Cryptogenic stroke patients have received a lot of research interest in the last couple of years. CRYSTAL AF and EMBRACE trials were published in June last year and made a strong case for longer duration of cardiac telemetry, especially in patients above age 55. Since these trials came out, several studies have found associations between EKG abnormalities such as QTc, as well as cardiac echo abnormalities such as left atrial enlargement with future detection of paroxysmal atrial fibrillation (PAF). For stroke providers, tools that help stratify a patient’s pretest probability assume special relevance as they may help tailor how long should a patient be monitored for PAF.



In the current edition of Stroke, Gladstone et al. used data from the EMBRACE trial to investigate the association between Holter-detected atrial premature beats (APB), in addition to age and left atrial enlargement, on the subsequent detection of AF by 30-day ECG monitoring, and during clinical follow-up over 90 days and 2 years. This was a pre-planned subanalysis of patients in the intervention arm of the trial, who received the AF auto-detect ECG loop recorder for up to 30 days. In 237 participants, they found that those who subsequently had AF detected had 629 APBs/24h vs. 45 in those without AF. APB count was a significant predictor of AF detection in a multivariable logistic regression analysis at all the time-points chosen (30 days, 90 days , and 2 years). The authors were also able to demonstrate a dose response relationship between APBs detected and AF detection rate.

The authors thus establish APBs as another biomarker that may help stratify a cryptogenic stroke patient’s risk of undetected PAF. The authors propose that those with with >1000APBs /24 hours may warrant longer than 30 day monitoring, and those with <500APBs/24 hours may need just 2 weeks. We are heading towards an era of individualized medicine, and novel biomarkers like this one are leading the way!