Hemodialysis in lithium poisoning: what is the evidence?

January 23, 2015, 11:59 pm

0Lithium Carbonate, 300mg, 1,000 Capsules per Bottle McGuffMedical.com★★★½☆

Extracorporeal Treatment for Lithium Poisoning: Systematic Review and Recommendations from the EXTRIP Workgroup. Decker BS et al. Clin J Am Soc Nephrol 2015 Jan 12 [Epub ahead of print]


The purpose of this review, from the Extracorporeal Treatments in Poisoning (EXTRIP) workgroup, was to give evidence-based recommendations related to use of hemodialysis and other extracorporeal modalities in the treatment of lithium toxicity.

The goal of enhanced elimination in lithium poisoning is to avoid the syndrome of irreversible lithium-effectuated nerutoxicity (SILENT), which causes persistent cerebellar dysfunction after prolonged exposure of the CNS to high levels of lithium.

Using elaborate methodology, the EXTRIP group reviewed relevant literature, identifying 166 papers (describing 228 patients) in multiple languages that merited inclusion. Most of these were case reports yielding, in the words of the authors, “a very low quality of evidence for all recommendations.” In other words, there is no real evidence, and the recommendations and (weaker) suggestions are based mostly on the opinions of the workgroup members.

Now, the EXTRIP group is comprised of some very smart clinicians from a variety of specialties, including medical toxicology, nephrology, and intensive care. Their opinions are definitely worth reading and considering carefully. For example, this is a summary of their proposed indications for extracorporeal treatment (ECTR – read hemodialysis) in lithium poisoning:

  • ECTR is recommended
    • if kidney function is impaired [serum creatinine > 2.0 mg/dL in adults or 1.5 mg/dL in the elderly] and lithium > 4.0 mEq/L
    • if there is decreased level of consciousness, seizures, or life-threatening dysrhythmias regardless of lithium level
  • ECTR is suggested:
    • if lithium level > 5.0 mEq/L
    • if there is confusion
    • if the expected time to reach lithium level < 1.0 mEq/L with optimal management is greater than 36 h

There are a number of problems with this list. Most importantly, I think, is that it considers hemodialysis in isolation without integrating the modality in the total care of the patient and other possible treatments. For example, if an adult took an acute lithium overdose and presented with a serum lithium level of 4.1 mEq/L and at serum creatinine of 2.1 mg/dL, but without signs or symptoms of toxicity, would it not be reasonable to administer fluids, observe carefully, and recheck levels in several hours? And if that makes sense, what if the lithium was 4.2 and the creatinine 2.2? Would the speculative potential benefit of hemodialysis in this situation outweigh the small chance of possible risks? Clinical judgment is still required, and strict cut-off criteria seem artificial — and certainly not evidence-based.

Some toxicologists have objected that the recommendations seem not to make a distinction between acute, chronic, and acute-on-chronic exposure. I think the authors would argue that inclusion of symptoms in their recommendations (if not in their weaker “suggestions”) corrects for not considering the pattern of exposure, and that the most important thing is whether the drug is in the CNS, not how it got there.

In any case, there are many pearls about lithium toxicity in this article, and it is well worth reading. Just take the recommendations with a grain of salt — but not a salt substitute.
Related posts:

Lithium-induced nephrotoxicity


  1. David Juurlink Says:


    I don’t think we should expect guidelines to address every clinical possibility.

    In your scenarios of Li+ levels of 4.1 or 4.2 mEq/L, a great deal depends on whether the accompanying renal insufficiency is likely to be readily reversed, allowing for a consequent increase in renal lithium clearance.

    This is sometimes not knowable a priori. If you have good reason to believe that it is reversible, supportive care and generous crystalloid administration may be defensible, particularly in a person with few comorbidities in whom you believe lithium absorption to be complete.

    If you’re not confident the creatinine can be promptly brought down promptly (either because of CKD or because of an inability to aggressively replete intravascular volume for any reason – LV systolic dysfunction, for example), or if there’s lithium remaining in the GI tract, not implementing HD early may be asking for trouble. As you correctly note, the most important thing is whether the lithium is in the CNS or not. Unlike the “wait and see” approach, early hemodialysis may clear lithium before it has the opportunity to fully distribute into that compartment.

    Dave Juurlink

  2. Leon Says:


    Thank you for your comments. I think we agree — clinical judgment is still required. My concern is that some will look at the recommendations and take the creatinine > 2 and lithium > 4 recommendation as an absolute indication for dialysis. Or, even worse, take the lithium > 5 “suggestion” as carved in stone. It seems to me that there is no real scientific basis for choosing that number and not, say, 4.9 or 5.1 or some other figure.

    The question is, would the speculative benefit of dialyzing immediately in the asymptomatic patient outweigh the small potential risks associated with dialysis such as those listed in the paper, including hypotension, drop in hemoglobin, upper extremity vein thrombosis, and sepsis.

    In my view the “recommendations” are not evidence-based because there’s really no good evidence to base them on. They’re opinions which, given the quality of the authors and other members of the EXTRIP workgroup, I respect tremendously. But opinions nonetheless, and open for debate.

  3. David Juurlink Says:


    Yes, we mostly agree. As for the risk/benefit tradeoff of early HD, who knows? It’s impossible to be definitive because the data are so limited. But it’s worth noting that the utility of most medical interventions increases with the severity of disease, as does the balance of benefits to risks.

    I don’t agree with your assertion that the recommendations (you’ll note the absence of quotation marks there) aren’t evidence-based. They are. Low quality evidence is something toxicology is sadly accustomed to, as you know, but this doesn’t preclude the crafting of recommendations, and sometimes even strong ones. It does mean that people are more likely to quibble with the final product, which is of course fine and usually a healthy thing to do. The EXTRIP group did its own share of quibbling as part of the process.

    Maybe I’ll pimp a little here and quote Dave Sackett (BMJ 1996): “Evidence based medicine is the conscientious, explicit, and judicious use of current best evidence in making decisions about the care of individual patients.” That’s exactly what EXTRIP did with each toxin we considered. Do you see an alternative?

    Dave Juurlink