Table 3.

Effect of smoking habit change with and without fasting serum glucose elevation on cardiovascular disease-related death.

		Continual smokers	Quitters	Ex-smokers	Never smokers
Cardiovascular disease-related death	Without FSG elevation*
	Events	133	34	72	127
	Person-years	194,009	58,546	114,444	206,670
	aHR (95% CI)	1.00 (reference)	0.74 (0.51–1.09)	0.68 (0.50–0.91)	0.59 (0.46–0.76)
	With FSG elevation
	Events	120	30	42	100
	Person-years	146,281	46,952	80,915	146,282
	aHR (95% CI)	1.00 (reference)	0.68 (0.46–1.03)	0.49 (0.34–0.70)	0.58 (0.44–0.77)
Myocardial infarction-related death	Without FSG elevation
	Events	68	16	33	51
	Person-years	194,009	58,546	114,444	206,670
	aHR (95% CI)	1.00 (reference)	0.68 (0.39–1.18)	0.67 (0.44–1.03)	0.50 (0.34–0.73)
	With FSG elevation
	Events	59	6	20	29
	Person-years	146,281	46,952	80,915	146,282
	aHR (95% CI)	1.00 (reference)	0.26 (0.11–0.62)	0.50 (0.30–0.84)	0.36 (0.22–0.57)
Total stroke-related death	Without FSG elevation
	Events	65	18	39	76
	Person-years	194,009	58,546	114,444	206,670
	aHR (95% CI)	1.00 (reference)	0.80 (0.47–1.35)	0.69 (0.46–1.03)	0.68 (0.48–0.96)
	With FSG elevation
	Events	61	24	22	71
	Person-years	146,281	46,952	80,915	146,282
	aHR (95% CI)	1.00 (reference)	1.11 (0.69–1.80)	0.48 (0.29–0.79)	0.79 (0.55–1.13)
Ischemic stroke-related death	Without FSG elevation
	Events	19	5	15	32
	Person-years	194,009	58,546	114,444	206,670
	aHR (95% CI)	1.00 (reference)	0.67 (0.25–1.82)	0.76 (0.38–1.53)	0.82 (0.45–1.50)
	With FSG elevation
	Events	14	10	5	524
	Person-years	146,281	46,952	80,915	146,282
	aHR (95% CI)	1.00 (reference)	2.02 (0.88–2.64)	0.48 (0.17–1.35)	1.16 (0.58–2.33)
Hemorrhagic stroke-related death	Without FSG elevation
	Events	25	2	15	22
	Person-years	194,009	58,546	114,444	206,670
	aHR (95% CI)	1.00 (reference)	—	0.81 (0.42–1.58)	0.59 (0.32–1.10)
	With FSG elevation
	Events	25	13	10	20
	Person-years	146,281	46,952	80,915	146,282
	aHR (95% CI)	1.00 (reference)	1.28 (0.79–3.15)	0.61 (0.29–1.30)	0.63 (0.34–1.18)

*FSG elevation defined by an elevation of fasting serum glucose of more than the upper limit of 95% confidence interval for FSG change among quitters (>4.17 mg/dL)

Hazard ratio calculated by Cox proportional hazards regression analysis with adjustments for age, socioeconomic status, physical activity, alcohol consumption, body mass index, baseline fasting serum glucose, blood pressure, total cholesterol, and Charlson comorbidity index

Acronyms: FSG, fasting serum glucose; aHR, adjusted hazard ratio; CI, confidence interval

Blank cells indicate no reliable values of hazard ratios and 95% confidence intervals due to the small number of events.