Mortality in relation to smoking: 50 years' observations on male British doctors.
Level 2 - randomized trial
Prospective cohort study with a dramatic effect size (tripling of age-specific mortality rate and 10-year loss of life expectancy)
PubMed 15213107 · doi:10.1136/bmj.38142.554479.AE
What was done
A prospective cohort study tracked 34,439 male British doctors in the United Kingdom across 50 years (1951 to 2001). Smoking habits were recorded in 1951 and updated periodically thereafter. Cause-specific and overall mortality were monitored and analyzed according to smoking habits across birth cohorts spanning from the 19th century to 1930.
What was found
Men born between 1900 and 1930 who continued smoking cigarettes died on average about 10 years younger than lifelong non-smokers. Quitting at ages 60, 50, 40, or 30 gained approximately 3, 6, 9, or 10 years of life expectancy, respectively. The probability of dying in middle age (35–69) for smokers versus non-smokers was 42% versus 24% (death rate ratio of 2.0) among those born in 1900–1909, and 43% versus 15% (death rate ratio of 3.0) among those born in the 1920s. Surviving from age 70 to 90 at 1990s rates was 7% for smokers versus 33% for non-smokers. Quitting at age 50 halved the excess hazard, while quitting at age 30 avoided almost all excess risk.
Why it matters
This landmark 50-year study provides definitive quantitative evidence of the massive mortality hazard of lifelong smoking, while proving that cessation even up to middle age substantially restores lost life expectancy.
Limits
The study cohort was restricted entirely to male British doctors, limiting direct generalizability to women, other occupational or socioeconomic backgrounds, and non-UK populations. Smoking status relied on periodic self-reporting.
Cited by
- supports Smoking decreases human life expectancy by approximately 10 years.