Published in Cancer Detection and Prevention 1997; 21(6):497-509.

Smoking History and Cancer Patient Survival: A Hospital Cancer Registry Study

Guo-Pei Yu, MD MPHa, Jamie S Ostroff, PhDb, Zuo-Fengzhang, MD Ph.Dc, Jian Tang MSc, Stimson P Schantz MDa

Departments of aSurgery, bPsychiatry and Behavioral Sciences, cEpidemiology and Biostatistics, Memorial Sloan-Kettering Cancer Center, New York, NY

Address all correspondence and reprint requests to: Stimson P. Schantz, M.D. Department of Surgery, Memorial Sloan-Kettering Cancer Center, 1275 York Ave., New York. NY 10021.

ABSTRACT: While tobacco use is clearly the most preventable cause of cancer, little is known about whether smoking adversely influences cancer patients survival. The goal of this study was to examine the effect of smoking history on survival among cancer patients. Data from Memorial Sloan-Kettering Cancer Center's tumor registry was used to identify 25,436 cases of cancer (12,447 male patients and 12,989 female patients). Information regarding smoking and alcohol consumption, histologic grade, tumor stage, and survival time was available. Proportional hazard analysis was used to examine the effect of smoking on the death from all causes among patients. Patients who had a history of smoking were found to have a lower rate of survival than nonsmokers. After controlling for age, race, alcohol use, and histologic grade, the risk ratios were 1.55 for males and 1.43 for females. A dose-response relationship was found between ever-smoking and cancer patient survival. The predictive effect of smoking on survival was significant for patients with oral, pancreatic, breast, and prostate cancers, but not for esophageal, stomach, colon, rectum, laryngeal, lung, cervix uteri, urinary bladder, and kidney cancers. Black patients with oral or breast cancer had a poorer prognosis associated with smoking compared with white and other nonwhite patients. The strongest effect of smoking on survival was found mainly among patients with breast cancer with a distant tumor stage or with prostate cancer with a regional tumor stage. Alcohol use alone was associated with a higher risk of death for nonsmokers with oral and pancreatic cancers than for similar cancer patients without a history of alcohol use. Smoking history plays a critical role in influencing cancer patient survival, especially for patients diagnosed with oral, pancreatic, breast, or prostatic cancers. In addition, alcohol consumption is independently related to survival in patients with oral and pancreatic cancers. Our study suggests that a potential means of improving cancer patient survival, especially from oral, pancreatic, breast, and prostatic cancers, may be achieved through smoking cessation.

KEY WORDS: breast, neoplasms, oral cavity, pancreas, prostate, smoking, survival.