INR management - target (not zone) and stats views

[pellicle]

Hi

probably this will have a limited audience of interest, but I thought I'd put my results up here in case anyone was interested.

I have been monitoring weekly with "ad hoc" monitors (of twice weekly) for about 2 years. More recently I have been extending my 'model' of analysis to include a factor which I call "metabolism" but its just a name. Its based on a scalar factor to apply to the estimation of warfarin in my body. The estimation of warfarin is based on the known estimates of half life of warfarin and then calculated over the last 4 days (at any given moment)
{for instance modeling in Excel this is a sample of the cell formula "=(C6+(0.75*C5)+(0.5*C4)+(0.3*C3)+(0.25*C2))" }

Of course this metabolism is NOT going to be stable (or even correct for any person in particular). Hence the scalar factor which is applied to it to then match the INR.

This can not be used for INR prediction (as I have no way to predict the metabolism scalar in advance) but it has enabled me to see interesting trends and make some gross estimations.

So, from over 100 sample points in the last 12 months I have seen that my "metabolism" has varied from a MIN figure of 6.15 to a MAX of 10.78

This is useful because I can then apply this pair of metabolism values to my current "accumulation of warfarin" and see what the likely outcome in INR will be should it reach that.

So, below is a table (based on my observed data for myself) which shows what my INR would be if my metabolism reached a MAX or a MIN value given being on a stabilised daily dose (in the header row)

dose	7	7.5	8	8.5
MAX INR	3.19	3.41	3.64	3.87
median INR	2.43	2.60	2.77	2.95
MIN INR	1.82	1.95	2.08	2.21

This proves valuable to me as it allows me to observe my trends in INR (well now I actually pay attention to trends in metabolism, which appears cyclic in nature) and if it is trending one way or another I know which dose to alter (and when I need to alter it).

Basically all the daily dose levels will see me within my desired 'range' of INR but when my metabolism trends high or low only one side of the dose rnage or the other will give me the desired outcomes.

So I could choose to sit on 8mg daily and be comfortable that my range would swing naturally between 2.08 and 3.64 ... 2.08 is a little low for my preference but its been rare and transient that such occurs.

Alternatively I could see that trend and adjust my dose to (say) 8.5 where the lowest INRwould be 2.21 (inside my preferred min value).

For those who keep electronic records you may find this helpful to examine yourself and see if you can discover that you have such natural limits to your trends. If you do then perhaps this will allow you some better insights into your own INR management. Please feel free to ask for more details on my calculations if you are interested (perhaps by PM so that I don't miss it posted in a thread).

Best Wishes

 

[Eva]

This is interesting! I'm going to ask my hubby to help with this !!

I/we did something similar for my synthroid medication plus acitomel. Hubby spent a long time perfecting it with notes regarding the dose/times and their effect and the repetition of a pattern. Since 1989, after I was treated with radio active iodine I-131, I have not been able to stay on the same dose for more than three or four months! Unfortunately, none of the doctors was interested...maybe they could not grasp it! So, I'm still suffering from "label thyroid".

Thanks for sharing. I'll let you know.

 

[Protimenow]

Pellicle:

You're probably right -- it may not be of much interest to a lot of people on this, or any, non-technical forum. It's interesting to me.

For some strange reason that I can't quite understand, I've gone from testing, like you did, more than once a week to testing every 10-14 days. This is contrary to advice that I give others - and myself. And instead of testing with two meters (Coag-Sense and CoaguChek XS, or Coag-Sense and hospital lab), I've been testing with just the Coag-Sense. Perhaps it's partially because the clinic decided that my range should be narrowed to 2.0 - 3.0, and my meter often tells me that I'm at 2.9 or so (which, at the lab may be closer to a perfectly safe 3.1 or 3.2).

For nearly two years, I've been having monthly (approximately) blood draws, and always slightly before or after testing with a meter. I plan to create a chart or table that compares these values using the same formulas as those seen on some of the papers that I've seen that compare meters to lab results.

These calculations are, I suppose, designed to find some 'truth' relevant to testing with meter or lab. Once I finally do the simple calculations (or have Excel do it for me), I'll probably post the results.

Your idea of calculating a 'metabolic' factor is interesting, although, of course, there are so many other factors (diet, exercise, weather, phase of the moon, sunspots, what your dog had for dinner, etc.) that can be included into this overall factor. With weekly testing, it may not reveal a lot if I apply this to my weekly results -- but it's certainly interesting.

 

[pellicle]

For the lurkers who read but don't prefer to post I have a little more on my blog now

http://cjeastwd.blogspot.com/2014/05/inr-management-goldilocks-dose.html

 

[newmitral]

Hi Pellicle,

Being an engineer myself, I have made similar calculations to try to understand the higher than typical variability of my INR.
I do find your methods interesting, and at the risk of boring non-technical types to tears, would like to try to clear up a few points about your model.

WARNING ***
Non-technical types should probably quit reading now and go get a beer or something.

I may misunderstand what you are saying, so please excuse me if that is the case.

I see two different factors coming together in your model - very similar to things I have also tried to use.

One is the "metabolism" factor, which is how quickly your metabolism clears the warfarin from your body. The other is the sensitivity of your body to the warfarin, which would be the scalar relationship factor between the amount of warfarin present and the resulting INR.

I am probably misreading what you said, but it seems as though what you are calling your "metabolism" factor is actually what I think of as the "sensitivity" factor. Since you say:

"I call "metabolism" but its just a name. Its based on a scalar factor to apply to the estimation of warfarin in my body."

You appear to have used a fixed rate of metabolizing the warfarin based on half-life, since you say:

"The estimation of warfarin is based on the known estimates of half life of warfarin and then calculated over the last 4 days (at any given moment)
{for instance modeling in Excel this is a sample of the cell formula "=(C6+(0.75*C5)+(0.5*C4)+(0.3*C3)+(0.25*C2))" }"

This latter concentration/decay of the amount of warfarin based on half-life is what I think of as a "metabolism" factor, since this is where the speed of clearing the warfarin from your blood is modeled. Your formula, as you say, is based on a constant 48 hour half-life, i.e. each preceding day's dose is reduced by 0.7071^d where "d" is the number of days. I have tried using a similar approach, but I apply the metabolic variability to the half-life factor so that each preceding day's dose is represented by
M^d, where M is a number which can vary to represent change in metabolism.

So, I would represent your cell formula as follows:
"=(C6+((M^1)*C5)+((M^2)*C4)+((M^3)*C3)+((M^4)*C2))"
and I would allow M to vary representing faster or slower metabolism factors. You appear to be using a fixed value of M=0.7071

Using a varying metabolic half-life factor leads to a similar estimate of current concentration of warfarin, which can then be multiplied by a different, perhaps also varying sensitivity factor, "S", which represents the sensitivity of the current INR to the current concentration of warfarin.

I believe this sensitivity factor, "S", is what you are calling a "metabolism" variation. I may be misreading what you have said, so please correct me if my interpretation of what you have described is incorrect.

Having said all that, I have not really found this type of modelling to be personally all that useful in managing my INR. I have just passed 3 years on warfarin, and in my particular case, my daily warfarin dose varies very little overall (12.0mg/day to 13.0mg/day). My INR is highly variable, often bouncing from one extreme of my range (2.5-3.5) to the other from one week to the next. I only really react and skip 1/2 or 1 dose if my INR spikes up over 4.0. It has never dropped low enough to cause concern.

I have been keeping all my INR statistics for the past 3 years in a MYSQL database accessed through an application interface page on my website.
This tool is like a spreadsheet on steroids. This allows me to plot my INR, similar to the way you plot yours in Excel. I also get a histogram plotted of the most recent weekly measurements along with a calculation of the mean and standard deviation. I normally have the plots and statistics default to the most recent 12 month period, but I can select any time frame I wish to analyze.

This tool has been useful to analyze the reduction in standard deviation as a result of changes in the daily low-dose K2 vitamin I am taking to try to reduce the variability of my INR. (see my posts in this thread: http://www.valvereplacement.org/forums/showthread.php?42046-Vitamin-K-tablets)

An additional feature of my web-based INR tracking/analytical tool is that I can, with a click of a button, generate a pdf FAX to send to my doctor with the last few weeks results. I FAX this to his office every month or so to keep him informed of my test results now that I have gone independent of the Alere weekly home-test program where I had to call my test result in to them each week and they relayed it on to my doctor - for an outrageous price.

I mention this web-based tracking and analytical tool which I have developed and have been using myself for the past few years, to see if there is any interest from others to use it. I would probably not want to take on the burden to host a service to let others use this tool on my own website, but I would be willing to share the webpage scripts (they're in PHP) and MYSQL database structure/design to others who might want to set up something similar for themselves. At the moment, there are a lot of things hard-coded for my own particular personal tastes and medical parameters, but I could make it more general if there were sufficient interest.

Pellicle, I can imagine that you may have an interest, but I'm not sure anybody else would. So, I am mentioning it here to see what sort of reaction is produced.

My apologies to the non-technical readers, but you have only yourself to blame if you stuck it out and read this whole thing through to the end.:biggrin2:

 

[Agian]

 

[pellicle]

Hi

that's classic, I love those guys (and these guys too)

 

[pellicle]

Hi

pardon the out of order reply, but:

Pellicle, I can imagine that you may have an interest, but I'm not sure anybody else would. So, I am mentioning it here to see what sort of reaction is produced.

I really loved Again's resonse


from here I'm going to repaste your warning

WARNING ***
Non-technical types should probably quit reading now and go iron the washing or something else more interesting.

...at the risk of boring non-technical types to tears, would like to try to clear up a few points about your model.

I may misunderstand what you are saying, so please excuse me if that is the case.

thankyou for your courtesy

I see two different factors coming together in your model - very similar to things I have also tried to use.

One is the "metabolism" factor, which is how quickly your metabolism clears the warfarin from your body. The other is the sensitivity of your body to the warfarin, which would be the scalar relationship factor between the amount of warfarin present and the resulting INR.

thats about it ...

I am probably misreading what you said, but it seems as though what you are calling your "metabolism" factor is actually what I think of as the "sensitivity" factor. Since you say:


You appear to have used a fixed rate of metabolizing the warfarin based on half-life, since you say:

I do ... its of course "wrong" because there is variance in that there are isomers of warfarin and both are processed at different rates in the body and at different rates in different people. So my model is intended to trowel over this variablity and produce a more simplified output with only 2 input parameters

This latter concentration/decay of the amount of warfarin based on half-life is what I think of as a "metabolism" factor,
...
You appear to be using a fixed value of M=0.7071

I'd need to reflect on that for a while but perhaps we are on different paths to the same answer. I have taken the view that when on a stabilised daily dose one can work out the residual amount of warfarin in the system (assuming if follows the half life) as basically 2.8 times the dose (which is what my formula results in if you run it through at least 4 iterations. So a dose of 7.5mg will result in an accumulation of 21.

The metabolism factor is the scalar required to match this value to my actual measured INR (which may be say 2.7). So in that case the "Metabolism" factor is about 7.77

I happen to find this somewhat intuitive to work with because an increased metabolism (say 7.9) will mean that my body is clearing the warfarin faster than the formula resulting in a lower INR.

Of course diet is all rolled into it and so as an end user you would need to factor that in to any INR effects too. This model is far more simple.

Having said all that, I have not really found this type of modelling to be personally all that useful in managing my INR. I have just passed 3 years on warfarin, and in my particular case, my daily warfarin dose varies very little overall (12.0mg/day to 13.0mg/day). My INR is highly variable, often bouncing from one extreme of my range (2.5-3.5) to the other from one week to the next.

well that's disappointing. It suggests that this is not easily transferable to others ... certainly not to all others.

do you take your daily dose totally stably?

I only really react and skip 1/2 or 1 dose if my INR spikes up over 4.0. It has never dropped low enough to cause concern.

I have been keeping all my INR statistics for the past 3 years in a MYSQL database accessed through an application interface page on my website.

good idea, I thought about doing that (although I'd have probably chosen PHP / Firebird as being an Oracle / PL/SQL guy I like the flexibility of that (personal taste).

My apologies to the non-technical readers, but you have only yourself to blame if you stuck it out and read this whole thing through to the end.:biggrin2:

:-D

If you feel inclined to discuss this, please PM me and we can dribble away about this without distubing everyone else ;-)

Thank you for your time in critique and analysis and your thoughtful replies.

Best Wishes

 

[Agian]

They say that 1% of people with mechanical valves have an 'event' every year. I guess that statistic relates to the majority of Warfarin-takers who test monthly and leave the decisions to the 'professionals'. I'm starting to think that doing what you guys do minimises the risk to something significantly less. I expect a quick tutorial when it's my turn. You guys are great.

 

[TheGymGuy]

Very good discussion. I have been self managing for almost 2 years now and test 1-2 times per week on average. I have not observed much of a variance in my results other than the food impact.
Once I got in range and doze with Vit. K in pill in the morning and night my range so far has been 2.2 - 2.8 pretty stably. I wrote about topping off your Vit. K stores which will reduce sensitivity to Vit. K from food sources within 200-300 mcgs.

In the past when I averaged at 2.0 - 2.1 for 2-3 weeks straight we upped my doze and I have kept it stable since. I wonder if my numbers will change when I cut my bodyweight by 10%, will keep in the loop when I do that.

 

[Protimenow]

I kind of wonder if TheGymGuy used doze (as in falling asleep) instead of dose because he read those discussions. I also loved Agian's response -- it was a great fit.

I've worked with MSSQL in the past, documenting some software that run under PHP. I'm not sure that I'll need to build a database around my INRs.

I've been self-managing for 5 years, and have pretty much ignored the advice from an anticoagulation clinic for about two years. I'm concerned that they've decided to switch my range from 2.0 - 3.5 to 2.0 - 3.0 -- other self-testers may successfully have reported INRs, when their actual INRs are 1.6 or 1.7. The only time I had a negative event was when I trusted my meter's 2.6 and was actually at 1.7 (according to the hospital that diagnosed my TIA).

Personally, I'm not sure that the variables for INR management and prediction can be controlled enough to accurately model the metabolic and other factors that are at play in establishing an accurate method of predicting INR. From meters that are probably unreliable to a certain degree to different bodily processes that may slightly alter our metabolism of warfarin to dietary factors that are hard to standardize to other things, I'm not sure that we can really do much modeling of INR related to these factors with any significant degree of confidence.

I WILL, one of these weeks, do an analysis of my two years of comparison between one or meters and the lab results.

 

[newmitral]

Hi Agian,

Great picture - I really liked it.

They say that 1% of people with mechanical valves have an 'event' every year..... I'm starting to think that doing what you guys do minimises the risk to something significantly less.

I have heard there is about a 1% chance of being in a fatal car accident in your lifetime (not annually). Just like driving defensively and paying attention while driving will help to keep you from being one of those fatalities, I am doing what I can to minimize my chances of having an "adverse incident" on warfarin. While it certainly is not difficult, and it has had no significant impact on my daily life, I do believe that knowledge of how your own body reacts to warfarin, and how your INR behaves will help stack the odds in your favor.

 

[Agian]

The more I learn, the less anxious I become.

 

[pellicle]

The more I learn, the less anxious I become.

It is when I read statements like this I feel that this forum has great worth.