Accuracy of the Apple Watch iECG in Children With and Without Congenital Heart Disease

Kobel M, Kalden P, Michaelis A, Markel F, Mensch S, Weidenbach M, Riede FT, Löffelbein F, Bollmann A, Shamloo AS, Dähnert I, Gebauer RA, Paech C.Pediatr Cardiol. 2022 Jan;43(1):191-196. doi: 10.1007/s00246-021-02715-w. Epub 2021 Sep 1.PMID: 34468775


Take Home Points:

  • Apple watch 4 iECG showed good quality as compared to lead I of a standard ECG in pediatric patients with and without congenital heart disease
  • Automatic rhythm classification of the Apple watch software was poor
  • Apple watch iECG may be a good product for detection of normal sinus rhythm, but a formal interpretation by a pediatric cardiologist would be necessary
  • Adequacy for determination of an arrhythmia on the iECG is still unknown

Dr Jared Hershenson

Commentary from Dr. Jared Hershenson (Greater Washington DC), section editor of Pediatric Cardiology Journal Watch: Over the past 5 years, use of smartwatches such as the Apple watch have exploded in popularity, and there have been increasing technologic advances of the software to allow for determination of abnormal heart rhythms. Currently, the iECG function of the Apple watch is FDA approved for detection of arrhythmias in adults > 22 years of age. Its utility and accuracy in the pediatric age population though is not yet known. This study compared the iECG of the Apple watch with a lead I rhythm strip of a standard 12-lead ECG in pediatric patients with and without congenital heart disease (CHD).


Methods: Patients aged 0-16 years presenting to a single cardiac center were eligible. An iECG and lead I rhythm strip were obtained within 10 minutes apart, and the cardiac rhythm, amplitudes and timing intervals were measured blindly by 2 pediatric cardiologists. The Apple watch was placed on the left wrist and a finger from the right hand placed on the knob. Parents or a nurse held the finger of the younger aged patients on to the knob to allow for recording.


Results: 215 patients, with a mean age of 8 +/- 4.33 years were enrolled, and an iECG was able to be obtained in all of them. 134 patients did not have CHD; the others had various diagnoses including shunt lesions (VSD/ASD), TGA, TOF, or single ventricles. Heart rate, P wave duration, QRS duration, QT interval, P wave amplitude, QRS amplitude, and T wave amplitude were compared and plotted via Bland-Altman analysis, showing a highly significant correlation of all values except for P wave duration and P wave amplitude which showed a weaker, yet still significant correlation. See Table 2. Multivariate analysis suggested that weight and BMI did affect the accuracy of the iECG. The majority of the rhythms were classified as sinus rhythm by the pediatric cardiologists (90%) while the automated Apple watch algorithm classified only 55% as sinus rhythm. Additionally, ~38% of iECGs were labeled as inconclusive by the algorithm, but when reviewed by the pediatric cardiologist, only ~5% of the tracings could not be assessed for rhythm, with those usually due to motion artifact/low quality. See Figure 3. There was no significant influence of the presence of CHD on heart rhythm classification.


Conclusions: Overall, there was good compatibility and quality of the iECG, but poor automated classification of the heart rhythm by the Apple watch. However, since sinus rhythm was the most common rhythm, and arrhythmias were limited in this sample, it remains to be seen whether the iECG will be accurate as a mobile event monitor in that scenario. As many practicing pediatric cardiologists know, it can be difficult to document and diagnose arrhythmias in young patients when they occur intermittently or last short durations. However, when possible, a pediatric cardiologist review of a tracing that shows sinus rhythm can potentially alleviate parental anxiety and possibly decrease unnecessary urgent care and ER visits.