Chromatography Forum

Sorry to bother, but I have a question that's been bugging me. I'm just a lay person, so if I say something stupid, please understand. Question: A very close friend had blood drawn (2 vials) for a BAC test. They were sent to the lab and analyzed using headspace chromatography. Each sample was tested with one returning a .086 result and the other a .092. I've always been told that the headspace chromatography was the most accurate test known to man yet that is a 7% difference between the two samples. Is this normal? His blood was "profoundly" acidic according to a doctor. Would that affect the results? I just can't believe that we're using a testing method that has an error rate of 7% or >. I'd appreciate your comments....Steve

I am familiar with the automated analyzer 'by Dupont' which uses headspace to determine blood alcohol content.

The test is very accurate and reliable. There is at least two issues which can affect the results, assuming the test was done on a similar type of instrument.

The first is sample homogeneity. Was the blood sample spun down sufficiently and was the serum homogeneous when it was sampled by the human operator?

The second is sampling error. Was the lab worker extremely precise in taking two samples of serum.

HS, from personal experience over almost a decade of research, is so reproducible that it is an excellent test of the reproducibility of technique of the analyst AND of the homogeneity of the sample.

Either one of these OR both can affect the result by several percent, as was seen in the results you have given us.

I am not surprised at the results, but I would have expected +/- 2%, not the +/- 3.5% given. Which cause is the reason for the variance? No one knows with only one set of data points.

best wishes,

Rod

Thanks for your response. It is my understanding that whole blood was tested, not just serum. The two vials were taken from the same place at the same time. Also, I might add that this person was severely burnt and according to him and witnesses who were with him and using the standard BAC calculator, his BAC at the time the blood was drawn should have been in the .01 range, not either .08 or .09. Because of the burns, as I said before, his blood was extremely acidic and he was also extremely hypovolemic, not from blood loss, but from the plasma loss due to the burns. Just trying to make sense of all this. Thanks again.

Hi

Well I do not think this is an analytical issue, rather a clinical chemistry/medical topic.

I have not been practising clincal chemistry analysis for 16 years but some things comes to mind:

A false high value may be related to that the result has not been compensated for the loss of blood volume (can be estimated as I recall)

Burned patients/vitims tends to be exposed to bacteria etc almost immediatly at the accident, causing risk for sepsis etc that can be a source of sligtly elevated alcohol.

If I recall right a sailor killed on USS Iwo Jima when a gun turret exploded was found to have a BAC of 1,5promille, this was ruled out (ie baterial caused) by complementary analysis of urine and/or eye globe. And there are similar cases like that.

My point really is that a single test result may not be adequate for complex cases/severly tramatized patients/victims to make a conclusion.

Thanks again for the responses. So, based on the situation of the person being severely burned, profound acidosis, possibly septic, and extremely hypovolemic, the the results of the blood drawn (by the way, it was 2 hours after the person was burned) may not accurately depict the person's true BAC at the time of his accident?.............Steve

bois8726,

Without knowing the ng's on column (a function of split ratio, loop volume, sample volume, etc...) it would be difficult for me to say if 7% is good or not. Having said that, I would add to Rod's comment that it would be good to know if the two samples were analyzed at the same time or if there was a gap which could account for the difference.

This sounds a lot more like a legal question (Doctor, what is the effect of the reduced blood volume? What is a BAC calculator? What experience do you have with the BAC analysis? etc....) than an analytical question to me.

Best regards,

AICMM

I am sorry to hear about your friend's injuries, here's hoping that he makes a full recovery.

If both results are above, or both results are below some legal limit, then the apparent discrepancy between the results would be legally irrelevant anyway, unless a clever lawyer fancies building a case that the tests must have been improperly conducted to have generated such discrepant results. I would be reasonably confident that the testing lab has reams of paperwork to show that they did in fact know what they were doing.

Bear in mind also, that the law of averages on which all statisitics are based also says that widely discrepant results can be produced (albeit at very low frequency) by a method that is very repeatable.

Peter

The two samples were tested at the same time..........thanks

This person was not charged with anything by the way, but does have some stigma attached to others thinking he was drunk when he had drank just 5 beers, waited 4 hours before the accident, extremely good health, 240lb male.

Apparently there have been no studies done on the affect of extreme burns on BAC.....thanks again to all.

Hi

As we say in Russia "order of numbers concidied" . When I worked as a paramedic, I was interesting "rather more or less than 0.1" . "Another song", if this disassembly in court, then accuracy is important .

Again, sorry about your friend.

A 7% difference between samples is not bad.
We do forensic BAC headspace here. We accept a sample containing 2-3 grey top tubes. We analyize one and save at least one for an independant test. We have not asked for the independant test results to compare with our results. I suspect do to the fact that it's never mentioned in court, that results from two tubes are the "same".
We analyize 2 aliquots of the tested tube and report the average of the two if each individual result is within 10% of the mean. If not two more aliquots are run and we report the average of all 4.
The biggest problem regarding the blood samples we recieve is that the sampler does not mix the tube. We some time recieve clotted samples.

Again, sorry about your friend.

A 7% difference between samples is not bad.
We do forensic BAC headspace here. We accept a sample containing 2-3 grey top tubes. We analyize one and save at least one for an independant test. We have not asked for the independant test results to compare with our results. I suspect do to the fact that it's never mentioned in court, that results from two tubes are the "same".
We analyize 2 aliquots of the tested tube and report the average of the two if each individual result is within 10% of the mean. If not two more aliquots are run and we report the average of all 4.
The biggest problem regarding the blood samples we recieve is that the sampler does not mix the tube. We some time recieve clotted samples.

Thanks for the info. I do find it a bit surprising that a 10% error is acceptable in a case where a persons future, insurance rates, driving privileges, etc. is on the line. If there were a 7% difference between two tests, who's to say that if a 3rd or 4th were run that the result wouldn't be 10% or 12%, etc.

Another question? Would it be important that when you did the testing that you knew the person being tested was extremely hypovolemic or acidic? Is there any case where you'd need to know that the person being tested's blood was compromised in order to get an accurate BAC? For instance, certain drugs, infections, etc.? .......Thanks again.

10% might sound like a large difference in analytical terms, but in everday terms pertinent to this case it is the difference between 10 and 11 beers. If you were drinking wine you would have to be half way through the second bottle before one extra glass would make a 10% difference to your blood alcohol.

Peter

In our case the largest error potential is in the sampling scheme. For our lab this is beyond our control. For our analytical standard solutions and the like, we are usually well within 5% usually 2-3%.
Before anyone is charged with DWI or DUI our state says that if the sample was drawn more than 2 hours after the incident , then the value must be related back to the incident time. This is where the "fun" starts.
In your case of a 240 Lb person only consuming 5 beers, there are still other issues. For his weight he would most likely add 0.016% bac/beer (4%v/v) so if he drank all 5 in an hour he would have been arround 0.8%.
For the average human the elimination rate is 0.015%/hr (zero order kinetics). So if he drank all his drinks in the first hour and then waited 4 hours he would have been at about 0.2% when he got into his car. (0.08-[ 4X 0.015 ]). The limit for truckers is 0.02%.
There are other considerations such as the Melenby effect, and as you mention blood volume that can be considered. These go beyond the scope of this discussion.
Again I'm sorry about your friend and am making no judgements thereof. I'm just giving you some information.

In our case the largest error potential is in the sampling scheme. For our lab this is beyond our control. For our analytical standard solutions and the like, we are usually well within 5% usually 2-3%.
Before anyone is charged with DWI or DUI our state says that if the sample was drawn more than 2 hours after the incident , then the value must be related back to the incident time. This is where the "fun" starts.
In your case of a 240 Lb person only consuming 5 beers, there are still other issues. For his weight he would most likely add 0.016% bac/beer (4%v/v) so if he drank all 5 in an hour he would have been arround 0.8%.
For the average human the elimination rate is 0.015%/hr (zero order kinetics). So if he drank all his drinks in the first hour and then waited 4 hours he would have been at about 0.2% when he got into his car. (0.08-[ 4X 0.015 ]). The limit for truckers is 0.02%.
There are other considerations such as the Melenby effect, and as you mention blood volume that can be considered. These go beyond the scope of this discussion.
Again I'm sorry about your friend and am making no judgements thereof. I'm just giving you some information.

Thanks for the info Bigbear. And thanks for being non-judgmental. I wouldn't blame anyone if they were judgmental given the circumstances. However, we are at a loss trying to figure out why his BAC results were so high compared to the facts. Either the testing, storage/handling, blood draw were in error, or something physiological happens during the first couple of hours after a severe burn that caused these results to be skewed in relation to what he should have registered at the time of the accident.

Is the blood volume something that needs to be known at time of testing, if it were dramatically altered by the burn?