Update on ferritin results variations between providers (Medichecks and NHS)

[jsy_girl]

hi all

I've posted about this many times before (link included below) but in summary I am in conversation with medichecks because I've had several blood tests with medichecks and the NHS, with (to me) large variations. Culminating in the final blood draw being on the same day, all taken at same time, by same NHS nurse. One test went off to the NHS, one went off to the Doctors laboratory (TDL) and one went to County Labs... the latter both medichecks. These came back with three different results - the highest at 53 ug/L (county) and lowest at 27 ug/L (NHS). (results in pic)

I now have a reply from the medichecks doctor who has discussed these results with a colleague at another university, and in conclusion they say this is an acceptable and normal variation. This "acceptable and normal" variation is a fluctuation of 23 ug/L - i.e. NHS often a bit "low" and medichecks often a bit "high". I guess if nothing else, you can read your own overlay onto whichever results you get.

They recommend sticking with one provider for your results to avoid that variation - but my concern is that this would be fine if my 'county lab' results followed the same trend of slightly increasing results over 2020 but they go up and down.

To add to this, from the previous discussion, the longer your blood draw is in transit, the more it degrades, leading to a higher concentration of ferritin and therefore a higher result.

For this reason I think i shall be asking medichecks to only send my results to TDL going forwards- and discounting results by a percentage if they take a little longer to get there.

Reply from medichecks:

I have discussed your results with the chair of our clinical governance committee, who is also a Consultant Chemical Pathologist at Nottingham University Hospital.

We have jointly looked at the external quality assurance data for the Beckmann Access analysers used by your NHS lab, and the Roche analysers used by both TDL and County Pathology in their work for Medichecks. As we have discussed previously the EQA scheme involves sending identical samples to labs around the country and asking them to perform a ferritin measurement using their analysers. Based upon this we were both in agreement that the variability seen between your various results was within the amount expected when comparing between laboratories.

In the latest round of results labs which were using Beckman Access analysers returned lower results than those which were using Roche. The Beckman analysers returned a result which was on average 23% lower than the average for all laboratories, and the Roche analysers returned a result which was on average 17% higher than the average for all laboratories. This fits with the difference that we see in your results, where your NHS results have generally returned lower levels than those you have received through Medichecks. When we look at the degree of variability seen between the 353 laboratory analysers which are assessed as part of this scheme a sample which has an average result of 123 ug/L has a standard deviation of 23ug/L. This means that ferritin levels are quite variable when compared between laboratories which are using different platforms. Whilst there is less variability when comparing between laboratories which are both using the same platform, the variability seen between your results from County and TDL is within the amount that is considered normal when compared to the external quality assurance data.

In summary ferritin is one of the results which varies quite widely when we try to compare results between laboratories, and so it is important to be consistent and rely on results from one laboratory when you are tracking results over time.

What do you all think? I haven't replied yet as, as always, keen to get your perspective on things.

Link to previous post on this (which links back to all the previous ones)

healthunlocked.com/thyroidu...

SeasideSusie humanbean SlowDragon bluebug helvella

[helvella Administrator]

I think I am quite impressed with the effort they appear to have put in to investigating the issues.

And very unimpressed by the analysers varying so much. Particularly on a very common test. (Might not make much scientific sense, but somehow seeing a rarely performed test varying wouldn't be quite so disturbing.)

I think that the labs should have a public statement of their accuracies for each and every test they perform. (With the presentation standardised to make them readily comparable.)

Further, it emphasises that ferritin alone is an inadequate approach to assessing iron levels. However, I'd like to know how well standardised all the other tests are!

Finally, I am very impressed by your keeping the bit between your teeth and getting this far. And thanks for letting us know.

[jsy_girl in reply to helvella]

Hello - thanks for your reply - sorry for the long response time, I wanted to set some time aside to respond to medichecks properly and respond to you as well.

I too am impressed that they have investigated so thoroughly but on the phone the previous time he did say I was not the usual customer complaining - we were having a challenging shall we say, but polite and interesting conversation - maybe he enjoyed it! who knows! Or maybe they are just worried about formal complaints.

I think you make an excellent point about the public nature of the accuracy of the test. I have included that in my response.

Thanks also for your praise. I am also storing up a whole load of thoughts on thyroid treatment in the UK which i shall at some point be channeling in the most productive way i can think of. Your support encourages me to do the same in that pursuit.

My reply to medichecks in full:

Thank you for getting back to me and also looking into this so thoroughly.

That is certainly interesting to understand the variability one gets with using different laboratories because of the different platform analysers, we did also see that the particular laboratories themselves, both using the same Roche analysers also had variations.

I wonder because of this is the information about the standard variation available to the public? I think that the labs should have a public statement of their accuracies for each and every test they perform. (With the presentation standardised to make them readily comparable.)

I also wonder if it could be communicated what the standard variation is for all blood markers - are they this wide for all other blood tests? Can you confirm whether these are publicly available, or if not, whether there would be appetite to publish them.

Your comment "It is important to be consistent and rely on results from one laboratory when you are tracking results over time"

If we go back to my blood results, the NHS trend, as with the TDL trend is that ferritin increases over time, which is consistent with my iron supplementation. However, the County results do not show this result . The county result decreases between August and December.

Further if we go back to an earlier conversation you proposed that, for the sample on 16 December, that due to the time it took for the County results to arrive with the lab, that the blood sample could have slightly degraded which would have led to the result being higher as ferritin was in greater concentration. However, it is lower than the previous county samples.

This county result does seem to fall outside the standard variation of being on average 17% higher.

I do still feel concerned about the county's results, versus TDL.

I do also still feel concerned about the range. I don't understand that if the NHS /Beckman analysers consistenly return results that are on average 23% lower than standard, how they can have a range that goes up so much higher. And the converse of the Roche analysers. Is that something to do with the customers of TDL/County labs not being representative of the population as a whole?

[helvella Administrator in reply to jsy_girl]

At one point, several years ago, we used to see around 150 as the top end of the ferritin reference intervals more often than not.

Then we started to see more and more using around 300 as the top end.

I have long wondered why this change occurred. I am suspicious that the reason is no-one does anything about results over 150, even over 300 for that matter. Action seems only to be taken when it zooms up even higher.

Therefore, I suggest the possibility that the top end was widely increased more as an administrative decision than anything. That is, a result of 200 would be reported as within reference interval and therefore more or less ignored. Rather than the GP having to see a result that was over range and then decide not to do anything! If you see what I mean.

It is really, really helpful when someone takes an issue such as this and hangs on as they work their way through the system.

You are quite right - these results are utterly useless. Even if some are truly accurate, you have no scientific basis for deciding which ones. It is adding in circumstantial detail (such as you supplementing iron) which makes some more likely than others. But that is not a good basis for serious decisions.

[jsy_girl in reply to helvella]

Yes I see what you mean about the ranges. Having seen another post where someone was at 333 and that was deemed okay still.

Well I'll keep going as long as I don't understand it and the explanation doesn't fully make sense.

I do also now think publishing the standard variability would be very helpful.

Imagine if the Thyroid panels have such variability. I dread to think.

That is the worrying thing re accuracy, how do you know which one? I guess you pick a range that works for both analysers and try to get both into that range, then you can be vaguely confident you're in the right area! I for one, will now always add 23% onto my NHS tests, and deduct 17% from a medichecks test and assume that is my correct result! I think I will also always ask for my tests to go to TDL.

[SeasideSusie Remembering]

Hi jys_girl

First of all, I admire your tenacity in trying to get to the bottom of this, I don't know if I would have had the patience or determination.

As you may remember, I do ferritin or iron panel due to trying to raise my ferritin level. I had ferritin tested as part of the full thyroid/vitamin panel in September 2020 with Blue Horizon, sent to County Pathology. All my Medichecks tests also go to County Pathrology and I have just done an iron panel 10 days ago with them. I had hoped to have seen a rise in ferritin level as I have been very diligent with weekly helpings of liver but my level has, disappointingly, reduced from 76.4 to 72.1 whereas I would have expected to see it rise after 4 months and be at least in the 80s if not more (but at least it's better than the 57 it was 12 months ago!). However, the rest of the iron panel actually changed with less serum iron and saturation % so maybe it's the iron content in the liver.

I think we have to accept a variation to a certain degree and I can possibly just about accept the difference between the County and TDL - 53 and 44 - but I find it hard to understand how 27 is also acceptable. The ranges are very different, which I've never understood, and my NHS range is the same as yours, although the unit of measurement is the same, so it would seem we can't go by percentages or the usual recommendation of half way through range that we say here.

As a matter of interest, looking at your changes since previous results, were you supplementing with iron tablets or eating liver to try and raise your level? I'm wondering why your percentage has changed.

[jsy_girl in reply to SeasideSusie]

Hi SeasideSusie

Thanks for your reply, sorry for the long time to respond i was wanting to make sure i set the time aside to respond properly to you and to medichecks. Thanks also for your praise. I am also storing up a whole load of thoughts on thyroid treatment in the UK which i shall at some point be channeling in the most productive way i can think of. Your support encourages me to do the same in that pursuit.

Yes I have been supplementing with iron and eating liver to raise my levels so i was expecting the results to tick up which they have for NHS and TDL.

Yes i remember you saying about using them for your own blood results, and quite frankly as you can see in my County results, they are the least consistent and with the widest variation from the other results.

My response to Medichecks - where i incorporated a further query about the ranges:

Thank you for getting back to me and also looking into this so thoroughly.

That is certainly interesting to understand the variability one gets with using different laboratories because of the different platform analysers, we did also see that the particular laboratories themselves, both using the same Roche analysers also had variations.

I wonder because of this is the information about the standard variation available to the public? I think that the labs should have a public statement of their accuracies for each and every test they perform. (With the presentation standardised to make them readily comparable.)

I also wonder if it could be communicated what the standard variation is for all blood markers - are they this wide for all other blood tests? Can you confirm whether these are publicly available, or if not, whether there would be appetite to publish them.

Your comment "It is important to be consistent and rely on results from one laboratory when you are tracking results over time"

If we go back to my blood results, the NHS trend, as with the TDL trend is that ferritin increases over time, which is consistent with my iron supplementation. However, the County results do not show this result . The county result decreases between August and December.

Further if we go back to an earlier conversation you proposed that, for the sample on 16 December, that due to the time it took for the County results to arrive with the lab, that the blood sample could have slightly degraded which would have led to the result being higher as ferritin was in greater concentration. However, it is lower than the previous county samples.

This county result does seem to fall outside the standard variation of being on average 17% higher.

I do still feel concerned about the county's results, versus TDL.

I do also still feel concerned about the range. I don't understand that if the NHS /Beckman analysers consistenly return results that are on average 23% lower than standard, how they can have a range that goes up so much higher. And the converse of the Roche analysers. Is that something to do with the customers of TDL/County labs not being representative of the population as a whole?

[MMaud]

This is quite enlightening, I think: bmj.com/content/368/bmj.m149

A few months ago, I had blood drawn for two sets of bloods on the same day - one panel requested by my GP, and the other by the Metabolic Bone bods, in readiness for a clinic appointment. The tests were largely different, but had a couple of common tests, including Vit D. Edited to add that I had one blood draw, by the same nurse, at the same time, from a single puncture.

My results were right a the top of the accepted ranges, but differed by some 13 points.

I asked both doctors to explain this to me. My request resulted in lots of wafflage, and "Trust me I'm a doctor" speak, which was disappointing.

[userotc]

Besides test readings, my comments on here 3 days ago attempted to show that acceptable and/or optimum ferritin levels alo appear to vary considerably healthunlocked.com/thyroidu...

My research uncovers figs of 70-150 for optimal health or 20-40 women and 50-70 men.

My personal readings (male) changed from 130 to 89 between Dec 16 and Jul 18 - but both were with NHS so maybe some other reason instead of lab variations (?).

[besttimes in reply to userotc]

As I understand it, ferritin levels can fluctuate in line with inflammation.

[userotc]

Yes. I'm seeing it as a positive despite apparent uncertainty re ferritin readings & limited results.

[Serendipitious]

I think I've just had the same experience as you. SeasideSusie referred me to your profile.

[jsy_girl in reply to Serendipitious]

Have replied on your post about the issue of ferritin results which is very similar to my experience (largely a repeat of the conclusion I’ve drawn from this post for the benefit of anyone who stumbles across this).