Prediction of adverse events in patients with initially medically treated type A intramural hematoma

Background: Prior studies provided limited data regarding natural history of initially medically treated type A intramural hematoma (IMH). Objectives: To develop predictive models for adverse aorta-related events in patients with type A IMH. Methods: We performed a retrospective pooled analysis of individual patient data, including baseline clinical and CT characteristics. All patients enrolled were followed up for adverse aorta-related events, defined as a composite of aortic disease-related death and the presence of aortic complications that required aortic invasive treatment. Results: A total of 172 patients (52.9% men) were included, with a mean age of 61.1 +/- 11.2 years. During a median follow-up time of 770.5 (45.3-1695.8) days, 60 patients (34.9%) experienced adverse aorta-related events. In Cox regression model for predicting adverse aorta-related events, hypertension (HR= 3.78, p =.067), MAD (HR= 1.05, p=.018), presence of ULP (HR= 2.43, p=.002) and pericardial effusion (HR= 1.65, p=.061) were independently associated with adverse aorta-related events. A majority of the adverse aorta-related events (n = 46, 76.7%) occurred within acute and subacute phase (90 days) of IMH. In predictive model for 90 days aortic events, MAD=50.7mm(OR= 2.79, p=.006) and presence of ULP (OR= 3.20, p=.002) were independent predictors. C statistic of the predictive model were 0.71 (p < .001). Conclusions: Predictive models including baseline clinical and CT characteristics as predictors allow for accurate estimation of risk of adverse aorta-related events in patients with type A IMH. The proposed predictive models are helpful for risk estimates and decision making. (c) 2020 Elsevier B.V. All rights reserved.