The metabolic consequences of acute coronary syndromes include the release of epinephrine, glucagon and other counter-regulatory hormones. These antagonise the effects of insulin, and cause further worsening of insulin resistance. In the myocardium this insulin resistance favours the utilisation of free fatty acids, which may have a deleterious effect on myocardial function. In non-diabetic subjects, this may cause temporary hyperglycaemia, sometimes called 'stress hyperglycaemia', that requires further investigation at a later stage for possible impaired glucose tolerance or diabetes (Chapter 11). In patients with established diabetes this can cause significant increases in hyperglycaemia, and may be a cause of metabolic decompensation and hyperglycaemic states.
The DIGAMI (Diabetes Mellitus Insulin-Glucose Infusion in Acute Myocardial Infarction) study was based on the hypothesis that the administration of a high-dose intravenous insulin infusion followed by intensive subcutaneous insulin administration would overcome the worsening insulin resistance, suppress free fatty acid release, reduce hyperglycaemia, improve glucose utilisation in the myocardium and thus reduce mortality (Malmberg etal., 1994). A small pilot study was performed to establish the safety of a high-dose insulin infusion. The insulin was added to a bag of 5% dextrose and infused at 30ml/h, which is roughly 5 units of insulin per hour (Malmberg etal., 1994). The main study was perfomed in 19 Swedish Coronary Care Units (CCUs) and randomised 620 patients to either insulin infusion followed by multidose subcutaneous insulin injection (usually four times daily) (n = 306), or conventional treatment (n = 314) (Malmberg etal., 1995). A further 620 subjects who were potentially eligible for inclusion in the study were seen in the CCUs during the time the study was running, but were not included either because of inability or unwillingness to comply with the study treatment protocol. Subjects were included if their random blood glucose was above 11.1 within the first 24 h. Most were previously diagnosed as having diabetes, but some were previously unknown. There was an insignificant reduction in total mortality (the primary endpoint) at hospital discharge and at 3 months, which was the initial proposed duration of follow-up. When the study was extended to 1 year of follow-up a significant absolute reduction in total mortality of 7% was observed in the intervention group compared to the conventional treatment group (Malmberg etal., 1995, 1996), and this was increased to an 11% absolute reduction at a mean follow-up of 3.4 years (Malmberg, 1997) (Figure 4.3). DIGAMI also confirmed that admission blood glucose and HbA1c were predictors of short-term and long-term mortality following MI (Malmberg etal., 1997, 1999).
Although the study was significantly positive, and the longer term follow-up results receved wide exposure in the medical literature, it did not lead to major changes in clinical practice for several reasons. Firstly, it was the only large study to show benefit of insulin treatment in diabetic patients, although some studies on the use of glucose-insulin-potassium therapy in non-diabetic subjects following MI had demonstrated similar benefits. Secondly, it was a complex intervention combining the use of intravenous followed by intensive subcutaneous insulin. The study intravenous insulin regimen was rather idiosyncratic and was quite difficult for nursing staff to administer in a busy CCU setting. Because the insulin was administered in a bag of dextrose, and not via a syringe pump, a lot of fluid was infused, raising worries about possible fluid overload. In routine clinical practice the education required to teach this frail group of patients to self-administer insulin was time consuming and required the involvement of the diabetes team, and as in the DIGAMI study many patients were unwilling or unable to start insulin injections following an MI. Finally, the mechanisms of possible benefit were unstudied in the original DIGAMI study.
A second DIGAMI study was established to address some of these difficulties (Malmberg etal., 2005). DIGAMI-2 contained three groups: one that received
Years in study
No. of patients at risk Control 314 232
Infusion 306 248
Years in study
No. of patients at risk Control 314 232
Infusion 306 248
Figure 4.3 Actuarial mortality curves during long-term follow-up in patients receiving insulin-glucose infusion and in control group among the total DIGAMI cohort. Absolute reduction in risk was 11%; relative risk 0.72 (0.55-0.92, P = 0.011). Reproduced from Malmberg K for the DIGAMI (Diabetes Mellitus, Insulin Glucose Infusion in Acute Myocardial Infarction) Study Group (1997). Prospective randomised study of intensive insulin treatment on long term survival after acute myocardial infarction in patients with diabetes mellitus. British Medical Journal 314: 1512-15. Figure 1, p. 1514.
intravenous followed by subcutaneous insulin as in DIGAMI-1 (although most got twice-daily insulin injection); a group that got conventional treatment all the way through, as in DIGAMI-1; and a group that received intravenous insulin followed by conventional treatment. In this way it was hoped to identify if subcutaneous insulin therapy was required for benefit or whether short-term intravenous insulin alone was sufficient. It was hoped to randomise 3000 subjects but the study was halted because of slow recruitment, and less than half this number of subjects was finally included. In an attempt to improve recruitment the protocol was amended slightly so that patients with known diabetes could enter with any blood glucose concentration, whereas in DIGAMI-1 it had to be above 11.1mol/l. Approximately half of the subjects had STEMIs and half had NSTEMIs.
The results were disappointingly negative, and if anything the total mortality at 2 years, which was the primary endpoint of the study, was lowest in the control group. Several problems can be identified with the study. As mentioned above the number of subjects recruited was low, there were baseline differences that favoured the control group, and an unexpectedly high number of non-cardiac deaths was observed in the intensive intervention group. The loosening of entry criteria to allow patients with known diabetes who had lower admission blood glucose readings meant that lower risk patients were included, as mortality is proportionate to admission blood glucose, and it also meant that there was a much smaller metabolic window for improvement with the intravenous infusion. On follow-up, blood glucose targets were not achieved, probably because of the use of a less intensive insulin regimen, and so there was no major difference in metabolic control, as measured by HbA1c, between the groups.
A recent study from six centres examined the effects of improved glycaemic control with an insulin and dextrose infusion for at least 24 h versus conventional therapy, including subcutaneous insulin, in 240 subjects with MI and a blood glucose concentration of > 7.8mmol/l. Half of the subjects had previously diagnosed diabetes. There was no significant effect on the primary endpoint of mortality at 3 and 6 months, but there was a significant reduction in cardiac failure and reinfarction within three months (Cheung etal., 2006).
Thus, at present the evidence for reductions in mortality with intravenous insulin following acute coronary syndromes in diabetic and non-diabetic subjects is not stong. Diabetic patients should be treated with a high-dose insulin infusion to try and minimise hyperglycaemia while avoiding hypoglycaemia, and thereafter the blood glucse concentration should be controlled as tightly as possible, again avoiding hypoglycaemia, but this does not have to be insulin based (see also Chapters 9 and 11). There is a need for large, long-term studies evaluating the effects of very intensive insulin therapy in patients with diabetes and MI, preferably with a treatment goal of normalising blood glucose concentrations.
Was this article helpful?