Lithium Toxicity and the Serotonin Syndrome

A drug with almost no margin for error

Lithium's problem is written into its pharmacology: a narrow therapeutic index. For most drugs there is a comfortable gap between the dose that works and the dose that harms. Lithium has almost none — the therapeutic window sits at roughly 0.6–1.0 mmol/L, and toxicity begins to creep in not far above it. That is why patients on lithium have their blood levels, kidney function and thyroid checked regularly; the drug demands surveillance in a way few others do. Lithium is not metabolised by the liver and is not protein-bound. It is a small ion that is handled almost entirely by the kidney, filtered and then reabsorbed alongside sodium. This single fact explains the whole toxicology: anything that makes the kidney hold on to more sodium and water — dehydration, renal impairment, or drugs that reduce renal perfusion — makes it hold on to more lithium too, and the level climbs.

Acute versus chronic — two different poisonings

It matters greatly whether the poisoning is acute or chronic. Acute overdose — a person who is not usually on lithium swallowing a large amount at once — presents early with gastrointestinal upset: nausea, vomiting and diarrhoea dominate, while the lithium is still in the gut and blood and has not yet penetrated the brain. The neurological features lag behind, because lithium enters cells and the central nervous system slowly. Chronic toxicity is the opposite and, in practice, the commoner and more dangerous scenario: a patient stable on lithium for years whose clearance falls — through dehydration, a vomiting illness, worsening kidney function, or a newly added drug. Here the body's tissues are already saturated, so the presentation is predominantly neurological from the outset, and it can be severe at a blood level that looks only modestly raised. A third pattern, acute-on-chronic, blends the two. The lesson: the number on the lab report means little without knowing the story behind it.

The precipitants worth memorising are almost all about the kidney. Three drug classes are the classic culprits, and all raise lithium by cutting its renal excretion. Thiazide diuretics deplete sodium, prompting the tubule to reabsorb more sodium — and lithium rides along with it. ACE inhibitors (and angiotensin-receptor blockers) drop the pressure driving filtration and reduce lithium clearance. NSAIDs constrict the afferent arteriole and cut renal blood flow, doing the same. Add a hot climate, a diarrhoeal illness or simple under-drinking, and a patient who has been safe for a decade can become toxic without changing their lithium dose at all. These interactions tie directly into the Cardiovascular and Endocrine chapters — the very drugs used for hypertension and heart failure are the ones that push lithium up.

The clinical picture: a brain in trouble

Lithium toxicity is, above all, a neurological syndrome, and it tends to progress in a recognisable order. Early on there is a coarse tremor — distinct from the fine, benign tremor of therapeutic lithium — together with nausea and diarrhoea. As it deepens, the cerebellar and cortical signs come out: ataxia and unsteady gait, slurred speech, muscle twitching and fasciculations, and brisk, exaggerated reflexes (hyperreflexia). Then confusion and agitation, and in severe poisoning seizures, a fluctuating conscious level, coma, and cardiac conduction changes. A minority are left with lasting cerebellar damage even after the level normalises — the SILENT syndrome (Syndrome of Irreversible Lithium-Effectuated NeuroToxicity). Two chronic effects, seen with long-term therapy rather than acute overdose, deserve their own mention: nephrogenic diabetes insipidus — lithium makes the collecting duct deaf to ADH, so the patient passes large volumes of dilute urine and, dangerously, dehydrates and concentrates the lithium further — and hypothyroidism, from lithium's interference with thyroid hormone release. Both link to the Endocrine section.

So how is lithium poisoning actually managed? The cornerstone is restoring renal excretion. Generous intravenous fluids — isotonic saline — rehydrate the patient and let healthy kidneys clear the lithium themselves; correcting the sodium and water deficit is doing the real work. At the same time you stop the drug and, just as importantly, stop the precipitant: withhold the thiazide, the ACE inhibitor, the NSAID that started the slide. Haemodialysis is reserved for the sickest: severe neurological features, very high levels, or a patient in renal failure who cannot clear the ion themselves. After dialysis, the level must be re-checked because of rebound. For an early, deliberate massive ingestion of a slow-release preparation, whole-bowel irrigation may be used to sweep tablets through before they are absorbed — but charcoal has no role.

Serotonin syndrome: too much of a good signal

This one is rarely a single drug — it is usually a collision. Serotonin syndrome is a state of excess serotonergic activity in the central nervous system, and its most important clue is in the drug history: it is typically precipitated by combining two or more serotonergic agents, or by adding one to another too soon. The offenders come from surprisingly varied classes. The antidepressants are the obvious ones — SSRIs and SNRIs, and above all MAOIs, whose combination with any other serotonergic drug is the most feared. But the list reaches well beyond psychiatry: the analgesics tramadol, pethidine and fentanyl; the antibiotic linezolid (itself a weak MAOI); the antiemetics and the migraine triptans; the recreational drugs MDMA and cocaine; and the herbal remedy St John's wort, which patients often take without telling anyone. It is the interaction that catches people out — starting an SSRI, adding tramadol for a sore back, and the two together tipping serotonin over the edge. These agents are covered individually in the Central Nervous System chapter; here it is their sum that matters.

The syndrome announces itself fast — usually within hours of the offending combination, which itself helps distinguish it from slower-building mimics. Clinically it is a triad. First, neuromuscular excitation: clonus is the standout finding — rhythmic jerking, classically inducible at the ankles (lower-limb predominant) and sometimes seen in the eyes as ocular clonus — along with hyperreflexia, tremor and rigidity. Second, autonomic instability: hyperthermia, tachycardia, sweating, flushing, dilated pupils, and often diarrhoea. Third, altered mental status: agitation, anxiety and confusion. Diagnosis is clinical, and the widely used Hunter criteria lean heavily on that neuromuscular hyperactivity — the presence of spontaneous or inducible clonus, or tremor with hyperreflexia, in someone taking a serotonergic drug is often enough to make the call. The hyperthermia is the part that kills: driven by relentless muscle activity, it can climb to dangerous levels.

Management runs from simple to aggressive with severity. The first and most important step costs nothing: stop every serotonergic drug. Most mild cases settle within a day or two on that alone plus supportive care. Beyond that, the priorities are aggressive supportive care and active cooling for hyperthermia, and benzodiazepines — which calm the agitation, reduce muscle activity and thereby the heat, and are the true workhorse of treatment. For moderate-to-severe cases that do not settle, the specific antidote is cyproheptadine, an antihistamine with serotonin-receptor antagonist (anti-5-HT2A) activity that directly opposes the excess signalling; it is given orally or by nasogastric tube. In life-threatening hyperthermia the muscles must be silenced fast: neuromuscular paralysis with intubation, and dantrolene to stop the muscular heat generation, mirroring the management of malignant hyperthermia. One antipyretic that does NOT help is paracetamol — the fever is muscular in origin, not a hypothalamic set-point problem, so cooling and stopping the muscle activity are what work.

Serotonin syndrome vs its mimics

Four hyperthermic, agitated, rigid states get confused at the bedside, and separating them changes the treatment. Neuroleptic malignant syndrome (NMS) is the closest mimic but almost a mirror image mechanistically: it is caused by dopamine blockade — antipsychotics, or abrupt withdrawal of a Parkinson's dopamine agonist — and evolves slowly, over days rather than hours. Its rigidity is generalised "lead-pipe" stiffness with bradyreflexia (reduced reflexes) and NO clonus, and it is treated by stopping the antipsychotic and giving dantrolene and the dopamine agonist bromocriptine. The antipsychotics that cause it are the same drugs whose immune-unrelated but overdose-relevant effects sit in the Central Nervous System chapter. The anticholinergic toxidrome shares the fever and agitation but the skin is the giveaway — dry and flushed rather than sweat-drenched, with the classic "mad, red, dry, hot, blind" picture and normal or reduced reflexes, not clonus. And malignant hyperthermia is a pharmacogenetic reaction to volatile anaesthetics and suxamethonium, unrelated to psychiatric drugs, also treated with dantrolene. The discriminators worth carrying: tempo of onset, the reflex/clonus pattern, and the drug history.