3.2-, and 3.8-fold higher, respectively, for the same age groups within stage I RCC (all p < 0.001; Fig. 2B); 1.4-, 1.4-, 1.6-, and 1.5-fold higher, respectively, for stage II RCC (all Table 2 Cause-of-death information within the entire population and stratified according to age group and renal cell carcinoma stage Overall <50 yr yr yr yr 80 yr Entire cohort Patients, no Total deaths, no Renal cell carcinoma Other-cause mortality Stage I Patients, no Total deaths, no Renal cell carcinoma Other-cause mortality Stage II Patients, no Total deaths, no Renal cell carcinoma Other-cause mortality Stage III Patients, no Total deaths, no Renal cell carcinoma Other-cause mortality Stage IV Patients, no Total deaths, no Renal cell carcinoma Other-cause mortality

4 [(Fig._1)TD$FIG] EUROPEAN UROLOGY 60 (2011) Fig. 1 Cumulative incidence plot depicting cancer-specific mortality (CSM) rates, stratified according to age categories (all Gray test p < 0.001) for (A) the entire population and for (B) stage I, (C) stage II, (D) stage III, and (E) stage IV renal cell carcinoma. p < 0.05; Fig. 2C); 1.1-, 1.1-, 1.3-, and 1.4-fold higher for stage III RCC (all p < 0.05, except for ages and yr; Fig. 2D); and 1.2-, 1.2-, 1.1-, and 1.4-fold higher respectively, for stage IV RCC (all p < 0.05, except for ages and yr; Fig. 2E). Similar findings were recorded in two subanalyses of patients with T1aN0M0 and T1 4N0M0 RCC. Specifically, the risk of CSM for the two groups was 1.5-, 3.0-, 4.8-, and 8.5-fold and 1.4-, 1.5-, 1.8-, and 2.1-fold higher in patients aged 50 59, 60 69, 70 79, and 80 yr, respectively (all p < 0.001). Figure 3A 3J illustrates the stage-for-stage and gradefor-grade age-specific HRs predicting CSM. Regardless of stage, more advanced age portended to higher CSM among patients with low- and high-grade disease. For example, relative to patients aged <50 yr, the risk of CSM was 1.4-, 1.6-, 2.1-, and 2.4-fold higher in patients with low-grade RCC, aged 60 69, 70 79, and 80 yr, respectively (all p < 0.001; Fig. 3A). Similarly, in patients with high-grade RCC, the respective HRs for the same age groups were 1.3, 1.3, 1.4, and 1.5, respectively (all p < 0.001, Fig. 3B). Within stage I, low-grade RCC (HR: 1.6, 2.2, 3.6, and 4.3) and highgrade RCC (HR: 2.2, 2.6, 2.4, and 3.0), more advanced age portended to higher CSM rates (all p 0.001; Fig. 3C and 3D). In stage II, more advanced age demonstrated significantly worse survival only in low-grade RCC (Fig. 3E). In all other stage and grade combinations, age failed to reach independent predictor status (Fig. 3G 3J). In a subanalysis of patients treated with RN, virtually the same results were recorded (Fig. 4). 4. Discussion Our objective was to examine the effect of age on CSM in patients with RCC of all stages, regardless of the treatment modality and after accounting for OCM. To date, the effect of age has never been studied using this type of appraisal. Previous age analyses were invariably performed on highly select patient subsets, in which the inclusion criteria relied on disease stage or treatment type, or both. These limitations undermined the generalizability of the findings. Most important, no study assessed the effect of age in a stage-for-stage and grade-for-grade fashion. In the current analysis, we attempted to circumvent these limitations. We included surgically and nonsurgically managed patients, as well as all disease stages. Finally, we used multivariable time-to-event modeling that controlled for other competing events. This consideration is essential in analyses addressing age, since a large proportion of elderly patients may succumb to OCM, which may critically confound the results. Our results first focused on the entire patient cohort. Subsequently, the data were stratified according to AJCC stage and tumor grade. Finally, the stratified analyses were repeated exclusively on patients treated with RN. The overall results showed that more advanced patient age predisposes to higher CSM rates, even after adjusting for all covariates, including that of OCM. Stage-specific analyses showed that the most substantial effect of advanced age on CSM was observed in individuals with stage I RCC.

5 1156 [(Fig._2)TD$FIG] EUROPEAN UROLOGY 60 (2011) [(Fig._3)TD$FIG] Fig. 2 Multivariable competing-risks regression predicting cancerspecific mortality (CSM) in patients with renal cell carcinoma of (A) all stages, (B) stage I, (C) stage II, (D) stage III, and (E) stage IV, stratified according to age categories. The x-axes denote the hazard ratios. The y- axes denote the age categories. The referent (ref.) category is patients aged <50 yr (hazard ratio: 1.0) across all examined subcohorts. The 95% confidence intervals are represented for each category. Furthermore, the effect of age was stronger in patients with low-grade RCC, regardless of stage. In other stages, the effect was substantially less. That being said, the effect of advanced age, although statistically significantly associated with worse survival, may be less pronounced in patients with stage III or IV RCC given the aggressive nature of the disease, beyond that of stage and grade, in younger patients, as was recently examined [15]. Fig. 3 Multivariable competing-risks regression predicting cancerspecific mortality (CSM) in patients with renal cell carcinoma of all stages, stage I, stage II, stage III, and stage IV, stratified according to low grade (I II) and high grade (III IV). The x-axes denote the hazard ratios. The y-axes denote the age categories. The referent (ref.) category is patients aged <50 yr (hazard ratio: 1.0) across all examined subcohorts. The 95% confidence intervals are represented for each category. Further stage-for-stage and grade-for-grade substratification revealed interesting findings. Specifically, the effect of advanced age in stage I patients is most prominent among individuals with low-grade tumors. It may be postulated that the detrimental effect of advanced age in individuals with low-grade disease was largely attributable to patients who were nonsurgically managed. Unfortunately, because

6 [(Fig._4)TD$FIG] EUROPEAN UROLOGY 60 (2011) Fig. 4 Multivariable competing-risks regression predicting cancerspecific mortality (CSM) in patients treated with radical nephrectomy for renal cell carcinoma of all stages, stage I, stage II, stage III, and stage IV, stratified according to low grade (I II) and high grade (III IV). The x-axes denote the hazard ratios. The y-axes denote the age categories. The referent (ref.) category is patients aged <50 yr (hazard ratio: 1.0) across all examined subcohorts. The 95% confidence intervals are represented for each category. of the low number of cases within the current cohort, we were unable to examine the effect of CSM in exclusively nonsurgically managed patients. Nonetheless, the detrimental effect of more advanced age remained highly important in surgically managed individuals with stage I low-grade tumors. Taken together, these data imply that the effect of more advanced age on CSM is particularly important in patients with stage I RCC and/or low-grade tumors. In contrast, the effect of age becomes secondary to disease characteristics in patients with tumors of more advanced stage (stage II IV) and/or high grade (III IV). These findings are particularly important given the rising incidence of localized RCC and increasing life expectancy of the population [16]. Several explanations may be proposed as to why the prognosis of elderly patients with low-stage and low-grade disease is significantly worse than that of their younger counterparts. First, the natural history of RCC may be more aggressive in the elderly. Second, elderly patients may be subject to a delay in treatment and suboptimal follow-up after diagnosis, particularly in the context of small renal masses, for which elderly patients may be more likely to be investigated or treated in a less timely fashion. Third, health insurance provider differences between elderly and younger patients may exist. Specifically, younger patients may be subject to periodic health examinations, whereas retired individuals may undergo less routine examinations. That said, given that the SEER database does not contain any information on tumor characteristics beyond stage and grade, and given that we did not have the time interval between diagnosis and treatment or health insurance information, the explanations proposed remain speculative. Furthermore, from a practical perspective, the current data suggest that active surveillance may not always be suitable in the elderly with stage I RCC, regardless of the grade of the tumor. This observation contradicts existing North American guidelines, in which older age is considered as a condition allowing less aggressive treatment management [17,18]. In contrast, the most up-to-date European guidelines do not account for the effect of age on prognosis in patients with RCC [19]. Currently, the rationale for active surveillance derives from the hypothesis that surgical intervention for small renal masses may not confer a survival advantage in elderly patients. For example, the Cleveland Clinic Foundation demonstrated a lack of survival benefit of both PN (HR: 0.67; 95% CI, ) and RN (HR: 0.75; 95% CI, ) relative to active surveillance in patients aged 75 yr ( p=0.2) [20]. Moreover, the majority of small renal masses may be benign or low-grade disease with indolent behavior [21]. However, existing series on active surveillance are limited by small sample size and short follow-up, and they originate from single-institution data. Furthermore, the current study as well as previously published reports showed that more advanced age represents a strong risk factor for CSM in patients with localized RCC [22,23]. However, more advanced age is a predictor for surveillance [20]. Although existing paradigms suggest that there is overtreatment in localized RCC, undertreatment represents an equally important consideration. Consequently, at least in some elderly patients with localized RCC, the potential for curative disease management should not be denied. More historical prognostic models for prediction of CSM do not incorporate age. Specifically, the University of California, Los Angeles, system opted for the inclusion of performance status in its model over age [24]. Similarly, the Mayo Clinic series relies on disease stage, tumor grade, and necrosis to generate a score to predict outcome [25]. A postoperative nomogram for prediction of recurrence also fails to account for age, as well as performance status [26]. However, more contemporary models have revealed important prognostication for patient age at diagnosis.

7 1158 EUROPEAN UROLOGY 60 (2011) For example, Karakiewicz et al. [9] developed a highly accurate preoperative nomogram predicting CSM-free survival using age, sex, symptoms, tumor size, tumor stage, and distant metastasis status. In summary, there is a growing but nonetheless discordant body of literature focusing on the relationship between age and CSM in RCC. Some studies have shown an association between young age and survival. For example, Sánchez-Ortiz et al. [27] recorded a survival benefit in patients aged 40 yr relative to those aged yr. Others recorded a 5-yr CSM-free survival benefit in younger patients (30 39 yr) versus older patients (70 79 yr, p=0.01) for localized RCC [5]. Jung et al. [28], as well as Komai et al. [7], recently corroborated these findings within two single institutional reports. Conversely, Thomson et al. [29] and Gillett et al. [30] reported a lack of statistically significant difference between young and old individuals in regard to CSM. Similarly, Hollingsworth et al. [10] failed to detect a significant difference for CSM for individuals with advanced age using univariable competing-risks analyses. Our manuscript is not without limitations. First, the SEER database does not allow the differentiation between sporadic versus hereditary RCC. This information is important, given that when RCC is diagnosed in patients aged <50 yr, a hereditary case may be suspected. Since the natural history of RCC in von Hippel-Lindau disease is different between hereditary and sporadic RCC [31], this differentiation may have provided important insights into the difference in prognosis according to age. Second, the version of the SEER database used does not contain information on baseline comorbidity status or performance status, both of which are important in the selection of surgical candidates. The lack of comorbidities represents an important limitation to the current study. The combination of age and comorbidities allows the estimation of life expectancy, which is an essential proxy for determining the survival outcomes of patients with RCC [16]. The use of competing-risks regression analysis may have compensated for the lack of baseline comorbidity and performance status. Nonetheless, the sole adjustment for OCM remains insufficient. Future studies examining the effect of age on survival should attempt to incorporate this information. Third, information on other minimally invasive treatment approaches (eg, percutaneous radio frequency ablation, cryoablation, microwave ablation, laser ablation, and high-intensity focused ultrasound ablation), immunotherapy, and targeted therapy was not available within the current database. This limitation may have induced a selection bias. Nonetheless, previous investigators have shown that thermal ablation was low (3.8%) in the SEER database [32]. Moreover, older patients were more likely to receive such treatment relative to their younger counterparts. Consequently, worse survival in elderly patients may not be entirely explained according to such bias. Fourth, we were unable to examine the effect of age on patients who were exclusively nonsurgically managed because of the small sample size. Existing literature on the role of active surveillance remains controversial. No study examining active surveillance and survival with adequate sample size and sufficient follow-up time exists. Fifth, the lack of information on adjuvant therapies likely affected survival data, especially in patients with more advanced disease stage. Sixth, as in all retrospective observational data, a coding misattribution with regard to cause of death may have been operational. However, this misattribution should be equally applicable to all patients, regardless of their age or stage of disease. Thus the limitation cannot completely explain the observed effect, where more advanced age portended to a worse survival, particularly in patients with low-stage/low-grade disease. Similarly, lack of central pathology, other coding misattribution, and entry error may have been operational. However, all previous studies examining population-based cohorts were also affected by this limitation [10,23,33]. 5. Conclusions The current data show that increasing age is an important prognostic factor for worse survival outcomes in patients with low-stage, low-grade RCC, despite adjusting for available covariates, including OCM. However, given the lack of information on comorbidities and performance status, the current study cannot recommend an undiscriminated surgical approach for all elderly patients with a localized disease. Nonetheless, further studies are necessary to confirm or refute the current findings. Author contributions: Maxine Sun had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Sun, Abdollah, Bianchi, Karakiewicz. Acquisition of data: Sun, Tian. Analysis and interpretation of data: Sun, Abdollah, Trinh, Jeldres, Karakiewicz. Drafting of the manuscript: Sun, Abdollah, Karakiewicz. Critical revision of the manuscript for important intellectual content: Jeldres, Shariat, Widmer, Zorn, Menon, Montorsi, Perrotte, Karakiewicz. Statistical analysis: Sun, Tian. Obtaining funding: Perrotte, Karakiewicz. Administrative, technical, or material support: Perrotte, Karakiewicz. Supervision: Shariat, Menon, Montorsi, Perrotte, Karakiewicz. Other (specify): None. Financial disclosures: I certify that all conflicts of interest, including specific financial interests and relationships and affiliations relevant to the subject matter or materials discussed in the manuscript (eg, employment/affiliation, grants or funding, consultancies, honoraria, stock ownership or options, expert testimony, royalties, or patents filed, received, or pending), are the following: Pierre I. Karakiewicz is partially supported by the University of Montreal Health Centre Urology Specialists, Fonds de la Recherche en Santé du Québec, the University of Montreal Department of Surgery and the University of Montreal Health Centre (CHUM) Foundation. Funding/Support and role of the sponsor: None. References [1] Sun M, Thuret R, Abdollah F, et al. Age-adjusted incidence, mortality, and survival rates of stage-specific renal cell carcinoma in North America: a trend analysis. Eur Urol 2011;59:

8 EUROPEAN UROLOGY 60 (2011) [2] Jemal A, Bray F, Center M, et al. Global cancer statistics. CA Cancer J Clin 2011;61: [3] Hollingsworth JM, Miller DC, Daignault S, et al. Rising incidence of small renal masses: a need to reassess treatment effect. J Natl Cancer Inst 2006;98: [4] U.S. interim projections by age, sex, race, and Hispanic origin: US Census Bureau Web site. gov/population/www/projections/usinterimproj. [5] Scoll BJ, Wong Y-N, Egleston BL, et al. Age, tumor size and relative survivalofpatientswithlocalizedrenal cellcarcinoma: a Surveillance, Epidemiology and End Results analysis. J Urol 2009;181: [6] Muramaki M, Miyake H, Sakai I, et al. Age at diagnosis as a powerful predictor for disease recurrence after radical nephrectomy in Japanese patients with pt1 renal cell carcinoma. Int J Urol 2011;18: [7] Komai Y, Fujii Y, Iimura Y, et al. Young age as favorable prognostic factor for cancer-specific survival in localized renal cell carcinoma. Urology 2011;77: [8] Karakiewicz PI, Jeldres C, Suardi N, et al. Age at diagnosis is a determinant factor of renal cell carcinoma-specific survival in patients treated with nephrectomy. Can Urol Assoc J 2008;2: [9] Karakiewicz PI, Suardi N, Capitanio U, et al. A preoperative prognostic model for patients treated with nephrectomy for renal cell carcinoma. Eur Urol 2009;55: [10] Hollingsworth JM, Miller DC, Daignault S, et al. Five-year survival after surgical treatment for kidney cancer: a population-based competing risk analysis. Cancer 2007;109: [11] Sobin L, Gospodarowicz M, Wittekind C, editors. International Union Against Cancer (UICC). TNM classification of malignant tumors. ed 6. Oxford, UK: Wiley-Blackwell; [12] Fine J, Gray R. A proportional hazards model for the subdistribution of a competing risk. J Am Stat Assoc 1999;94: [13] Gray R. A class of K-sample tests for comparing the cumulative incidence of a competing risk. Ann Stat 1988;16: [14] Pintilie M. Competing risks: a practical perspective. Hoboken, NJ: John Wiley & Sons; [15] Bianchi M, Sun M, Jeldres C, et al. Distribution of metastatic sites in renal cell carcinoma: a population-basedanalysis. Ann Oncol. Inpress. [16] Walter LC, Covinsky KE. Cancer screening in elderly patients: a framework for individualized decision making. JAMA 2001;285: [17] Campbell S, Novick A, Belldegrun A, et al. Guideline for management of the clinical T1 renal mass. J Urol 2009;182: [18] Kidney cancer (cited V ). National Comprehensive Cancer Network Web site. gls/f_guidelines.asp#site. [19] Ljungberg B, Cowan NC, Hanbury DC, et al. EAU guidelines on renal cell carcinoma: the 2010 update. Eur Urol 2010;58: [20] Lane BR, Abouassaly R, Gao T, et al. Active treatment of localized renal tumors may not impact overall survival in patients aged 75 years or older. Cancer 2010;116: [21] Volpe A, Cadeddu JA, Cestari A, et al. Contemporary management of small renal masses. Eur Urol 2011;60: [22] Lughezzani G, Sun M, Budäus L, et al. Population-based external validation of a competing-risks nomogram for patients with localized renal cell carcinoma. J Clin Oncol 2010;28: [23] Kutikov A, Egleston BL, Wong Y-N, et al. Evaluating overall survival and competing risks of death in patients with localized renal cell carcinoma using a comprehensive nomogram. J Clin Oncol 2010; 28: [24] Zisman A, Pantuck AJ, Wieder J, et al. Risk group assessment and clinical outcome algorithm to predict the natural history of patients with surgically resected renal cell carcinoma. J Clin Oncol 2002; 20: [25] Thompson RH, Leibovich BC, Lohse CM, et al. Dynamic outcome prediction in patients with clear cell renal cell carcinoma treated with radical nephrectomy: the D-SSIGN score. J Urol 2007;177: [26] Kattan MW, Reuter V, Motzer RJ, et al. A postoperative prognostic nomogram for renal cell carcinoma. J Urol 2001;166:63 7. [27] Sánchez-Ortiz RF, Rosser CJ, Madsen LT, et al. Young age is an independent prognostic factor for survival of sporadic renal cell carcinoma. J Urol 2004;171: [28] Jung E-J, Lee HJ, Kwak C, et al. Young age is independent prognostic factor for cancer-specific survival of low-stage clear cell renal cell carcinoma. Urology 2009;73: [29] Thompson RH, Ordonez MA, Iasonos A, et al. Renal cell carcinoma in young and old patients is there a difference? J Urol 2008;180: , discussion [30] Gillett MD, Cheville JC, Karnes RJ, et al. Comparison of presentation and outcome for patients 18 to 40 and 60 to 70 years old with solid renal masses. J Urol 2005;173: [31] Clark PE, Cookson MS. The von Hippel-Lindau gene: turning discovery into therapy. Cancer 2008;113: [32] Choueiri TK, Schutz FAB, Hevelone ND, et al. Thermal ablation vs surgery for localized kidney cancer: a Surveillance, Epidemiology, and End Results (SEER) database analysis. Urology 2011;78: [33] Lin DW, Porter M, Montgomery B. Treatment and survival outcomes in young men diagnosed with prostate cancer: a populationbased cohort study. Cancer 2009;115: