Register Login

POC: When to Validate High INRs

George received this question from an anticoagulation clinic coordinator last week: What value would you suggest to repeat a high fingerstick INR to a venous blood, >4.0, >5.0, etc. and why? Thank you. Here is the answer George provided:
“Hello, and thank you for your question. Point of care (POC) instruments provide rapid PT/INR turnaround with reasonable precision, making them indispensable tools for anticoagulation clinics. Medscape provides informative comments from Dr. Jack Ansell, founder of the Anticoagulation Forum,  on the use of these instruments. The precision and linearity of conventional plasma-based “central laboratory” PT/INRs is superior to POC and the managers of most anticoagulation clinics that use POC instruments choose to reflex the assay to the lab for confirmation if there is a significant change from a patient’s previous PT/INR result or if a result is unexpectedly high. Of course, this means the patient now endures a venipuncture and is delayed an hour or two. I just returned from speaking engagements in two locations, and with your question in mind, I asked audience members, all medical lab scientists, which INR they consider to be critical, 4.0 or 5.0. In both locations the audiences were evenly divided!

I recommend you use 4.0 based on data published in Turpie AGG. New oral anticoagulants in atrial fibrillation. Eur Heart J 2008;29:155-65. Turpie shows that the risks of a serious bleed begin to rise dramatically at 4.0. However, there exist no published data that delineate the benefits and adverse outcomes of choosing 4.0 vs. 5.0 as the reflex value for POC, nor do the operators’ manuals of the various POC instruments provide a conclusion. Of course, at 4.0 you inconvenience more patients and incur greater expense, at 5.0 you risk sending home a patient with a bleeding risk. Since I’ve found no documented conclusion, I have posted your question as a “Quick Question” survey on Please watch for the survey and check the answers that it accumulates over the next few days.

Comments (3)
Alan Neal
Sep 11, 2012 2:39am

Why is there a requirement to check POC INR‘s with laborator
Why is there a requirement to check POC INR‘s with laboratory results? If local evaluations are performed and give good correlation and interferences are excluded, for instance high HCT or presence of LA, then if INR is raised, surely intervention is required based on clinical findings and INR value checked by repeat analysis. We find (published data) that INR values at supratherapeutic levels frequently give poor interlaboratory correlation with main analysers due to different thromboplastin sources and their different sensitivities. We frequently see recombinant thromboplastins performing differently to tissue extract thromboplastins, with INR‘s varying as much a 2 units with higher result around 8, though the ISI‘s of the reagents are both around 1.0-1.1. Which result is correct? NB: both testing laboratories are high performing accredited services, with good EQA. I would also suggest that POC INR analysers using thromboplastins with ISI‘s of approximately 2 should be reviewed, as many POC INR systems with ISI‘s of approximately 1.0 have better EQA performance than the main laboratory’s, per UK NEQAS information.
Thank you, Mr Neal. Yes, practice standards require that point of care (POC) instruments that are used in anticoagulation clinics be calibrated to the main lab’s plasma-based assay, and indeed, many POC instrument results are mathematically modified by internal circuitry to more closely match reference method results even though they typically operate on whole blood. In our institution, the POC instruments are backed up by the main lab in the sense that an unexpected result is confirmed by what is perhaps the more reproducible system, so we try for reasonable compatibility between the methods. Any time the INR rises to above 5 or 6, however, all bets are off. You would have difficulty finding agreement between any two unlike reagent-instrument combinations, be they POC or plasma-based. I would advocate that clinicians recognize the need for intervention any time the INR exceeds 4, and that a 6, 8, or 10 all be regarded as having the same clinical significance. Again, thank you for your comments.

Apr 29, 2012 5:33am

This is a helpful comment, thank you. It looks like 4.0 has
This is a helpful comment, thank you. It looks like 4.0 has taken the early lead in our poll. Geo.

Apr 26, 2012 4:15am

As with most laboratory tests where two or more systems are
As with most laboratory tests where two or more systems are used for the same test, I believe it is common to validate less robust systems by correlation with a reference system. I am not sure if relying on poll results to establish linearity would turn out to be a good idea when the inspectors come around!

Within our hospital/outreach system, we check linearity of INRs on POC analyzers by paired testing against our hospital analyzers. We do indeed find that linearity gets much worse around an INR of 3.5-4.0. Our protocol is to have those POC patients with INRs above 3.5 to have a venous draw that is run at the hospital.

There are published articles regarding this problem. They point out the fact that, in general, POC INR analyzers generally have higher ISIs, often approaching 2.0 or higher. Larger analyzers are usually around 1.0. They point out that bad linearity with higher INRs would be expected with anything using high ISIs.

Leave a Reply