AIM: To examine the association between treatment-achieved HbA1c values and incidence of both coronary artery disease (CAD) and severe eye disease with different diabetes treatments. METHODS: Associations of treatment-achieved HbA1c were investigated in various treatment groups [diet only; insulin; sulphonylurea (SU) alone; SU with glinides; and antihyperglycaemic agents other than glinides, SU or insulin] taken from a nationwide claims database of 14,633 Japanese diabetes patients. Cox's regression analysis examined risks over a 5.1-year follow-up. RESULTS: A significant linear trend was associated with HbA1c levels and CAD events in the diet-only group, and CAD risks were significantly higher in insulin and SU groups with HbA1c ≤ 7.0% and > 8.0% than in the diet-only group with HbA1c ≤ 7.0%. In contrast to CAD, a linear association was observed regardless of treatment modality between achieved HbA1c levels and risk of severe diabetic eye disease, but with no significant difference in eye disease risk between groups with HbA1c ≤ 7.0% and 7.1-8.0% in those treated with either SU alone, SU with glinides, or insulin. CONCLUSION: These findings suggest that the relationship between treatment-achieved HbA1c and incidence of both CAD and severe diabetic eye disease differed according to treatment, based on a large-scale real-life database. More research is now needed to confirm these findings and to further investigate the underlying mechanisms.