Troponin measurements, myocardial infarction diagnoses and outcomes. An analysis of linked data from New South Wales, Australia.

Abstract

Cardiac troponin assays are very widely requested tests, particularly in emergency departments. Thus, many seriously ill patients who may not have heart disease as their primary discharge diagnosis have undergone troponin testing during hospitalisation. To determine associations between cardiac troponin levels and mortality, including from cardiovascular and non-cardiovascular causes, among patients hospitalised in New South Wales, Australia over a 2-year period. Over a 2-year period (2006-2008), 172 753 hospitalised patients had a quantitative or qualitative troponin assay performed in New South Wales (Australia). The associations were examined, using data linked to late outcomes, between elevations in levels of troponins T or I and 1-year mortality. Mortality was determined for International Statistical Classification of Diseases 10th Revision diagnostic groups. Of 172 753 patients undergoing troponin testing, 44 357 (25%) had a cardiovascular diagnosis of whom (7% had myocardial infarction) 47 827 (28%) had a probable ischaemic heart disease diagnosis if the 23 873 (14%) of patients coded with 'chest pain' were included. In patients with a cardiovascular diagnosis and elevated troponin 3060 (1.8%) died in 12 months in comparison to 6262 (3.6%) in those with a non-cardiovascular disease diagnoses and elevated troponin. The 1-year mortality hazard with respect to a troponin elevation was 2.5 (95% confidence interval 2.3-2.7) and 2.0 (95% confidence interval 1.99-2.01) for those with a cardiovascular and non-cardiovascular diagnosis respectively. In a very large state-wide hospitalised patient cohort, among patients with elevated troponin levels and non-cardiac diagnoses, mortality was higher than in those with cardiovascular diagnoses, including MI.