Decreased Overall and Bladder Cancer–Specific Mortality with Adjuvant Chemotherapy After Radical Cystectomy: Multivariable Competing Risk Analysis

EURURO-6294; No. of Pages 4 EUROPEAN UROLOGY XXX (2015) XXX–XXX

available at www.sciencedirect.com journal homepage: www.europeanurology.com

Brief Correspondence

Decreased Overall and Bladder Cancer–Specific Mortality with Adjuvant Chemotherapy After Radical Cystectomy: Multivariable Competing Risk Analysis Michael Froehner a,*, Rainer Koch b, Ulrike Heberling a, Vladimir Novotny a, Sven Oehlschlaeger a, Matthias Hu¨bler c, Gustavo B. Baretton d, Oliver W. Hakenberg e, Manfred P. Wirth a a

Department of Urology, University Hospital ‘‘Carl Gustav Carus,’’ Technische Universita¨t Dresden, Dresden, Germany; b Department of Medical Statistics and

Biometry, University Hospital ‘‘Carl Gustav Carus,’’ Technische Universita¨t Dresden, Dresden, Germany; c Department of Anesthesiology, University Hospital ‘‘Carl Gustav Carus,’’ Technische Universita¨t Dresden, Dresden, Germany;

d

Department of Pathology, University Hospital ‘‘Carl Gustav Carus,’’ Technische

Universita¨t Dresden, Dresden, Germany; e Department of Urology, University of Rostock, Rostock, Germany

Article info

Abstract

Article history: Accepted June 29, 2015

Adding chemotherapy to radical cystectomy (RC) may improve outcome. Neoadjuvant treatment is advocated by guidelines based on meta-analysis data but is severely underused in clinical practice. Adjuvant treatment of patients at risk could be an alternative. We analyzed a sample of 798 patients who underwent RC between 1993 and 2011 for high-risk superficial or muscle-invasive urothelial or undifferentiated bladder cancer, of which 23% received adjuvant cisplatin-based chemotherapy and %5 received neoadjuvant chemotherapy. The use of adjuvant chemotherapy was an independent predictor of decreased overall mortality (hazard ratio [HR]: 0.50; 95% confidence interval [CI], 0.38–0.66; p < 0.0001) and bladder cancer–specific mortality (HR: 0.71; 95% CI, 0.52–0.97; p = 0.0321), but it was not associated with competing mortality. Similar figures were obtained when analyzing the number of cisplatin-containing cycles administered or when restricting the analysis to patients with lymph node–positive or extravesical but lymph node–negative disease, suggesting a mortality-reducing treatment effect after adjusting for several patient- and tumor-related confounders. Future trials should directly compare the concepts of neoadjuvant and adjuvant application of chemotherapy in candidates for RC. Patient summary: Adjuvant chemotherapy may decrease overall and bladder cancer– specific mortality after radical cystectomy (RC). Future trials should directly compare the concepts of neoadjuvant and adjuvant application of chemotherapy in candidates for RC. # 2015 European Association of Urology. Published by Elsevier B.V. All rights reserved.

Associate Editor: James Catto Keywords: Urologic neoplasms Comorbidity Cisplatin Adjuvant Chemotherapy Mortality Bladder Cystectomy Competing risk analysis Proportional hazards model

* Corresponding author. Department of Urology, University Hospital ‘‘Carl Gustav Carus,’’ Technische Universita¨t Dresden, Fetscherstrasse 74, D-01307 Dresden, Germany. Tel. +49 351 4587462; Fax: +49 351 4584333. E-mail address: [email protected] (M. Froehner).

Adjuvant chemotherapy after radical cystectomy (RC) is controversial. In contrast to neoadjuvant chemotherapy, the 2013 version of the guidelines on muscle-invasive and metastatic bladder cancer (BCa) did not recommend the use of adjuvant chemotherapy outside of clinical trials [1].

Nevertheless, in clinical practice, only a minority of patients actually receive neoadjuvant treatment [2–4], and adjuvant treatment is used considerably more frequently [2,3]. We studied 798 of 871 consecutive patients with high-risk superficial or muscle-invasive urothelial or undifferentiated

http://dx.doi.org/10.1016/j.eururo.2015.06.053 0302-2838/# 2015 European Association of Urology. Published by Elsevier B.V. All rights reserved.

Please cite this article in press as: Froehner M, et al. Decreased Overall and Bladder Cancer–Specific Mortality with Adjuvant Chemotherapy After Radical Cystectomy: Multivariable Competing Risk Analysis. Eur Urol (2015), http://dx.doi.org/10.1016/ j.eururo.2015.06.053

EURURO-6294; No. of Pages 4 2

EUROPEAN UROLOGY XXX (2015) XXX–XXX

BCa who underwent RC for between January 1993 and December 2011 at our institution. Patients had complete data on age (continuous variable); gender; Charlson score (continuous variable); American Society of Anesthesiologists (ASA) physical status classification (classes 3–4 vs 1–2); number of lymph nodes (LNs) removed (continuous variable); local tumor stage (bladder confined vs extravesical disease [ie, invasion beyond the detrusor muscle]); LN status (positive vs negative); LN density (continuous variable); any neoadjuvant chemotherapy (yes vs no); adjuvant cisplatinbased chemotherapy (yes vs no); and, alternatively, the number of cycles of adjuvant cisplatin-based chemotherapy (continuous variable). The median age was 69 yr, and median follow-up of the censored patients was 7.1 yr (interquartile range: 4.3–11.4 yr). Eligible patients with extravesical and/or node-positive BCa were routinely offered adjuvant cisplatin-based chemotherapy after RC. After gemcitabine became available as a combination partner of cisplatin, if feasible, four and six cycles were striven for in patients with extravesical LNnegative and LN-positive disease, respectively. Treatment was reduced or stopped in cases of unacceptable toxicity or at the patient’s request. Chemotherapy was considered adjuvant when at least one cisplatin-containing cycle was given prior to documented tumor progression after RC. The cumulative incidence of deaths from BCa was determined by univariate and multivariable competing risk analysis. Cox proportional hazards models for competing risks according to Fine and Gray were used to study combined effects of the variables on overall, competing, and BCaspecific mortality. The analyses were done with the SAS v.9.4 statistical package (SAS Institute, Cary, NC, USA). Patients who had received adjuvant chemotherapy were somewhat younger (mean age: 66 vs 68 yr; p < 0.0001) and healthier (Charlson score 2: 31% vs 40%; p = 0.0199; ASA classes 3–4: 37% vs 41%; p = 0.34) and had somewhat more LNs removed (20 vs 17; p = 0.0218). Adjuvant chemotherapy consisted of gemcitabine and cisplatin (75%), methotrexate and cisplatin with or without epirubicine and vinblastine (22%), or other combinations with cisplatin (3%). Further demographic data are shown in Supplementary Table 1. In the univariate analysis, adjuvant chemotherapy was associated with increased BCa-specific mortality (Fig. 1) and overall mortality (10-yr overall mortality: 65.0% vs 55.5%; Mantel-Haenszel hazard ratio [HR]: 1.31; 95% confidence interval [CI], 1.04–1.64; p = 0.0210; other univariate HRs not shown). In the multivariable analysis shown in Table 1, the opposite effect was seen, with decreased overall and BCa-specific mortality risks (HRs <1) in patients who received adjuvant chemotherapy. Controlling for multiple confounders, particularly for tumor-related risk factors, may explain this HR conversion. Substituting the variable adjuvant chemotherapy by the number of given cisplatin-based cycles or restricting the analysis to patients with LN-positive disease or with extravesical but LN-negative disease gave similar results (Supplementary Table 2–4). No relationship between adjuvant cisplatin-based chemotherapy and competing mortality was observed. Because neoadjuvant chemotherapy was

Fig. 1 – Cumulative bladder cancer–specific mortality curves stratified by application of adjuvant cisplatin-based chemotherapy in the whole study sample and in patients with positive lymph nodes and extravesical lymph node–negative disease, respectively (univariate competing risk analysis).

disadvantaged by application to patients with particularly poor risks, it was associated with adverse outcome even after controlling for confounders (Table 1; Supplementary Table 2–4). Although patients who received adjuvant chemotherapy were somewhat younger and/or healthier, the consistent identification of adjuvant chemotherapy as an independent predictor of reduced mortality in various models may most likely be explained, at least in part, by a true treatment effect. These observations support the concept of offering adjuvant chemotherapy to high-risk patients after RC, particularly since they were made in a largely unselected sample with a meaningful proportion of patients who were unfit or unwilling to undergo chemotherapy.

Please cite this article in press as: Froehner M, et al. Decreased Overall and Bladder Cancer–Specific Mortality with Adjuvant Chemotherapy After Radical Cystectomy: Multivariable Competing Risk Analysis. Eur Urol (2015), http://dx.doi.org/10.1016/ j.eururo.2015.06.053

EURURO-6294; No. of Pages 4 3

EUROPEAN UROLOGY XXX (2015) XXX–XXX

Table 1 – Optimal Cox proportional hazards models for overall, bladder cancer–specific, and competing mortality in the whole sample using the categorical variable adjuvant cisplatin-based chemotherapy Category End point: Overall mortality Age (continuous variable, per-year increase) Charlson score (continuous variable, per-unit increase) ASA classes 3–4 (vs 1–2) Extravesical disease (vs bladder confined disease) Positive lymph nodes (vs negative lymph nodes) Lymph node density (continuous variable, per-10% increase) Any neoadjuvant chemotherapy (vs none) Adjuvant cisplatin-based chemotherapy (vs none) End point: Bladder cancer–specific mortality Extravesical disease (vs bladder confined disease) Positive lymph nodes (vs negative lymph nodes) Any neoadjuvant chemotherapy (vs none) Adjuvant cisplatin-based chemotherapy (vs none) End point: Competing mortality Age (continuous variable, per-year increase) Charlson score (continuous variable, per-unit increase) ASA classes 3–4 (vs 1–2) Positive lymph nodes (vs negative lymph nodes)

HR

95% CI

p

1.02 1.12 1.64 2.11 2.24 1.14 1.87 0.50

1.01–1.04 1.06–1.18 1.32–2.04 1.69–2.62 1.64–2.62 1.07–1.21 1.28–2.75 0.38–0.66

0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 0.0014 <0.0001

2.10 3.83 2.47 0.71

1.58–2.79 2.81–5.22 1.54–3.97 0.52–0.97

<0.0001 <0.0001 0.0002 0.0321

1.04 1.16 1.86 0.49

1.02–1.06 1.07–1.25 1.33–2.61 0.32–0.74

<0.0001 0.0001 0.0003 0.0006

ASA = American Society of Anesthesiologists; CI, confidence interval; HR, hazard ratio. The decreased competing mortality in patients with positive lymph nodes may be explained by the high level of bladder cancer mortality in this subgroup. When the analysis was restricted to patients without neoadjuvant chemotherapy, adjuvant chemotherapy was also an independent predictor of decreased overall (HR: 0.44; 95% CI, 0.32–0.59; p < 0.0001) and bladder cancer–specific mortality (HR: 0.67; 95% CI, 0.48–0.94; p = 0.0204).

In comparable studies investigating patients with positive LNs, adjuvant chemotherapy was identified as an independent predictor of lower BCa-specific mortality [5], and similar cumulative mortality curves were observed [6]. In a population-based study, a similar hazard ratio for the impact of adjuvant chemotherapy on BCa-specific mortality (HR: 0.73; 95% CI, 0.64–0.84 [3]) but a somewhat lower impact on overall mortality (HR: 0.71; 95% CI, 0.62– 0.81 [3]) was seen. A somewhat greater impact of adjuvant chemotherapy on BCa-specific mortality (HR: 0.56; 95% CI: 0.39–0.80) was observed in 756 consecutive cystectomy patients [7]. As in a recently published randomized trial [8], when stratified by the application of adjuvant chemotherapy in this study, the mortality curves separated more clearly for patients with extravesical but LN-negative disease rather than for those with LN-positive disease (Fig. 1). In patients with extravesical but LN-negative disease in this randomized trial, the HR (0.37; adjusted 95% CI, 0.16–0.83 [8]) was similar to this study (Supplementary Table 4) but higher in patients with LN-positive disease (HR: 0.94; 95% CI, 0.65–1.34 [8]). This study has several limitations. Up to >20 years after treatment, it was difficult to determine how much adjuvant chemotherapy was actually given, and no data on toxicity or treatment after progression were obtainable. The complexity of data acquisition was prone to classification errors (even though most likely not relevant for interpretation because of the relatively large sample size). Inclusion of more factors or exclusion of others could produce different models. Finally, despite controlling for several patient-, tumor-, and treatment-related parameters, it could not be ruled out that further hidden confounders influenced the results. In conclusion, in this study, adjuvant cisplatin-based chemotherapy offered to patients at risk was associated with decreased overall and BCa-specific mortality after

adjusting for several patient- and tumor-related confounders in a routine clinical setting, with a greater effect in extravesical, node-negative disease. Future trials should directly compare the concepts of neoadjuvant and adjuvant application of chemotherapy in candidates for RC. Author contributions: Michael Froehner 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: Froehner. Acquisition of data: Froehner, Novotny, Heberling, Hu¨bler, Baretton. Analysis and interpretation of data: Koch, Wirth, Froehner. Drafting of the manuscript: Froehner, Koch, Heberling, Novotny, Oehlschlaeger, Hu¨bler, Baretton, Hakenberg, Wirth. Critical revision of the manuscript for important intellectual content: Froehner, Koch, Heberling, Novotny, Oehlschlaeger, Hu¨bler, Baretton, Hakenberg, Wirth. Statistical analysis: Koch, Froehner. Obtaining funding: None. Administrative, technical, or material support: Wirth. Supervision: Wirth. Other (specify): None. Financial disclosures: Michael Froehner certifies 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: None. Funding/Support and role of the sponsor: None.

Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j. eururo.2015.06.053.

Please cite this article in press as: Froehner M, et al. Decreased Overall and Bladder Cancer–Specific Mortality with Adjuvant Chemotherapy After Radical Cystectomy: Multivariable Competing Risk Analysis. Eur Urol (2015), http://dx.doi.org/10.1016/ j.eururo.2015.06.053

EURURO-6294; No. of Pages 4 4

EUROPEAN UROLOGY XXX (2015) XXX–XXX

References

[6] Lucca I, Roupreˆt M, Kluth L, et al. Adjuvant cisplatin-based combination chemotherapy for lymph node-positive urothelial carcino-

[1] Witjes JA, Compe´rat E, Cowan NC, et al., European Association of

ma of the bladder following radical cystectomy: a retrospective

Urology. EAU guidelines on muscle-invasive and metastatic bladder

international study of more than 1500 patients. BJU Int 2015;115:

cancer: summary of the 2013 guidelines. Eur Urol 2014;65:778–92.

722–7.

[2] David KA, Milowsky MI, Ritchey J, Carroll PR, Nanus DM. Low inci-

[7] Kwon T, Jeong IG, Lee J, et al. Adjuvant chemotherapy after radical

dence of perioperative chemotherapy for stage III bladder cancer

cystectomy for bladder cancer: a comparative study using inverse-

1998 to 2003: a report from the National Cancer Data Base. J Urol

probability-of-treatment weighting. J Cancer Res Clin Oncol 2015;

2007;178:451–4.

141:169–76.

[3] Booth CM, Siemens DR, Li G, et al. Perioperative chemotherapy for

[8] Sternberg CN, Skoneczna I, Kerst JM, et al., European Organisation

muscle-invasive bladder cancer: a population-based outcomes

for Research and Treatment of Cancer Genito-Urinary Cancers Group; Groupe d’Etude des Tumeurs Uroge´nitales; National Can-

study. Cancer 2014;120:1630–8. [4] Leow JJ, Martin-Doyle W, Rajagopal PS, et al. Adjuvant chemother-

cer Research Institute Bladder Cancer Study Group; National

apy for invasive bladder cancer: a 2013 updated systematic review

Cancer Institute of Canada Clinical Trials Group; German Associa-

and meta-analysis of randomized trials. Eur Urol 2014;66:42–54.

tion of Urologic Oncology (AUO). Immediate versus deferred

[5] Clifton MM, Psutka SP, Boorjian SA, et al. Cancer-specific mortality

chemotherapy after radical cystectomy in patients with pT3–pT4

following radical cystectomy for bladder cancer with lymph node

or N+ M0 urothelial carcinoma of the bladder (EORTC 30994): an

involvement: impact of pathologic disease features and adjuvant

intergroup, open-label, randomised phase 3 trial. Lancet Oncol 2015;

chemotherapy. World J Urol 2015;33:373–9.

16:76–86.

Please cite this article in press as: Froehner M, et al. Decreased Overall and Bladder Cancer–Specific Mortality with Adjuvant Chemotherapy After Radical Cystectomy: Multivariable Competing Risk Analysis. Eur Urol (2015), http://dx.doi.org/10.1016/ j.eururo.2015.06.053