Level of agreement between laboratory and point-of-care prothrombin time in patients after stopping or continuation of acenocoumarol anticoagulation: A comparison of diagnostic accuracy

BACKGROUND: Procedures requiring optimisation of the coagulation status of patients using vitamin K antagonists are frequently postponed due to the late availability of laboratory international normalised ratio (INR) test results. A point-of-care (POC) alternative may facilitate early decision-making in peri-operative patients.

OBJECTIVES: To assess the level of agreement between the POC-INR and the laboratory INR in patients who continue or stop vitamin K antagonists to determine whether the POC test may be a good alternative to the laboratory INR.

DESIGN: Study of diagnostic accuracy.

SETTING: Single-centre study at Zaans Medical Centre, The Netherlands.

PATIENTS: Included patients were scheduled for cardioversion (these continued taking vitamin K antagonists), or a surgical procedure (these stopped taking vitamin K antagonists).

MAIN OUTCOME MEASURES: The level of agreement and clinical acceptability of the laboratory and POC-INR results, evaluated by Bland-Altman analysis and error grid analysis.

RESULTS: The surgical and cardioversion groups consisted of 47 and 46 patients, respectively. The bias in the INR in the surgical group was -0.12 ± 0.09 with limits of agreement of -0.29 to 0.05, whereas the cardioversion group showed a bias in the INR of -0.22 ± 0.36 with limits of agreement from -0.93 to 0.48. The percentage errors between methods in the surgical and cardioversion groups were 16 and 21%, respectively. Error grid analysis showed that the diagnostic accuracy of the POC prothrombin time is clinically acceptable as the difference did not lead to a different clinical decision in the surgical group with INR values less than 1.8.

CONCLUSION: The current study shows a good level of agreement and clinical accuracy between the laboratory and POC-INR in patients who stopped anticoagulation intake for surgery. However, in patients who continued their anticoagulation therapy, the agreement between the two methods was less accurate.