Factors Associated With Hospital Mortality in Acute Myocardial Infarction

Abstract

Aim To determine clinical and laboratory parameters associated with in-hospital mortality in patients with acute myocardial infarction and to develop a multifactorial prognostic model of in-hospital mortality.Material and methods This was a study based on the 2019-2020 Registry of acute coronary syndrome of the Tyumen Cardiology Research Center, a branch of the Tomsk National Research Medical Center. The study included 477 patients with ST-segment elevation acute myocardial infarction (AMI), 617 patients with non-ST segment elevation AMI, and 26 patients with unspecified AMI. In-hospital mortality was 6.0 % (n=67). Clinical and laboratory parameters were assessed on the day of admission. The separation power of indicators associated with in-hospital mortality was determined using a ROC analysis. The data array of each quantitative parameter was converted into a binary variable according to the obtained cut-off thresholds, followed by creation of a multifactorial model for predicting in-hospital mortality using a stepwise analysis with backward inclusion (Wald). The null hypothesis was rejected at p<0.05.Results The multivariate model for prediction of in-hospital mortality included age (cut-off, 72 years), OR 3.0 (95 % CI: 1.5-5.6); modified shock index (cut-off threshold, 0.87), OR 1.5 (95 % CI: 1.1-2.0); creatine phosphokinase-MB (cut-off threshold, 32.8 U / L), OR 4.1 (95 % CI: 2.2-7.7); hemoglobin (121.5 g / l), OR 1.7 (95 % CI: 1.2-2.3); leukocytes (11.5×109 / l), OR 1.9 (95 % CI: 1.3-2.6); glomerular filtration rate (60.9 ml / min), OR 1.7 (95 % CI: 1.2-2.2); left ventricular ejection fraction (42.5 %), OR 4.1 (95 % CI: 2.0-8.3); and size of myocardial asynergy (32.5 %), OR 2.6 (95 % CI: 1.4-5.0).Conclusions Independent predictors of in-hospital mortality in AMI are age, modified shock index, creatine phosphokinase-MB, peripheral blood leukocyte count, hemoglobin concentration, left ventricular ejection fraction, size of myocardial asynergy, and glomerular filtration rate. The in-hospital mortality model had a high predictive potential: AUC 0.930 (95 % CI: 0.905-0.954; p <0.001) with a cutoff threshold of 0.15; sensitivity 0.851, and specificity 0.850.