STEMI: ST-segment elevation myocardial infarction; T2DM: Type 2 Diabetes mellitus; MACCE: Main adverse cardiovascular and cerebrovascular events; ROC: Receiver operating characteristic; LVEF: Left ventricular ejection fraction; PCI: Percutaneous coronary intervention; TIMI: Thrombolysis in myocardial infarction.

Research showed that no matter patients have diabetes or not, acute myocardial infarction patients often have stress hyperglycemia in admission, which can strengthen ischemic myocardial injury [1] and closely associated with mortality and complications in hospital [2,3]. Acute myocardial infarction patients with diabetes always have poor prognosis and increased mortality accordingly. Our research aimed to research the relationship between stress hyperglycemia, in-hospital death, in-hospital adverse cardiovascular events and one-year MACCE in patients with ST-segment elevation myocardial infarction, and assess the predictive value of stress hyperglycemia in hospital and in the long term.

General information

We did a single center prospective cohort study, enrolling 456 STsegment elevation myocardial infarction patients in prospective randomized way from January 2015 to January 2016 at Zhongda Hospital affiliated to Southeast University, including 347 males and 109 females. The inclusion criteria were as followed: 1) willing to sign informed consent; 2) Aged more than 18; 3) Meet the diagnostic criteria of “2017 ESC guide: management of ST-segment elevation acute myocardial infarction patients” [4], the exclusion criteria were: 1) patients with severe cardiac insufficiency (require large vasoactive drugs to maintain blood pressure), severe valvular heart disease and hemodynamic instability; 2) severe liver dysfunction, malignancy, severe infectious disease and stroke patients; 3) information such as random admission blood glucose, follow-up, etc., was not complete.

The study was approved by the Ethics Committee of Zhongda Hospital affiliated to Southeast University (Nanjing, Jiangsu, China). Written informed consent was obtained from the patients or a legal representative.

Groups

There is no unified definition of stress hyperglycemia, the American Diabetes Association took fasting glucose >6.9 mmol/L or random blood glucose >11.1 mmol/L without diabetes history as the criteria of hyperglycemia in 2009 [5]. (2) Based on this, we recorded the baseline data, then used propensity matching (a ratio of 1:2) to balance the confounding variables and rule out the bias, we divided patients into two groups: hyperglycemia group (random admission glucose >11.1 mmol/L) 71 cases and non-hyperglycemia group (admission random glucose ≤ 11.1 mmol/L) 142 cases.

Data collection and follow-up

We recorded general information, past history, admission random glucose, troponin, blood routine, biochemical index, LVEF, platelet aggregation inhibitor, anticoagulant, beta blocker, statin, coronary angiography situation, the postoperative TIMI flow grade, hospital stay, hospital cost, in-hospital death, malignant arrhythmia, temporary pacemaker implantation, cardiogenic shock, congestive heart failure, one-year follow-up MACCE were implemented by telephone and outpatient clinic. MACCE mainly includes: all-cause death, target vessel revascularization, non-fatal myocardial infarction, unstable angina and heart failure that need hospital treatment, cerebral apoplexy, or transient cerebral hemorrhage.

Statistical analysis

SPSS 22.0 statistical software was used for data analysis. We used propensity matching (a ratio of 1:2) to balance the confounding variables and rule out the bias. The numerical data were expressed as X ± S, and compared using independent sample t test, non-normally distributed numerical data were expressed as median and 25-75 interquartile range, and compared using rank-sum test. Classification data were expressed as number and percentage and compared using X² test. ROC curve was to assess the correlation between hyperglycemia and in-hospital death. The predictive value of random blood glucose for prognosis of STEMI patients is evaluated by Kaplan-Meier curve and Cox regression. P<0.05 was considered statistically significant.

Comparison of general data and past history between hyperglycemia group and non-hyperglycemia group

There were no statistical differences in gender, age, BMI, systolic blood pressure, diastolic blood pressure, heart rate, Killip classification, smoking, hypertension, diabetes, chronic kidney disfunction, cancer, previous myocardial infarction history, previous stent implantation history and atrial fibrillation history (P ≥ 0.05) (Table 1).

Table 1: Comparison of general data and past history between hyperglycemia group and non-hyperglycemia group.

Comparison of auxiliary inspection index between hyperglycemia group and non-hyperglycemia group

White blood cell count and random blood sugar were statistically significant in the two groups (P=0.043, P<0.001, respectively), none of troponin, red blood cell count, hemoglobin, platelet count, potassium ion, albumin, alanine aminotransferase, aspartate transaminase, lactate dehydrogenase, creatine kinase, creatinine, eGFR, uric acid, serum cystatin C, triglycerides, cholesterol, HDL-C, LDL-C, left ventricular ejection fraction was statistically significantly different (P ≥ 0.05) (Table 2).

Table 2: Comparison of auxiliary inspection index between hyperglycemia group and non-hyperglycemia group.

Comparison of medications, coronary angiography and hospitalization situation between hyperglycemia group and non-hyperglycemia

The number of coronary lesion vessels in hyperglycemia group is higher than the control group (P=0.038), hospital stay and hospital cost are also higher (P=0.036, 0.028, respectively). There were no significant difference between the two groups in aspirin, clopidogrel, low molecular heparin, statin, the proportion of thrombus aspiration, coronary stent implantation, and TIMI flow grades after coronary angiography (P ≥ 0.05) (Table 3).

Table 3: Comparison of medications, coronary angiography and hospitalization situation between hyperglycemia group and non-hyperglycemia group.

Comparison of major adverse cardiovascular events between hyperglycemia group and non-hyperglycemia group

In hyperglycemia group, the incidence of hospital mortality, malignant arrhythmia and cardiogenic shock incidence were higher than control group (P=0.003, 0.023, 0.030, respectively), the incidence of temporary pacemaker implantation and congestive heart failure between the two groups had no significant difference (P ≥ 0.05) (Table 4).

Table 4: Comparison of major adverse cardiovascular events between hyperglycemia groups and non-hyperglycemia group.

The ROC curve of admission random blood glucose predicting in-hospital deaths for non-diabetic patients and diabetic patients

In non-diabetic patients, 8.1% of patients with hyperglycemia died in hospital, the ROC curve of admission random blood glucose predicting in-hospital deaths showed the area under curve and 95% CI were 0.627, 0.438 ~ 0.837 (P=0.043), in patients with diabetes, 9.6% of hyperglycemia patients died in hospital, the AUC and 95% CI were 0.786, 0.586 ~ 0.913 (P=0.013) , the cut-off value is 17.67 mmol/L, the sensitivity and specificity were 80.0% and 76.4% accordingly (Figures 1 and 2).

Figure 1: The ROC curve of admission random blood glucose predicting in-hospital deaths for non-diabetic patients.

Figure 2: The ROC curve of admission random blood glucose predicting in-hospital deaths for diabetic patients.

Kaplan-Meier curve and Cox regression model

The incidence of one-year follow-up MACCE in hyperglycemia group was 41.7%, compared with 27.8% in the control group (P=0.018). Gender, age, diabetes, hypertension, past history of myocardial infarction, hyperglycemia, left ventricular ejection fraction were included in Cox regression model. Stress hyperglycemia is an independent predictor of MACCE (OR=4.398, 95% CI=2.869 ~ 7.483, P<0.001) (Table 5 and Figure 3).

Table 5: Cox regression model for one-year follow-up MACCE.

Figure 3: The incidence of one-year follow-up MACCE in hyperglycemia group and non-hyperglycemia group.

AMI patients with diabetes have higher mortality than those with normal blood glucose [6], several large sample cohort studies showed a significant increased cardiovascular risk in patients with impaired glucose tolerance [7,8]. A 20-year follow-up in Whitehall Study, Paris Prospective Study and Helsinki Policemen Study that enrolled over 20,000 non-diabetic patients showed that hyperglycemia significantly increased the risk of cardiovascular events, all-cause mortality and cancer incidence [9]. Oswald et al. [10,11] found that in acute myocardial infarction patients with or without diabetes, admission random blood glucose levels were closely associated with mortality [10,11]. In patients with acute myocardial infarction, the incidence of stress hyperglycemia is high, and the elevated level of admission random blood glucose is often associated with poor long-term prognosis [12].

Stress hyperglycemia is transient hyperglycemia occurred during disease, the American diabetes association (ADA) defined stress hyperglycemia in 2009 as, for patients without diabetes, fasting glucose >6.9 mmol/L or random blood glucose >11.1 mmol/L [5]. Norhammar A et al. [13] did a research published in Lancet, claiming that in the early stages of acute myocardial infarction, increased blood glucose can be a marker for early risk assessment [13]. Stress hyperglycemia can be seen as a compensatory response to ischemic myocardial cells when acute myocardial infarction occurred, the rise in blood sugar can increase the availability of glucose, reduce the apoptosis, and meanwhile, metabolic disorders can lead to extensive myocardial injury. Hyperglycemia has direct adverse effects on ischemic myocardium, including oxidative stress, inflammation, apoptosis, endothelial dysfunction, hypercoagulability, and platelet aggregation [1]. Stubbs PJ found serum insulin is weakly correlated with glucose levels after AMI [14], Oliver EF put forward hyperglycemia is related with the hormone that raises blood sugar rather than decreased insulin level [15].

Lonborg et al. [16] found STEMI patients with hyperglycemia have significantly greater myocardial infarction size than control group, stress hyperglycemia can be used as a marker to assess myocardial injury severity [16]. A multi-center, prospective cohort study [17] enrolled 816 AMI patients with cardiogenic shock confirmed admission blood glucose is an independent risk factor for mortality in non-diabetic patients, Deckers JW [18] included 11324 patients from 1985 to 2005 with AMI, the conclusion is that the incidence of hyperglycemia and that in non-diabetic patients are increasing, the 30- day mortality in stress hyperglycemia group was 3.6 times compared with control group (95% CI=2.9 ~ 2.9).

In this study, 18.9% of patients with STEMI are consolidated with hyperglycemia on admission, statistically significant differences can be seen in gender, Killip classification, smoking, diabetes, history of previous myocardial infarction, white blood cell count, albumin, alanine transaminase, triglycerides, admission random blood glucose, LVEF, hospital days, hospital cost, in-hospital deaths, malignant arrhythmia, temporary pacemaker implantation, cardiogenic shock between hyperglycemia group and non-hyperglycemia group. Binary analysis showed that stress hyperglycemia is an independent risk factor for in-hospital death. There were 86 patients with hyperglycemia, of which 65 were diabetic. In non-diabetic patients, the ROC of admission random blood glucose predicting in-hospital deaths showed the area under curve is 0.670, while the area is 0.746 in patients with diabetes, which has higher predictive value. One-year follow-up MACCE showed that the incidence of main adverse cardiovascular and cerebrovascular events was significantly higher in hyperglycemia group, stress hyperglycemia is included in Cox regression model for one-year follow-up MACCE, in other word, hyperglycemia is an independent discriminator of MACCE.

Stress hyperglycemia is closely associated with poor outcome in patients with acute myocardial infarction, therefore, we need to pay high attention to the admission random blood glucose to help evaluate both in-hospital and long-term prognosis in STEMI patients. GONG SU et al. [19] found that elevated blood glucose variability was more predictive for MACCE in AMI patients. Teraguchi I also put forward the blood glucose changes from the peak to valley may be a potential therapeutic targets [20]. More clinical researches need to be carried out to confirm intervention measures and their curative effects.

A large amount of STEMI patients have hyperglycemia on admission, our research showed that hyperglycemia is closely related to poor prognosis both in hospital and in the long term in STEMI patients. Admission blood glucose can be regarded as a good discriminator for predicting in-hospital and long-term prognosis, and we should pay more attention to it in clinical practice.

This study was supported by the National Natural Science Foundation of China (Grant no. 81600227), Trial registration: ChiCTR1800014763.

The authors have declared that no competing interest exists.

Yuhan Qin and Gaoliang Yan contributed equally to this work.