Quality Assurance At the Point of Care: PT/INR: What Is the Gold Standard for Patient Treatment?.

The Prothrombin time (PT) can be performed in laboratories by instruments such as BCS (Siemens), and at the Point of Care using devices such as the Protime 3 (ITC) and the Coaguchek XS (Roche). Laboratory based instruments such as the BCS have been regarded as the Gold Standard for PT/INR assuming results can be obtained in under 24 hours from blood drawn in an uncentrifuged sample. A previous study performed in the Netherlands recommends that INR testing be completed within 6 hours of blood draw. This recommendation has not been universally adopted. In order to assess this effect, a study comparing the laboratory instrument (BCS) to POC instruments was performed while simultaneously comparing lab based results at or greater than 6 hours.

The INR values of the BCS in two separate laboratories, Coaguchek XS, and Protime 3 were compared to each other. The INR values and turnaround times for each were recorded. The results were compared using paired observation T-tests and regression analysis including all data and within sub-therapeutic (<2), therapeutic (2.0–3.0), and supratherapeutic range (>3.0). Additionally, point of care results were compared to BCS test results before and after 6 hours.

60 patients had 3 INR test methods performed. All samples were tested in two different reference laboratories, both using BCS testing method. Reference lab A performed the tests under optimal conditions including a turnaround times of less than 2 hours, while Reference lab B performed the tests under usual workflow conditions. While the total Protime 3 results were significantly different from Reference lab A, the Coaguchek XS results were not. In the therapeutic range, the Coaguchek XS and Reference lab B had significant correlation (r = .556 and .634 respectively) to reference lab A while the Protime 3 did not (r = .281).

Reference lab B results were significantly different from Reference lab A when testing was performed after 6 hours (P-value = .0252), but had similar results to Reference lab A when performed within 6 hours (P-value = .402). The Coaguchek XS showed no significant differences when compared to Reference lab A results (p-value = .1262). Another indicator of lack of concordance between Reference lab A and B, when testing is performed greater than 6 hours, was the change in directionality of the bias (p-value = .0396). Additionally, Protime 3 results in the therapeutic range correlated significantly better with Reference lab B results when they were resulted within 6 hours (r= .191 and r= .782).

Based on these findings, the BCS method appears to lose reliability when the test is performed on uncentrifuged blood greater than 6 hours after blood draw. Interestingly, the Coaguchek XS correlated best with Reference lab A, BCS testing under optimal conditions. The Protime 3, on the other hand, was less well correlated with Reference lab A. The change in directionality of the bias also suggests that the BCS results were flawed when testing was performed after 6 hours. Although additional preanalytic factors (i.e. temperature) may have had an effect on these results, these data suggest that laboratory based INR testing should be performed within six hours of sample collection.

No relevant conflicts of interest to declare.