EMCrit Podcast 32 – Treatment of Severe Hyperkalemia

>>> Update: For a new take on kayexalate, see Mak Moayedi’s Lecture

Hey folks. As I get ready for ACEP, I just wanted to get a quick podcast put up. One of the listeners requested an episode on the treatment of hyperkalemia in the ED.

There was a fantastic article published in Critical Care Medicine on the topic by a Dr. Weisberg. I go through my management and discuss some of the pearls from the article.

Weisberg LS. Management of severe hyperkalemia. Crit Care Med. 2008 Dec;36(12):3246-51.

Additional References added Feb 2012

ECG is insensitive and non-specific for severe hyperkalemia issues; essentially is crap (Clin J Am Soc Nephrol 3: 324-330, 2008). ECG peaked T waves, that resolved after K normalized were noted in only 1 of the 14 hyperkalemic patients who went on to have arrhythmia or cardiac arrest. Only half of them had any T-wave changes.

Calcium Gluconate doesn’t require Hepatic Metabolization before it is active



and now to the podcast…


You finished the 'cast,
Now get CME credit

Already an EMCrit CME Subscriber?
Click Here to Get CME Credit for the Episode

Not a subcriber yet? Why the heck not?
By subscribing, you can...

  • Get CME hours
  • Support the show
  • Write it off on your taxes or get reimbursed by your department

Sign Up Today!


Subscribe Now

If you enjoyed this post, you will almost certainly enjoy our others. Subscribe to our email list to keep informed on all of the ED Critical Care goodness.

This Post was by , MD, published 5 years ago. We never spam; we hate spammers! Spammers probably work for the Joint Commission.


  1. emcrit says

    The Trueginator wrote me with this ACEP Abstract:

    Beneficial Effect of Calcium Treatment for Hyperkalemia Is Mediated by Calcium-Dependent Conduction, Not “Membrane Stabilization”

    Abstract seems to indicate we should be saying, “cardiac protection” instead of “membrane stabilization.”

  2. says

    thanks for reviewing hyperK, as this comes up virtually every shift.

    I disagree that the ECG should not be the initial intervention when potassium is reported as high.

    The Weisberg paper supports the assertion that “the first cardiac manifestation of hyperkalemia may be ventricular fibrillation” with a single reference, a study by Dodge from 1953.

    The Dodge study is amazing. These guys gave a huge slug of oral potassium to a bunch of different types of subjects and watched the ECG. They remark on the ECG changes across all these subjects in excruciating detail for many pages, and then, at the end of the paper, they mention that, by the way, an hour after they gave their 15 grams of KCl, one study subject XXXXX DIED.

    “The fatality occurred in a 47 year old white man who complained of exertional dyspnea for nine years and intermittent claudication for three years. There was no history of angina pectoris, chest pain, or congestive heart failure. His cardiac examination was entirely normal with the exception of the electrocardiogram which showed a normal QRS loop but an abnormally directed T vector, producing inverted T waves in Leads I and Va to Vg. One hour following the oral administration of 15.0 Gms. of potassium chloride the patient developed short runs of ventricular tachycardia at a time when there were only slight potassium effects as evidenced by increase in T-vector magnitude. Eighty minutes following potassium administration the patient developed fatal ventricular fibrillation. Just prior to the onset of the ventricular fibrillation the T vector was still only slightly increased in magnitude with no change in direction. None of the other electrocardiogram signs of potassium intoxication appeared, such as P-R interval abnormalities, absence of P waves, or intraventrirular block. Post-mortem examination revealed a marked degree of coronary artery sclerosis with many focal areas of myocardial fibrosis, but no area of frank myocardial infarction.”

    Can you imagine what these guys were thinking when they killed this guy?

    So the assertion that there were no changes prior to vfib is not true, even based on this single 1953 study, but more importantly, we know that ***the ECG changes of hyperkalemia are not reliable when there are existing T wave or QRS changes.*** We just had a fabulous case of a gentleman with a host of pre-existing ECG changes who presented with hyperK and marked bradycardia that responded to hyperK treatment. His ECG demonstrated loss of the P but no other signs, however his QRS was widened at baseline and Ts abnormal at baseline.

    The majority of elevated potassium values are spurious. I’m not sure if you meant to suggest that all these hyperKs should receive calcium chloride before an ECG, but I believe that would be an incorrect and dangerous strategy.

    If someone is critically ill, and a high K is resulted, it’s reasonable to treat. Patients who are critically ill are more likely to have true hyperK, are in closely monitored settings being cared for by the best nurses.

    But in the average well emergency patient whose labs show elevated K, the right response is ECG. In patients with a relatively normal baseline tracing, both the 1953 Dodge paper and the accumulated experience of everyone I’ve talked to on this subject agree that if the ECG is OK, you don’t need to reach for calcium.

    I don’t interrupt colleagues who give the calcium gluconate to patients without ECG changes, because even though I don’t think it has any benefit (it does not change the serum K and has a brief duration of action), it is very unlikely to cause harm beyond diverting attention away from more appropriate interventions. But the chloride salt is a different story, you can hurt someone with CaCl2.

    I submit that the correct response to high K in most situations is an ECG. If there are ECG changes consistent with hyperK or existing ECG changes that confound hyperK analysis, commence treatment, watching the ECG and serum K carefully. I believe that in the vast majority of patients the ECG will correctly guide your approach. There are exceptions, and liberal use of hyperK therapies does not bother me as long as the potential toxicities of these therapies are understood and weighed against the likelihood of benefitting the patient.

    • emcrit says

      I think this excerpt sums it up best
      These classic ECG changes, which have been well described in the literature, are not always present.3–5 The relation between ECG changes and serum potassium will vary between people. It is therefore important to realise that the ECG is not always a reliable test for mild to moderate hyperkalaemia.4 Even in severe hyperkalaemia, only minimal changes may be seen on the ECG.5
      (from Emerg Med J 2002;19:74-77)

      3. Yu AS. Atypical electrocardiographic changes in severe hyperkalaemia. Am J Cardiol1996;77:906–8.
      4. Wrenn KD, Slovis CM, Slovis BS. The ability of physicians to predict hyperkalaemia from the ECG. Ann Emerg Med1991;20:1229–32.
      5. Martinez-Vea A, Bardaji A, Garcia C. Severe hyperkalaemia with minimal electrocardiographic manifestations: a report of seven cases. J Electrocardiol1999;32:45–9.

      I believe (I hope) in the podcast I made clear that we should be sure the elevated potassium is not spurious before treating with calcium.

      But once you are assured of that, I maintain that high Ks (as to what is a high K, it probably depends on the scenario) should probably receive calcium. I was taught the same as you and I believed that a normal EKG in the setting of very high Ks means the patient is safe, but do we have evidence of that?

      Can you elaborate on the additional harms posed by CaCl over CaGluc aside from the extravasation issue mentioned in the podcast.

  3. says

    you CENSORED me? sheesh.

    my only concern with CaCl2 vs. CaGlu is extravasation. a rare but serious complication, and if there is no chance the patient will benefit…

    I understand that the ECG does not correlate well with the level, but the point is which predicts a patient-oriented outcome, the level or ECG?

    • emcrit says

      The patient oriented outcomes are dysrhythmia and death. It would seem that the ECG is specific, but insensitive for these patient oriented outcomes.

      I think we can agree that if the patient looks well, not-so-high K (6.0-7.0), you believe the lab is real, and no EKG changes, that if you are going to give calcium, it should be in the form of calcium gluconate.

  4. says

    Scott, do you have any knowledge of whats new with IV calcium along with digoxin. The old myth warns about “stone heart” but I’ve heard evidence is weak and based on very old cases in the literature.


  5. emcrit says

    A listener wrote to send me an article demonstrating that CaGluc doesn’t require hepatic breakdown to achieve its ionized form. Great job myth busting, Greg! Article is
    Anesthesiology 1990;73:62.

  6. emcrit says

    Have gotten a few questions regarding Digoxin and Hyperkalemia. Here is the answer from one of my crit care pseudo-fellows:

    On Tue, Oct 5, 2010 at 5:17 AM, Amanda Holland Yang wrote:

    So, according to Levine in the journal of emergency medicine, don’t worry about the calcium… they did a chart review of all pts diagnosed by a clincian with dig toxicity. 23 had gotten calcium, and there was no increased incidence of life threatening dysrhythmias or death. The poison center and north shore had a problem with this in their follow up letter though, arguing that maybe the patients weren’t truly dig toxic, just had elevated levels.

    Levine says that if someone is known to be dig toxic, obviously just give them digibind rather than calcium, but if you are treating someone for hyperkalemia and don’t know that they are on digoxin, it should not be a problem. I think this sounds reasonable.

    DIGOXIN TOXICITY Journal of Emergency Medicine

  7. mark albert says

    Very good!! Many clinicians rely only on EKG as an indicator, also the tx for the fast remediation is the Ca for membrane stabiliztion, specifically the gluconate salt. The myth of grab NAHCO3 although could be use as an adjunct, depending on the clinical situation, has to be dispelled. Very often, there is an overkill, then the pendulum goes to the other side, and u may see hypokalemia- so prudent use of these pharmacological interventions and understaning of mechanism of actions, onset and duration is critical.

  8. Derek says

    I was treating a patient with hyperkalemia (7.2) and bradycardia. I gave 3 amps of calcium gluconate, which my nurses refused to push. They said they only had ever given one amp.

    The article used for this podcast states the dose of calcium is 10ml of Ca gluconate. However, if I would normally use 10ml of CaCl, then shouldn’t I give 3 amps of calcium gluconate?

    • emcrit says

      1 amp of CaCl or 3 amps of CaGluconate just as you say. Thing is CaGluconate should not be pushed. Each amp should drip in over ~10 minutes. Perhaps this is why the nurses balked?

      • Derek says

        Thanks for the quick response. I did make a mistake as I pushed it instead of dripped it in. Glad there were no adverse outcomes

        The nurses balked because they said they had never given 3 amps before, only 1. What confuses me is that the table posted on hyperkalemia above states to give 10ml (1 amp) of calcium gluconate.

        • emcrit says

          Pushing CaGluc just results in occasional hypotension. Even if it happens, you can fix it with some fluids and it will be short-lived. So no big deal. The comments Mark makes below actually mostly apply to calcium chloride which can cause irritation and thrombophelbitis in extravasation. CaGluc can actually be given sub-q for HF toxicity and is ok in extravasation.

          • emcrit says

            The dose in the table is for immediate stabilization, but I would give at least 2-3 grams (amp) total.

  9. mark albert says

    As a pharmacist, it is the medication of choice, being a membrane stabilizing agent, but pushing it to fast can lead: cardiotoxicity, hypotension, local thrombophlebitis, tingling sensation, calcium taste, flushing, nausea, vomiting, sweating and
    irritation and potential for extravasation (especially if it is a peripheral stick)

  10. mark albert says

    so in essence, cagluc is the agent of choice? faster onset than the cloride salt, no hepatic breakdown to become pharmacodynamic, no tissue necrosis issues in peripherial sticks- should be on our rigs opposed to cacl then, if pts are presenting with hyperkalemia, and in arrest and the medic is doing their H and T and found a pt may have some high T waves (PEA arrests) in consult with Medical control,

    • emcrit says

      Calcium Chloride if they are crashing because you can slam it in, but you better have a good vein. CaCl also agent of choice if you have a central line. Pretty much everyone else, CaGluc is prob. safer.

  11. Cherinor Sillah says

    Hi Scott,
    I have been listening to your program and has change my perspective in medicine. I am family medicine trained with fellowship in sports medicine. I do sports medicine in my newly open private practice and work some nights in the ER. I listen to your podcast on hyper k . Great. You made a comment about given insulin subcutaneously. Do you give insulin subq or IV in the ER in patient that has hyperglycemia after fluid and normal potassium that is not in Dka or hyperosmolar nonketotic hyperglycemia?
    Thanks for your respond in advance

    C. Sillah
    Macon , Ga

  12. Jeff Siegler says


    Can Calcium Chloride be given safely through an IO or is it risky like peripheral stick?



    • says

      Party line is anything that can go IV can go IO, but you better be pretty sure of your IO being properly placed before giving CaCl. May be safer to use gluconate in these cases–that’s what I do unless it is a cardiac arrest.

  13. Chris says

    Hey Scott,

    Wicked podcast as always, and I’m wondering about another possible myth with digoxin toxicity. I’ve been told about the “irritable myocardium” and no placement of CVLs other than femorals, as you could precipitate VF by tickling the already unstable myocardium with an IJ or subclavian. I can’t find any literature on this, do you know anything to support it?



  14. Chris says

    Hey Scott,

    After looking it up and checking Rosen’s and Goldfrank’s, I came upon this article, which seems to be the basis for the argument of no IJs/Subclavians:

    CLINICAL TOXICOLOGY7 31(2), 261-273 (1993)

    I’ve attached the key excerpt from the discussion here as well:

    “Safety of cardiac pacing in the treatment of digitalis intoxication was
    assessed in 39 pacing-treated patients from Groups 1 and 3. Fourteen adverse effects (36%) were recorded. These iatrogenic accidents were pacing-induced arrhythmias (6 cases), pacing defects (6 cases), and infectious complications (2 cases). The six pacing-induced arrhythmias occurred during or just after insertion (1 ventricular tachycardia, 3 VF) or subsequent to pacemaker adjustment (1 VA after a brief pause of pacing to study the underlying rhythm; 1 VF during reduction of ventricular rate from 80 bpm to 60 bpm). The six pacing defects occurred after ambulance transport (1 VA), external cardiac massage (1 patient), or accidental removal of the pacemaker by a confused
    patient, while no causes were found in three cases. The two infectious
    complications were staphylococcus epidermidis septicemias. One septicemia was complicated by fatal septic shock. Five out of these accidents (13 %) had a fatal outcome (2 VF, 2 VA, 1 septic shock). Thus, the overall pacing-induced mortality was 42% (5 out of 12 fatalities).”

    So, they showed that pacing someone with a dig OD is bad, and that complications of pacing often occurred “shortly after insertion”, which they don’t define in the paper in terms of time. I presume this leads to the inference that it may be the guide wire of the CVC kit that irritates the myocardium and precipitates the arrythmia/death. I’m not sure I can completely make that leap, as it seems to me that repeatedly shooting electricity into the already electrically abnormal heart will be a bigger problem than tickling it with a wire.
    Some other things to consider are that the pacers were put into patients receiving lower/no digibind, and that the pacers were put in ahead of time at another hospital, and patients were then transferred to the ICU center in the study. Interestingly the cardiology center put in the pacers, but had no Digibind on hand. The digibind was only available at the ICU center where the patients would be transferred after having their pacer inserted.

    Alright, overall I’m feeling like a CVC is probably fine, especially in an appropriately Digibound patient. However, pacing (especially without giving Digibind) looks to be a suboptimal plan.


  15. Decal says

    I know I am a little late to this post but I have seen references give doses up to 3 gms of Calcium Gluconate IV Push or 1 gram of Calcium Chloride. What I was trying to find if there was data to support these doses? I have tried to look at the AHA CPR guidelines back to 2000 and there is no reference of primary literature supporting them.

    Thanks for the discussion and great job with the podcast.

  16. Shaz says

    A patient with hyperkalemia and third degree AV block. which one should be the first priority- calcium gluconate or pacemaker??

  17. Jim Feldman says

    In regards to the rate of administration of calcium-from the table in reference cited:
    the AHA guidelines reference:
    Shift potassium into cells: http://circ.ahajournals.org/content/112/24_suppl/IV-121.full

    Calcium chloride (10%): 500 to 1000 mg (5 to 10 mL) IV over 2 to 5 minutes to reduce the effects of potassium at the myocardial cell membrane (lowers risk of ventricular fibrillation [VF])

    Although pharmacy guidelines may very, I believe that it is useful to reference the AHA source in support of one’s decision making.

    In terms of the evidence base that guides decision making in level and treatment, evidence based reviews indicate lack of evidence for many treatments

    example- Two studies investigated infusion of sodium bicarbonate for the management of hyperkalemia30,31 (Appendix 5). One reported a small absolute reduction in serum potassium from baseline of 0.47 (standard deviation [SD] 0.31) mmol/L (p = 0.001) at 30 minutes;31 in the other trial, there was no significant change in serum potassium at 30 and 60 minutes, relative to placebo.30 The trials of bicarbonate infusion had the lowest reduction in mean serum potassium among all single and combined modalities, including salbutamol and insulin.

    Management of patients with acute hyperkalemia
    Meghan J. Elliott, Paul E. Ronksley, Catherine M. Clase, Sofia B. Ahmed, Brenda R. Hemmelgarn
    CMAJ. 2010 October 19; 182(15): 1631–1635. doi: 10.1503/cmaj.100461
    PMCID: PMC2952010

  18. J Kim says

    I am glad you mentioned about using SC insulin rather than IV push insulin for routine hyperglycemia management, thank God! I’ve felt the same way for long time and it never made sense to see other docs pushing IV insulin to treat pure chronic hyperglycemia in ED. It never made pharmacokinetic sense to me. We don’t tell patients to get PICC line and push IV insulin at home forever for their diabetes every few hours during day and why are we pushing IV insulin for chronic hyperglycemia? It’s like pushing IV antihypertensives in ED to every anxious chronic hypertension patient who comes to ER every other week for hypertension after checking BP at Walmart automatic blood cuff machine during grocery shopping, as if that IV medicine at ER every other week is extremely essential to manage Pt’s chronic HTN. Thank you for validating my pet peeve and great blog again!


  1. […] tedavisinin mükkemmel bir analizi için Scott Weingart’?n ses yay?n?na göz atman?z? tavsiye […]

Speak Your Mind (Along with your name, job, and affiliation)