BACKGROUND: An international comparison has indicated that the association between smoking and cardiovascular disease (CVD) differs according to total cholesterol (TC) levels. However, little has been published about the relationship between smoking and CVD mortality among populations with various cholesterol levels. METHODS: We calculated the adjusted relative hazard (RH) of smoking for CVD mortality among 8912 Japanese individuals without a history of stroke or heart disease, who were separated according to TC levels of >or=5.40, 4.81-5.39, 4.26-4.80 and <4.25 mmol/l into groups Q4, Q3, Q2 and Q1, respectively. The P-values for multiple interactions between TC and smoking status for CVD mortality were calculated using TC as a continuous variable, dichotomized smoking status (never vs current), and by including cross-product terms in the regression models. RESULTS: After 19 years of follow-up, 313 men and 291 women died of CVD. The RH of CVD mortality among men who currently smoked compared with those who never smoked was increased with higher TC (RH = 2.36 in Q4) and decreased in those with lower TC (RH = 0.85 in Q1) (interaction, P < 0.01). The profiles for coronary heart disease (CHD) mortality and ischaemic CVD (composite endpoint of CHD and ischaemic stroke) in men and for ischaemic CVD mortality in women were identical. The interaction might be explained by a biological mechanism and by frailty of those who have never smoked with lower TC. CONCLUSIONS: Counteractive measures should be implemented against smoking targeted towards Japanese with elevated TC levels.
BACKGROUND: An international comparison has indicated that the association between smoking and cardiovascular disease (CVD) differs according to total cholesterol (TC) levels. However, little has been published about the relationship between smoking and CVD mortality among populations with various cholesterol levels. METHODS: We calculated the adjusted relative hazard (RH) of smoking for CVD mortality among 8912 Japanese individuals without a history of stroke or heart disease, who were separated according to TC levels of >or=5.40, 4.81-5.39, 4.26-4.80 and <4.25 mmol/l into groups Q4, Q3, Q2 and Q1, respectively. The P-values for multiple interactions between TC and smoking status for CVD mortality were calculated using TC as a continuous variable, dichotomized smoking status (never vs current), and by including cross-product terms in the regression models. RESULTS: After 19 years of follow-up, 313 men and 291 women died of CVD. The RH of CVD mortality among men who currently smoked compared with those who never smoked was increased with higher TC (RH = 2.36 in Q4) and decreased in those with lower TC (RH = 0.85 in Q1) (interaction, P < 0.01). The profiles for coronary heart disease (CHD) mortality and ischaemic CVD (composite endpoint of CHD and ischaemic stroke) in men and for ischaemic CVD mortality in women were identical. The interaction might be explained by a biological mechanism and by frailty of those who have never smoked with lower TC. CONCLUSIONS: Counteractive measures should be implemented against smoking targeted towards Japanese with elevated TC levels.