Smoking negatively impacts renal cell carcinoma overall and cancer-specific survival

Authors

  • Nils Kroeger MD,

    1. Institute of Urologic Oncology, David Geffen School of Medicine at the University of California-Los Angeles, Los Angeles, California
    Search for more papers by this author
  • Tobias Klatte MD,

    1. Institute of Urologic Oncology, David Geffen School of Medicine at the University of California-Los Angeles, Los Angeles, California
    Search for more papers by this author
  • Frédéric D. Birkhäuser MD,

    1. Institute of Urologic Oncology, David Geffen School of Medicine at the University of California-Los Angeles, Los Angeles, California
    Search for more papers by this author
  • Edward N. Rampersaud MD,

    1. Institute of Urologic Oncology, David Geffen School of Medicine at the University of California-Los Angeles, Los Angeles, California
    Search for more papers by this author
  • David B. Seligson MD,

    1. Department of Pathology and Laboratory Medicine, David Geffen School of Medicine at the University of California-Los Angeles, Los Angeles, California
    Search for more papers by this author
  • Nazy Zomorodian MSN,

    1. Institute of Urologic Oncology, David Geffen School of Medicine at the University of California-Los Angeles, Los Angeles, California
    Search for more papers by this author
  • Fairooz F. Kabbinavar MD,

    1. Institute of Urologic Oncology, David Geffen School of Medicine at the University of California-Los Angeles, Los Angeles, California
    2. Department of Medicine, David Geffen School of Medicine at the University of California-Los Angeles, Los Angeles, California
    Search for more papers by this author
  • Arie S. Belldegrun MD,

    1. Institute of Urologic Oncology, David Geffen School of Medicine at the University of California-Los Angeles, Los Angeles, California
    Search for more papers by this author
  • Allan J. Pantuck MD

    Corresponding author
    1. Institute of Urologic Oncology, David Geffen School of Medicine at the University of California-Los Angeles, Los Angeles, California
    • Associate Professor of Urology, Institute of Urologic Oncology, David Geffen School of Medicine at UCLA, 924 Westwood Boulevard, Suite 1050, Los Angeles, CA 90095-7207
    Search for more papers by this author
    • Fax: (310) 794-3513


Abstract

BACKGROUND:

Tobacco use is a leading cause of premature death, yet few studies have investigated the effect of tobacco exposure on the outcome of patients with renal cell carcinoma (RCC). The authors of this report retrospectively studied the impact of smoking on clinicopathologic factors, survival outcomes, and p53 expression status in a large cohort of patients with RCC.

METHODS:

Eight hundred-two patients (457 nonsmokers and 345 smokers) who had up to 232 months of follow-up were compared for differences in their clinicopathologic features and survival outcomes. Immunohistochemical differences in p53 expression were correlated with smoking status.

RESULTS:

Smokers presented more commonly with pulmonary comorbidities (P < .0001) and cardiac comorbidities (P = .014) and with a worse performance status (P = .031) than nonsmokers. Smoking was associated significantly with tumor multifocality (P = .022), higher pathologic tumor classification (P = .037), an increased risk of bulky lymph node metastases (P = .031), and the presence of distant metastases (P < .0001), especially lung metastases (P < .0001). Both overall survival (OS) (62.37 months vs 43.64 months; P = .001) and cancer-specific survival (CSS) (87.43 months vs 56.57 months; P = .005) were significantly worse in patients who smoked. The number of pack-years was retained as an independent predictor of CSS and OS in patients with nonmetastatic disease. Mean expression levels of p53 were significantly higher in current smokers compared with former smokers and nonsmokers (P = .012).

CONCLUSIONS:

In patients with RCC, a history of smoking was associated with worse pathologic features and survival outcomes and with an increased risk of having mutated p53. Further investigation of the genetic and molecular mechanisms associated with decreased CSS in patients with RCC who have a history of smoking is indicated. Cancer 2012;. © 2011 American Cancer Society.

INTRODUCTION

It is estimated that smoking is responsible for 443,000 deaths annually in the United States.1 The evidence supporting a causal relation between smoking and the development of kidney cancer has been presented by both the US Surgeon General's report and the International Agency for Research on Cancer.2, 3 However, most previous studies focused primarily on the risk of developing kidney cancer from exposure to tobacco; by contrast, fewer studies investigated the influence of tobacco exposure on survival outcome. Furthermore, previous studies4-8 often were limited by small sample sizes and/or limited numbers of patients who presented with or developed metastatic disease, were unable to detect an independent impact of tobacco on the survival of patients with renal cell carcinoma (RCC), or were lacking in their ability to provide molecular or genetic mechanistic explanations to explain the different survival outcomes in patients with a history of tobacco exposure versus lifetime nonsmokers.4, 6, 7 Finally, those studies were unable to convincingly investigate the influence of tobacco exposure level on patients' survival outcome.

Many solid human tumors demonstrate mutations of the p53 tumor suppressor gene.9 The prognostic value of immunohistochemically detectable (ie, mutated) p53 expression in RCC has been demonstrated in 4 large studies.10 Those studies demonstrated the importance of p53 on the recurrence of localized and metastatic RCC,11-13 disease-specific-survival,14 and disease progression15 in patients with RCC. However, to our knowledge, none of those studies investigated differing p53 expression with respect to tobacco exposure history, and the mechanisms that are responsible for these observations are not well understood. Therefore, we hypothesized that the investigation of a relation between tobacco exposure and p53 expression levels in patients with RCC may prove instructive. Therefore, the overall objectives of this study were to evaluate the influence of a history of tobacco exposure on pathologic features and on the cancer-specific survival (CSS) and overall survival (OS) of patients with clear cell RCC (ccRCC) and to investigate the influence of smoking exposure on p53 expression.

MATERIALS AND METHODS

Clinical and pathologic data were gathered from the Institutional Review Board-approved University of California-Los Angeles (UCLA) kidney cancer database. One thousand eight hundred nine patients who were treated for RCC between 1989 and 2007 were retrospectively reviewed. Eight hundred two patients with histologically confirmed ccRCC who had a complete data set of necessary clinical and pathologic variables, including data on tobacco exposure and dosage history, formed the study cohort. Patients were divided into 2 groups: lifetime nonsmokers (457 patients) and patients who had a positive history of tobacco exposure (345 patients). Patients were assigned to the smokers group if they had a tobacco exposure history of >1 pack-year, which we defined as smoking 1 pack of cigarettes (20 manufactured cigarettes) per day for 1 year. Because smoking history in the main clinical database coded patients as smokers or nonsmokers but did not specify time from smoking cessation, patients were categorized only as smokers or nonsmokers as in previous studies in the literature.16

Clinical data extracted for the study included age, sex, Eastern Cooperative Oncology Group performance status (ECOG PS),17 and data on comorbidities, such as hypertension, diabetes mellitus, chronic obstructive pulmonary disease (COPD), and coronary artery disease. Pathologic data included tumor stage, tumor size, tumor grade, and the presence of vascular invasion, satellite lesions, and sarcomatoid features. The tumor-lymph node-metastasis (TNM) staging classification was defined according to 2002 American Joint Committee on Cancer criteria,18 and nuclear grade was assigned according to the Fuhrman grading scheme.19 Localized tumors were defined as N0M0, whereas metastatic RCC was defined by the presence of lymph node involvement (N+) and/or distant metastasis (M1).

Tissue Microarray Construction

A custom kidney cancer tumor microarray (TMA) was constructed using tumor tissue from the 311 patients who were treated at UCLA between 1989 and 2000 for ccRCC. Only a subset of patients who were included in the TMA analysis also was part of the main clinical study cohort. In the TMA cohort, sufficient data were available to divide patients into current smokers (65 patients), former smokers (112 patients), and lifetime nonsmoker (134 patients). Patients were assigned to the group of current smokers if they were active smokers at the time of diagnosis and to the group of former smokers if they had ceased smoking before they were diagnosed. Formalin-fixed, paraffin-embedded primary tumor specimens were obtained from the Department of Pathology and Laboratory Medicine. Three 0.6-mm core tissue biopsies were taken from selected, morphologically representative regions of each paraffin-embedded specimen and precisely arrayed using methods previously described.20 Representative 4-μm-thick sections of the resulting tumor TMA block were transferred to glass slides using the paraffin sectioning aid system (adhesive-coated slides PSA-CS4x, adhesive tape, and ultraviolet lamp; Instrumedics, Inc., Richmond, Ill) to support the cohesion of the 0.6-mm array elements.

Immunohistochemistry

Four-micrometer-thick, formalin-fixed, paraffin-embedded tissue array sections were deparaffinized in xylene and then rehydrated in a graded ethanol series. After blocking of endogenous peroxidase activity, tissues were pretreated in a steamer for 25 minutes for retrieval of antigens. And 3% normal goat serum was used to block nonspecific protein binding. Immunohistochemical (IHC) staining was then carried out using an anti-p53 monoclonal murine antibody, immunoglobulin G2b (IgG2b) kappa isotype clone D0-7 (DAKO, Carpenteria, CA) at a dilution of 1:100. The antibody was then observed using the Envision Plus horseradish peroxidase (3,3′-diaminobenzidine) staining system, (DAKO). The anti-p53 primary antibody was replaced with nonimmune mouse IgG2b antibody (DAKO) at the same concentration to serve as a negative control. A single pathologist (D.B.S.), who was blinded to clinicopathologic variables and clinical outcomes, performed quantitative assessment of p53 expression. The extent of expression (“staining frequency”) was recorded as the percentage of the entire tumor sample that stained positive. The overall score used for subsequent statistical analysis was the pooled mean from the 3 spots sampled from each tumor.

Statistical Analyses

Clinical and pathologic variables were compared using the chi-square test and the Fisher exact test, as appropriate; and continuous variables were compared with the Student t test or the Mann-Whitney U test, as appropriate. The primary endpoint was CSS, which was calculated from the date of nephrectomy to either the date of death from RCC or the date of last contact recorded by our institution. OS was calculated from the date of nephrectomy to either the date of death for any reason or the date of last contact. Three patients who were missing follow-up information but had an otherwise complete clinicopathologic data set were included for the analysis of clinicopathologic features. CSS and OS were estimated using the Kaplan-Meier method, and group differences were compared using the log-rank test. Multivariate Cox proportional hazards models using a backward stepwise selection method with the likelihood ratio criterion (inclusion/exclusion: P ≤ .05/P ≥ .10) were used to determine the independent influence of variables on CSS or OS. The hazard ratio (HR), 95% confidence interval (CI), and P value for each removed variable were obtained during the removal step. All variables were investigated as continuous variables to avoid over-fitting of the model. The predictive accuracy of the prognostic models was assessed by using the concordance index (C index). All statistical tests were 2-sided and were performed at a significance level of .05 using the PASW (SPSS) 18.0 software package (IBM Corporation, New York, NY).

RESULTS

Clinical and Pathologic Characteristics:

The main clinical characteristics of the study population are listed in Table 1. Smokers had a median tobacco exposure history of 30 pack-years (range, 1-135 pack-years). Smokers had a higher incidence of COPD (P < .0001), cardiac disease (P = .014), and impaired ECOG PS compared with nonsmokers (P = .031). The mean number of all comorbidities, however, did not differ between the 2 groups.

Table 1. Patient and Tumor Characteristics
 SmokersNonsmokers 
CharacteristicNo.%No.%P
  1. Abbreviations: COPD, chronic obstructive pulmonary disease; ECOG PS, Eastern Cooperative Oncology Group performance status; SD, standard deviation.

Follow-up: Mean±SD, mo42.76±43.59 43.15±44.35 .900
Age: Mean (range), y58.92 (24-88) 59.02 (22-89) .915
Sex     
 Men24470.726056.9<.0001
 Women10129.319743.1 
Pack-years: Median (range)30 (1-135)   
ECOG PS     
 014140.923451.2 
 1155.121146.2.031
 2133.8112.4 
 310.310.1 
Comorbidities     
 Diabetes mellitus3610.45010.9.819
 Hypertension13639.417037.2.819
 COPD216.140.9<.0001
 Coronary artery disease5114.8429.2.014
Total no. of comorbidities: Mean (range)1.05 (0-4) 0.96 (0-6) .193

Table 2 lists the pathologic features of smokers and nonsmokers. Smoking was associated with higher pT classification (P = .037), higher pN status (P = .031), and a greater likelihood of distant metastases (P < .0001), especially lung metastases (P < .0001). Furthermore, smokers more often had vascular involvement (P = .025) and multifocal lesions (P = .022). Although no differences were observed in terms of Fuhrman grade (P = .174), the presence of sarcomatoid features (P = .199), or mean tumor size (P = .285), smokers were more often placed into higher risk groups of the UCLA Integrated Staging System21 because of a combination of worse clinical and pathologic features (P < .0001).

Table 2. Pathologic Features
 SmokersNonsmokers 
FeatureNo.%No.%P
  1. Abbreviations: SD, standard deviation; UCLA, University of California-Los Angeles.

Pathologic tumor classification
 pT112034.818740.9 
 pT245137616.6.037
 pT315946.117738.7 
 pT4216.1173.7 
Pathologic lymph node status
 pN0/pNx28783.240488.7 
 pN1205.8265.7.031
 pN23811275.9 
Metastasis classification
 M017951.931869.6 
 M116648.113930.4<.001
 Lung metastases13739.710122.1<.0001
Fuhrman grade
 14412.85812.7 
 215544.923150.5 
 312435.913228.9 
 4226.4367.9 
Tumor size: Mean±SD, cm7.18±4.09 7.49±3.92  
Sarcomatoid features215193.4.199
Vascular invasion12034.913128.7.025
Multifocal lesions6418.65812.7.022
UCLA integrated staging system
 Nonmetastatic     
  Low risk5014.58919.5 
  Intermediate risk9226.716636.3 
  High risk257.2439.4 
 Metastatic    <.0001
  Low risk267.5265.7 
  Intermediate risk13940.311825.8 
  High risk133.8153.3 

Survival Analysis

The median follow-up for the entire cohort was 39 months (range, 0-232.17 months) in patients with nonmetastatic disease and 15 months (range, 13-196.60 months) in patients with metastatic disease. At the time of analysis, 349 patients had died from RCC. Tobacco smoking history was associated with a 35% greater risk (HR, 1.35; 95% CI, 1.09-1.67; log-rank P = .005) of dying from kidney cancer (Fig. 1). Ninety patients died from causes other than kidney cancer. Patients with a positive tobacco exposure history had a 36% greater risk of dying from any cause compared with nonsmokers (HR, 1.36; 95% CI, 1.13-1.64). After 5 years, 86% ± 3% of smokers had died compared with 75% ± 2% of nonsmokers (log-rank P = .001 for OS) (Fig. 1 and Table 3). In a multivariate analysis that included all evaluable patients, ECOG PS, tumor stage, distant metastases, Fuhrman grade, multifocality, and sarcomatoid features, but not tobacco dosage, were independent predictors of CSS.

Figure 1.

(Top) Cancer-specific survival and (Bottom) overall survival estimates are illustrated. HR indicates hazard ratio.

Next, additional analyses of patients with nonmetastatic and metastatic disease were conducted to evaluate the influence of smoking on CSS in these separate groups. In total, 464 patients (298 nonsmokers and 166 smokers) presented with localized disease, and 338 patients (159 nonsmokers and 179 smokers) presented with metastatic disease. In multivariate Cox regression analysis, tobacco dosage classified by pack-years was retained as an independent predictor of CSS in patients with nonmetastatic disease. Each pack-year was associated with a 1% increase in the risk of death from RCC. The final model for CSS in patients with nonmetastatic disease included pT classification, Fuhrman grade, ECOG PS, sarcomatoid features, tumor size, sex, and the number of pack-years (Table 4). The predictive accuracy for the model increased significantly when the amount of tobacco exposure was added (the C index increased from 79.9% to 80.6%; P = .0151). For predicting OS, the number of pack-years was 1 of the strongest predictors in patients with nonmetastatic disease (Table 4). The C index increased significantly from 67.6% to 68.7% (P = .0007) when the number of pack-years was added to the model. Although the clinical significance of this difference is unclear, it underscores the prognostic importance of tobacco exposure on RCC survival. Conversely, the number of pack-years was not an independent predictor of either OS or CSS in patients with metastatic disease.

Table 3. Five-Year Survival Probability
 Survival Probability: Mean±SE, %
 CSSOS
SurvivalSmokersNonsmokersSmokersNonsmokers
  1. Abbreviations: CSS, cancer specific survival; OS, overall survival; SE, standard error.

1-Year77±283±275±280±2
2-Year65±374±262±370±2
3-Year57±367±352±362±3
4-Year53±362±347±356±3
5-Year27±335±414±325±3
Table 4. Multivariate Cox Proportional Hazards Analyses
FactorHR95% CIP
  1. Abbreviations: CI, confidence interval; CSS, cancer-specific survival; ECOG PS, Eastern Cooperative Oncology Group performance status; HR, hazard ratio; M0, no metastasis; N0, negative lymph node status; OS, overall survival.

Multivariate analysis of OS for N0M0
 Sex1.150.87-1.61.430
 Tumor classification1.5151.272-1.805<.0001
 Fuhrman grade1.2931.038-1.601.022
 ECOG PS2.0561.519-2.738<.0001
 Sarcomatoid features2.4931.206-5.153.014
 Tumor size1.010.96-1.06.604
 Pack-years of smoking1.0111.006-1.017<.0001
Multivariate analysis of CSS for N0M0
 Sex1.9171.210-3.037.006
 Tumor classification1.6421.284-2.099<.0001
 Fuhrman grade1.8691.364-2.562<.0001
 ECOG PS2.2861.494-3.498<.0001
 Sarcomatoid features3.3081.464-7.475.004
 Tumor size1.0781.018-1.141.010
 Pack-years of smoking1.0101.003-1.017.008
Multivariate analysis of p53 expression for all patients
 Tumor classification1.571.25-1.98<.0001
 Lymph node status1.260.974-1.63.079
 Metastasis classification2.851.85-4.39<.0001
 Fuhrman grade1.250.962-1.623.095
 ECOG PS1.821.36-2.42<.0001
 p53 expression1.011.01-1.02.003

Immunohistochemistry

Three hundred eleven patients who had a maximum follow-up of 11.8 years (median, 2.12 years) were analyzed for p53 expression. The TMA study population was adjusted for pT classification, pN status, M classification, Fuhrman grade, and ECOG PS. Expression of p53 was detected in 70.8% of current smokers, 52.7% of former smokers (P = .018), and 53% of nonsmokers (P = .017). Current smokers had significantly higher total p53 expression compared with both former smokers (P = .010) and nonsmokers (P = .012) (Fig. 2). In multivariate analysis, p53 expression was retained as an independent predictor of CSS for all patients (HR, 1.013; 95% CI, 1.005-1.022; P = .002).

Figure 2.

Expression of the p53 tumor suppressor gene is illustrated according to patients' smoking history.

DISCUSSION

We examined the outcome of patients who had ccRCC with respect to their tobacco exposure history for differences in clinicopathologic features at the time of presentation and for differences in survival outcome. The main finding of the study is that patients who have a positive history of tobacco exposure present with more aggressive forms of RCC and experience worse CSS and OS. Smokers presented more commonly with impaired ECOG PS and with worse medical comorbidities, including cardiac and pulmonary disease, compared with nonsmokers. In addition, smoking was associated significantly with tumor multifocality, worse pT classification, and increased incidence of lymphatic and distant metastases. Moreover, the amount of tobacco exposure, as measured by number of pack-years smoked, was identified as an independent predictor of CSS and OS in patients with nonmetastatic ccRCC, and each pack-year of smoking was associated with a 1% increase in mortality for patients with nonmetastatic RCC. However, tobacco exposure was not an independent predictor of CSS or OS in patients with metastatic RCC. Finally, p53 expression was detected more often and at higher mean expression levels in current smokers than in former smokers or nonsmokers and was retained as an independent predictor of CSS in patients with nonmetastatic disease.

Tobacco exposure has been reported as a factor that significantly and negatively influences CSS in numerous cancers, including cancers of the colon, rectum,22 prostate,23 and bladder.24 Hunt et al, in a meta-analysis of 24 studies, observed a dose-response relation between tobacco exposure and the relative risk for RCC development.16 Moreover, epidemiological studies have suggested that smoking is associated with a worse survival outcome in patients with RCC.4-8 Sweeney and Farrow reported an increased risk of overall and cancer-specific death (HR, 1.7; 95% CI, 1.2-2.5) for current smokers, but only during the first 6 months of their course of disease. However, that study had several limitations in terms of a short follow-up period, lack of availability of pathologic data, and the inability to analyze tobacco exposure quantitatively.7 In a large study of patients with ccRCC, Parker et al also reported a 31% greater risk of dying from RCC for current smokers compared with nonsmokers (HR, 1.31; 95% CI, 1.09-1.58; P = .004). However, in that study, the difference was no longer significant when patients were adjusted for TNM classification.4

Smokers presented with significantly higher rates of cardiac disease and had a worse ECOG PS at diagnosis. Both of these findings serve as potential explanations for why OS was worse in smokers compared with nonsmokers. Even more noteworthy is our observed that the incidence of both pulmonary metastases and COPD was significantly greater in smokers compared with nonsmokers. Chronic tobacco exposure is a major etiologic factor in the development of COPD,25 which is caused in part by the destructive enzyme secretion of activated alveolar macrophages, whose numbers are increased several times in smokers compared with nonsmokers.26 Macrophages also are able to generate reactive oxygen species and to initiate angiogenesis and invasion.27, 28 The chronic alteration of the lungs caused by tobacco exposure provokes an inflammatory microenvironment that is an ideal soil for metastases,29 which may explain in part why smokers were significantly more likely to have pulmonary metastases compared with nonsmokers. Chronic tobacco exposure also alters multiple immunologic functions, including the innate and adaptive immune system.26 RCC is known as a tumor that is sensitive to immunotherapy.30 However, we were not able to demonstrate differences in the response to immunotherapy in smokers compared with nonsmokers (data not shown).

In this study, current smokers had a significantly higher frequency of detectable p53 than both former smokers and nonsmokers: The amount of p53 expression was significantly higher in current smokers, and p53 expression was retained as an independent predictor of CSS for all patients. The p53 tumor suppressor gene regulates the cell cycle and induces apoptosis when DNA damage occurs. Studies in lung cancer previously demonstrated that the relative risk of harboring a p53 mutation was up to 13 times greater in lifetime heavy smokers compared with lifetime nonsmokers.1 Furthermore, similar to the current data being presented currently for RCC, a study in lung cancer has demonstrated that p53 mutations are detected in 47.5% of never smokers, 55.6% of former smokers, and 77.4% of current smokers. Moreover, that study indicated that 93% of tumors specifically with p53 missense mutations demonstrate positive IHC staining.31 Unfortunately, the authors of that report did not provide mechanistic insights into the observed difference in p53 expression between current smokers and former smokers. Excess p53 accumulation identified by IHC expression analysis has been commonly accepted as a surrogate for p53 mutation status. Moreover, previous reports indicated that p53 expression is detectable by IHC in 20% to 50% of renal tumors,11-13, 15 and the amount of p53 expression reportedly is an important prognostic factor in ccRCC.10 However, in some studies, p53 mutation analyses have suggested that the increased p53 expression observed in RCC can occur independent of p53 mutations,10 but this observation was based only on 2 small analyses in which the presence of p53 mutations was correlated with p53 staining in mixed histologic subpopulations.32, 33

In the current study, p53 expression was noted in 70.8% of current smokers, in 52.7% of former smokers, and in 53% of nonsmokers. Therefore, the frequency of p53 expression in current smokers was greater than the frequency of expression reported in previous studies that did not stratify patients according to their smoking status. It is noteworthy that p53 expression appears to normalize after smoking cessation to the same level as that in nonsmokers. This supports the hypothesis that p53 expression detected by IHC represents mutated protein. However, to date, it remains unclear why mutation frequencies in current smokers and former smokers differ significantly. It seems reasonable to expect that the mutations caused by smoking would persist beyond the tobacco exposure period. It also is possible that p53 damage may occur only after prolonged tobacco exposure or that p53 function can be restored through cellular repair mechanisms that are limited in their repair capacity when under the continuous influence of tobacco smoke. For example, it was demonstrated that 1 key carcinogen of tobacco smoke and environmental pollution, benzo(a)pyrene diol epoxide (BPDE), a metabolite of benzo(a)pyrene, is a major candidate for p53 mutations in lung cancer. BPDE forms adducts at major hotspots of p53 mutations and, thus, is a candidate for smoking-related p53 mutations.34 Denissenko et al demonstrated that repair mechanisms are able to remove these adducts from DNA in a time-dependent manner.35 DNA repair mechanisms are controlled enzymatically; and, like every enzymatic reaction, these processes follow saturation kinetics. Thus, it appears that mutations may become more readily apparent when under a continuous influence of BPDE than after smoking cessation.

One limitation of the current study is that data on smoking and on the number of pack-years of tobacco exposure were obtained from a self-reported patient interview. It is possible, therefore, that the estimates for tobacco exposure used in this study are only rough approximations. Sweeney and Farrow and Nelson et al have reported that, in direct patient interviews, smoking history often is underestimated.7, 36 Thus, it is reasonable to assume that self-reported tobacco exposure history also may be underestimated. Conversely, it has been suggested that patients who are smokers are likely to overestimate lifetime tobacco use. However, despite these uncertainties, we were able to demonstrate that smoking is associated with molecular changes that may explain the worse survival outcomes and clinicopathologic features of smokers compared with nonsmokers.

In conclusion, this study demonstrates a link between smoking, worse pathologic features, aberrant p53 expression, and diminished survival outcomes in patients with ccRCC, and each additional pack-year of smoking is associated with a 1% increase in the risk of mortality in patients with nonmetastatic disease. These findings warrant further investigation into the genetic and molecular mechanisms associated with decreased CSS and OS observed.

FUNDING SOURCES

No specific funding was disclosed.

CONFLICT OF INTEREST DISCLOSURES

The authors made no disclosures.

Ancillary