9 replies to this topic

[johnhemming]

I do weekly blood tests. I use a few labs for complex historical reasons and I wanted to run a test where a sample taken through the same needle was sent to three labs simultaneously. I think there are important lessons from this that are useful to people. Labs tend to have slightly different panels, but I have stuck to the default panel. I will not identify the labs. I had the blood taken at about 10.20am on 3rd November and I drove the sample to the first lab. I am not sure when they started the test, but I got the results at 5.15pm via my GP so they must have been sent out something like 5pm. The second sample went via a courier and I had the results on Friday 4th November at 17.10 again this was via my GP so they would have been produced a bit earlier. The third lab went via post and the results were available by 6.53pm today.

I am not going to give all the results from all the tests. The units are UK units and I will generally give a result when I had a result from two or more labs.

Haemoglobin g/L 144/144/145 - happy with this
RBC count E12/L 4.49/4.55/4.59 - not as good but not too bad. Delays cause problems for blood cell values
MCV fl 96.1/94.1/95.9 - I wonder if the labs with delay bodge their figures a bit for the delay. This varies a bit too much for my liking.
MCH pg 32.1/31.6/31.6 - OK
MCHC g/l 33.5/33.6/33 - OK
RDW 12.8/12.9/12.9 - OK
Platelets E9/L 173/190/178 - this is a bit varied
WCC E9/L 2.83/2.95/3 - a bit varied
Neutrophils 1.83/1.98/1.97 - a bit varied
Lymphocytes 0.7/0.69/0.71 - OK
Monocytes 0.25/0.26/0.28 -
Eosinphils 0.01/0/0 - ok
Basophils 0.02/0.02/0.02 - perfecto
Na 136.8/140/? - not really happy with this as it is an ion
K 4.8/4.7.? - ok
Urea 4.39/4.6/4.9 - looks like this goes up with time before testing.
Creatinine 73.43/83/93 - this is one that makes me really unhappy because it is clear that testing delays cause the variation.
Bilirubin 6.73/7/6.5 - ok ish
ALP 74.15/67/69
AST 24.63/20/?
ALT 22.73/19/18
GGT 22.73/24/23
Total protein 60.05/67/68
Albumin 38.54/42/38.7
Globulin ?/25/29.3 - too much variation
Ca 2.28/2.3/? - ok
Phosphate 1.26/1.12/? this metabolises
Uric Acid 316.33/298/306
Triglyceride 1.6/1.7/1.74
Cholesterol 4.71/4.7/4.86
HDL 1.73/1.7/2.95 - that last figure looks wrong
LDL 3.01/2.2/2.42 - this also makes me uncomfortable as to the later two labs although it may be that the first is wrong. Still too much variation.
Fe ?/14.9/14.3
TIBC ?48/46.7
Ferritin 176.7/298/291 - these do seem to be the same units and the variation is far too high. My guess is that this is intransit metabolism.
I got an HbA1C from the third lab that is out of kilter with my other HbA1C values and I think is caused by in transit metabolism.
Free Thyroxine 15.16/17.3/17.2
25OHD 147.18/153/147

In conclusion. This gives me a figure I can use to adjust prior values from the same labs to give an inter lab comparison. I knew there were problems with metabolism and Creatinine, but it appears there are other similar problems.

I have not tried to research Ferritin metabolism, but it strikes me that there may be an issue there.

The cholesterol figures are particularly concerning particularly that people get quite excited about this figure. I wonder if it gets metabolised away in transit, but the variation is too big. Looking at the total cholesterol, however, is appears to be a question of deciding what is LDL and what is not.

I have reviewed the history of my HbA1c tests and correlated this to the delays in testing. I estimate that one day’s delay in testing results in the % going up by about 0.5. My lowest HbA1c result this year was 4.18% equivalent to 22 mmol/mol. This was one where I took the sample to the lab and it was tested within an hour or two.

The highest was the one where the test took a few days and was 5.3% 34 mmol/mol. This sort of thing worries me (as does creatinine) because people rely on this data to identify whether or not they have T2 diabetes (or CKD) and it could clearly produce false positives.

[Mind]

What great information. Many suspect that one will get different results from different labs.

 

Most concerning is the wide variation in HDL and LDL, because most doctors are on a hair-trigger to prescribe statins the instant anyone's level rises above a certain level.

[johnhemming]

Its difficult to identify the most concerning. Speaking to a nephrologist recently it appears that creatinine of 122 could result in a biopsy to test the kidney.

 

HbA1c appearing to vary by 1% depending upon test time is also an issue.

[albedo]

This is a bit old but I recollect reading it when I was thinking about slightly different topics related to labs and comparison eg between EU and US labs and also for other biomarkers with wide variability as hormones.

I also regularly test but not at your frequency. Maybe you find something useful:

 

Ceriotti F. Quality specifications for the extra-analytical phase of laboratory testing: Reference intervals and decision limits. Clinical Biochemistry. 2017;50(10-11):595-598.

 

Abstract: Reference intervals and decision limits are a critical part of the clinical laboratory report. The evaluation of their correct use represents a tool to verify the post analytical quality. Four elements are identified as indicators. 1. The use of decision limits for lipids and glycated hemoglobin. 2. The use, whenever possible, of common reference values. 3. The presence of gender-related reference intervals for at least the following common serum measurands (besides obviously the fertility relate hormones): alkaline phosphatase (ALP), alanine aminotransferase (ALT), creatine kinase (CK), creatinine, gamma-glutamyl transferase (GGT), IgM, ferritin, iron, transferrin, urate, red blood cells (RBC), hemoglobin (Hb) and hematocrit (Hct). 4. The presence of age-related reference intervals. The problem of specific reference intervals for elderly people is discussed, but their use is not recommended; on the contrary it is necessary the presence of pediatric age-related reference intervals at least for the following common serum measurands: ALP, amylase, creatinine, inorganic phosphate, lactate dehydrogenase, aspartate aminotransferase, urate, insulin like growth factor 1, white blood cells, RBC, Hb, Hct, alfa-fetoprotein and fertility related hormones. The lack of such reference intervals may imply significant risks for the patients.
Keywords: Reference values; Reference intervals; Decision limits; Quality specifications

[johnhemming]

Thanks for this.  My big concern is that it is known that some biomarkers metabolise after the sample is taken.  This makes material differences to the results, but is often ignored.

[albedo]

Thanks for this.  My big concern is that it is known that some biomarkers metabolise after the sample is taken.  This makes material differences to the results, but is often ignored.

 

Absolutely, a known problem. E.g. see:

 

Murphy JM, Browne RW, Hill L, Bolelli GF, Abagnato C, Berrino F, Freudenheim J, Trevisan M, Muti P. Effects of transportation and delay in processing on the stability of nutritional and metabolic biomarkers. Nutr Cancer. 2000;37(2):155-60. doi: 10.1207/S15327914NC372_6. PMID: 11142087.
 

[johnhemming]

It is a known, but generally ignored problem.

[albedo]

It is a known, but generally ignored problem.

 

Agree, your experiment is a very interesting one.
 

[johnhemming]

Agree, your experiment is a very interesting one.
 

 

I am concerned to get accurate results and it is surprising what a mess the system is.  The Labs should know of the issues, but although you can find research papers there is neither a warning from the labs or some mechanism for tracking the timing of sampling/testing etc.

 

In the mean time, however, I now have pigmented facial hairs on my cheeks where previously there was not facial hair and also a CRP of <0.16mg/L (which is below the measurable limit of my current lab - who at least know what their measurable lower limit is). The former means that my biochemistry has been such that cellular differentiation is improved (this is also evidenced in some low creatinine and HbA1c values), The latter indicates a really low senescent cell load - the figure is a biochemistry average prior to 30s really.

[kurdishfella]

blood tests are usless