Background: Three randomised controlled trials have investigated prolonged monitoring methods for the detection of Atrial Fibrillation (AF) in stroke patients. All have shown that more patients with Atrial Fibrillation are detected by prolonged monitoring, but none has so far investigated if biomarkers identify patients who benefit from the intervention. BNP is of special interest, has retrospective analyses from the WARSS trial showed that BNP identifies patients who might benefit from oral anticoagulation in comparison to Aspirin (Longstreth, Stroke 2013).
Methods: We analysed brain natriuretic peptide (BNP) measured at randomisation in participants of the Find-AF randomised trial. In brief, Find-AF randomised randomised 402 stroke patients aged 60 years and above with a recent stroke (< 7 days to onset of symptoms) to enhanced and prolonged monitoring (3x10 day Holter ECG) or usual care. The primary endpoint was the detection of AF by 6 months.
Results: BNP could be measured in 373 of 402 participants (93 %). Patients with AF detected by 6 months had significantly higher BNP values (median 53; IQR 30;169) as compared to patients without AF (median 28; IQR 12;62) (p<0.001). BNP >= 100 pg/ml versus < 100 pg/ml showed a significant interaction with the intervention (p=0.01). The number needed to screen in the overall study was 11, while it was 3 for BNP>100pg/ml and 18 for BNP<100 pg/ml (see Figure).
Conclusion: BNP identifies a high-risk subgroup of stroke patients with undetected AF.