What is ground truth: home meter or lab venipuncture?

[pem]

The phrase '1 and 1/2 times as slow' is one of those confusing statements that has no meaning for me.

My understanding of the way INR is measured is that it's a linear measure of time to clot. A person with an INR of 2.0 takes twice as long (or one time longer) to clot as a person with an INR of 1.0. An INR of 2.5 takes 2.5 times as long (or 1.5 times longer) to clot than a person with an INR of 1.0. The INR was developed in order to make sense of the varying prothrombin times that labs received based on differing thrombin reagents.

So - in answer to your question, INR, as I understand it, is linear, relating to the standard of 1.0.

I'm going to the lab for a blood draw this afternoon - inconvenient and probably unnecessary - but in the interest of science, I'm doing it. Before I go, I'll test my blood using strips from the same batch, on an InRatio and an InRatio2. Results may be interesting...

Sorry for the confusion: "1 and 1/2 times as slow" means the same thing as "1.5 times longer".

Thanks! So if the INR relationship is linear with clotting time as you suggest, then I guess what I wonder about is the relationship between clotting time and the likelihood of adverse effects. Ultimately, I'm wondering if it is better to err high than to err low by the same amount.

Looking forward to your lab vs meter results!

Best,
pem

 

[Protimenow]

I just did a test on both meters -- same lot of strips. The InRatio gave me a 3.8 (PT 37.9), The InRatio2 - a few secnds later, a different finger was a 3.8 (PT 38.0). Even with a minuscule margin of error, these results are exactly the same.

Now I'm off to the lab. Let me guess -- they'll tell me to reduce my dosage.

I'll share my results and instructions. (And, in response to Pem's question -- I would be more comfortable with an INR on the lower side than I am on the high side. In spite of my recent stroke, where the INR had to be below 2 for some time (in spite of what my meter told me), I would be more concerned with hemorraghic issues than ischemic ones.

 

[Protimenow]

I had my blood drawn roughly 90 minutes ago, so it's probably too soon to expect a call from the clinic. I would be surprised if he lab's draw doesn't also put me out of range.

I would also be somewhat surprised if the clinic doesn't start monkeying with my dosage. Because the clinic believes that all I have are the 7.5 mg tabs that they gave me, I wonder if they'll start having me hold a dose - perhaps every three days or so. This will, of course, make my next finger stick on Tuesday a somewhat iffy value (because it will be either two or three days since the dropped dosage). Of course, I may be wrong. They may want me to make the drive to the clinic so I can wait an hour or so to get some 5s and 2s.

Then, too, these clinics probably don't trust the poor little ignorant patient to be able to figure out how to take TWO pills a day instead or one, and to be able to read the labels and tell a 5, say, from a 2 or 2.5.

I'll report on what the clinic's blood draw results were and how the clinic deals with my slightly high INR.

 

[pem]

Thanks for checking in. I sense your frustration. I know the clinics do their best to avoid "user error" and also how trying that can be

Will be interested to see your lab value.
Pem

 

[Protimenow]

Not frustrated, a bit bemused. I actually went a day early - but the lab was happy to draw my blood.

In fact, my meters said my INR was a bit high -- but would my new dosage actually keeps me at the high end of my range? I didn't make any modifications last night, so I'll be interested in what the clinic says.

 

[newmitral]

Unfortunately, I cannot access the link - it seems to require a username/password.

That page did not require login when I accessed it before, but it is still able to be accessed via the Alere website:

http://www.alere.com/EN_US/products/alere-inratio-2-pt-inr-monitoring-system-poc-test/index.jsp

click on the "documentation" tab then click on the "Product Manual" link.

I'd really like to know what they consider the "validated operating range".

I don't think that range is actually spelled out in the manual, but this web page:
http://webserver.pa-ucl.com/wwwdocs/cltm/Point%20Of%20Care/INR%20Alere%20INRatio2%20(POCT).htm#TopOfPage

makes the following statement in the "Limitations" section on their Inratio2 page:

" 3. A hematocrit higher than 55% or lower than 30% can cause inaccurate results."

I don't know how authoritative that website might be, or where they got their valid range numbers. I would recommend speaking with Alere directly to get the values straight from them.

I will be very curious to learn the results of your own side-by-side test.

Me too. Unfortunately, it will likely be a month or more before I can get a venipuncture lab INR test side by side with my home meter INR test. I will post the results when I get them.

 

[pem]

Thanks very much for the other link and the validated hematocrit range. I'll see if I can get Alere to corroborate that.

Best,
Pem

 

[Protimenow]

I've started a new thread about my INRs -- briefly - my InRatio and InRatio2 both reported 3.8. 90 minutes later, a blood draw was taken, with a result of 3.09.

What do I do about it?

(I'll look for responses on the new thread)

 

[Protimenow]

I've started a new thread with a warning. I invite you to read my post.

In answer to the question about which value is true -- I'll paraphrase what Alere's tech support person told me: Trust the lab.

The lab procedure spins the blood, separating the plasma and drugs from the solids. It's a more rigorous test than a look at a few ml of capillary blood.

Although there may be questions about proper handling of the blood after it's been drawn, or the accuracy of the reagents used to do the testing, in general, the lab seems to be the gold standard.

I found an issue with the meters - I don't know if it's InRatio specific or if the CoaguChek XS and ProTime meters would have the same error - but this may be the reason I had a stroke -- a medication I was taking fooled the meter into reporting a higher INR than I actually had. Adjusting my dosage to this erroneous INR put me in a dangerous range.

There may be other things that can similarly cause false meter results.

When in doubt, it's probably best to trust the labs. (If lab and meter results are both in range, you're doing okay).

 

[pem]

I've started a new thread about my INRs -- briefly - my InRatio and InRatio2 both reported 3.8. 90 minutes later, a blood draw was taken, with a result of 3.09.

What do I do about it?

(I'll look for responses on the new thread)

Interesting. That's about what lab value my model would predict based on you meter results. So far all parallel tests reported in this thread support that model. Yours, mine, and the other (can't recall the user name, but you know who I mean - the articulate one).

I did another meter test today with venipuncture sent to two different labs. I'll report the results tomorrow.

Which medication affects meter but not lab results?

Best
Pem

 

[pem]

(can't recall the user name, but you know who I mean - the articulate one).

Newmitral is who I meant

 

[Protimenow]

Pem:

In the past, I've never had THAT MUCH variation between meter and lab. On the day of my stroke, the meter showed 2.6 -- 36 hours or so later, the lab had 1.7. Although there was a time gap, the gap in values seemed to be beyond your model's predictions.

I noticed that the Quercetin I had taken previously seemed to raise the INR. I'm pretty comfortable in the thought that, for some reason, the Quercetin did something to alter the way that the InRatio arrived at a value. I was thinking of stopping it until I get a finger stick for an ITC meter on Tuesday - but think I'll keep taking it until then so I can see what their meter says.

 

[pem]

Latest results from parallel testing:
InRatio2: INR 3.1
Predicted lab value: ~2.6
Lab1 (ISI 1.8): INR 2.4
Lab2 (ISI 1.0): INR 2.4

Pem

 

[Protimenow]

Pem:

This is indeed interesting. I wouldn't have expected to see that much variance between meter and lab. I'm interested in seeing if others have had similar experience with your predictive model.

The papers that I've read comparing meter results with lab results have all been MUCH closer than the algorithm that you provided. It would be interesting to see what meters other than the InRatio may report. I know that it's not practical, but it would be good if there can be some kind of database that compares lab and meter readings (there IS a database - it's called the anticoagulation clinic). I would NOT expect that big a variance in the majority of tests.

I'm continuing to take Quercetin - the substance that I think fooled my meter - until Tuesday, when I get a fingerstick and INR from an ITC meter. (I'll also do a parallel test on an InRatio before I go to the clinic). It may be interesting to see if there's a difference reported by the two different metering methods.

(As you noted before, my 3.8, according to your algorithm matched the predicted lab value. I wonder if this predictive method is valid only for InRatio or can be extended to other meters. I wonder, too, that if this formula accurately predicts the lab values, why the developers of the InRatio meters didn't/don't somehow program it into their meters. It would sure help sell a lot of meters if this level of agreement with the labs was possible)

 

[pem]

(As you noted before, my 3.8, according to your algorithm matched the predicted lab value. I wonder if this predictive method is valid only for InRatio or can be extended to other meters. I wonder, too, that if this formula accurately predicts the lab values, why the developers of the InRatio meters didn't/don't somehow program it into their meters. It would sure help sell a lot of meters if this level of agreement with the labs was possible)

If you scroll up through my earlier posts, you will see that I did make one side-by-side meter to meter comparison between InRatio2 and CoaguChek XL. The result was almost identical: InRatio2=3.0, CoaguCheck=2.9. That's only a single data point, but it was enough to convince my cardiologist to have me rely solely on labs if that's my preference (the lab results were consistent with each other but predictably different than the two meter results).

I will make one more meter to lab1 to lab2 comparison and if everything is as I expect, I will consider using my meter to generate a predicted lab value and then go by that.

My next order of business is to find out if my hematocrit tends to lie in the "validated operating range" as per newmitral's post. I need to ascertain from Alere exactly what that range is and how hematocrit might affect results. Furthermore, I need to ascertain the stability of my hematocrit over time. If I'm going to rely on my meter to predict the equivalent lab value and that model depends ultimately on a stable hematocrit, I need to know that. Does that all make sense? I think I'm eeking into the territory of diminishing returns with this level of diligence, but I just need to be in that "peace of mind" zone before I trust my health to a piece of hardware or methodology.

Thanks for your ongoing correspondence. Looking forward to your subsequent results.

Best,
pem

 

[Protimenow]

Quoting from the InRatio 2 manual, regarding Hematocrit:
"A hematocrit (percentage of your blood that is red blood cells) that is higher or lower than the validated operating range of the monitoring system can cause an inaccurate result. Refer to the test strip package for more information. Verification of our hematocrit with your health care provider will help ensure the reliability...." Apparently, the range of acceptable hematocrits may vary from strip lot to strip lot. It's too bad that they don't have a home hematocrit test system -- this way, we'd be able to confirm that results are accurate. (In my case, perhaps the quercetin that I'm blaming for the elevated INRs was responsible for modifying my hematocrit enough to get bad results. I may never know)

I didn't go back through your earlier messages, but is there a reason that you're concerned about your hematocrit going out of range enough to influence the InRatio 2's test results?

Also -- regarding confidence in my meters:

My first meter was a ProTime meter that had been previously used in a clinic. I didn't have blood draws to compare it to, at first. My wife's blood - which I used kind of as a control - had an INR of 1.0. I changed from the ProTime to a newer ProTime 3, and parallel test results were very close. I had a mild concussion in December 2010 and the INR at the hospital lab was 2.92. The day before, my INR on my ProTime 3 meter was 2.9. I maintained confidence in the ProTime meters.

When I switched to my InRatio, I again did parallel testing. The INRs were in close enough agreement that I was comfortable with my switch to InRatio. (The InRatio was smaller and easier to use, so I stuck with it). The last test with my InRatio and InRatio 2 - done perhaps a minute apart - yielded equal INRs and almost equivalent prothrombin times. I was confident with their reliability (a statistical term meaning that the results are the same -- regardless of accuracy).

However -- with my recent experience of large differences between meter and lab, I'm a bit concerned. The algorithm proposed for adjusting InRatio2 results to lab results (InRatio value to the .85 power) is a very interesting one. It's one that I may actually create a column in my INRDiary spreadsheet for, and track predicted lab values. IF this formula can be reliably used to predict lab values, it's an important one to know. It may show that my assumption about Quercetin changing the INR reported was incorrect.

If I had the funds for repeated lab tests, and a doctor who could prescribe them, I'd check out the differences that I thought were caused by Quercetin. (That 2.6 versus a lab's 1.7 is still troubling).

Like you, pem, I was the same age at replacement as you were. I've been clicking for quite a while.

 

[pem]

I've analyzed my recent data and improved the model that predicts my lab values from my home meter values. I now have a new model called, appropriately, "Model 2" The original model (Model 1) just adjusted for an overall expansion of the INR, but didn't consider a linear offset. The consequence of this was that Model 1 would consistently overpredict lab values when the INR was high. Model 2 retains the features of Model 1 but also adjusts for an observed offset. So the new model is simply:

(Meter Value raised to the power of 0.75) + 0.25. So, compared with Model 1, I've lowered the value of the exponent from 0.85 to 0.75 and added an offset of 0.25.

Model 1 has a mean absolute error of 0.2
Model 2 has a mean absolute error of 0.15 (that means, on average, the predicted lab INR value will differ from the actual lab INR value by 0.15)

The attached graph shows the meter values ("InRatio2"), predicted lab values ("Model 2"), and actual lab values ("Lab-CPL").

View attachment 9037

I am confident that this model works for me and my meter. I hope to explore further the extent to which this model works for other people and other home meters. So far it is consistent with what newmitral and protimenow have reported in this thread, but there are too few data points to draw any conclusions from that.

Best,
pem

 

[Protimenow]

pem:

It's interesting that you've applied this amount of rigor to attempting to predict a lab value versus your meter's value. It'll probably work well for me.

However-according to InRatio Technical Bulletin 109 (they sent me a copy), the Clinical and Laboratory Standards Institute (CLSI) calls for the following "minimum acceptable accuracy" for 'Professional Operators:'

"In the INR range below 2.0, 90% of the allowable difference between results from the POC (point of care = meter) system and the reference (lab result) shall be +/- 0.5%

In the INR range or 2.0 to 4.5, 90% of the allowable differences between results from the POC system and the reference result shall be +/- 30%"

Clearly, according to your new model, these values seem to be outside of that range. My recent 3.8s SHOULD have been matched by a lab result of AT LEAST 3.6 -- and the actual lab value was 3.09. Your reported values were also outside of the range.

I wonder if you're taking something that artificially increases the INR reported by your meter (just as I was), or if it's a wider issue. (It doesn't really matter WHAT your meter says, as long as it's reliable and particular values can be associated with actual lab values. If, for example, my meter said 4.5 every time my INR was 2.25, and 7.6 when my actual INR was 3.8, I'd be satisfied just to divide my values in half. If I take a medication that consistently raises my INR by .8, I'll just shoot for a range that's .8 higher than it's supposed to be).

I hope that we haven't lost too many readers here. It APPEARS as if the InRatio - at least in our comparing to lab results - is out of the acceptable range in our non-professional hands.

 

[pem]

I think what your predictive model (lab=meter^0.85) is giving you is an additional correction of the meter sensitivity to your Thalassemia condition. This is purely speculation on my part, but past postings on this forum seem to indicate that others believe the meter may be sensitive to that condition, and could give wrong readings in that situation.
The Professional User Guide for the meter, available at:
https://sdmctrlprod.biosite.com/MC/main/mastercontrol/vault/view_doc.cfm?ls_id=1637AE860C782040C3
does say "A hematocrit (percentage of blood that is red blood cells) that is higher or lower than
the validated operating range of the INRatio®2 system can cause an inaccurate result."

Newmitral,

Thanks for your insights on this. I ran your hypothesis to ground. I spoke with Alere technical support today and corroborated your findings that the "validated operating range" for hematocrit is 30-55%. I then looked through old medical records and found, to my pleasant surprise, numerous test results indicating my hematocrit levels in 2008 and 2009. I have plotted those levels here (I can't seem to remember how to display an inline image here, so it is an attachment link instead - sorry):



As you can see, my hematocrit level tends to stay in a narrow range from about 38 to 42, which is well within the validated operating range for the InRatio2:

mean 40.15
stdev 1.39
min 38.1
max 42.3

This seems to refute your hypothesis insofar as hematocrit is concerned. I suppose it is possible that other aspects of the blood profile stemming from the Thalassemia trait could be at play here, but apparently those have not been studied by Alere and are not part of their validation criteria.

Thanks again for helping me to rule out a potential source of variance.

Best,
pem

 

[pem]

protimenow,

It sounds like the Tech Bulletin is saying that below an INR of 2.0 the meter is extremely accurate, and above an INR of 2.0 there is lots of variability. I would be interested to know if they further subdivide that range so we can know the confidence interval for the 2.0 to 3.0 range and from 3.0 to 4.5. Can you make that tech bulletin available to us somehow?

It has occurred to me that other factors (e.g., medication, lifestyle, etc.) could be at play in artificially increasing the meter results and not the lab results, but an answer to that eludes me. The only medication I take is coumadin. My diet tends to be varied but without extremes. I exercise some days and not others.

By the way 3.8 * 70% = 2.66. In other words, 2.7 as a lab value is within 30% of 3.8 as a meter value. Not sure how you came up with 3.6 in your analysis above. Also keep in mind that that is a 90% confidence interval. 10% of the results will fall even farther outside of that range according to the tech bulletin. I will check this confidence interval against my own data.

If my meter results varied inconsistently compared to the lab results, I'd be more interested in the tech bulletin confidence intervals. But my meter values and the lab values correlate so well (Pearson correlation coefficient of 0.85) that I am starting to wonder if Alere needs to revisit their calibration model.

Thanks,
pem