Kidney Week

Abstract: TH-PO345

Role of Artificial Intelligence-Enabled Electrocardiography in the Management and Outcome Prediction of Coexisting Hyperkalemia and Bradycardia

Session Information

• Sodium, Potassium, and Volume Disorders: Clinical
  October 24, 2024 | Location: Exhibit Hall, Convention Center
  Abstract Time: 10:00 AM - 12:00 PM

Category: Fluid, Electrolytes, and Acid-Base Disorders

• 1102 Fluid, Electrolyte, and Acid-Base Disorders: Clinical

Authors

• Chen, Chien-Chou, Tri-Service General Hospital Songshan Branch, Taipei, --- Select One ---, Taiwan

Chien-Chou Chen, MD

• Lin, Chin, National Defense Medical Center, Taipei City, Taiwan

Chin Lin, PhD

• Shih, Chi Wei, Tri-Service General Hospital Department of Medicine, Taipei, Taiwan

Chi Wei Shih, MD

• Lu, Ang, Tri-Service General Hospital Department of Medicine, Taipei, Taiwan

Ang Lu, MD

• Lin, Shih-Hua P., Tri-Service General Hospital Department of Medicine, Taipei, Taiwan

Shih-Hua P. Lin, MD

Background

Concomitant hyperkalemia and bradycardia as a life-threatening urgency requires rapid management. Although artificial intelligence-enabled electrocardiography (AI-ECG) has been developed to rapidly detect hyperkalemia and bradycardia, its application in management and outcome prediction has not been evaluated.

Methods

This retrospective study was performed at the emergency department of a single medical center over 8 years. Patients with hyperkalemia>5.5 mmol/L and ECG showing bradycardia less than 50 bpm were included. Both the cardiologists and nephrologists were consulted for pacemaker placement (PP) and emergent hemodialysis (HD) evaluation. Patients' characteristics, treatment including PP and HD, and outcome were examined. AI-ECG-K⁺ quantified by ECG12Net analysis showing ECG-K⁺≥5.5 was defined as ECG-K⁺ hyperkalemia.

Results

A total of 156 patients who met the inclusion criteria had mild (5.5-5.9 mmol/L, n=59), moderate (6.0-6.4 mmol/L, n=44), and severe hyperkalemia (≥6.5 mmol/L, n=53). Their mean heart rate (HR) was 39.6±8.1 bpm, with atrioventricular block in 68 patients. Most of them had acute kidney injury, chronic kidney disease, or chronic HD. Approximately 50% of them had the drugs affecting HR. The sensitivity for AI-ECG to predict mild, moderate, and severe hyperkalemia was 74.6%, 88.6%, and 96.2%, respectively. Patients with positive ECG-K⁺ (ECG-K⁺≥5.5) had significantly higher rate of emergent PP (29.9% vs 18.2%), HD (43.3% vs 13.6%), and in-hospital mortality (19.4% vs 9.1%) compared to those with negative ECG-K⁺ (ECG-K⁺<5.5).

Conclusion

AI-ECG-K⁺ may help offer management suggestion and predict the outcomes in the patients with coexisting hyperkalemia and bradycardia, even in the mild to moderate hyperkalemia.