Background and objective: Multiple oral therapies are required long term for the majority of patients with type 2 diabetes mellitus to achieve acceptable glycaemic levels; alternatively, insulin therapy has to be initiated. This study investigated the addition of acarbose to maximum doses of sulfonylurea in very poorly controlled type 2 diabetes patients and assessed its effect in delaying further glycaemic deterioration.
Study design: In this 78-week, double-blind, placebo-controlled European study, patients were randomised to receive acarbose, titrated to a maximum dose of 100mg three times daily, or matching placebo. Concomitant sulfonylurea treatment (glibenclamide/gliclazide) was to remain unchanged throughout the study. A sample size of 171 patients per treatment arm was calculated. The primary efficacy analysis was intention to treat.
Methods: The change in glycosylated haemoglobin (HbA(1c)) levels from baseline to the end of the study was regarded as the primary efficacy variable. Patients whose HbA(1c) levels increased above 10.5% on two consecutive visits terminated the study prematurely because of insulin administration. Secondary efficacy variables included the changes in blood glucose and C-peptide, both at fasting and at the 1h-postprandial level.
Patients: A total of 330 patients (acarbose 164, placebo 166) were valid for the efficacy analysis. Patients were generally overweight (body mass index 29.0 kg/m(2)) and showed very poor metabolic control (HbA(1c) >9%, fasting blood glucose >200 mg/dL, and 1h-postprandial blood glucose >300 mg/dL).
Results: Acarbose significantly improved HbA(1c) levels compared with placebo (least square mean [LS-mean] difference -0.54%, 95% CI -0.86 to -0.22; p = 0.001). A number of patients had to discontinue the study prematurely because of insulin administration (24.5% in the placebo and 14.2% in the acarbose group). There was a significant LS-mean difference of -14.8 mg/dL (p = 0.0195) in fasting blood glucose levels and highly significant differences in 1h-postprandial blood glucose (LS-mean difference -33.4 mg/dL, p < 0.0001) and in the rise in blood glucose from fasting to 1h-postprandial (LS-mean difference -19.6 mg/dL, p = 0.0001), all in favour of acarbose. Acarbose was shown to have a good safety profile and was generally well tolerated.
Conclusion: Acarbose was shown to have the potential to delay further deterioration of glucose control in type 2 diabetes patients who are very poorly controlled with maximum sulfonylurea doses.