A comparison of international normalized ratio results by point-of-care device and clinical laboratory analyzers in a vascular surgery department

Purpose This prospective comparative study aims to evaluate the accuracy and validity of CoaguChek XS point-of-care system (Roche Diagnostics GmbH, Mannheim, Germany), compared with the standard central laboratory international normalized ratio (INR) results, in a vascular surgery ward. Methods Data based on 54 INR pairs, measured by the central laboratory analyzer and CoaguChek XS, were compared using the intraclass correlation coefficient. Other outcomes included the time and cost of the service from the patient's perspectives. Correlation was assessed using Pearson correlation test. Bland-Altman analysis was done to determine mean difference and 95% limits of agreement. Cohen κ test was done to assess agreement between the 2 methods. Results The mean INR values for CoaguChek XS and standard laboratory (Sysmex Ca-1500 system; Siemens Healthcare Diagnostics, Erlangen, Germany) were 1.21 ± 0.28 and 1.24 ± 0.43, respectively. The Pearson correlation coefficient in all 108 measurements was (r) = 0.591 (95% confidence interval, 0.38-0.74; P < 0.0001), indicating a moderately positive correlation. Overall intraclass correlation coefficient was 0.803 (95% confidence interval, 0.324-0.707) indicating good reliability. Bland-Altman analysis showed a bias of 0.035 and a 95% limits of agreement ranging from -0.63 to 0.70. There was substantial agreement between the 2 methods (κ = 0.694; SE, 0.108; 95% confidence interval, 0.481-0.907). CoaguChek XS results were obtained, on average, 213 minutes faster than the standard laboratory. Conclusions CoaguChek XS appears to have positive correlation, substantial agreement, and good reliability when compared with the standard laboratory. Thus, it is recommended to be used in hospital settings.