Metformin

Why this poison is interesting

Metformin is the library's clearest example of a drug whose overdose contradicts its pharmacology. It is an antidiabetic, so the expectation is hypoglycaemia — and that expectation is wrong. Metformin does not drive insulin release, and hypoglycaemia has not been observed even after very large ingestions.² Its danger is a lactic acidosis: metformin inhibits mitochondrial respiration, lactate accumulates, and the patient who dies is the one whose pH falls while the glucose stays normal.³

• It does not cause hypoglycaemia. The Glucophage label records that hypoglycaemia has not been seen with metformin doses of up to very large amounts, although lactic acidosis has occurred in such circumstances.² This is the single most important thing that separates it from the sulfonylureas, which share its clinic and nothing else about its overdose.
• The lesion is metabolic, and it builds. Metformin raises plasma lactate in a concentration-dependent way by inhibiting mitochondrial respiration, predominantly in the liver;³ as the drug accumulates, the acidosis widens.
• There is no antidote. Nothing blocks or reverses the biguanide. Management is supportive, and in the severely acidotic patient the specific treatment is the dialyser — which removes both metformin and the lactate at once.¹²

Metformin is the drug that reminds you to read the gas, not the glucose: its overdose hides in the pH.

The toxic principle

There is one toxic principle, and it is the parent drug acting on the mitochondrion. Everything clinical follows from lactate.

• Metformin inhibits mitochondrial respiration. It suppresses complex I of the respiratory chain, predominantly in hepatocytes, which shifts the cell toward anaerobic glycolysis and raises lactate production in a plasma-concentration-dependent manner.³ Established
• It inhibits hepatic gluconeogenesis. The liver's normal use of lactate — converting it back to glucose — is impaired, so lactate that is overproduced is also under-cleared.³ The two effects compound: more lactate made, less lactate removed. Established
• It does not release insulin. Metformin is not a secretagogue; it sensitises tissues to insulin rather than stimulating its release, which is why overdose does not reliably cause hypoglycaemia and why a normal or high glucose does not reassure.² Established

Toxicokinetics

Metabolism and the metabolites

Metformin is the rare monograph with nothing to say under this heading, and the absence is itself informative. It is not metabolised — it is excreted unchanged, with no human metabolites identified.² There is no toxic metabolite, no bioactivation, no minor pathway that becomes major in overdose.

What changes the answer

• Renal impairment is the dominant modifier — it is the difference between a cleared drug and an accumulating one, and it is why metformin is contraindicated in significant renal impairment in therapeutic use.³ Established
• A second insult multiplies the lactate. Sepsis, cirrhosis, cardiac failure or any cause of hypoperfusion adds its own lactate and impairs clearance, and MALA in therapeutic dosing usually requires one.³ Established
• Co-ingestants that lower blood pressure or harm the kidney accelerate the spiral; the overdose rarely travels alone.

Elimination and accumulation

Metformin is cleared by the kidney and nowhere else. There is no hepatic route to fall back on, no metabolite to excrete, and negligible protein binding to buffer the free concentration. When the kidney works, a metformin overdose is largely a gastrointestinal nuisance; when it does not, the drug accumulates and the acidosis follows.

The EXTRIP position

EXTRIP recommends extracorporeal treatment in severe metformin poisoning (1D) — one of the strongest recommendations in its whole catalogue, and a marked contrast to the against recommendations that dominate this band.¹

• Indicated if any of: lactate >20 mmol/L (1D) or >15 mmol/L (2D); pH ≤7.0 (1D) or ≤7.1 (2D); failure of standard supportive measures (1D); shock (1D); impaired kidney function (1D); decreased level of consciousness (2D); liver failure (2D).¹
• Stop when lactate is <3 mmol/L and pH >7.35 (both 1D).¹
• Intermittent haemodialysis with a bicarbonate buffer is preferred initially (1D); continuous kidney replacement therapy is an alternative if haemodialysis is unavailable (2D); repeat sessions may use either (1D).¹
• Keep monitoring lactate and acid–base after dialysis — the large volume of distribution means levels and acidosis can rebound and a further session may be needed.¹

Target organs — and why those

Timeline of effects

Metformin's timeline is slow at the front and steep at the back: gastrointestinal upset first, a lactic acidosis building underneath, and — if the accumulation is not interrupted — shock, coma and renal failure. The dangerous interval is the quiet one, while the lactate is rising and the patient still looks well.

What the mechanism predicts at the bedside

Why the lactate, not the glucose, is the number to follow

Because the toxic principle is inhibition of mitochondrial respiration and gluconeogenesis, the disease expresses itself as lactate and pH, not as glucose.²³ A metformin overdose with a normal sugar and a rising lactate is behaving exactly as the mechanism predicts, and the lactate trend is the severity marker that drives the decision to dialyse.¹ Established

Why there is no antidote, only removal

Nothing is bioactivated, consumed or bound, so there is no cofactor to replace and no enzyme to block.² The only way to stop the mitochondrial effect is to take the molecule out of the patient — which, because metformin is barely protein-bound and freely filtered, a dialyser does efficiently, clearing the accumulated lactate in the same session.¹

Why renal function decides everything

Metformin is cleared only by the kidney and has no metabolic escape route.² A normal kidney turns an overdose into a GI illness; a failing kidney turns a therapeutic dose into MALA. This is why impaired renal function is simultaneously the main risk factor, a marker of severity and an EXTRIP indication for dialysis.¹³ Established

Why bicarbonate buffering and dialysis go together

EXTRIP prefers intermittent haemodialysis with a bicarbonate buffer,¹ because the machine is being asked to do two jobs at once — remove the drug and correct the acidosis. Treating the acidosis while leaving the drug in place invites rebound; removing the drug while ignoring the acidosis leaves the immediate killer untreated. The bicarbonate-buffered dialysis does both.

The antidote, from the poison's side

Metformin has no antidote, and the reason is mechanistic rather than historical: there is no step in its toxicity for an antidote to occupy. This section therefore describes the specific treatment that stands in an antidote's place — extracorporeal removal — and why it fits the poison so exactly.

• Haemodialysis is the specific treatment, because metformin's negligible protein binding and small size make it readily removable, and the same session clears the lactate and corrects the pH.¹² It is the functional antidote — a machine doing what no molecule can. Established
• Supportive care carries the rest — airway, ventilation, fluids, vasopressors for shock, and correction of acidosis — and EXTRIP's indications explicitly include failure of standard supportive measures, recognising that supportive care alone may not outrun the accumulation loop.¹
• There is nothing to give that reverses the biguanide. No drug blocks its mitochondrial effect; looking for a pharmacological antidote wastes the time that should go into the gas, the lactate and the decision to dialyse.²
• Expect to repeat. The large volume of distribution means drug and acidosis can rebound, so monitoring continues after the first session and a second may be needed.¹

Critical appraisal

1. The mitochondrial-respiration account of the lactic acidosis is well established and clinically load-bearing.³ It explains why the acidosis is the disease, why the glucose is irrelevant, and why dialysis — which removes drug and lactate together — is the specific treatment. DeFronzo's review states the concentration-dependent mechanism and the predominantly hepatic site directly.³
2. That hypoglycaemia is not a feature is stated by the label itself,² Established and it is the most useful single fact on the page — it redirects attention from the sugar to the gas. The contrast with the sulfonylureas is deliberate and the two pages are written to be read together.
3. The MALA-versus-MILA distinction is real but can be overdrawn. The mechanism is identical; the difference is only the source of accumulation (failing clearance versus a large dose).³ The page keeps them under one heading on purpose, because the bedside response — follow the lactate, dialyse the severe — does not depend on which label applies.
4. The requirement for a second insult applies to therapeutic-use MALA, not necessarily to massive overdose.³ A very large acute ingestion can accumulate enough metformin to acidify a previously well patient without sepsis or cirrhosis, so the 'second insult' framing is a description of the common case, not a safety rule.
5. The EXTRIP recommendation is unusually strong for this band — a recommend for (1D) where most of Band F2's agents draw a recommend against.¹ The quality of evidence is still very low, and the thresholds (lactate 20, pH 7.0) are consensus cut-points from case-level data; they should be read alongside the clinical picture, not as validated population triggers.
6. The latent phase is genuine and is the page's main safety message.²³ Its timing depends on dose, preparation and renal function, so 'the acidosis can build' is an instruction to repeat the gas, not a fixed interval to wait out.

References

1. 1
   EXTRIP Workgroup. Extracorporeal treatment for metformin poisoning: systematic review and recommendations. Recommendation set at extrip-workgroup.org/metformin; full systematic review: Calello DP, Liu KD, Wiegand TJ, et al. Crit Care Med 2015;43(8):1716–30 (PubMed 25860205). Source of the strong (1D) recommendation to perform extracorporeal treatment in severe poisoning; the lactate >20 mmol/L (1D) and >15 mmol/L (2D) and pH ≤7.0 (1D) and ≤7.1 (2D) thresholds; shock, impaired kidney function and failure of supportive measures as indications; the lactate <3 mmol/L and pH >7.35 cessation criteria; and the preference for intermittent haemodialysis with a bicarbonate buffer. Read from the workgroup's published recommendation page; verified 13 Sep 2026.
2. 2
   Glucophage 500 mg film-coated tablets — Summary of Product Characteristics, Merck Serono. emc product 987. §4.9 and §5.2 fetched and read in full. Source of the statement that hypoglycaemia has not been seen even after very large metformin ingestions while lactic acidosis has, that haemodialysis is the most effective method to remove lactate and metformin, the negligible protein binding, the 50–60% bioavailability and ~2.5 h peak, the saturable/incomplete absorption, the 63–276 L volume of distribution, the absence of human metabolites, the renal clearance >400 mL/min by filtration and tubular secretion, the ~6.5 h half-life, and its prolongation with renal impairment. Verified 13 Sep 2026.
3. 3
   DeFronzo R, Fleming GA, Chen K, Bicsak TA. Metformin-associated lactic acidosis: current perspectives on causes and risk. Metabolism 2016;65(2):20–29. PubMed 26773926. Source of the concentration-dependent inhibition of mitochondrial respiration predominantly in the liver, the requirement in therapeutic use for an accumulating metformin level plus a secondary disrupting event (cirrhosis, sepsis or hypoperfusion), and the high mortality of established MALA. Verified 13 Sep 2026 from the abstract.
4. 4
   TOXBASE — metformin; biguanides. National Poisons Information Service. toxbase.org (NHS login required. NPIS: 0344 892 0111. The authoritative UK source for lactate and pH thresholds, dialysis criteria and resuscitation targets. Login-gated, so not quoted here.)