OBJECTIVE
The primary objective of the study was to investigate the prognostic factors of AMI in the ICU. The additional objectives were to analyze the mortality of AMI in the ICU of a tertiary teaching hospital over a period of 17 years, discriminate AMI as a primary or secondary cause of ICU admission, and compare arterial and venous AMI.
MATERIAL AND METHODS
We conducted a retrospective, observational, non-interventional, single-center study including a cohort of patients who stayed in the ICU of the Cliniques Universitaires St-Luc (Brussels, Belgium) between 2000 and 2017. The patients were identified using the RCM (résumé Clinique minimum) database and the ICD9-CM coding system. The patients had to fulfil two criteria: ICU admission and a positive diagnosis of AMI. The diagnosis was based on at a least one of the following methods: abdominal computed tomography (CT) and/or surgery (laparoscopy or laparotomy) and/or arteriography and/or autopsy.
RESULTS
The 30-day mortality rate among the 228 included patients was 64.9%. Three models of multivariate logistic regression were developed to assess the prognostic criteria in AMI of arterial origin. In the first model, markers of cardiac dysfunction (OR = 7.523; 95% CI = 1.283-44.119; p=0.025), mechanical ventilation 24h before diagnosis of AMI (OR = 6.417; 95% CI = 1.481-27.794; p=0.013), arterial lactate 24h after diagnosis (OR = 1.888; 95%CI = 1.277-2.79; p=0.001), and anticoagulation (OR = 0.125; 95% CI = 0.016-0.954; p=0.045) were found to be associated with mortality. In the second model, anticoagulation was associated with a decrease in mortality (OR = 0.119; 95% CI = 0.018-0.786; p=0.027), and other factors were correlated with higher mortality, namely arterial delta lactate (OR = 1.299; 95% CI = 1.071-1.574; p=0.008), mechanical ventilation 24h before diagnosis (OR = 5.451; 95%CI = 1.439-20.656; p=0.013), cardiac dysfunction (OR = 4.72; 95% CI = 1.047-21.268; p=0.043), and increase of INR values (OR = 4.298; 95% CI = 1.17-15.787; p=0.028). The last model showed three factors associated with an increase or decrease in mortality: maximal dose of vasopressors (VPmax) administered to the patient (OR = 1.201; 95% CI = 1.085-1.329; p<0.001), arterial delta lactate (OR = 1.244; 95%CI = 1.048-1.477; p=0.012), and anticoagulation (OR = 0.19; 95% CI = 0.043-0.841; p=0.029). The risk of mortality (ROM) could be expressed by a new variable: ROM = 1.40 + 1.20 (VPmax) + 0.19 (anticoagulation) + 1.24 (arterial delta lactate); with an AUC of 0.872 (95% CI = 0.819-0.925). AMI was the primary cause of ICU admission in 97 patients (42.5%), while 131 patients (57.5%) developed AMI during their ICU stay. In our study, 14 patients (6.1%) presented venous AMI, with a 30-day mortality rate of 21%.
CONCLUSION
At 30 days, the mortality rate of the 228 patients identified from the database was 64.9%. Independent factors of poor prognosis in arterial AMI were identified: markers of cardiac dysfunction, need for high doses of vasopressors, mechanical ventilation 24h before diagnosis of AMI, increase of INR values, increased arterial lactate concentration 24h after diagnosis, and further increase of arterial lactate concentration 24h following the diagnosis. Anticoagulation was identified as a factor of good prognosis. Finally, a new score was proposed to predict the prognosis of AMI in the ICU.