Challenges in Clinical Electrocardiography
March 13, 2023
Computer-Estimated Heart Rate vs True Heart Rate—Be Aware of the Software
Laszlo Littmann
JAMA Intern Med. Published online March 13, 2023. doi:10.1001/jamainternmed.2022.6613
Apatient in their 60s was admitted to the intensive care unit for gastrointestinal bleeding and shock associated with liver cirrhosis, portal hypertension, and ruptured esophageal varices. Emergency treatment included balloon tamponade of the esophagus, blood transfusions, octreotide, and vasopressin. Despite fluid resuscitation and the use of pressors, the blood pressure continued to fall, and the kidney function progressively deteriorated. Computer interpretation of a 12-lead electrocardiogram (ECG) indicated supraventricular tachycardia with right bundle-branch block and right superior QRS axis (Figure, A). The interpretation software also reported frequent premature ventricular complexes. The patient’s baseline ECG recorded a few months earlier was normal.

Questions: Based on the clinical history, ECG findings, and their computer interpretation (Figure, A), what was a likely contributing factor to the patient’s refractory hypotension and shock? What could have been a simple life-saving intervention?
The 12-lead ECG (Figure, A) showed sinus tachycardia at 115 beats per minute (bpm). Upright P waves were noted in most of the leads with negative and biphasic P waves in V1 and V2. The PR interval was normal, and the QRS complexes were slightly widened at 120 milliseconds (ms). In V1 and V2, there was a slightly atypical right bundle-branch block pattern with the initial upgoing components of the QRS complexes as tall or slightly taller than the terminal upgoing components. The frontal plane demonstrated a right superior QRS axis. Importantly, the interpretation software calculated a ventricular rate of 230 bpm, twice the true heart rate. There were no premature ventricular complexes. In severely ill patients, heart rate overcounting by computerized software, especially in combination with wide QRS complexes, is a strong indication of possible severe hyperkalemia.1-3
The large gap between the computer-estimated heart rate and the true heart rate and its significance were not recognized. Approximately 40 minutes later, the patient experienced a near-asystolic cardiac arrest. Telemetry revealed a very slow wide-complex rhythm without P waves (Figure, B). The QRS morphologic findings were similar to the lead II QRS morphologic findings in the previous 12-lead recording, but the QRS complexes had further widened from 120 ms to approximately 180 ms. There were prominent and tented T waves. The true ventricular rate was 26 to 30 bpm, and in this instance too, the computer erroneously calculated a heart rate of 61 bpm. The cause of double counting was obvious because computer-generated annotation marks V were seen above both the QRS complexes and T waves. The software interpreted the T waves to be ventricular complexes.
The patient underwent protocol-driven cardiac resuscitation, which included several rounds of intravenous calcium gluconate. Spontaneous circulation returned within minutes. An electrolyte panel drawn just minutes before the cardiac arrest showed a serum potassium level greater than 9.0 mmol/L (normal, 3.6-5.4 mmol/L). The patient had a prolonged hospital stay that included placement of a transjugular intrahepatic portosystemic shunt. Eventually, the patient survived to hospital discharge.
Hyperkalemia affects up to 10% of hospitalized patients, and in the intensive care setting, the percentage is probably higher. The signs and symptoms of critical illness and hyperkalemia—hypotension, shock, kidney failure, and cardiac dysrhythmias frequently overlap, and therefore clinically, even severe hyperkalemia can be overlooked.4 The lethal effects of hyperkalemia are more closely associated with the rate of rise in serum potassium and the effect of hyperkalemia on the ECG findings than on the absolute laboratory value.5,6 Unfortunately, physicians have poor ability to recognize ECG signs of hyperkalemia.7
In addition to peaking of the T waves, there are several less appreciated ECG signs that should prompt practitioners to consider hyperkalemia. These include widening of the QRS complexes, axis shifts, flattening of the P waves, pseudoinfarction pattern and Brugada-type ST-elevation in the anterior chest leads.8 Most of these changes are associated with hyperkalemia-induced transmembrane sodium-channel blockade. The P waves can be absent even during supraventricular rhythms, a condition referred to as sinoventricular conduction. Wide QRS complex pulseless electrical activity and nonshockable regular wide-complex tachycardia are frequently indicators of extreme hyperkalemia and should prompt immediate treatment with intravenous calcium.9 Recently, it was found that overcounting of the heart rate by ECG interpretation software, especially when combined with wide QRS complexes, can also be an indication of severe hyperkalemia.1-3
The cause of hyperkalemic heart rate double counting is uncertain, but it is most likely due to software not being able to distinguish between widened QRS complexes and tall T waves. The present case strongly supports this hypothesis because during cardiac arrest, both the QRS complexes and the T waves were marked as ventricular complexes (Figure, B). Double counting of the heart rate has been seen with different ECG systems using a variety of computer algorithms.1-3 Overcounting of the heart rate is not fully specific for the diagnosis of hyperkalemia given that pacemaker spikes, artifact, tall P waves, and physiologically tall T waves too can be mistaken for QRS complexes. However, in a severely ill patient, heart rate overcounting, especially when combined with QRS complex widening, should prompt consideration of severe hyperkalemia. In such cases, hemodynamic deterioration or severe dysrhythmia may warrant emergency treatment with intravenous calcium, even when confirmatory laboratory results are not yet available.
Mistakes made by interpretation software can be troublesome and may lead to inappropriate treatments. The present case demonstrates that occasionally, recognizing certain software errors can actually provide clues to important clinical diagnoses.
- Double counting of the heart rate by interpretation software, especially when combined with wide QRS complexes and axis shifts, points to the possibility of severe hyperkalemia.
- In critically ill patients who experience hemodynamic deterioration or dysrhythmia, recognition of heart rate overcounting may warrant emergency treatment with intravenous calcium, even when confirmatory laboratory results are not yet available.
- Practitioners should not routinely ignore computerized ECG interpretations because even erroneous readings can occasionally provide important diagnostic clues.