Determine atrial fibrillation burden with a photoplethysmographic mobile sensor: the atrial fibrillation burden trial: detection and quantification of episodes of atrial fibrillation using a cloud analytics service connected to a wearable with photoplethysmographic sensor

Figures

• 

  Figure 1. Conceptual illustration of atrial fibrillation burden classification per patient. First, the atrial fibrillation burden is assessed independently using photoplethysmography and electrocardiogram signals (upper and middle row). Using the electrocardiogram-based assessment as gold standard, the atrial fibrillation burden assessment is then classified into true/false positive/negative (bottom row). The time unit of this classification is 1 s.

• 

  Figure 2. Bland–Altman analysis shows the agreement of atrial fibrillation burden in minutes per patient measured by photoplethysmography and electrocardiogram. Each dot represents the atrial fibrillation burden for each patient in the atrial fibrillation group. On the x-axis of the plot, the mean atrial fibrillation burden in minutes of the two measurement methods is displayed. On the y-axis, the difference between the atrial fibrillation burden in minutes measured by electrocardiogram and photoplethysmography is displayed. The true positive atrial fibrillation burden measured by the photoplethysmography was subtracted from the atrial fibrillation burden measured by the electrocardiogram. A positive difference between the two measurements is therefore resulting from a higher atrial fibrillation burden measured by the electrocardiogram than measured by the photoplethysmography. The black line represents the average value of difference between the atrial fibrillation burden in minutes measured by the electrocardiogram and the PPG. The red dashed lines represent the upper and lower 95% confidence interval of the difference between the measurement methods.

• 

  Figure 3. Bland–Altman analysis shows the agreement of non-atrial fibrillation time in minutes per patient measured by photoplethysmography and electrocardiogram. Each dot represents the non-atrial fibrillation time for each patient with non-atrial fibrillation time. On the x-axis of the plot, the mean non-atrial fibrillation time in minutes of the two measurement methods is displayed. On the y-axis, the difference between the non-atrial fibrillation time in minutes measured by electrocardiogram and photoplethysmography is displayed. The true negative non-atrial fibrillation time measured by the photoplethysmography was subtracted from the non-AF time measured by the electrocardiogram. A positive difference between the two measurements is therefore resulting from a higher non-atrial fibrillation time measured by the electrocardiogram than measured by the photoplethysmography. The black continuous line represents the average value of difference between the non-atrial fibrillation time in minutes measured by the electrocardiogram and the photoplethysmography. The red dashed lines represent the upper and lower 95% confidence interval of the difference between the measurement methods.

• 

  Figure 4. Bland–Altman analysis shows the agreement of percentage of time in atrial fibrillation per patient measured by photoplethysmography and electrocardiogram. Each dot represents the percentage of time in atrial fibrillation for each patient in the atrial fibrillation group. The percentage of time in atrial fibrillation per patient was calculated by dividing the individual amount of detected atrial fibrillation burden in minutes in the electrocardiogram for each patient and accordingly correctly identified atrial fibrillation burden in the PPG by the evaluable monitoring time for the respective patient. On the x-axis of the plot, the mean percentage of time in atrial fibrillation in minutes of the two measurement methods is displayed. On the y-axis, the difference between the percentage of time in atrial fibrillation measured by electrocardiogram and photoplethysmography is displayed. The black line represents the average value of difference between the percentage of time in atrial fibrillation measured by the electrocardiogram and the photoplethysmography. The red dashed lines represent the upper and lower 95% confidence interval of the difference between the measurement methods.