Association of Smoking Initiation and Cessation Across the Life Course and Cancer Mortality: Prospective Study of 410 000 US Adults

Smoking remains the leading cause of cancer mortality in the US.¹ Prior research has shown that smokers who start younger have greater mortality risk than those who start later, and quitting (especially at younger ages) substantially reduces that risk.^2,3 However, the relevance of age at smoking initiation and cessation to cancer mortality in contemporary populations, particularly across the life course, is not well established.

In this prospective cohort study, we investigated the association between age at smoking initiation and cessation and cancer mortality (C00-C97 in the International Classification of Diseases, 10th Revision) at ages 25 to 79 years using data from the US National Health Interview Survey for 1997 to 2014 linked to the National Death Index, with follow-up through December 31, 2015.⁴ Self-reported current daily smokers were categorized by age at smoking initiation (<10, 10-14, 15-17, 18-20, or ≥21 years), and ex-smokers were categorized by age at quitting (15-34, 35-44, 45-54, or 55-64 years). Current nondaily smokers (4% of cohort) and ex-smokers who quit at ages younger than 15 years or 65 years and older (1%) were excluded.

Cox regression was used to estimate cancer mortality rate ratios (RRs) for current and ex-smokers vs never smokers. To account for differences between smoking groups (shown in eTable in the Supplement), analyses adjusted for age at risk (11 groups), sex, race and ethnicity (5 groups), education (4 groups), region (4 groups), and alcohol consumption (5 groups). Race and ethnicity were self-reported and grouped into 5 categories to account for previously reported differences in smoking habits and mortality risk by race and ethnicity. Race and ethnicity are collected in the National Health Interview Survey to report the health status of the nation according to major demographic factors. To limit the bias of quitting because of ill health,⁵ those who quit within 5 years of death were recategorized as current smokers. Attributable fractions among current smokers and excess risk avoided among those who quit were also estimated. The study was based on deidentified publicly available data, which is considered nonhuman participants research under the US Department of Health and Human Services’ Office for Human Research Protection (the Common Rule) and does not require institutional review board review or informed consent. Analyses were performed using Stata, version 15.1 (StataCorp LLC) and R, version 4.0.5 (R Project for Statistical Computing).

Among 410 231 participants included (mean [SD] age, 48 [15] years; 56% female), there were 10 014 cancer deaths at ages 25 to 79 years during 3.7 million person-years of follow-up (mean [SD], 10 [5]). Compared with never smokers, the overall cancer mortality RR associated with current smoking was 3.00 (95% CI, 2.86-3.16). For those starting at ages younger than 10, 10 to 14, 15 to 17, 18 to 20, and 21 years and older, the RRs (95% CI) were 4.01 (3.33-4.82), 3.57 (3.29-3.87), 3.15 (2.94-3.37), 2.86 (2.65-3.08), and 2.44 (2.24-2.67), respectively (Figure 1). Therefore, if these excesses were interpreted as largely causal, smoking would account for 75% of cancer deaths among those starting before age 10 years and 59% among those starting at age 21 years and older (Figure 2). By contrast, compared with never smokers, the RRs (95% CI) for quitting at ages 15 to 34, 35 to 44, 45 to 54, and 55 to 64 years were 0.95 (0.88-1.04), 1.23 (1.13-1.33), 1.45 (1.34-1.57), and 1.88 (1.73-2.05), respectively (Figure 1). Thus, those who quit smoking at ages 15 to 34, 35 to 44, 45 to 54, and 55 to 64 years avoided an estimated 100%, 89%, 78%, and 56% of the excess cancer mortality risk associated with continued smoking, respectively (Figure 2).

In this contemporary US population, current smoking was associated with 3 times the cancer mortality rate of never smoking. Although delayed smoking initiation was associated with substantial reductions in cancer mortality risk, those who began smoking at ages 21 years and older nevertheless had more than twice the risk compared with never smokers. Quitting, even at older ages, was associated with a substantial reversal of these risks.

In December 2019, the federal legal age to purchase tobacco was raised from 18 to 21 years.⁶ Given that most cancer deaths among smokers were attributable to smoking, even among those who started smoking at ages 21 years and older, the effect of such legislation on US cancer mortality will depend on the extent to which it both delays smoking initiation and prevents initiation entirely.

These findings underscore that starting to smoke at any age is extremely hazardous, but smokers who quit—especially at younger ages—can avoid most of the cancer mortality risk associated with continued smoking.

Accepted for Publication: July 26, 2021.

Published Online: October 21, 2021. doi:10.1001/jamaoncol.2021.4949

Corresponding Author: Blake Thomson, DPhil, Cancer Disparity Research, Department of Surveillance and Health Equity Science, American Cancer Society, 3380 Chastain Meadows Pkwy, Ste 200, Kennesaw, GA 30144 (blake.thomson@cancer.org).

Author Contributions: Dr Thomson had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Thomson, Peto.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Thomson.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Thomson, Peto.

Administrative, technical, or material support: Lacey.

Supervision: Emberson, Lewington, Islami.

Conflict of Interest Disclosures: Drs Thomson and Islami reported being employed by the American Cancer Society during the conduct of the study. Dr Emberson reported receiving a grant to the University of Oxford for the EMPA-KIDNEY trial from Boehringer Ingelheim, grants to the University of Oxford for research related to the Mexico City Prospective Study from Regeneron and AstraZeneca, and other (performance of genetic assays of 150 000 baseline samples from the Mexico City Prospective Study) from Regeneron outside the submitted work. Dr Lewington reported grants from UK Medical Research Council and CDC Foundation (with support from Amgen) outside the submitted work. No other disclosures were reported.

Funding/Support: This work was supported by the Intramural Research Department of the American Cancer Society.

Role of the Funder/Sponsor: The American Cancer Society 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; or decision to submit the manuscript for publication.