Short Report
Health Consequences of Smoking Among Canadian Smokers: An Update

Larry F Ellison, Howard I Morrison, Margaret de Groh and Paul J Villeneuve

Abstract

The expected number of deaths in four hypothetical Canadian cohorts (male current smokers, male never-smokers, female current smokers and female never-smokers) was examined by constructing abridged life tables. The expected number of premature deaths (before age 70)  among lifelong smokers was found to be about twice that expected among lifelong never-smokers for both males (2.3) and females (1.9). The higher number of premature deaths in the smoking cohorts resulted mainly from cancer and coronary heart disease. The results of this paper highlight the dramatic impact that smoking has on premature mortality.

Key words: Canada; coronary disease; life table; mortality; neoplasms; smoking

Introduction

Premature death due to tobacco use is the most important public health problem facing Canadians today. It has been estimated that at least one quarter of all deaths among persons aged 35-84 in Canada are attributable to tobacco use.¹ An estimated 45,000 deaths in Canada in 1991 were caused by smoking.²

This study was designed to update and refine previous work^3,4 exploring premature mortality (before age 70) attributable to smoking. The most recent Canadian study modelling such mortality⁴ has become outdated, in part because of changes in mortality rates, particularly for coronary heart disease. We modelled the expected number of deaths in four hypothetical cohorts (male current smokers, male never-smokers, female current smokers and female never-smokers), starting from age 15.

Methods

Age-specific prevalence rates of current smokers were estimated from the 1996 National Population Health Survey.⁵ Relative risk estimates for smoking-related diseases were derived from data from the American Cancer Society's Cancer Prevention Study II (American Cancer Society, personal communication, 1998), which is examined in detail elsewhere.⁶ Mortality rates according to age, sex and cause were calculated using death counts retrieved from the Canadian Mortality Database of Statistics Canada and Canadian population data, adjusted for census undercount, also obtained from Statistics Canada. It was assumed that the cohorts would be subject to these mortality rates over the course of their lifetime.

The method outlined by Mattson et al.⁷ was used to calculate mortality rates for current smokers and never-smokers using the aforementioned relative risk, prevalence and mortality rates. Abridged life tables were constructed⁸ for each of the four cohorts (i.e. male current smokers, male never-smokers, female current smokers and female never-smokers). From these life tables, the probabilities of dying in each age interval were retrieved. Expected numbers of deaths were then estimated by multiplying age-specific probabilities of death by the number of surviving members of the respective cohort, a figure taken from the constructed life tables. The expected total number of deaths for each cohort was based on an initial population of 100,000 persons aged 15 years. Those in the smoking cohorts were assumed to be smokers for the duration of their lives. Deaths attributable to smoking were calculated by subtracting the number of deaths among never-smokers from the number of deaths among current smokers.

For each cause of death, probabilities of death for each age interval were calculated in a similar fashion to that already described. Expected numbers of deaths were estimated by multiplying age- and cause-specific probabilities of death by the number of surviving members of the respective cohort. Risks to smokers were assumed to be equivalent to never-smokers for homicide, motor vehicle traffic accidents, HIV/AIDS and suicide. To form a direct basis of comparison, the process was repeated using smoking prevalence rates⁹ and mortality figures from 1990.

Results

More than a third of the cohort of 100,000 male smokers aged 15, and almost a quarter of a similar group of females, were anticipated to die before age 70 (Table 1). The expected number of premature deaths among smokers was found to be about twice that expected in similar cohorts of never-smokers for both males (2.3) and females (1.9). The higher number of premature deaths among the smoking cohorts resulted mainly from cancer and coronary heart disease.

Smoking accounted for 56% and 48% of premature mortality among male and female smokers respectively. Among male smokers, approximately 3.5% of premature deaths were due to suicide, 2.1% to motor vehicle accidents, 1.4% to HIV/AIDS and 0.4% to homicide. Similarly, among female smokers, 1.5% of premature deaths were due to motor vehicle accidents, 1.6% to suicide, 0.3% to homicide, and 0.2% to HIV/AIDS.

Table 2 presents the same type of information as Table 1 but uses 1990 rather than 1996 smoking prevalence and mortality data. The number of premature deaths attributable to smoking was 5.5% lower using 1996 compared with 1990 data. The largest decreases were observed for cardiovascular disease; coronary heart disease was 18% lower, and cerebrovascular disease was 8% lower. For chronic obstructive pulmonary disease and cancer, particularly the former, the number of premature deaths attributable to smoking declined among males but increased among females.

Discussion

The results of this paper highlight the dramatic impact that smoking has on premature mortality. Compared with non-smokers, the risk of premature death is more than double among males and almost double among females who begin smoking by age 15. Over half of the expected premature deaths would be attributable to smoking as compared with less than 6% from suicide, motor vehicle traffic accidents, HIV/AIDS and homicide combined. Like other researchers,^3,4 we observed that the higher number of premature deaths among the smoking cohorts resulted mainly from cancer and coronary heart disease.

The estimate of the expected premature mortality attributable to smoking was over 10% lower than that previously reported in a similar hypothetical study³ using 1990 smoking prevalence and mortality data. However, methodologic differences between these two studies made a direct comparison problematic. Accordingly, we duplicated our analysis using data from 1990. Our 1996 estimates represent a 6% lower number of smoking-attributable premature deaths than the 1990 estimates. Differences between the two sets of estimates can be explained largely by changes in mortality rates that have occurred between 1990 and 1996, and they may reflect improved survival for such conditions as coronary heart disease and cerebrovascular disease. While the mortality rate for chronic obstructive pulmonary disease has decreased over this period among men, it has increased among women, particularly women aged 65-69.

We observed nearly twice as many smoking- attributable deaths among males than among females. This is, in part, the result of sex differences in age-specific mortality rates and relative risk estimates. The relative risk estimates used in this analysis partially reflect past differences in patterns of smoking (e.g. age at initiation, number of cigarettes smoked daily) between men and women.¹⁰ Although forecasting changes in the prevalence of smoking, mortality rates and underlying relative risks was beyond the scope of this report, it is reasonable to assume that if sex-specific patterns of smoking were similar then differences in smoking- attributable death counts between males and females would be considerably narrowed.

A range of interventions, from educating the public about the adverse health effects of tobacco use to advertising restriction and other legislative initiatives (e.g. taxation), contributed to significant decreases in the prevalence of tobacco use during the 1980s.^10,11 Nevertheless, 30% of Canadians 15 years and older still smoke, and teen smoking rose sharply during the early 1990s.^12-14 Since most smokers start this highly addictive habit during adolescence,¹⁰ the pattern of smoking among youth will shape the future health care burden and the number who will prematurely die. A compre- hensive approach to preventing youth smoking is necessary to reduce the number of Canadians who will die from smoking-related diseases.^10,15

Acknowledgements

We would like to thank the American Cancer Society for providing data on the Cancer Prevention Study II.

References

1. Collishaw NE, Tostowaryk W, Wigle DT. Mortality attributable to tobacco use in Canada. Can J Public Health 1988;79:166-9.

2. Ellison LF, Mao Y, Gibbons L. Projected smoking-attributable mortality in Canada, 1991-2000. Chronic Dis Can 1995;16(2):84-9.

3. Villeneuve P, Morrison H. Health consequences of smoking in Canada: an update. Chronic Dis Can 1994;15(3):102-4.

4. Mao Y, Morrison H, Nichol RD, Pipe A, Wigle D. The health consequences of smoking among smokers in Canada. Can J Public Health 1988;79:390-1.

5. Statistics Canada. National Population Health Survey, 1996/97. Household component, user's guide for the public use microdata files. Ottawa, 1998; Cat 82-M0009GPE.

6. Garfinkel L. Selection, follow-up, and analysis in the American Cancer Society prospective studies. Natl Cancer Inst Monogr 1985;67:49-52.

7. Mattson ME, Pollack ES, Cullen JW. What are the odds that smoking will kill you ? Am J Public Health 1987;77:425-31.

8. Selvin S. Statistical analysis of epidemiologic data. New York: Oxford University Press, Inc., 1991:263-9.

9. Health and Welfare Canada (Stephens T, Fowler Graham D, eds). Canada's Health Promotion Survey 1990: technical report. Ottawa, 1993; Cat H39-263/2-1990E.

10. US Department of Health and Human Services. Reducing the health consequences of smoking: 25 years of progress. A report of the Surgeon General. Rockville (MD): Public Health Service, Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health; 1989; DHHS Pub No (CDC) 89-8411.

11. Lewitt EM, Coate D, Grossman M. The effects of government regulation on teenage smoking. J Law Economics 1981;24:545-69.

12. Clark W. Youth smoking in Canada. Can Social Trends 1996;43: 2-7. (Statistics Canada Cat 11-008E).

13. Poulin C, Wilbur B. Nova Scotia student drug use 1996: technical report. Province of Nova Scotia, 1996.

14. Adlaf EM, Ivis FJ, Smart RG, Walsh GW. Ontario Student Drug Use Survey, 1977-1995. Toronto: Addiction Research Foundation, 1995.

15. Health Canada. The directional paper of the National Strategy to Reduce Tobacco Use: an update, 1993. Ottawa, 1993. 

[Previous][Table of Contents] [Next]