Abstract Cardiac troponins (cTn) are the most sensitive and specific biomarkers of myocardial injury. The rise and/or fall of cardiac troponins above the 99th percentile of the upper reference limit (URL) is required in the assessment of acute myocardial infarction. In our institution, high sensitivity troponin I (hsTnI) is performed on four Access 2 Immunoassay Systems and three DXI 800 Access Immunoassay Systems. As a regional medical center, we accept patients from outside hospitals where contemporary cardiac troponin I (cTnI) were performed. The results from different assays and platforms may disagree and lead to incorrect explanation of the results as false rise and/or fall of cTn. To harmonize results among these assays, we analyzed the variation between the two assays of cTnI and hsTnI, and the concordance of same assay hsTnI on different instruments using patient specimens. To investigate the between-method variance, 424 patient specimens with troponin levels ranging from 0-23370 ng/L were used to measure both cTnI and hsTnI on DXI 800 instruments. To investigate the within-method agreement, 43 patient specimens with troponin levels ranging from 2-114394 ng/L were used to measure hsTnI on Access 2 and DXI instruments. The between-method comparison of cTnI and hsTnI on DXI showed a slope of 1.003 with intercept of -43.31 and correlation coefficient (CC) of 0.9590. The mean bias was -5.1%, which was within the analytical imprecision range determined in our laboratory (2.2%-10.7%). However, comparison of cTnI and hsTnI in patients with troponin <20ng/L (n=103), the 99th percentile of the URL for males, showed the slope was 1.325 with intercept of -0.97 and CC of 0.7462. The mean bias was 1.91 (21.53%), which was much larger than the analytical imprecision. This is consistent with the fact that cTnI results are not reliable at the lower level due to the imprecision problem. Comparison of cTnI and hsTnI in patients with troponin >20ng/L (n=321) showed the slope was 1.006 with intercept of –60.09 and CC of 0.9576. The mean bias was -54.30 (-5.23%), which was within the analytical imprecision range. The within-method comparison of hsTnI on Access 2 and DXI 800 showed the slope was 1.138 with intercept of -633.0 and CC of 0.9888. The mean bias was 2126.4 (10.6%), within the analytical imprecision range. These data indicate that cTnI results above the 99th percentile URL are comparable with hsTnI results if both are tested on DXI 800, and hsTnI results tested on Access 2 and DXI 800 are comparable. Thus, to use the cardiac troponins assay properly, we suggest the laboratory report cardiac troponins with the assay and instrument names to reduce confusion during results comparison.