Eight RCTs (n=22,215 participants, range 118 to 20,201) of GIK based on insulin focus and three RCTs (n=1,649, range 240 to 1253) of insulin-glucose based on glycaemia focus were included in the review. There was a discrepancy in the numbers of participants reported in the text and a table. Most studies had a Jadad score of 3 and two studies had a Jadad score of 2.
Trials using GIK (eight RCTs): There was no evidence of a reduction in mortality with use of GIK compared to control (eight trials) with considerable heterogeneity between trials (I2=48%). In subgroup analysis, GIK before reperfusion was not associated with a reduction in mortality compared to control (five trials, I2=24%). GIK after reperfusion was also not associated with a reduction in mortality compared to control (three trials, I2=48.4%). Sensitivity analysis that excluded one trial that appeared to be an outlier, removed the statistical heterogeneity and did not markedly change the results.
Post hoc sensitivity analysis was undertaken to remove two trial subgroups that had greater weight in the analyses. The summary effect measure for the subgroup analysis of GIK before reperfusion was not changed markedly. However, GIK was associated with a significant increase in mortality compared to control when given after reperfusion (RR 1.78, 95% CI 1.13 to 2.80; two trials).
Trials using insulin-glucose (three RCTs): There was no evidence of a reduction in mortality with use of insulin-glucose compared to standard therapy in treatment of acute myocardial infarction with diabetes (three trials, no significant heterogeneity). Subgroup analysis according to whether glycaemic control was achieved found no evidence of a significant difference in mortality between the two groups.
There was no evidence of publication bias by inspection of a funnel plot and use of Begg correlation.