Which Poison Cannot Be Detected in Post-Mortem?

No poison is truly undetectable in every circumstance, but several come remarkably close. Potassium chloride, inert gases like helium, the anesthetic succinylcholine, and the party drug GHB all share a common trait: they either mimic substances the body produces naturally, break down too fast to catch, or simply do not show up on standard toxicology screens. The gap between what forensic science can theoretically identify and what it routinely does identify is wider than most people realize, and that gap is where poisons go unnoticed.

Potassium Chloride and the Problem of Normality

If you had to pick one substance that best fits the label “undetectable post-mortem poison,” potassium chloride is the strongest candidate. Injected intravenously in sufficient quantity, it stops the heart almost immediately by disrupting its electrical rhythm. The problem for forensic investigators is that potassium floods out of every cell in the body as soon as death occurs. Cell membranes lose their integrity, and intracellular potassium, which is normally many times higher than the level in blood, pours into surrounding fluids. By the time an autopsy is performed, potassium levels in blood and even in vitreous humor (the fluid inside the eye, which is more resistant to post-mortem change) are elevated regardless of the cause of death.

A study examining fatal intravenous potassium injections found that even when potassium measurements were taken from several different biological samples of reasonable quality within a relatively short window after death, the results were essentially uninterpretable. The researchers concluded that circumstantial evidence, not biochemistry, provided the greatest diagnostic contribution in suspected potassium poisoning cases.¹ In plain terms, if someone is found dead and the only poison used was potassium chloride, current forensic chemistry alone cannot prove it. Investigators need to find syringes, injection marks, witness testimony, or other physical evidence to build the case.

Inert Gases Leave Almost Nothing Behind

Helium and nitrogen kill by displacing oxygen. You breathe them in, your blood oxygen drops, and within minutes you lose consciousness and die. Neither gas is toxic in the traditional sense. They do not bind to hemoglobin the way carbon monoxide does, they do not damage tissue, and they do not leave metabolites. Helium is colorless, odorless, and cannot be detected using standard toxicological analysis.² At autopsy, the findings look like generic asphyxiation, which could have dozens of causes.

There is a narrow window of opportunity, though. If samples are collected quickly and properly during autopsy, headspace gas chromatography can detect helium in blood and lung tissue. A study of three suicidal helium asphyxiation cases demonstrated exactly this, but the authors stressed that helium escapes from samples extremely easily. If sampling is delayed, sloppy, or if nobody thinks to look for an inert gas in the first place, the evidence vanishes.³ The practical reality is that unless investigators find a helium tank, tubing, or an exit bag at the scene, there may be no reason to even suspect an inert gas, and nothing in a routine toxicology screen would flag it.

GHB and the Body’s Own Chemistry

Gamma-hydroxybutyrate presents a uniquely frustrating puzzle. GHB is a naturally occurring substance in the human body. Every person has trace levels of it circulating in their blood. After death, bacteria in the gut and decomposing tissue produce additional GHB, and this post-mortem production increases with the time between death and autopsy.⁴ So if a toxicology screen detects GHB in a corpse, the question is always: did this person ingest GHB, or did their body make it after they died?

Forensic labs deal with this by setting post-mortem cutoff values, concentration thresholds above which a result is reported as positive. But these cutoffs are imprecise because the range of endogenous and post-mortem GHB formation varies significantly between individuals and depends on how decomposed the body is.⁵ A person who took a lethal dose of GHB might have blood levels that overlap with the range a decomposing body produces on its own. A person who died of something else entirely might have levels that look suspicious simply because their body sat undiscovered for a few days.

Making matters worse, most forensic toxicology labs historically did not even screen for GHB unless it was specifically requested. When one medical examiner’s office implemented universal GHB screening using modern mass spectrometry, they identified 15 GHB-associated deaths over five years, and in roughly three-quarters of those cases, nobody had suspected GHB involvement beforehand.⁶ Those deaths would have gone unexplained, or been attributed to something else, at labs still running older testing protocols.

Succinylcholine Destroys Itself

Succinylcholine is a neuromuscular blocking agent used in hospitals to temporarily paralyze muscles during surgery. Injected outside a medical setting, it paralyzes the diaphragm, and the victim suffocates while fully conscious. The drug has earned an outsized reputation in true-crime circles for good reason: it breaks down incredibly fast. Enzymes in the blood (pseudocholinesterase) chop it into succinylmonocholine and then into succinic acid and choline, both of which are normal metabolites found in every human body.

Detection is technically possible but extremely difficult. Succinylmonocholine, the intermediate breakdown product, is the best forensic marker, and urine is the preferred sample because the drug and its metabolites linger longer there than in blood.⁷ In three homicide cases where darts tainted with succinylcholine were used, forensic scientists detected succinylmonocholine in the victims’ blood at concentrations as low as 0.45 nanograms per milliliter, and at much higher levels in urine.⁸ But those cases involved known circumstances (dart injuries on the body, a clear suspect) that pointed investigators toward the right test. Without that suspicion, standard post-mortem toxicology would not look for succinylmonocholine at all.

Insulin and Endogenous Hormones

Insulin is the body’s own hormone, produced by the pancreas to regulate blood sugar. An overdose of injected insulin causes severe hypoglycemia, which can kill. The forensic challenge is distinguishing injected insulin from the insulin the body was already producing. After death, insulin levels shift unpredictably, and the pancreas may continue to release stored insulin into surrounding tissue.

Modern forensic labs approach this by measuring both insulin and C-peptide, a molecule that is co-released with insulin when the pancreas produces it naturally. When someone injects pharmaceutical insulin, their blood insulin spikes but their C-peptide does not, because injected insulin does not come packaged with C-peptide. A wildly abnormal ratio between the two is a red flag. In one confirmed case of homicidal insulin injection, the measured insulin concentration was over 5,000 micro-units per milliliter, and the ratio of insulin to C-peptide was 111 to 1, far beyond anything the body could produce on its own.⁹ That extreme ratio made the case clear, but subtler overdoses, or cases where the body has had time to metabolize the excess insulin before death, are much harder to pin down. Newer synthetic insulin analogues add another layer of complexity because their structures differ from human insulin and may require targeted testing to identify.

Botanical and Fungal Poisons

Plant and mushroom toxins occupy an awkward blind spot in forensic toxicology. They are diverse in structure, uncommon in poisoning cases (at least in most Western countries), and consequently absent from the drug panels that labs run routinely. Aconitine, the toxin in monkshood, can be quantified in post-mortem tissue using targeted mass spectrometry methods, as demonstrated in case reports of suicidal aconitine ingestion.¹⁰ But the operative word is “targeted.” Somebody has to ask for the test.

Mushroom poisonings are even less well-served. A review of forensic cases involving toxic fungi, including those containing amatoxins (the compounds in death cap mushrooms) and orellanine, found that these poisonings are only rarely documented by toxicological analysis. The reasons include the sheer number of toxic mushroom species, the chemical diversity of their toxins, gaps in knowledge about how these toxins distribute and metabolize in the body, and a genuine scarcity of validated analytical methods for biological samples.¹¹ A death cap mushroom kills by destroying the liver over several days. If a pathologist sees liver failure at autopsy and there is no clinical history to suggest mushroom ingestion, the death might be attributed to hepatitis or another cause without toxicological confirmation ever being sought.

Ricin and Protein Toxins

Ricin, extracted from castor beans, is one of the most feared biological toxins. It kills by shutting down protein synthesis inside cells, leading to organ failure that looks a lot like severe sepsis. The toxin itself is a large protein that degrades in the body, but a smaller molecule called ricinine (present in castor beans alongside ricin) serves as a biomarker for exposure. Ricinine has been consistently detected in blood, urine, and post-mortem tissues in confirmed ricin poisoning cases.¹² However, that same systematic review noted that current detection limits need improvement, that distinguishing lethal from non-lethal exposures is difficult, and that ricin poisoning can be confused with other conditions that cause septic shock-like symptoms. In a scenario where an investigator does not suspect ricin, the autopsy findings alone would not point toward it.

Why Routine Screening Misses More Than You Would Think

Most forensic toxicology labs start with a screening panel, a set of tests designed to flag the most common drugs and poisons. These panels typically cover opioids, benzodiazepines, amphetamines, cocaine, alcohol, and sometimes a few others. A validated screening method might cover around 40 to 50 drugs simultaneously.¹³ That sounds comprehensive until you consider that there are thousands of pharmacologically active compounds in the world, hundreds of plant toxins, dozens of novel synthetic opioids appearing on the market each year, and a growing list of designer drugs that may not even have reference standards available for comparison.

Novel synthetic opioids illustrate the problem well. Fentanyl analogues and other synthetic opioids emerge so rapidly that forensic labs struggle to keep their reference libraries current. A substance that does not match anything in the lab’s database simply will not be identified. Factors like post-mortem redistribution and the stability of these drugs in stored samples add further complications to interpretation.¹⁴ The result is that some deaths caused by new synthetic drugs are initially classified as “cause undetermined” simply because the toxin was not on anyone’s radar yet.

A similar challenge exists with chemical weapons like Novichok nerve agents. Their extreme toxicity means lethal doses are minuscule, and the lack of standardized reference materials and field-deployable detection methods makes identification an ordeal even for specialized laboratories.¹⁵ In both the Skripal and Navalny cases, detection required referral to military-grade or specialized chemical weapons analysis labs, not standard forensic toxicology.

Post-Mortem Redistribution Muddies Everything

Even when a poison can be detected, interpreting what the numbers mean is its own minefield. After death, drugs do not stay where they were when the heart stopped beating. They leak out of organs, diffuse through tissues, and concentrate or dilute depending on the drug’s chemistry. This phenomenon, called post-mortem redistribution, can make drug levels look far higher or lower than they were at the time of death.¹⁶

How dramatic can the distortion be? In one study comparing drug levels measured in blood collected before death with levels measured in the same person’s femoral blood at autopsy, the ratio varied enormously. Tricyclic antidepressants showed post-mortem levels roughly three times higher than the living level on average, but individual cases ranged from barely changed to nearly 50 times higher.¹⁷ This means a toxicologist looking at a post-mortem blood result cannot simply compare it to a clinical reference range and declare “overdose” or “therapeutic.” The number is inherently unreliable as a measure of what was circulating at the time of death.

Genetic variation makes this worse. People metabolize drugs at very different rates depending on their version of the cytochrome P450 enzyme family. Someone who is an ultra-rapid metabolizer might convert codeine into morphine so efficiently that their post-mortem morphine-to-codeine ratio looks like heroin use when it was actually a normal dose of codeine.¹⁸ Layer that individual metabolic variation on top of unpredictable post-mortem redistribution, and you begin to see why forensic toxicology results are often described as providing supporting evidence rather than definitive proof.

When Standard Samples Are Gone

Decomposition adds yet another dimension of difficulty. As a body breaks down, blood becomes unrecoverable or hopelessly contaminated, urine evaporates or leaks, and tissue architecture disintegrates. In advanced decomposition, toxicologists turn to alternative matrices. Vitreous humor, the gel-like fluid inside the eyeball, is relatively isolated from bacterial contamination and post-mortem diffusion, making it useful for qualitative drug screening even when blood is compromised.¹⁹ Bone marrow, buried inside long bones and shielded from the environment, has also been shown to retain detectable concentrations of numerous drugs when blood and urine are unavailable.²⁰

But even these alternatives have limits. An exhumation case where toxicological analysis was performed on tissue more than two and a half years after burial found that gas chromatography-mass spectrometry was negative for drugs of abuse, and protein-based analysis could only confirm the presence of a naturally occurring substance (creatine), not the prescription medications the deceased had been taking in life.²¹ Time is the enemy of forensic toxicology. Every hour that passes between death and sample collection reduces the chance that a substance will be found, and for some poisons, the window is measured in minutes rather than days.

The Polonium Lesson

The 2006 poisoning of Alexander Litvinenko with polonium-210 is perhaps the most dramatic illustration of how a substance can evade detection simply because nobody thinks to look for it. Litvinenko was sick for weeks before anyone considered a radioactive poison. Standard toxicology screens returned nothing. It was not until day 22 of his illness that a urine sample was analyzed with gamma-ray spectroscopy and revealed a characteristic energy signature. Confirmatory testing, completed only after his death, showed polonium-210 concentrations roughly a billion times higher than normal background levels.²²

Polonium-210 is an alpha emitter. Alpha particles are easily stopped by skin and do not trigger radiation detectors of the type commonly used to screen people. A Geiger counter would not have flagged it. Only specialized spectroscopy on body fluids could, and that kind of analysis is not part of any forensic toxicology workflow. If Litvinenko had died more quickly, or if clinicians had not pursued increasingly exotic explanations for his deterioration, the cause of death might never have been identified. The lesson is not that polonium is undetectable. It is that the detection requires an investigator who has the imagination to ask the right question and the access to equipment that almost no forensic lab possesses.

When Context Does the Work That Chemistry Cannot

For substances like potassium chloride and succinylcholine, where the chemistry alone cannot prove poisoning, forensic pathologists rely heavily on what surrounds the death. Injection site marks, pharmacy records, the deceased’s medical history, witness accounts, surveillance footage, and the suspect’s access to the substance all become critical. Case series in forensic pathology have demonstrated that DNA analysis, histopathology, radiology, and toxicology together solve complicated cases that any single discipline would miss on its own.²³ A systematic and thorough death investigation compensates for the blind spots in any one analytical method.

This is an important corrective to the popular notion of a “perfect poison.” In crime fiction, the emphasis falls on whether a substance shows up on a toxicology report. In real forensic investigations, the question is broader: can the manner and cause of death be determined by any combination of available evidence? A poison that leaves no chemical trace can still be identified through the circumstances of its administration, the pattern of organ damage it produces, or the digital footprint of the person who purchased it. The truly “undetectable” poison is less a property of chemistry and more a failure of investigation, one where the right questions were never asked and the right tests were never run.