Validity of heart rate measurements by the Garmin Forerunner 225 at different walking intensities

The accuracy of wrist worn heart rate monitors based on photoplethysmography (PPG) is not fully clinically accepted. Therefore, we aimed to validate heart rate measurements of a commercially available PPG heart rate monitor, i.e. the Garmin Forerunner^® 225. Twelve healthy volunteers (six women; mean age: 28 years) performed a treadmill protocol consisting of: five minutes sitting, five minutes standing, 10 minutes walking at 4 km/h, 10 minutes walking at a gradient of 5% and intensity of 4-6 metabolic equivalents (METs), 10 minutes walking at a gradient of 8% and intensity of seven METs or more. Walking speeds were individually determined. Walking bouts were separated by a standardised five minute rest period. Heart rate was measured as the average of the last three minutes standing and of each walking bout. A three lead patch-based electrocardiogram (ECG; Zensor^®) was used as criterion method. Statistical analyses included Pearson's correlation (r), paired t-tests, root mean squared error (RMSE) and Bland?Altman plots. The mean values per three minutes of every condition did not differ significantly between the Garmin Forerunner^® 225 and the Zensor^®. RMSE was 3.01 beats per minute (bpm) or 2.89%. The Bland-Altman bias was 1.57 bpm. Limits of agreement (LoA) were wide, ranging from 32.53 to 29.40 bpm. However, Pearson's r ranged from 0.650 to 0.868 suggesting moderate to strong validity. Generally, mean heart rates, r values, RMSE and the Bland-Altman bias indicated good overall agreement in this sample of healthy adults, but wide LoA are making it difficult to trust individual measurements.