Association of childhood smoking and adult mortality: prospective study of 120 000 Cuban adults

Abstract

Background: The average age at which people start smoking has been decreasing in many countries, but insufficient evidence exists on the adult hazards of having started smoking in childhood and, especially, in early childhood. We aimed to investigate the association between smoking habits (focusing on the age when smokers started) and cause-specific premature mortality in a cohort of adults in Cuba.

Methods: For this prospective study, adults were recruited from five provinces in Cuba. Participants were interviewed (data collected included socioeconomic status, medical history, alcohol consumption, and smoking habits) and had their height, weight, and blood pressure measured. Participants were followed up until Jan 1, 2017 for cause-specific mortality; a subset was resurveyed in 2006-08. We used Cox regression to calculate adjusted rate ratios (RRs) for mortality at ages 30-69 years, comparing never-smokers with current smokers by age they started smoking and number of cigarettes smoked per day and with ex-smokers by the age at which they had quit.

Findings: Between Jan 1, 1996, and Nov 24, 2002, 146 556 adults were recruited into the study, of whom 118 840 participants aged 30-69 years at recruitment contributed to the main analyses. 27 264 (52%) of 52 524 men and 19 313 (29%) of 66 316 women were current smokers. Most participants reported smoking cigarettes; few smoked only cigars. About a third of current cigarette smokers had started before age 15 years. Compared with never-smokers, the all-cause mortality RR was highest in participants who had started smoking at ages 5-9 years (RR 2·51, 95% CI 2·21-2·85), followed by ages 10-14 years (1·83, 1·72-1·95), 15-19 years (1·56, 1·46-1·65), and ages 20 years or older (1·50, 1·39-1·62). Smoking accounted for a quarter of all premature deaths in this population, but quitting before about age 40 years avoided almost all of the excess mortality due to smoking.

Interpretation: In this cohort of adults in Cuba, starting to smoke in childhood was common and quitting was not. Starting in childhood approximately doubled the rate of premature death (ie, before age 70 years). If this 2-fold mortality RR continues into old age, about half of participants who start smoking before age 15 years and do not stop will eventually die of complications from their habit. The greatest risks were found among adults who began smoking before age 10 years.

Funding: UK Medical Research Council, Cancer Research UK, British Heart Foundation, US Centers for Disease Control and Prevention (CDC) Foundation (with support from Amgen).

Figure 1

Smoking prevalence in the baseline survey (1996–2002) by year of birth Proportion of participants at baseline who were current smokers of cigarettes, current smokers of cigars only, or ex-smokers (of either cigarettes or cigars). Years of birth are divided into the following categories: before 1935, 1935–39, 1940–44, 1945–49, 1950–54, 1955–59, and 1960 or after.

Figure 2

All-cause mortality at ages 30–69 years for cigarette smokers versus never-smokers by age when starting to smoke (A) and by usual number of cigarettes smoked per day (B) Box area is inversely proportional to the variance of the log risk. Error bars denote group-specific 95% CIs. Analyses were adjusted for age, sex, education, alcohol consumption, body–mass index, and province. (A) Age at which participants started smoking regularly was divided into the following categories: 5–9 years, 10–14 years, 15–19 years, and 20 years or older. (B) Participants were categorised by the number of cigarettes smoked per day at baseline (<10, 10–19, 20, and >20) and the mortality rate ratios plotted against the mean number of cigarettes smoked per day at resurvey in these baseline-defined groups.

Figure 3

Cause-specific mortality at ages 30–69 years for cigarette smokers versus never-smokers by age when starting to smoke Box area is inversely proportional to the variance of the log risk. Error bars denote group-specific 95% CIs. Analyses were adjusted for age, sex, education, alcohol consumption, body-mass index, and province. COPD=chronic obstructive pulmonary disease. IHD=ischaemic heart disease.

Figure 4

All-cause mortality at ages 30–69 years for ex-cigarette smokers versus never-smokers by age when stopping smoking Box area is inversely proportional to the variance of the log risk. Error bars denote group-specific 95% CIs. Analyses adjusted for age, sex, education, alcohol consumption, body–mass index, and province. Age when stopping smoking was divided into the following groups: 25–34 years (45 deaths), 35–44 years (61 deaths), and 45–54 years (45 deaths).

Figure 5

All-cause mortality in the USA at ages 30–69 years for cigarette smokers versus never-smokers by age when starting to smoke Box area is inversely proportional to the variance of the log risk. Error bars denote group-specific 95% CIs. Age when starting to smoke regularly was divided into the following categories: younger than 10 years (233 deaths), 10–14 years (1566 deaths), 15–17 years (2237 deaths), 18–20 years (1539 deaths), and older than 20 years (1116 deaths). Data sources were the US National Health Interview Surveys and US National Death Index. Analyses were adjusted for age at risk, sex, race, education, alcohol consumption, and region.