Q. Qiao 1, M. Tervahauta 2, A. Nissinen 2 and J. Tuomilehto 1. Introduction

Transcription

1 European Heart Journal (2000) 21, doi: /euhj , available online at on Mortality from all causes and from coronary heart disease related to smoking and changes in smoking during a 35-year follow-up of middle-aged Finnish men Q. Qiao 1, M. Tervahauta 2, A. Nissinen 2 and J. Tuomilehto 1 1 Department of Epidemiology and Health Promotion, National Public Health Institute, Helsinki and 2 Department of Public Health and General Practice, University of Kuopio, Kuopio, Finland Aims The risk of early and late death in relation to smoking and ex-smoking were studied. Methods and Results A cohort of 1711 Finnish men born between 1900 and 1919 were recruited in 1959 and followed up for 35 years. Information on smoking status was collected at each of six examinations made from 1959 to 1989 using a standardized questionnaire. Vital status at the end of 1994 was collected for every man. The effect of smoking on mortality was assessed using Cox proportional hazards model. Adjusted ratios for 35-year all-cause mortality were 1 62 (95% CI ) in current smokers and 1 13 (CI ) in former smokers compared with non-smokers. The hazards ratios for 35-year coronary heart disease mortality were 1 63 (CI ) and 1 39 (CI ), respectively. The risk for 10 year mortality was stronger than for 35 year mortality among both former and current smokers, given the same amount of cigarettes consumed. Men smoking persistently were most at risk, while those who persisted in quitting had no increased risk of death compared with non-smokers. Conclusion Smoking increases the risk of premature death in middle-aged men and giving up smoking earlier in life can prevent smoking attributable premature death. (Eur Heart J 2000; 21: , doi: /euhj ) 2000 The European Society of Cardiology Key Words: Smoking, smoking cessation, mortality, coronary heart disease. See page 1570 for the Editorial comment on this article Introduction Revision submitted 1 February 2000, and accepted 2 February Correspondence: Qing Qiao, Department of Epidemiology and Health Promotion, National Public Health Institute, Mannerheimintie 166, FIN-00300, Helsinki, Finland X/00/ $35.00/0 Smoking is known to be a major contributor to mortality from a variety of conditions such as coronary heart disease and lung cancer [1 9]. Former smokers are also at higher risk of death than those who have never smoked, an effect which may be seen for many years after smoking has been given up [10 16]. Despite the decline in the prevalence of cigarette use in many countries, smoking remains the leading cause of preventable death [17 21]. Although a large body of evidence has been found relating cigarette smoking to mortality, results are conflicting [2,4,5,13,15,22,23]. Cigarette smoking was not independently predictive of coronary heart disease in older people in the Framingham study [23], nor in another American study of people aged 50 years or older [22], although in the latter it was a strong predictor of all-cause mortality. Other studies found that the deleterious effects of smoking on total mortality as well as coronary heart disease mortality extend well into later life although the relative risk weakens after the age of 65 [2,4] or 75 years [5]. Total and coronary heart disease mortality in relation to current and former smoking were studied in a cohort of elderly men born in eastern and western Finland between 1900 and In this study we tried to ascertain whether long-term smokers or former smokers who survive to the age of 75 years or older still remain at higher risk of death than non-smokers, and whether the relative risk is higher for the 10-year than for the 35-year death rate. In other words, do the deleterious effects of 2000 The European Society of Cardiology

2 1622 Q. Qiao et al. smoking on total mortality as well as coronary heart disease mortality decline with ageing? In addition, we estimated how much findings from a study based on the data collected at one point have been influenced by changes in smoking habits during a long-term follow-up. Subjects and methods A cohort of 1711 men born in eastern and western Finland between 1900 and 1919 were first examined in 1959 in connection with the Seven Countries Study [24,25]. They were re-examined in 1964, 1969, 1974, 1984 and Information about their smoking habits was collected at each of the six examinations using a standardized questionnaire developed for the Seven Countries Study. According to the answers to the questions, subjects were divided into non-smokers, ex-smokers and current smokers of cigarettes, pipe and cigars. Men who had given up smoking less than 1 year before were classified as current smokers. Daily consumption of cigarettes was expressed as current or former smoking of 1 9, and 20 cigarettes per day. Questionnaires about smoking were completed by 1673 of the 1711 men at the initial examination, giving a response rate of 98%. Of these men 1496 (87%) reported that they had never smoked a pipe or cigar, and these men constituted the population for analysis of cigarette smoking. A new variable, taking into account changes in smoking habits during the period was formed for men who had attended all 1959, 1964 and 1969 examinations (n=1346, 79%). The new variable contained five categories: men who had never smoked before 1969, persistent former smokers from 1959 to 1969, smokers in one of the three examinations, smokers in two of the three examinations and persistent current smokers from 1959 to Mortality data The data on vital status and cause of death were collected over the 35-year follow-up. The diagnostic classification was based on the review of death certificates, collection of medical information and clinical records from hospitals, and interviews with physicians and relatives of the deceased or any other witnesses of fatal events. After the 15th year of follow-up, only the reviewing and coding of official death certificates was performed. These data enabled central reviewers to assign the final underlying cause of death, following standard criteria and using the World Health Organization s International Classification of Diseases, Eighth Revision [26]. In the case of coexisting causes of death, coronary heart disease took precedence over stroke in the hierarchy that assigned underlying cause of death. At the end of the 35 years the vital status of every man in the cohort was known. Procedures for the measurements of other chronic heart disease risk factors such as blood pressure, body mass index and serum cholesterol have been described in detail elsewhere [24,25,27,28]. Statistical analysis Death rates per 1000 person-years from all causes and from coronary heart disease were calculated with adjustment for age and area of residence (east versus west). The mean age of death and the difference between groups were calculated and compared by simple factorial ANOVA adjusted for age, body mass index, systolic blood pressure and serum cholesterol. Cox proportional-hazards regression analyses were performed adjusted for age, area of residence, body mass index, systolic blood pressure and serum cholesterol at baseline. Occupational status at baseline and the interaction of cholesterol and smoking were also adjusted but not included in the final model because they were not significantly associated with mortality. Hazards ratios for death from all causes and from coronary heart disease according to the smoking status at baseline and changes in smoking habits through the first 10 years of the follow-up were estimated separately. Subjects known to have coronary heart disease at baseline were excluded from all analyses related to coronary heart disease mortality. The statistical analyses were performed using SPSS for Windows program version 8 0. Results In 1959, 66% of the cohort were current smokers, 16%, former smokers and 18% men who had never smoked. By 1989, the corresponding proportions were 15%, current, 55%, former and 30%, men who had never smoked. The prevalence of smoking decreased dramatically over this period of 30 years among survivors, whereas less than 1% of the non-smokers and 1% of the former smokers at baseline were smoking in Among the 1673 respondents at baseline examination, a total of person-years was accumulated by the end of 1994 after 35 years of the follow-up. The study observed 1366 deaths from all-causes and 433 from coronary heart disease in men not having coronary heart disease at baseline. The 35-year death rates and the age-adjusted hazards ratios for deaths from all causes and from coronary heart disease are presented in Table 1, stratified by 10-year age groups and by area of residence. An increased risk of death was observed in current smokers and in former smokers who were 50 years or older. Multivariate adjusted hazards ratios show that smokers, both current and former, had higher hazards of death than men who had never smoked before 1959 (Table 2). The risk of death at 10 years was higher than that at 35 years and it was higher for 10-year coronary heart disease mortality than for 10-year all-cause mortality.

3 Smoking and premature death 1623 Table 1 35-year mortality (per 1000 person years) from all causes and coronary heart disease according to smoking status at baseline stratified by areas and age groups (years) at the start of the study in middle-aged Finnish men Area, age and smoking status Number of deaths All-cause mortality (95% CI)* Number of deaths CHD mortality (95% CI)* East years Non-smoker Ex-smoker ( ) ( ) Current smoker ( ) ( ) years Non-smoker Ex-smoker ( ) ( ) Current smoker ( ) ( ) West years Non-smoker Ex-smoker ( ) ( ) Current smoker ( ) ( ) years Non-smoker Ex-smoker ( ) ( ) Current smoker ( ) ( ) *Age-adjusted from Cox regression analyses. CI=confidence interval. CHD=coronary heart disease. The hazards ratios of deaths were also estimated according to the number of cigarettes consumed a day at baseline and the results are presented in Table 3. Hazard ratios for the 35-year mortality increased significantly in former smokers who had smoked 20 or more cigarettes a day and in current smokers who consumed 10 or more cigarettes a day compared with the non-smokers. Generally, the risk for 35-year mortality tended to increase with the increasing number of cigarettes consumed in former and in current smokers compared with the non-smokers. The risk for the 10-year mortality, however, increased in both former and current smokers in spite of the number of cigarettes consumed. Risks did Table 2 s (95% CI) of 10-year and 35-year mortality from all-cause and coronary heart disease in relation to smoking status at baseline Smoking status at baseline Total mortality (n=1673) CHD mortality (n=1463) 10-year mortality Ex-smoker 1 74 ( ) 7 37 ( ) Current smoker 2 16 ( ) 6 80 ( ) 35-year mortality Ex-smoker 1 13 ( ) 1 39 ( ) Current smoker 1 62 ( ) 1 63 ( ) Adjusted for age, area of residence, body mass index, systolic blood pressure, serum cholesterol at baseline. not differ significantly between smokers with regard to the number of cigarettes consumed a day. Table 4 shows the adjusted mean age at death according to the number of cigarettes smoked at baseline. Life expectancy was 2 years shorter for former smokers and 4 5 years shorter for current smokers than for men who had never smoked before The youngest age at death was seen in current smokers who consumed 10 or more cigarettes per day. The impact of changes in smoking habits between the first and the third examinations ( ) was evaluated in terms of mortality at the end of As shown in Table 5, men who were smoking persistently between 1959 and 1969 had the highest hazards ratios of death from all causes and from coronary heart disease in all men. Those who smoked in two of the three examinations had a hazard ratio for all-cause mortality next to that of the persistent current smokers. There was no increased risk of death in men who smoked only in one of the three examinations in those who quit smoking since 1959 compared with the non-smokers. Discussion The cohort of the study was well organized and carefully followed up at 5-year intervals for more than 30 years. Data on smoking status were ascertained not only at baseline but also at each of five examinations during the long-term follow-up. This enabled us to study the longterm deleterious effect of smoking and to examine the extent to which the findings based on the baseline data

4 1624 Q. Qiao et al. Table 3 s (95% CI) of 10-year and 35-year mortality from all-cause and coronary heart disease in relation to the number of cigarettes consumed a day at baseline Baseline cigarettes consumption (number per day) Total mortality (n=1496) CHD mortality (n=1305) 10-year mortality Ex-smoker ( ) 8 18 ( ) ( ) 7 32 ( ) ( ) 7 91 ( ) Current smoker ( ) 7 13 ( ) ( ) 8 28 ( ) ( ) 7 16 ( ) 35-year mortality Ex-smoker ( ) 1 21 ( ) ( ) 1 47 ( ) ( ) 1 60 ( ) Current smoker ( ) 1 13 ( ) ( ) 1 82 ( ) ( ) 1 96 ( ) Adjusted for age, area of residence, body mass index, systolic blood pressure, serum cholesterol at baseline. could be influenced by changes in smoking habits during a long-term follow-up. On the basis of either the baseline smoking status or on the changes in smoking habits, the study indicated that smoking increased the risk of death from all causes and from coronary heart disease in middle-aged men. Because a large proportion of the men gave up smoking during the follow-up, the subsequent quitting might have weakened the observed association of smoking with deaths in current smokers. It is also possible that recidivism among former smokers reduced the benefits of abstinence. However, the number of former smokers reverting to smoking was quite small: among men alive in 1989 only 1% of the former smokers at baseline smoked again in Their influence was small and could not have had a substantial impact on our findings. The risk of death related to smoking was stronger for 10-year mortality than for 35-year mortality, given the same number of cigarettes consumed. A significantly increased risk for 10-year mortality was observed in all smokers including those with minimum tobacco consumption. However, the risk in the light smokers disappeared and in the heavy smokers weakened after 35 years of follow-up. This implies that, with time, the effect of smoking on the risk of death can be reduced and evened out to the level of men who never smoked. The main reason for the fall in hazards ratios in smokers after 35 years of follow-up might be due to the subsequent quitting among the smokers. At the end of 30 years of follow-up, the current smokers only accounted for 15% of the elderly survivors, whereas it was 66% at baseline. In part, the reduction in mortality might be due to selective deaths of those particularly susceptible to the deleterious effects of smoking. Another possible explanation is that other age-associated factors might play a more important role in mortality with ageing, and the impact of smoking therefore become less pronounced. The time required for former smokers to reduce the risk of death to the level of non-smokers differs across studies. It ranges from 10 to 30 years from cessation for all-cause and from 2 to 30 years for coronary heart disease mortality [2,10,12,16,29,30]. In our study, after 35 years of follow-up an excess risk for all-cause and for coronary heart disease mortality was still retained among former smokers who had smoked 20 or more cigarettes a day. However, it was lower than that among Table 4 Adjusted age at death from all causes and from coronary heart disease according to the number of cigarettes consumed per day at baseline Cigarettes consumption (number per day) Total mortality (n=1219) Deaths Mean age (95% CI) (years) CHD mortality (n=388) Deaths Mean age (95% CI) (years) Non-smoker ( ) ( ) Former smoker ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) Current smoker ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) Adjusted for age, body mass index, serum cholesterol and systolic blood pressure at baseline, by ANOVA.

5 Smoking and premature death 1625 Table 5 s (95% CI) of deaths from all-cause and coronary heart disease in relation to the changes in smoking status between 1959 and 1969 Changes in smoking status between 1959 and 1969 Total mortality (n=1346) Number of deaths (%) CHD mortality (n=1205) Number of deaths (%) Non-smoker 253 (18 8) (18 8) 1 Persistent former smoker 176 (13 1) 0 97 ( ) 156 (12 9) 1 12 ( ) Smoked in one of three examinations 116 (8 6) 1 13 ( ) 97 (8 0) 1 07 ( ) Smoked in two of three examinations 143 (10 6) 1 18 ( ) 128 (10 6) 1 08 ( ) Persistent current smoker 658 (48 9) 1 84 ( ) 597 (49 5) 1 84 ( ) Adjusted for age, area of residence, mean of body mass index, systolic blood pressure and serum cholesterol measured between 1959 and current smokers who consumed the same number of cigarettes. The different findings among studies about the time needed for a mortality risk in a former smoker to be reduced to the level of non-smokers may be caused partly by factors such as the lack of ascertainment of smoking status after enrolment. The discrepancy between studies could also be explained by differences in the number of cigarettes consumed and the duration of smoking before giving up. Other risk factors may also play a role; they may not have been fully adjusted for in all studies or there might be interaction/effect modification. Previous studies have found that some smokers are likely to stop because of newly diagnosed disease, and an excess mortality risk within the first few years after giving up smoking has been attributed to the ill quitter effect [16,29]. We have checked the prevalence of coronary heart disease at baseline according to the smoking status at baseline and no significant difference in coronary heart disease prevalence was found between former smokers and others. Also, men who were known to have coronary heart disease at baseline were not included in the analyses related to coronary heart disease mortality. Therefore, the increased risk of coronary heart disease mortality in former smokers was unlikely to be the consequence of the increased coronary heart disease at baseline. Nor would one expect any ill quitter effect among former smokers who had given up for the more than 35 years, although the ill quitter phenomenon might result in an under-estimation of the short-term benefits of giving up. Among the study cohort all-cause mortality increased sharply with ageing and the duration of follow-up, as expected. It reached 30% in the first 15 years of the follow-up and 70% and 82% by 30 and 35 years of follow-up, respectively. In order to avoid the substantial loss of observation due to deaths, when changes in smoking habits were being taken into account we used the data from the first 10 years of the follow-up only. In this way, 79% (n=1346) of the cohort were included in the analyses. The power of the study was reduced and there might have been a selection bias in the results. Nevertheless, the results based on the judgement of change were consistent with those calculated from the baseline data among the whole study population. The results illustrated clearly the association of smoking with the risk of death. The more cigarettes consumed and the longer the duration of smoking the higher the risk of death. It may be argued that the increased risk of death among quitters might have been over-estimated because the recidivism usually occurs within the first few years of quitting among heavy smokers [14,31]. However, in this study less than 10% of the quitters had given up smoking less than 1 year from baseline and they were defined as current smokers in the analyses. Moreover, the results did not change after excluding these men from the analyses. In conclusion, our observations are consistent with those of other studies among middle-aged populations [16,32] and confirm that men who were smokers at any time during middle ages of life run an increased risk of premature death from all causes and from coronary heart disease. The deleterious effects of smoking extended to the age of 75 years or older among the long-term smokers as well as former smokers who had consumed more than 20 cigarettes a day. The risk, however, weakened with ageing. Altogether, our study suggests that giving up smoking early in life or reducing the amount of smoking may prevent smoking attributable premature death and preserve life expectancy. This study was supported by grants from Yrjo Jahnsson s Foundation, Academy of Finland, the Sandoz Foundation for Gerontological Research and the National Institute of Health (AG-08762). References [1] U.S. Public Health Service. Smoking and Health. Report of the Advisory Committee to the Surgeon General of the Public Health Service. Department of Health, Education, and Welfare, Public Health Service, Centre for Disease Control, PHS Publ. no Washington, DC: Government Printing Office, 1964.

6 1626 Q. Qiao et al. [2] Hammond EC. Smoking in relation to the death rates of one million men and women. National Cancer Institute Monograph No. 19. Bethesda, MD, U.S.A.: NIH, 1966: [3] Carstensen JM, Pershagen G, Eklund G. Mortality in relation to cigarette and pipe smoking: 16 years observation of Swedish men. J Epidemiol Community Health 1987; 41: [4] Doll R, Peto R. Mortality in relation to smoking: 20 years observations on male British doctors. Br Med J 1976; ii: [5] LaCroix AZ, Lang J, Scherr P et al. Smoking and mortality among older men and women in three communities. N Engl J Med 1991; 324: [6] Tverdal A, Thelle D, Stensvold I, Leren P, Bjartveit K. Mortality in relation to smoking history: 13 years follow-up of Norwegian men and women years. J Clin Epidemiol 1993; 46: [7] Hammond EC, Horn D. Smoking and death rates-report on forty-four months of follow-up of men. J Am Med Assoc 1958; 166: [8] Reid DD, Hamilton PJS, McCartney P, Rose G. Smoking and other risk factors for coronary heart-disease in British civil servants. Lancet 1976; (7993): [9] Doll R, Hill AB. Lung cancer and other causes of death in relation to smoking. A second report on the mortality of British doctors. Br Med J 1956; ii: [10] Dobson AJ, Alexander HM, Heller RF, Lloyd DM. How soon after quitting smoking does risk of heart attack decline? J Clin Epidemiol 1991; 44: [11] Cook DG, Shaper AG, Pocock SJ, Kussick SJ. Giving up smoking and the risk of heart attacks. Lancet 1986; 2: [12] Ben-Shlomo Y, Smith GD, Shipley MJ, Marmot MG. What determines mortality risk in male former cigarette smokers? Am J Public Health 1994; 84: [13] Halpern MT, Gillespie BW, Warner KE. Patterns of absolute risk of lung cancer mortality in former smokers. J Natl Cancer Inst 1993; 85: [14] Ockene JK, Kuller LH, Svendsen KH, Meilhan E. The relationship of smoking cessation to coronary heart disease and lung cancer in the multiple risk factor intervention trial (MRFIT). Am J Public Health 1990; 80: [15] Gordon T, Kannel WB, McGee D, Dawber TR. Death and coronary attacks in men after giving up cigarette smoking. A report from Framingham Study. Lancet 1974; ii: [16] Kawachi I, Colditz GA, Stampfer MJ et al. Smoking cessation in relation to total mortality rates in women. Ann Intern Med 1993; 119: [17] Austoker J, Sanders D, Fowler G. Smoking and cancer: smoking cessation. Br Med J 1994; 308: [18] Editorial. Smoking and health: a 25-year perspective. Am J Public Health 1989; 79: [19] Pierce JP. International comparisons of trends in cigarette smoking prevalence. Am J Public Health 1989; 79: [20] Gritz ER. Reaching toward and beyond the year 2000 goals for cigarette smoking. Cancer 1994; 74: [21] American Thoracic Society. Cigarette smoking and health. Am J Respir Crit Care Med 1996; 153: [22] Barrett-Connor E, Suarez L, Khaw K-T, Criqui MH, Wingard DL. Ischemic heart disease risk factors after age 50. J Chronic Dis 1984; 37: [23] Gordon T, Castelli WP, Hjortland MC, Kannel WB, Dawber TR. Predicting coronary heart disease in middle-aged and older persons. The Framingham Study. J Am Med Assoc 1977; 238: [24] Karvonen MJ, Blomqvist G, Kallio V. Epidemiological studies related to coronary heart disease: characteristics of men aged in seven countries. C4: men in rural east and west Finland. Acta Med Scand 1966; 460 (Suppl): [25] Keys A. Coronary heart disease in seven countries. New York: American Heart Association, Monograph No. 29, [26] WHO: International Classification of Disease. Eighth Revision. Geneva, Switzerland: World Health Organisation, [27] Nissinen A, Tervahauta M, Pekkanen J et al. Prevalence and change of cardiovascular risk factors among men born : the Finnish cohort of the Seven Countries Study. Age Ageing 1993; 22: [28] Nissinen A, Kivelä S-L, Pekkanen J et al. Levels of some biological risk indicators among elderly men in Finland. Age Ageing 1985; 15: [29] U.S. Department of Health and Human Services. Washington, DC: The Health benefits of smoking cessation. U.S. Department of Health and Human Services, Public Health Service, Centers for Disease Control and Prevention, Centers for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health, DHHS Pub. no. (CDC) , [30] Rogot E, Murray JL. Smoking and causes of death among U.S. veterans: 16 years of observation. Public Health Rep 1980; 95: [31] Sorlie PD, Kannel WB. A description of cigarette smoking cessation and resumption in the Framingham study. Pre Med 1990; 19: [32] Peto R, Lopez AD, Boreham J, Thun M, Heath C Jr. Mortality from tobacco in developed countries: indirect estimation from national vital statistics. Lancet 1992; 339: