Skip to main content
  • Home
  •  Synopsis Article Listing
  •  Predictors of atrial fibrillation development in patients with embolic stroke of undetermined source: An analysis of the RE-SPECT ESUS trial

Predictors of atrial fibrillation development in patients with embolic stroke of undetermined source: An analysis of the RE-SPECT ESUS trial

What was known?

Embolic stroke of undetermined source (ESUS) is associated with a stroke recurrence rate of 3% to 6% per year. Patients with ESUS have also been found to have a high incidence of atrial fibrillation (AF), suggesting that AF may be an underlying cause of stroke recurrence. Known predictors of AF after stroke include clinical characteristics, findings on Holter monitoring and echocardiographic measurements. Detection of AF in these patients is important because it affects the management strategy.


The aim of this study was to assess clinical factors predicting the development of AF and associated recurrent stroke in patients with ESUS.

What this study adds:

This study has identified a patient subgroup that can be identified through simple clinical variables and may benefit from prolonged cardiac monitoring for detection of AF. 

Clinical implications:

Among patients with ESUS, around 7.5% will develop AF over the next 19 months. Risk factors for development of AF in ESUS patients include older age, higher body mass index (BMI), hypertension, absence of diabetes, and N-terminal prohormone of brain natriuretic peptide (NT-proBNP) levels. For patients with these risk factors for developing AF after ESUS, long-term cardiac monitoring may be beneficial.

Study design:

This is a secondary analysis of the RE-SPECT ESUS study (NCT02239120), which was an international, double-blind Phase III trial in which patients (n=5390) with recent ESUS were randomly assigned to receive dabigatran and aspirin placebo or aspirin and dabigatran placebo. AF was identified through investigators’ standard of care and by 1-year electrocardiography (ECG). NT-proBNP level was evaluated as an AF predictor for a subset of patients (n = 1,117) who had a baseline level available.

Primary endpoints:

  • Recurrent stroke of ischaemic, haemorrhagic, or unspecified type, assessed in a time-to-event analysis.
  • Major bleeding according to International Society on Thrombosis and Haemostasis  criteria, assessed in a time-to-event analysis

Secondary endpoints:

  • Adjudicated ischaemic stroke

Inclusion criteria:

  • Ischaemic stroke with a brain lesion visualized by neuroimaging (either brain Computed Tomography [CT] or Magnetic Resonance Image [MRI])
  • The index stroke must have occurred either up to 3 months before randomisation (Modified Rankin Scale[mRS] ≤3 at randomisation) or up to 6 months before randomisation (mRS ≤3 at randomisation) in selected patients aged ≥60 years, plus at least one additional risk factor for recurrent stroke.

Notable exclusion criteria:

  • Modified Rankin Scale of ≥4 at randomisation or inability to swallow medications.
  • Major risk cardioembolic source of embolism including: intracardiac thrombus diagnosed by transthoracic or transesophageal echocardiography; paroxysmal, persistent or permanent AF;  atrial flutter; prosthetic cardiac valve (mitral or aortic, bioprosthetic or mechanical); atrial myxoma: other cardiac tumours; moderate or severe mitral stenosis; recent (< 4weeks) myocardial infarction; valvular vegetations, or infective endocarditis


  • Of the 5390 participants, 403 (7.5%) developed AF during follow-up
  • In the multivariate model, independent predictors of developing AF after ESUS during the study were: older age (odds ratio [OR] for 10-year increase, 1.99; 95% confidence interval [CI], 1.78–2.23]; p<0.001), hypertension (OR, 1.29; 95% CI, 1.16–1.44; p=0.0304), diabetes (OR, 0.74; 95% CI, 0.56–0.96; p=0.0226), and body mass index (OR  for 5-U increase, 1.29; 95% CI, 1.16–1.43; p<0.001)
  • In a sensitivity analysis of 1117 patients with baseline NT-proBNP measurements, only older age and higher NT-proBNP were significant independent predictors of AF
  • CHA2DS2-VASc scores were higher in patients with AF compared with those without AF.


In this study, older age, higher BMI, hypertension, and absence of diabetes were associated with a higher risk of developing AF. When baseline NT-proBNP was available, only older age and elevation of this biomarker were significant predictors of AF. The rate of recurrent stroke in patients who developed AF was higher than in the entire group. This emphasises the importance of detection and treatment of AF in patients who have had ESUS. Simple clinical variables and, when available, levels of NT-proBNP could help identify a patient population who may benefit from prolonged cardiac monitoring. However, these data should be considered hypothesis-generating and not conclusive, as they are vulnerable to detection bias. Further study is warranted to assess the clinical utility of NT-proBNP as a biomarker of risk in this setting.


  • This is a secondary analysis that is hypothesis-generating only
  • Detection of AF during follow-up was not done systematically but according to standard of care, apart from the systematically collected, yearly 12-lead ECGs
  • Biomarkers were only available in 20% of the patient population, limiting assessments of the association of NT-proBNP with the risk of stroke