Objective: This study evaluated the diagnostic performance of four point-of-care (POC) cardiac troponin I (cTnI) assays compared to a central laboratory cTnI assay for detecting myocardial injury and diagnosing acute myocardial infarction (AMI).
Design and methods: Plasma obtained at admission, 3 h, and 6 h post-admission in 169 patients presenting with symptoms suggestive of acute coronary syndrome (ACS) was studied. cTnI concentrations were measured on the Instrumentation Laboratory prototype GEM Immuno, Radiometer AQT90, Mitsubishi PATHFAST, Abbott i-STAT and the Ortho-Clinical Diagnostic Vitros assays. MI was determined based on 99th percentiles according to Universal MI guidelines.
Results: For ruling in MI at presentation (0 h), the GEM Immuno (sensitivity 63%, specificity 85%) and PATHFAST (sensitivity 53%, specificity 86%) were comparable to the OCD (sensitivity 68%, specificity 81%), and significantly better (p<0.05) than the AQT90 (sensitivity 26%, specificity 93%) and i-STAT (sensitivity 32%, specificity 92%). cTnI concentrations and serial rising patterns after MI differed by each assay. Negative predictive values were >90% and ROC AUCs were >0.90 after 6h for all assays. Detection of myocardial injury in non-ischemic pathologies accounted for lower than 100% specificity for MI.
Conclusion: cTnI is a sensitive biomarker for detection of myocardial injury. The analytical variability that exists between POC cTnI assays demonstrates substantial diagnostic differences for ruling in and ruling out MI in patients presenting with symptoms suggestive of ACS.
Keywords: Acute coronary syndrome; Acute myocardial infarction; Cardiac troponin; Diagnostic accuracy; Point of care testing.