Point-of-care measurement of activated clotting time for cardiac surgery as measured by the Hemochron signature elite and the Abbott i-STAT: Agreement, concordance, and clinical reliability

Abstract

Background: Since inadequate heparin anticoagulation and insufficient reversal can result in complications during cardiopulmonary bypass (CPB) surgery, heparin anticoagulation monitoring by point-of-care (POC) activated clotting time (ACT) measurements is essential for CPB initiation, maintainance, and anticoagulant reversal. However, concerns exist regarding reproducibility of ACT assays and comparability of devices. Methods: We evaluated the agreement of ACT assays using four parallel measurements performed on two commonly used devices each (i.e., two Hemochron Signature Elite (Hemochron) and two Abbott i-STAT (i-STAT) devices, respectively). Blood samples from 30 patients undergoing cardiac surgery on CPB were assayed at specified steps (baseline, after heparin administration, after protamine administration) with four parallel measurements (two of each device type) using commercial Kaolin activated assays provided by the respective manufactures. Measurements were compared between identical and different device types using linear regression, Bland-Altman analyses, and calculation of Cohen's kappa coefficient. Results: Parallel i-STAT ACTs demonstrated a good linear correlation (r = 0.985). Bias, as determined by Bland-Altman analysis, was low (- 3.8 s; 95% limits of agreement (LOA): - 77.8 -70.2 s), and Cohen's Kappa demonstrated good agreement (kappa = 0.809). Hemochron derived ACTs demonstrated worse linear correlation (r = 0.782), larger bias with considerably broader LOA (- 13.14 s; 95%LOA:-316.3-290 s), and lesser concordance between parallel assays (kappa = 0.554). Although demonstrating a fair linear correlation (r = 0.815), parallel measurements on different ACT-devices showed large bias (-20s; 95% LOA: - 290-250 s) and little concordance (kappa = 0.368). Overall, disconcordant results according to clinically predefined target values were more frequent with the Hemochron than i-STAT. Furthermore, while discrepancies in ACT between two parallel iSTAT assays showed little or no clinical relevance, deviations from parallel Hemochron assays and iSTAT versus Hemochron measurements revealed marked and sometimes clinically critical deviations. Conclusion: Currently used ACT point-of-care devices cannot be used interchangeably. Furthermore, our data question the reliability of the Hemochron in assessing adequacy of heparin anticoagulation monitoring for CPB.