On-X Aortic Heart Valves: Safer with Less Warfarin


No announcement yet.

Coag-Chek meter information wanted

  • Filter
  • Time
  • Show
Clear All
new posts

  • #16
    I do a lab test about monthly with a draw, but I stick myself every week. I have noticed that the higher my INR, the farther off the Coaguchek XS is. As long is it is consistent I am ok with the error. I can adjust for that, and I just want to know that I am very low or excessively high. Luckily, I am very consistent and have only had to adjust my dosage a few times in five years.
    Diagnosed with a bicuspid aortic valve at age 25.
    Replaced with St. Judes mechanical valve on 12-20-11 Mayo Clinic at age 47.


    • #17
      Originally posted by Jamey T View Post
      .... I have noticed that the higher my INR, the farther off the Coaguchek XS
      This study on the device would seem to support that finding.

      I suppose in the "design tradeoff" phase they (Roche) may have been more willing to accept that given that being out more at higher INR is of less clinical significance than at lower INR (especially if you take a Log view of the signiicnce of higher INR)

      Let me know if you want the PDF of the study


      • #18
        A few years ago, when I was more actively researching the correlation between meter results and lab results, I ran across some papers that also stated that both InRatio and Coaguchek XS had a higher variance from the lab values when the INRs were high. In my testing, when I tried to determine the most accurate meter for my personal testing, I also encountered this. The higher my INR got, the higher the error in results on these meters.

        Because the variance increased in value, it wasn't possible to accurately predict the lab values - for example, if the lab and meter matched at 2.2, you'd expect a match at 3.5 - but with a meter, the 3.5 lab value may be read as 4.0 by the meter (I'm making up numbers here). And a 3.0 by blood draw may show up as 3.3. Clearly, it wasn't just a matter of subtracting an equal number from any meter's result in order to determine lab values.

        A few years ago, one of the people on this forum calculated an algorithm to predict, based on the meter's results, what the lab results would be. I don't have that formula, but suffice it to say that if a mathematical formula is required to figure out what a blood draw would determine, the meters would be difficult for most people to easily use to predict their 'actual' INRs.

        Of course - it's good to know the meter, and to be confident that, at a certain INR reported by the meter, the actual blood draw would repeatedly and reliably report a particular value. (For example, if the meter says 4.0, it would be good to KNOW that the lab would almost always say 3.5). This is kind of like a speedometer with a known error - if it says 20, you're going 20, if it says 70, you're doing 65 - KNOWING this makes the errors almost acceptable.

        With that said - you really have to KNOW the actual value that relates to a value reported by the meter - and for many of us, this makes these meters somewhat less acceptable.

        One thing that should be pointed out is that not all labs could necessarily be trusted to report accurately, either. I've had some INR results from labs that I chose to only use ONCE that were far different from the meter (and, occasionally, other labs that I used within hours of the erroneous lab). Not all labs can be always trusted to give good values.

        FWIW - this error is one of the reasons that I chose to use the Coag-Sense. The Coag-Sense actually does a physical test for clotting - it doesn't use an electronic effect to determine when a clot forms. A meter that does an actual physical test, detecting a physical clot, doesn't rely on the same kind of decision making in designing the meter (can we make errors at the high end of the scale? as hypothesized by Pellicle) because there's no electronic phenomenon to be concerned with. In the case of my Coag-Sense meters, the correlation between meter and lab has been well within limits.
        My meter often reports values that are slightly lower than those of the lab - and often give the same prothrombin times as the lab. This suggests that the values of the reagent for lab or meter are slightly different.

        Personally, I prefer a meter that gives me a slightly lower number than the labs than I would to have a meter that reports slightly higher than the lab. With a meter that gives me a 2.0, I'd much rather assume that the lab value is closer to 2.2 than I would to have a meter tell me 2.0 and worry how much lower the lab value ACTUALLY is.


        • #19
          Early on when I routinely performed comparison tests my XS and the lab agreed more than once. Usually there was a .1 to .2 difference--not enough to be concerned about and I don't remember whether higher or lower. The instruction booklet that came from Roche advised anything up to and including a .8 difference was accurate. Once when it took the lab 3 days to respond the comparison was .8 difference. At that point I considered comparison testing a waste of time the 3 days time frame being useless. Just my experience and feeling.
          Mitral Valve Replacement Surgery, 1999
          Home test weekly since January 2004-Coaguchek S
          October 2006-Coaguchek XS
          INR managed by anti-coagulation clinic


          • #20
            I can't imagine a .8 difference being acceptable in all cases. If a test result was 2.3, does this mean that an actual INR of 1.5 is acceptable? I believe that the acceptable standard is +/- 30%. This is what most meters can do - in my experience CoaguChek XS had a higher degree of variance at high INRs - and I believe this has been well documented.

            When I got my InRatio, I trusted it completely - and paid the price. When I compared it a few times to lab draws, I contacted Alere and was asked which to believe - their answer was always 'trust the lab.'

            For my own testing, I use the Coag-Sense. It has always been very close to MOST labs. I say MOST labs, because some labs don't seem to get an accurate result. I've verified this by having blood draws at two different labs, plus testing with my meter, all within a few hours of each test.

            I'm now trusting my Coag-Sense meter, and doing a blood draw to confirm continued accuracy, every few months.

            If you keep records of your INR results - both meter and lab - it shouldn't matter too much how long it takes for the lab to give you results, once you develop a validated trust in your meter. For me, I record the results of both lab and meter. The last time I compared the two, the INR was the same, and the prothrombin times were within .4 seconds of each other.

            If you aren't already doing so, keep a record of your test results, your warfarin dosages, and anything else that you think may have affected your INR. If you get your blood draw the same day as your self-test, you'll be able to plug lab values into the record - whenever you get the results - and be able to make the comparison.


            • #21
              Originally posted by Protimenow View Post
              I can't imagine a .8 difference being acceptable in all cases. If a test result was 2.3, does this mean that an actual INR of 1.5 is acceptable? I believe that the acceptable standard is +/- 30%. This is what most meters can do - in my experience CoaguChek XS had a higher degree of variance at high INRs - and I believe this has been well documented.
              well firstly I guess that it makes a difference if that 0.8 difference is from an absolute or both are variance around another number which is somehow the "true" INR.

              We know well that INR is a rubbery indicator at best and INR is dependent on the reagents used ... indeed its not even a number that we know affords any specific case by case "protection" against clots.

              For what its worth my own testings have never been greater than 0.3 on any lab vs my coagucheck I've ever done (and I've done a few)

              (*Note: this is not time series data, but each point is just a date where I made a sample ... somehow I've lost some of the sample points of comparison with the lab ... don't know why)

              What matters is not what individuals report, its what you find for yourself with your gear.

              I always encourage people to do baseline checks not just once, but perhaps once per year ... and if variance is found request information on reagents.

              From a Roche publication:

              There are known blood disorders (which are uncommon but exist) which can influence the INR readings.

              Discuss with your doctor and if in doubt just compile some data and look at your own situation


              • #22
                I've found that the lab's reagents can make a significant difference in reported INR. I have probably fifty or more comparisons between meters and labs, and may produce a series like the one that Pellicle posted, when I kick myself into gear and create one.

                On the subject of reagents - INR is calculated by dividing the prothrombin time (the time it takes to form a clot) by a value that is specific to the reagent (reagent in the strip, or reagent at the lab). For at least a year, my Coag-Sense was giving an INR value that was usually .1 or .2 below that reported by the lab. I asked the meter's manufacturer about the difference (although I was completely satisfied by the reliability of the difference - always .1 or .2 off), I thought I'd get a response. The manufacturer's response was that the lab was probably using a different reagent.

                I had a test at the lab last week - after going six months without bothering to make a comparison with the lab. The results were almost exactly the same - INR values matched, and prothrombin times were within .4 seconds of each other. I haven't confirmed that this match will happen again, but it strongly suggests that the lab has changed to a different reagent whose value matches the one in my strips. Producing a reagent value takes a bit of accurate guesswork - there's no solid standard for creating an 'exact' reagent value.

                When I get a chance, I'll see about putting together a lab versus meter comparison chart.