Design of a new mortality indicator in acute coronary syndrome on admission to the Intensive Care Unit.

OBJECTIVE: To design a mortality indicator in acute coronary syndrome (ACS) in the intensive care unit (ICU). DESIGN: A multicenter, observational descriptive study was carried out. PARTICIPANTS: Patients with ACS admitted to the ICUs included in the ARIAM-SEMICYUC registry between January 2013 and April 2019. INTERVENTIONS: None. MAIN VARIABLES OF INTEREST: Demographic parameters, time of access to the healthcare system, and clinical condition. Revascularization therapy, drugs and mortality were analyzed. Cox regression analysis was performed, followed by the design of a neural network. A receiver operating characteristic curve (ROC) was plotted to calculate the power of the new score. Lastly, the clinical utility or relevance of the ARIAM indicator (ARIAM's) was assessed using a Fagan test. RESULTS: A total of 17,258 patients were included in the study, with a mortality rate of 3.5% (n = 605) at discharge from the ICU. The variables showing statistical significance (P < .001) were entered into the supervised predictive model, an artificial neural network. The new ARIAM's yielded a mean of 0.0257 (95%CI: 0.0245-0.0267) in patients discharged from the ICU versus 0.27085 (95%CI: 0.2533-0.2886) in those who died (P < .001). The area under the ROC curve of the model was 0.918 (95%CI: 0.907-0.930). Based on the Fagan test, the ARIAM's showed the mortality risk to be 19% (95%CI: 18%-20%) when positive and 0.9% (95%CI: 0.8%-1.01%) when negative. CONCLUSIONS: A new mortality indicator for ACS in the ICU can be established that is more accurate and reproducible, and periodically updated.