© The European Society of Cardiology 2014.BACKGROUND: Myonecrosis provoked by illness unrelated to unstable coronary plaque is common, but uncertainty about a cause-effect relationship with future events challenges the appropriateness of initiating therapies known to be effective in cardiac conditions. We examined the causal relationship between troponin elevation in non-coronary diagnoses and late cardiac events using the Bradford Hills criteria for causality.METHODS AND RESULTS: Patients presenting acutely to South Australian public hospitals receiving at least one troponin between September 2011-September 2012 were included. Diagnoses were classified as coronary, non-coronary cardiac and non-cardiac using the International Classification of Diseases, version 10 Australian Modified, codes. The relationship between peak in-hospital troponin, using a high-sensitivity troponin T assay and adjudicated cardiac and non-cardiac mortality, and subsequent myocardial infarction (MI) was assessed using competing-risk flexible parametric survival models. Troponin results were available for 38,161 patients of whom, 12,645 (33.6%), 3237 (8.5%), and 22,079 (57.9%) patients were discharged with coronary, non-coronary cardiac and non-cardiac diagnoses, respectively. Troponin >14 ng/l was observed in 43.6%. The relationship between troponin and cardiac mortality was stronger among the non-coronary diagnosis group (troponin 1000 ng/l: coronary hazard ratio: 5.1 (95% confidence interval (CI) 4.0-6.6) vs non-coronary hazard ratio: 16.3 (95% CI 12.6-22.4)). The temporal hazard for cardiac death was marked within 30 days in both groups. Among non-coronary diagnoses, the hazard for recurrent MI was higher but did not vary with time.CONCLUSIONS: Consistency with causal criteria between secondary myonecrosis and cardiac events suggest the potential benefit for extending cardiac specific interventions to this population if supported in trials appropriately designed to address competing risks. Troponin elevation precipitated by non-coronary events is common and demonstrates an associations with late mortality that are analogous to spontaneous MI resulting from unstable coronary plaque. These observations help inform the design of randomized clinical trials exploring the benefits and risk of therapies with established benefits in other cardiac conditions. Such studies will need to appropriately account for competing risks in this population of patients.