Background: The relationship between glycated hemoglobin A(1c) (HbA(1c)) and stroke has not been fully elucidated. In addition, there have been few reports from Asia, and few trials have been conducted for each stroke subtype.
Methods: A prospective cohort study was performed involving 32,726 participants (9,558 men and 23,168 women) aged 40-79 years at baseline with approximately 6 years of follow-up in a general population of Japan. The end points included the incidence of all, ischemic and hemorrhagic strokes. Multivariate-adjusted hazard ratios by 6 HbA(1c) categories were calculated using a Cox proportional hazard model.
Results: During the follow-up, 393 participants developed a stroke, including 240 ischemic and 139 hemorrhagic stroke cases. The adjusted hazard ratios of all strokes and ischemic stroke showed tendencies to increase with HbA(1c) level, and the relationships were independent of other confounders. The adjusted hazard ratios of ischemic stroke incidence showed an apparent and graded increase from a relatively mild HbA(1c) level (> or =6.0%); the HRs (95% CI) were 1.79 (1.10-2.93) for HbA(1c) 6.0-6.4%, 2.20 (1.21-4.00) for HbA(1c) 6.5-6.9% and 2.91 (1.91-4.44) for HbA(1c) > or =7.0% compared with the category of HbA(1c) 5.0-5.4%. For hemorrhagic stroke incidence, no significant increase due to rises in HbA(1c) level was observed.
Conclusions: The relationship of the risk of stroke, especially ischemic stroke, to HbA(1c) in the general population appears to be graded without any apparent threshold. The ischemic stroke risk would increase from a relatively mild HbA(1c) level of > or =6.0%.
Copyright 2008 S. Karger AG, Basel.