Clinical and public health implications of glycemic relapse in type 2 diabetes

Background Although intensive therapeutic regimens can improve glycemic control in type 2 diabetes, glycemic relapse might subsequently occur. Objectives To determine the incidence of relapse following treatment of type 2 diabetes and to characterize factors that predict deterioration of glycemic control. Design and Intervention This was an outpatient study of adults with type 2 diabetes who were referred to an intensive diabetes care program. The 12-week program provided instruction in diabetes self-management and adjustment of medication. Insulin therapy was started if glycemic goals were not met with lifestyle modifications and hypoglycemic drugs. Patients initially received a bedtime dose of 0.3 U/kg of intermediate-acting or long-acting insulin, which was adjusted until a fasting plasma glucose level of <= 6.9 mmol/l was attained. If postprandial hyperglycemia persisted, patients were also given a rapid-acting insulin analog before meals. Patients were encouraged to self-adjust their insulin dose in response to home glucose monitoring. Glycated hemoglobin (HbA(1c)) levels were measured at study entry, at 12 weeks, and every 6 months thereafter. The nadir HbA(1c) level was defined as the lowest HbA(1c) level achieved within 6 months of completing the program. Outcome Measures The primary outcome measure was glycemic relapse, defined as the interval between the nadir HbA(1c) level and the first episode of hyperglycemia ( HbA(1c)>= 7%). Results A total of 393 patients were included in the analysis. Baseline characteristics of the study group included a mean age of 54.8 years, 50% female participants, 25% black participants, a mean BMI of 35 kg/m(2), a median duration of diabetes for 1.1 years, and a mean HbA(1c) level of 9.3%. Overall, 126 patients were receiving insulin. The mean duration of diabetes was 5.3 years in patients receiving insulin and 0.3 years in those not receiving insulin (P<0.001). Additionally, patients receiving insulin had a mean weight gain of 1.1 kg during the program, whereas patients not receiving insulin had a mean weight loss of 2.2 kg (P<0.001). Treatment with insulin was associated with a high risk of relapse (hazard ratio [HR] 1.5). The probability of relapse in the group receiving insulin was 56% after 1 year and 86% after 3 years, compared with 38% and 70%, respectively, for the group not receiving insulin. The median time to relapse was 7.7 months and the incidence rate for relapse was 0.84 per person-year for patients receiving insulin, whereas for patients not receiving insulin the median time to relapse was 19.9 months and the incidence rate for relapse was 0.42 per person-year. Duration of diabetes >= 1 year also predicted relapse (HR 1.6); however, weight loss >= 4.5 kg was associated with a decreased risk of relapse among patients not receiving insulin (HR 0.6). Conclusion There was a high incidence of relapse in patients after intensive treatment of type 2 diabetes; duration of disease and insulin treatment both predicted loss of glycemic control.