Truace
Subscribe

AI heart tracing misses low potassium more often than bedside blood tests in dialysis patients

Study title: Comparison of an AI-enabled single-lead ECG monitor and point-of-care testing for detecting intradialytic hypokalemia

Researchers tested an AI model that estimates blood potassium from a single-lead ECG in people receiving maintenance hemodialysis. They compared it with point-of-care blood testing, using central laboratory results as the reference standard.

Read the study abstract

Background: Intradialytic hypokalemia is an underrecognized complication of maintenance hemodialysis (MHD). While point-of-care (POC) testing offers rapid biochemical confirmation, it remains limited by its intermittent nature. Artificial intelligence (AI)-enabled electrocardiography (ECG) provides a potential noninvasive alternative for trend monitoring; however, its performance in MHD has not been compared with POC testing against a central laboratory standard. Methods: This single-center prospective study com…

By the Truvace desk1 min read1 sourceEvidence: Moderate · 64/100 — Controlled or validated comparison · n = 74 · 1 independent sourcePeer-reviewedTRV-2026-2321· cite & verify
Comparison of an AI-enabled single-lead ECG monitor and point-of-care testing for detecting intradialytic hypokalemia
Kidney dialysis patients : a population at undue risk? : hearing before the Special Committee on Aging, United States Senate, One Hundred Sixth Congress, second session, Washington, DC, June 26, 2000 by United States. Congress. Senate. Special Committee on Aging. Public domain

In brief

The AI tool tracked normal potassium levels reasonably well and could flag some low readings, but bedside testing was much more precise. Whether the tool improves patient outcomes or can safely replace some blood tests is not yet known and needs further study.

Main points

  1. The algorithm was trained on ECG-potassium pairs from 45 patients (2021-2022) and tested in 29 patients (2023).
  2. Mean absolute error was 0.479 mmol/L in the non-hypokalemic range but 1.000 mmol/L during hypokalemia.
  3. Point-of-care testing had a mean absolute error of 0.201 mmol/L and a correlation of 0.977 with laboratory values.
  4. The AI-ECG model reached an AUC of 0.722 for hypokalemia screening, versus 0.996 for point-of-care testing.

The problem

The AI-enabled ECG was far less accurate than point-of-care testing for diagnosing low potassium (AUC 0.722 vs. 0.996), and its error rose to 1.000 mmol/L during hypokalemia.

The rundown

The study was a single-center prospective comparison during maintenance hemodialysis. Concurrent laboratory, point-of-care and ECG measurements allowed direct comparison against central laboratory assays.

Point-of-care testing gave the closest match to laboratory results, but it is limited by being intermittent. The AI-ECG approach is noninvasive and could support continuous trend monitoring.

The authors describe the AI model's screening potential as preliminary. They state that formal prospective decision-impact studies are needed to show it improves clinical outcomes or safely reduces invasive testing.

Sources

  1. Peer-reviewedRenal Failure2026-10-30

The debate