Urological Oncology BMI AND RCC HAFERKAMP et al.

The inuence of body mass index on the long-term survival of patients with renal cell carcinoma after tumour nephrectomy
Axel Haferkamp, Maria Pritsch*, Jens Bedke, Nina Wagener, Jesco Ptzenmaier, Stephan Buse and Markus Hohenfellner
Departments of Urology and *Medical Biometry, University of Heidelberg, Germany
Accepted for publication 27 September 2007

Study Type Prognosis (case series) Level of Evidence 4 OBJECTIVE To assess whether under- or overweight at the time of surgery has any effect on the survival of the patients with renal cell carcinoma (RCC), as obesity increases the risk of developing RCC. PATIENTS AND METHODS We prospectively evaluated 780 patients who had nephrectomy for RCC between 1990 and 2005. We used uni- and multivariate Cox proportional hazards models to assess the effect of body mass index (BMI), tumour

stage, Fuhrman grade, age, sex, histological type and performance status on cancerspecic survival (CSS). Patients were grouped according to BMI (in kg/m2), as underweight (<18.5), normal (18.5<25), overweight (25<30) and obese (30). RESULTS The median (range) follow-up was 5.3 (0.515.4) years, the patients being followed until June 2006; 254 patients died during the follow-up. Multivariate analyses of all patients showed that tumour stage, Fuhrman grade, Karnofsky performance status, age, sex and BMI were independent prognostic factors for CSS. While underweight patients had a signicantly worse prognosis than those of normal weight, overweight or obese patients

had a similar outcome to that of patients of normal weight. In a subgroup analyses including patients with localized RCC only, there was a strong tendency to less aggressive disease in the overweight group (P = 0.081). CONCLUSIONS Being underweight is an unfavourable and new risk factor for CSS in patients with RCC treated by nephrectomy. Although not signicant, there seems to be a limited favourable prognostic effect of overweight on CSS in patients with localized RCC. KEYWORDS renal cell carcinoma, body mass index, nephrectomy, long-term survival

INTRODUCTION Several epidemiological studies have consistently suggested that obesity is related to an increased risk of developing RCC in women and men [15]. The hypothetical reasons for this increased risk include higher levels of oestrogen, insulin and growth factors in adipose tissue, and hypertension and immune malfunction [6]. Although obesity appears to double the risk of developing RCC, some evidence showed that patients with RCC and a higher body mass index (BMI) at diagnosis have better survival than those with a lower BMI [711]. To date, the data supporting this inverse association remain somewhat questionable because of specic limitations in the existing studies. These include the retrospective design of the studies, the absence of data of important covariates, and the evaluation of subgroups only. In addition, all authors have focused on

overweight and obese patients only, and compared them with a group of combined normal and underweight patients (BMI < 25 kg/m2). No study of RCC has described the prognostic inuence of underweight. This is surprising, as Flegal et al. [12] showed that being underweight was associated with greater mortality than in those of normal weight in the National Health and Nutrition Examination Survey. Therefore the aim of the present study was to evaluate the prognostic inuence of BMI (in kg/m2), grouped as underweight (<18.5) , normal (18.5<25), overweight (25<30) and obese (30) on cancer-specic survival (CSS) of prospectively assessed patients with RCC.

radical nephrectomy at the authors institution between 1990 and 2005 were entered into a prospective database. The median (range) age of the patients was 61.6 (14.689.0) years; 319 were aged <60 years and 461 were 60 years, and 494 were male and 286 were female. According to the 1997 American Joint Committee on Cancer TNM staging system, the tumour stage was I in 412 patients (52.8%), II in 77 (9.9%), III in 141 (18.1%) and IV in 150 (19.2%). The tumour histological type and grade were determined according to the 1997 WHO classication and Fuhrmans criteria. The carcinomas were clear-cell in 646 specimens (86.1%), chromophilic in 70 (9.3%), chromophobic in 30 (4.0%) and collecting duct in four (0.5%), with 30 not classied by the pathologist. The Fuhrman grade was 1 in 186 specimens (24.0%), 2 in 431 (55.7%), 3 in 156 (20.3%) and 4 in one (0.1%), with six not classied by the pathologist. At the time of

PATIENTS AND METHODS In all, 780 patients with RCC who had no previous malignant tumour and who had a

2008 THE AUTHORS

JOURNAL COMPILATION

2 0 0 8 B J U I N T E R N A T I O N A L | 1 0 1 , 1 2 4 3 1 2 4 6 | doi:10.1111/j.1464-410X.2007.07375.x

1243

H A F E R K A M P ET AL.

TABLE 1 Univariate and multivariate analysis of prognostic risk factors for long-term CSS of all patients with RCC, and of those with localized RCC (stage I and II) HR (95% CI), P Univariate 1.42 (1.091.83), 0.008 1.68 (1.282.20), <0.001 2.63 (1.833.77), <0.001 3.82 (2.336.26), <0.001 6.39 (4.279.55), <0.001 25.5 (17.537.2), <0.001 1.90 (1.292.80), 0.001 7.46 (5.0211.1), <0.001 1.69 (1.092.62), 0.019 1.48 (0.544.03), 0.443 0.77 (0.591.02), 0.067 0.70 (0.481.01), 0.057

Variable Age (60 vs <60 years) Gender (male vs female) Karnofsky PS (80% vs <80%) Tumour stage II vs I III vs I IV vs I Fuhrman grade 2 vs 1 3 vs 1 Histological type (clear cell vs other) BMI Underweight vs normal Overweight vs normal Obese vs normal

Multivariate (721) 1.43 (1.081.89), 0.013 1.53 (1.142.05), 0.004 1.73 (1.162.56), 0.007 3.25 (1.945.46), <0.001 4.58 (2.947.13), <0.001 19.05 (12.528.9), <0.001 1.16 (0.771.75), 0.473 2.20 (1.433.40), <0.001 0.93 (0.591.48), 0.765 4.27 (1.4712.4), 0.008 1.00 (0.751.34), 0.989 1.11 (0.741.65), 0.613

Multivariate, localized RCC (456) 2.50 (1.424.39), 0.002 2.06 (1.183.60), 0.011 1.09 (0.373.22), 0.872 3.26 (1.875.67), <0.001

1.33 (0.732.43), 0.352 2.34 (1.025.38), 0.046 2.42 (0.956.12), 0.063 3.15 (0.7712.8), 0.110 0.60 (0.341.07), 0.081 0.68 (0.321.43), 0.306

surgery 714 patients (91.5%) had a Karnofsky performance status (PS) of 80% and 66 of <80%. The BMI was classied as a four-level categorical variable according to the USA National Institutes of Health (NIH); 10 patients were underweight at the time of surgery, 245 were normal, 361 were overweight and 141 were obese, with 23 not evaluated. The median (range) BMI was 26.6 (15.746.3) kg/m2. Of the 10 underweight patients, eight had stage I tumours, one stage II and one stage III. According to Kim et al. [13,14], ve of these patients had signs of cachexia, including weight loss (one), hypoalbuminaemia (three), anorexia (one) or malaise (one). Patients were prospectively evaluated every 3 months for the rst 2 years after treatment, every 6 months for the next 3 years, and yearly thereafter, with a chest X-ray or thoracic CT, abdominal ultrasonography or CT or MRI and serum chemistry. Patients were followed until June 2006. Survival was calculated from the date of nephrectomy. The survival endpoint was the date of the last follow-up or death. Kaplan-Meier estimates were used to describe survival rates, including point-wise asymptotic 95% CI. Patients who were conrmed to have died from other tumour were censored. Furthermore, assuming independence of the occurrence of RCC and other tumours in the same patient, patient survival was censored at the time of

occurrence of a second malignancy. The study was approved by the institutional review boards. The following clinical and pathological features were assessed for their prognostic relevance to the long-term survival the patients with RCC: age (60 vs <60 years), gender, Karnofsky PS (<80% vs 80%), tumour stage (IIIV vs I), Fuhrman grade (2 or 3 vs 1), histological type (clear cell vs other) and BMI (underweight, overweight or obesity vs normal). Univariate and multivariate analyses of prognostic factors were used within the Cox proportional hazards model. For each prognostic factor the hazard ratio (HR) in the univariate analysis and the adjusted HR in the multivariate analysis are given, including the 95% CI. In all tests, P < 0.05 was considered to indicate signicance. RESULTS The median (range) follow-up was 5.3 (0.515.4) years; to June 2006, 254 patients (32%) had died from their disease. The CSS rate (95% CI) at 5 years after surgery for all patients was 67.3 (63.771)%; it was 47.6 (11.084.3)% in underweight patients, 62.1 (55.368.7)% in those of normal weight, 69.8 (64.475.0)% in overweight and 70.5 (62.278.7)% in obese patients. On univariate analyses, the risk of dying from RCC for overweight patients was reduced to 75% of that in those of normal weight (HR 0.77, 0.591.02, P = 0.067), and to about two-

thirds for obese patients (HR 0.70, 0.481.01, P = 0.057). Being underweight had no signicant inuence on CSS in the univariate analysis (HR 1.48, 0.544.03, P = 0.443). Age, gender, Karnofsky PS, tumour stage, Fuhrman grade and clear cell subtype were signicant univariate prognostic factors of CSS in these patients. In the multivariate model of all patients, which included the univariate prognostic factors noted above and the four-level categorical variable BMI, the association between overweight or obesity and death from RCC was no longer evident. By contrast, being underweight at the time of surgery signicantly increased the HR by more than four times that of patients of normal weight. Age >60 years, male sex, Karnofsky PS of <80%, tumour stage and Fuhrman grade 3 remained signicant prognostic factors for dying from RCC in the multivariate model. The HRs, 95% CI and P values of the univariate and multivariate analyses for all patients are shown in Table 1. In a multivariate subgroup analysis of 456 patients with localized RCC, including tumour stage I and II, there was a strong tendency (P = 0.081) for being overweight to reduce the risk of death from RCC to 60% of that in patients of normal weight. In obese patients the risk reduction was similar, at 68%, but this was not signicant (P = 0.306). There was also a strong tendency for being underweight at surgery to worsen the patients prognosis 3.1 times. Age, gender, tumour stage, and

1244

JOURNAL COMPILATION

2008 THE AUTHORS

2008 BJU INTERNATIONAL

BMI AND RCC

Fuhrman grade remained signicant prognostic factors for dying from RCC in this multivariate subgroup model. Table 1 also shows the HRs, 95% CIs and P values for the multivariate analyses of patients with localized RCC. DISCUSSION Based on the ndings of several large epidemiological studies, obesity is considered a risk factor for developing RCC in adults [15]; in the present study we evaluated the prognostic inuence of BMI on the CSS of patients with RCC treated with nephrectomy. The results indicate that, despite a strong trend for a better prognosis (as long-term CSS) for overweight or obese patients in the univariate analyses, this effect was no longer present in the multivariate model, where other known prognostic factors, i.e. gender, age, Karnofsky PS, tumour stage, histological subtype and Fuhrman grade, were included. This result is in accordance with the retrospective studies [10,11]; Parker et al. [11] evaluated 970 patients with RCC and were unable to identify obesity (BMI 30 kg/m2) as a prognostic factor (HR 0.90, 0.651.23, P = 0.488) for CSS in their multivariate analysis, which also included the prognostic factors Mayo Clinic Stage, Size, Grade and Necrosis score, TNM stage groups, nuclear grade and tumour necrosis. They concluded that BMI offers little additional prognostic information beyond the accepted prognostic features. Donat et al. [10] evaluated 1137 patients with RCC who had had a radical or partial nephrectomy. In their multivariate analysis, only age >65 years, systemic symptoms, surgery type and pathological stage affected overall survival. While obesity (BMI 30 kg/m2) did not affect overall survival (HR 0.90, 0.621.30, P = 0.58), overweight (BMI 25<30) was almost signicant (HR 0.69, 0.481.00, P = 0.05). Although we were unable to conrm these results in the multivariate analysis of all the present patients, the ndings were similar, with a strong tendency to less aggressive disease in the overweight group, in the multivariate subgroup analysis including patients with localized RCC. Perhaps a BMI of 25<30 kg/m2 has a limited positive prognostic effect in this group of patients. As the long-term survival rate of patients with localized RCC is so high, more patients than included in the present study (456) need to be

evaluated to clearly identify being overweight as a positive prognostic factor in patients with localized RCC. Two studies [7,8] reported ndings indicating that overweight and/or obese patients had a more favourable prognosis than patients with a normal BMI. Yu et al. [7] evaluated 349 patients with RCC and available BMI data, and a median follow-up of 4.4 years. They reported a favourable prognosis in obese patients for disease-free survival (HR 0.43, 0.190.98) and overall survival (0.68, 0.38 1.22), but the study ndings were limited by the retrospective design of the study. A more contemporary review by Kamat et al. [8] of 400 patients with non-metastatic, node-negative RCC also showed a more favourable prognosis in overweight and obese patients than in those with a normal BMI for disease-free, overall and CSS. Unfortunately, that study also had some important limitations; it was retrospective, the median follow-up was relatively short (32 months) and the study did not control for other known prognostic factors, e.g. PS or histological subtype. Despite these limitations the results in that study underline our impression that being overweight at the time of surgery might have a limited positive prognostic effect in patients with localized disease. To our knowledge no other study has focused on the effect of being underweight (<18.5 kg/ m2) in the prognosis of patients with RCC. Most studies evaluating BMI in these patients dened normal differently from the NIH denition of as a BMI of <25 kg/m2, therefore including patients who were underweight [8,10,11]. In the present study we used the NIH denition and identied 10 patients with a BMI of <18.5 kg/m2 as an additional subgroup. Our results indicate that being underweight at the time of surgery worsened the prognosis of patients by more than four times. One reason for underweight can be cachexia, which is characterized by a complex, multilevel pathogenesis. It involves upregulated tissue catabolism and impaired anabolism, release of tumour-derived catabolic factors and inammatory cytokines, and neuroendocrine dysfunction. These culminate to create an energy-inefcient state characterized by wasting, chronic inammation, neuroendocrine dysfunction and anorexia [1517].

Cachexia-related ndings, including weight loss of 2.5 kg within 3 months, hypoalbuminaemia, malaise or anorexia, were previously identied as an independent predictor of survival in patients with RCC [13,14]. Kim et al. [13] reported a prognosis of disease-free survival of three times worse, and 4.4 times worse for CSS, in cachectic than in non-cachectic patients when evaluating 250 with T1N0M0 RCC. In the present group of 10 underweight patients, ve presented with cachexia-like symptoms, as described above; in these patients cachexia was most likely the reason for the low BMI, but cachexia-like symptoms did not account for the low BMI in the other ve patients. Several reasons for underweight other than tumour cachexia, including lifestyle and malnutrition, have been reported [18,19] and are probably the reason for being underweight in these ve asymptomatic patients. These patients can only be identied as being at greater risk of dying from RCC when their BMI is calculated. This indicates that the groups of underweight and tumour cachexia partly overlap, with some patients in both groups and others in just one group. Comparing the present study with published reports, there are advantages and limitations. As opposed to all other published reports, the present study was prospective, with a standardized follow-up after surgery for all patients. The present results also enhanced previous work in this area by evaluating a previously unstudied subgroup of BMI. Potential weaknesses include a possible selection bias associated with the referral pattern to a tertiary-care centre, and the relatively few patients compared to previous epidemiological studies. In conclusion, the present results indicate that BMI offers additional prognostic information beyond the accepted prognostic features. Being underweight (BMI < 18.5 kg/m2) represents a new and unfavourable risk factor for CSS in patients with RCC treated by nephrectomy. Although not statistically signicant, there was a limited but favourable prognostic effect of overweight (BMI 25<30 kg/m2) on CSS in patients with localized RCC. The importance and prognostic implications of this nding need to be addressed in a large multi-institutional pooled analysis.

2008 THE AUTHORS

JOURNAL COMPILATION

2008 BJU INTERNATIONAL

1245

H A F E R K A M P ET AL.

CONFLICT OF INTEREST None declared.

8 REFERENCES 1 Bergstrom A, Hsieh CC, Lindblad P, Lu CM, Cook NR, Wolk A. Obesity and renal cell cancer a quantitative review. Br J Cancer 2001; 85: 98490 Chow WH, Gridley G, Fraumeni JF Jr, Jarvholm B. Obesity, hypertension, and the risk of kidney cancer in men. N Engl J Med 2000; 343: 130511 van Dijk BA, Schouten LJ, Kiemeney LA, Goldbohm RA, van den Brandt PA. Relation of height, body mass, energy intake, and physical activity to risk of renal cell carcinoma: results from the Netherlands Cohort Study. Am J Epidemiol 2004; 160: 115967 Oh SW, Yoon YS, Shin SA. Effects of excess weight on cancer incidences depending on cancer sites and histologic ndings among men: Korea National Health Insurance Corporation Study. J Clin Oncol 2005; 23: 474254 Pischon T, Lahmann PH, Boeing H et al. Body size and risk of renal cell carcinoma in the European Prospective Investigation into Cancer and Nutrition (EPIC). Int J Cancer 2006; 118: 72838 Moyad MA. Obesity, interrelated mechanisms, and exposures and kidney cancer. Semin Urol Oncol 2001; 19: 270 9

10

11

12

13

14

Yu ML, Asal NR, Geyer JR. Later recurrence and longer survival among obese patients with renal cell carcinoma. Cancer 1991; 68: 164855 Kamat AM, Shock RP, Naya Y, Rosser CJ, Slaton JW, Pisters LL. Prognostic value of body mass index in patients undergoing nephrectomy for localized renal tumors. Urology 2004; 63: 4650 Schips L, Lipsky K, Zigeuner R et al. Does overweight impact on the prognosis of patients with renal cell carcinoma? A single center experience of 683 patients. J Surg Oncol 2004; 88: 5761 Donat SM, Salzhauer EW, Mitra N, Yanke BV, Snyder ME, Russo P. Impact of body mass index on survival of patients with surgically treated renal cell carcinoma. J Urol 2006; 175: 4652 Parker AS, Lohse CM, Cheville JC, Thiel DD, Leibovich BC, Blute ML. Greater body mass index is associated with better pathologic features and improved outcome among patients treated surgically for clear cell renal cell carcinoma. Urology 2006; 68: 7416 Flegal KM, Graubard BI, Williamson DF, Gail MH. Excess deaths associated with underweight, overweight, and obesity. JAMA 2005; 293: 18617 Kim HL, Han KR, Zisman A, Figlin RA, Belldegrun AS. Cachexia-like symptoms predict a worse prognosis in localized t1 renal cell carcinoma. J Urol 2004; 171: 18103 Kim HL, Belldegrun AS, Freitas DG et al. Paraneoplastic signs and symptoms of

15

16

17

18

19

renal cell carcinoma: implications for prognosis. J Urol 2003; 170: 17426 Saini A, Nasser AS, Stewart CE. Waste management cytokines, growth factors and cachexia. Cytokine Growth Factor Rev 2006; 17: 47586 Morley JE, Thomas DR, Wilson MM. Cachexia: pathophysiology and clinical relevance. Am J Clin Nutr 2006; 83: 735 43 George J, Cannon T, Lai V et al. Cancer cachexia syndrome in head and neck cancer patients. Part II. Pathophysiol Head Neck 2007; 29: 497507 Ali SM, Lindstrom M. Socioeconomic, psychosocial, behavioural, and psychological determinants of BMI among young women: differing patterns for underweight and overweight/obesity. Eur J Public Health 2006; 16: 32531 Hayashi F, Takimoto H, Yoshita K, Yoshiike N. Perceived body size and desire for thinness of young Japanese women: a population-based survey. Br J Nutr 2006; 96: 115462

Correspondence: Axel Haferkamp, Department of Urology, University of Heidelberg, INF 110, 69120 Heidelberg, Germany. e-mail: Axel.Haferkamp@med.uniheidelberg.de Abbreviations: BMI, body mass index; CSS, cancer-specic survival; HR, hazard ratio; PS, performance status; NIH, National Institutes of Health.

1246

JOURNAL COMPILATION

2008 THE AUTHORS

2008 BJU INTERNATIONAL