|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
© 2003 American Society for Clinical Oncology Impact of Smoking on Cancer Stage at Diagnosis
From the Roger Maris Cancer Center; Department of Pediatrics, MeritCare Children’s Hospital, Fargo; Department of Pediatrics, Faculty of Medicine, University of North Dakota, Fetal Alcohol Syndrome Center, Grand Forks, ND. Address reprint requests to Nathan Kobrinsky, MD, 820 4th St. North, Fargo, ND 58122, email: nathankobrinsky{at}meritcare.com.
Background: Studies evaluating the relationship between smoking and cancer spread are limited. Methods: We studied the relationship between cancer stage at diagnosis (local, regional, or metastatic) and smoking history (current, previous, or nonsmoker). For lung cancer, patterns of spread were also studied. Results: In a tumor registry for eastern North Dakota, northwestern Minnesota, and northern South Dakota, 11,716 cases were identified from 1986 to 2001. Current smokers (relative risk [RR], 2.11; 95% confidence interval, 1.93 to 2.32; P < .001) and previous smokers (RR, 1.56; 95% confidence interval, 1.42 to 1.72; P < .001) had an increased risk of metastatic disease at diagnosis. Current smokers (RR, 1.39; 95% confidence interval, 1.29 to 1.51; P < .001), but not previous smokers, also had an increased risk of regional disease. An increase in metastatic disease was most evident for prostate cancer (RR, 1.53; P = .003). An increase in regional disease was most evident for head and neck (RR, 3.53; P < .001), prostate (RR, 1.83; P = .030), and breast cancer (RR, 1.22; P = .005). Compared with previous smokers, current smokers with metastatic lung cancer were more likely to have involvement of the brain (33.6% v 23.0%; P = .004), bone marrow, adrenal gland, and pericardium (24.7% v 15.9%; P = .004). Conclusion: Previous or current smoking is a risk factor for increased cancer stage in a wide range of malignancies. Further study is required to determine whether this association is causal.
CIGARETTE SMOKING has been unequivocally established as the main causative factor for cancers of the head and neck, lung, and bladder and a major contributing factor for cancers of the esophagus, pancreas, and kidney.1,2 For this reason, tobacco settlement money in many states has been targeted for smoking cessation programs.3 Nonetheless, for subjects with newly diagnosed cancer, the recommendation for immediate smoking cessation is often given and received with ambivalence. Is the commonly expressed adage "it’s too late to quit" true? Or, does cigarette smoking have effects on the pathogenesis of cancer that go beyond carcinogenesis; that is, is cigarette smoking associated with the spread of cancer? If so, it would clearly never be too late to quit. Previous investigators have demonstrated an association between smoking and the spread of head and neck cancer.4 Smoking doubles the risk of myocardial infarction and ischemic stroke.5 In vitro, smoking confers a hypercoagulable state through effects on coagulation proteins, tissue factor expression, and platelets.6–8 Furthermore, many experimental systems demonstrate an association between hypercoagulability, tumor angiogenesis, tumor growth, and the metastatic spread of cancer.9 On the basis of these considerations, it was proposed that cigarette smoking may be a risk factor for cancer spread for a wide range of malignancies. To test this hypothesis, the relationship between cancer stage at diagnosis (local, regional, or metastatic) and smoking history (current, previous, or nonsmoker) was compared for a broad range of malignancies. This article describes the results of these analyses. Current smoking, and to a lesser extent previous smoking, was associated with an increased stage at diagnosis and adversely affected metastatic behavior.
Study Population Patients entered in a regional tumor registry from eastern North Dakota, northwestern Minnesota, and northern South Dakota from January 1, 1986, to October 1, 2001, were studied. Cancers studied included breast, bladder, colorectal, esophageal, head and neck, kidney, lung, melanoma, pancreatic, prostate, and stomach. Cases for which stage, pathology, or smoking history were unknown were excluded. Cases with a history of smokeless tobacco, pipe, or cigar smoking were also excluded. Cancer stage was classified using criteria outlined in the Surveillance Epidemiology and End Results (SEER) Summary Staging 1977 Manual.10 Stages 0, 1, and 2 were identified as "localized." Stages 3, 4, and 5 were identified as "regional." Stage 7 was identified as "metastatic." Cases were identified as nonsmokers, current smokers, or previous smokers. Current smokers were defined as regular smokers at diagnosis or within a year of diagnosis. Previous smokers were defined as having quit smoking a year or more before diagnosis.
Statistical Analysis Relative risks were calculated to examine the association between the extent of disease at diagnosis and current or previous smoking history. Stage-dose relationships between smoking and relative risks for regional and metastatic spread were calculated comparing current smokers with nonsmokers or previous smokers with nonsmokers where appropriate. Potential confounding effects of sex and age were evaluated using the Cochran-Mantel-Haenszel statistic.11 Potential interactions between sex, age, smoking history, and tumor stage were also examined by logistic regression.
To evaluate the effects of smoking on the patterns of lung cancer spread, the relationship between tumor type (ie, small-cell, adenocarcinoma, or squamous cell carcinoma), specific sites of metastatic involvement at diagnosis, and smoking history were evaluated by
Patient distribution by disease, sex, and age is summarized in Table 1  . Overall, 11,716 patients were evaluated. Fifty-four percent were male and 46% were female. The mean age was 66 years (range, 12 to 99 years). Twenty-five percent (n = 2,929) were current smokers, 31.1% (n = 3,638) were previous smokers, and 43.9% (n = 5,149) had never smoked. Males were more likely to be current or previous smokers than females for all types of cancer. Cancer was localized in 60.7% (n = 7,116), regionally spread in 21.5% (n = 2,516), and metastatic in 17.8% (n = 2,084) of patients.
Smoking History and Cancer Stage at Diagnosis The relationship between smoking history and cancer stage at diagnosis is summarized in Table 2 , where 24.8% of current smokers, 19.8% of previous smokers, and 12.4% of nonsmokers had metastatic disease at diagnosis. Similarly, 25.1% of current smokers had regional disease at diagnosis whereas 21.0% of nonsmokers had regional disease at diagnosis (P < .001). Differences were particularly evident for cancers of the breast (P = .034), head and neck (P < .001), prostate (P < .001), and stomach (P = .037).
A possible dose effect of cigarette smoking is demonstrated in Table 2  . Particularly for cancers of the head and neck, prostate, stomach, bladder, and melanoma and to a lesser extent, other cancers, metastatic disease was most prevalent in current smokers, less prevalent in previous smokers, and least prevalent in nonsmokers.
Smoking History and the Relative Risk of Cancer Spread
Current smokers had a more than twofold increase in the risk of developing metastatic disease compared with nonsmokers (relative risk [RR], 2.11; P < .001). An increase was also evident for prostate cancer (RR, 1.53; P = .003). Previous smokers also had an increased risk of developing metastatic disease compared with nonsmokers (RR, 1.56; P < .001). Of note, previous smokers with squamous cell lung cancer had a 10% decrease in the risk of developing metastatic disease (P = .043). The potential public health implications of these data can be portrayed by the effects of smoking on cancer stage and also by the number of additional metastatic patients potentially caused by smoking. For all cancers evaluated in this data set, current smoking was associated with an additional 383 patients with metastatic spread, and previous smoking was associated with an additional 258 patients with metastatic spread of cancer. Current smokers also had an increase in the regional spread of cancer at diagnosis (RR, 1.39; P < .001). Increases were evident for breast cancer (RR, 1.22; P = .005), head and neck cancer (RR, 3.53; P < .001), and prostate cancer (RR, 1.83; P = .030). Previous smokers had an increased risk of regional spread for head and neck cancer (RR, 2.39; P = .003). An increased risk of regional spread of other cancers at diagnosis was not observed; however, an interaction with sex was demonstrated (P < .001). When stratified by sex, males demonstrated an increased risk of regional spread (RR, 1.62; 95% confidence interval, 1.39 to 1.89; P < .001). An increased risk for females was not observed. In this data set, current smoking was associated with 207 additional patients with regional cancer spread.
Smoking History and Pattern of Metastatic Spread in Lung Cancer
Current smokers were more likely to have involvement of the CNS (33.6%) than previous smokers (23.0%) or nonsmokers (25.0%; P = .004). This association was observed for adenocarcinoma (P = .007) and for squamous cell carcinoma (P = .017) but not for small-cell carcinoma (P = .996). Current smokers were also more likely than previous or nonsmokers to have involvement of the adrenal gland, bone marrow, and pericardium (P = .004). This effect was equally observed for all pathologies of lung cancer. Other effects of smoking on the pattern of spread of lung cancer were observed. An increase in metastases to the peritoneum (P = .065) and pleura (P = .065) was identified in current and previous smokers compared with nonsmokers. Also, an apparent increase in metastases to the skin (P = .047) was identified in current smokers with adenocarcinoma compared with previous and nonsmokers. These findings are of uncertain significance in view of the small numbers compared.
Effects of Age and Sex For prostate cancer, 3.4% of nonsmokers versus 5.8% of current smokers had regional disease and 11.4% of nonsmokers versus 17.0% of current smokers had metastatic disease at diagnosis (P < .001). Whereas the RR of metastatic spread was significant, the RR of regional spread may have been confounded by age (P = .073). The mean (±SB) ages for current smokers, previous smokers, and nonsmokers with prostate cancer were 68.15 ± 8.54, 70.45 ± 7.51, and 70.38 ± 8.31 years, respectively. The difference among the three groups was about 2 years, a clinically insignificant difference.
Although the association between smoking and the spread of stomach cancer in Table 2 For cancer of the pancreas, an interaction with age (P = .002) and sex (P = .010) was observed; however, a relationship between stage and smoking history was still not identified for this cancer after stratifying for these variables.
The association between smoking and the spread of breast cancer shown in Table 2
Table 5
There was a significant confounding relationship between age and spread of breast cancer. Younger women were more likely to have regional spread of breast cancer than localized or metastatic disease (F = 16.27, P < .001).
Information on the effects of smoking on the spread of cancer is limited. Trigg et al12 reported a retrospective review of 499 subjects with laryngeal cancer. They found an incremental increase in the risk of advanced (stage III and IV) disease with increasing tobacco use and with heavy alcohol use.13 In the present study, we found an association between current smoking and, to a lesser extent, previous smoking, with an increase in regional and distant cancer spread at diagnosis when compared with nonsmokers. Previously reported effects of smoking vary for different organ systems. In this study, strong associations with stage at diagnosis were observed for cancers of the head and neck, kidney, bladder, and prostate and melanoma. An association with breast cancer was also observed, though this was confounded by age. Thus, the observation of increased regional spread in a population that is a mean 7 years younger at diagnosis (ie, smokers v nonsmokers) may be an age effect rather than a smoking effect. No association between smoking status was observed for the gastrointestinal malignancies. An association with the stage of lung cancer was not observed; however, an association with adverse patterns of metastatic spread in lung cancer was evident. Associations between the various lung cancer histologies and smoking history have been long appreciated. Small-cell lung cancer is the most rapid and widely metastatic type of lung cancer. It is particularly associated with heavy, current smokers who favor unfiltered cigarettes. The incidence of small-cell lung cancer decreases dramatically in subjects who quit smoking and is minimal in nonsmokers. Conversely, adenocarcinoma of the lung is the most commonly observed histology in previous smokers. In this study, metastatic behavior as well as lung cancer histology was also related to smoking history. Current smokers with adenocarcinoma and squamous cell but not small-cell carcinoma experienced an increased risk of CNS metastases when compared with previous smokers and nonsmokers. Similarly, current smokers with adenocarcinoma but not squamous or small-cell carcinoma experienced an increased risk of skin metastases when compared with previous smokers and nonsmokers. The organ and tissue specificity of these smoking-related observations indicates but does not prove an underlying biologic basis for these effects.
Limitations This study does not determine whether smoking in fact increases cancer spread and adversely affects the patterns of metastatic behavior at diagnosis, or whether smoking is closely associated with a population of patients far less likely to seek medical attention until symptoms are advanced. These data identify an important public health issue that requires further research. If smoking facilitated cancer spread before diagnosis, smoking cessation at diagnosis would clearly be warranted. Although smoking is an important risk marker, further study is required to determine whether smoking may also be a risk factor or part of a complex causal mechanism for cancer spread. This hypothesis that smoking may facilitate cancer spread should be tested in a prospective trial comparing the outcome of patients matched by disease and stage who quit smoking with patients who keep smoking from the time they are diagnosed with cancer. Studies addressing this issue are limited and conflicting. In one such study of head and neck cancer, continued smoking did not affect outcome.15 In another head and neck cancer study4, continued smoking increased the risk of second malignant neoplasms; however, an effect on recurrence of the primary tumor was not reported. The findings of this study are also consistent with but do not provide causal evidence for the hypothesis that a hypercoagulable state induced by cigarette smoking may facilitate the spread of cancer. Cigarette smoking could also increase tumor angiogenesis or suppress host immune surveillance mechanisms. In future research on this hypercoagulable state hypothesis, the effects of antiplatelet or anticoagulant therapies introduced at the time of diagnosis on disease outcome would need to be prospectively evaluated to establish a causal link and establish biologic specificity of smoking as a factor in cancer spread. In summary, smoking is associated with cancer spread at diagnosis. This finding has major implications for future research and offers support for the notion that, with regard to cigarette smoking, "it’s never too late to quit."
We thank Kathy Hanish, RN, MS, CNA, Executive Director, and John Leitch, MD, Medical Director, Roger Maris Cancer Center; and Ron Miller, MD, Head of the Department of Pediatrics, MeritCare Children’s Hospital, for their support of this research project. We thank Ralph Levitt, MD, and Ted Kleiman, MD, for their critical review of the manuscript.
Supported by the Roger Maris Cancer Center, MeritCare Children’s Hospital, and the Children’s Miracle Network, Fargo, ND.
1. Burns DM, ed. Tobacco and the Clinician: Interventions for Medical and Dental Practice. Bethesda, MD, National Cancer Institute, NIH publication 94-3693, 1994 2. U. S. Smoking and Health: A Report of the Surgeon General. Washington, DC, Government Printing Office, Department of Health, Education, and Welfare, publication 79-50066, 1979
3. Kessler D, Myers ML: Beyond the tobacco settlement. N Engl J Med 345:535–537, 2001
4. Khuri FR, Kim ES, Kee JJ, et al: The impact of smoking status, disease stage, and index tumor site on second primary tumor incidence and tumor recurrence in the head and neck retinoid chemoprevention trial. Cancer Epidemiol Biomarkers Prev 10:823–829, 2001
5. Ockene IS, Miller NH: Cigarette smoking, cardiovascular disease and stroke: A statement for healthcare professionals from the American Heart Association. Circulation 96:3243–3247, 1997
6. Maretzky S, Tani S, Kangavari S, et al: Smoking increases tissue factor expression in atherosclerotic plaques: Implications for plaque thrombogenicity. Circulation 102:602–604, 2000 7. Miller GJ, Bauer KA, Cooper JA, et al: Activation of the coagulant pathway in cigarette smokers. Thromb Haemost 79:549–553, 1998[Medline] 8. Benowitz NL, Fitzgard GA, Wilson M, et al: Nicotine effects on eicosanoid formation and hemostatic function: Comparison of transdermal nicotine and cigarette smoking. J Am Coll Cardiol 22:1159–1167, 1993[Abstract] 9. Rickles FR, Levine MN, Dvorak HF: Abnormalities of hemostasis in malignancy, in Colman RW, Hirsh J, Marder VJ (eds). Hemostasis and Thrombosis: Basic Principles and Clinical Practice (ed 4). Philadelphia, PA, Lippincott, Williams and Wilkins, 2001, pp 1131–1152 10. Shambaugh EM, Weiss MA (eds). Summary Staging Guide: Cancer Surveillance Epidemiology and End-Results Reporting: SEER Program 1977. Bethesda, MD, Public Health Service, NIH publication 98-2313, 1998 11. Schelesselman JJ. Case Control Studies: Design, Conduct, Analysis. New York, Oxford University Press, 1982 12. Trigg DJ, Lait M, Wenig BL: Influence of tobacco and alcohol on the stage of laryngeal cancer at diagnosis. Laryngoscope 110:408–411, 2000[CrossRef][Medline]
13. Henschke CI, McCauley DI, Yankelevitz DF, et al: Early lung cancer action project: A summary of the findings on baseline screening. Oncologist 6:147–152, 2001 14. Wells CK, Peduzzi PN, Feinstein AR: Presenting manifestations, cigarette smoking, and detection bias in age at diagnosis of lung cancer. Ann Epidemiol 11:239–247, 2001[CrossRef][Medline] 15. de Graeff A, de Leeuw JRJ, Ros WJG, et al: Sociodemographic factors and quality of life as prognostic indicators in head and neck cancer. Eur J Cancer 37:332–339, 2001[CrossRef][Medline] Submitted May 16, 2002; accepted November 12, 2002. This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2003 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
TOP
ABSTRACT
INTRODUCTION
2 analysis in a three-by-three contingency table. Where expected values were less then five for two or more cells, regional and metastatic groups were combined. Significance at the 0.05 level was noted. All analyses were performed using the SAS software, version 8.0 (SAS Institute, Cary, NC). 
