Methods

Study subjects

The study subjects were Japanese males and females aged 40–79 years who participated in one of three cohort studies conducted between 1980 and 1990 in Japan. These studies were the Japan Public Health Center-Based Prospective (JPHC) Study,18 the Three-Prefecture Cohort Study (TPCS)19 and the JACC study.8 20 21

The study design of these three cohort studies is described in detail elsewhere.8 18–21

In brief, the JPHC Study started in 1990 for the first group (JPHC Study Cohort 1) and in 1993 for the second group (JPHC Study Cohort 2). Study subjects, aged 40–59 years for JPHC Study Cohort 1 and 40–69 for JPHC Study Cohort 2, were recruited from all of the residents in the target areas served by selected public health centres in Japan. The JACC Study cohort was established from 1988 to 1990. Residents aged 40–79 in 45 study areas throughout Japan were enrolled. They were mostly from health checkups provided by the municipalities or from all residents living in a target area. The TPCS collected data from 1983 to 1985 in selected areas of three prefectures in Japan (Miyagi, Aichi and Osaka). An additional study population was sampled in 1990 in one city in Osaka Prefecture. The study population included all residents aged 40 years or older, but those aged 80 years or over were excluded in our study. Study subjects were followed up to the end of 2000 (JPHC Study 1), the end of 2003 (JPHC Study 2), the end of 1999 (JACC Study) or the end of 1993 to 1995 (the end of February 2000 for one city in Osaka prefecture) (TPCS).

The study participants in all three cohort studies filled out a self-administered questionnaire which included questions on smoking behaviour. In the present study, we omitted 1719 males and 2564 females from the JACC Study because they lived in a study area covered by the TPCS. Of the remaining 147 210 males and 177 581 females, we excluded 7184 males and 20 771 females with missing information on smoking behaviour. The number of subjects in this study was 296 836 (140 026 males and 156 810 females).

The study was approved by the human ethics review committees of the National Cancer Center, Tokyo, Japan (No 17-91, 2006).

Definition of smoking status

In the JPHC Study 1, subjects who stated that they had never smoked, ever smoked and currently smoked at the baseline survey were classified as lifelong non-smokers, former smokers and current smokers, respectively. In the JPHC Study 2, the subjects who stated that they currently smoked at baseline were classified as current smokers. Those who stated that they were non-smokers at baseline with responses regarding smoking cessation were classified as former smokers. The remaining subjects were classified as lifelong non-smokers. In the JACC Study, the subjects who stated that they had never smoked at baseline, had quit smoking before the baseline survey, smoked at baseline were classified as never smokers, former smokers and current smokers, respectively. The TPCS used a similar definition of smoking to that in the JACC Study; however, one area of the TPCS used a slightly different definition of lifelong non-smoker; lifelong non-smokers were defined as those who had smoked very little or never.

Mortality surveillance

In each cohort, residential status including survival status was confirmed annually via the residential registry in each area. Mortality surveillance in each cohort was carried out by investigators systematically reviewing death certificates. Cause of death was used to define mortality for total cardiovascular disease (ICD-9 codes 390–459, ICD-10 codes I01-I99), total coronary heart disease (ICD-9 codes 410-414, ICD-10 codes I20–25), total stroke (ICD-9 codes 430–438, ICD-10 codes I60–69) and subtypes of stroke.

Statistical analysis

The gender-specific hazard ratios (HRs) and their 95% CIs were calculated according to smoking status after adjustment for age and cohort using the Cox proportional hazards model.

We used lifelong non-smokers as a reference category for the analysis of HRs among current and former smokers. We then used current smokers as a reference category for the analysis of HRs among former smokers according to the number of years since quitting in which the number of smoking years was further adjusted for. The number of years since quitting was how many years have passed since smoking cessation at baseline.

All analyses were conducted using the SAS statistical package Version 9.1.

Results

Of 296 836 males and females followed up for 2 855 395 person-years (an average of 9.6 years), 7210 participants died from total cardiovascular disease. Of these deaths, 1497 died from coronary heart disease and 3131 died from stroke. The proportions of current smokers and former smokers were 54.4% and 25.0%, respectively, among males and 8.1% and 2.4%, respectively, among females (table 1).

There was a significant association between smoking and cardiovascular disease mortality in both males and females and the magnitude of the effect of smoking was stronger in females than in males (table 2). The interactions with regard to gender were statistically significant (p=0.02 for total cardiovascular disease, p=0.23 for coronary heart disease and p=0.005 for stroke). The magnitude of the effect of smoking varied with stroke subtype.

In both males and females, excessive mortality associated with current smoking was greater in middle-aged subjects than in elderly subjects for total cardiovascular disease, coronary heart disease, stroke and stroke subtypes. The interaction with regard to age groups was statistically significant for all endpoints in both males and females.

A dose-response relation between the number of smoking years and mortality was not evident (table 3). Dose-response relations between the average number of cigarettes smoked per day and mortality from total cardiovascular disease, coronary heart disease and stroke were evident among females, regardless of the number of smoking years. A similar relation for coronary heart disease mortality was identified among males. Male current smokers who started smoking at age 19 years or younger had higher mortality from total cardiovascular disease, coronary heart disease and stroke than those who started at an older age regardless of the number of smoking years. A similar relation for coronary heart disease mortality was identified among female current smokers.

Among male former smokers, the age-adjusted and cohort-adjusted HRs of total cardiovascular disease decreased by 13% at 5–9 years after smoking cessation at baseline, the point at which the benefit of cessation was attained (table 4). At 10–14 years after smoking cessation at baseline, the mortality of former smokers did not differ significantly from that of lifelong non-smokers. The age-adjusted and cohort-adjusted HRs of coronary heart disease decreased by 37% at 10–14 years and by 53% at 15 years or more after smoking cessation at baseline, at which the maximum benefit of cessation was achieved. The age-adjusted and cohort-adjusted HRs of stroke decreased by 35% at 10–14 years after smoking cessation at baseline, at which the benefit of cessation was achieved and mortality risk did not differ from that of lifelong non-smokers. These associations did not differ significantly after adjusting for years of smoking/average number of cigarettes smoked per day.

Among female former smokers, the age-adjusted and cohort-adjusted HRs of total cardiovascular disease increased 2–4 years after smoking cessation at baseline and decreased 5–9 years after smoking cessation at baseline. The age-adjusted and cohort-adjusted HRs of total cardiovascular disease decreased at 5–9 years after smoking cessation at baseline and the maximum benefit of smoking cessation was attained at 15 years or more after smoking cessation at baseline. The age-adjusted and cohort-adjusted HRs of coronary heart disease decreased at 10–14 years after smoking cessation at baseline, when the mortality rate did not differ from that of lifelong non-smokers and the maximum benefit of cessation was attained at 15 years or more after smoking cessation at baseline. The age-adjusted and cohort-adjusted HRs of stroke decreased at 10–14 years after cessation when the mortality rate did not differ from that of lifelong non-smokers; however, no statistically significant benefit of cessation was observed.

To analyse the effect of early deaths among recent quitters, we examined the HRs of total cardiovascular disease mortality by excluding the early deaths that occurred within 3 years from the baseline. The age-adjusted and cohort-adjusted HRs for death from cardiovascular disease among male former smokers according to the number of years since they quit smoking were 0.88 (95% CI 0.68 to 1.15), 0.99 (95% CI 0.85 to 1.16), 0.80 (95% CI 0.69 to 0.92), 0.69 (95% CI 0.58 to 0.81) and 0.55 (95% CI 0.48 to 0.63) at less than 2 years, 2–4 years, 5–9 years, 10–14 years 15 years and more after smoking cessation at baseline, respectively. The respective HRs were 1.03 (95% CI 0.51 to 2.08), 0.87 (95% CI 0.51 to 1.46), 0.73 (95% CI 0.46 to 1.15), 0.55 (95% CI 0.29 to 1.03) and 0.56 (95% CI 0.35 to 0.89) among female former smokers.

Discussion

The present study is the first to identify the effects of smoking cessation on coronary heart disease mortality in Japanese female former smokers. In addition, this study is one of the few studies to identify the dose-response relation between the number of cigarettes smoked per day, age at smoking initiation and cardiovascular disease mortality in Japanese females.

We identified an excess risk of mortality from total cardiovascular disease, coronary heart disease and stroke associated with current smoking in both males and females. The excess risks were more evident in males and females aged 40–64 years than in those aged 65–79 years. A strong association between smoking and mortality from coronary heart disease and subarachnoid haemorrhage was identified in both genders. These findings are consistent with previous Japanese studies.5 6

Our pooled analysis identified a dose-response relation between the number of cigarettes smoked per day and mortality from coronary heart disease but not stroke in males, which was consistent with previous Japanese studies.5 6 22 For female smokers, significant dose-response relations were identified for both coronary heart disease and stroke in the present study. Few studies have examined the dose-response relation in females because of the small number of cases among female smokers. The Ibaraki cohort that consisted of 32 705 males and 63 959 females aged 40–79 years who were residents of Ibaraki prefecture in Japan, reported higher mortality from total cardiovascular disease, particularly coronary heart disease, among female smokers who smoked 20 cigarettes per day compared with those who smoked less than 19 cigarettes per day.23 No dose-response relation regarding total stroke mortality was identified in that study.

Mortality from total cardiovascular disease, coronary heart disease and stroke among current male smokers was higher owing to initiation of smoking at an earlier age, regardless of the number of smoking years. The Hirayama study, which consisted of approximately 260 000 males and females aged 40 years and older throughout Japan who were followed between 1965 and 1978, reported that mortality from coronary heart disease was particularly high in smokers who started smoking aged 15 years or younger.5 Our analysis showed that female smokers who started smoking at 19 years or younger tended to have higher mortality from coronary heart disease compared with smokers who started smoking at 20 years or older. Our results were consistent with those of previous studies in the USA. The Nurse's Health Study reported that female smokers who started smoking before the age of 15 years had the highest mortality from total cardiovascular disease.10 Heavy smokers and smokers who started smoking at a younger age may have poorer health behaviour profiles, which could be a possible reason for clear dose-response relation with cardiovascular mortality.

We found that smoking cessation led to a decline in the risk of mortality from total cardiovascular disease, coronary heart disease and stroke among male former smokers. The estimated risk did not differ between male former smokers and lifelong non-smokers at 10–14 years after smoking cessation. These results were consistent with those from large prospective cohort studies previously conducted in Western countries13 23 and Asian countries5 (figure 1).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Decreasing patterns of coronary heart disease (CHD) and stroke mortality in males after quitting smoking.

Among female former smokers, a reduced risk of total stroke was observed 2–4 years after smoking cessation, while the risk of coronary heart disease mortality increased at 2–4 years after smoking cessation and decreased at 5–9 years after smoking cessation. The estimated risk of total cardiovascular disease, coronary heart disease and stroke among female former smokers did not differ from that among lifelong non-smokers at 10–14 years after smoking cessation. The estimated risk reduction for coronary heart disease at 10 years since smoking cessation was similar to that in previous studies conducted in Western countries (figure 2).24 No other studies have examined the reduction in coronary heart disease mortality resulting from smoking cessation in female former smokers in Asian countries.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Decreasing patterns of coronary heart disease (CHD) and stroke mortality in females after quitting smoking.

Wakai et al25 reported significant reduction of lung cancer mortality risk as a result of smoking cessation using the same pooled analysis data; the age-adjusted and cohort-adjusted HRs of lung cancer decreased by 74% at 15–19 years after smoking cessation at baseline. The benefit of smoking cessation appeared on cardiovascular disease mortality earlier compared to lung cancer in the present pooled analysis. This time lag could be because differences in mechanisms in which smoking causes cardiovascular disease and lung cancer. Smoking effect on cardiovascular disease include short-term effect, such as accelerated thrombus formation through increased plasma fibrinogen,26 increased platelet aggregability,27 increased haematocrit,28 as well as long-term effect such as atheroma formation through direct injury of endothelial cells29 and low high density lipoprotein cholesterol levels.30 On the other hand, smoking effects on lung cancer include mainly long-term effects, such as the carcinogenetic effect.31

The effect of smoking on cardiovascular mortality risk was larger in females than in males. This finding is consistent with the results of previous studies in Western countries24 32 33 and in Asian countries.4 The anti-oestrogenic effect of smoking among females as well as the thrombogenic effect of smoking, together could be a reason for the stronger effect of smoking in females.34 Another reason for the stronger effect of smoking in females may be related to the participants' characteristics. Smoking was relatively uncommon among females, particularly in middle-aged and older females. Female smokers may be significantly different from non-smoking females in other health behaviours and in social situations which could confound the association between smoking and cardiovascular disease mortality. Unfortunately, information on these factors was not obtained.

The magnitude of risk reduction of smoking cessation in JACC study8 was much larger than our results; significant reduction in stroke and coronary heart disease mortality was observed two years after smoking cessation in these studies. On the other hand, we observed delayed benefits of smoking cessation 5 years after smoking cessation. The limited adjustment for confounding factors may underestimate the benefits of smoking cessation in this study, because previous studies showed that the adjustment for the confounding factors lowered the HR of cardiovascular disease among quitters.10 35

Passive smoking at home, to which Japanese females were susceptible owing to the prevalence of male smokers, would have diminished the risk of smoking and benefits of smoking cessation among females. If we could have deleted the effect of passive smoking, we might have identified a larger risk of smoking and larger risk reduction by smoking cessation among females than the observed effects in the present study. Unfortunately, we do not have information on passive smoking at home.

Recent quitters include a disproportionate number of ill people as they quit smoking because of their illness.36 Therefore, short-term benefits of smoking cessation could be obscured by that ill quitter effect. The excess risk of cardiovascular disease mortality among females who had quit smoking for 4 years or less in this study could have been the result of the ill quitter effect. Our additional analysis deleting deaths within 3 years since the beginning of study yielded the larger benefit of smoking cessation among recent quitters.

There are other methodological limitations in the present pooled study. First, misclassification of smoking status could have occurred because of self-report of smoking status. Under-reporting smoking behaviour is expected to generate the underestimation of smoking effect on cardiovascular mortality.37 Second, people who had quit smoking many years before our baseline survey were less likely to be included as participants owing to death or a diagnosis of cardiovascular diseases or cancer. Third, smoking habits could have changed during the follow-up period. Both current smokers who stopped smoking and former smokers who resumed smoking during the follow-up period could have led to the underestimation of smoking cessation effect.

The present study confirmed the association between smoking and mortality from cardiovascular disease in both males and females. Moreover, the findings from this study provided empirical evidence that smoking cessation led to a risk reduction in mortality from cardiovascular disease, coronary heart disease and stroke in both males and females, regardless of the duration of smoking or the number of cigarettes smoked per day. Smoking cessation is a crucial preventive measure against death from cardiovascular disease.