Patients with type 2 diabetes (history of treatment discontinuation, n=78; no history of treatment discontinuation, n=516) were prospectively observed for 5 years, and the incidence of treatment discontinuation was investigated. Age and a history of treatment discontinuation were associated with treatment discontinuation in the overall population (hazard ratio: 2.17). HbA1c (hazard ratio: 2.12) was significantly associated with treatment discontinuation in patients with a history of discontinuation, while age and BMI were significantly associated with treatment discontinuation in patients without a history of treatment discontinuation. The treatment discontinuation rate after 5 years was 23 % in patients with a history of treatment discontinuation, which was significantly higher in comparison to that in patients without a history of treatment discontinuation (11 %). At the start of the observation period, the HbA1c value was higher in patients with a history of treatment discontinuation; however, after the first year it remained similar to that in the group without a history of treatment discontinuation. The prevalence of retinopathy and nephropathy remained higher in the group with a history of treatment discontinuation. A history of treatment discontinuation is a risk factor for re-discontinuation, and even if treatment is continued, vascular complications cannot be sufficiently reduced. Therefore, it is important to prevent discontinuation from the start of treatment.