Tobacco smoking and risk of 36 cardiovascular disease subtypes: fatal and non-fatal outcomes in a large prospective Australian study

Abstract

Background: Tobacco smoking is a leading cause of cardiovascular disease (CVD) morbidity and mortality. Evidence on the relation of smoking to different subtypes of CVD, across fatal and non-fatal outcomes, is limited.

Methods: A prospective study of 188,167 CVD- and cancer-free individuals aged ≥ 45 years from the Australian general population joining the 45 and Up Study from 2006 to 2009, with linked questionnaire, hospitalisation and death data up to the end of 2015. Hazard ratios (HRs) for hospitalisation with or mortality from CVD among current and past versus never smokers were estimated, including according to intensity and recency of smoking, using Cox regression, adjusting for age, sex, urban/rural residence, alcohol consumption, income and education. Population-attributable fractions were estimated.

Results: During a mean 7.2 years follow-up (1.35 million person-years), 27,511 (crude rate 20.4/1000 person-years) incident fatal and non-fatal major CVD events occurred, including 4548 (3.2) acute myocardial infarction (AMI), 3991 (2.8) cerebrovascular disease, 3874 (2.7) heart failure and 2311 (1.6) peripheral arterial disease (PAD) events. At baseline, 8% of participants were current and 34% were past smokers. Of the 36 most common specific CVD subtypes, event rates for 29 were increased significantly in current smokers. Adjusted HRs in current versus never smokers were as follows: 1.63 (95%CI 1.56-1.71) for any major CVD, 2.45 (2.22-2.70) for AMI, 2.16 (1.93-2.42) for cerebrovascular disease, 2.23 (1.96-2.53) for heart failure, 5.06 (4.47-5.74) for PAD, 1.50 (1.24-1.80) for paroxysmal tachycardia, 1.31 (1.20-1.44) for atrial fibrillation/flutter, 1.41 (1.17-1.70) for pulmonary embolism, 2.79 (2.04-3.80) for AMI mortality, 2.26 (1.65-3.10) for cerebrovascular disease mortality and 2.75 (2.37-3.19) for total CVD mortality. CVD risks were elevated at almost all levels of current smoking intensity examined and increased with smoking intensity, with HRs for total CVD mortality in current versus never smokers of 1.92 (1.11-3.32) and 4.90 (3.79-6.34) for 4-6 and ≥ 25 cigarettes/day, respectively. Risks diminished with quitting, with excess risks largely avoided by quitting before age 45. Over one third of CVD deaths and one quarter of acute coronary syndrome hospitalisations in Australia aged < 65 can be attributed to smoking.

Conclusions: Current smoking increases the risk of virtually all CVD subtypes, at least doubling the risk of many, including AMI, cerebrovascular disease and heart failure. Paroxysmal tachycardia is a newly identified smoking-related risk. Where comparisons are possible, smoking-associated relative risks for fatal and non-fatal outcomes are similar. Quitting reduces the risk substantially. In an established smoking epidemic, with declining and low current smoking prevalence, smoking accounts for a substantial proportion of premature CVD events.

Keywords: Arrhythmia; Atrial fibrillation; Cardiovascular disease; Cardiovascular mortality; Cerebrovascular disease; Coronary heart disease; Heart failure; Ischaemic heart disease; Smoking; Tobacco.

Fig. 1

Grouped CVD subtype outcomes in relation to smoking status at baseline. a Combined fatal and non-fatal outcomes. b Fatal outcomes only. *Age-sex-standardised rate per 1000 person-years. RR (relative risk) adjusted for age, sex, region of residence, alcohol consumption, annual household income and education attainment; adjustment for sex is through stratification for outcomes of IHD, AMI, non-AMI, IHD, heart failure (stratified for income as well) and major CVD in a and for IHD and major CVD in b. They are plotted on a log scale and are represented by squares of areas proportional to the natural logarithm of the number of events, indicating the amount of statistical information available. Ischaemic heart disease (IHD) (ICD-10-AM codes I20–I25); acute myocardial infarction (I21); IHD, excluding AMI (I20, I22–I25); cerebrovascular disease (I61–I67 and I69); heart failure (I50); peripheral arterial disease (I70–I74); other major CVD (any major CVD except ischaemic heart disease, cerebrovascular disease, heart failure, and peripheral arterial disease); and any major CVD (as defined in reference [15]). Note that the number of events in subtypes will not add up to the total events in any major CVD as the subtypes are not mutually exclusive

Fig. 2

Combined fatal and non-fatal outcomes for specific CVD subtypes (level 3 ICD-10-AM codes) with at least 50 events, in current versus never smokers. *Age-sex-standardised rate per 1000 person-years. RR (relative risk) adjusted for age, sex, region of residence, alcohol consumption, annual household income and education attainment; adjustment for sex is through stratification for outcomes I21, I25, I48, I50 (stratified for income as well) and I10 (stratified for alcohol consumption and income as well). I10 and K55 are not part of “any major CVD” but were included in this table due to their relevance as outcomes related to smoking, see Fig. 1 for details of RR plots

Fig. 3

Grouped CVD subtype outcomes for current smokers, by smoking intensity, compared to never smokers at baseline. a Combined fatal and non-fatal outcomes. b Fatal outcomes only. *Age-sex-standardised rate per 1000 person-years. RR (relative risk) adjusted for age, sex, region of residence, alcohol consumption, annual household income and education attainment; adjustment for sex is through stratification for outcomes of IHD, AMI, non-AMI IHD and major CVD in a. p_trend < 0.0001 for all outcomes. Categories of smoking intensity (never smokers, 1 to 14 cigarettes/day, 15 to 24, and ≥ 25) are based on smoking behaviour reported at baseline. RRs are plotted on a log scale against the median number of cigarettes per day within each of the pre-defined categories, based on smoking intensity reported at follow-up among current smokers. Note different scales used for y-axis, see Fig. 1 for details of RR plots and outcome definitions

Fig. 4

Major CVD outcomes in current smokers, by fine categories of smoking intensity, compared to never smokers at baseline. a Combined fatal and non-fatal outcomes. b Fatal outcomes only. *Age-sex-standardised rate per 1000 person-years. RR (relative risk) adjusted for age, sex, region of residence, alcohol consumption, annual household income and education attainment; adjustment for sex is through stratification for major CVD in a. p_trend < 0.0001 for all outcomes. Categories of smoking intensity (never smokers, 1 to 3 cigarettes/day, 4 to 6, 7 to 9, 10 to 14, 15 to 24, and ≥ 25) are based on smoking behaviour reported at baseline. RRs are plotted on a log scale against the median number of cigarettes per day within each of the pre-defined categories, based on smoking intensity reported at follow-up among current smokers, see Fig. 1 for details of RR plots and outcome definitions

Fig. 5

Grouped CVD subtype outcomes by age at smoking cessation, among past smokers, compared to never smokers. a Combined fatal and non-fatal outcomes. b Fatal outcomes only. *Age-sex-standardised rate per 1000 person-years. RR (relative risk) adjusted for age, sex, region of residence, alcohol consumption, annual household income and education attainment; adjustment for sex is through stratification for outcomes of IHD, AMI, non-AMI IHD, heart failure and major CVD in a and IHD, non-AMI IHD and major CVD in b. They are plotted on a log scale against median values of age categories for stopping smoking (< 25 years old, 25 to 34, 35 to 44, 45 to 54). Dotted lines above the reference lines show RRs for current smokers relative to never smokers. Note different scales used for y-axis, see Fig. 1 for details of RR plots and outcome definitions

Fig. 6

Combined fatal and non-fatal acute myocardial infarction in current smokers versus never smokers, in various population subgroups. *Median number of cigarettes per day smoked by current smokers in the baseline data. **Age-sex-standardised rate per 1000 person-years for all subgroups except for subgroups of age and sex, where crude rates are presented. RR (relative risk) adjusted for age, sex, region of residence, alcohol consumption, annual household income and education attainment; adjustment for sex is through stratification in all models except for subgroups of age and income. The dotted line next to the reference line shows the RR of total current smokers compared to total never smokers, see Fig. 1 for details of RR plots and outcome definition