Customize your JAMA Network experience by selecting one or more topics from the list below.
In Reply We agree with Dr Fantz and the FDA guidance1 that confirmation venipuncture INR testing should be performed, at minimum, for all patients with POC INR testing results of 4.5 or greater. However, we recommend that each institution examine whether a lower threshold for performing POC INR testing (such as >3.5) would better meet patient safety goals. Donaldson et al2 conducted a single-institution study comparing POC INR values using 2 different POC devices (CoaguChek XS Plus and i-STAT) vs venipuncture INR (Stago). Based on prior studies,3 the authors defined a clinically significant INR difference (one that would result in differing dosing plans) as when the INR measurement was within therapeutic range on one device and out of therapeutic range on the other device or when both device INR measurements were out of range with an INR difference of 0.5 or greater. They also assessed the percentage of times that the device INR values differed by 0.4 INR units or more. The CoaguChek POC INR values were significantly different, with at least a 0.4-unit difference, from the venipuncture INR values in 33% of cases and the i-STAT POC INR values were significantly different, with at least a 0.4-unit difference, than the venipuncture INR values in 54% of cases.
It is true, as Fantz states, that there are many variables that contribute to discrepancies between POC and venipuncture INR measurements, including, but not limited to, the thromboplastin reagents used and the assay international sensitivity index (ISI), as well as the variables delineated in our article.4
At the University of Chicago, where the patient in the article was treated, the coagulation laboratory uses rabbit brain–derived thromboplastin (Diagnostica Stago STA Neoplastin CI Plus) with an ISI of approximately 1.3. In contrast, the POC test (CoaguChek XS Pro) used at the hospital uses a thromboplastin-containing recombinant human tissue factor with an ISI of approximately 1.0. Based on the differences in these reagents, discrepancies between the POC INR instruments and venipuncture INR systems are exacerbated compared with what would be expected based on the distinct sample types.
In 2015-2016, we performed a 9-month correlation of all POC INR values of 3.5 or greater with simultaneous venipuncture INR values (n = 130). The POC INR instrument result was, on average, 0.98 INR units higher than the venipuncture INR (95% CI, 0.78-1.19). Of these 130 data points, 88 were POC INR values between 3.5 and 4.5; in that subgroup, 74 of 88 (84%) had a POC INR value that was at least 0.5 units higher than the venipuncture INR value (ie, clinically significant per Donaldson et al2), while 39 of 88 (44%) had a venipuncture INR confirmation value between 2.0 and 3.0 (ie, therapeutic).
As Fantz mentions, POC instrument users must be aware of the correlation of their POC INR testing instrument with their laboratory INR test results. Such correlations may be less robust at a particular institution, given reagent and instrument factors. Because of this, we recommend that clinical sites that use POC INR monitoring should perform internal correlation studies with their venipuncture INR assays to determine the “ideal” higher-end INR cutoff that will require a venipuncture INR confirmation.
Corresponding Author: Irsk Anderson, MD, Department of Internal Medicine, 5841 S Maryland Ave, MC3051, Chicago, IL 60637 (firstname.lastname@example.org).
Conflict of Interest Disclosures: Dr Wool reported receiving personal fees from Diagnostica Stago. No other disclosures were reported.
Anderson I, Wool GD, Madden W. Confirming Point-of-Care INR Test Results—Reply. JAMA. 2020;323(12):1191–1192. doi:10.1001/jama.2020.0951
Coronavirus Resource Center
Create a personal account or sign in to: