Author + information
- Received July 7, 2008
- Revision received March 12, 2009
- Accepted March 20, 2009
- Published online May 1, 2009.
- Dan E. Høfsten, MD, PhD⁎,⁎ (, )
- Brian B. Løgstrup, MD⁎,
- Jacob E. Møller, MD, PhD†,
- Patricia A. Pellikka, MD‡ and
- Kenneth Egstrup, MD, DMsc⁎
- ↵⁎Reprint requests and correspondence:
Dr. Dan E. Høfsten, Department of Medical Research, Funen Hospital Svendborg, 5700 Svendborg, Denmark
Objectives We studied the influence of abnormal glucose metabolism on left ventricular (LV) function and prognosis in 203 patients with acute myocardial infarction.
Background Abnormal glucose metabolism is associated with increased mortality after acute myocardial infarction. This appears to be particularly attributable to an increased incidence of post-infarction congestive heart failure. A relationship between glucose metabolism and LV function could potentially explain this excess mortality.
Methods In patients without known diabetes, glucose metabolism was determined using an oral glucose tolerance test before discharge. LV function was assessed using echocardiographic measurements (LV end-diastolic volume, LV end-systolic volume, LV ejection fraction, restrictive diastolic filling pattern, early transmitral flow velocity to early diastolic mitral annular velocity ratio [E/e′], and left atrial volume index) and by measuring plasma N-terminal pro-B-type natriuretic peptide levels.
Results After adjustment for age and gender, a linear relationship between the degree of abnormal glucose metabolism was observed for each marker of LV dysfunction (ptrend < 0.05) with the exception of left atrial volume index (p = 0.10). During a median follow-up of 21 months, 32 patients died, and 39 patients met the secondary end point of death or hospitalization for heart failure. After adjustment for differences in LV function, as well as other relevant characteristics, newly detected, as well as known diabetes were independent predictors of both all-cause mortality (hazard ratios [HR]: 4.2 [95% confidence interval (CI): 1.1 to 17.1] and HR: 5.7 [95% CI: 1.3 to 25.2], respectively), and the composite of death or hospitalization for heart failure (HR: 4.3 [95% CI: 1.2 to 15.6] and HR: 5.8 [95% CI: 1.5 to 22.3], respectively). Comparable nonsignificant trends were observed for patients with impaired glucose tolerance.
Conclusions Although perturbations in glucose metabolism were linearly associated with impairment of LV function in the early phase of acute myocardial infarction, this relationship alone did not explain the excess mortality in patients with newly detected or known diabetes.
Diabetes mellitus (DM) is a risk factor for the development of ischemic heart disease (1,2), and patients with DM who suffer an acute myocardial infarction (AMI) have a worse outcome than nondiabetic patients do (1,3). Furthermore, perturbations in glucose metabolism below the diagnostic threshold of DM are associated with a risk of ischemic heart disease that exceeds that of patients with normal glucose metabolism but is lower than that of patients with overt diabetes (4). Thus, the severity of abnormal glucose metabolism appears to represent a continuum that translates into a gradual increase in the risk of cardiac morbidity. Although fasting glucose levels have been demonstrated to predict outcome in nondiabetic patients with AMI (5,6), glucose metabolism may be misclassified in as many as 41% of nondiabetic patients with ischemic heart disease, when the diagnosis is solely based on fasting glucose levels (7). The use of an oral glucose tolerance test (OGTT) has therefore been advocated (7–9).
In patients with AMI, the increased mortality rates associated with DM seem especially to be due to an increased risk of post-infarction congestive heart failure (3,10). This suggests that abnormal glucose metabolism may play a pivotal role in left ventricular (LV) remodeling. It is, however, uncertain how various degrees of dysglycemia affect LV function in patients without known DM. Consequently, it is not known whether the relationship between dysglycemia and prognosis in patients with AMI is caused by underlying differences in LV function.
The aims of our study were to assess the relationship between glucose metabolism and LV function in the early phase of AMI, and to determine whether the adverse prognosis in patients with dysglycemia is attributable to LV dysfunction.
From August 2004 to June 2006, we consecutively included 203 patients admitted to the coronary care unit at Funen Hospital, Svendborg, Denmark, with a validated diagnosis of AMI based on a dynamic rise in troponin-T >0.1 μg/l, as well as either typical symptoms, characteristic electrocardiographic changes, or both. The study was approved by the Regional Ethics Committee of Southern Denmark and The Danish Data Protection Agency, and written informed consent was obtained from each patient.
Assessment of glucometabolic control
In patients with no history of diabetes, a 2-h OGTT was performed before discharge. After an overnight fast, a 75-g glucose solution was administered and capillary blood glucose levels were measured using a HemoCue 201+ glucose analyzer (HemoCue AB, Ängelholm, Sweden). Patients were classified according to the 1999 World Health Organization criteria for whole blood glucose levels (11) as having newly detected DM if fasting blood glucose was ≥6.1 mmol/l (110 mg/dl) or if the 2-h post-load blood glucose was ≥11.1 mmol/l (200 mg/dl). Impaired glucose tolerance (IGT) was defined as a 2-h post-load blood glucose ≥7.8 mmol/l (140 mg/dl) unless fasting blood glucose met the criteria for newly detected DM. Impaired fasting glycemia (IFG) was defined as a fasting blood glucose ≥5.6 mmol (100 mg/dl) unless 2-h post-load blood glucose met the criteria for newly detected DM.
Assessment of LV function
Echocardiography was performed immediately after enrollment, using a Vivid 7 ultrasound system (GE Medical Systems Inc., Horten, Norway). Images were blinded and analyzed offline by a single investigator.
The LV end-systolic and end-diastolic volumes were measured from standard 2-dimensional apical 4- and 2-chamber projections using Simpsons biplane method, and LV ejection fraction (LVEF) was calculated (12). Mitral inflow was assessed in the apical 4-chamber view, using pulsed-wave Doppler echocardiography, with the Doppler beam aligned parallel to the direction of inflow and the sample volume at the leaflet tips during diastole. From the mitral inflow profile, the E-wave peak velocity and E-wave deceleration time (EDT) were measured (13). The transmitral filling pattern was considered restrictive when EDT was abnormally abbreviated (EDT <140 ms), which is suggestive of elevated LV filling pressures (14). Pulsed-wave tissue Doppler imaging of the mitral annulus was obtained from the apical 4-chamber view, using a 2-mm sample volume placed in the lateral mitral valve annulus, and the ratio of early transmitral flow velocity to early diastolic mitral annular velocity (E/e′) was calculated. Previous studies have shown a good correlation between E/e′ and LV end-diastolic pressure, and a strong association with outcome in patients with AMI (15). Doppler recordings from at least 5 consecutive heartbeats were analyzed (10 in patients with atrial fibrillation). Ectopic and post-ectopic beats were disregarded. From apical 4- and 2-chamber projections, left atrial volume was measured in end-systole using the biplane area-length method and indexed for body surface area. The resulting volume index was dichotomized at 32 ml/m2 in statistical analyses (16).
Blood samples were drawn in the morning on the day OGTT was performed or on the subsequent weekday. Plasma N-terminal pro-B-type natriuretic peptide (NT-proBNP) levels were analyzed using an ELECSYS proBNP immunoassay (Roche Diagnostics GmbH, Mannheim, Germany). The NT-proBNP concentrations were logarithmically transformed in all statistical analyses.
Follow-up and end points
Survival status of all patients was obtained from the Danish Civil Registration System 12 months after inclusion of the last patient, yielding a follow-up period ranging from 12 to 34 months. The study had a primary end point of all-cause mortality and a secondary end point being a composite of death or hospital admission for congestive heart failure.
Analyses were conducted with STATA/MP 10.0 (StataCorp LP, College Station, Texas). Baseline characteristics are presented as median [interquartile range] and compared using Kruskal-Wallis equality-of-populations rank tests for continuous variables, whereas prevalence is presented as counts (percentage) and compared using Pearson chi-square tests for categorical variables. Trend analyses investigating the relationship between the degree of dysglycemia and LV function, were performed using linear regression for continuous variables, and logistic regression for dichotomous variables. For outcome, hazard ratios were estimated using Cox proportionate hazards regression. Covariates, all selected prior to the inspection of data, were age, gender, pre-existing heart failure, Killip class at admission (I vs. II to IV), LVEF, EDT <140 ms, E/e′, left atrial volume index >32 ml/m2, and NT-proBNP and glucometabolic status. Assumptions of proportionality and linearity of hazards were met for all variables. A 2-sided p < 0.05 was considered statistically significant in all analyses.
Prevalence of abnormal glucose metabolism
A previous diagnosis of DM was present in 30 (15%) patients. In the remaining 173 patients, OGTT was performed a median of 3 days (interquartile range: 2 to 4 days) following hospital admission. From the total cohort, 45 (22%) patients were classified as newly detected DM, 56 (28%) as IFG/IGT, and 72 (35%) as having normal glucose regulation. Thus, among patients without a previous diagnosis of diabetes, 58% presented with abnormal glucose regulation. Baseline characteristics are shown in Table 1. Patients with abnormal glucose metabolism were generally older and more frequently presented with clinical signs of heart failure (i.e., Killip class ≥II). In-hospital treatment, which was administered at the discretion of the treating physicians, is shown in Table 2.
Glucose metabolism and LV function
Echocardiography was performed a median of 2 days (interquartile range: 2 to 4 days) after admission. Figures 1 and 2⇓ show LV function and NT-proBNP levels according to glucometabolic state. After adjustment for age and gender, a statistically significant linear increase was observed with increasing degree of dysglycemia for all variables except left atrial volume index (Figs. 1 and 2), also after further adjustment for atrial fibrillation, treated hypertension, and hyperlipidemia (E/e′: p < 0.001; restrictive filling: p = 0.04; NT-proBNP: p < 0.001); LVEF: p = 0.03; LV end-diastolic volume: p = 0.01; LV end-systolic volume: p = 0.02). Similar linear trends were also seen when limiting analyses to patients with a first AMI (E/e′: p < 0.001; restrictive filling: p = 0.01; NT-proBNP: p < 0.001; LVEF: p = 0.04; LV end-diastolic volume: p = 0.13; LV end-systolic volume: p = 0.10).
Prognostic implications of abnormal glucose metabolism
During a median follow-up of 21 months (interquartile range: 16 to 28 months), 32 patients died. The secondary end point was met in 39 patients. Kaplan-Meier event-free survival curves are shown in Figure 3. After adjusting for the pre-specified potential confounders, a gradual increase in hazard was observed with increasing degree of dysglycemia (Table 3). Although IFG/IGT was not an independent predictor of outcome, an apparently linear increase in hazard ratios was observed across the spectrum of dysglycemia (Fig. 4).
In a population with AMI, we demonstrated a linear relationship between LV function and abnormal glucose metabolism from impaired glucose tolerance to overt diabetes. This association could represent an important pathophysiological link between dysglycemia and cardiovascular mortality. However, even after adjustment for LV systolic and diastolic function, abnormal glucose metabolism was still associated with an increased risk of death and hospitalization for congestive heart failure. This implies that the relationship between glucose metabolism and prognosis extends beyond differences in LV function.
Recently, several studies have reported a prevalence of newly detected DM ranging from 22% to 31%, and IGT from 32% to 39% in patients with AMI (8,9,17,18) and stable coronary artery disease (17). This is comparable to our findings of 26% and 32%, respectively, and substantially higher than the prevalence in a recently published large community-based study (19). This emphasizes the significance of dysglycemia as a risk factor for development of cardiovascular disease.
Previous studies have shown that known DM is associated with impaired LV function after AMI. In the MILIS (Multiple Center Investigation of the Limit of Infarction) study, LVEF was comparable in the acute phase of AMI, but by 10 days, it was significantly lower among patients with known DM compared with nondiabetic patients (10), indicating that dysglycemia has an unfavorable influence on early LV remodeling. Our data extend this observation to markers of LV filling and furthermore suggest a linear relationship with dysglycemia that continues below the diagnostic threshold of DM. This is in accordance with a recent study by Henareh et al. (20), in which a correlation between post-OGTT glucose levels and LV myocardial performance index and e′-velocities was reported. This could indicate that dysglycemia may, over time, have deleterious effects on the heart, rendering it more vulnerable to acute myocardial ischemia, possibly through alterations in myocardial structure and microcirculation (21).
The role of glucose metabolism in LV dysfunction is not limited to patients with AMI. For decades, DM has been recognized as an independent risk factor for the development of heart failure (22), and a relationship between fasting glucose and the risk of hospitalization for congestive heart failure has been demonstrated among patients at high cardiovascular risk, with or without DM (23). Several pathophysiological mechanisms have been proposed, including myocardial hypertrophy and fibrosis accelerated by hyperinsulinemia, deleterious effects of advanced glycosylation end products, as well as a metabolic shift toward increased myocardial free fatty acid consumption as a consequence of insulin resistance. Any of these hypotheses could similarly be responsible for the association between dysglycemia and prognosis observed in our study.
Acute physiological stress, such as AMI, is associated with increased release of catecholamines and cortisol, impairing insulin sensitivity and thus inducing hyperglycemia. It is therefore controversial whether a reliable estimate of glucometabolic status can be made as early as 3 days after AMI. On the other hand, a similar distribution of abnormal glucose metabolism has been observed in a population with stable coronary heart disease (17). In another study of patients with AMI, the prevalence of abnormal glucose metabolism remained relatively unchanged when an OGTT was repeated 3 months after discharge (18). Thus, there is substantial evidence to suggest that the prognostic implications of abnormal glucose metabolism detected using an OGTT at least 3 days following AMI is associated with pre-existing perturbations in glucose regulation, rather than simply being an epiphenomenon of infarct size.
Neither patients nor their treating physicians were blinded to glucometabolic status at baseline. This may have affected subsequent patient management in a way that cannot easily be adjusted for in statistical analyses.
We included patients consecutively with few exclusion criteria, resulting in a very heterogeneous population. Therefore, particularly in patients with prior myocardial infarction, we cannot distinguish the magnitude of pre-existing LV dysfunction from the direct impact of acute ischemia. However, we found similar, associations between LV function and dysglycemia when limiting analyses to patients with a first AMI and after adjusting for atrial fibrillation, hypertension, and hyperlipidemia.
In addition, it should be stressed that the association between dysglycemia and outcome for patients in the IFG/IGT group did not reach statistical significance. However, when comparing the hazard ratios observed for these patients with those of the other glucometabolic groups, the association between the degrees of dysglycemia was strikingly linear (Fig. 4), which suggests that the absence of statistical significance in these patients is related to insufficient statistical power introduced by analyzing glucose metabolism as a categorical variable. This was however necessary to also include patients with known DM, where an oral OGTT for safety reasons could not performed. The number of covariates used in survival analyses was selected before the collection of data, and because we observed slightly lower event rates than anticipated, these models may also be somewhat overfitted.
The present study demonstrates that although abnormal glucose regulation is associated with impairment of LV function in the early phase of AMI, this relationship does not alone explain the excess mortality in patients with newly detected or known DM. A multidisciplinary approach to the individual patient, including assessment of both LV function and glucose metabolism, may improve risk stratification and outcome after AMI.
Supported by grants from the Danish Heart Foundation, Copenhagen, Denmark.
- Abbreviations and Acronyms
- acute myocardial infarction
- diabetes mellitus
- E-wave deceleration time
- early transmitral flow velocity to early diastolic mitral annular velocity ratio
- impaired fasting glycemia
- impaired glucose tolerance
- left ventricle
- left ventricular ejection fraction
- N-terminal pro-B-type natriuretic peptide
- oral glucose tolerance test
- Received July 7, 2008.
- Revision received March 12, 2009.
- Accepted March 20, 2009.
- American College of Cardiology Foundation
- Timmer J.R.,
- Ottervanger J.P.,
- Thomas K.,
- et al.
- Suleiman M.,
- Hammerman H.,
- Boulos M.,
- et al.
- O'Sullivan J.J.,
- Conroy R.M.,
- Robinson K.,
- Hickey N.,
- Mulcahy R.
- Bartnik M.,
- Ryden L.,
- Malmberg K.,
- et al.
- Ishihara M.,
- Inoue I.,
- Kawagoe T.,
- et al.
- Stone P.H.,
- Muller J.E.,
- Hartwell T.,
- et al.,
- MILIS Study Group
- WHO consultation,
- Alberti K.,
- Zimmet P.
- Lang R.M.,
- Bierig M.,
- Devereux R.B.,
- et al.,
- Chamber Quantification Writing Group; American Society of Echocardiography's Guidelines and Standards Committee; European Association of Echocardiography
- Nishimura R.A.,
- Tajik A.J.
- Moller J.E.,
- Pellikka P.A.,
- Hillis G.S.,
- Oh J.K.
- Hillis G.S.,
- Moller J.E.,
- Pellikka P.A.,
- et al.
- Moller J.E.,
- Hillis G.S.,
- Oh J.K.,
- et al.
- Bartnik M.,
- Ryden L.,
- Ferrari R.,
- et al.
- Barr E.L.,
- Zimmet P.Z.,
- Welborn T.A.,
- et al.
- Galderisi M.
- Held C.,
- Gerstein H.C.,
- Yusuf S.,
- et al.