Illustration: A translucent heart inside a pale shield of light, copper spheres pressing against the shield from outside.
← EMA Cited
EMA Cited · Electrolytes

Hyperkalemia

Updated Oct 15, 2026Adults · ED management21 sources · 15 checked in PubMed5 figures

Bottom line

  • Read the ECG, not just the number. In severe hyperkalemia, QRS widening, heart rate under 50, and junctional rhythm predicted short-term adverse events. Peaked T waves did not.3 A normal ECG does not rule out danger.4
  • ECG changes mean IV calcium first. It works in minutes, wears off within about an hour, and does not lower potassium.1
  • Shift potassium with insulin plus glucose and nebulized albuterol 10–20 mg. Expect a drop of roughly 0.7–1.2 mEq/L.1,5
  • Hypoglycemia is the most common treatment harm. It typically appears around 2 hours after insulin, so check glucose for 6 hours.1,12
  • Remove potassium with a binder (SZC) or dialysis. Sodium polystyrene sulfonate (Kayexalate) is slow, unpredictable, and linked to bowel injury.1,16
Watch · 0:29

Hyperkalemia in 30 seconds

Every spoken line and on-screen number comes from a cited line in this article. The pictures are AI illustrations, not clinical images.

Severity

Mild5.5–5.9mEq/L
Moderate6.0–6.4mEq/L
Severe≥ 6.5mEq/L

Bands from the UK Kidney Association, which uses the European Resuscitation Council definition.1 No single treatment threshold is agreed in US emergency practice.2 For potassium, 1 mmol/L equals 1 mEq/L.

First 10 minutes

  1. Monitor and get a 12-lead ECG. Get an ECG for any potassium ≥ 6.0. Use continuous cardiac monitoring for ≥ 6.5, for any ECG changes, or for 6.0–6.4 in a patient who is unwell or whose level is rising.1
  2. Confirm fast. Run potassium on a blood gas analyzer while the lab result is pending.1 A high level with a normal ECG should raise the question of a spurious result.19
  3. ECG changes or instability: give IV calcium now.1,19
  4. Shift: insulin with glucose, plus nebulized albuterol for severe hyperkalemia.1
  5. Plan removal and find the cause. Start a binder. In anuric or dialysis patients, call for dialysis early.19,1
  6. Recheck potassium at 1, 2, 4, 6 and 24 hours, aiming for < 6.0 within 2 hours. Follow the glucose schedule below.1
FIG 1 · Decision flowCited
  1. Potassium ≥ 6.0. Get a 12-lead ECG, start monitoring and confirm on a blood gas analyzer.1 A high level with a normal ECG and normal renal function may be spurious.19
  2. ECG changes or instability?
    Yes
    IV calcium now: gluconate 10% 30 mL over 10 min, or chloride 10% 10 mL over 5 min in arrest or peri-arrest.1,19
    Repeat if ECG changes persist.1
  3. Shift: regular insulin 10 units IV with 25 g glucose, plus nebulized albuterol 10–20 mg.1
  4. Pre-treatment glucose under 126 mg/dL?
    Yes
    Run D10 at 50 mL/h for 5 hours.1
  5. Check glucose at 0, 30, 60, 90, 120, 180, 240, 300 and 360 minutes after insulin.1
  6. Dialysis patient, anuric, or not responding?
    Yes
    Call for dialysis early. Urgent at K ≥ 6.5 in dialysis patients and for refractory cases.1,19
  7. Remove: SZC 10 g PO three times daily for K ≥ 6.5; consider it for 6.0–6.4.1
  8. Recheck potassium at 1, 2, 4, 6 and 24 hours, aiming for < 6.0 within 2 hours.1
FIG 1The first hours of hyperkalemia. Drawn for this page from the UK Kidney Association guideline and REBEL EM. Each box carries its own citations.1,19

ECG

Typical changes are peaked T waves, flattened P waves with a longer PR interval, a widening QRS, and finally a sine wave and arrest. They do not always appear in that order, and several can appear at once.1 See the LITFL ECG library for more examples.17

FIG 2How the ECG changes, in schematic. Drawn for this page from the sequence the UK guideline describes. It is not to scale, and real tracings don't follow the order reliably or match fixed potassium levels.1
12-lead ECG at potassium 7.6 with tall narrow T waves in the inferior and precordial leads
FIG 3Potassium 7.6. Tall, narrow T waves, clearest in II, III, aVF and V3–V6.James Heilman, MD · Wikimedia Commons · CC BY-SA 3.0
12-lead ECG at potassium 8.2 in a patient with kidney disease, showing peaked T waves across V2 to V6
FIG 4Potassium 8.2, kidney disease. Peaked T waves dominate V2–V6. Note how ordinary the rhythm strip looks at a glance.M-J. F. Agbayani & E. Gonzales · Wikimedia Commons · CC BY 4.0
Finding (K ≥ 6.5, n = 188)Adverse events within 6 h
Heart rate < 50RR 12.3 (6.7–22.6)
Junctional rhythmRR 7.5 (5.3–11.1)
QRS prolongationRR 4.7 (2.0–11.2)
Peaked T wavesRR 0.77 (0.35–1.70), not significant

Single-center retrospective study. Adverse events (mostly symptomatic bradycardia) occurred in 15% of patients. Every patient who had one also had at least one hyperkalemic finding on the preceding ECG.3

The ECG is insensitive. In 90 admitted patients with potassium ≥ 6, the ECG missed many cases, and readers agreed only moderately. The authors advise managing by the clinical picture and serial potassium levels, not by the ECG alone.4

Treatment

AgentAdult doseOnsetLastsNotes
1 · Protect the heart
Calcium gluconate 10%30 mL (3 g) IV over 10 min~3 min30–60 minUse outside cardiac arrest. One 10 mL ampule is only a third of this dose.1,19
Calcium chloride 10%10 mL (1 g) IV over 5 min~3 min30–60 minPreferred in arrest or peri-arrest. Same calcium (6.8 mmol) as 30 mL gluconate.1
2 · Shift potassium into cells
Regular insulin + glucose10 units IV + 25 g glucose (50 mL D50)< 15 min, peak 30–60~2 h, then reboundLowers potassium 0.65–1.0. If pre-treatment glucose < 126 mg/dL, follow with D10 at 50 mL/h for 5 h.1
Albuterol, nebulized10–20 mg (4–8 standard nebs)≤ 30 min≥ 2 hUse as an add-on only. 20–40% of patients drop less than 0.5. Beta-blockers blunt the effect.1,7,19
Sodium bicarbonateNot routine——No effect on potassium in pooled data.1,5
3 · Remove potassium
Sodium zirconium cyclosilicate (SZC)10 g PO TID, up to 72 h< 1 h—Give for K ≥ 6.5. Consider for 6.0–6.4.1
Patiromer8.4 g PO daily4–7 h—An option for K ≥ 6.0.1
Sodium polystyrene sulfonate (Kayexalate)Not recommended acutely> 4 h—Unpredictable effect. Risk of intestinal injury.1,16
DialysisNephrology or critical care decision—rebound afterDefinitive treatment. Urgent for dialysis patients with K ≥ 6.5 and for refractory cases.1

Unreviewed doses. Verify against the current label and your institution's protocol before giving anything.

All potassium values in mEq/L. Combined, insulin-glucose and albuterol lower potassium about 0.7–1.2 mEq/L (low to very low certainty).5

FIG 5When each treatment works. Calcium wears off before insulin peaks, and the hypoglycemia window outlasts insulin's effect on potassium. Dashed lines mark the recommended potassium rechecks. Times are approximate and vary between patients.1,12 The values are the same as in the table above.

Calcium

  • Repeat the dose if ECG changes persist, or if potassium stays high after the effect wears off.1
  • Give calcium for ECG changes even when dialysis is already planned.1
  • The evidence is weak. An ILCOR review found no evidence of clinical benefit. The UK guideline downgraded its recommendation in July 2026 but still advises giving calcium.5,1
  • Digoxin toxicity is not a contraindication. In 23 digoxin-toxic patients given calcium, no life-threatening dysrhythmias occurred within an hour, and mortality matched patients who didn't get calcium (22% vs 20%).6 The "stone heart" warning rests on a handful of case reports.18
Sources disagree

How long calcium lasts: about 30–60 minutes per the UK Kidney Association1, and 20–30 minutes per REBEL EM.19 Plan potassium-lowering treatment as if the shorter figure is right.

Binders

  • In ED patients (n = 70), adding SZC to insulin-glucose gave a larger drop at 2 hours (−0.72 vs −0.36 mEq/L). At 4 hours the difference was small and not statistically clear (−0.41 vs −0.27).14
  • In outpatients (mean K 5.6), SZC brought potassium to normal in a median of 2.2 hours.15
  • In adults aged 66 and older, Kayexalate was associated with roughly double the risk of a serious GI event within 30 days (HR 1.94). The absolute numbers were small.16
  • Loop diuretics can help patients who still make urine and are volume replete. This advice comes from the chronic-management section of the UK guideline.1
Histology slide showing angular purple Kayexalate crystals within inflammatory exudate next to a perforated colon
FIG 6Kayexalate crystals in a colonic perforation. The angular, purple, striated fragments are resin crystals within inflammatory exudate.Atlas of Medical Foreign Bodies · Wikimedia Commons · CC BY-SA 2.0

Avoiding hypoglycemia

  • Who is at risk: patients without diabetes, a low pre-treatment glucose, female sex, low body weight, and renal dysfunction.13,12
  • When it happens: in dialysis patients, at a median of 2 hours after insulin, lasting a median of 2 hours.12 Across 62 studies, the median onset was 124 minutes.1 Many patients have left your attention by then.
  • Check glucose at 0, 30, 60, 90, 120, 180, 240, 300 and 360 minutes after insulin.1
  • Pre-treatment glucose under 126 mg/dL (7 mmol/L): run D10 at 50 mL/h for 5 hours after the bolus.1

Lower insulin doses

  • In 675 ED patients with renal insufficiency, 5 units vs 10 units: hypoglycemia 19.5% vs 28.6%, with the same potassium drop (−1.0 mEq/L).8
  • A meta-analysis of 10 retrospective cohorts (n = 3,437) found that 5 units, 0.1 units/kg, or anything under 10 units cut the odds of hypoglycemia (OR 0.55) and severe hypoglycemia (OR 0.41), with no difference in potassium lowering.9
  • In a 700-patient ED cohort, doses under 10 units also caused less hypoglycemia (11.2% vs 17.6%), but lowered potassium somewhat less (0.94 vs 1.11 mEq/L).10
  • Weight-based dosing used 0.1 units/kg, capped at 10 units.11
Guideline position

The UK Kidney Association keeps 10 units. It judges the evidence insufficient to change the insulin dose, and finds that giving more glucose reduces hypoglycemia more consistently than giving less insulin.1 All the lower-dose data are retrospective.9

Dialysis patients

  • Dialyze urgently at K ≥ 6.5.1
  • Give calcium for toxic ECG changes even when dialysis is immediately available.1
  • Use standard medical therapy while waiting. Up to 40% of patients with end-stage renal disease do not respond to albuterol.1
  • Expect potassium to rebound after a dialysis session.1

Cardiac arrest

Give calcium chloride when hyperkalemia is known or suspected as the cause. This is a weak recommendation (grade 2C).1 The ILCOR review found no evidence that calcium improves clinical outcomes.5

Pseudohyperkalemia

A falsely high result is common. REBEL EM calls it the leading cause of a high potassium.19 Suspect it when renal function and the ECG are normal, or when the patient has a blood disorder.1

  • Causes: delayed or cold transport, difficult draws, hemolysis, high platelet counts, erythrocytosis, EDTA carry-over from the wrong order of draw, and drawing from an arm with a potassium-containing infusion.1
  • Confirm it with paired lithium-heparin (plasma) and clotted (serum) samples from a large vein. A serum value more than 0.4 above the plasma value confirms pseudohyperkalemia.1

Causes

  • Reduced excretion: renal failure. Drugs include ACE inhibitors, ARBs, spironolactone, NSAIDs, and trimethoprim.19,1
  • Shift out of cells: acidosis, succinylcholine, and digoxin toxicity.19,6
  • Cell death: tumor lysis, rhabdomyolysis, burns, crush injury, and hemolysis.19

Disposition

  • Severe (≥ 6.5) and stable: admit to a bed with continuous cardiac monitoring.1
  • Airway, breathing, circulation or consciousness problems: refer to critical care.1
  • Refractory to medical therapy: nephrology or critical care decides urgently on renal replacement therapy.1
  • Found incidentally in a well patient: recheck mild levels within 3 days and moderate levels within 1 day. Assess severe levels in hospital.1

Pitfalls

  • Ordering "one amp" of calcium gluconate gives a third of the guideline dose.1,19
  • Giving albuterol at asthma doses. Hyperkalemia needs 10–20 mg.7,19
  • Handing off or discharging before the glucose low point, around 2 hours after insulin.12
  • Being reassured by the absence of peaked T waves. Watch the QRS, the rate, and the rhythm.3
  • Missing the rebound. Shifted potassium comes back as treatment wears off, so recheck at 4 and 6 hours.1
  • Using milk of magnesia or Fleet enemas, which risk magnesium and phosphate toxicity.20

How strong is the evidence?

Almost every acute treatment rests on low or very low certainty evidence.5 Calcium is supported mainly by physiology and case reports.1 Insulin dosing data are retrospective.9 The expert panel convened by the American College of Emergency Physicians calls the common drugs temporizing at best.2

On-shift teaching

Questions to ask a learner at the bedside. Tap a question to see the answer. Every answer uses facts cited above.

Q1In severe hyperkalemia, which ECG findings actually predicted adverse events in the next 6 hours?
Heart rate under 50 (RR 12.3), junctional rhythm (RR 7.5) and QRS prolongation (RR 4.7). Peaked T waves did not (RR 0.77, not significant).3
Q2How many 10 mL ampules of 10% calcium gluconate equal one ampule of 10% calcium chloride?
Three. 30 mL of gluconate and 10 mL of chloride each deliver about 6.8 mmol of calcium. That's why "one amp" of gluconate is a third of the guideline dose.1,19
Q3Why do patients become hypoglycemic hours after insulin for hyperkalemia, and how long should you watch?
Insulin can outlast the dextrose given with it.19 The median onset is about 2 hours (124 minutes across 62 studies), so check glucose at 0, 30, 60, 90, 120, 180, 240, 300 and 360 minutes.1,12
Q4A digoxin-toxic patient has a wide QRS and a high potassium. Is calcium contraindicated?
No. In 23 digoxin-toxic patients given calcium, none had a life-threatening dysrhythmia within an hour, and mortality matched those who didn't get it (22% vs 20%).6 The "stone heart" warning rests on a handful of case reports.18
Q5Why shouldn't albuterol be used on its own?
Its effect is unpredictable: 20–40% of patients drop less than 0.5 mEq/L, up to 40% of patients with end-stage renal disease don't respond, and beta-blockers blunt it. Use it alongside insulin-glucose.1
Q6How do you prove a high potassium is spurious?
Send paired lithium-heparin (plasma) and clotted (serum) samples from a large vein. A serum value more than 0.4 above the plasma value confirms pseudohyperkalemia.1

References

Evidence base

What this page is built on

21 sources, cited 111 times. Dots are sources, filled by how far each was read; stronger color means a larger share of the citations. Tap a tier to see its sources.

Read in full · 2Abstract read · 15Record only · 0Link checked · 4

Checked by the writer, not a physician. The dots show how each source was read, not that anyone has confirmed the claims.

How this page was made
  • Written from scratch. No source text was copied; facts are restated and cited.
  • Each PubMed citation was resolved through NCBI E-utilities, and each claim was checked against that paper's abstract or the guideline's full text.
  • Sources searched: PubMed, the UK Kidney Association guideline, LITFL, REBEL EM, Emergency Medicine Cases. emDocs pages were found but could not be fetched, so they are not cited.
  • Figures 1 and 4 were drawn for this page from cited data. Figures 2, 3 and 5 are openly licensed images from Wikimedia Commons, credited under each one, and they keep their own licenses.
  • Not for clinical use.
Not for clinical use

AI-generated draft. Not reviewed by a physician. This page is for education only and may contain errors, including in doses. Do not use it to diagnose, treat or make decisions about any patient. Rely on your own clinical judgment, your institution's protocols and current drug labeling. Reading this page does not create a physician-patient relationship. Full terms