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


Acarbose was effective in maintaining glycemic control in type 2 diabetes mellitus but was associated with adverse effects

ACP J Club. 1999 Nov-Dec;131:74. doi:10.7326/ACPJC-1999-131-3-074

Source Citation

Holman RR, Cull CA, Turner RC, on behalf of the UKPDS Study Group. A randomized double-blind trial of acarbose in type 2 diabetes shows improved glycemic control over 3 years (U.K. Prospective Diabetes Study 44). Diabetes Care. 1999 Jun;22:960-4. [PubMed ID: 99300837]



What is the effectiveness of acarbose for maintaining glycemic control in patients receiving established therapy for type 2 diabetes mellitus?


Randomized (concealed), double-blind, placebo-controlled trial with follow-up of 3 years.


23 hospital-based diabetes clinics in the United Kingdom.


1946 patients (mean age 60 y, 63% men, mean body mass index 30 kg/m2, median hemoglobin [Hb] A1c level 7.9%, median fasting plasma glucose [FPG] level 8.7 mmol/L) who had type 2 diabetes, were already receiving therapy, and did not have gastrointestinal problems or life-threatening illness. Follow-up was 83%.


Patients were allocated to acarbose (n = 973), titrated to a maximum dose of 100 mg 3 times/d, or placebo (n = 973). Patients continued their preexisting therapy.

Main outcome measures

HbA1c levels, FPG levels, body weight, compliance with study medication, adverse effects, β-cell function, insulin sensitivity, and major clinical events were measured every 4 months.

Main results

Analysis was by intention to treat. At 3 years, patients in the acarbose group had a lower median HbA1c level {8.3% vs 8.6%, P = 0.003}*, a lower rate of compliance with the study medication (39% vs 58%, P < 0.001), and higher rates of flatulence (30% vs 12%, P < 0.001) and diarrhea (16% vs 8%, P < 0.001) than did those in the placebo group. No difference existed between groups for FPG level, body weight, albuminuria, β-cell function, insulin sensitivity, and rate of major clinical events. When patients who were receiving the study medication at 3 years were analyzed alone, those in the acarbose group had a 0.5% lower median HbA1c level (8.1% vs 8.6%, P < 0.001) than did those in the placebo group.


Acarbose more effectively maintained glycemic control in patients with established type 2 diabetes mellitus on preexisting therapy but had higher rates of noncompliance and adverse effects than did placebo.

Source of funding: Bayer U.K.

For correspondence: Professor R.R. Holman, Diabetes Trials Unit, University of Oxford, Radcliffe Infirmary, Woodstock Road, Oxford OX2 6HE, England, UK. FAX 44-1865-723826.

*Data provided by author.


Optimal glycemic control in type 2 diabetes may require multiple forms of therapy, and an increasing number of hypoglycemic agents are now available. The study by Holman and colleagues describes the glucose-lowering properties of acarbose, an α-glucosidase inhibitor that is used as an adjunct to preexisting therapy.

This study confirms earlier reports of the efficacy of acarbose (1). Based on its mechanism of action, which is the delay of carbohydrate absorption, acarbose should predominantly affect postprandial hyperglycemia. This is supported by the current data showing that acarbose had no effect on FPG level but was associated with lowered HbA1c levels at 3 years. Although increased gut carbohydrate absorption has not traditionally been seen as a contributor to postprandial hyperglycemia, recent data suggest that splanchnic glucose uptake may be altered (2). Delayed entry of glucose from the gut by α-glucosidase inhibition may alleviate these altered splanchnic responses.

The major problem is that acarbose tends to cause flatulence and diarrhea. This side effect resulted in a > 50% noncompliance rate at 12 months in the current study, a rate that is higher than the 27% dropout rate in an earlier 12-month study (1). Inhibition of intestinal brush-border enzymes is the mechanism of both the drug's effect and its major side effect. Attempts to alleviate the latter (with use of the enzyme α-galactosidase) have been associated with reduction of the glucose-lowering effect (3). Slow titration of the drug is important to optimize compliance. Among patients who can tolerate acarbose, its lack of associated weight gain is a desirable feature. The future challenge is to establish the optimal choice of oral agents for individual patients.

Sean F. Dinneen, MD
Mayo Clinic and FoundationRochester, Minnesota, USA


1. Chiasson JL, Josse RG, Hunt JA, et al. The efficacy of acarbose in the treatment of patients with non-insulin-dependent diabetes mellitus. A multicenter controlled clinical trial. Ann Intern Med. 1994;121:928-35.

2. Ludvik B, Nolan JJ, Roberts A, et al. Evidence for decreased splanchnic glucose uptake after oral glucose administration in non-insulin-dependent diabetes mellitus. J Clin Invest. 1997;100:2354-61.

3. Lettieri JT, Dain B. Effects of Beano on the tolerability and pharmacodynamics of acarbose. Clin Ther. 1998;20:497-504.