May 14th, 2014

Informed Interpretation of the 20-Year Results of the DIGAMI Trial

During the years 1990 through 1993 the Swedish DIGAMI (Diabetes Mellitus Insulin Glucose Infusion in Acute Myocardial Infaction 1) trial randomized 620 MI patients with elevated glucose levels to either acute high-dose insulin infusion followed by daily subcutaneous multi-dose insulin treatment or conventional therapy. Earlier results from the trial showed beneficial effects, including improved survival, for patients in the insulin treatment arm.

Now, a paper published in The Lancet Diabetes and Endocrinology, presents 20-year follow-up results showing an average 2.3 year increase in survival for patients in the insulin-treatment arm (median survival 7.0 years versus 4.7 years, HR 0.83, CI 0.70-0.98, p=0.27).

The finding, writes Denise Bonds in an accompanying editorial, is consistent with the evolution of the field in recent decades. CardioExchange asked Darren McGuire, Professor of Internal Medicine, Director of the Cardiology Clinical Trials unit, and Director of the Parkland Hospital and Health System Outpatient Cardiology clinics at University of Texas Southwestern Medical Center in Dallas, to comment on the study. — Larry Husten

McGuire: There are a few key caveats that are always important to remember when interpreting results of the DIGAMI trial:

1. Most folks misinterpret this trial as a glucose control trial, as does Dr. Bonds in the accompanying editorial. This is a trial that compared two management strategies: insulin vs. no insulin; it did not compare two levels of glucose control. Beyond the first 48 hours during the hyperglycemic/hyperinsulinemic “GIK-like” infusion protocol, contrasts of glucometrics were not statistically different at any subsequent trial timepoint. So, this trial sheds no light on whether glucose control is effective; rather, it shows that a strategy using insulin is better than a strategy not using insulin.

2. The acute infusion was not targeted to glucose control — it was an infusion of dextrose designed to support delivery of high dose insulin (starting 5 units/hr) targeted to hyperglycemic targets (glucose of 126-180 mg/dL by protocol).

3. Most importantly, one can interpret the trial as showing superiority in the insulin-treated patients that was driven by adversity in the “control group.” In the DIGAMI era, sulfonylurea medications were about the only alternative (along with metformin), and concern remains about the cardiovascular safety of sulfonylureas.

My personal interpretation of DIGAMI is that randomization to insulin protected participants from being treated with sulfonylureas, thereby improving outcomes. This is buttressed by two specific considerations: First, the “treatment benefit” in DIGAMI was most evident in the patients entering the trial not treated with insulin (i.e., the “least sick” of any given diabetes cohort); this is counter to the ever-prevailing concept that the sickest patients benefit most from effective therapies. In this case, those not entering on insulin were most likely during trial treatment to receive sulfonylureas if randomized to the “control” arm. Second, insulin glargine has been shown to be almost identical to placebo for cardiovascular efficacy in the ORIGIN trial. Though not an acute post-ACS trial, ORIGIN had a large representation of patients with prior MI. This suggests the contrast in DIGAMI is not because insulin was better, but that “control” was worse.


Comments are closed.