TOPLINE

In a large retrospective analysis of US emergency department (ED) patients, higher initial serum potassium levels were associated with higher short-term risks for cardiac arrest, ventricular tachycardia (VT), or ventricular fibrillation (VF) as well as ICU admission, intubation or mechanical ventilation, and death within 48 hours. Borderline potassium elevation was also associated with worse overall short-term outcomes, although the rates of VT and VF did not increase.

METHODOLOGY

• This retrospective cohort study used the TriNetX Research Network data on adults who visited US EDs between 2010 and 2024. Of 141,938 patients with potassium levels ≥ 5.5 mmol/L, 140,288 were included after propensity score matching.
• The initial potassium levels were used to classify patients as normokalemic (3.5-5.0 mmol/L), borderline hyperkalemic (5.01-5.49 mmol/L; exploratory group), or hyperkalemic (further classified into four tiers: 5.5-5.99, 6.0-6.49, 6.5-6.99, and ≥ 7.0 mmol/L).
• The primary outcome was a composite of cardiac arrest, VT, or VF within 48 hours of the index measurement for potassium. The secondary outcomes were ICU admission, intubation or mechanical ventilation, and death within 48 hours and hypoglycemia (blood glucose levels < 54 mg/dL) within 24 hours.
• Each hyperkalemia stratum was independently compared with propensity score-matched normokalemic control individuals.
• An exploratory analysis examined borderline potassium elevation (5.01-5.49 mmol/L), wherein 1,249,213 patients were matched to the same number of normokalemic control individuals.

TAKEAWAY

• Within 48 hours, cardiac arrest, VT, or VF occurred in 3.05% of patients with potassium levels of 5.5-5.99 mmol/L, rising to 6.68% in those with potassium levels ≥ 7.0 mmol/L compared with 0.77%-0.90% in matched control individuals (relative risk [RR], 3.97-8.08; absolute increase, 2.28-5.85 percentage points; all P < .0001).
• Hypoglycemia within 24 hours was more frequent in the hyperkalemia groups (RR, 3.32-6.55; all P < .0001), though without a consistent gradient. Mortality within 48 hours rose from 2.04% in those with potassium levels of 5.5-5.99 mmol/L to 5.79% in those with levels ≥ 7.0 mmol/L compared with 0.18%-0.21% in control individuals (RR, 9.83-27.4; all P < .0001).
• ICU admissions were reported in 8.46% of patients with potassium levels of 5.5-5.99 mmol/L, 11.23% of those with levels of 6.0-6.49 mmol/L, 13.28% of those with levels of 6.5-6.99 mmol/L, and 12.23% of those with levels ≥ 7.0 mmol/L compared with about 4% in control individuals (RR, 2.12-3.37; all P < .0001). Intubation or mechanical ventilation followed a similar pattern of requirement with increasing potassium severity with each hyperkalemia tier, at 5.00%, 6.86%, 7.85%, and 6.97% (RR, 3.79-5.84; all P < .0001).
• Among patients with borderline potassium elevation (5.01-5.49 mmol/L), the composite outcome was observed in 1.52% vs 0.85% of control individuals (RR, 1.79; 95% CI, 1.75-1.83), although the rates of isolated VT and VF did not increase.

IN PRACTICE

"These findings support the use of potassium as a readily available risk marker in acute care, but they do not establish potassium-specific causality or demonstrate that more aggressive treatment would reduce adverse outcomes," the authors wrote.

"Future studies incorporating vital signs, electrocardiographic data, serial potassium measurements, and treatment timing are needed to clarify the independent contribution of hyperkalemia to early ED outcomes," they added.

SOURCE

The study was led by Joseph O'Brien, Denver Health Medical Center, Denver, Colorado. It was published online on September 16, 2026, in The American Journal of Emergency Medicine.

LIMITATIONS

The study likely had residual confounding by acute illness severity, as it did not account for mental status, presenting complaints, ECG abnormalities, lactate levels, sepsis severity, or validated illness severity scores. Residual imbalance in creatinine and bicarbonate continued to persist after matching, particularly in the 6.0-6.49 mmol/L stratum. The 48-hour outcome window may have included events occurring after considerable changes in potassium levels or following hyperkalemia-directed treatment. Serial potassium measurements, treatment timing, and the exact timing of outcome onset were unavailable.

DISCLOSURES

The study received no funding. The authors reported having no conflicts of interest.

This article was created using several editorial tools, including AI, as part of the process. Human editors reviewed this content before publication.