Correlation of the roche coaguCheck to laboratory plasma international normalized ratio at high international normalized ratio values: A quality assurance assessment of 18 months of data and clinical implications

Point-of-care devices have facilitated the task of monitoring international normalized ratio (INR) values for patients taking warfarin. In previous validation studies performed at the Montefiore Medical Center, the CoaguChek was proven accurate for INR readings below 3.5. Because of this information, 498 consecutive CoaguChek INR results of 3.5 or greater were sent into the laboratory for confirmation. Eighteen months of correlation data were compiled to compare the CoaguChek with the central laboratory values. Analysis of the data showed that there was a quantifiable bias for the POC test giving higher readings than the central lab. Approximately 81.5% of the POC readings were higher than results obtained from the central laboratory, 11.6% were lower, and 6.9% were the same as the central laboratory. The bias increased progressively with increasing POC INRs. This correlation pattern raises the possibility that the international sensitivity index for this POC instrument may require adjustment for prothrombin time. Correction of this progressive bias may provide more accurate POC measurements of INR values and aid physicians to make more accurate decisions regarding the anticoagulant medications for their patients.