Poisoning and Dialysis
On this page
Direct answer
A poison is dialysable when it is small, water-soluble, poorly protein-bound and confined largely to the extracellular space — the profile of methanol and ethylene glycol (with their toxic acids), lithium, salicylates, theophylline, phenobarbitone and metformin-associated lactic acidosis. Dialysis is indicated not merely for a high level but for level plus end-organ threat: coma, seizures, refractory acidosis (classically pH below about 7.25-7.3 in toxic alcohol poisoning), visual symptoms in methanol, or oliguria. Heavily protein-bound drugs — phenytoin, carbamazepine — and organophosphates leave the dialyser untouched and go to haemoperfusion instead. In India the classic presentations are methanol from adulterated liquor, organophosphates and aluminium phosphide from agriculture, and suicidal drug cocktails.
What you must remember
- Dialysability criteria: low molecular weight, water solubility, low protein binding, small volume of distribution, low natural clearance — and a native clearance that is failing.
- Methanol: metabolises to formic acid — visual blurring, scotomata, blindness, severe high-anion-gap acidosis; treat with fomepizole (or ethanol) to block alcohol dehydrogenase, and dialyse for acidosis, visual signs or level above about 50 mg/dL — dialysis removes methanol and formate both.
- Ethylene glycol: metabolises to oxalic acid — calcium oxalate crystals in urine, renal failure, profound acidosis; same antidote logic and dialysis triggers as methanol.
- Lithium: dialyse for level above about 4 mEq/L in acute overdose (chronic toxicity lower, around 2.5, especially symptomatic), or any level with coma, seizures or renal failure; rebound after dialysis is expected — repeat sessions and monitor.
- Salicylates: dialyse chronic toxicity above about 60 mg/dL and acute above about 100 mg/dL, or earlier with coma, seizures, pulmonary oedema or refractory acidosis — haemodialysis outperforms alkaline diuresis once these thresholds cross.
- Metformin: the metformin itself is dialysable and lactic acidosis clears with it — dialysis is the definitive therapy in metformin-associated lactic acidosis.
- Not dialysable (know the logic): organophosphates and phenytoin, carbamazepine (protein-bound — haemoperfusion territory); digoxin (huge volume of distribution, antidote Fab fragments instead); distal thin-membrane poisons such as hydrocarbons.
- Practical pearls: high blood flow (300-400 mL/min) maximises extraction; a large-surface dialyser, repeat sessions for rebound-prone toxins, and never let dialysis delay atropine in organophosphate poisoning.
A methanol outbreak walk-through
A district hospital receives six men over one weekend after a hooch batch: blurred vision, vomiting, staggering, and a graveyard-high anion gap with osmolar gap. Triage runs on physiology. Those with visual symptoms, acidosis below pH 7.3 or methanol levels above roughly 50 mg/dL go straight to the dialysis list; the rest receive fomepizole (where funded) or intravenous ethanol, thiamine and folate as cofactor support, and hourly monitoring — because methanol itself is only mildly intoxicating while its metabolite formic acid blinds and kills. The machine set-up is conventional but aggressive: bicarbonate bath to fight the acidosis, blood flow 350-400, and sessions repeated until the osmolar gap closes and acidosis resolves, since tissue methanol keeps re-equilibrating into plasma as dialysis pulls it out.
The Indian follow-through matters as much as the acute care: reporting to the district surveillance unit, forensic samples of the liquor, and — on the preventive register — the reason poison-control resources and antidote stocking vary so widely between states. For the exam, this single scenario encodes most of the page: dialysability logic, antidote-plus-dialysis as parallel therapies, and the thresholds that convert a biochemistry number into a machine decision.
Where students slip
The first slip is modality mismatch — sending phenytoin or carbamazepine to haemodialysis when protein binding demands haemoperfusion, or conversely "haemoperfusion for lithium", which charcoal barely adsorbs. The second is forgetting that dialysis clears the parent alcohol and the toxic metabolite but does not stop metabolism, so fomepizole or ethanol runs alongside — antidote and machine are partners, not alternatives. The third is rebound: a reassuring post-dialysis lithium or methanol level rises again as tissue stores re-equilibrate, so the answer is planned repeat sessions and continued monitoring, not celebration. And the reflex worth suppressing: dialysing every overdose reflexively when supportive care and activated charcoal already suffice for most.
Frequently asked questions
What properties make a poison amenable to removal by dialysis?
Small size, water solubility, low protein binding and a small volume of distribution, together with low endogenous clearance — the profile of lithium and the toxic alcohols.
Which findings trigger dialysis in methanol poisoning?
Metabolic acidosis (pH roughly below 7.25-7.3), visual symptoms, methanol level above about 50 mg/dL, or renal failure — with fomepizole or ethanol running to block metabolism meanwhile.
Why is digoxin toxicity not treated with dialysis?
Digoxin is huge in volume of distribution and largely tissue-bound, so extracorporeal removal is negligible; digoxin-specific antibody fragments are the therapy.
Which poisoned patients need haemoperfusion rather than haemodialysis?
Those with heavily protein-bound drugs such as phenytoin, carbamazepine and theophylline in regions of severe toxicity, where adsorption outperforms diffusive clearance.
Why do lithium levels rise again after dialysis?
Lithium re-equilibrates from the intracellular compartment into plasma after the session, so repeat treatments and serial levels are planned rather than a single run.