Updated July 2026 · 9 min read
Part of the ICU Emergencies Hub — browse every related guide in one place.
A critical potassium comes back on a patient who looks fine, and the clock starts. Hyperkalemia is one of the few labs that can stop a heart in minutes, and the treatment is a specific, ordered sequence that every ICU nurse should be able to run from memory. The order matters as much as the drugs — do it out of sequence and you've protected nothing.
Everything about hyperkalemia treatment makes sense once you hold the framework: stabilize, shift, eliminate. Calcium stabilizes the cardiac membrane so a lethal rhythm is less likely while you work — it does not lower the potassium at all. Insulin, albuterol, and (in the acidotic patient) bicarbonate temporarily push potassium from the blood into cells, dropping the serum number for a few hours — but the potassium is still in the body and will come back out. Only binders, diuresis, and dialysis remove potassium. A treatment that stops at "shift" is a treatment that will fail in a few hours, which is why disposition — especially the dialysis question — is part of the plan from the start.
The number matters, but the ECG tells you how much time you have. Hyperkalemia changes the ECG in a roughly progressive order, and any of these findings turns a lab abnormality into an emergency:
| Roughly as K+ rises | ECG change |
|---|---|
| Earliest | Peaked, narrow-based T waves |
| Next | Flattened/absent P waves, PR prolongation |
| Then | Widening QRS |
| Late / peri-arrest | Sine-wave pattern, then VF or asystole |
Calcium raises the threshold potential and restores the electrical gradient across the cardiac cell membrane, making a lethal arrhythmia less likely. It works within minutes but lasts only about 30–60 minutes, and it does nothing to the potassium level — it buys time, it doesn't fix the problem. You'll give either calcium gluconate (peripheral-friendly) or calcium chloride (about three times the elemental calcium, but caustic and best through a central line). Give it whenever there are ECG changes or the potassium is dangerously high. The classic caution is the patient on digoxin, where calcium was traditionally given more cautiously — know your protocol. For the chloride-vs-gluconate and line decision, see the calcium chloride vs gluconate comparison.
Insulin + dextrose is the backbone shift. Insulin drives potassium into cells within 15–30 minutes and lasts a few hours; you give dextrose with it to prevent hypoglycemia (a regular IV insulin dose with an amp or infusion of dextrose, unless the patient is already hyperglycemic). Check the glucose — hypoglycemia is the most common and most dangerous complication of hyperkalemia treatment, and it can occur an hour or two later, so monitor beyond the initial check.
Albuterol (nebulized, at a dose higher than a routine breathing treatment) also shifts potassium into cells and works additively with insulin. It causes tachycardia, so it's used thoughtfully in the cardiac patient. Sodium bicarbonate has a limited, selective role — it may help the patient who is acidotic, but it's not an effective standalone shift in most patients and shouldn't be relied on. For dextrose specifics, see the D50 and hypoglycemia guide.
This is the only step that actually lowers total-body potassium, and it's where the real disposition decision lives.
Dialysis is the definitive removal, and it's the answer for the patient in renal failure, the severely elevated potassium, or the patient who isn't responding — if the kidneys can't excrete potassium, the temporizing measures are just a bridge to the dialysis machine. Loop diuretics (furosemide) increase renal potassium excretion in the patient who still makes urine and has functioning kidneys. Potassium binders in the GI tract remove potassium over hours: the newer agents (patiromer, sodium zirconium cyclosilicate) are better tolerated and faster than the old sodium polystyrene sulfonate (Kayexalate), which works slowly and carries a real risk of bowel injury. Binders are for ongoing and subacute control, not the crashing patient.
While you're treating the number, someone has to answer why. Common drivers include renal failure (the big one), medications (ACE inhibitors, ARBs, potassium-sparing diuretics, potassium supplements, NSAIDs), tissue breakdown (rhabdomyolysis, tumor lysis, hemolysis, crush injury), acidosis, and adrenal insufficiency. And rule out the pseudohyperkalemia of a hemolyzed or difficult draw before you commit to an aggressive workup — but never let "it might be hemolyzed" delay treatment in a patient with ECG changes. Stop the offending potassium sources (hold the supplement, the ACE inhibitor, the potassium-containing fluids) as part of the treatment, not after it.
Hyperkalemia is a choreography problem. You're getting the patient on a monitor and watching the ECG evolve, pushing calcium the moment it's warranted, giving insulin/dextrose and albuterol and then chasing the glucose for hours, calling the renal team early if dialysis is on the table, and making sure the removal step doesn't get forgotten once the number looks better. The single most common failure isn't giving the wrong drug — it's stopping at the shift, watching the potassium drift back up, and being surprised. Stabilize, shift, eliminate, and find the cause: run all four and the emergency stays an emergency you controlled.
Get the ICU Notebook
Free investing strategies built for nurses. One email per week, no fluff.
Yes, send it freeNo spam. Unsubscribe any time.