Validity and reliability of an smartwatch single-lead ECG for monitoring QT interval prolongation: an interventional study with moxifloxacin
L Hoek, A H Maass, A A Voors, B Darpo, R Kleiman, J L P BrouwerAbstract
Background
QTc interval monitoring is recommended when initiating QT-prolonging therapies. Traditionally, QTc measurement relies on 12-lead electrocardiograms (ECGs) obtained during clinical visits, which limits feasibility for frequent or continuous monitoring. Commercial smartwatches can record single-lead mobile ECGs, offering potential for remote QTc monitoring with greater convenience.
Objective
To evaluate the validity and reliability of an smartwatch in detecting changes in measuring QTcF after administration of a known Qc prolonging drug, moxifloxacin, in comparison with the gold standard 12-lead ECG.
Methods
In a thorough QT/QTc study, 22 healthy subjects underwent simultaneous single-lead smartwatch ECG recordings in three positions—left wrist (SW lead I), left lower abdomen (SW lead II), and chest at the location of precordial lead V2—together with standard 12-lead ECGs recorded before and 1, 2, and 3 hours after a 400 mg oral dose of moxifloxacin. All tracings were analyzed by an independent core laboratory. For each smartwatch configuration, comparisons were made with standard ECG lead II, which is commonly used as the reference in thorough QT studies. Change from baseline (ΔQTcF) was calculated and assay sensitivity was deemed to be demonstrated in case the lower bound of the 90% confidence interval (CI) of the mean ΔQTcF was above 5 ms. Agreement between smartwatch and ECG measurements was assessed using linear mixed-effects models, intraclass correlation coefficients (ICC), and Bland–Altman analysis of mean difference and 95% limits of agreement (LOA).
Results
The subjects had a mean age of 27.4 years with 63.6% female. Mixed-effects modelling showed similar QTcF patterns across smartwatch and standard ECG recordings after moxifloxacin, with statistically significant prolongation observed for standard ECGs (p < 0.001) and smartwatch lead II (p < 0.001). Among smartwatch configurations, SW lead II demonstrated the highest ICC (0.699), indicating moderate reliability. Bland–Altman analysis showed strong agreement with the standard ECG, with mean differences of −9.47 ms ((LOA) −34.3 to +15.1 ms) for SW lead I, −4.9 ms (−26.8 to +16.9 ms) for SW lead II, and −0.15 ms (−28.1 to +27.8 ms) for SW lead V2. Assay sensitivity was only demonstrated with the standard ECG, with a mean ΔQTcF of 8.7 ms with 90% CI 6.97–10.4 ms.
Conclusions
Smartwatch ECG recordings can be considered for long-term, remote clinical QT monitoring. Due to large variability, these recordings are not yet suitable for regulatory use in phase I trials or formal QT assessments of new drugs.