RESUMO
AIM: To evaluate the safety of four insulin titration algorithms in a homogeneous population of insulin-naïve type 2 diabetic patients. METHODS: We conducted a 24-wk, open, single-center study with 92 insulin-naïve type 2 diabetes patients who failed treatment with one or two oral drugs. The patients were randomized to one of the four following algorithms: LANMET (n = 26) and LANMET PLUS (n = 22) algorithms, whose patients received a fixed initial insulin dose of 10 U, and DeGold (n = 23) and DeGold PLUS (n = 21) algorithms, whose patients' initial insulin dose was based on their body mass index (BMI). In addition, patients in the PLUS groups had their insulin titrated twice a week from 2 to 8 U. In the other two groups, the titration was also performed also twice a week, but in a fixed increments of 2 U. The target fasting glucose levels for both groups was 100 mg/dL. RESULTS: There was no significant difference in efficacy parameters. There was no significant difference when comparing moderate hypoglycemia events in algorithms starting with a 10 U fixed dose and algorithms based on BMI. However, there was a significant increase in moderate hypoglycemia events among the PLUS treated patients when the LANMET and DeGold algorithms were compared with the 2 fast-titration PLUS algorithms. We observed 12 hypoglycemia events in the first group, which corresponded to 0.94 events/patient per year, and we observed 42 events in the second group, which corresponded to 2.81 events/patient per year (P < 0.037). No further significant differences were observed when other comparisons between the algorithms were carried out. CONCLUSION: Starting insulin glargine based on BMI is safe, but fast titration algorithms increase the risk of moderate hypoglycemia.