Metabolic Modulation of Acute Myocardial Infarction
The ECLA Glucose-Insulin-Potassium Pilot Trial
Background—Several trials have been performed in the past using glucose, insulin, and potassium infusion (GIK) for the treatment of acute myocardial infarction (AMI). Because of continuing uncertainty about the potential role of this therapeutic intervention, we conducted a randomized trial to evaluate the impact of a GIK solution during the first hours of AMI.
Methods and Results—Four hundred seven patients with suspected AMI admitted within 24 hours of symptoms onset were enrolled. In a ratio of 2:1, 268 patients were allocated to receive GIK (high- or low-dose) and 139 to receive control. Phlebitis and serum changes in the plasma concentration of glucose or potassium were observed more often with GIK. A trend toward a nonsignificant reduction in major and minor in-hospital events was observed in patients allocated to GIK. In 252 patients (61.9%) treated with reperfusion strategies, a statistically significant reduction in mortality (relative risk [RR] 0.34; 95% CI: 0.15 to 0.78; 2P=0.008) and a consistent trend toward fewer in-hospital events in the GIK group were observed.
Conclusions—Our results confirm that a metabolic modulation strategy in the first hours of an AMI is feasible, applicable worldwide, and has mild side effects. The statistically significant mortality reduction in patients who underwent a reperfusion strategy might have important implications for the management of AMI patients. It is now essential to perform a large-scale trial to reliably determine the magnitude of benefit.
The pathophysiological mechanism of acute myocardial infarction (AMI) is the rupture of an atherosclerotic plaque in an epicardial coronary artery, exposing subendothelial tissue to a subsequent thrombogenic response and leading to a complete obstruction of the vessel.1 2 Strategies directed toward achieving early and sustained reperfusion of the infarct-related artery have reduced mortality;3 4 5 6 7 thrombolytics, aspirin, and, in selected centers, primary angioplasty are used as routine treatment during the early hours of AMI. The concept of metabolic protection of the ischemic myocardium hinges on the fact that glucose can be metabolized anaerobically and can thereby provide glycolytic ATP in the cytosolic compartment.8 Because glucose uptake is promoted by insulin, infusions of glucose and insulin with potassium (GIK) were initially evaluated in AMI by Sodi Pallares in 1962.9
Because the potential role of this intervention is still uncertain, we conducted a pilot clinical trial to evaluate the feasibility of GIK administration in contemporary practice and to assess its effect on clinical end points to develop a rationale for performing a large-scale trial of GIK infusion during the acute phase of myocardial infarction.
Recruitment of Patients
To encourage recruitment, the trial was kept simple and eligibility was broad. Patients with a suspected AMI within 24 hours of symptom onset were eligible regardless of age or ECG findings. There were no restrictions on ancillary treatments. Overall coordination was performed by the ECLA (Estudios Cardiológicos Latinoamérica) Coordinating Center in Rosario, Argentina.
The absolute contraindications to GIK infusion were severe renal impairment or hyperkalemia. No other contraindications were specified by the protocol but instead were left to the discretion of the responsible physician. Patients were to be entered in the trial only once.
Eligible patients were randomized in each participating center by a central telephone randomization system located at the ECLA Coordinating Center, except for patients in Brazil where, because of language differences, randomization was done using sequentially numbered sealed envelopes prepared centrally.
Informed consent was obtained from individual patients according to the requirements of local ethics committees, which approved the study.
Active treatment (GIK infusion) and control (usual care) were allocated within each center in a ratio of 2:1.
Two GIK solutions concentrations were used. High-dose GIK (HD GIK) has been advocated to suppress FFA levels.10
HD GIK consisted of 25% glucose, 50 IU soluble insulin (either human or not human) per liter and 80 mmol KCl per liter at an infusion rate of 1.5 mL/(kg · h) over 24 hours. Low-dose GIK (LD GIK) was a mixture of 10% glucose, 20 IU soluble insulin (either human or not human) per liter and 40 mmol KCl per liter at an infusion rate of 1.0 mL · kg−1 · h−1 over 24 hours. GIK infusion was initiated immediately after randomization and interrupted only if it was thought by the attending physician to be clearly indicated. Physicians were free to use whatever additional therapy they considered necessary, including the method of reperfusion (thrombolytics or primary PTCA).
After discharge, a clinical record form was sent to the ECLA Coordinating Center. Serial plasma concentrations of glucose and potassium were collected and recorded at randomization, and at 6, 24, and 48 hours after randomization.
The protocol stated that the population would be analyzed based on the treatment groups to which patients were randomized (HD GIK, LD GIK, and control), both GIK groups versus control, and both GIK groups versus control stratifying the population into those receiving and those not receiving a reperfusion strategy. All analyses were done on an intention-to-treat basis. The statistical analyses for continuous variables were done using ANOVA, and the comparisons for categorical variables were expressed as the relative risk (RR) with corresponding 95% CI. Probability values are 2-sided throughout and 2P>0.05 was considered not significant.
In 29 hospitals in 6 Latin American countries, 407 patients with a diagnosis of suspected AMI were randomized over 14 months. They were randomized as follows: 135 (33.2%) patients to HD GIK, 133 (32.7%) to LD GIK, and 139 (34.2%) to the control group. GIK infusion (HD or LD) was not started in 2 patients (0.7%) and was completed in 238 (88.8%) of the patients allocated to GIK. Among the 252 (61.9%) patients treated with a reperfusion strategy, 95% received lytic therapy and 5% were treated with primary angioplasty.
The mean infused volume was 2422 mL per 24 hours for HD GIK and 1795 mL/24 h for LD GIK in patients who received the infusion in accordance with the protocol (24-hour infusion). In 384 (94.3%) patients, myocardial infarction was enzymatically confirmed.
Tables 1⇓ and 2⇓ show the baseline characteristics of the randomized population. A family history of coronary heart disease and smoking were more frequently observed in the GIK group. All other variables were well matched and did not differ significantly.
Because of a lack of significant differences in the rate of adverse effects and the impossibility of detecting any difference in benefit between the 2 dose regimens , owing to the small number of patients (Table 3⇓), the results are reported as a comparison of the combination of both GIK regimens versus control.
The hospital length of stay, an indirect measurement of morbidity, did not differ in the treatments groups (9.8 days GIK versus 10.5 days control). Most of the population received a peripheral line (83%) ; phlebitis was reported in 45 (16.8%) patients allocated to GIK, although only 5 were identified as severe (1.9%). In no case did phlebitis cause an excess rate of morbidity (eg, sepsis) or any other complication.
Signs and symptoms of fluid overload were not different among the treatment groups. As shown in Table 4⇓, the incidence of severe heart failure (Killip class>2) was not significantly different between the GIK and control groups, and a nonsignificant trend toward a reduction in the incidence of severe heart failure was observed in the GIK group.
Figure 1⇓ shows the mean plasma levels of glucose and potassium at prerandomization and 6, 24, and 48 hours after randomization. No significant differences were observed in the glucose levels between the GIK and control groups, and significantly higher potassium levels were observed 24 hours (GIK 4.25 mmol/L versus control 4.04 mmol/L, P=0.0001) and 48 hours (GIK 4.24 mmol/L versus control 4.08 mmol/L, P=0.0027) after randomization in the GIK group.
Table 4⇑ shows the in-hospital events stratified by treatment group. The overall mortality rate during the hospital phase was 8.35% (95% CI 5.66 to 11.04). A trend toward a non–statistically significant reduction of major and minor in-hospital events was observed in patients allocated to GIK (Figure 2a⇓). The reported incidence of electromechanical dissociation was higher in the control group (5.8% versus 1.5%, 2P=0.016).
The baseline characteristics of 252 (61.9%) patients treated with a reperfusion strategy are shown in Table 2⇑. A higher prevalence of men and smokers was observed in the GIK group; all other variables were well matched. As shown in Table 4⇑, a significant reduction in mortality (RR 0.34; 95% CI 0.15 to 0.78; 2P=0.008) and electromechanical dissociation (RR 0.18; 95% CI 0.036 to 0.909; 2P=0.02) as well as a nonsignificant trend toward fewer in-hospital events was observed in the GIK treatment group (Figure 2b⇑).
Using a more specific analysis of the data and focusing on hard events, we stratified the end points of death, severe heart failure, and ventricular fibrillation in reperfused and nonreperfused patients. A 47% nonsignificant reduction in mortality for any death was observed (2P=0.10); using the 99% CI, the lower limit for mortality reduction in the reperfused patients was still below the unit (odds ratio 0.27; 99% CI 0.08 to 0.96). Interestingly, the heterogeneity test was significant, (χ2=4.68, P=0.03), probably reinforcing the concept that GIK produces different effects in patients who do or do not undergo a reperfusion strategy. A 30% nonsignificant reduction of any severe heart failure was observed, which was more pronounced in reperfused patients; and a striking 56% borderline-significant (P=0.07) reduction in any ventricular fibrillation was observed both in reperfused and nonreperfused patients (Figure 3⇓).
A significant reduction in the composite end point of death, nonfatal severe heart failure (Killip class >2), and nonfatal ventricular fibrillation was observed for both the overall and reperfused patients (Table 4⇑, Figure 2⇑).
We observed a nonsignificant trend toward more frequent use of invasive procedures during the in-hospital phase (CABG/nonprimary PTCA) in the GIK group (RR 1.36; 95% CI 0.61 to 3.00), and this trend was more pronounced in the reperfusion group (RR 2.28; 95% CI 0.68 to 7.66).
Although we have not done post hoc analyses due to the risk of true or false-positive or -negative results, we found a strong relation between the time from symptom onset and the impact of the infusion. A significant reduction in mortality was observed in patients treated within 12 hours of symptom onset (RR 0.43; 95% CI 0.2 to 0.9; P=0.02) both for the overall population and for patients who underwent a reperfusion strategy.
Among patients in whom the infusion was completed, we observed a significant reduction in mortality in the overall GIK group (RR 0.44; 95% CI 0.21 to 0.90) and in those patients who underwent any reperfusion strategy (RR 0.21; 95% CI 0.08 to 0.58) (Figure 4⇓).
The 1-year follow-up Kaplan-Meier curves show a dilution of the treatment effect in both the overall and reperfused populations. A nonsignificant mortality reduction of 19% and 33% for the overall and reperfused patients, respectively, was observed. Although this attenuation of the treatment effect, when analyzing each of the treatment arms, a consistent statistically significant reduction in mortality was still present for the HD GIK group in reperfused patients (RR 0.37, 95% CI 0.14 to 1.00; log-rank test 0.046) (Figure 5⇓).
This pilot trial demonstrates that GIK infusion is feasible and could have worldwide applicability. It requires neither human nor technological resources that could limit its wide use, has few side effects (which are easily managed), and does not produce major morbidity or mortality in AMI patients (compared with treatments of proven benefit but life-threatening side effects).
Direction of the Effect
Important clinical events such as mortality, severe heart failure, ventricular fibrillation, and electromechanical dissociation (Figure 2a⇑ and 2b⇑), which could be influenced by the GIK infusion (infarct size limitation or protection of ischemic/viable myocardium and incidence of arrhythmias), demonstrated a consistent trend toward benefit. This effect was more impressive and reached statistical significance in patients who underwent reperfusion strategies. When examining only patients who were actually treated (not on an intention-to-treat basis), which is a useful approach for in-depth analysis in pilot studies, the reduction in events was in the same direction, more pronounced, and it reached statistical significance for mortality in the overall population as well as in patients who underwent a reperfusion strategy (Figure 4⇑).
The only end point that went in the opposite direction was the use of in-hospital invasive procedures (CABG/nonprimary PTCA). Although we cannot draw any conclusion about this observation, we can speculate that patients who received GIK, especially those with reperfusion therapy, have a larger amount of viable myocardium leading to a higher demand for revascularization procedures.
Regarding the impact of GIK in patients within 12 hours of symptom onset (a post hoc analysis), the explanation could be that the earlier the solution is infused, the more evident is the effect obtained. Although this hypothesis agrees with experimental data, confounding factors actually modify this interpretation, because most patients (72%) studied before 12 hours from the onset of symptoms received a reperfusion strategy.
Magnitude of the Effect
The impressive magnitude of the reduction in mortality in patients who underwent reperfusion (Figure 2b⇑) deserves cautious interpretation mainly because most treatments in AMI produce only moderate effects on major outcomes such as mortality (eg, 15% to 20%),11 and much larger numbers of patients are needed for clinical trials to yield a reliable estimate of the magnitude of the treatment effect.12
However, experimental data support the concept that GIK infusion in the context of ischemic/reperfused myocardium produces an impressive impact on the evolution of the injured but viable myocardium when added early during the ischemic period.13 Thus, as shown in our trial, an important impact of GIK on the reperfused myocardium cannot be excluded.
The surprising high mortality rates for control patients who underwent any reperfusion strategy is difficult to explain. The only plausible explanation is the small number of patients in that group, 79, and the high number of fatal events, 12, perhaps reflecting the play of chance. Although the mortality rate in this group was 15.2%, the 95% CI was 7.22 to 23.14. The same explanation could be applied to the low mortality rate of the reperfused patients in the GIK group, with a mortality of 5.2% and a 95% CI between 1.89 and 8.51; one could, therefore, expect a lower magnitude of effect in a larger trial. An alternative explanation of the high death rates of the control-reperfused patients could be a selection bias (sicker patients underwent reperfusion therapy), because reperfusion therapy was not randomized.
Reinforcing the concepts that GIK mainly induces beneficial effects in the subgroup of reperfused patients and that HD GIK appears to be superior to LD GIK, the 1-year follow-up in HD GIK –reperfused patients showed a statistically significant reduction in mortality (Figure 5⇑).
Our pilot study was similar in size to or larger than most trials conducted to test the efficacy of early primary angioplasty for the treatment of AMI before the GUSTO II substudy.14 15 16 17 18 19 The impressive magnitude of benefit of these early small studies led to the incorporation of such a strategy as routine care of AMI patients in some selected centers.
Our study is, however, the largest trial with GIK published in the thrombolytic era. Our findings of an impressive impact on clinical events and few, manageable side effects could be of great importance for the clinical management of AMI patients.
Proposed Mechanism of Action of GIK
Major metabolic changes occur during the early hours of AMI20 ; these include increased secretion of catecholamines,21 increased concentration of circulating free fatty acids (FFA),22 and glucose intolerance.23 Each of these abnormalities might adversely influence the outcome of AMI, either by provoking arrhythmias or by compromising the survival of ischemic tissue.24 25
Under normal conditions, myocardium depends on aerobic metabolism, and the preferred fuels for myocardial oxidative metabolism are FFA.26 The high concentration of FFA during the early stages of AMI, a consequence of high sympathetic activity,21 could be deleterious, increasing the incidence of potentially malignant arrhythmias22 and adversely influencing the outcome of ischemic but viable tissue.24 25
A detailed review of the mechanism of action of GIK has recently been written by Apstein and Taegtmeyer.13 A meta-analysis of all properly randomized trials27 of GIK in AMI that involves about 2000 patients and 350 deaths showed a consistent reduction in mortality of 28% (95% CI 14 to 83, P=0.004). Our findings concur with those results, although the suggested benefit was more pronounced in our pilot study, reaching statistical significance in patients who underwent a reperfusion strategy. The trials included in the meta-analysis were conducted in the prethrombolytic era, which could explain the differences in the magnitude of the effect. When we add our HD GIK data to the 4 trials in the meta-analysis that used HD GIK, those results become clearer, with an odds reduction of 43% (SD: 22), 2P=0.05.
Supporting the concept that long-term metabolic control in diabetic patients is of critical importance in preventing vascular complications, one clinical trial, the Digami trial,28 found 1-year mortality of diabetic patients with AMI to be reduced significantly with an in-hospital glucose-insulin infusion followed by multidose subcutaneous insulin.
Limitations of Our Study
The lack of a blinded design introduced a potential bias in end point assessment or ancillary care that may have affected patient outcomes. The large volume of fluid that had to be administered made blinding difficult, because the blinding procedure would probably have been detrimental to the control group.
We also were unable to explore the mechanistic action of the GIK infusion, mainly because we did not measure plasma FFA or insulin levels.
The results of our pilot clinical trial confirm that a metabolic modulating strategy is feasible in the first hours of an AMI with a GIK infusion. Such a strategy is also applicable worldwide and is associated with mild side effects that do not lead to an increase in morbidity or mortality. The impressive magnitude of the benefit in terms of mortality reduction and other in-hospital clinical events is in agreement with other data and could have a major impact on the clinical management of AMI patients.
Because of the provocative results of this clinical trial in contemporary practice and the consistency of the evidence between the meta-analysis and our study, we have designed and started a full-scale clinical trial to definitively confirm the convenience of incorporating this simple and inexpensive therapy for the routine care of AMI patients.
The following centers and investigators collaborated in the performance of this study (values in parentheses denote the number of patients enrolled).
ECLA Steering Committee: Ernesto Paolasso (Chairman), Rafael Díaz, Leopoldo S. Piegas, Carlos Tajer, Andres Orlandini.
ECLA Coordinating Center: Graciela Romero, María Inés Genisans, Jorgelina Cirasa, Andrea Pascual, Mirta Abraham, Daniel Wojdyla, Cristina Cuesta.
Argentina: Norberto Nordaby, Hospital Frances (38), Buenos Aires; David Ryba, Hospital Santojanni (1), Buenos Aires; Luis Guzman, Sanatorio Allende (7), Cordoba; Osvaldo Perrino, Hospital Italiano (25), La Plata; Victor Arregui, Instituto Medico Platense (1), La Plata; Gustavo Irusta, Hospital Lagomaggiore (20), Mendoza; Roberto Balado, Clinica Maria Auxiliadora (4), Olavarria; Fernando Colombo Berra, Sanatorio Quilmes (15), Quilmes; Miguel Cartia, Hospital De Emergencias (37), Rosario; Ariel Dogliotti, Instituto Cardiovascular de Rosario (11), Rosario; Alberto Navarro Zaval, Instituto de Cardiologia (2), Tucuman.
Brazil: Joao B. Lopes Loures, S.C. Mis. Juiz De Fora (12), Juiz De Fora; Luis M. Fragomeni, H.S. Vicente De Apulo (15), Passo Fundo; Rui Fernando Ramos, Inst. Dante Pazzanese (18), San Pablo; Luiz Gomes Da Silva, Sancor (15), Santos.
Chile: Ramon Corbalan, Hospital PTS Universidad Catolica (9), Santiago; Eduardo Chavez, Asistencia Publica (2), Santiago.
Mexico: Juan Carlos Ramirez, Hospital Miguel Hidalgo (12), Aguascalientes; Sergio Gonzalez, Hospital General De Durango (6), Durango; Isaac Medina Godinez, Hospital Civil De Guadalajara (40), Guadalajara; Jesus Zuñiga Sedan, Hospital Del Carmen (8), Guadalajara; Manuel Barrera Bustil, Clinica De Merida (1), Merida; Lauro Carrillo, Hospital Español (9), Mexico DF; Jesus Martinez Sanche, Hospital Ingles Abc (8), Mexico DF; Jorge Bahema Cuevas, Hospital Universitario (2), Monterrey.
Uruguay: Gabriel Vanerio, Hospital Britanico (6), Montevideo.
Venezuela: Jesus Isea, Félix Sanchez, Hospital Domingo Luciani (70), Caracas; Leonardo Madrid, Centro Clínico Orinoco (13), Ciudad Bolivar.
We are indebted to patients who took part in this study, the investigators and nurses who collaborated wholeheartedly with the national coordinators in each country and Prof Robert Califf (Duke University) for useful comments on the early draft of the manuscript and Penny Hodgson for her helpful assistance in the correction of the manuscript.
↵1 Investigators who collaborated in this trial are listed in the Appendix.
- Received March 3, 1998.
- Revision received May 28, 1998.
- Accepted July 24, 1998.
- Copyright © 1998 by American Heart Association
Califf RM. The rationale for thrombolytic therapy. Eur Heart J. 1996;17(suppl E):2–8.
Gruppo Italiano per lo Studio della Streptochinasi nell’Infarto Miocardico (GISSI). Effectiveness of intravenous thrombolytic treatment in acute myocardial infarction. Lancet. 1986;i:397–402.
ISIS-2 Collaborative Group. Randomised trial of intravenous streptokinase, oral aspirin, both, or neither among 17 187 cases of suspected acute myocardial infarction: ISIS-2. Lancet. 1988;ii:349–360.
Fibrinolytic Therapy Trialists’ Collaborative Group. Indications for fibrinolytic therapy in suspected acute myocardial infarction: collaborative overview of early mortality and major morbidity results from all randomised trials of more than 1000 patients. Lancet. 1994;343:311–322.
Stanley AW, Moraski RE, Russell RO, Rogers WJ, Mantle JA, Kriesberg RA, McDaniel HG, Rackley CE. Effects of glucose-insulin-potassium on myocardial substrate availability and utilisation in stable coronary artery disease: studies on myocardial carbohydrates, lipid and oxygen arterial-coronary sinus differences in patients with coronary artery disease. Am J Cardiol. 1975;36:929–937.
Yusuf S, Flather M. Magnesium in acute myocardial infarction. BMJ. 1995;310:751–752.
Apstein CS, Taegtmeyer H. Glucose-insulin-potassium in acute myocardial infarction: the time has come for a large, prospective trial. Circulation. 1997;96:1074–1077.
Grines CL, Browne KF, Marco J, Rothbaum D, Stone GW, O’Keefe J, Overlie P, Donohue B, Chelliah N, Timmis GC, Vlietstra RE, Strzeleck M, Puchrowiczs S, O’Neill WW, for the Primary Angioplasty Myocardial Infarction Study Group. A comparison of immediate angioplasty with thrombolytic therapy for acute myocardial infarction. N Engl J Med. 1993;328:673–679.
Gibbons RJ, Holmes DR, Reeder GS, Bailey KR, Hopfenspirger MR, Gersh BJ, for the Mayo Coronary Care Unit, and Catheterization Laboratory Groups. Immediate angioplasty compared with the administration of a thrombolytic agent followed by conservative treatment for myocardial infarction. N Engl J Med. 1993;328:685–691.
Ribeiro EE, Silva LA, Carneiro R, D’Oliveira LG, Gasquez A, Amino JG, Tavares JR, Petrizzo A, Torossian S, Duprat Fo R, Buffolo E, Ellis SG. Randomized trial of direct coronary angioplasty vs intravenous streptokinase in acute myocardial infarction. J Am Coll Cardiol. 1993;22:376–380.
Elizaga J, Garcia EJ, Bueno H. Primary coronary angioplasty vs systemic thrombolysis in acute anterior myocardial infarction: in-hospital results from a prospective randomized trial. Eur Heart J. 1993;14:118. Abstract.
Ribichini F, Steffenino G, Dellavalle A, Meinardi F, Vado A, Feola M, Uslengui E. Primary angioplasty vs thrombolysis in inferior acute myocardial infarction with anterior ST-segment depression: a single-center randomized study. J Am Coll Cardiol. 1996;27:221A. Abstract.
Sowton E. Cardiac infarction and the glucose-tolerance test. BMJ. 1962;1:84–86.
Opie LH. Metabolic response during impending myocardial infarction, I: relevance of studies of glucose and fatty acid metabolism in animals. Circulation. 1972;45:483–490.
Oliver MF, Rowe MJ, Vetter N. Metabolic intervention. In: Oliver MF, Julian DG, Donald KW, eds. Effect of Acute Ischaemia on Myocardial Function. Edinburgh, Scotland: Churchill Livingstone; 1972:354.
Fath-Ordoubadi F, Beatt KJ. Glucose-insulin-potassium therapy for treatment of acute myocardial infarction: an overview of randomized placebo-controlled trials. Circulation. 1997;96:1152–1156.
Malmberg K, Ryden L, Efendic S, Herlitz J, Nicol PTS, Waldenstrom A, Wedel H, Welin L, on behalf of the DIGAMI study group. Randomized trial of insulin-glucose infusion followed by subcutaneous insulin treatment in diabetic patients with acute myocardial infarction (DIGAMI): effect on mortality at 1 year. J Am Coll Cardiol. 1995;26:57–65.