Predicting Outcome in Metformin Lactic Acidosis
A 56-year-old male with diabetes and hepatitis comes in after an intentional overdose of his own metformin. On evaluation, he is awake and alert but mildly tachycardic. Lab evaluation demonstrates elevated lactate at 4 mmol/L and a pH of 7.3. Will he die?
Metformin is one of the most commonly prescribed drugs for Type 2 diabetes, and overdose is a common scenario encountered in toxicology. Derived from the French lilac (Galega officinalis) metformin is a biguanide that treats non-insulin-dependent diabetes via inhibition of gluconeogenesis, facilitating cellular glucose uptake, and improving insulin responsiveness. It replaced its more hazardous predecessor phenformin, the notorious “P” in MUDPILES, which caused an unacceptably high rate of “lactic acidosis” (a concept I’ll tackle a little later on). While preliminary data suggested that metformin didn’t cause lactic acidosis at all1, that is clearly untrue. Although 20 times less likely than with phenformin2, metformin-associated lactic acidosis (also known as MALA) is definitely real, both after an acute overdose and in persons taking it therapeutically who have impaired metformin or lactate clearance. And in these scenarios, sometimes patients die.
Pyruvate gone bad
Pyruvate + NADH + H+ ⇔ Lactate + NAD+
To understand how metformin produces hyperlactatemia, first, let’s talk about lactate itself. Lactate (CH3CH(OH)COOH) exists in a balance with pyruvate. Under normal physiologic conditions, pyruvate enters gluconeogenesis and the Krebs cycle via a number of enzymatic reactions that largely depend on adequate cellular stores of NAD+.
If something goes awry, pyruvate is converted to lactate. This increase in lactate production is also sometimes termed a Type B, or non-hypoxic, lactic acidosis. This is in contrast to Type A (hypoxic) lactic acidosis, which occurs in the context of tissue hypoxia: ischemia, shock, and whatnot. And as long as we’re talking about the alphabet soup of lactate, don’t forget about D-lactate, the isomeric form overproduced in small bowel dysfunction or resection, particularly in the context of a high carbohydrate meal. This form of lactate is not picked up by the lab as lactate but can still produce a Type B lactic acidosis. The lactate we’re talking about with metformin is L-lactate, the isomer most commonly involved in hyperlactatemia.
Type B lactic acidosis is the more common mechanism of xenobiotic-induced lactic acidosis. Many drugs, such as antiretrovirals, linezolid, cyanide, propofol, salicylates, epinephrine, and propylene glycol, impair mitochondrial respiration and back up the pyruvate pathway to generate excess lactate. Also, anything which creates a higher NADH / NAD+ ratio, such as alcohol, slows down the other reactions and favors the production of lactate. Of note, thiamine is an essential cofactor for pyruvate dehydrogenase, so thiamine deficiency contributes to lactate buildup, and hyperlactatemia is a marker for Wernicke’s encephalopathy.3,4
LACTATE ≠ ACIDOSIS
Lactate alone does not cause a drop in pH because it is not itself a proton donor. Some really dislike the term “lactic acidosis” for this reason – the very presence of lactate does not mean there is an acidosis. If you infuse lactate into a healthy subject, the pH does not fall. The link between lactate and acidosis primarily stems from the fact that lactate is a marker for decreased cellular respiration.5 So while hyperlactatemia does not necessarily mean acidosis, they often happen at the same time in the same patient.
Metformin causes both hyperlactatemia and acidosis, through several mechanisms (Figure 2)6:
- First, it causes acidosis via inhibition of Complex 1 in the electron transport chain. This shuts down the generation of ATP, creating a buildup of protons in the mitochondria, which creates the acidosis. This also prevents the utilization of NADH as a substrate in the electron transport chain.
- NADH builds up, NAD+ is not produced, and reactions that require NAD+ as a substrate slow down or stop. Pyruvate accumulates.
- In contrast, the conversion of pyruvate to lactate by lactate dehydrogenase, which is favored by excess NADH, takes off. Excess lactate is generated from pyruvate.
- Metformin blocks the entry of pyruvate into gluconeogenesis. More lactate is generated as pyruvate accumulates.
- If that were not enough, metformin decreases hepatic lactate uptake and utilization. Lactate clearance is impaired.
What results is acidosis and increased lactate, attributable at least partly to metformin, that is often termed Metformin-Associated Lactic Acidosis (MALA). Some use the label Metformin-Induced Lactic Acidosis (MILA) when the blame clearly lies with the drug and not other exacerbating factors. For the purposes of this discussion, we will stick with MALA to mean lactic acidosis when metformin is involved, and is, at least partly, responsible.
First and foremost, the approach to MALA involves correcting any underlying concurrent conditions. Fluid resuscitation for hypovolemia, treatment of septic shock and heart failure, and so on.
Then comes the question of sodium bicarbonate. A tempting option when facing a patient with a pH of 6.9 and a lactate of 20, this is tricky business. Potential benefits include raising the pH to optimize physiology, organ function, and perfusion. But criticisms abound, including the potential to worsen intracellular acidosis, the absence of evidence that it improves outcome, the increase in serum osmolarity if not administered carefully, the decrease in intracellular calcium that in turn decreases cardiac contractility, and the potential to actually increase lactate production.7 All in all, it’s a therapy best used sparingly, and probably only at extremely low pH (<7.0).
The most definitive therapy is hemodialysis. Metformin has pretty good characteristics for dialysis: low molecular weight (165 Da), high water solubility, low protein binding, and it is excreted unchanged by the kidney. While total body clearance is normally pretty brisk (500ml/min), this falls significantly when GFR drops, so dialysis can still appreciably improve on the endogenous clearance of the drug.8 Other benefits include the ability to rapidly correct pH, electrolyte abnormalities, and hypothermia, and enhance endogenous clearance of lactate. EXTRIP recommends this therapy be instituted when lactate is over 20 mmol/L and pH <7.0, with more liberal thresholds if co-morbid conditions are present.8
Risk factors for MALA morbidity
How do we identify which patients to be concerned about? When should we institute hemodialysis, and on whom? Although anything that increases serum concentrations of metformin can be fatal, patients with acute metformin overdose generally have a more favorable prognosis than patients who develop an intercurrent illness which leads to MALA. Anytime a patient’s clearance of metformin or lactate is impaired, or tissue hypoxia occurs, MALA can creep in. Any condition impacting renal clearance or blood flow can cause an accumulation of the drug and subsequent lactate overproduction. A concomitant Type A lactic acidosis, such as in the face of septic shock or ischemia, can easily become a mixed Type A/B acidosis with MALA. And, since 70% of lactate clearance takes place in the liver, hepatic insufficiency can also be bad.9 The mortality rate for MALA in these circumstances approaches 50%.7
Acute metformin overdoses clearly fare better, although mortality does occur. This is the patient who presents after the overdose and may initially look pretty good, but progressively (and sometimes rapidly) deteriorates. It’s important to not be fooled if initial lab evaluation is reassuring. A recent study demonstrated it can take up to 12 hours for MALA to develop after an overdose. One of the most confusing features of these cases is that the rate of rise is not predictable.10
Prediction and Timing of Mortality
Several studies have taken a stab at predicting which patients with MALA are at the highest risk of mortality. Generally, the lower the pH and the higher the lactate, the worse the outcome, although the association is not clear.2 In patients with intercurrent illness and MALA, serum metformin levels correlate with outcomes – but are not practical to obtain in real-time. One study of mixed patients (both acute and chronic MALA) demonstrated the one independent predictor of mortality to be decreased prothrombin activity – presumably because the liver clears lactate.7 Another study looking at only overdose patients concluded that no patients died unless they reached a nadir pH of 6.9 and a peak lactate of 25mmol/L.11
Here’s the trouble: a pH of 6.9 and a lactate of 25 is a destination, but how do we predict who is on their way? How long do they take to get there? A recent poison center study demonstrated that everyone who had MALA had an initial lactate greater than 2 mmol/L, but there was no clear relationship between the peak lactate and how long it took to get there.10 Also, some patients didn’t manifest MALA until more than 12 hours after presentation. Time and time again, investigators have tried to pin down exactly who is at greatest risk, and when. Alas, the answers about timing evade us. I’d love to see a study looking not just at the numbers, but the slope of the line. A girl can dream.
In the meanwhile, here’s what we have:
- Metformin overdose MALA is better than intercurrent illness MALA.
- Lactate does not cause acidosis, but metformin causes both acidosis and hyperlactatemia.
- Measuring pH and lactate are some help, so you should monitor them both frequently.
- MALA can take a while to develop, so don’t dispo patients before at least 12 hours if you’re concerned.
- Correct the underlying condition, and perform hemodialysis if things continue to worsen. Avoid bicarbonate unless the pH is truly awful.
- Don’t be fooled in a patient who initially looks pretty good. There is plenty of time to die.
- 1.Salpeter S, Greyber E, Pasternak G, Salpeter E. Risk of fatal and nonfatal lactic acidosis with metformin use in type 2 diabetes mellitus. Cochrane Database Syst Rev. 2010;(4):CD002967. doi:10.1002/14651858.CD002967.pub4
- 2.Blumenberg A, Benabbas R, Sinert R, Jeng A, Wiener S. Do Patients Die with or from Metformin-Associated Lactic Acidosis (MALA)? Systematic Review and Meta-analysis of pH and Lactate as Predictors of Mortality in MALA. J Med Toxicol. January 2020. doi:10.1007/s13181-019-00755-6
- 3.Donnino M, Miller J, Garcia A, McKee E, Walsh M. Distinctive acid-base pattern in Wernicke’s encephalopathy. Ann Emerg Med. 2007;50(6):722-725. doi:10.1016/j.annemergmed.2006.10.022
- 4.Butterworth R. Effects of thiamine deficiency on brain metabolism: implications for the pathogenesis of the Wernicke-Korsakoff syndrome. Alcohol Alcohol. 1989;24(4):271-279. doi:10.1093/oxfordjournals.alcalc.a044913
- 5.Wang G, Hoyte C. Review of Biguanide (Metformin) Toxicity. J Intensive Care Med. August 2018:885066618793385. doi:10.1177/0885066618793385
- 6.Protti A, Fortunato F, Monti M, et al. Metformin overdose, but not lactic acidosis per se, inhibits oxygen consumption in pigs. Crit Care. 2012;16(3):R75. doi:10.1186/cc11332
- 7.Seidowsky A, Nseir S, Houdret N, Fourrier F. Metformin-associated lactic acidosis: a prognostic and therapeutic study. Crit Care Med. 2009;37(7):2191-2196. doi:10.1097/CCM.0b013e3181a02490
- 8.Calello D, Liu K, Wiegand T, et al. Extracorporeal Treatment for Metformin Poisoning: Systematic Review and Recommendations From the Extracorporeal Treatments in Poisoning Workgroup. Crit Care Med. 2015;43(8):1716-1730. doi:10.1097/CCM.0000000000001002
- 9.Kraut J, Madias N. Lactic acidosis. N Engl J Med. 2014;371(24):2309-2319. doi:10.1056/NEJMra1309483
- 10.Theobald J, Schneider J, Cheema N, DesLauriers C. Time to development of metformin-associated lactic acidosis. Clin Toxicol (Phila). November 2019:1-5. doi:10.1080/15563650.2019.1686514
- 11.Dell’Aglio D, Perino L, Kazzi Z, Abramson J, Schwartz M, Morgan B. Acute metformin overdose: examining serum pH, lactate level, and metformin concentrations in survivors versus nonsurvivors: a systematic review of the literature. Ann Emerg Med. 2009;54(6):818-823. doi:10.1016/j.annemergmed.2009.04.023