OUP user menu

Newly detected abnormal glucose tolerance: an important predictor of long-term outcome after myocardial infarction

M. Bartnik, K. Malmberg, A. Norhammar, Å. Tenerz, J. Öhrvik, L. Rydén
DOI: http://dx.doi.org/10.1016/j.ehj.2004.09.021 1990-1997 First published online: 2 November 2004

Abstract

Aims Recent data revealed that patients with myocardial infarction (MI) have a high prevalence of previously unknown diabetes mellitus (DM) and impaired glucose tolerance (IGT). The added prognostic importance of this finding has not been prospectively explored. To investigate whether a newly detected abnormal glucose tolerance (IGT or DM) assessed early after an MI, is related to long-term prognosis.

Methods and results Patients (n=168; age 63.5±9.3 years) with MI, no previous DM and admission blood glucose <11.0 mmol/l were followed for major cardiovascular events defined as the composite of cardiovascular death, non-fatal MI, non-fatal stroke or severe heart failure (HF). According to an oral glucose tolerance test (OGTT) before hospital discharge, 55 patients had normal and 113 abnormal glucose tolerance (GT). During the follow-up of median 34 months there were eight cardiovascular deaths, 15 patients had a recurrent MI, six had a stroke and ten severe HF. All patients who died from cardiovascular causes had abnormal GT. The composite cardiovascular event occurred in 31 (18%) patients. The probability of remaining free from cardiovascular events was significantly higher in patients with normal than abnormal GT (p=0.002). Together with previous MI, abnormal GT was the strongest predictor of future cardiovascular events (hazard ratio 4.18; CI 1.26-13.84; p=0.019).

Conclusions Abnormal glucose tolerance is a strong risk factor for future cardiovascular events after myocardial infarction. Since it is common and possible to detect even during the hospital phase it may be a target for novel secondary preventive efforts.

See page 1969 for the editorial comment on this article (doi:10.1016/j.ehj.2004.10.003)

Keywords Prognosis; Myocardial infarction; Survival; Abnormal glucose tolerance

Introduction

People with impaired glucose tolerance have a cardiovascular mortality rate twice that of their counterparts with normal glucose tolerance.1,2 It is only in the recent decade that blood glucose has been recognized as an independent risk factor for cardiovascular morbidity and mortality.3,4 Moreover, there are strong indications that the risk increases in a continuous manner starting well below the current threshold for the diagnosis of diabetes mellitus.3,5,6 Elevated admission blood glucose in patients admitted for myocardial infarction is associated with increased risk for both in-hospital complications and long-term mortality.7–9 The knowledge about patients with myocardial infarction and newly diagnosed abnormal glucose tolerance, however, is still limited.

The GAMI (Glucose tolerance in Patients with Acute Myocardial Infarction) study verified the primary hypothesis that impaired glucose metabolism is common in patients with myocardial infarction and possible to detect already during the early course of the disease.10 This report deals with the secondary objective of the GAMI study, which was to relate long-term outcome to the newly diagnosed glucometabolic state of these patients.

Patients and methods

Patients admitted to the coronary care units at Karolinska and VästerÃ¥s Hospitals in Sweden with confirmed myocardial infarction and without known diabetes were prospectively enrolled as described in detail elsewhere.10 In brief, the inclusion criteria were age <80 years, capillary blood glucose at admission <11.0 mmol/l and serum creatinine <200 μmol/l. The study cohort comprised 181 patients. Patients were followed with regards to the occurrence of nonfatal re-infarction, non-fatal stroke, severe heart failure necessitating hospitalisation and mortality due to cardiovascular disease or other causes. Each event was recorded at the first occurrence only. All participants were followed until death or for at least two years. The follow-up consisted of outpatient visits after 3 and 12 months and a prospective review of the hospital admission registries in the Stockholm and VästerÃ¥s regions. In case of hospitalization the records were reviewed for diagnoses and medical interventions. Information on mortality and causes of death was obtained from the Swedish national death registry, hospital records, copies of death certificates and, when available, autopsy reports. Patients who had moved into another region of the country, were interviewed by telephone concerning their health status and any hospital admissions and a copy of hospital records was obtained if applicable. Follow-up data were complete for all patients.

The study complies with the Declaration of Helsinki. The Ethics Committee of the Karolinska Institute approved the study protocol and all recruited patients gave their written informed consent.

Patients

Demographic data and clinical characteristics at admission were recorded according to the study protocol.10 Glucose metabolism was assessed by means of fasting capillary blood glucose measured daily during the hospital stay and a standard oral glucose tolerance test (OGTT; 75g glucose in 200 ml water) during stable conditions before hospital discharge (day 4 or 5).11 Patients were classified, as defined below, into one of the following categories: diabetes mellitus, impaired glucose tolerance or normal glucose tolerance. Compared to previous reports, four further patients were classified. Two developed overt hyperglycaemia during hospitalization and two were previously reported as missing by logistic reasons.10

Assays

Admission blood glucose was measured with a photometer immediately after sampling (HemoCue AB, Ängelholm, Sweden). Glycosylated hemoglobin A1c (HbA1c) was analysed by high pressure liquid chromatography (HPLC) from blood applied to filter paper (normal values <5.3%; Boehringer Mannheim Scandinavian AB, Bromma, Sweden).12 Venous blood was sampled fasting on the first morning following admission (within 24 hours of admission) and during the OGTT (at 0 and 120min) before discharge. Plasma obtained by centrifugation was stored at –70 °C until assayed.

Plasma lipids, total cholesterol, high-density lipoprotein cholesterol (HDL-cholesterol) and triglycerides, were measured by standard methods and low-density lipoprotein cholesterol (LDL-cholesterol) was calculated according to Friedewald's formula. Apolipoprotein B was analysed by turbidometry with specific antibodies and Syncron LX Systems APO Calibrator (Beckman Coulter Fullerton, CA, USA). Lipoprotein(a) (Lp(a)) was analysed with the TintElize Lp(a) kit (Biopool International, Umeå, Sweden). Free fatty acids (FFA) were determined with an enzymatic colorimetric method from Wako Chemicals GmbH (Neuss, Germany).

Concentrations of insulin and intact pro-insulin were quantified with immunoassays from DAKO Diagnostics Ltd. (Cambridgeshire, UK). C-reactive protein (hs-CRP) was determined by an ultra sensitive latex-enhanced immunoassay (BN II instrument, Dade Boehring, Marburg GmbH, Germany). Plasma PAI-1 activity was measured with Chromolize PAI-1 kits (Biopool International, Umeå, Sweden). Urine concentration of albumin was analysed on Image Immunochemistry systems (Beckman Instrument, Fullerton, CA, USA) and creatinine with use of a Victor 950 (Johnson and Johnson Clinical Diagnostics, Rochester, NY, USA).

Definitions

Myocardial infarction was diagnosed according to the joint recommendations of the ESC and ACC.13 Myocardial infarction was diagnosed if markers of myocardial ischaemia exceeded the upper reference limit on two occasions (troponin T>0.05 g/l or CK-MB >10 μg/l) in the presence of typical symptoms (chest pain >15 min, pulmonary oedema in the absence of valvular heart disease, cardiogenic shock, ventricular tachycardia or ventricular fibrillation) or new Q-waves in at least two of the 12 standard an ECG leads, or ECG indicating acute ischaemia (ST-segment elevation, ST-depression or T-wave inversion).

History of hypertension was recorded if treated prior to enrolment.

Dyslipidaemia was defined in case of HDL-C <0.9 mmol/l in men (<1.0 in women) and/or triglycerides ≥1.7 mmol/l.11 Micro-albuminuria was recognized based on the albumin: creatinine ratio (ACR) analysed in the morning urine sampled before discharge according to the WHO recommendations (ACR> 30mg/g).11

The glucometabolic state was classified based on the WHO criteria for capillary whole blood. Normal glucose tolerance (NGT) was recognized as a fasting blood glucose <6.1 mmol/l and 2-h post-load glucose <7.8 mmol/l. Impaired glucose tolerance (IGT) was defined as fasting blood glucose <6.1 mmol/l and 2-h post-load glucose of 7.8–11.0 mmol/l and diabetes mellitus (DM) as a fasting blood glucose >6.0 mmol/l or a 2-h post-load glucose >11.0 mmol/l.11 The term abnormal glucose tolerance is used to describe the presence of newly detected type 2 diabetes mellitus or IGT.

All patients were treated according to the clinical presentation following the established national and international guidelines without any attention to the glucose tolerance assessed immediately before hospital discharge.14

Outcome measures

Cardiovascular death was defined as death from myocardial infarction, stroke and sudden death without any obvious reason.

A non-fatal re-infarction was defined as a non-fatal myocardial infarction occurring later than 72 h after the index infarction.

Stroke was defined according to the WHO criteria as a neurological deficit observed by a physician and persisting >24 hours without any other disease explaining the symptoms.15 Severe heart failure was recognized when causing hospital admission including intensified, or additional treatment.

A composite outcome was defined as major cardiovascular event representing the first occurrence of stroke, re-infarction, severe heart failure or cardiovascular death.

Statistical analysis

Continuous variables are presented as median (lower and upper quartile) and categorical variables as counts and proportions (%). In the statistical analyses regarding the relation of various clinical and biochemical parameters to the occurrence of outcome events, only patients whose glucometabolic state was classified (n=168), were included. The 13 unclassified patients are described on an individual basis (Table 2). Differences of baseline characteristics between patients grouped by glucose tolerance status (normal and abnormal) were compared by the χ2 test, two-tailed Fisher's exact test or Wilcoxon rank-sum test using STATISTICA version 6.1 (Tables 1, 3). Kaplan–Meier curves were computed for the composite endpoint of major cardiovascular events and Gehan's generalized Wilcoxon test was used to compare patients with normal and abnormal glucose tolerance.

View this table:

Table 2. Details on patients (n=13) that could not be glucometabolically classified before discharge

CharacteristicPatient number
12345678910111213
Sex (male/female)FFMFMMFFMFFFF
Age (years)78675178665179777277736256
BMI (kg/m2)23.321.825.129.427.820.923.527.122.929.025.225.122.5

Medical history
Myocardial infarctYYYY
Heart failureY
HypertensionYY
<(NPL)YY

Treatment
ReperfusionYYYY
β-BlockadeDISYYYYYYY
ACEDISY
StatinDISYYYYYY
AspirinDISYYYYYYY

Biochemical measurements
Admission BG (mmol/L)5.76.54.26.14.68.27.86.27.57.46.45.07.7
FBG day 2 (mmol/L)4.85.04.55.64.96.07.85.45.95.35.25.9
FBG day 3 (mmol/L)3.85.04.54.75.06.36.85.85.55.45.7
FBG day 4–5 (mmol/L)3.65.35.2
HbA1c (%)5.44.94.75.05.75.05.05.04.45.75.65.05.3
Insulin (pmol/l)10260329941305322421163664
Pro-insulin (pmol/l)11.911.84.519.695.56.74.926.56.22.76.8
Cholesterol (mmol/l)5.76.16.85.68.34.96.43.15.26.811.96.5
HDL-C (mmol/l)1.41.01.11.31.01.01.50.90.91.31.61.2
TG (mmol/l)2.32.63.81.14.01.81.01.72.62.06.01.5
HsCRP day 2 (mg/L)2.416.21.66.05.725.81.527248.418.615.23.6
HsCRP day 4–5 (mg/L)10225.768.8

GT classified 3-monthNGTDMNGT
GT classified after 1 yearIGTDMIGT
Follow-up time (days)1318440114911391131999224175320829799
Time to 1st event (days)131844011498611319992241753202799
Major eventDeathMIDeathDeathDeathDeathDeathStroke
Cause of deathNPLMIStrokeNPLMIHF

M: male; F: female; NPL: malignancy; reperfusion: thrombolysis or primary PCI; DIS: at discharge; TG: triglycerides; hsCRP: high sensitivity C-reactive protein; admission BG: capillary blood glucose at admission; FBG: fasting blood glucose; Y: Yes; empty spaces stand for No; –: not available; GT: glucose tolerance (OGTT performed during follow-up); MI: myocardial infarction; HF: heart failure.

View this table:

Table 1. Clinical characteristics

VariablesGlucose tolerancep
Normal (n=55)Abnormal (n=113)
Age (years; median; Q1–Q3)60 (54–67)65 (57–71)0.004
Male (sex)44 (80%)76 (67%)0.080
BMI (kg/m2; median; Q1–Q3)26 (23–28)27 (24–30)0.108

Medical history
Myocardial infarction7 (13%)25 (22%)0.135
Heart failure1 (2%)12 (11%)0.063
Hypertension (treated)17 (31%)38 (34%)0.724
Revascularization (CABG or PCI)6 (11%)15 (13%)0.660
Stroke1 (2%)5 (4%)0.665
Concomitant malignancy3 (6%)3 (3%)0.394

Risk factors for coronary artery disease
Current smoking22 (40%)37 (33%)0.357
Dyslipidaemia38 (69%)78 (70%)0.942
Microalbuminuria2 (4%)12 (14%)0.085

Treatment
Reperfusion therapy28 (51%)42 (38%)0.100
β-BlockadeDIS51 (93%)101 (92%)1.0
ACE-inhibitor or ARB DIS9 (16%)16 (15%)0.760
StatinDIS46 (84%)63 (57%)<0.001
AspirinDIS52 (95%)102 (93%)0.753

Data presented are numbers (%) if not stated otherwise; Q1–Q3: lower and upper quartiles; dyslipidaemia: HDL-cholesterol <0.9 mmol/l (men) or <1.0 (women) and/or triglycerides ⩾1.7 mmol/l; microalbuminuria: albumin/creatinine ratio in the morning urine ⩾30 mg/g; reperfusion therapy: thrombolysis or primary percutaneous coronary intervention (PCI) during current hospital stay. DIS: medication at discharge; comparison between patients with normal and abnormal glucose tolerance by means of the χ2 test, two-tailed Fisher's exact test or Wilcoxon rank-sum test for categorical and continous variables, respectively.

The main analysis used a Cox proportional hazard regression model (SAS version 8.2 PROC PHREG) to find risk factors for the composite endpoint, major cardiovascular event.16 The model is semi-parametric in the sense that no assumption concerning the distribution of the event-free survival times are necessary. To check the proportional hazard assumption, i.e. that the hazard ratio for two subjects with fixed predictors is constant over time, log(-log[survival probability]) for different categories was plotted against time to ensure that the curves were reasonably parallel (continuous variables were categorized into tertiles).

The following model selection strategy was used. Initially Cox regression models that contained each of the candidate predictors one at a time were fitted to determine which sole predictors had some importance (p value <0.2). Then all variables with a p value <0.2 (Table 5), together with current smoking and statin treatment at discharge, which are known to influence survival, were fitted together with abnormal glucose tolerance and forced into the model. All possible combinations of predictors were then fitted and the models were compared using the Akaike information criterion (AIC), which is minus two times the log likelihood plus a penalty function of two times the number of predictors in the model; the smaller the value of this criterion the better the model. The rationale behind this criterion is that if the only difference between two models is that a chance predictor has been included the values of the AIC for the two models will not differ much, but would rather tend to increase. Moreover, the AIC is an approximate measure of the prediction accuracy. The penalty function accounts for the fact that the same data is being used to fit the model and assess it through the likelihood. With this strategy, in contrast to a standard forward or backward selection, the problem that predictors are usually related to each other is avoided. We also avoided the problems with a high number of tests, only getting a single final model.17–19 From a set of almost equal models, with respect to the AIC, the final model was chosen on medical grounds, only including predictors with a significant contribution (p<0.05) to avoid over-parametrization.

A two-sided p value less than 0.05 was regarded as statistically significant.

Results

Pertinent characteristics of the patient material are presented in Table 1. A total of 36 (20%) patients had a previous myocardial infarction, 14 (8%) had a previous history of heart failure and six (3%) a previous stroke. According to the final diagnosis, 70 (39%) patients had Q-wave infarctions and 44 (24%) were located in the anterior wall. Glucose tolerance was classified before hospital discharge (n=168) as normal in 55 (35%) patients and abnormal in 113 (67%), comprising impaired glucose tolerance (n=58) and newly detected diabetes (n=55). The remaining 13 (7%) patients could not be classified due to death (n=2), prolonged unstable condition (n=5), consent withdrawal (n=2) or other condition (n=4) that did not permit the OGTT. A detailed description of these 13 patients is given in Table 2.

Patients with abnormal glucose tolerance were more likely to be older but did not differ regarding the prevalence of cardiovascular risk factors from their counterparts with normal glucose metabolism. Abnormal glucose tolerance was associated with higher plasma pro-insulin, hs-CRP and free fatty acids at discharge, as presented in Table 3.

View this table:

Table 3. Biochemical characteristics

VariablesGlucose tolerancep
Normal (n=55)Abnormal (n=113)
Admission BG (mmol/l)5.8 (5.1–7.1)6.4 (5.8–7.4)0.002

Measurements taken fasting on day 2
Fasting BG (mmol/l)5.2 (4.6–5.7)5.7 (5.1–6.3)0.001
HbA1c (%)4.8 (4.5–5.2)5.0 (4.6–5.3)0.165
Insulin (pmol/l)49 (31–79)68 (42–99)0.027
Pro-insulin (pmol/l)6.1 (4.5–9.7)9.7 (6.5–16.4)<0.001
Pro-insulin/Insulin ratio (%)12 (9–17)17 (10–25)0.026
Cholesterol (mmol/l)6.0 (5.3–7.0)5.9 (5.1–6.7)0.187
HDL-cholesterol (mmol/l)1.17 (1.01–1.39)1.10 (1.00–1.30)0.301
LDL-cholesterol (mmol/l)3.7 (3.1–4.4)3.7 (2.9–4.4)0.556
LDL/HDL ratio2.9 (2.5–4.3)3.2 (2.5–4.0)0.960
Triglycerides (mmol/l)2.3 (1.5–3.2)2.1 (1.5–3.0)0.494
Apolipoprotein B (g/l)1.15 (0.87–1.30)1.05 (0.89–1.23)0.215
Lipoprotein(a) (g/l)0.14 (0.05–0.36)0.18 (0.04–0.43)0.370
hs-CRP (mg/l)12.3 (4.9–20.1)11.4 (4.5–28.7)0.683

Measurements before discharge (day 4–5)
PAI-1 activity (IU/ml)6.4 (3.3, 14.4)10.4 (4.0, 20.9)0.103
hs-CRP (mg/l)12.7 (5.3, 28.0)23.0 (9.4, 65.2)0.002
Free fatty acids (mEq/l)0.39 (0.30, 0.61)0.57 (0.36, 0.73)0.009

Oral glucose tolerance test before discharge (day 4–5)
Fasting BG (mmol/l)4.8 (4.5–5.3)5.3 (4.8–5.7)<0.001
BG 60min (mmol/l)9.1 (7.8–10.6)11.9 (10.3–13.0)<0.001
BG 120min (mmol/l)6.5 (5.9–7.1)10.3 (8.8–11.9)<0.001
Insulin (pmol/l)52 (33–70)56 (37–90)0.110
Pro-insulin (pmol/l)4.8 (3.9–6.8)6.8 (4.6–9.3)<0.001
Pro-insulin/insulin ratio (%)10 (9–14)12 (8–17)0.142

Values presented as median (lower–upper quartiles); BG: blood glucose; hs-CRP: high sensitivity C-reactive protein; PAI-1 activity, plasminogen activating inhibitor type 1 activity; comparison between patients with normal and abnormal glucose tolerance by means of Wilcoxon rank-sum test.

Outcomes

Table 4 summarizes all events that were recorded during a median (lower, upper quartile) follow-up time of 34 (27, 40) months. During this time 20 patients died. Twelve deaths were attributed to cardiovascular disease, six to cancer and two to other diseases. Cardiovascular deaths occurred only in patients with abnormal glucose tolerance (eight deaths) or those who could not be classified due to their unstable condition (four deaths). The crude relative frequencies of various cardiovascular events are presented in Fig. 1. Among the 13 unclassified patients two died within 72 h following admission and a further two within four weeks. No events occurred in patients who withdrew their consent, neither in patients, who due to other conditions did not perform an OGTT before discharge.

Fig. 1

Major cardiovascular events in relation to glucose tolerance. Relative frequency of each major cardiovascular event among patients with normal (NGT; open bars) or abnormal glucose tolerance (AGT=IGT or DM; hatched bars). Reported events comprise severe heart failure, non-fatal re-infarction, non-fatal stroke, cardiovascular death and total number of events and are reported as crude proportions.

View this table:

Table 4. Events in relation to glucometabolic status

Type of eventGlucose tolerance (n)Total
Normal (55)Impaired (58)Diabetes (55)
Death36514
Non-cardiovascular3216
Cardiovascular0448
Stroke1416
Non-fatal re-infarction29415
Severe heart failure16310
Sum of Major cardiovascular event4231239

Table summarizes events that occurred until death or Dec 13, 2002. Each event was recorded only once. Major cardiovascular event represents a non-fatal stroke, non-fatal re-infarction, severe heart failure (necessitating hospitalisation) or cardiovascular death.

Out of 168 patients, who were classified according to glucose tolerance before hospital discharge, 31 (18%) experienced at least one major cardiovascular event (Table 4, Fig. 1). The composite endpoint occurred significantly more in patients with abnormal glucose tolerance 28 (17%) than those with normal glucometabolic state 3 (2%) (p=0.003). Kaplan–Meier curves representing time to the first cardiovascular event diverged widely for patients with normal and those with abnormal glucose tolerance (Fig. 2). The probability of remaining free from cardiovascular events was significantly poorer in patients with abnormal than normal glucose tolerance (p=0.002).

Fig. 2

Time to the first major cardiovascular event Kaplan–Meier curves for patients with normal (dashed line) and abnormal glucose tolerance (solid line). The Wilcoxon test for difference between groups, p=0.0021. Major cardiovascular events represent the first occurrence of a non-fatal stroke, non-fatal re-infarction, severe heart failure necessitating hospitalisation or cardiovascular death. Numbers below graph are the number of patients at risk at different times of observation.

The incidence of major cardiovascular events was related to several variables identified by a simple Cox regression (Table 5). The best prediction of a future cardiovascular event by a multiple Cox regression was achieved in a model including previous myocardial infarction, previous stroke and abnormal glucose tolerance (Table 6). Apart from previous myocardial infarction, abnormal glucose tolerance was the second strongest predictive factor for future major cardiovascular events with a hazard ratio of 4.2 (95% CI, 1.2–13.8; p=0.019).

View this table:

Table 5. Candidate predictors related to major cardiovascular events

VariablepHHazard ratio95% Confidence interval
Previous myocardial infarction<0.0013.481.69–7.14
Previous heart failure0.0023.861.65–9.08
Abnormal glucose tolerance0.0081.711.15–2.55
Age (years)0.0151.051.01–1.10
Reperfusion therapy0.0240.380.16–0.89
Pro-insulin/insulin ratio (day 2)0.0536.250.96–40.71
Previous stroke0.0593.150.95–10.45
Fasting blood glucose (day 2)0.0811.250.97–1.62
Blood glucose (120min)0.0871.100.99–1.24
Free fatty acids (mEq/l)0.1512.050.77–5.51
PAI-1 activity (IU/ml)0.1800.980.95–1.01

Results of simple Cox regression for variables with p<0.20.

120 min after glucose load (day 4–5); reperfusion therapy: thrombolysis or primary PCI during current hospitalisation.

View this table:

Table 6. Adjusted risks for major cardiovascular events

VariablepHHazard ratio95% CI
Previous myocardial infarction0.0013.381.62–7.04
Abnormal glucose tolerance0.0194.181.26–13.84
Previous stroke0.0373.681.08–12.49

Cox proportional hazard regression.

Discussion

The main finding in this study is that abnormal glucose tolerance, newly identified before hospital discharge of patients with myocardial infarction, characterised individuals with a high likelihood for major cardiovascular events. Together with a previous myocardial infarction or stroke, the glucometabolic state was the strongest predictor of future cardiovascular events.

An advantage with the current report is that evaluation of the glucometabolic state was achieved early and at the same time in all patients. According to our knowledge, the GAMI trial is the first to assess glucose metabolism by means of an OGTT within the first week in a contemporary population with myocardial infarction. As previously reported there was a very high correlation between OGTT performed before hospital discharge and three months thereafter opening a possibility for risk prediction already within the very early post-myocardial infarction period.20 Moreover, classification of glucometabolic state by means of OGTT has a particular value for the risk prediction of cardiovascular events in elderly patients.21–23

The high incidence of cardiovascular events in the early post-infarction period as illustrated by rapidly diverging Kaplan-Meier curves resembles the impact of impaired fasting glucose on all-cause mortality reported in a large registry of patients with coronary artery disease.24

A substantial proportion of cardiovascular events would not be identified if classification were conducted several weeks after hospital discharge as emphasised by the LAMBDA investigators.25

In agreement with other investigators, the difference between the incidence of cardiovascular events among patients with newly detected diabetes and less pronounced impairment of glucose tolerance was negligible.4,22–24 The majority of events did, in fact, occur in patients with impaired glucose tolerance as also observed by the DECODE Investigators.4,23 This underlines the continuous risk with increasing blood glucose already apparent at levels considered fairly normal.3

The overall mortality of 11% is lower than the one year mortality of 13% described in non-diabetic patients followed by the Register of Information and Knowledge about Swedish Heart Intensive Care Admissions (RIKS-HIA).26 However, the GAMI patients were younger and did not have admission hyperglycaemia >11 mmol/l or renal insufficiency and in addition they were able to participate in an OGTT on day 4–5. Thus, the patients studied in this report represent a relatively healthy subgroup of those admitted with myocardial infarction in every-day CCU settings. In this perspective the GAMI cohort is well suited to generate hypotheses to be confirmed in a larger setting.

In the multivariate analysis, glucometabolic status, together with a previous myocardial infarct and stroke, remained as the only risk predictors. Classical cardiovascular risk factors that contribute to the development and extent of arteriosclerosis did not critically influence the tempo in which subsequent major cardiovascular events occurred in contrast to the glucometabolic milieu. Abnormal glucose tolerance by its multiple interactions with inflammatory response, pro-oxidative stress and not the least pro-coagulatory properties may play an important role for promoting adverse events.9,27 In this context, the completeness of follow-up and the extended metabolic characterisation, including hs-CRP, PAI-1 activity, pro-insulin, insulin levels and plasma free fatty acids can be accounted for additional advantage of this study.

Reperfusion therapy, which was less commonly applied in patients who experienced major cardiovascular events, together with age and use of statins, did not improve the Cox multiple regression model. In contrast, abnormal glucose tolerance was associated with four times higher risk for the composite of cardiovascular death, re-infarction, stroke or severe heart failure during a median follow-up time of 34 months.

The important role of fasting glycaemia and the unexpected strength of its impact on the mortality rate after percutaneous coronary interventions has very recently been proven and discussed by Muhlestein et al., who found the prevalence of abnormally elevated fasting glucose of 61% in patients with CAD undergoing PCI procedures.28

Study limitations

The GAMI patient population is fairly small. Thus despite a seemingly convincing message, this observation needs confirmation in a larger study. The Euro Heart survey on diabetes and the heart29 offers such opportunity. Preliminary one-year follow-up data in 2000 of the almost 5000 included patients in total showed an increased mortality among patients with newly detected pertubations in the glucose metabolism (Rydén et al., unpublished data on file).

The rigorous protocol applied in the GAMI study resulted in a loss of 13 patients from further analysis. Several of these patients would probably have had abnormal glucose tolerance if tested. This assumption is supported by their rather high average age, fairly high BMI and that several of them had a history of previous myocardial infarction, heart failure or hypertension (Table 2). Glucometabolic perturbations were in fact subsequently detected in three of them and 5 out of 13 had elevated HbA1c. Although this loss of patients decreased the statistical power, it was perceived appropriate to limit observations to patients in whom the metabolic status was precisely characterised rather than applying an 'intention to treat' principle as in trials addressing treatment modalities.

Clinical implications

An oral glucose tolerance test can easily be added to the standard risk evaluation procedures in the hospital setting and may become an important value for future planning of enhanced secondary prevention. Abnormal glucose tolerance is a strong risk factor for future cardiovascular events after myocardial infarction. As it is common and possible to detect already during the hospital phase, it may be a target for novel secondary preventive efforts.

Acknowledgments

The authors are grateful to Camilla Hage and Marja-Leena Ojutkangas for taking care of patients and collecting the follow-up data.

Footnotes

  • The study was supported by the Swedish Heart and Lung Foundation, AFA Insurance, King Gustaf V's and Queen Victoria's Foundation and Aventis Pharma. These grants, obtained in competition, were totally unrestricted and unconditional.

References

  1. [1]
  2. [2]
  3. [3]
  4. [4]
  5. [5]
  6. [6]
  7. [7]
  8. [8]
  9. [9]
  10. [10]
  11. [11]
  12. [12]
  13. [13]
  14. [14]
  15. [15]
  16. [16]
  17. [17]
  18. [18]
  19. [19]
  20. [20]
  21. [21]
  22. [22]
  23. [23]
  24. [24]
  25. [25]
  26. [26]
  27. [27]
  28. [28]
  29. [29]
View Abstract