Author + information
- Received November 19, 2002
- Revision received March 9, 2003
- Accepted March 20, 2003
- Published online September 3, 2003.
- Iwan C.C van der Horst, MD*,
- Felix Zijlstra, MD, PhD, FACC‡,* (, )
- Arnoud W.J van't Hof, MD, PhD*,
- Carine J.M Doggen, MSc, PhD∥,
- Menko-Jan de Boer, MD, PhD*,
- Harry Suryapranata, MD, PhD, FACC*,
- Jan C.A Hoorntje, MD, PhD*,
- Jan-Henk E Dambrink, MD, PhD*,
- Rijk O.B Gans, MD, PhD§,
- Henk J.G Bilo, MD, PhD†,
- Zwolle Infarct Study Group
- ↵*Reprint requests and correspondence:
Dr. Felix Zijlstra, Department of Cardiology, University Hospital Groningen, Hanzeplein 1, Postbus 30.001, 9700 RB Groningen, The Netherlands.
Objectives In this study we considered the question of whether adjunction of glucose-insulin-potassium (GIK) infusion to primary coronary transluminal angioplasty (PTCA) is effective in patients with an acute myocardial infarction (MI).
Background A combined treatment of early and sustained reperfusion of the infarct-related coronary artery and the metabolic modulation with GIK infusion has been proposed to protect the ischemic myocardium.
Methods From April 1998 to September 2001, 940 patients with an acute MI and eligible for PTCA were randomly assigned, by open-label, to either a continuous GIK infusion for 8 to 12 h or no infusion.
Results The 30-day mortality was 23 of 476 patients (4.8%) receiving GIK compared with 27 of 464 patients (5.8%) in the control group (relative risk [RR] 0.82, 95% confidence interval [CI] 0.46 to 1.46). In 856 patients (91.1%) without signs of heart failure (HF) (Killip class 1), 30-day mortality was 5 of 426 patients (1.2%) in the GIK group versus 18 of 430 patients (4.2%) in the control group (RR 0.28, 95% CI 0.1 to 0.75). In 84 patients (8.9%) with signs of HF (Killip class ≥2), 30-day mortality was 18 of 50 patients (36%) in the GIK group versus 9 of 34 patients (26.5%) in the control group (RR 1.44, 95% CI 0.65 to 3.22).
Conclusions Glucose-insulin-potassium infusion as adjunctive therapy to PTCA in acute MI did not result in a significant mortality reduction in all patients. In the subgroup of 856 patients without signs of HF, a significant reduction was seen. The effect of GIK infusion in patients with signs of HF (Killip class ≥2) at admission is uncertain.
Since the early 1960s, glucose-insulin-potassium (GIK) infusion has been advocated as therapy in the early hours after acute myocardial infarction (MI) (1–4). An overview involving nine of these early studies with 1,932 patients described that GIK might play an important role in reducing in-hospital mortality after acute MI (5). In four studies in which a high dose was used, a nonsignificant reduction in mortality from 12% in the control group to 6.5% in the GIK group was observed (6–9). High-dose GIK denotes an infusion of 30 g of glucose, 50 U of insulin, and 80 mmol potassium per liter at a rate of 1.5 ml/kg body weight/h (10). This suppresses arterial free fatty acid levels and myocardial free fatty acid uptake, and induces a maximal increase in myocardial glucose uptake. The main effect of the GIK infusion was considered to be the beneficial effect of administration of glucose to the ischemic myocardium (11–14).
At present, primary coronary transluminal angioplasty (PTCA) is regarded to be the most effective reperfusion treatment of acute MI (15–17). The complementary role of GIK to reperfusion therapy has been proposed (18,19). Therefore, we performed a randomized trial to investigate the value of GIK when combined with PTCA for acute MI.
All consecutive patients with symptoms consistent with acute MI of >30 min duration, presenting within 24 h after the onset of symptoms and with an ST-segment elevation of more than 1 mm (0.1 mV) in two or more contiguous leads on the electrocardiogram, or new onset left bundle branch block, were evaluated for inclusion in this single-center study. Patients presented at our center and patients referred for treatment of high-risk myocardial infarction (MI) from nine referring hospitals without angioplasty facilities were included. Patients were excluded when pre-treated with thrombolysis or when an illness associated with a marked restricted life expectancy was present. Before randomization baseline characteristics were recorded. The research protocol was reviewed and approved by the medical ethics committee of our hospital, and patients were included after informed consent.
After admission, patients were randomly assigned, with the use of a computerized randomization program, to either GIK infusion or no infusion. In the GIK group, a continuous infusion of 80 mmol potassium chloride in 500 ml 20% glucose with a rate of 3 ml/kg body weight/h over an 8- to 12-h period in a peripheral venous line was given, as soon as possible. A continuous infusion of short-acting insulin (50 U Actrapid HM, Novo Nordisk, Copenhagen, Denmark) in 50 ml 0.9% sodium chloride was started with the use of a pump (Perfusor-FM, B. Braun, Melsungen, Germany). Baseline insulin-infusion dose and hourly adjustments of the insulin dose were based on a nomogram (20)to obtain blood-glucose levels between 7.0 and 11.0 mmol/l, based on measurements of whole-blood glucose (Modular System, Roche/Hitachi, Basel, Switzerland). Additional treatment in all patients consisted of intravenous heparin, nitroglycerin, and aspirin.
After the infusion was started, all patients went to the catheterization laboratory where both coronary arteries were visualized; PTCA was performed with standard techniques if the coronary anatomy was suitable for angioplasty. Successful reperfusion was defined by Thrombolysis In Myocardial Infarction grade 3 blood flow in combination with myocardial blush grades 2 and 3 (21,22). After sheath removal, low-molecular-weight heparin was given for 1 to 3 days.
The primary end point of the study was 30-day mortality. Secondary end points were recurrent infarction, repeat coronary angioplasty, and the composite incidence of death, recurrent infarction, or repeat coronary angioplasty. The incidence of death was evaluated in predefined subgroups including age <60 years versus ≥60 years, men versus women, anterior MI versus nonanterior MI, diabetics versus nondiabetics, Killip class 1 versus Killip class ≥2, and successful reperfusion versus no successful reperfusion. Analyses were performed according to the intention-to-treat principle. Relative risks (RRs) were calculated by dividing the cumulative incidence rate in the GIK group by the cumulative incidence rate in the control group. Differences between group means at baseline were assessed with the two-tailed Student ttest. Chi-square analysis or Fisher exact test was used to test differences between proportions. Survival was calculated by the Kaplan-Meier product-limit method. The log-rank test was used to evaluate differences in survival curves between the two treatment groups. The Cox proportional-hazards regression model was used to calculate RRs adjusted for differences in baseline characteristics. Statistical significance was considered a two-tailed p value < 0.05. The SPSS Version 10.1 (SPSS Inc., Chicago, Illinois) was used for all statistical analysis.
We postulated that GIK would induce a risk reduction of 66% (19). With an estimated absolute mortality rate of 6% in the control group, a power of 0.80 and an alpha of 0.05, 820 patients needed to be included in the study. In order to have 820 patients in each major predefined subgroup (patients in Killip class 1, patients without diabetes mellitus, and patients with successful reperfusion), we planned to enroll 940 patients. Interim analysis with regard to safety was planned and performed after inclusion of 240 patients, and the study was continued according to protocol. Enrollment of patients started in April 1998 and ended September 2001.
Of the 940 patients enrolled, 476 patients were randomly assigned to the GIK group, and 464 patients to the control group. Table 1shows clinical and demographic characteristics of the groups; with the exception of male gender, there were no statistically significant differences between both groups. The GIK infusion was started 15 to 20 min after admission. Average door-to-balloon time was 45 min in the GIK group and 48 min in the control group. After coronary angiography, 860 patients (90.5%) underwent PTCA, 38 patients (4.0%) were referred for coronary artery bypass grafting within 7 days after initial stabilization, and 42 patients (4.5%) were treated conservatively. There were no major differences in blood-glucose levels between the GIK group and the control group at admission (median blood-glucose level 8.5 mmol/l in both groups) and 16 h after admission (median blood-glucose level 7.7 mmol/l in the GIK group and 8.1 mmol/l in the control group). Side effects such as hypoglycemia, hyperkaliemia, and severe phlebitis were not observed.
Of the 940 patients, 50 (5.3%) had died at 30 days, 23 of 476 patients (4.8%) in the GIK group versus 27 of 476 patients (5.8%) in the control group (RR 0.82, 95% confidence interval [CI] 0.46 to 1.46) (Table 2, Fig. 1). In 856 of the 940 patients (91.1%) without signs of heart failure (HF) (Killip class 1), the mortality rate was 5 of 426 patients (1.2%) in the GIK group versus 18 of 430 patients (4.2%) in the control group (RR 0.28, 95% CI 0.1 to 0.75) (Tables 2 and 3, Fig. 2). ⇓In this subgroup of patients, a higher number of patients died of HF in the control group (0.7% in the GIK group vs. 2.8% in the control group) (Table 4). In the 84 patients (8.9%) with signs of HF (Killip class ≥2), mortality rate was 18 of 50 patients (36%) in the GIK group versus 9 of 34 patients (26.5%) in the control group (RR 1.44, 95% CI 0.65 to 3.22). In this subgroup of patients, 34.0% in the GIK group versus 20.6% in the control group died of HF (Table 4). In 99 patients (10.5%) with a history of diabetes mellitus, 2 of the 50 patients (4.0%) in the GIK group died versus 6 of 49 patients (12.2%) in the control group (RR 0.30, 95% CI 0.06 to 1.56). Mortality rates in the other predefined subgroups are shown in Table 2. Figure 3shows RRs and 95% CIs in each subgroup. Causes of death are represented in Table 4.
Recurrent infarction occurred in 11 patients: 4 patients (0.8%) in the GIK group and 7 patients (1.5%) in the control group (RR 0.55, 95% CI 0.16 to 1.90) (Table 5). The composite end point was comparable in the GIK group and the control group (RR 0.79, 95% CI 0.50 to 1.24). In patients without HF (Killip class 1), the composite end point showed a significant advantage of GIK (RR 0.48, 95% CI 0.27 to 0.87). Gender, age, previous MI, history of diabetes mellitus, smoking status, Killip class, anterior MI, multivessel disease, and successful reperfusion were associated with 30-day mortality in the overall population. In multivariate analysis in the overall study population, age, previous MI, Killip class, and successful reperfusion were independent prognostic factors for 30-day mortality. After adjustment for these factors in the overall population, RR of mortality with GIK became 0.61 (95% CI 0.34 to 1.10). In the group of Killip class 1 patients, the beneficial effect of GIK remained significant with an adjusted RR of 0.28 (95% CI 0.10 to 0.77).
The 30-day mortality rate was lower in the GIK group (4.8%) compared with the control group (5.8%), but was not statistically significant. In the large predefined subgroup of 856 patients without signs of HF (Killip class 1), a significant reduction of mortality was seen (1.2% in the GIK group compared with 4.2% in the control group). In the 84 patients with signs of HF (Killip class ≥2), we did not observe a significant difference in mortality between the two groups. In the other predefined subgroups, such as patients with successful reperfusion, GIK did not result in a significant beneficial effect (p = 0.14) (Table 2). The definition of several subgroups, particularly with small numbers of patients, carries the risk of finding significance due to chance alone, that is, either beneficial or detrimental. Nevertheless, the principal finding of benefit of GIK in patients without signs of HF and, in particular, the RR and CI (Fig. 3) are unlikely caused by chance.
Previous studies with GIK
The beneficial effect of GIK in acute MI has been suggested by several clinical studies (2,5). The first small study that combined GIK with thrombolysis reported an improved ejection fraction and wall motion (9). In the Estudios Cardiologicos Latinoamerica (ECLA) pilot trial, a beneficial effect on in-hospital mortality was observed primarily in the subgroup of patients in which GIK was combined with thrombolysis (19). However, in the Polish-Glucose-Insulin-Potassium (Pol-GIK) trial, with 954 patients using low-dose GIK, mortality in the GIK group was even significantly higher than in the control group (8.9% vs. 4.8%, p = 0.01) (23).
When compared with the results of the ECLA pilot trial of the subgroup of 252 patients treated with GIK and reperfusion, the mortality rate reduction in our study, that is, 1% in the overall population, was much smaller (19). Several facts could explain the difference between both studies. First, all patients in both groups of our study were treated with PTCA compared with prior studies that used preferentially thrombolysis or no reperfusion therapy. Second, the mortality rate is low, due to the effect of our standard care for patients with acute MI, that is, the short time from onset of symptoms to treatment, the use of additional therapies, and a high rate of stenting (57.1%). Third, in our study the number of patients with signs of HF at admission was much higher than in the ECLA pilot trial. In this subgroup of patients, an increased mortality rate was observed in our study. It is harder to reduce the mortality even more in a population with such a low mortality rate. The mortality rate in the control group in the ECLA pilot trial was 15.2%, a percentage that was even higher than the mortality rate of patients in the control group of patients that did not receive reperfusion therapy.
GIK in patients with signs of HF
The effect of GIK infusion in patients with signs of HF (Killip class ≥2) is uncertain. Despite liberal use of intra-aortic balloon pumping (9.3%), more patients in the GIK group needed mechanical ventilation (2.5% in the GIK group vs. 0.9% in the control group). Heart failure was an important cause of death in patients admitted with Killip class ≥2 (Table 5). Therefore, it is a possibility that in some patients the potential positive effects of GIK on metabolism were overruled by a negative effect of the volume load on the hemodynamic state (24–26). For a patient of 80 kg in our study, our infusion rate resulted in the infusion of 1,920 ml in 8 h. Administering such a volume load to patients with an impaired left ventricular function may induce hemodynamic instability, even though the metabolic effect may be beneficial. However, we did not expect this infusion rate to be detrimental because in early studies hemodynamic disturbances, such as pulmonary congestion, dyspnea, and edema, occurred in only 1% to 3% of patients treated with GIK (6). In experimental and clinical studies, the infusion of insulin and glucose did increase the contractile force (27–29), although others found that GIK made no differences on hemodynamic parameters (8).
Studies with strict metabolic regulation
In the Diabetes Insulin-Glucose in Acute Myocardial Infarction (DIGAMI) study with 620 patients with an acute MI and diabetes mellitus or glucose at admission >11.0 mmol/l, the combination of insulin-glucose infusion followed by intensive insulin treatment resulted in better glycemic control in the insulin-glucose group (20). At one-year follow-up, mortality was lower in the insulin-glucose group, with 18.6% versus 26.1% in the control group (p = 0.03) (30).
It is suggested that GIK is mainly beneficial by the effects of insulin on the myocardium and the whole body (31,32). The results of a study of intensive insulin therapy in critically ill patients to maintain blood glucose below 6.1 mmol/l gives direction for future research in which metabolic interventions in acute MI are directed towards normoglycemia (33).
GIK and time to reperfusion
It has been postulated that through reduction in the extent of ischemic myocardial damage and suppression of free fatty acid levels, GIK therapy prevents reperfusion injuries that may occur after successful revascularization (6). Protection of the cell membrane of ischemic cardiac cells may also improve reflow after reperfusion and protect against no-reflow phenomenon by reducing cell swelling and microvascular compression. It has been estimated that GIK therapy has the potential to protect ischemic myocardium before reperfusion for 10 h or even longer (34). Our study offered the opportunity to analyze the effect of GIK in patients with different ischemic time, that is, time to admission and door-to-balloon time. We could not find a clear relation between the time delays and the effect of GIK (Tables 2 and 3) (34).
With the possible exception of volume overload in patients with Killip class ≥2, there were no adverse effects of GIK. Phlebitis did not occur in our study, even though the majority of patients received the infusion via a peripheral venous line. It is most likely that this is due to the short infusion period, maximally 12 h. In the ECLA pilot trial, 17% developed mild phlebitis, and only 2% developed severe phlebitis after 24 h of GIK infusion (19).
With the observed low mortality rates in the GIK group and control group in the overall population, our study could not detect a significant difference in mortality. A more realistic risk reduction for future research may be a risk reduction of 30% in patients with Killip class 1. Therefore, in future studies, over 2,000 patients have to be included. By including 940 patients in our present study, the number of patients in some of the subgroups, such as in patients with a history of diabetes mellitus, were too small to draw definite conclusions. The subgroups of patients in our study were predefined, but we did not use a method to correct for multiple comparisons to analyze them. The role of GIK in patients with signs of HF is unclear, and the volume load may be detrimental. A more tailored approach in patients with HF at admission seems warranted, such as the admission of glucose 30%, an infusion of no more than 500 ml, or a period of infusion of 12 to 24 h (6). Alternatively, in these patients the GIK infusion should be combined with other measures, such as continuous monitoring of hemodynamic parameters (28). Our study had an open-label design, because the need to administer a large volume of fluid, frequent blood-glucose monitoring, and adjustments of the insulin dose in the study group is both unethical and unpractical to organize in a placebo-controlled study.
Glucose-insulin-potassium infusion as adjunctive therapy to PTCA in acute MI seems promising. In the large subgroup of patients without signs of HF (over 90% of all patients), we did find a significant beneficial effect. The favorable results of GIK in patients without signs of HF did encourage us to start a large, multicenter, randomized trial to determine the optimal metabolic management in patients with a heart rate <90 beats/min, a systolic blood pressure >100 mm Hg, and no third heart sound or pulmonary rales. Alternative regimens for metabolic modulation may be necessary for patients with signs of HF.
☆ Supported by a grant from the Netherlands Heart Foundation (99.028).
- confidence interval
- Estudios Cardiologicos Latinoamerica pilot trial
- heart failure
- myocardial infarction
- primary coronary transluminal angioplasty
- relative risk
- Received November 19, 2002.
- Revision received March 9, 2003.
- Accepted March 20, 2003.
- American College of Cardiology Foundation
- Fath-Ordoubadi F.,
- Beatt K.J.
- Heng M.K.,
- Norris R.M.,
- Singh B.N.,
- Barratt Boyes C.
- Stanley A.W.,
- Prather J.W.
- Cave A.C.,
- Ingwall J.S.,
- Friedrich J.,
- et al.
- Apstein C.S.,
- Gravino F.N.,
- Haudenschild C.C.
- Vanoverschelde J.L.,
- Janier M.F.,
- Bakke J.E.,
- Marshall D.R.,
- Bergmann S.R.
- Diaz R.,
- Paolasso E.A.,
- Piegas L.S.,
- et al.
- Malmberg K.,
- Efendic S.,
- Ryden L.E.
- van't Hof A.W.,
- Liem A.,
- Suryapranata H.,
- Hoorntje J.C.A.,
- de Boer M.J.,
- Zijlstra F.
- Ceremuzynski L.,
- Budaj A.,
- Czepiel A.,
- et al.
- Sjöstrand F.,
- Edsberg L.,
- Hahn R.G.
- Wildenthal K.,
- Mierzwiak D.S.,
- Mitchell J.H.
- Sasso F.C.,
- Carbonara O.,
- Cozzolino D.,
- et al.
- Malmberg K.,
- Ryden L.,
- Efendic S.,
- et al.
- Eberli F.R.,
- Weinberg E.O.,
- Grice W.N.,
- Horowitz G.L.,
- Apstein C.S.