Navigation
14 April 2010     Court of Arnhem
Acquittal in case Lucy de Berk

The DIGOXIN “EVIDENCE”

# "Expert" witness

Professor De Wolff indicated in the trial that there is "no doubt that digoxin played a role in the death of the patient. There was digoxin present that wasn't supposed to be present." He accuses others of quasi-scientific arguments.

De Wolff furthermore says he never suggested that Lucia de B. must have been the one to administer digoxin to the patient. It would be, according to him, a deadly sin for anyone in the court to make suggestions on her guilt or innocence in this instance.

When professor De Wolff says that there is no doubt possible that digoxine played a role in the death of Child A, he suggests that as a pharmacologist he was knowledgeable enough about the medical condition of the child and the situation before and during death. However, he himself publicly stated that he did not know that the heart was not contracted. Furthermore, other relevant information was not given to him or requested by him.

# Medical data

The patient concerned had a complex heart problem, decompensation cordis, pulmonary hypertention, alectase, a brain dysfunction, necrotising enterocolitis and other problems. The child had a heart operation 48 days before death and seemed to be improving slowly. But the problems remained and even seemed to increase in the final days. Diuretic medication and oxygen had to be increased. The day before death the child suffered from general malaise with severe stinking diaorrhea, bloating and problems with tube feeding. About 45 minutes before death the child was examined by doctors. The diagnosis of bowel infection was made and a potassium IV was inserted.

Lucia had connected up the monitor with a fingercup before the examination by the doctors. Breathing rate and pulse were elevated and quickly decreased shortly before death. Reanimation did not help. The paediatrician signed a death certificate and requested an autopsy.

The next day suspicions of Lucia did the rounds and it was decided that the death was "unnatural". The municipal coroner was called and the judicial investigation starts.

# Diagnostics in retrospect

The clinical data mentioned above seems to have been kept in the background in the case of baby A. Even when the hospital and NFI autopsy and the toxicological tests showed no suspicious results the public prosecutor continued to consider it a case of deliberate poisoning. "After all, it couldn't have been a coincidence".

During the police investigation potassium poisoning was at first suspected. A high concentration of potassium was found in the eye fluid. However, this is not considered reliable because of the relatively long time between death and NFI autopsy.

Almost a year after death a police officer produced a jar with gauzes for the investigation. These gauzes were retrieved from the abdominal cavity at the NFI autopsy. Their origin is unclear; especially as pathologist Spaander of the first autopsy said he doesn't use gauzes. Professor Uges immediately stated that no valid tests can be done on these gauzes. The gauzes contained a bloody fluid, not blood as is often said. The gauzes were frozen and defrosted for tests four times over a period of 3 years.

When testing the gauzes two similar (immuno-assay) tests showed digoxine concentrations of 22 and 25 microgram/liter and the HPLC-method showed 7 microgram/liter. The HPLC method is the only method that is selective enough in this situation for the correct digoxin determination. The Court chose the average of both immuno-assays because they were similar. The other (more reliable) test result was ignored. According to professor De Wolff a deadly concentration is more than 15 microgram/liter.

In a (chronic) digoxin poisoning the heart film shows a "broad heart complex". The junior doctor made a drawing of the monitor display that he saw shortly before death. This supposedly shows a broad heart complex. On this broad heart complex and the gauzes the foundation of for the digoxin evidence was laid.

The Court eventually stated that Lucia must have given the child a deadly digoxin injection about an hour-and-a-half before death via the IV valve.

There are no traces of digoxin in the IV system, or other remains of this act. Nobody saw Lucia give an injection.

# Discussion between expert witnesses

The statement where Professor De Wolff indicated that he "did not suggest that Lucia was the one to administer digoxin by injection" is entirely unbelievable. cf:

On February 5 2004, during the court session De Wolff states:

the fact that Emit 2000 and the IMx-assay measured similar digoxin concentrations is on its own enough [for him] to say that that makes it 100% certain... that the high concentration [of digoxin in the blood]... is not compatible with normal life.

In his report of March 16 2004 De Wolff writes:

From the answer to question 2 is should be clear the that concentrations measured post mortem in blood and liver ... can only be explained by an acute high dosis of digoxin.

On May 11 2004 during the court session De Wolff states:

the [clinical] presentation at the patients death is compatible with an acute overdose of digoxin. This could be considered proof in a scientific sense.

Professor De Wolff made these statements on the basis of test results from the Emit 2000 and IMx, both so-called immuno-assays, which are not selective enough for this test. He left the much lower results from the HPLC-MS test out of his consideration, whilst also indicating himself that this was the "gold standard method".

# Dasgupta, who has published a great many articles about digoxin, writes:

"Scientifically you cannot rely on those Emit 2000 and IMx tests. You must use the 7 μg/L of the HPLC-MS method."

In general immunoassay for digoxin is subject to many interferences including DLIS, while more sophisticated analytical technique such as HPLC/MS is free from such interference. This is because digoxin molecule is identified by its mass spectral characteristics which is also the fingerprint of the molecule. In an American court of law it is most likely the Judge and Juries would be extremely concerned regarding the discrepancy between digoxin results obtained by the Gold Standard, HPLC/MS and two immunoassays. Many references in the scientific literature including research by our group for last 18 years clearly show that both EMIT 2000 and IMX digoxin are subjected to DLIS and other interferences. Moreover, HPLC/MS where an extraction is necessary prior to analysis also eliminates any potential matrix effect where immunoassays are affected by matrix other than serum or plasma and hence may explain the discrepancy.

In other words, a real scientific expert should use the 7μg/l from the HPLC-MS method and not the higher results from the Emit 2000 and the IMx assay as De Wolff does.

# Expert Koren writes:

A 48hrs postmortem level of 7 ng/ml by HPLC can well be within the therapeutic range during life, because postmortem redistribution can be different to the order of several fold.
After 48hrs the elevation can be much higher than after 24hrs, because more digoxin is released from tissues (where it was in high concentrations) into the blood (where concentrations are low). Also, this represents an AVERAGE from many observations, so the elevation in particular case can be much higher.
The post mortem level in this case, based on the HPLC method you mentioned, could well be within the therapeutic range during life. If the verdict of murder was based on this level, there is a risk of major injustice and terrible violation of human rights here.

The post mortem distribution mentioned by Koren does not appear in the arrest. There is a casual reference to possible increase of concentration by evaporation and contamination. The process of distribution which occurs after death and through which digoxin is released from tissues is not mentioned. In the first 24 hours after death therefore 5 microgram/liter (approximately) should be subtracted from the 7. After 48 hours this number is "much higher". With this reasoning, only a small amount of digoxin is left. Not toxic, not even therapeutic.

There are very few scientifically based facts known on the secretion of digoxin. Recent Polish research indicates that the concentration of digoxin in tissue is not 30x higher than in blood, but that this can also be 100x or more in blood.

In the article "Forensic science in the dock" in the British Medical Journal (BMJ 2004;329;636-637) Drummer, Forrest, Goldberger and Karch write:

...we have long known that blood sampled from the heart of a dead person who had been on long term digoxin treatment may contain a seemingly toxic concentration of digoxin when, in fact, the actual blood concentration immediately before death was the appropriate non-toxic therapeutic concentration. Even if it could be shown that blood concentrations after death were the same as concentrations at the time of death, which blood sample should be used? Drug concentrations are likely to have changed after death...

If the blood concentration at the time of death cannot be known with certainty, then how is it possible to extrapolate the time and amount of drug ingested before death? The simple answer is that such extrapolations are prone to considerable error and generally should be viewed as unreliable and not evidence based. Despite these limitations, such calculations are frequently and wrongly produced during court proceedings, even though the problems we outline have been widely known for many years...

The international experts on digoxin, such as G. Koren and A. Dasgupta find the investigation unacceptable and find it "quite shocking" that Lucia was convicted on the grounds of these incorrect conclusions. Professor De Wolff states that he discussed the case with "several experienced" colleagues thereby illustrating his conscientious method. Dasgupta and Koren's criticism did not change his opinion.

# Lab result from Straatsburg

At the time of the investigation everyone around him was convinced that Lucia had "done it" and should be convicted. And hopes for proof of her guilt were firmly pinned on the digoxin test. In April of 2004 Lushof literally asked for a miracle to determin digoxin poisoning in a letter to the laboratory in Strasbourg. In that very same letter Lusthof indicates that the gauze-material was not suited for testing and that the reliability of a test was influenced by the process of freezing and defrosting the gauzes.

That the court did not wait for the French test results of the "miracle" requested and convicted Lucia in June of 2004 using digoxin poisoning as the main evidence - followed by a chain of "analog" facts - is also a miracle.

De Wolff and Lusthof stated in court in May of 2004 that the data from Strasbourg probably wouldn't be all that relevant anyway.

There are many more strange aspects to the performance of both experts. Both indicated that the HPLC-ms method was the most reliable for digoxin. That they still used the average result of the two immuno-assays when this test showed a result that is too low to indicate poisoning poisoning - 7 μg/L in bloody fluid and 0 μg/L in the organs - was strange to say the least. After all, the "nice similarities between these tests" and the "frequent good use in hospitals" (expert quotes) have nothing to do with the exact scientific evidence gathering on the digoxin levels in these circumstances. De Wolff knew, or should have known, that the immuno-assays were not reliable for this specific use. In a baby and in diseases of the heart muscle, digoxin-like immunoreactive substance (DLIS) is present in the blood and immuno-assays cannot distinguish sufficiently between this DLIS and digoxin. The "good use" on a normal man or woman with digoxin therapy says nothing about the quantitative determination in a baby with a heart muscle disease.

Whereas in 2002 everyone in Court agreed that the time between injection and death was estimated to be roughly between 15 minutes to half an hour, in the high court Lusthof and De Wolff stated that the time between injection and death must be one to one-and-a-half hours. On the basis of this timing and the data in the trend-tables the court reconstructed the time which they said pointed at Lucia as the person administering the injection. That the trend-tables served as evidence and the more accurate trend-graphs - that showed clearly that on the time in question a medical treatment was taking place - were ignored, was also accepted by the experts during their reconstruction. The pharmacologists also implicitly pointed at Lucia as the guilty party.

# Questions posed by the treatment of the digoxin evidence

Questions and problems posed by the treatment of the digoxin evidence are summarised below:

  • Why was there a change in the investigation a year after the death and why was the attention focused on digoxin then?
  • When and by whom was the ECG made and why was it not noted?
  • Why is there no ECG print? Bearing in mind that an autopsy is routinely performed for deaths after heart operations it is strange that there is no ECG. However there is a strange drawing from a junior doctor who was on his first shift and said he didn't know anything about cardiology at the time.
  • Why was De Wolff not clearly told about the decompensation cordis, the right ventricle hypertrophia, the pulmonary hyptertension and the necrotising entero-collitis, the atelectas, the brain anomaly and other symptoms? In view of this, why did the information provided to him make as its focus the successful operation and the imminent return home of the child?
  • Why was there no attention given to the necessity to increase oxygen support and diuretic medication in the last days before death?
  • Why did De Wolff not receive the first autopsy report which clearly states that the heart was not contracted? (In the second autopsy report that was given to De Wolff this is no longer an issue due to the lapse of time and the actions taken).
  • The question about the origin of the gauzes, which were found in the body during the second autopsy was never answered sufficiently, not in the Court and not to De Wolff either. Why was this acceptable?
  • The gauzes probably did not contain blood, but the (bloody) fluid from the abdomen. However, they are always referred to as blood.
  • The material already contained alcohol due to the start of decomposition; how much influence does this have on the determination of the levels of digoxin (which is more soluble in alcohol than in water)?
  • The DNA tests made to answer the "correct origin" of the tissue indicate (of course) that the DNA patterns in the gauzes resembled those of A after being in the body. However this says nothing about the origin and the way they got there.
  • The temperature under which the material was stored was 20ºC over a period of a three year investigation. Nowadays it is believed necessary to keep storage temperatures lower.
  • The few drops (!) of fluid extracted from the gauzes were probably frozen and defrosted at least 4 times and this freezing and defrosting could easily have changed the test results which are therefore not scientifically reliable.
  • The process of post mortem distribution is mentioned as an aside in the criminal trial; it shows an elevation of 5 microgram/l in the first 24 hours! - with clear plus and minus margins.
  • The immuno-assays should not have been used as determination of the level of digoxin due to the known facts about DLIS.
  • The adamant denial of De Wolff on the possible remnants of digoxin in the body as a result of chronic administration until the day of the operation in July cannot be substantiated with hard fact. There is very little known about the half-life and "behaviour" of digoxin in the various organ systems. International research has shown that digoxin can be present in organs in high concentrations.
  • The negative liver test at the NFI and in Strasbourg excludes an acute administration of 1 to 1.5 hours before death, as calculated by the court.
  • Why did no one investigate the expected effects of dixogin administered 1 to 1.5 hours before death on the heart?
  • It was clear to De Wolff cs that Lucia herself (out of concern) had connected up a monitor - with finger contact! - even before the doctors where consulted. This registered heart rate, breathing rate, oxygen tension and saturation. But not an ECG.
  • Why was a medical error excluded, while after all publications we should know better?
  • Why were the trend-table and the trend-graph not studied more in depth? It would have been clear then that the time line used by the court was incorrect, or would have immediately pointed to a digoxin injection administered by a third party.

Hopefully the parties involved and the experts have the courage to ask these questions of themselves and answer them honestly. Why did things transpire in this case as they did? And why is Lucia in jail for Christmas for the seventh year in a row?

# The doctor, death and justice

On March 30 'Medisch Contact' (Medical Contact) published an article about "the doctor, death and justice" by physician-lawyer Dirk van der Wedden. He states that the criteria the court used in the evidence gathering could have far reaching consequences for medical practice. In these criteria it concerns:

  1. a sudden death,
  2. that is medically inexplicable,
  3. in the presence of a suspect.

Using real life examples Van der Wedden shows that none of these criteria are tenable. Many sudden deaths are natural, whilst unnatural deaths can happen over a period of time. The term "medically inexplicable" supposes more knowledge from doctors about causes of death than is realistic. An autopsy does not always find a reason for a sudden and unexpected death. The case of the polonium poisoning in London where the perpetrator was long gone also shows that the third critera - in the presence of a suspect - is not valid.