Register Login

No More PT in Seconds?

From Dr. Larry BraceEdward Hospital, Naperville, IL: I need an opinion. Here is my view on prothrombin time (PT) in seconds. We should no longer have to do a normal range for the PT in seconds. We know that the general consensus is that the international normalized ratio (INR) is considered normal to at least 1.2 (maybe 1.25 if you calculate to 2 decimal points). I think that the PT in seconds should be eliminated and we should view the INR in the same way we view cholesterol. We don’t do normal reference intervals for cholesterol, we just adopt them based on ATP guidelines because cholesterol is a “standardized” assay. Isn’t the INR a “standardized” assay? We say it is, and that every laboratory that reports INR should match every other laboratory that reports INR. If someone would really need to have a normal reference interval for the PT in seconds (and I don’t know why they would), the upper end of the normal range would be the PT in seconds that correlates to an INR of 1.2 and the lower end of the reference interval would be the PT in seconds that corresponds to an INR of 0.9.

If you adopt as normal an INR of 0.9–1.2, but determine independently the reference interval for the PT in seconds, you invariably get situations in which the INR is normal but the PT in seconds is not, or vice versa. This causes no end of grief because it is almost impossible to explain to clinicians why one result is flagged as normal and the other as abnormal.

I bring this up because during my most recent CAP inspection at Edward, the CAP inspectors had a fit about the upper end of our normal range in seconds (which corresponded to an INR of 1.25). (Yes, we still report both, but I don’t know for how much longer). BTW, the upper limit of 1.25 was based on clinician input because they were unhappy chasing clinically irrelevant “abnormal” results. Personally, I would be just as happy with an INR upper limit of 1.2 and will lobby for that with the newest lot of reagent (see below). When we did our last normal range verification, it comes as no surprise that our normal range in seconds did not match the upper and lower limits based on the INR. I simply adjusted the PT in seconds to match the INR. We just completed our evaluation of a new lot of PT reagent. We have 3 Stago instruments in 2 locations – the Satellite is at an off-site. Here are the results of the reagent validation:

Compact range: NEW LOT PT= 11.9–13.5, INR 0.92–1.08; OLD LOT PT= 12.2–15.4, INR 0.93–1.26
Evolution range: NEW LOT PT= 12.2–13.4, INR 0.94–1.06; OLD LOT PT= 12.2–15.4, INR 0.93–1.26
Satellite range: NEW LOT PT= 11.6–13.6, INR 0.90–1.10; OLD LOT PT= 12.4–15.7, INR 0.93–1.26

What would you do? There is no way that an INR upper limit of 1.1 would ever work.

From Geo: Dr. Brace and I have had this discussion on one or two occasions in the past, and I’ve stubbornly hewed (hewn?) to “conventional wisdom,” asserting that the INR was only developed for monitoring Coumadin. To screen for extrinsic pathway coagulopathies, we must use and report the PT in seconds compared to our locally developed PT reference interval. However, I’ve never been able to refute his argument, presented above with eloquence. It is true that most clinicians who have entered the profession in the past 20 years only know the test by “INR,” and, in fact, the INR is nothing more than a ratio computed directly from the PT in seconds, so why not just use it? I’d like to hear from our participants on this subject.

Comments (6)
Anticoagulant Therapy
Mar 19, 2013 4:51pm

With respect to Dr Brace, the INR is calculated mathematical
With respect to Dr Brace, the INR is calculated mathematically but the calculation is based on 2 measured components, the PT and the ISI. As we know, the ISI is determined by performing PTs on a group of stabilized warfarinised patients and some normals calibrating against one of the International thromboplastin standards. For the INR to be used in liver disease and other groups of patients one would have to be sure that the ISI as determined on warfarinised patients, is also applicable to these other groups of patients and there is evidence that this is not the case. See The international normalized ratio to prioritize patients for liver transplantation: problems and possible solutions A. TRIPODI, V. CHANTARANGKUL and P. M. MANNUCCI Journal of Thrombosis and Haemostasis, 6: 243–248.
Interestingly, Tripodi has also written in JTH (doi:1111/jth. 12166) that for rivaroxaban monitoring one should use an ISI determined specifically for the drug (by the manufacturer hopefully) or the old fashioned PT ratio (patient PT/MNPT) rather than the conventional INR. Regarding the interpretation of the PT in seconds, I don’t see that this is a problem as long as the reference interval is always reported and of course this can be a pain to determine and check with new batches of reagents. But this is the case with almost all analytes.

Mar 19, 2013 11:35am

In an article published in McGlasson DL. Laboratory vari
In an article published in McGlasson DL. Laboratory variables that may affect test results in prothrombin times (PT)/international normalized ratios (INR). Lab Med Feb 2003 34:124-9 we discovered that if you didn’t locally calibrate the INR there would still be huge differences in comparing six different reagent/combinations in INR results. With one combination we found that the same specimen gave an INR result of 3.95 with one reagent/instrument combination and a 7.16 with another test set-up. In a large group of subjects on oral anticoagulant therapy (OAT) we found that the the difference in INR means ranged from as low as 0.8% to 16.4%. In fact we found that when the INR gets above 3.0 it tends to flatten out and a subject with an INR in the therapeutic range of 2.0-3.0 that the chromogenic FX level was the same as some INR results from 6.0-12.0. See McGlasson DL, Romick BG, Rubal BL. Comparison of a chromogenic factor X assay with international normalized ratio for monitoring oral anticoagulation therapy. Blood Coagulation and Fibrinolysis 2008. 19:513-7. Rosborough saw similar results in a subsequent publication Rosborough TK, Jacobsen JM, Shepherd MF. chromogenic factor X and INR differs during early warfarin initiation compared with chronic warfarin administration. The INR probably should be replaced with well established ranges for chromogenic FX and we could replace a lot of variables that are inherent in the method.

Mar 18, 2013 2:38pm

I started this conversation with the hope it would lead to g
I started this conversation with the hope it would lead to good discussion of the topic. So far so good. Like most laboratorians, I know that the INR was initially developed to monitor patients on stable oral anticoagulant therapy, but its use has morphed way beyond that, and in my opinion for good reason.

I would like to remind everyone that the INR is nothing more than a mathematical manipulation of the PT in seconds that accounts for the variability of PT reagents. It does not matter if the patient is taking an oral anticoagulant, if there is “PT” factor deficiency or if there is a derangement of the coagulation system due to liver disease, etc. Math is math! The fact that we have accounted for PT reagent variability leaves us with a “standardized” test. The PT in seconds is no such thing. If I tell a clinician who practices at 3 different locations with labs that use 3 different PT reagents that his/her patient has a PT of 17 seconds, will the clinician really know what that means? The answer is no, because the time in seconds is completely reagent dependent. A 17 second PT could correspond to an INR of 1.3 to 2.0 depending on the reagent used. The importance of this is is the potential for inappropriate treatment of patients (e.g. use of FFP, etc.)based on a mis-understanding of the meaning of the PT in seconds. If I tell the clinician that the INR on the same patient is 1.3, the meaning is completely understood. At an International Society on Thrombosis and Haemostasis meeting where Armando Tripodi (one of the principal international proponents of the INR) was speaking more than a decade ago, I posed this same situation to him in a open forum (1 clinician, 3 locations, 3 different labs using 3 different reagents). He agreed that the INR was a better lab value for the clinician than the PT in seconds. Again, I emphasize that the INR is nothing more than a mathematical manipulation of the PT in seconds that reflects the degree of derangement of the coagulation system no matter what the cause. Why should we have a value in seconds that can mean very different things and have serious impact on patient care? Why are we still messing around with the PT in seconds? I realize that this approach may be uncomfortable to many, but the same was true when we accepted “standardized” lipid reference intervals. When was the last time your lab did a cholesterol normal range study? And if you did, why?

As a lab director, I can tell you that coag is not the only lab area experiencing some of these same issues. Immunochemistry is the same. If I measure growth hormone by 3 different FDA approved assays, I will almost assuredly get 3 widely different values, so the meaning of the test result is test method/kit dependent. Efforts are currently underway in the international community to standardize (harmonize) test methods so the the same value is obtained wherever the test is performed, just like we did for cholesterol. We have already done that for the PT – it’s called the INR.

George Fritsma
Mar 18, 2013 4:11am

Posted on behalf of Dave McGlasson, Wilford Hall USAF Medica
Posted on behalf of Dave McGlasson, Wilford Hall USAF Medical Center, San Antonio:

George, We quit reporting seconds for the PT/INR result some years ago when we went to lower ISI reagents.  Every time we would report a PT above 14.0 seconds our surgeons would start ordering FFP. Dr. John Olson adopted the practice of only reporting the INR and it almost completely dropped the useless ordering of FFP because as Dr. Brace noted a PT of 15.0 seconds was translating to an INR of around 1.2.  There should be little argument on this issue.

Mar 16, 2013 10:18am

As a bench tech in a hospital, I have had questions from oth
As a bench tech in a hospital, I have had questions from other clinicians regarding this apparent “duplicity” with PTs and INRs. On the one hand, it does seem like we should be using one or the other. By using both, along with different sets of reference ranges (secs vs. INR ratios, therapuetic vs. non-therapuetic), leads to this type of confusion.

It seems to me, though, that the concept of a “normal” INR is perhaps not useful for two reasons. One, as mentioned above, is that the INR, and the method used to calculate it, was developed to help monitor coumadin patients across platforms. The other reason is that if one is going to use any analyte (derived or not) to monitor a patient, one needs to establish a reference range based on the normal patient population. Using any particular “normal” range of INRs to calculate backwards to a normal PT range seems like we are over thinking things a bit. If a “normal” range for INR is to be established at all, it must come from real patients.

I am afraid to say that we should stick with the status quo, which is to use the INR for monitoring coumadin patients, and the PT for everything else.


Mar 14, 2013 6:22pm

Dear George,
This is a quick comment on this subject. i agr

Dear George,
This is a quick comment on this subject. i agree with your statement regarding ‘conventional wisdom”. In our institution we have never (ie never) issued an reference interval for an INR – only a “suggested” therapeutic interval. We do not report the PT when an INR is requested. So no auditor has ever been able to criticise us for our non existent reference interval for the INR.
Reporting only INRs can have a potential problem when the increased PT is due to liver disease or a drug such as rivaroxaban where conventional wisdom suggests that the ISI for the calculation of the INR should be indepently determined for these patient populations and who do we know does this?
I know that many labs report INRs only but only because it simplifies the IT side.

Leave a Reply