Abstract

Background: The 4th Universal Definition of Myocardial Injury (UDMI) recognizes several categories of myocardial injury, including acute myocardial infarction (MI) which is further sub-classified into five types. However, data on these different types of myocardial injury and their risk factors is limited. Methods: In the MESA study of 6814 participants, 15905 clinical events were identified over the first 14 years, 4079 of which meet MESA criteria for physician adjudication for a possible cardiovascular event. Herein, we developed a standardized data format and a REDCap tool with an interactive robust logic algorithm to re-adjudicate all 4079 cases for the presence and classification of all 9 types of myocardial injury as defined by the 4 th UDMI. Adjudication process as shown in Figure 1 . The prevalence of myocardial injury types was evaluated using descriptive statistics, and adjudicator agreement was assessed using Cohen’s kappa (κ) statistics and percent agreement. Results: Out of 4079 events, adjudication is completed on 2282, of which 15% classified into subtypes of myocardial injury. Adjudication was achieved for 91% of the events in phase 1, 7% in phase 2, 2% in phase 3. The overall agreement between two adjudicators for the presence of myocardial injury was 91% (κ: 0.67), but the agreement for the specific subtype was 53% (κ: 0.38). The most common events were Type 1 MI (N= 114), followed by Type 2 MI (N= 95), and Acute non-ischemic myocardial injury (N= 85) ( Figure 2 ). Compared to the original MESA adjudication for the presence of MI, 97% (N= 72) of probable MI and 8% (N= 174) of no MI were reclassified into five and six types of myocardial injury events respectively. Conclusion: This study highlights the complexities in identifying subtypes of myocardial injury based on current definition. This study provides a novel dataset to explore diverse correlations with these myocardial injury subtypes.