Background Cardiac troponin is integral to the diagnosis of myocardial infarction, but is elevated in many patients without acute coronary syndrome. We evaluate the impact of a clinical assessment of pre-test probability on the positive predictive value (PPV) of cardiac troponin for a diagnosis of myocardial infarction.

Methods In 1,054 consecutive patients undergoing blood sampling in the Emergency Department, cardiac troponin I was measured in excess serum using a high-sensitivity assay. Patients were classified as type 1 or type 2 myocardial infarction, or myocardial injury, and the prevalence determined in the entire cohort and in those where testing was requested by the attending physician. We determine the predicted PPV for type 1 myocardial infarction across different pre-test probabilities.

Results Cardiac troponin was requested in 136 patients identifying 15 (11.3%) with type 1 myocardial infarction. Testing everyone identified 2 further patients with type 1 (17/1,054 [1.7%]), but 127 (12%) patients with type 2 myocardial infarction or myocardial injury. In patients selected for testing the PPV for type 1 myocardial infarction was 50.0% (95% confidence interval [CI] 32.8%-67.2%), falling to 13.6% (95% CI 8.0%-19.7%) when measured across the cohort. The PPV was heavily influenced by the pre-test probability (Figure 1). Irrespective of diagnosis, cardiac troponin was a independent predictor of death (hazard ratio 1.26 per 2-fold increase, 95% CI 1.06 to 1.49) (Figure 2).

• Download figure
• Open in new tab
• Download powerpoint

Abstract 112 Figure 1

• Download figure
• Open in new tab
• Download powerpoint

Abstract 112 Figure 2

Conclusions Cardiac troponin concentrations are elevated in one in eight patients in the Emergency Department. If testing is performed without an assessment of pre-test probability, interpretation of cardiac troponin will be challenging and may adversely impact on the diagnosis of acute myocardial infarction.