Investigations into the circumstances surrounding the death of David Kelly have led to the exchange of acrimonious views including allegations of conspiracy and murder. David Kelly, a government scientist and weapons expert, committed suicide by cutting his wrist and taking painkillers after he was identified in newspapers as the man the UK government believed was the source for a BBC report on Iraq. Impetus for the debate stems mainly from conflicting views about the cause of death, including issues that relate to postmortem toxicology results and their interpretation. Controversy occurs from the mistaken notion that postmortem laboratory measurements, taken in isolation, can be interpreted effectively.

The current controversy illustrates some universally held, but mistaken, notions about the process of death investigation in the United Kingdom and elsewhere. Many assume that forensic pathology is as evidence based as other branches of medicine. This assumption is not accurate.

In the course of caring for living patients, doctors who interpret hospital laboratory tests know, or can quickly find out, the “normal” value for any particular drug. But most doctors (as well as the general public) would be surprised to learn that there are few if any “normals” in postmortem toxicology. Non-circulating blood after death is not the same thing as circulating blood before death, and evidence that the concepts of normal or therapeutic drug concentrations can be applied to blood from dead bodies is severely lacking.

Even in living bodies, interpretation of a single blood concentration measurement is impossible without considering route of administration, number of doses taken, and the amount of drug actually in the body. Such information is almost never available to investigators, making it impossible to determine the cause of death solely by comparing a single postmortem drug concentration measurement with a range of published values, originally derived from measurements made in living people. With chronic use, tolerance occurs, and tolerance cannot be measured or estimated after death. Healthy patients enrolled in methadone maintenance programmes, for example, may have blood methadone concentrations in excess of other, non-tolerant methadone users examined on the autopsy table.¹ Similarly, 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.²

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.³ For many drugs, including those found in David Kelly, concentrations may increase by as much as 10-fold.⁴ Furthermore, drug concentrations in blood samples from cadavers are site dependent, higher in some locations and lower in others.⁵ Should the site yielding the lowest or highest result be used? Or should an average value for three sites be used? Nobody knows because the process has never been studied systematically.

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.⁶ 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.

Postmortem measurements of drug concentration in blood have scant meaning except in the context of medical history, the sequence and circumstances surrounding death, and necropsy findings. The paucity of evidence based science, coupled with the pretence that such science exists in regard to postmortem toxicology, leads to the abuse of process, almost certainly to the miscarriage of justice, and possibly even to false perceptions of conspiracy and cover up.

We have written this editorial partly because of the Kelly matter, where the central issue concerned the interpretation of the toxicology results. Death investigation and forensic pathology are also not immune to misinterpretation. Poor or inadequate death investigation and incomplete or misinterpreted forensic pathology studies may also result in wrong conclusions. All aspects of the medicolegal death investigation triad—investigation (history), pathology, and laboratory results—are essential and must be evaluated in context with one another. We have formed an ad hoc group to address this issue. A detailed analysis of the problem with suggestions for reform is in preparation.