Tricyclic Antidepressant Overdose: Bicarbonate and the Wide QRS

One drug, four poisons

A tricyclic in overdose is not a clean toxin — it is a shotgun that hits several receptors at once. At therapeutic doses TCAs relieve depression mainly by blocking the reuptake of noradrenaline and serotonin — the pharmacology covered in the Central Nervous System chapter. In overdose that mechanism is almost irrelevant; what kills is everything else the molecule does. Four actions run in parallel. First, anticholinergic (antimuscarinic) blockade produces the classic toxidrome — delirium and agitation, tachycardia, dry mouth, flushed hot skin, dilated pupils, urinary retention: "mad as a hatter, dry as a bone, red as a beet, hot as a hare." Second, alpha-adrenergic blockade dilates vessels and drops blood pressure. Third, antihistamine (H1) blockade adds sedation and coma. And fourth — the lethal one — a fast sodium-channel blockade in the heart. Layered on top, TCAs lower the seizure threshold, so convulsions are common and dangerous. It is the collision of all these effects that makes the overdose so treacherous.

The real killer: jamming the cardiac sodium channel

Every heartbeat begins with a rapid inflow of sodium through fast voltage-gated sodium channels — the steep upstroke (phase 0) that fires the impulse from cell to cell. TCAs block these channels, and they do it preferentially when the channel is open or the heart is beating fast, so the poison bites harder exactly when the heart is stressed. This is called a membrane-stabilising or quinidine-like effect, and it is shared with the class Ia and Ic anti-arrhythmics and with local anaesthetics — the same family discussed under Cardiovascular / Arrhythmias. The consequence is slowed conduction through the ventricle. The impulse crawls, so the QRS complex — the electrical signature of ventricular depolarisation — widens. Push it further and conduction fragments into ventricular tachycardia or fibrillation, while the drug's negative inotropy and alpha-blockade collapse the blood pressure. Slowed conduction, wide QRS, ventricular arrhythmia, hypotension: this quartet is what stops the heart.

The ECG is the prognosis

In tricyclic poisoning you do not need a drug level — you need an ECG. The width of the QRS is the single most useful measurement in the whole poisoning, because it tracks how much sodium channel is blocked — and therefore how sick the patient is. Two thresholds are worth carrying: a QRS wider than about 100 ms flags a real risk of seizures, and wider than about 160 ms flags a real risk of ventricular arrhythmia. Alongside the widening, the blockade drags the terminal part of the QRS rightward, which produces a characteristic pattern in lead aVR: a tall terminal R wave with a deep, slurred S wave in leads I and aVL. That rightward terminal axis, together with sinus tachycardia from the anticholinergic effect, is close to a fingerprint for tricyclic cardiotoxicity. So the ECG is not just diagnostic — it is the running scorecard you treat against, watching the QRS narrow as bicarbonate takes effect.

The antidote: sodium bicarbonate, a two-hit rescue

Sodium bicarbonate is the specific antidote, and its beauty is that it attacks the sodium-channel block from two directions at once. First, the sodium: a concentrated bolus floods the extracellular space with sodium ions, and that surge of substrate helps drive current through whatever channels remain open — brute force overcoming the blockade. Second, the bicarbonate: it raises the blood pH into a mildly alkalotic range, and alkalinisation reduces the drug's affinity for the sodium channel, so more of the poison lets go. A sodium load plus an alkaline shift — the two hits reinforce each other, which is why bicarbonate works where a plain saline sodium load or hyperventilation alone is weaker. The indications are the cardiotoxic signs, not the mere fact of ingestion: give it for a widening QRS, for ventricular arrhythmia, or for hypotension. You titrate to effect — pushing boluses and watching the QRS narrow on the monitor — aiming for a mildly alkalotic arterial pH rather than any fixed dose. It treats the killing mechanism directly while the anticholinergic and sedative effects wear off on their own.

Everything else: support, and what not to give

Around the bicarbonate, the rest of the care is supportive — and half of it is knowing which "obvious" antidotes are traps. Seizures are treated with benzodiazepines, which raise the seizure threshold and are also the treatment for agitation — and controlling seizures matters beyond the fit itself, because the acidosis they generate worsens sodium-channel binding. Phenytoin is specifically avoided: as a sodium-channel blocker it is at best useless and at worst additive to the cardiotoxicity. Hypotension that does not respond to fluids and bicarbonate is supported with vasopressors. Because much of the danger is anticholinergic, gut absorption is slow, and a single dose of activated charcoal is reasonable in an early presentation with a protected airway — decontamination principles that belong to the Enhanced-elimination and Antidotes chapter. Note what does not help: TCAs are large, highly protein-bound and lipophilic with a huge volume of distribution, so haemodialysis removes almost nothing. Two antidotes are actively contraindicated. Flumazenil must not be used — even if a co-ingested benzodiazepine is suspected — because removing that GABA-ergic brake can unmask seizures in a patient already primed to convulse and already at risk of fatal arrhythmia. And physostigmine, the anticholinesterase that reverses the anticholinergic delirium, has precipitated asystole and seizures in tricyclic overdose and is contraindicated; the danger of "fixing" the anticholinergic picture is a theme that recurs in the Anticholinergic chapter. For arrhythmia refractory to bicarbonate, magnesium and, increasingly, intravenous lipid emulsion are used as rescue — the lipid "sink" pulling the lipophilic drug out of the tissues, the same principle applied to local anaesthetic toxicity.