Classical Mistakes in Forensic Toxicology Made by Forensic Pathologists

Introduction

One key difference between a routine hospital-based autopsy and a forensic autopsy is toxico-logical analysis. In forensic death investigation, toxicological results are important for a variety of reasons. They may help determine the cause, contributing cause, or manner of death and also may be useful for legal matters in civil litigation and criminal prosecutions. Therefore, the proper interpretation of these results is an important duty of the forensic pathologist.

In 1956, Dr. Alan R. Moritz gave the Ward Burdick Award address at the 35th annual meeting of the American Society of Clinical Pathologists in Chicago. His address, entitled “Classical mistakes in forensic pathology,” was subsequently published and remains a masterwork in the field of forensic pathology for the newly trained as well as the experienced pathologist (1). It describes errors in autopsy practice, interpretations, and forensic reasoning. In one section he discusses “mistakes of not exercising good judgment in the taking and handling of specimens for toxicological examination” (1). These include contamination of specimens, inadequate samples, poorly selected specimens, and lack of adequate clinical information for the forensic toxicologist. The growth of technology and research in the field of forensic toxicology during the subsequent 56 years has added new issues and problems. This review will examine them with a focus on the interpretation of postmortem toxicology results.

Mistake of Determining the Cause of Death Based Solely on the Drug Concentration

One of the greatest mistakes that a forensic pathologist can make with toxicology is opining the cause of death based solely on a drug concentration without consideration of the entire case (2–9). If the cause of death could be determined just from the toxicology results, forensic toxicologists, not forensic pathologists, would determine the cause of death. There is a need for a physician with medical and forensic training to interpret the toxicology results in the context of the entire case. This also highlights the myth of the fatal concentration.

There are a variety of pharmacology and toxicology references that provide “toxic” or “fatal” doses or concentrations (10, 11). As Adelson noted, the term “fatal” dose has three meanings that are used “indiscriminately and interchangeably” in the scientific literature (12). They are: 1) the minimum dose that always causes death, 2) the minimum dose that generally or usually causes death, and 3) the minimum dose known to cause death (12). These variations hinder the value of the concept of a “fatal” concentration. Most experienced forensic pathologists have seen people die from mechanical violence (e.g., gunshot wound) with blood cocaine concentrations ten times greater than others whose deaths were clearly attributable to cocaine intoxication. Recreational concentrations of many drugs overlap so-called “fatal” concentrations. Few drugs of abuse have a predictable dose-response curve and user issues are key (e.g., habituation/tolerance, cardiovascular health). In addition, studies that list ranges of drug concentrations in fatal intoxications may fail to consider other factors, such as how many were instances of multidrug intoxications, survival intervals, decomposition, tolerance, methodologies, sample site, etc. These caveats also apply to attempts to calculate the precise antemortem dose based upon the body weight, volume of distribution, and postmortem blood concentration since this formula is intended for living patients. As long as one is aware of the pitfalls, postmortem drug concentrations may be useful in certain circumstances. Forensic judgment, however, must be exercised with these interpretations.

The proper determination of the cause of death due to an intoxication requires three factors. First, the autopsy fails to disclose a disease or physical injury whose extent or severity is inconsistent with continued life. Second, the toxicology results are in the range typically encountered in such deaths. And third, the history and circum-stances are consistent with a fatal intoxication. The first criterion raises the issue of the performance of an autopsy in suspected intoxication deaths. Since in most instances, there is no anatomic or microscopic evidence of an intoxication death, one may ask why an autopsy is needed. Intoxication deaths, however, are in most instances a diagnosis of exclusion. Performing toxicologic analysis on specimens collected from a decedent who does not undergo an autopsy is not a universally accepted practice among forensic pathologists (13). Some offices perform toxicology on all deaths, while others only do it on those that undergo autopsy or a subset (13).

Autopsy standards by the National Association of Medical Examiners (NAME) state that an autopsy should be done if the death is “by apparent intoxication by alcohol, drugs, or poison” (14). In certain situations, a limited approach may be justified. For example, a 55-year-old man with well-documented advanced coronary disease is witnessed to clutch his chest and collapse. The decedent has a remote history of substance abuse. The family objects to an autopsy but agrees to toxicology. If the toxicology subsequently detects cocaine, the inclusion of cocaine with the heart disease makes pathophysiologic sense. But what if the same patient is instead found dead in bed and the subsequent toxicology detects oxycodone not cocaine? Instead of a substance abuse history, he has a poorly-defined chronic pain syndrome. Did he die from heart disease, an intoxication, or a combination? An autopsy may not reveal the answer in all instances but it may provide additional information to consider. For example, a ruptured myocardial infarct or coronary artery thrombus is a compelling finding and would trump the opioid intoxication diagnosis (13). The benefits of the autopsy include the documentation of the extent of natural disease and the ability to obtain a variety of quality toxicological specimens. It also provides a stronger foundation for the ultimate conclusion of the cause of death.

Some may argue that in a particular case, if one decides that an autopsy is not needed to certify the death, then toxicology testing also need not be performed. That is, if the pathologist is convinced that a death is due to disease why spend resources on a test that will provide a result when it is too late to perform an autopsy. If there is a concern for substance abuse, then an autopsy is appropriate. The risk of performing toxicological analyses on “external only” examinations, is the delayed dilemma of how to interpret an opiate result days or weeks later when an autopsy is no longer an option. Without other compelling clinical information about the death, one may not be able to interpret the toxicology result. Toxicological data are no substitute for an entire case study and the exercise of medical judgment in the evaluation of death.

Rapid toxicology screening tests that can be performed on urine at the autopsy table are available. Use of these in instances when an autopsy is performed may allow some natural deaths to be certified without having to wait for the formal toxicologic analysis. Therefore, some additional natural deaths can be certified on the day of the autopsy without having to pend the certification until the toxicology is finalized. For example, consider our previously discussed 55-year-old man with a history of coronary disease and remote substance abuse who is witnessed to collapse. The autopsy demonstrates advanced coronary artery disease. The history of the sudden collapse with coronary disease and a negative rapid screening test, may be sufficient to reassure the forensic pathologist that this is a natural death that can be certified that day. This provides the next-of-kin with an immediate final death certificate that can be used to resolve insurance issues, etc. Due to the limitations of these rapid tests (see below), however, they are not a substitute for formal postmortem toxicologic analysis since these rapid tests may include instances of false positive or false negative results. If the rapid screening test is positive, the toxicological interpretation and death certification is deferred until the formal toxicologic analysis is completed. In rare instances, the death certificate may need to be amended if the subsequent toxicologic analysis demonstrates that the rapid screening test was a false negative.

Failure to Consider the Postmortem Nature of the Specimen when Interpreting the Result

Postmortem alterations result in qualitative as well as quantitative changes to toxicological findings (15–20). That is, depending upon the postmortem interval and other factors, the detection and quantitation of a substance after death can be different from the premortem (just before death) analysis. After death, some toxicological substances may disappear while others may be produced such as ethanol by colonic bacteria (21–23). In one study, blood alcohol concentrations of over 0.2 gm% were documented in decedents who had not ingested ethanol before death (21). Analysis of the vitreous or urine may assist to distinguish postmortem ethanol production from premortem ingestion (21). Cocaine may undergo spontaneous hydrolysis to benzoylecgonine which is minimized by the addition of a preservative (such as potassium fluoride) and specimen refrigeration (2).

Qualitative detection of certain substances (e.g., ethylene glycol, arsenic) may be sufficient for the interpretation of their role in the cause of death regardless of the concentration. Since these poisons are without abuse or clinical-use potential, their detection alone without quantitation is usu-ally sufficient. But determination of the role of medications or drugs of abuse in the cause of death is more complex. This is particularly true when the decedent has advanced underlying disease that is capable of explaining the death. In these instances, the forensic pathologist considers the concentration and the context of the entire case. Questions as to whether the postmortem concentration can be used reliably to help determine the cause of death raise the issue of how postmortem concentrations are related to ante-mortem concentrations.

Studies of fentanyl fatalities have examined postmortem and antemortem concentrations (24, 25). Postmortem femoral blood concentrations from 118 decedents with therapeutic use of fentanyl were compared to serum concentrations of 27 living patients with therapeutic fentanyl patches. The researchers reported postmortem fentanyl blood concentrations that were on average up to nine times higher than in vivo serum concentrations at the same dose and concluded that postmortem blood concentrations cannot be directly compared with in vivo serum concentrations (24). Another study demonstrated that post-mortem fentanyl concentrations increased with increasing postmortem interval even in femoral blood (26). In this study, postmortem femoral blood was collected from seven decedents at two postmortem intervals, shortly after death (FB1) and at autopsy (FB2) with mean collection times of 4.0 and 21.6 hours, respectively. The fentanyl concentrations for FB1 and FB2 ranged from undetectable to 14.6 μg/L (mean, 4.6 μg/L) and 2.0 to 52.5 μg/L (mean, 17.3 μg/L), respectively.

Postmortem redistribution was first described in 1975 by Holt and Benstead who identified different concentrations of digoxin in samples of blood collected in three different sites: heart, neck, and leg (27). They recommended that postmortem blood should be collected from the leg veins if assessment of antemortem digitalis toxicity is needed. In 1985, Koren and MacLeod used the term “postmortem redistribution” in a study of digoxin concentrations in rats examined at the time of death and 12 hours after death (28). Post-mortem redistribution of tricyclic-antidepressant medications also has been well studied (29, 30).

During life, drug concentrations are essentially the same regardless of location of the blood draw due to the continuous circulation of blood. Therefore, after death something must change to account for these variations in concentrations by sample site. After death there are a variety of changes that occur including a decrease in blood pH and increased permeability of membranes. Due to these factors, drug concentrations in blood do not remain static and may shift along concentration gradients. Postmortem redistribution includes both movements in the vasculature that may result in site specific differences but also shifts of drugs between tissues and drug reservoirs (31). The redistribution of drugs may not be solely due to drug gradients between tissues because some drugs with a large volume of distribution (e.g., cannabinoids) have not shown consistent increases in blood concentrations after death (5).

Drug distribution in the body may be altered due to transfer of drugs across barriers that normally maintain a concentration gradient during life (32). Diffusion through previously impermeable barriers are a particular issue with small nonpolar molecules (33). Drug concentrations can be altered by physiochemical change in the postmortem environment such as alterations in pH which may change the ionized state of the substance (32). For example, free morphine concentrations may increase due to postmortem hydrolysis of the conjugated metabolites (34). Postmortem redistribution of drugs occurs as a result of diffusion of a drug from a higher concentration to a lower one following disruption of cellular membranes. Drugs with high lipid solubility or high tissue concentrations (i.e., large volume of distribution) are particularly susceptible. In addition, other factors can affect site to site variability. These include size of the blood sample, postmortem interval, sample type (whole blood vs. plasma), volume of distribution, tolerance, protein-binding, and postmortem stability of the drugs (7, 35–37). The volume of a peripheral blood sample should not be too large as it may include blood permeating from central sites. Some recommend ligation or compression of the proximal femoral/iliac vessels to prevent contamination of the sample with central blood.

Failure to Communicate with the Toxicology Laboratory and Understand their Limitations

The forensic pathologist must not treat the toxicology laboratory as a black box to which specimens are sent and results are returned. The more information that is provided to the toxicology laboratory, the better the result will be. No toxicology laboratory routinely tests for every drug and poison. Toxicology laboratories typically use screening tests for commonly abused drugs, alcohols, and medications. One needs to inform them of specific drugs of concern (e.g., what drugs were found at the scene). As Moritz noted: “Too often the toxicologist receives specimens with no information, but only a request that he test for a poison” (1). The forensic pathologist must inform the forensic toxicologist of the circumstances. Let the toxicologist know if the decedent was a depressed jeweler who was found dead in his apartment with an empty cup next to him. Otherwise, the laboratory may not test for cyanide. Let the toxicologist know if the body had been embalmed prior to the autopsy. Foren-sic pathologists should have a general knowledge of what their toxicology laboratory routinely tests for so they understand what may or may not be detected on routine analysis. One should not conclude that because the toxicology was “negative” that all drugs and poisons were excluded.

This is a two-way communication. The toxicology laboratory also has a responsibility to provide detailed results. Accreditation of a laboratory may assure that the results are reported appropriately. But not all forensic toxicology laboratories are accredited. Accrediting groups such as the American Society of Crime Laboratory Directors (ASCLD) and the American Board of Forensic Toxicology (ABFT) require protocols for re-cording, reviewing, and reporting results. These accreditations are largely voluntary and only a few jurisdictions (New York, Texas, Oklahoma) require ABFT or ASCLD accreditation although other States are developing regulations. In addition, the National Institute of Justice has funded forensic scientific working groups (e.g., Scientific Working Group for Forensic Toxicology) to develop and disseminate consensus standards for various forensic practices including forensic toxicology. It should not be acceptable for a toxicology report simply to state: “Blood: no drugs detected.” Minimal report information typically includes the specimen (e.g., blood), method of analysis (e.g., gas chromatography/mass spectrometry), the substance that was tested for (e.g., fentanyl), and the result (7). Although forensic pathologists may not have direct control of this, they need to recognize this inadequacy and seek a more detailed report. Forensic pathology training programs need to provide fellows with adequate exposure to forensic toxicology, and forensic pathologists need to pursue continuing education about developments in forensic toxicology that may affect their practice.

Failure to Collect Proper Specimens Including Hospital Specimens

Just because a specimen is collected, it does not mean that it must be tested. For example, even though stomach contents may be collected at each autopsy, it may not be analyzed in every instance. Providing additional samples to the forensic toxicologist will maximize their testing ability. It is better for the forensic toxicologist to ultimately discard unused samples than for the forensic pathologist to not collect them. Storage limitations may be problematic for some offices. Analysis of stomach contents, however, may help to distinguish a suicidal ingestion from an accident. Since a suicidal history is not always known or suspected at the time of the autopsy, the benefit of collecting and retaining a specimen that can be tested later is invaluable. Pill fragments are not always grossly visible in the gastric contents with a suicidal ingestion. This analysis may make the difference between an accident, suicide, and undetermined manners of death.

Instances in which there is a prolonged survival interval may make toxicological analysis of autopsy specimens less useful due to continued metabolism during the agonal period or hospitalization. The extent of intoxication at the time of injury is usually what is of forensic interest. Therefore, obtaining residual hospital-admission-blood samples may be useful. The American Association of Blood Banks (AABB) recommends that a cross-match specimen is retained for seven days, however, some laboratories will keep it for 10 or 14 days. Care must be taken in noting the acquired hospital specimen type (e.g., serum vs. plasma vs. whole blood) since drug concentrations will vary among them. For example, the concentration of ethanol in plasma will be approximately 1.2 times higher than the whole blood sample. Since forensic pathologists typically interpret results from a single point in time, it may not be possible to know when the decedent began the post-absorptive phase (down-ward portion of the ethanol concentration curve). Therefore, a postmortem ethanol concentration following a survival interval of two hours may equate to a concentration that was the same, lower, or higher at the time of the injury depending upon where the decedent was on the ethanol curve at the time of the injury. In some instances, the vitreous ethanol concentration can assist with this interpretation (8). At autopsy, collection of subdural hemorrhage and bile also can help with these delayed deaths (38, 39).

Many hospitals have the ability to perform limited in-house toxicology analysis. This may include a urine screen for common drugs of abuse, a blood alcohol, and salicylate/acetaminophen blood analysis. Forensic toxicology, however, is different from hospital-based toxicology. Forensic toxicology, because it is often used in a legal arena, has a more rigorous methodology and confirmation requirements. For example, a forensic toxicology laboratory will confirm a result either by a different test methodology on the same specimen or apply the same methodology on a different specimen. Hospital laboratories usually do not. So, how reliable are the results from an unconfirmed hospital urine toxicology rapid screening test? These tests may detect only metabolites (e.g., benzoylecgonine) and yet the result may be reported as the parent drug (e.g., cocaine). This may be mistakenly interpreted as evidence of an acute intoxication which is not justified based on this one test. The lack of a second confirmatory test also is a problem forensically. This is one more reason to obtain the hospital's samples to allow for more rigorous forensic toxicologic analysis.

Including all Intoxications on the Death Certificate Even if They DID Not Make a Pathologic Contribution to Death

Some Medical Examiners/Coroners include the toxicology results on the death certificate as a contributing condition even when the person died of mechanical injury (i.e., gunshot wound, blunt force from a motor vehicle collision) and not an intoxication. While these intoxications may partly help explain the circumstances of why a driver crashed his/her car, the intoxications do not play a pathologic role in the death (40). That is, the person would have died from the mechanical violence with or without the intoxication. Except in exceedingly rare situations (e.g., an adult who drowns in a bathtub which does not make sense without an incapacitating physical or chemical injury), an intoxication should not be included as a cause or contributing condition to explain why a person sustained the lethal injury. To include these results on the death certificate is speculative, potentially misleading, and arbitrary. Both intoxicated and non-intoxicated drivers are involved in motor vehicle collisions. The contributing cause of death section of the death certificate clearly states that it is for conditions contributing to death but not resulting in the underlying cause of death. The toxicology results are part of the autopsy file and are readily accessible for any legal matters.

Failure to Identify and then Properly Certify an Intoxication Death

There has been a dramatic increase in the number of intoxication deaths in the United States in the last decade. In 2008, poisonings (by drugs and toxins) became the leading cause of injury death in U.S. even surpassing motor vehicle collision deaths (41,000 vs. 38,000) (41). Of these poisoning deaths, 89% were due to drugs, particularly opioid analgesics. The proper identification and certification of these deaths is an important public health issue and involves four steps. The first is the initial identification of a suspected intoxication death which typically occurs during the initial scene and history investigation (42). The second is the identification of intoxicants which is a combination of the pathologist collecting and submitting the appropriate samples and the toxicology laboratory performing and reporting the appropriate analysis (43). The third is the proper certification of the cause of death with consideration of the complexity of postmortem toxicological interpretation. A position paper on the proper certification of cocaine intoxication fatalities has been published in order to improve consistency (44). The fourth is the proper terminology on the death certificate. Standard, explicit terminology benefits record tracking by public health agencies. Listing the drugs (i.e., acute intoxication due to the combined effects of oxycodone, fentanyl, and diazepam) provides more useful information than simply writing “multidrug intoxication.” Since it is often not possible to tease out individual drugs’ roles, it is customary to include all of the relevant drugs in the cause of death.

Conclusion

This review describes potential errors made by the forensic pathologist with regard to toxicology analysis encountered with death investigation. These include incorrectly determining the cause of death based solely on a drug concentration and/or failing to consider the postmortem nature of the specimen when interpreting results. Proper communication with the toxicology laboratory, an understanding of analytical limitations, and the proper collection of specimens are duties of the forensic pathologist. The forensic pathologist interprets the toxicology results in the setting of the entire death investigation which requires consideration of the decedent's medical history, the circumstances surrounding death, the environment of the death, the autopsy findings, and the results of the analytical toxicology. Toxicologic results should be included on the death certificate only when they make a pathologic contribution to death.