Increased periodic repolarization dynamics and supraventricular premature complexes frequency independently predicted atrial fibrillation detected after ischemic stroke and significantly improved risk prediction compared with the clinical CHASE-LESS score.
Key Findings
Results
AFDAS occurred in 9.5% of acute ischemic stroke patients during a median follow-up of 17 months.
25 of 263 patients developed AFDAS
Median follow-up was 17 months (interquartile range, 12–26 months)
All 263 patients were in sinus rhythm at baseline without prior or current atrial fibrillation
AFDAS was identified through continuous in-hospital monitoring, 6-month follow-ups, and electronic health record review
Results
Patients who developed AFDAS had significantly higher PRD values than those who did not.
PRD values: 5.80 deg² in AFDAS patients versus 2.93 deg² in non-AFDAS patients
P=0.005
PRD was measured via a standardized 30-minute high-resolution ECG performed within 5 days after acute ischemic stroke
PRD reflects sympathetic activity-associated oscillations in cardiac repolarization
Results
PRD independently predicted AFDAS in multivariable Cox regression analysis.
Hazard ratio for PRD: 1.27 (95% CI, 1.03–1.44); P=0.037
The analysis was adjusted for other covariates including supraventricular premature complexes frequency
PRD remained an independent predictor after multivariable adjustment
Results
Supraventricular premature complexes frequency independently predicted AFDAS in multivariable Cox regression analysis.
Hazard ratio for supraventricular premature complexes per hour: 1.02 (95% CI, 1.01–1.03); P<0.001
Supraventricular premature complexes per hour was assessed during the same standardized 30-minute high-resolution ECG as PRD
Both PRD and supraventricular premature complexes frequency were independently significant predictors in the same multivariable model
Results
A model combining PRD and supraventricular premature complexes per hour showed significantly better discrimination for AFDAS than the CHASE-LESS clinical score alone.
Area under the curve (AUC) for PRD + supraventricular premature complexes model: 0.82
AUC for CHASE-LESS score alone: 0.67
Difference in AUC was statistically significant: P=0.005
The CHASE-LESS score comprises coronary heart disease, heart failure, age, stroke severity, hyperlipidemia, diabetes, and prior stroke/transient ischemic attack
Methods
The study enrolled patients prospectively from the STROKE-CARD cohort between February 2021 and May 2023.
263 patients with acute ischemic stroke were enrolled
Enrollment period: February 22, 2021 to May 5, 2023
Heart rate variability was also assessed alongside PRD and supraventricular premature complexes during the standardized ECG
All patients were in sinus rhythm at baseline with no prior or current atrial fibrillation
What This Means
This research suggests that a novel heart rhythm measurement called Periodic Repolarization Dynamics (PRD) — which captures irregular fluctuations in the heart's electrical recovery phase linked to the nervous system — can help identify stroke patients who are likely to develop atrial fibrillation (an irregular heartbeat) after their stroke. The study followed 263 stroke patients for a median of about 17 months and found that roughly 1 in 10 developed this condition, known as atrial fibrillation detected after stroke (AFDAS). Patients who later developed AFDAS had nearly double the PRD values of those who did not, and PRD remained a significant predictor even after accounting for other risk factors.
The researchers also found that the frequency of premature heartbeats originating in the upper chambers of the heart (supraventricular premature complexes) was another independent predictor of AFDAS. When PRD and premature heartbeat frequency were combined into a single prediction model, it outperformed a standard clinical risk score (the CHASE-LESS score) by a substantial margin — achieving an accuracy measure (AUC) of 0.82 compared to 0.67 for the clinical score alone.
This research suggests that a simple, non-invasive 30-minute ECG recording performed within days of a stroke could meaningfully improve doctors' ability to identify which patients are most at risk of developing atrial fibrillation afterward. This matters because undetected atrial fibrillation after stroke is associated with a higher risk of recurrent stroke, and earlier detection could open the door to timely preventive treatments. Both PRD and premature heartbeat counts could potentially be added to existing clinical tools to improve patient monitoring and care decisions after stroke.