Autopsy: First Batches

by Dr. Alan D. Barbour (contributor)

M. Henri Paul died Sunday, 31 August 1997.
Autopsy at Paris morgue:
Blood samples drawn directly from the heart.
Urine, eye fluid and bile sampled.

Urine and vitreous humor samples can be used for alcohol analysis - values slightly higher than for blood at equilibrium; of no use for CO analysis. Vitreous humor not often taken in such a case (fresh body, other fluids readily available), but a good precaution. Can also be used for some other tests. Bile is, in my opinion (others may differ), largely useless for drug analysis.

Stomach was intact, as were other organs. Aorta ruptured.

Ruptured aorta consistent with side impact at greater than forty-odd mph. As I recall, the Mercedes did not hit the column head-on.

No liver damage or signs of alcoholism.

Chronic alcoholism should be manifested in the liver, and detectable by examination of tissue sections. Long before cirrhosis sets in there is fatty liver, and before that comes alcoholic hepatitis (inflammation of the liver).

Autopsy: Second Batches

4 September 1997 a separate blood sample was drawn, at the request of the family of M. Paul. It is reported that when Stephan saw the first lab. test results and realized their import, he personally attended the taking of a second batch of bodily fluid samples, had the process photographed, and used different labs for tests.

It is, at least in form, what a conscientious magistrate would do; and the more complicated a conspiracy, the more people who are involved in it, the less the chance that it can be kept secret.

First test results

a. Drugs
First results announced afternoon of Monday, 1 September 1997
Blood, hair, eye fluid and organ tissues "examined under microscopes, shaken up in test tubes and centrifuges, and combined with chemical reagents..."

Hair analysis? For past drug use? Quite up-to-date--cutting edge, even.

b. Alcohol
Police lab analysis reported 1.87 g/l blood-alcohol; a private lab as a control reported 1.74 g/l.

Identical results, within the limits of experimental uncertainly. The second decimal place should not be reported, because the first decimal place is uncertain by plus or minus one in a good lab.

Alcohol testing is very standardized; these days it would almost surely be done by gas chromatography. The traces of drugs detected would probably have been measured by gas chromatography-mass spectrometry (GC/MS) which can be very good in capable hands or quite the opposite, but said traces of drugs are probably less important than the carbon monoxide.

c. Carbon Monoxide
The principal methods of testing for CO are:

1) A type of automated or semi-automated analyzer intended to measure CO in the blood of living subjects. (Commonly called a "co-oximeter" in the U.S.) Common in hospitals of areas where CO poisonings are frequent. Generally accepted as reliable for freshly and specially collected blood from living subjects; not acceptable for coroners' bloods.

2) Manual spectrophotometric methods where a blood sample is split, the aliquots are treated in two different ways, and spectrophotometric absorbence readings are taken at different wavelengths. The CO level is calculated in a rather intricate manner that involves a small difference between two large numbers. The chemical treatments (some of which are time-sensitive, despite what the textbooks say to the contrary), extensive manipulation and the inherent uncertainties in the measurements combined with the calculation render this method very questionable.

3) Liberation of the CO from the hemoglobin in a closed vessel and its consequent measurement. I know of one lab that measures the liberated CO with an infrared detector. The classical method is to use Conway diffusion cells, which allow one to liberate the CO and then take it up in a solution of palladium chloride, which is thereby decolorized in proportion to the amount of CO in the sample. The CO liberating solution should contain lead acetate to prevent liberation of hydrogen sulfide from putrefied bloods. As long as enough time is allowed for diffusion to take place, this is very reliable (caveat infra).

CAVEAT: Calculation of the carbon monoxide saturation requires measurement of the amount of hemoglobin in the sample tested. The "cyanmethemoglobin" method is the most accurate, and the manual cyanmethemoglobin method works well. However, the cyanmethemoglobin method is often used in automated blood cell counters (e.g., "Coulter" counters), and many people assume it is reliable for measuring hemoglobin in coroners' samples as well as in living subjects. This is not true in the case of blood samples containing significant amounts of carbon monoxide, because they need more time to react than the automated analyzers allow, thus giving falsely low hemoglobin readings and falsely high CO saturations.

CAVEAT #2: Standardization of most carboxyhemoglobin assays is difficult, perhaps unreliable at best. Only the diffusion dish method is readily standardized. Quality control is also not terribly easy. (I used to make my own quality control specimens by bubbling automobile exhaust fumes through a bottle of blood, hoping to escape notice in the parking lot! Since the introduction of low-pollution engines it has been more difficult to make controls by this method--one must search out an old car or use a lawn mower.)

If the CO results were duplicated at two different labs that would seem to vindicate the results UNLESS the two labs used the same analytical method. If they used significantly different methods AND obtained equivalent results, the CO concentration of the blood sample is quite likely correct.

On the other hand, as you say, CO dissipates quickly. If the 20% level is verified, this would mean that his CO level at 10 pm could have been much higher than 20%.

The CO level goes down quickly IF one is given 100% oxygen to breathe rather than air. In competition with air (21% oxygen), CO will win out. After death CO levels will be quite stable.

Second Test Results

10 September, an official communique from Stephan announced:

Unless one assumes group perjury or evidence tampering of incredible perfection, that would seem to pretty well settle the matter of blood alcohol levels.

The article said the original two blood alcohol tests were done on heart blood form M. Paul, and a third test was done on arterial blood. The question is, were the blood specimens collected at the same time or different times? If at different times, it is unlikely in the extreme that the test results are incorrect (at least with respect to alcohol). What is more likely, however, is that the heart blood and arterial blood were collected into different bottles at the same time. In such a case confusion is possible if more than one autopsy was performed by the same pathologist on the same day and the specimens were returned to the lab together.

Further blood tests revealed a "therapeutic level" of antidepressant fluoxetine (Prozac) and "infra-therapeutic" level of the tranquiliser tiapride. Tests on hair determined Paul took Prozac regularly since May and Tiapridal since July 1997.

I dare say these were probably tests developed specially for this case. (Actually, development of custom tests is something competent coroners' laboratories do on a fairly common basis. With the advent of practical gas chromatography-mass spectrometry in the early to mid 1980's, such custom test development became much faster and easier.)

Private Testing

17 September 1997 the private lab under Dr. Gilbert Pepin "mandated to do more sophisticated blood analysis" reported that Paul "had been in a state of moderate chronic alcoholism for a minimum of one week (the limit of the test's reliability)."

This strikes me as an absolutely extraordinary statement (perhaps "fantastic" would be a better adjective). Absent compelling evidence for the reliability of the testing (whatever it was) one must completely discount the statement, in my opinion.

How did CO get into Paul's blood?

In general:
The CO analysis was unreliable, OR there was significant leakage of CO into the passenger/driver compartment of the car. We cannot distinguish between these possibilities because (a) at this time we cannot estimate the probability that the CO analysis was properly performed by a reliable method and (b) the blood of the other people in the car was not tested (and even if it had been, the previous problem applies). Mr. Trevor-Jones' memory cannot be relied upon for symptoms of CO poisoning. If we were to find how the CO analyses were performed, we might be able to make a somewhat better estimate. The greater the procedural differences in the CO testing methods, the greater the probability that results in agreement are correct.

I can certainly understand the Princess of Wales' body being spared the indignities of an autopsy, but collection of a blood sample would have been a nice precaution. Still, it would probably have been considered an unreasonable precaution. It was an unfortunate miscalculation on the part of Dodi's family to not have a blood sample collected from him. The only other (admittedly tenuous) source of information that readily occurs would be someone who maintained or often drove the crashed vehicle, and who might remember something consistent with CO poisoning. I should expect the magistrates are way ahead of us on this, and of course conspiracy theorists would make the mechanic a part to the conspiracy.

One more thing rankles: WHY was a CO test done on M. Paul? I doubt it is regularly done on traffic accident victims (although I do not know French practice). There should have been some indication for it: observation by the pathologist conducting the autopsy of an unusual cherry-red color to the tissues, or observation of an unusual red color to the tissues in the bottles a day or two after the autopsy (when they normally appear brown).

A. Mickey Finn

There's no way a drink could be concocted which would produce a high level of CO in the blood. And even if there were, what would be the point? Good Mickey Finns exist without resorting to such awkward complexities, and as has been amply demonstrated, a high level of CO would start people asking questions. The only practical method of getting a lot of CO into the blood is to breathe it -- for some time.

B. CO and Airbags

"Another medical expert considering the suspected presence of carbon monoxide in the blood of driver Henri Paul concludes that the substance is due to inhalation of the airbag inflating gas."

Two problems with this idea: the decedent had no time to breathe the gas, and as I recall the airbags (at least in the U.S.) are inflated by the explosion of an azide, which decomposes to NITROGEN gas.

I have tested for the presence of carbon monoxide in the blood of people whose bodies were recovered from burnt automobiles ("crispy critters" in the jocular slang of U.S. medical examiners and their cohorts). I have usually talked with the pathologists who did the autopsies as well. If the decedent has a significant amount of CO in his/her blood this is taken as proof that he/she did not die instantly in the crash, but died (at least in part) from the burns, inhaling smoke while dying. This is supported by the presence of soot in the lungs. Note that organic materials burn to carbon dioxide in the presence of excess air; if air is limited they will burn (at least in part) to carbon monoxide. Note also that the pathologists would be expected to save samples of lung tissue so it could be examined many years later, but since there was no fire there would be no reason to find CO in the blood or soot in the lungs.

I think we can now junk the "airbag produced the CO that was found in HP's blood" idea, quite apart from the extraordinary difficulty, under the circumstances, of getting much CO into HP's blood even if pure CO was being sprayed into his face from a malignantly tampered-with airbag. Why can we junk the idea? Because:

a) the airbag under normal operation did not produce CO,
b) it was alleged to produce CO if it caught fire, but the description makes one strongly suspect the common confusion of CO and CO2,
c) and anyway, it *didn't* catch fire!

C. A leak in the exhaust pipe of the Mercedes S-280?

The French did not perform an autopsy of Diana - police report: "in accordance with instructions received."

TR-J survived, so his blood was not tested; and the body of Dodi was flown straight home, to London. We do not know if COHb was in their blood.

And evidence of a leaky passenger compartment would be destroyed in the accident. The likelihood of carbon monoxide poisoning in an auto is significantly less if the auto meets the sort of pollution control standards current in California, which are the strongest in the U.S., which in turn is (according to what I have on hearsay) quite a bit further down that particular road than the European countries. When people try to kill themselves by inhaling exhaust fumes piped into their cars, it takes at least several minutes to reach the kind of concentration found in HP's blood, a concentration, incidentally, which would have been fairly disabling. Also, if CO were in the passenger compartment of the car (or at least in the driver's compartment in the case of separate driver and passenger compartments) it couldn't have failed to be present in the blood of others in the car. All one would need to do to dissipate the CO would be to open the car's vents or turn on the air conditioner (which probably comes on when the defroster is used). If a reasonable person wants to bump someone off he would use a direct, reasonably certain method. But the operative word here is "reasonable." We must recall the Central Stupidity Agency's attempt to liquidate Castro with poison cigars! D. A switched blood sample? The mix of antidepressants, alcohol, and CO "found" in HP's blood is typical of a suicide by exhaust gas inhalation, who are frequently depressed, taking antidepressants, and typically have a few drinks before committing suicide. Since it is very hard to explain how HP could have walked and talked normally with his blood in that state, or even walked at all, the simplest explanation is that it wasn't his blood, but that of a car-exhaust suicide, of which it *is* typical.

That is certainly at least an academic possibility; I cannot venture an opinion on the probability. I should expect that because of the prominence of the case the pathologists would make a point of autopsying only that one case and then returning to their offices/laboratory with the specimens, but they might not. If they autopsied more than one case and returned to the lab with all the specimens in the same bag the chance of a labelling error increase, and can only be offset by a consistent practice of labelling the specimen containers AT THE TIMES THEY ARE FILLED. The pathologists' log books and the accession logs of the laboratory should show us what the case is.

I honestly can't imagine how anyone can seriously claim that this preposterously intoxicated blood sample came from HP. My guess is that the authorities know full well that someone must have switched the blood samples, simply because whatever fool did it stupidly picked a preposterously intoxicated sample, but they are stuck with accepting it now as really HP's blood because to question it opens the conspiracy can of worms.

Here's another possible thing to consider: was the decedent's blood typed by the lab? In my experience that is normally only done on bloods from homicide victims, and only the ABO type is readily detectable (if you are lucky); and since about 40-some percent of people (in the U.S., anyway) are type A and the same percentage type O, there is roughly a 50-50 chance a random blood will match the ABO type of any randomly selected decedent. If the blood submitted to the lab under M. Paul's name was not ABO-typed one is out of luck (unless one can resort to DNA testing or some such), but if it was one might get lucky. M. Paul's blood type should have been documented in his military records or at a blood bank or hospital somewhere; perhaps even in his hospital birth records.

"Quick, Jean-Paul, draw 200cc of boozed up blood while the Doc's gone to the toilet! Take that one in the corner, smells like a brewery, and hurry!"

Or--and this does not account for the purported CO--might the syringes have been washed/stored in alcohol and reused? You would think such a thing well-nigh impossible, but I know of a particular coroner's office where it happened.

I don't recall when M. Paul died, but if the weather were cold there is an increased possibility that the morgue might have in it the body of someone who died due to CO exposure from a faulty heater. We needn't postulate a suicide or a crispy critter.

E. Contaminated sample

"...given the violent nature of the death, the possibility of contamination of the blood is real and must be seriously considered."

He is probably thinking of the well-known phenomenon of a ruptured gut spilling alcohol into body cavities, where it might mix with blood from ruptured blood vessels. Sometimes such body cavity fluids are all one has to work with, but in this case there were also urine from an intact bladder and vitreous humor.

A post to the news group (alt.conspiracy.princess-diana) claims:
"At the request of HP's family, three separate forensic teams looked at HP's post-mortem. They consisted of Dr. Peter Vanezis, Regis Chair of Forensic Medicine at Glasgow University, a team from St. Georges Hospital in London, and a third team from Lausanne, Switzerland. All stated the tests done on HP were fraught with inconsistencies, incompetence, breaks in the chains of custody. Basically, they stated what they saw was one long train of negligence that made it impossible to conclude that HP was drunk that night."

It seems telling that three separate groups would say the same thing, but of course I do not know what information they based their conclusions upon. I have seen adversarial "experts" in court who made such statements (their large fees not influencing their judgment in the least, of course), and professional rivalries sometimes produce equally biased statements. Still, to not find one honest person out of three such would seem unlikely. But how to reasonably reconcile these statements with the multiple sample collections, the multiple tests on different body fluids, the reference laboratories? Conspiracy on such a grand scale should be well nigh impossible. And admit the possibility that even a chain of possession with uncertainties does not mean the specimens were incorrectly identified or substituted.

Fayed employed specialists who are very critical of the lab. tests.

But I should expect they were denied access to the labs' records and manuals.

This matter is an affair of state. I understand your assumption that your peers are honourable and no doubt you are right, but the question may arise as to what they honour most.

I don't rule out the dishonorable or incompetent, but neither do I assume it. I wasn't born yesterday, and I have seen some pretty incredible things.

Paul's burial

2 September 1997, Stephan issued a burial permit and banned cremation, to allow exhumation if needed.

Impressively cautious.

Could Henri Paul have been drunk without others knowing? Certainly possible.

[Aside: the extraordinary amount of soda found at M. Paul's apartment has troubled me. Closet alcoholics often consume highly unusual amounts of other beverages. Do you know anyone who keeps hundreds of cans of soda in stock at home?]

It has been my experience that people often consume the entire contents of a bottle of alcoholic beverage. Would there be in France a standard-sized container with as much alcohol in it as ten ounces (295 ml) 86-proof liquor? We might look for 300-ml bottles of 86-proof (43%) liquor or 250-ml bottles of 100-proof (50%) liquor. This hypothetical would result in a fairly steady blood alcohol level of about 0.15-0.17 g/dl on the job, which would imply constant behavior, and at a level where intoxication MIGHT not be evident to a casual observer, especially if the subject were a chronic drinker. A trained observer or a person familiar with the subject's non-intoxicated behavior would be expected to notice the intoxication.

What quantity of alcohol is needed to be consumed to account for the reported blood-alcohol level?
To summarize, we may start with the ASSUMPTION M. Paul was stone sober at the time he picked up his passengers at the airport, had an opportunity to drink unnoticed between 7 and 10 PM (although no empties were reported found at his apartment), had two drinks at the Ritz between 10 and midnight, and died at 0030 hrs with a blood alcohol of 0.18g/dl.

Making a few reasonable assumptions about metabolic rate, etc., I figure that we must account for another six average-size drinks (equivalent to 1.25 oz. 86-proof whiskey) for a total of eight PLUS another drink for each hour that passed between their consumption and 2330 hrs. If we hypothesize that the consumption took place between 1900 and 2200 hrs, we must add about three more drinks for a total of eleven. Running it through a computer program for calculating blood alcohol curves I see that I overestimated by about one drink .

Diana assasination exposed

Diana first Doctor

Diana killed by specialists

Diana Press Reports

Three more photographers detained

The People report

Diana was truly happy with Dodi

The Blood Test

Diana Finished Off

Interview With Diana


Diana Was Unconcious

Diana Doctors

Diana Pregnant?

Diana Under Surveillance


New Light

The Fiat Uno

Diana was killed by MI5

Henri Paul's Biography

Dianas last words

Man Arrested over Murder Claims

Medical Expert Opinions

Outtakes from BBC Diana Interview

Papparazi Reports

Diana's Engagement Ring


What Happened When

Diana was warned

Why Diana had to Die