What level of hyperglycaemia on admission indicates a poor prognosis in patients with myocardial infarction treated invasively?Kardiol Pol 2012; 70(6):564-72KP
Stress hyperglycaemia on admission is a predictor of mortality in patients with acute myocardial infarction (MI).
To establish what level of hyperglycaemia on admission indicates a significantly poorer long-term prognosis in patients with MI treated invasively.
Glycaemia on admission was measured in patients with both ST-segment elevation MI (STEMI) and non-ST- -segment elevation MI (NSTEMI) treated with percutaneous coronary intervention (PCI). In-hospital and late mortality were evaluated during a 679.3 ± 202 day follow-up.
We enrolled 794 patients (564 men; 71%), mean age 63.8 ± 11.3 years. One per cent of the patients died during initial hospitalisation, and 10% during the two-year follow-up. The mean value of glycaemia in the whole population was 115 ± 36 mg/dL (6.32 ± 1.98 mmol/L). Admission glycaemia in patients who died in hospital was 194 ± 71 mg/dL (10.67 ± 3.91 mmol/L), while in the patients discharged home it was 114 ± 35 mg/dL (6.27 ± 1.93 mmol/L) (p 〈 0.0001). In terms of two-year mortality, the patients who died had also significantly higher glycaemia on admission (145 ± 48 mg/dL; 7.98 ± 2.64 mmol/L) vs 112 ± 31 mg/dL (6.16 ± 1.71 mmol/L, p 〈 0.0001). Apart from admission hyperglycaemia, we found the following risk factors of late mortality in univariate analysis: age, heart rate (HR), left ventricular ejection fraction (LVEF), glomerular filtration rate (GFR), creatinine level, number of significantly narrowed coronary vessels other than the infarct related artery (IRA), and unsuccessful PCI. In multivariate analysis, the following parameters correlated with death in the two-year follow-up: glycaemia on admission, age, HR, LVEF, GFR, creatinine level, total cholesterol, number of significantly narrowed coronary vessels other than the IRA, and unsuccessful PCI. Hyperglycaemia on admission was an independent risk factor of death even after adjustment for confounding variables such as age, sex and LVEF. We compared the areas under ROC curve for in-hospital mortality and the areas under ROC curve for late mortality according to glycaemia on admission. Both were significantly different from those of a random model (p 〈 0.001 and p 〈 0.001, respectively). A glycaemia value of 205 mg/dL (11.28 mmol/L) calculated from ROC curve had the highest sensitivity and specificity for late mortality. Apart from these findings, we observed a linear correlation between glycaemia and mortality.
The best cut-off value for stress hyperglycaemia determined by ROC curve in patients with acute MI treated invasively is 205 mg/dL (11.28 mmol/L). Patients with glucose levels 〉 205 mg/dL (11.28 mmol/L) on admission have significantly higher late mortality compared to those with glucose levels 〈 205 mg/dL (11.28 mmol/L). Our results suggest that hyperglycaemia is a reliable marker of poor outcome in acute MI patients with and without previously diagnosed diabetes mellitus. This level of glucose may be used in risk stratification in patients with acute MI.