Background: The present study was conducted in order to determine the impact of different regimens of anti-hyperglycemic agents on all-cause mortality.
Materials and Methods: The Thailand diabetes registry cohort was a prospective 3-year follow-up of the Thai Diabetes
Registry project, registered 9,419 diabetes from 10 diabetic clinics in tertiary medical centers in Bangkok and the major
provinces between April, 2003 and February, 2006. The authors included type 2 diabetic patients over 30 year-old. Patients whose serum creatinine was >3 mg/dl were excluded to avoid a bias of choosing insulin and metformin. Since few patients took glinide or alpha-glucosidase inhibitors, our 8,109 (86.09%) patients were classified by their respective regimen of antihyperglycemic agents: i.e., whether sulfonylurea monotherapy (1,151; 14.2%); metformin monotherapy (794; 9.8%); sulfonylurea combined with metformin (3,665; 45.2%); thiazolidinedione [TZD] monotherapy or combined with antihyperglycemic agents (393; 4.9%); insulin treatment combined with metformin or TZD (1,273; 15.7%); or, insulin treatment without metformin or TZD (833; 12.6%).
Results: The mortality rate for those who received insulin treatment without metformin or TZD combination (49.3 per 1,000 patient-years) were the highest among anti-hyperglycemic treatment regimens followed by sulfonylurea monotherapy (22.7 per 1,000 patient-years), insulin treatment combined with metformin or TZD (13.6 per 1,000 patient-years), sulfonylurea combined with metformin (12.2 per 1,000 patient-years), TZD (8.1 per 1,000 patient-years) and metformin monotherapy (7.0 per 1,000 patient-years). Compared with patients who received sulfonylurea monotherapy, those who received metformin monotherapy, TZD, sulfonylurea combined with metformin had a respective lower mortality rate of 53% (adjusted hazard ratio [HR] 0.47 (0.25 to 0.88)), 56% (adjusted HR 0.44 (0.19 to 1.00)), and 31% (adjusted HR 0.69 (0.50 to 0.95)). The mortality of patients who received insulin treatment without metformin or TZD combination was increased by 53% (adjusted HR 1.58 (1.10 to 2.28)), while the harmful effects of insulin was reduced when combined with metformin or TZD (adjusted HR 0.83 (0.53 to 1.29)).

Keywords: Diabetes, Anti-hyperglycemic agents, Metformin, Sulfonylurea, Insulin, Thiazolidinedione, Mortality rate, Death
rate