DIGAMI-2: Chronic insulin therapy no better than usual care in reducing mortality in diabetics post-AMI

Shelley Wood

September 15, 2004

Munich, Germany - The Diabetes Mellitus Insulin-Glucose Infusion in Acute Myocardial Infarction 2 (DIGAMI-2) trial has failed to show that long-term intensive insulin treatment improves survival in diabetic patients, compared with usual care, but investigators say the results nonetheless highlight the importance of glucose lowering.

"I think the main message is that glucose still matters and we need new tools to treat these postinfarction diabetic patients," coprincipal investigator Dr Klas Malmberg (Karolinska University, Stockholm, Sweden) told heartwire .

The DIGAMI-2 results were presented by Malmberg and Dr Lars Rydén (Karolinska University) at the European Association for the Study of Diabetes 2004 meeting earlier this month.

DIGAMI-2 designed to clarify earlier results

The DIGAMI-2 study set out to extend the findings of the original DIGAMI trial. Published in 1999, DIGAMI compared usual-care treatment of diabetics post-AMI with intensive insulin therapy involving an acute insulin-glucose infusion during the first 24 hours post-AMI, followed by thrice-daily insulin injection for at least three months[1]. While the intensive-treatment group had a lower mortality than the usual-care group at one year, it was unclear whether the benefit was due to the initial insulin-glucose infusion or the chronic insulin therapy.


The main message is that glucose still matters.


In DIGAMI-2, patients were thus randomized to one of three groups:

  • Acute insulin-glucose infusion followed by insulin-based long-term glucose control.

  • Insulin-glucose infusion followed by standard glucose control.

  • Routine metabolic management according to physician discretion.

The trial was initially designed to include 3000 patients but was ultimately capped at 1253 after poor enrollment, due in part, says Malmberg, to several "troublesome" aspects of the trial.

"This is actually the largest randomized study so far in patients with diabetes and AMI, but we had hoped to get double the number of patients," he told heartwire . Despite keen investigators, the "academic" trial couldn't compete against "more profitable studies" with more incentives to enroll patients, he noted. As well, the treatment being studied in DIGAMI-2, requiring insulin injections several times daily in group 1, "is more difficult than just giving pills," Malmberg added.

Commenting on the DIGAMI-2 results for heartwire , Dr Darren McGuire (UT Southwestern, Dallas) observed that the results had been highly anticipated by the cardiology community. "Although enrollment fell well short of planned, the investigators should be congratulated for amassing one of the largest diabetes and MI studies reported to date."

No significant differences

After a mean of two years, the authors found no significant differences in mortality between groups 1 and 2 (the primary end point of the study) or between groups 2 and 3 (the secondary endpoint). As well, there were no differences in nonfatal reinfarction and strokes between the three groups.

Mortality differences in DIGAMI-2


Group 1

Group 2

Group 3

Mortality (%)

23.4 22.6* 19.3**

*p=0.831, group 1 vs 2
**p=0.203, group 2 vs 3

Of note, patients in group 1 did not achieve the glucose-lowering targets established in the trial design. Also complicating interpretation of the findings, 41% of patients in group 3 received extra insulin in the hospital and 14% received insulin infusions. Over the long term, multidose insulin for glucose lowering was given to 15% of patients in group 2, 13% of patients in group 3, and only 42% of patients in group 1.


It doesn't really matter how you lower glucose, and that's of course good for clinicians to know.


"Despite the fact that many of the patients in group 1 had insulin and many fewer patients in group 3 had insulin, physicians were able to achieve and maintain almost the same glucose control and there were no differences in mortality," Malmberg commented. "That means that it doesn't really matter how you lower glucose, and that's of course good for clinicians to know: they don't need to use this rather troublesome multidose insulin injection. If they could manage the glucose control by different measures it actually has the same effects on outcome, and this we believe is a very important result."

Disappointing enrollment may or may not have been important

Malmberg does not believe that better enrollment would have changed the outcomes of DIGAMI-2. "I don't think it would have made a difference. I think the key thing is that we didn't manage to lower the glucose as much as we wanted in group 1, and that means that there is really an unmet need for better tools for lowering glucose in these types of patients."

According to McGuire, the trial's low enrollment, paired with a lower-than-expected mortality rate, rendered the trail "dramatically underpowered, and therefore the negative results must be interpreted with caution." However, he continued, "the concerning trend toward detrimental effects associated with the insulin-infusion arms remains unexplained."

McGuire notes that some centers have been routinely employing the DIGAMI infusion strategy for diabetic patients with MI since the publication of the earlier results. "The present negative results from a study more than twice as big as DIGAMI and with almost twice as many evaluable deaths should serve to extinguish much of that enthusiasm, at least until more information is available."

One of the most surprising aspects of the DIGAMI-2 results was the finding of no apparent benefit of the acute insulin-glucose infusion, a strategy employed not so much for tight glycemic control as for hypothetical metabolic benefits for the heart, as is being investigated in the glucose-insulin-potassium (GIK) trials.

"The overall negative results of DIGAMI 2 are all the more surprising given the accumulated data suggesting benefit of GIK, even in the present era of MI care," McGuire noted. "Hopefully, some insight will be gained from the diabetes subset of the CREATE/ECLA GIK 2 study results, a 20000 patient, international randomized study evaluating GIK therapy in acute MI." Results from CREATE/ECLA GIK 2 will be presented at the upcoming 2004 American Heart Association Scientific Sessions.

Appropriate drugs must be used

One final point worth emphasizing from DIGAMI-2, said Malmberg, is that drugs with known cardiovascular benefit are effective in diabetic patients. At hospital discharge in DIGAMI-2, beta blockers were given to more than 80% of all patients, aspirin to almost 90%, ACE inhibitors to about 65%, and statin treatment to 65%.

"I want to emphasize that concomitant treatments, which have often been withheld from people with diabetes, were very helpful in this trial, and they lower mortality in these patients," Malmberg commented. "The overall mortality over two years was 20% in DIGAMI-2, and that's very high for an acute-MI trial. But if you look back a couple years, it was 30%, so it has improved, but there is still much to do."


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.