Assessment of heart rate measurements obtained from smartphone electrocardiography as compared to 24-h Holter monitoring in dogs with atrial fibrillation

Abstract

Introduction/Objectives

The aim of this study was to evaluate the ability of smartphone electrocardiogram (ECG) to estimate heart rate (HR) in dogs with atrial fibrillation (AF) and assess its agreement with the 24-h mean HR obtained from continuous ambulatory electrocardiography (Holter monitoring).

Animals, Materials and Methods

Eleven dogs were fitted with a Holter monitor, while owners recorded 5-min ECGs using smartphone application–based electrodes, both attached and manually placed. Recordings were analyzed for the application’s QRS detection accuracy. Bland-Altman and bias plots compared smartphone-derived HR with Holter-derived HR. Receiver operating characteristic analysis assessed the smartphone ECG’s ability to differentiate between 24-h mean HRs above or below 125 beats per min (bpm).

Results

The smartphone ECG underestimated HR due to undersensing of QRS complexes by a mean of 7.15 bpm (standard deviation = 22.13) for the attached electrode and 2.6 bpm (standard deviation = 6.09) for the manually placed electrode. Bland-Altman analysis showed poor agreement, with a differential bias of −34.69 bpm (95% confidence interval = −73.25, 3.86) and proportional bias of 1.25 bpm (95% confidence interval = 0.988, 1.50). Receiver operating characteristic analysis identified a smartphone HR threshold of 122.5 bpm with 100% sensitivity and 99.96% specificity for distinguishing 24-h mean HRs above or below 125 bpm.

Study Limitations

The limitations of this study were the small sample size of large-breed dogs and limited generalizability to all AF etiologies.

Conclusions

Smartphone ECG underestimated HR in dogs with AF and showed poor agreement with Holter-derived HR. However, it may help identify dogs with 24-h mean HR above or below 125 bpm.