The Poisoned Patient: A Systematic Approach
The instinct in poisoning is to reach first for the name of the drug — to find the antidote, the reversal, the magic bullet. But the patient in front of you rarely arrives with a label, and the poison is almost never the first thing that kills them. What kills them first is a lost airway, a stopped breath, a collapsing circulation — problems you can treat without ever naming the toxin. The whole discipline of toxicology begins with a discipline of restraint: resuscitate the patient, not the poison, and let a structured risk assessment — not a panicked search for an antidote — tell you what happens next.
A 19-year-old is wheeled in at 2 a.m. after her flatmate found her drowsy on the sofa beside an empty blister pack she can't identify. She is rousable but slurring, breathing shallow, her oxygen saturation drifting down. The team's first question is not "what did she take?" — it is "is she protecting her airway?" They sit her up, give oxygen, put on monitoring, and secure a line. Only once she is breathing safely does the second question come: how much, of what, and how long ago. She admits to "a lot" of paracetamol tablets around midnight — and here is the trap, because right now she looks far better than she is. The paracetamol has not yet begun to hurt her liver. The patient you can see is not the danger; the patient you will see in three days is.
Resuscitate the patient, not the poison
In the first minutes, the name of the toxin is a distraction — the airway is the emergency. Every poisoned patient is approached the same way any critically ill patient is: ABCDE — Airway, Breathing, Circulation, Disability, Exposure. This ordering is not a formality; it encodes what actually kills. A patient who overdoses on a sedative most often dies not from the drug's chemistry but from a simple mechanical consequence — an unprotected airway and a suppressed respiratory drive. Fix the airway and the breathing and you have bought all the time in the world to work out the toxin. The poison can wait; hypoxia cannot. This is the single most important idea in the whole subject, and the reason a toxicology work-up never begins with the antidote cupboard.
Circulation deserves its own thought. Many overdoses drop blood pressure — through vasodilation, myocardial depression, or fluid loss — and the first, cheapest intervention is almost always intravenous fluid, not a specific antidote. "D" for disability means a rapid neurological check and, critically, a bedside glucose: hypoglycaemia is a great mimic of poisoning and a treatable cause of coma you must never miss. "E" for exposure means looking at the whole patient — temperature (poisons cause both hyper- and hypothermia), the skin, the pupils, patches, injection sites. Only when the patient is stable does toxicology proper begin.
The coma cocktail — reborn as targeted therapy
For the undifferentiated comatose patient, an older generation was taught a reflex "coma cocktail" given blindly to everyone: oxygen, glucose, naloxone, and thiamine. The modern practice keeps the ingredients but discards the blindness — each is now given for a reason, guided by the bedside findings. Oxygen for hypoxia. Glucose for a documented (or strongly suspected) hypoglycaemia — a stick test takes seconds and reverses a coma no antidote could touch. Naloxone when the picture fits opioids: pinpoint pupils and a low respiratory rate; it is titrated to restore adequate breathing, not to fully wake the patient, because a large bolus in a dependent patient precipitates violent, distressing withdrawal. Thiamine before or alongside glucose in the malnourished or alcohol-dependent, to avoid tipping a deficient brain into Wernicke's encephalopathy.
Think of the poisoned patient as a house filling with smoke. Your first job is not to identify which appliance caught fire — it is to get the occupants breathing: open a window, carry them to air, keep their hearts beating. Naming the source matters, but it is the second job, and a house full of smoke will kill long before you've traced the fault to the toaster. Resuscitation is opening the window. Identifying the poison is inspecting the wiring — essential, but only once everyone is breathing.
Risk assessment: the toxicologist's core skill
A single structured question answers everything: what will this exposure do to this patient, and when? Once resuscitation is under way, the central intellectual task is a risk assessment — a prediction of the expected clinical course from a handful of variables. The agent: what was taken (and remember that co-ingestants are the rule, not the exception — most deliberate overdoses involve more than one substance). The dose: how much, taken honestly as a worst-case estimate. The time since ingestion: the single most important variable, because it tells you where on the poison's timeline the patient sits. The formulation: an immediate-release tablet and a modified-release (sustained-release) version of the same drug behave like two different poisons — the modified-release form is absorbed slowly, so peak toxicity is delayed and prolonged, and a reassuring early level can lull you into discharging a patient whose real peak is still hours away. Finally the patient factors: age, weight, pregnancy, and the organs that clear the drug — impaired kidneys or liver turn an ordinary dose into a dangerous one.
Why does the time-course matter so much? Because some of the deadliest poisons have a lethal latent period — a treacherous window in which the patient looks and feels well while the injury is already under way at the cellular level. Paracetamol is the classic: for the first day the patient is nearly asymptomatic, yet the toxic metabolite NAPQI is quietly depleting the liver's glutathione, and hepatic failure declares itself only two to four days later. The toxic alcohols — methanol and ethylene glycol — behave the same way: the parent alcohol is relatively harmless; it is the slowly formed acid metabolites that blind or kill, sometimes many hours after a patient who "just seems drunk" first presents. A patient who looks fine is not the same as a patient who is fine. This principle is developed in depth in the Paracetamol and Toxic-alcohol chapters, where the whole management hinges on treating during the silent window, before the damage is done.
The most dangerous poisoned patient is often the one who looks the best. A calm, comfortable patient a few hours after a large paracetamol or ethylene-glycol ingestion is not reassuring — they are in the latent period, and the window in which treatment prevents catastrophe is open now and closing. Reassurance from a normal early examination, or a low early drug level in a modified-release overdose, is one of the classic ways toxicology kills. Trust the risk assessment and the timeline over how the patient looks.
- Approach every poisoning as a critically ill patient: ABCDE first, identify the toxin second.
- The immediate killers are mechanical — lost airway, respiratory depression, hypotension — not the drug's chemistry.
- The modern coma cocktail (O₂, glucose, naloxone, thiamine) is given for a reason, not blindly.
- Risk assessment = agent + dose + time + formulation + co-ingestants + patient factors.
- Time since ingestion is the pivotal variable — it places the patient on the poison's timeline.
- A lethal latent period (paracetamol, toxic alcohols) means a well-looking patient can be gravely poisoned.
History, collateral, and the myth of the drug screen
The history in poisoning is notoriously unreliable — patients may be drowsy, confused, ashamed, or deliberately vague, and self-reported quantities are guesses at best. So collateral is gold: the ambulance crew's description of the scene, empty packets, pill counts against a pharmacy record, a call to a relative, the patient's regular medications and what a household member might also have access to. Piecing together what was truly available and taken often tells you more than the patient can. But note the boundary — this is information-gathering, not acting on unverified instructions; a claim about the dose is a lead to corroborate, not a fact to bank.
The urine "tox screen" is the most over-ordered and least useful test in poisoning. Students expect a single panel to name the poison; reality disappoints. The standard urine drug-of-abuse screen detects a narrow list of substances, is riddled with false positives and negatives, reports what was taken days ago as readily as today, and — crucially — almost never changes management. You treat the clinical picture (the toxidrome) in front of you, not a delayed qualitative urine result. Knowing that a screen is "positive for benzodiazepines" does not tell you the patient is poisoned by them, and a negative screen does not exclude a dangerous ingestion the panel simply doesn't cover. The tests that actually change management are specific quantitative levels and physiological measurements, and they are a short, deliberate list.
The investigations that actually change management
A focused panel earns its place. A paracetamol level in every deliberate overdose — this is non-negotiable, because paracetamol is common, silent early, and eminently treatable, and patients under-report or don't know it is in the tablets they took. A salicylate (aspirin) level for the same reasons. An ECG to read the QRS width and the QTc: a widening QRS is the fingerprint of sodium-channel blockade (as in tricyclic-antidepressant toxicity) and predicts seizures and arrhythmia, while a prolonged QTc warns of torsades. A bedside glucose, always. A venous or arterial blood gas to reveal a metabolic acidosis and its severity. And the two calculated gaps: the anion gap (raised by acid-producing poisons) and the osmolar gap (raised early by the unmetabolised toxic alcohols) — together they are a powerful clue to methanol or ethylene-glycol poisoning before a specific level returns. Each of these is ordered because a positive result redirects treatment, which is the only good reason to order any test in toxicology.
Regardless of what the patient says they took: a timed paracetamol level (interpreted against the treatment nomogram once ≥4 hours post-ingestion), a salicylate level, a bedside glucose, an ECG (QRS and QTc), and a blood gas with electrolytes to calculate the anion and osmolar gaps. A urine drug-of-abuse screen is deliberately not on this list — it is the test students order and toxicologists ignore. The pattern of these results, read together with the toxidrome, usually identifies the danger long before any specialised assay returns.
Supportive care is the antidote you almost always have
For most poisons there is no antidote — and most patients recover anyway, on good supportive care alone. It is a myth, fed by television, that poisoning is treated with a specific antidote pushed into a vein. Antidotes exist for only a minority of toxins, and even where one exists it usually buys time rather than working alone. The true mainstay of poisoning care is meticulous supportive care: a protected airway and adequate oxygenation, fluids and vasopressors to hold the blood pressure, control of seizures (with benzodiazepines first), correction of temperature and electrolytes, and attentive monitoring while the body clears the poison itself. Get the supportive care right and the great majority of poisoned patients walk out well. The narrow toolbox of things that go beyond support — stopping further absorption, speeding elimination, and the specific antidotes — is developed in the chapters that follow this one, but none of it replaces the fundamentals; it is built on top of them.
This is also where the broader map of the subject comes together. The pattern of vital signs, pupils, skin and mental state you elicited in the ABCDE and Exposure steps often falls into a recognisable Toxidrome — the pattern-recognition shortcut that names a class of poison from the bedside, covered in the next chapter. Whether to stop further uptake of the drug is the domain of Decontamination, and whether the body's own clearance can be accelerated (by urinary alkalinisation or dialysis) is the domain of Enhanced elimination — and both decisions lean directly on the pharmacokinetics you meet in the Principles of Pharmacology chapter: how a modified-release formulation prolongs absorption, and how a drug's half-life and volume of distribution decide whether removing it is even possible.
When to call the poisons centre
No clinician holds the toxicology of thousands of agents in their head, and the honest, safe move is to ask early. National poisons information services and their databases (such as TOXBASE) exist precisely to give current, agent-specific advice on expected toxicity, thresholds for concern, and management. Call sooner rather than later for: an unfamiliar agent, a large or mixed ingestion, a modified-release preparation, a child, a paediatric "one pill can kill" concern, any patient who is symptomatic out of proportion, or whenever the risk assessment is uncertain. Reaching for the phone is not a sign of weakness; in toxicology it is standard, expected practice — and it is far better to ask about a well-looking patient in the latent period than to explain later why you didn't.
- The history is unreliable — corroborate with collateral (paramedics, packets, pharmacy records, relatives).
- A urine drug-of-abuse screen rarely changes management — treat the toxidrome, not the screen.
- Order a paracetamol level in every deliberate overdose, plus salicylate, glucose, ECG, and a gas with the gaps.
- Supportive care is the mainstay; most poisonings have no antidote and recover with good support.
- Call the poisons centre / TOXBASE early — for unfamiliar, large, mixed, modified-release, or paediatric exposures.
- This chapter sets up the next four: toxidromes, decontamination, enhanced elimination, and antidotes.
- Hunting for the antidote before securing ABCDE — the airway, breathing and circulation kill first, and are treatable without ever naming the toxin.
- Being reassured by a well-looking patient or a low early level in a paracetamol, toxic-alcohol, or modified-release ingestion — the latent period and delayed absorption hide the true peak.
- Ordering a urine drug screen and letting it steer (or falsely reassure) management — it is rarely actionable; the paracetamol level, ECG, and gas are what count.
A 25-year-old is brought in two hours after taking an unknown quantity of tablets. She is drowsy with a low respiratory rate and oxygen saturation of 88% on air. Which is the most appropriate first action?
- Resuscitate the patient, not the poison: ABCDE and a bedside glucose come before identifying any toxin — the immediate killers are mechanical.
- Risk assessment (agent, dose, time, formulation, co-ingestants, patient) predicts the course; time and a lethal latent period matter most — a well-looking patient can be gravely poisoned.
- The drug screen rarely helps; the useful tests are a paracetamol level in every deliberate overdose, plus salicylate, ECG (QRS/QTc), glucose, and a gas with the anion and osmolar gaps.
- Supportive care is the mainstay — most poisonings have no antidote and recover with it; call the poisons centre early, and read on for toxidromes, decontamination, elimination and antidotes.
- Goldfrank's Toxicologic Emergencies — Initial evaluation and general management of the poisoned patient.
- Rang & Dale's Pharmacology — Principles of toxicology and drug overdose.
- Katzung Basic & Clinical Pharmacology — Management of the poisoned patient.
- British National Formulary (BNF) — Emergency treatment of poisoning; and the UK National Poisons Information Service database (TOXBASE).
- Murray L, et al. Toxicology Handbook — Risk assessment-based approach to poisoning.
- UpToDate — General approach to drug poisoning in adults.
