Siegel RL, Jacobs EJ, Newton CC, Feskanich D, Freedman ND, Prentice RL, Jemal A. Deaths Due to Cigarette Smoking for 12 Smoking-Related Cancers in the United States. JAMA Intern Med. 2015;175(9):1574-1576. doi:10.1001/jamainternmed.2015.2398
The 2014 US Surgeon General’s Report provided the estimated annual number of smoking-attributable deaths during 2005 to 2009 from cancer overall and lung cancer specifically but not separately for the 11 other cancers found to be caused by smoking.1 Current estimates of smoking-attributable mortality for specific cancer sites are based on data from 2000 to 2004.2 Updated estimates are needed because smoking patterns and the magnitude of the association between smoking and cancer death have changed in the past decade. From 2000 to 2012, smoking prevalence decreased from 23.2% to 18.1%.3 In contrast to this favorable trend, recently published data revealed that the risk of cancer death among smokers can increase over time.4 Therefore, we estimated the number and proportion of deaths in the United States in 2011 attributable to cigarette smoking for 12 cancers caused by smoking.
For 12 cancers established as caused by smoking, we used the standard formula5 to calculate the population-attributable fraction (PAF) within strata defined by sex and age group (35-54, 55-64, 65-74, and ≥75 years) using SAS statistical software, version 9.3 (SAS Institute Inc). The PAFs were calculated using smoking prevalence (current, former, or never) from the 2011 National Health Interview Survey3 and age- and sex-specific relative risks (RRs) for former and current smoking from the Cancer Prevention Study II (CPS-II)4 (for the 35- to 54-year age group, follow-up from 1982-1988) or the pooled contemporary cohort (PCC)4 (for other age groups, follow-up from 2000-2011). The National Health Interview Survey provides smoking prevalence estimates based on in-person interviews of a nationally representative sample of US adults. The CPS-II and the 5 studies that compose the PCC are large US cohort studies that ascertained smoking from self-administered questionnaires and are described in detail elsewhere.4 The RRs were adjusted for age, race, educational level, and alcohol use; the RRs from the PCC were additionally adjusted for cohort. The RRs from CPS-II appear in the 2014 US Surgeon General’s Report,1 and the RRs from the PCC appear in the article by Carter et al.6 For each cancer site, smoking-attributable deaths in each age and sex group were calculated by multiplying age- and sex-specific PAFs by age- and sex-specific death counts in 2011 derived from the National Vital Statistics System. Smoking-attributable deaths were then summed across age and sex groups to determine the total number of smoking-attributable deaths, which was divided by the total number of deaths to calculate the overall PAF. These statistical methods and data sources match those used in smoking-attributable mortality calculations in the 2014 US Surgeon General’s Report but use slightly updated data.1
Of the 345 962 deaths among adults 35 years and older in 2011 from the 12 cancer sites examined, we estimate 167 805 (48.5%; 95% CI, 46.2%-51.2%) were caused by cigarette smoking. The largest proportions of smoking-attributable deaths were for cancers of the lung, bronchus, and trachea (125 799 [80.2%]; 95% CI, 79.2%-81.1%) and larynx (2856 [76.6%]; 95% CI, 68.7%-83.5%) (Table and Figure). Approximately half of the deaths from cancers of the oral cavity, esophagus, and urinary bladder were due to smoking.
Cigarette smoking continues to cause numerous deaths from multiple cancers despite half a century of decreasing prevalence. The smoking downturn is likely reflected in the generally lower proportions of deaths caused by smoking in 2011 than in 2000 to 2004 for the 10 overlapping cancer sites. Earlier estimates excluded colorectal and liver cancers (10 029 smoking-attributable deaths in 2011) because they were not yet established smoking-related sites.
One limitation of our study is the cohort populations, which are less racially diverse and more educated than the US population. In addition, tobacco exposures other than cigarettes were not included in our analysis. These exposures include secondhand smoke, which is estimated to cause an additional 5% of lung cancer deaths,1 and the use of cigars, pipes, and smokeless tobacco, which undoubtedly account for a considerable proportion of deaths from cancers of the oropharynx, larynx, and esophagus. Although differences in exposures (eg, diet and exercise) between smokers and nonsmokers may have potentially confounded our results, these differences have minimal effect on smoking-attributable risk.1
Continued progress in reducing cancer mortality, as well as deaths from many other serious diseases, will require more comprehensive tobacco control, including targeted cessation support.
Corresponding Author: Rebecca L. Siegel, MPH, Surveillance and Health Services Research Program, Intramural Research Department, American Cancer Society, 250 Williams St NW, Atlanta, GA 30303.
Published Online: June 15, 2015. doi:10.1001/jamainternmed.2015.2398.
Author Contributions: Dr Jacobs and Ms Newton had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Siegel, Jacobs, Jemal.
Acquisition, analysis, or interpretation of data: Siegel, Jacobs, Newton, Feskanich, Freedman, Prentice.
Drafting of the manuscript: Siegel.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Jacobs, Newton, Feskanich, Prentice.
Administrative, technical, or material support: Siegel, Freedman.
Study supervision: Jemal.
Conflict of Interest Disclosures: None reported.
Funding/Support: Data analysis for this work was funded by the American Cancer Society. The Intramural Research Program of the National Cancer Institute of the National Institutes of Health (NIH) provided support for the NIH-AARP Diet and Health Study. The CPS-II nutrition cohort was supported by the Intramural Research Programs of the American Cancer Society. The Nurses’ Health Study and the Health Professionals Follow-up Study were supported by grants P01 CA87969 and UM1 CA167552, respectively, from the National Cancer Institute. The Women’s Health Initiative program was supported by contracts N01WH22110, N01WH24152, N01WH32100-32102, N01WH32105, N01WH32106, N01WH32108, N01WH32109, N01WH32111, N01WH32112, N01WH32113, N01WH32115, N01WH32118, N01WH32119, N01WH32122, N01WH42107-N01WH42126, N01WH42129-N01WH42132, and N01WH44221 from the National Heart, Lung, and Blood Institute.
Role of the Funder/Sponsor: The funding sources had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and the decision to submit the manuscript for publication.
Additional Contributions: Brian D. Carter, MA, MPH, helped create the pooled contemporary cohort (PCC) data set, and Dana Flanders, MD, Dsc, provided statistical guidance and editorial critique. Mr. Carter and Dr. Flanders were compensated for their work as an employees of the American Cancer Society.