Poison Test: How It Works and What It Detects

Poison testing is a broad term for laboratory methods that detect harmful substances in biological samples like blood, urine, or hair. In practice, it almost always follows a two-stage approach: a fast initial screen that flags potential toxins, followed by a slower but far more precise confirmatory analysis that identifies the exact substance and its concentration. The range of detectable poisons has expanded enormously since the 1830s, when the first reliable chemical test for arsenic transformed criminal investigation. Modern labs can now identify hundreds of drugs, heavy metals, gases, and biological toxins in a single analytical run, though the testing landscape comes with real limitations that are worth understanding.

Screening First, Confirmation Second

The typical poison test begins with immunoassay screening. These tests use antibodies designed to bind to specific drug classes or toxin groups. When a target substance is present in a urine or blood sample, the antibody reaction produces a measurable signal. Immunoassays are fast, relatively cheap, and widely available, which is why they dominate initial testing in emergency departments and workplace drug programs.¹ They can return results within minutes, which matters when a doctor needs to decide on treatment for someone who may have overdosed.

The catch is that immunoassays are blunt instruments. They identify drug classes rather than individual compounds, and they miss entire categories of substances. A study comparing on-site immunoassay screening against the gold-standard laboratory method in emergency department patients found that roughly half of the patients had ingested substances the screening device could not detect at all, and about one in five results disagreed with the confirmatory test.² That rate of discrepancy is why any positive screening result in a serious poisoning case gets sent for confirmatory testing.

The confirmatory step almost always involves some form of mass spectrometry. Techniques like gas chromatography-mass spectrometry and liquid chromatography-mass spectrometry separate a sample into its individual chemical components and then identify each one by its molecular weight and fragmentation pattern. These instruments can detect substances at concentrations measured in parts per billion or lower, and they can distinguish between structurally similar compounds that an immunoassay would lump together.³ The tradeoff is time: mass spectrometry results can take hours or even days, depending on the lab’s workload and the complexity of the analysis.

What Gets Sampled

Blood and urine are the workhorses of poison testing, but they are not the only options. Blood captures what is circulating in the body right now and allows concentration measurements that clinicians can compare against known toxic thresholds. Urine catches drugs and their breakdown products over a longer window, typically hours to a few days, which makes it the preferred matrix for workplace testing and abstinence monitoring.

Hair offers something neither blood nor urine can: a timeline stretching back weeks or months. As hair grows, substances get incorporated into the shaft, creating a record of exposure that can be sectioned and analyzed to estimate when a substance was used. Other alternative specimens include oral fluid, sweat, nails, and vitreous humor from the eye. Each has tradeoffs involving how easily a sample can be collected, how much drug gets into that matrix, and how long the detection window lasts.⁴ Blood-based analysis is often preferred when reference concentration ranges are available for clinical comparison, while hair and urine are more useful when the question is whether someone was exposed at all over a longer period.⁵

Drugs and Medications

The most common reason for poison testing is suspected drug exposure, whether from an intentional overdose, accidental ingestion, or substance abuse. Standard immunoassay panels typically screen for a set of common drug classes: amphetamines, opiates, benzodiazepines, barbiturates, cocaine, cannabis, phencyclidine, and tricyclic antidepressants. Some expanded panels add synthetic opioids like fentanyl or newer substances like buprenorphine.

Acetaminophen (paracetamol) poisoning deserves special mention because it is one of the most common causes of acute liver failure in the developed world, and its treatment depends entirely on a precisely timed blood level. Doctors plot the measured concentration against the number of hours since ingestion on a treatment nomogram. If the level falls at or above a threshold of 150 micrograms per milliliter at four hours post-ingestion, the antidote acetylcysteine is started. A higher line, starting at 300 micrograms per milliliter at four hours, marks patients at elevated risk of liver injury.⁶ This is one of the clearest examples of how a quantitative poison test directly determines treatment.

Heavy Metals

Testing for toxic metals like lead, mercury, arsenic, thallium, and cadmium requires a different analytical approach. Immunoassays do not work for individual elements, so laboratories use inductively coupled plasma mass spectrometry, which vaporizes the sample in an extremely hot plasma and then measures the mass of each element present. This technique is considered the standard for multi-element analysis because of its sensitivity and ability to measure many metals simultaneously.⁷

A single validated method can quantify upward of 30 elements in blood, plasma, urine, or hair from one preparation procedure, covering metals ranging from lithium and beryllium to lead, mercury, and uranium.⁸ This breadth matters for cases where the toxic agent is unknown. A clinician suspecting heavy metal poisoning does not need to guess which metal to look for; the lab can screen for dozens at once and report anything that falls outside normal reference ranges.

Gases and Volatile Substances

Carbon monoxide poisoning is one of the most common toxic gas exposures, and detecting it involves measuring how much hemoglobin in the blood has been bound by carbon monoxide rather than oxygen. Clinicians check this using a device called a CO-oximeter, which reads the blood’s light absorption at specific wavelengths. For more precise or forensic-grade results, gas chromatography can separate and measure the carbon monoxide directly. Portable sensor-based gas chromatography systems have also entered forensic practice, offering easier handling in the field.⁹

Hydrogen cyanide is another gas tested for, particularly in fire fatalities and certain industrial exposures. In aviation accident investigations, for example, blood samples from fatalities are routinely analyzed for both carbon monoxide (as carboxyhemoglobin) and cyanide.¹⁰ Other volatile substances that can be detected include alcohols like methanol and ethylene glycol, various solvents, and anesthetic gases, all of which have distinct signatures when passed through a gas chromatography system.

Plant and Fungal Toxins

Biological toxins from mushrooms, plants, and other organisms present a different detection challenge. Amatoxins, the lethal compounds in death cap mushrooms, are a good example. These peptides are present in tiny amounts and act on the liver hours after ingestion, when the patient may not yet feel seriously ill. Rapid lateral flow immunoassay devices have been developed that can detect alpha-amanitin in mushroom tissue at concentrations as low as about 0.3 nanograms per milliliter, a sensitivity comparable to laboratory-grade liquid chromatography-mass spectrometry methods.¹¹

Other biological threat agents, including ricin and T-2 mycotoxin, can also be screened in food samples using commercially available immunoassay kits. These devices have proven capable of detecting all three toxins at concentrations well below what would pose a health concern in a single serving of food, using straightforward extraction and dilution steps.¹² These tools are designed more for food safety and bioterrorism screening than for clinical diagnosis in a patient, but the underlying technology is similar.

When Screening Tests Get It Wrong

False positives are the most practically important limitation of immunoassay screening. Because the antibodies in these tests bind to molecular shapes rather than exact chemical identities, structurally similar but unrelated compounds can trigger a positive result. A large-scale analysis of electronic health records identified previously unknown cross-reactivities affecting immunoassay screens for amphetamines, buprenorphine, cannabinoids, and methadone.¹³ The problem is widespread: a comprehensive review of the literature found reported false-positive interferences for every major drug class tested by immunoassay, including amphetamines, opiates, benzodiazepines, cannabinoids, tricyclic antidepressants, phencyclidine, LSD, and barbiturates.¹⁴

Common culprits include over-the-counter cold medications containing pseudoephedrine or dextromethorphan (which can trigger amphetamine screens), certain antidepressants (which can trigger phencyclidine or tricyclic antidepressant screens), and proton pump inhibitors (which have been linked to false-positive THC results). These errors matter beyond clinical medicine. A false-positive drug screen can lead to loss of employment, child custody problems, or criminal charges if the result is not confirmed by a more specific method. Ethical concerns around testing without informed consent and the lack of routine confirmatory testing in many settings make this an area where the technical limitations have real human consequences.¹⁵

Rapid Bedside Testing in Emergency Rooms

Point-of-care poison tests are designed to give emergency physicians a fast read on what a patient may have taken. These are typically small cassette-based devices that run an immunoassay panel on a urine sample right at the bedside. Their performance is generally strong for detecting presence or absence of common drug classes. In a study of patients suspected of drug toxicity, a rapid bedside device showed sensitivity above 90% for cocaine, opiates, barbiturates, benzodiazepines, and tricyclic antidepressants, with specificity above 95% for every class except tricyclic antidepressants.¹⁶

A comparative study of three different point-of-care devices found that sensitivity across the board ranged from about 97% to 100%, though specificity varied more widely depending on the device, ranging from 50% on one device’s expanded panel to about 90% on the others.¹⁷ Cannabis and benzodiazepines have consistently shown the most variability in rapid test accuracy across devices and studies.¹⁸ For emergency clinicians, these tools are most useful for ruling out a substance rather than definitively confirming one. A negative result carries more weight than a positive, which still needs to be interpreted cautiously and, in ambiguous cases, confirmed.

Testing After Death

Post-mortem toxicology is substantially harder than testing in a living patient, mainly because of a phenomenon called postmortem redistribution. After death, drugs do not stay where they were when the heart stopped. They leak out of organs like the liver, lungs, and heart into surrounding blood and tissue. Cell membranes break down, blood pools and coagulates, and decomposition accelerates the whole process.¹⁹ The result is that a drug concentration measured in post-mortem blood may be significantly higher or lower than what was actually circulating at the time of death.

The longer the interval between death and sampling, the more advanced the decomposition and the greater the distortion of drug levels.²⁰ Forensic toxicologists try to minimize this by collecting blood from peripheral sites like the femoral vein, which is farther from the major organs that act as drug reservoirs. They also compare concentrations across multiple sample types and correlate findings with any clinical information available from before death.²¹ Despite decades of research into this problem, interpreting post-mortem drug levels remains one of the most contested areas in forensic science.

Untargeted Screening for Unknown Poisons

Sometimes the question is not “is substance X present” but “what on earth is in this sample.” High-resolution mass spectrometry makes it possible to take an untargeted approach, recording the full chemical profile of a sample without needing to specify what to look for in advance. One method applies this to plasma from poisoned patients: the instrument captures a complete spectral dataset, and then a data-mining technique subtracts out normal background compounds, leaving behind a “xenobiotic fingerprint” of foreign substances unique to each individual sample.²² The high mass accuracy of these instruments allows compounds to be identified with minimal ambiguity, even when the analyst had no prior suspicion of what might be there.²³

This capability is particularly relevant when novel or unusual toxins are involved. Traditional targeted methods can only find what they are programmed to look for. If someone has been poisoned with an exotic plant alkaloid or a newly synthesized compound, a targeted screen would miss it entirely. Untargeted approaches do not solve every problem, since identifying an unknown compound still requires matching its spectral data against reference libraries, and those libraries have gaps. But they represent a fundamentally different philosophy of testing: cast a wide net first, then narrow down.

The Problem of Novel Synthetic Substances

New psychoactive substances, sometimes called designer drugs, are a persistent headache for poison testing. These compounds are deliberately modified at the molecular level to evade both legal controls and laboratory detection. Their chemical structures shift constantly, which means standard immunoassay panels may not recognize them, and even mass spectrometry reference libraries may not contain their spectral data.²⁴ Synthetic cannabinoids, synthetic cathinones, and novel benzodiazepine analogs are among the most common categories. In practice, a patient who has taken a brand-new synthetic compound may present with serious symptoms while every standard poison test comes back negative.

Laboratories are responding by expanding their mass spectrometry libraries more frequently, developing broader screening panels, and using untargeted high-resolution methods that can flag unknown compounds for further investigation. But there is an inherent lag: a new substance has to appear on the street, be identified in a laboratory somewhere, and have its spectral data shared before other labs can detect it reliably. During that window, cases go undiagnosed or are attributed to the wrong substance.

Workplace and Environmental Exposure Monitoring

Poison testing extends well beyond the emergency department and the forensic morgue. Occupational health programs rely on biological monitoring to assess whether workers are being exposed to harmful substances at dangerous levels. Organophosphate pesticide exposure, for example, is tracked by measuring the activity of cholinesterase enzymes in red blood cells. These enzymes are inhibited by organophosphates, so a drop in their activity serves as a biomarker for exposure. A study of workers exposed to organophosphates found a clear association between occupational exposure and cholinesterase inhibition, underscoring the value of routine monitoring.²⁵

For industrial chemicals like benzene, testing traditionally relied on measuring specific metabolites in urine. Newer metabolomic approaches go further by profiling the full range of metabolic changes in exposed workers, including disruptions to energy metabolism and lipid metabolism that might flag health damage before obvious symptoms appear.²⁶ These approaches have also been applied to dioxin exposure, where untargeted metabolomic analysis of urine identified metabolic patterns linked to dioxin toxicity and revealed compounds with potential as clinically meaningful biomarkers.²⁷

Metabolomics and Where Testing Is Headed

The metabolomic approach represents a broader shift in poison testing philosophy. Instead of asking “is toxin X present,” metabolomics asks “what has this person’s body chemistry been doing.” By analyzing thousands of small molecules in blood or urine using high-resolution mass spectrometry, researchers can detect characteristic metabolic disturbances that point to a specific exposure even when the toxin itself has already been cleared from the body. This is especially promising for chronic low-level exposures, where the traditional approach of measuring the toxin directly may miss the problem entirely because concentrations have dropped below detection limits between sampling events.

A review of metabolomics applied to occupational pollutants concluded that the continued application of these techniques to worker populations has the potential to discover new biomarkers for chemical exposures, identify possible toxic mechanisms, and improve understanding of how workplace chemicals increase disease risk.²⁸ The practical barrier is complexity: metabolomic datasets are enormous, and distinguishing toxin-related metabolic shifts from normal biological variation across different people, diets, and health conditions remains an active research problem. The technique is not yet a routine clinical tool, but it is increasingly used in research settings and may eventually transform how chronic poisoning and environmental exposure cases are investigated.

From Arsenic Mirrors to Molecular Fingerprints

The history of poison testing tracks a steady march from detecting one substance at a time to profiling hundreds simultaneously. In 1836, the chemist James Marsh developed a test that could reveal arsenic by converting it to arsine gas and depositing a telltale metallic mirror on a glass tube. By 1840, this technique had become the first analytical method in toxicology introduced during a criminal trial, fundamentally changing how poisoning cases were prosecuted. Before the Marsh test, arsenic was nicknamed “inheritance powder” because it was virtually undetectable. Once a reliable test existed, arsenic poisoning cases plummeted.

Modern laboratories bear almost no resemblance to Marsh’s bench setup, but the core logic has not changed: take a biological sample, separate or isolate the suspected substance, and identify it with enough certainty to hold up under scrutiny. What has changed is the scale of what can be detected. A single mass spectrometry run can now screen for hundreds of drugs, dozens of metals, and unknown compounds that have never been catalogued, all from a fraction of a milliliter of blood. Meanwhile, the same fundamental tension Marsh faced persists: the gap between what testing can detect and what people are actually poisoned with. Whether the unknown is a new synthetic drug or a rare plant toxin, no test is useful unless it can recognize what it finds. That gap keeps shrinking, but in a world where new toxic substances emerge faster than reference libraries can grow, it has not closed.