Impact of Admission Glycemic Variability, Glucose, and Glycosylated Hemoglobin on Major Adverse Cardiac Events After Acute Myocardial Infarction

This was a single-center, prospective follow-up study. Consecutive patients admitted to the cardiology department of Beijing An Zhen Hospital of Capital Medical University for AMI between July 2010 and February 2011 were selected. The inclusion criteria were: 1) confirmed admission diagnosis of AMI, 2) admission glucose <16.7 mmol/L, and 3) without diabetic ketosis or nonketotic hyperosmolar coma. AMI was defined according to the universal definition of myocardial infarction. Myocardial infarction was defined as acute if the time elapsed between the first symptom and admission was ≤72 h. To enable long-term follow-up and repeated visits to our outpatient clinic, only patients under the age of 80 years and living within the hospital’s catchment area were eligible. The exclusion criteria were severe noncardiac disease with expected survival of <1 year and unwillingness to participate. A patient could only be included once. Data, including information on previous clinical history, cardiovascular risk factors, and medication, were collected in hospital. Type 2 diabetes mellitus (T2DM) was diagnosed according to the American Diabetes Association criteria or the use of insulin or glucose-lowering medication. Hypertension was defined as systolic blood pressure ≥140 mmHg and/or diastolic blood pressure ≥90 mmHg or treatment with oral antihypertension drugs. Hyperlipidemia was diagnosed according to the modified National Cholesterol Education Program-Adult Treatment Panel III. The estimated glomerular filtration rate (eGFR) value was calculated by Modification of Diet in Renal Disease equation (9). The Global Registry of Acute Coronary Events (GRACE) risk score were calculated as admission (10), which is recommended by the National Institute for Health and Clinical Excellence to assess risk in patients with acute coronary syndrome. Patients were categorized according to the mean amplitude of glycemic excursions (MAGE) level on admission (<3.9 and ≥3.9 mmol/L), based on reference values for continuous glucose monitoring in Chinese subjects (11), according to admission glucose levels (the upper tertile ≥8.61 mmol/L and the other two-thirds <8.61 mmol/L), and according to HbA_1c (<6.5 and ≥6.5%) (12). The study protocol was approved beforehand by the Medical Ethics Committee of Beijing An Zhen Hospital of Capital Medical University, and the procedures followed were in accordance with the institutional guidelines. The study complied with the Declaration of Helsinki, and informed consent was obtained from all patients.

All patients were equipped with a CGMS (Medtronic MiniMed; Medtronic) and monitored for 48 consecutive h after admission. A CGMS sensor was inserted into the subcutaneous abdominal fat tissue, calibrated according to the standard Medtronic MiniMed (Medtronic) operating guidelines. During CGMS monitoring, patients checked their blood glucose level with a self-monitoring of blood glucose device (Medisafe Mini; Terumo, Tokyo, Japan) at least four times per day. Then, they entered the self-monitoring of blood glucose data and time of each meal into the CGMS. After monitoring for 48 h, the recorded data were downloaded into a personal computer for analysis of the glucose profile and glucose excursion parameters with MiniMed Solutions software. After downloading the recorded data, the MAGE was analyzed from the intermediate 24 h of recording to avoid bias due to insertion and removal of the CGMS or insufficient stability of the monitoring system (13). Since measurable range of glucose by CGMS was mechanically limited from 2.2 to 22.2 mmol/L, the case showing the data out of this range was excluded from the study. The MAGE was calculated by measuring the arithmetic mean of the differences between consecutive peaks and nadirs, provided that the differences are >1 SD of the mean glucose value. Patients would maintain antihyperglycemic therapy as usual and avoid glucose infusion during the CGMS monitoring period. Otherwise, the patient would be excluded from the study.

Blood samples were collected on admission and after overnight fasting and stored at −70°C prior to analysis. Blood glucose, creatinine, total cholesterol, and triglyceride levels were measured by automatic biochemical analyzer (Hitachi 747; Hitachi, Tokyo, Japan). Serum concentration of HbA_1c was determined by high-performance liquid chromatographic method using an automatic HbA_1c analyzer (Tosoh HLC-723G7; Tosoh Corporation, Tokyo, Japan).

Patients were followed up prospectively for about 12 months. During the follow-up period, incidences of major adverse cardiac event (MACE) were registered, including new-onset myocardial infarction, acute heart failure, and cardiac death. All MACE data were adjudicated by an experienced cardiovascular physician blinded to clinical details and outcomes.

All statistical analyses were performed by using SPSS for Windows 13.0 (SPSS Inc, Chicago, IL). Data are presented as frequencies and percentages for categorical variables and mean ± SD for continuous variables, unless otherwise indicated. Differences between two groups were assessed by using the χ² and unpaired t tests. Correlation between continuous variables was determined by Spearman correlation coefficients. Admission MAGE was included as a continuous and as a categorized (<3.9 and ≥ 3.9 mmol/L) variable. Admission glucose and HbA_1c levels were also included as continuous and categorized (admission glucose: <8.61 and ≥8.61 mmol/L; HbA_1c: <6.5 and ≥6.5%) variables. Kaplan-Meier survival curve analysis was used to represent the proportional risk of MACE for the admission MAGE, glucose, and HbA_1c values, and the log-rank test was performed to assess differences between high levels and low levels of those variables. To ascertain the independent contribution to MACE, multivariate regression analysis was made. A value of P < 0.05 was considered statistically significant.

During the study period, 222 patients with complete data were included in the final analysis. Mean age was 62 ± 10 years, 62.6% were male, and 53.6% had diabetes. Participants were treated conservatively (7%), with percutaneous coronary intervention (77%), or with coronary artery bypass surgery (16%). MAGE level was <3.9 mmol/L in 143 patients (64.4%) and ≥3.9 mmol/L in 79 (35.6%). Admission glucose level was <8.61 mmol/L in 148 patients (66.7%) and ≥8.61 mmol/L in 74 (33.3%). HbA_1c was <6.5% in 132 (59.5%) and ≥6.5% in 90 (40.5%). The GRACE risk score ranged from 75 to 235 with a mean of 142 ± 34. Baseline characteristics of patient groups based on admission MAGE, glucose, and HbA_1c are shown in Table 1. The correlations of admission MAGE with admission glucose and HbA_1c were significant (Spearman r = 0.570 and r = 0.355, respectively; all P < 0.001).

At the end of 12-month follow-up, 10 patients had died (4.5%) for cardiac causes, 12 patients had new-onset myocardial infarction (5.4%), and 9 patients had acute heart failure (4.1%). AMI patients with high admission MAGE (≥3.9 mmol/L) levels had a significantly higher incidence of MACE compared with low MAGE level patients (24.1 vs 8.4%; P = 0.001). The higher rates of MACE were also observed in AMI patients with admission glucose level ≥8.61 mmol/L (21.6 vs. 10.1%; P = 0.020). No significant rates of adverse cardiovascular events were observed between patients with HbA_1c levels ≥6.5% and patients with HbA_1c levels <6.5% (18.7 vs. 10.7%; P = 0.091). AMI patients with a higher MAGE level had significantly higher cardiac mortality compared with AMI patients with MAGE levels <3.9 mmol/L (8.9 vs. 2.1%; P = 0.02). There were no significantly different rates of cardiac mortality, new-onset myocardial infarction, and acute heart failure in AMI patients between high and low levels of admission glucose or HbA_1c (Fig. 1). Kaplan-Meier survival curves for the two patient groups by admission MAGE, admission glucose, and HbA_1c are shown in Fig. 2.

To investigate the associations of admission MAGE, glucose, and HbA_1c with incidences of MACE with respect to baseline characteristics, we used multivariate analysis. Inclusion variables were age, sex, and all variables that were significantly different among MAGE, admission glucose, or HbA_1c categories (current smoking, diabetes history, previous coronary artery disease (CAD), eGFR, total cholesterol, triglyceride level, left ventricular ejection fraction (LVEF), oral antihyperglycemic agents, insulin, and diuretics). The independent predictors of MACE were age, previous CAD, LVEF, and MAGE (Table 2). Admission glucose ≥8.61 mmol/L (hazard ratio [HR] 2.256 [95% CI 0.957–5.271]; P = 0.077) or HbA_1c ≥6.5% (HR 1.584 [0.557–4.504]; P = 0.387) were not significantly associated with MACE. After adjustment for the GRACE score, MAGE was found to be still associated with incidences of MACE (HR 2.419 [1.273–9.100]; P = 0.017), but admission glucose (HR 2.025 [0.820–4.999]; P = 0.126) and HbA_1c (HR 1.508 [0.526–4.329]; P = 0.445) were not.

We investigated the association between admission GV, glucose, HbA_1c level, and 1-year MACE in patients with AMI. Our study demonstrated that elevated MAGE on admission was a strong and independent predictor of increased risk of MACEs in patients with AMI, but HbA_1c was not.

There were major differences in baseline characteristics according to admission MAGE. Patients with MAGE ≥3.9 mmol/L more often had diabetes or a history of CAD. They had lower LVEF and eGFR levels. There were also differences in baseline characteristics according to admission glucose or HbA_1c levels. Patients with higher glucose or HbA_1c levels were more found to have diabetes or CAD history, lower LVEF and eGFR, and higher total cholesterol and triglyceride levels. There are a higher proportion of patients with high MAGE, glucose, or HbA_1c level on diuretic therapy. Many diuretics, particularly thiazide diuretics that act on a particular part of the kidneys, have the potential to cause hyperglycemia (14), so we have adjusted the models by diuretics to prevent its affection. After adjusting for these different baseline characteristics, admission MAGE was significantly associated with poor outcomes, but HbA_1c was not.

Numerous prior studies have established that elevated admission glucose and HbA_1c are powerful predictors of survival and increased risk of cardiovascular complications in AMI patients both with and without diabetes (4,5,15). However, glycemic disorders are not solely limited to sustained hyperglycemia but can be extended to glycemic excursions that include both upward and downward acute glucose changes. Admission glucose is only a blood glucose value of point-in-time, which cannot reflect the status of acute glucose swings. HbA_1c is a convenient marker of long-term glycometabolic status. Patients with similar mean glucose or HbA_1c levels can have markedly different glycemic excursions (16). Fluctuations of glucose seem to have more deleterious effects than sustained hyperglycemia in the development of cardiovascular complications as acute glucose excursions activate the oxidative stress (17,18).

More and more evidences have been found that GV may be an important role in resolving potential vascular problems in diabetes. Some researchers suggested that postchallenge glucose excursion is independently related to carotid intima-media thickness and may contribute to the development of atherosclerosis in individuals with T2DM independent of other risk factors (6,19). In our previous study, we found that GV is an important contributing factor in the severity of CAD, which is independent of the average level of blood glucose (7). The Verona Diabetes study (20) reported that fasting GV is an independent predictor of mortality in T2DM patients. Some studies concluded that GV was a significant predictor of mortality in critically ill patients independently from mean glucose level and severity of illness (21–23). In the current study, patients with a higher MAGE level (≥3.9 mmol/L) have higher GRACE risk scores. After 1-year follow-up, a significantly higher incidence of MACEs and cardiac mortality was found in those patients. The results indicate that high glucose fluctuations may be associated with the risk of future adverse cardiovascular events in patients with AMI. Multivariate analysis disclosed that in the AMI population, GV (i.e., MAGE ≥3.9 mmol/L) was an independent predictor of MACEs, even after adjusting for GRACE risk score.

Although stress-induced hyperglycemia can partly explain the relation between admission GV and outcomes, glucose excursion itself can also be harmful. Ceriello et al. (17) reported that intermittent hyperglycemia induced a higher degree of apoptosis in endothelial cells than chronic hyperglycemia. Quagliaro et al. (24) showed that the apoptosis of endothelial cells exposed to intermittent high glucose may be related to a reactive oxygen species overproduction, through protein kinase C–dependent activation of NADPH oxidase. Glucose excursion may also be an important mediator in inflammatory responses (25). In vitro studies indicate that glucose fluctuations can activate the nuclear factor-κB and protein kinase C pathway, leading to a greater expression of the adhesion molecules and excess formation of advanced glycation end products than stable high glucose (26,27). Moreover, severe glycemic disorders may adversely affect sympathetic dysfunction, which is associated with mortality and morbidity of cardiovascular disease (28); the thrombotic properties of platelets are increased in a hyperglycemic environment, and this can result in additional cardiovascular complications (29).

Although admission glucose, HbA_1c, and GV may all be associated with adverse prognosis, our study shows that increased admission MAGE is more important. A significantly lower rate of event-free survival in patients with high admission glucose levels was observed in our Kaplan-Meier survival analysis. That result has confirmed the prognostic influence of admission glucose on outcomes of AMI patients as findings of previous trials. However, in the multivariate analysis, admission glucose is not found to be associated with the incidence of MACE, especially after adjustment for GRACE risk score. Acute hyperglycemia is an important contributing factor in glycemic excursions on admission. So there may be some interaction between admission glucose and GV. As a blood glucose value of point-in-time, admission glucose cannot accurately show dynamic changes of blood glucose as GV and can be influenced by many random factors. That may be why admission glucose does not show the independent association with incidence of MACE when analyzed with MAGE together in our study. Increased HbA_1c represents long-term glucose regulation, whereas elevated admission glucose excursion is not only a symptom of glucose dysregulation, but also of stress and general poor health. Carmen Wong et al. (30) found cortisol level is correlated with acute hyperglycemia in patients with AMI. In our Kaplan-Meier survival-curve analysis, only a trend toward lower event-free survival rate in AMI patients with HbA_1c ≥6.5% (P = 0.055) was observed. The less clear association between HbA_1c and MACE could be due to a limited number of patients with a relatively short follow-up in present study. There is a clear association to be found between HbA_1c and long-term outcome in AMI patients after 3.3 years of follow-up (31). So HbA_1c may have limited predictive power on short-term prognosis in patients with AMI, but its association with long-term prognosis may be stronger.

There is still an extensive debate about GV as a risk factor for cardiovascular complications independent of HbA_1c (32). Siegelaar et al. (33) performed reanalysis of the data of the HEART2D study, which shows that targeting postprandial glucose–decreased intraday GV would not be beneficial in reducing adverse cardiovascular events in AMI patients. However, the HEART2D study was not designed to determine the impact of GV on the risk of MACEs, and the MAGE and SD levels were not found significantly different between two contrasting groups in the study. In addition, the method of calculating GV from self-measured blood glucose profiles may be not very accurate. Overall, more well-designed studies are needed to investigate whether GV will play an important role in the prognosis of AMI.

The sample size was relatively small, so that some subgroup comparisons may have lacked power to detect significant differences for selected variables. Due to the lack of microvascular complications data, we did not include those risk factors in analysis. Although we had maintained the patients’ antihyperglycemic therapy as usual and avoided glucose infusion during the CGMS monitoring period, some factors, such as different diets, physical and emotional stress, etc., which may affect levels of admission glucose fluctuations, could not all be prevented. In addition, tests to detect diabetes were not routinely done, so some cases of diabetes may have been missed. However, if the observed relation between glucose excursion and MACE was due to undiagnosed diabetes, one would have expected a more distinct association between HbA_1c and outcomes.

Elevated admission MAGE might predict risk of MACE in our AMI patients, but HbA_1c did not. Acute glucose excursion in glycemic metabolism would seem to be of greater importance than admission glucose and long-term derangements of glucose metabolism in predicting 1-year outcomes in AMI. The results of this study further support the view that glucose variability should be one of the targets of treatment for the glycemic disorders encountered in AMI patients.

This work was supported by a key grant from the Beijing Health Special Foundation (JING 09-08).

No potential conflicts of interest relevant to this article were reported.

G.S. designed the study, collected data, analyzed data statistically, and drafted the manuscript. S.-h.M. designed the study, contributed to the analysis with suggestions and advice, and critically reviewed and edited the manuscript. H.T. designed the study, contributed to the analysis with suggestions and advice, and critically reviewed and edited the manuscript. Z.L. collected data. H.-X.Y. did continuous glucose monitoring. H.Z. collected data. Y.Z. critically reviewed and edited the manuscript. L.T. contributed to the analysis with suggestions and advice and critically reviewed and edited the manuscript. G.S. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

The authors appreciate the professional technical assistance from the laboratory technicians at the Clinical Laboratory Center. The authors thank the volunteer patients for their participation, and the study nurses Jin Wu and Suyun Liu for their skills and devotion to the patient care.