"Acarbose considerably reduces the risk of myocardial infarction", headlines the Ärzte Zeitung (1), referring to a metaanalysis (2) of randomised studies that include a total of 2,180 diabetes patients, which appeared at the beginning of the year. According to this analysis, the relative risk of cardiovascular complications when taking acarbose (GLUCOBAY) is said to be reduced by 35% and the relative risk of myocardial infarction even by 64%. Adverse events (2) reported in the original studies were evaluated.

The information cannot be deduced from looking at the study results published. Of the seven studies assessed in the metaanalysis only three have been published in full (3-5). In two of these publications no serious cardiovascular events were reported either for acarbose or for placebo (3,5). The third study describes one serious cardiovascular event - heart failure - with acarbose, but none with placebo (4). The largest and most comprehensive published double blind study with acarbose in type 2 diabetes, the UKPDS 44* (a-t 1998, no. 10: 88-90), was not included at all in the metaanalysis. 1,946 patients took part in the UKPDS 44 study, almost as many patients as were assessed in the present metaanalysis. Over three years, no significant advantage is shown for acarbose in the UKPDS, for any pre-defined endpoint. The relative risk of any outcome associated with diabetes when taking acarbose, including cardiovascular events such as myocardial infarction or stroke, is 1.00, compared with placebo (95% confidence interval 0.81-1.23) (6).

The metaanalysis is thus based essentially on the manufacturer's data which are obviously not accessible and can therefore not be verified. By excluding the UKPDS 44 the authors are furthermore performing an arbitrary and incomprehensible selection of data. The study is therefore scientifically worthless and devoid of any relevance.

As is the case for people with impaired glucose-tolerance (STOP-NIDDM*; a-t 2003; 34: 73-4), assertions about a cardio-protective effect of acarbose (GLUCOBAY) for patients with diabetes mellitus have not been verified from valid study data (7).

(R = randomised study, M = metaanalysis)

1

Ärzte Ztg. of 01 June 2004

2

HANEFELD, M. et al.: Eur. Heart J.JJ. 2004; 25: 10-6

3

CHIASSON, J.-L. et al.: Ann. Intern. Med. 1994; 121: 928-35

4

HASCHE, H. et al.: Diab. Nutr. Metab. 1999; 12: 277-85

5

JOSSE, R.G. et al.: Diabetes Res. Clin. Pract. 2003; 59: 37-42

6

HOLMAN, R.R. et al.: Diabetes Care 1999; 22: 960-4

7

KAISER, T., SAWICKI, P.T.: Dtsch. Med.med. Wochenschr. 2004; 129: 1156

STOP-NIDDM = Study To Prevent Non-insulin dependent Diabetes Mellitus; UKPDS = United Kingdom Prospective Diabetes Study (44 = acarbose group)