Current issues of ACP Journal Club are published in Annals of Internal Medicine


Acarbose reduced glycosylated hemoglobin levels in IDDM

ACP J Club. 1997 Jul-Aug;127:13. doi:10.7326/ACPJC-1997-127-1-013

Source Citation

Hollander P, Pi-Sunyer X, Coniff RF. Acarbose in the treatment of type I diabetes. Diabetes Care. 1997 Mar; 20:248-53.



To determine the safety and efficacy of acarbose in combination with diet and insulin therapy in patients with insulin-dependent diabetes mellitus (IDDM).


36-week randomized, double-blind, placebo-controlled trial.


Multiple centers in the United States.


264 patients who were ≥ 18 years of age and had an established diagnosis of IDDM and stable body weight. Patients were excluded if they had a disease or condition that complicated the diabetic state or jeopardized their ability to adhere to the protocol, including gastrointestinal (GI) disease. Valid data were available for 236 patients (89%) (mean age 37 y, 65% men).


Patients followed a diabetic diet during a 6-week run-in period and were allocated to acarbose (n = 114) or placebo (n = 122). Acarbose was given at an initial dose of 50 mg 3 times/d (week 0) and titrated at 6-week intervals to 300 mg 3 times/d at week 18. Patients' insulin regimens were stabilized during the 6-week run-in period and adjusted after randomization based on hypoglycemic events at the investigator's discretion.

Main outcome measures

Mean absolute change in glycosylated hemoglobin (HbA1c) levels (normal range 3.6% to 4.9%), mean percentage change in daily insulin requirements, mean change in the number of hypoglycemic episodes, and adverse effects.

Main results

Patients who received acarbose had a mean decrease in HbA1c levels of 0.3%, and patients who received placebo had a mean increase of 0.18% {95% CI for the -0.48 difference in mean change -0.70 to -0.26}*. No differences existed between groups for mean change in insulin dose or regimen, number of hypoglycemic episodes, or body weight. Acarbose was associated with more adverse events, particularly diarrhea, flatulence, and abdominal pain. 84% of patients who received acarbose reported ≥ 1 adverse event compared with 49% of patients who received placebo ({CI for the 35% absolute risk increase 24% to 45%}*, P = 0.01). 19% of patients who received acarbose stopped treatment compared with 5% of patients who received placebo (P < 0.001).


The addition of acarbose to diet and insulin in patients with insulin-dependent diabetes mellitus reduced gly-cosylated hemoglobin and plasma glucose levels. It did not result in increased hypoglycemic events but was associated with gastrointestinal events.

Source of funding: Bayer Pharmaceutical Company.

For article reprint: Dr. P. Hollander, Ruth Collins Diabetes Center, Baylor University Medical Center, 3500 Gaston Avenue, Dallas, TX 75246, USA. FAX 214-820-4840.

*Numbers calculated from data in article.


The Diabetes Control and Complications Trial has shown that more normal, long-term glycemic control reduces the risks for microvasculopathy and neuropathy in patients with IDDM (1). This causal relation does not require a glycemic threshold (1), meaning that treatments that lower elevated HbA1c levels in patients with IDDM by even modest amounts (e.g., 0.5%) are worth a look. Acarbose may be such a treatment.

The study by Hollander and colleagues is currently the largest and longest placebo-controlled trial of acarbose for IDDM. The patients had suboptimal glycemic control (mean HbA1c level approximately 6.6%) and appeared typical of the patients seen in daily practice. The increased GI adverse effects and mild increases in transaminase levels seen in patients who received acarbose was expected (2) and emphasizes the need to avoid the drug in patients with GI disease, including diabetic gastropathy. The predictable GI effects also raise the question of whether the lower HbA1c level ascribed to acarbose was in fact secondary to co-interventions, including diet, exercise, and insulin therapy. However, the comparable weights and similar insulin programs in the acarbose and placebo groups make this unlikely. Although no differences were seen in hypoglycemia rates between the 2 groups, this remains a concern that may demand lowering insulin doses in some patients receiving acarbose. Further data on risk for hypoglycemia, and whether acarbose maintains improved glycemic control in patients with IDDM beyond 24 weeks, will be important.

Acarbose is not approved for use in IDDM in North America. However, the result of this trial and the efficacy of acarbose shown for NIDDM with its apparent lack of major irreversible adverse effects (2) make a closely monitored trial of acarbose (including measurement of liver enzyme levels and frequent measurement of glucose levels at home) an option I would now carefully consider for selected patients with IDDM.

Jeffery Mahon, MD
University HospitalLondon, Ontario, Canada

Jeffery Mahon, MD
University Hospital
London, Ontario, Canada


1. The absence of a glycemic threshold for the development of long-term complications: the perspective of the Diabetes Control and Complications Trial. Diabetes. 1996;45: 1289-98.

2. Hollander P. Safety profile of acarbose, an alpha-glucosidase inhibitor. Drugs. 1992; 44 (Suppl 3):47-53.