Relationship Between Pulse Transit Time, PPG Features, and Blood Pressure in Atrial Fibrillation

Atrial fibrillation (AF) is a common arrhythmia associated with increased cardiovascular risk. Continuous blood pressure (BP) monitoring is critical in AF management; however, most cuffless BP estimation methods are designed for sinus rhythm (SR), limiting their effectiveness in AF populations. This study explores the relationship between pulse transit time (PTT), photoplethysmogram (PPG) morphological features, and BP in AF patients. Synchronized electrocardiogram (ECG), PPG, and invasive BP signals were collected from 43 patients undergoing ablation procedures, resulting in 43,219 AF beats and 10,995 SR beats. Beat-to-beat variations in PTT, PPG features, and BP were analyzed. Feature-BP correlations were quantified, and individualized ridge regression models were developed for BP estimation. Results indicate that the AF group exhibited significantly higher beat-to-beat variability in BP, PTT, and PPG features than the SR group. In AF, descending time and descending area showed stable and strong negative correlations with diastolic BP (-0.75 ± 0.09 and -0.69 ± 0.12), while PTT and pulse intensity ratio correlated well with systolic BP (-0.40 ± 0.20 and -0.35 ± 0.18). These features were identified as effective markers of BP fluctuations in AF. Although estimation errors were higher in AF, the ridge regression model more accurately tracked dynamic BP trends (r = 0.748 for SBP, 0.734 for DBP), indicating enhanced sensitivity to BP variability. In contrast, lower BP variability in SR enabled accurate predictions but with limited responsiveness to subtle BP fluctuations. These findings provide critical insights into developing cuffless BP measurement models tailored to AF patients.