Neurol Sci. 2026 Jun 17;47(7):571. doi: 10.1007/s10072-026-09180-4.
ABSTRACT
PURPOSE: Atrial fibrillation (AF) drives cardioembolic stroke, yet paroxysmal AF (PAF) often escapes detection in embolic stroke of undetermined source (ESUS). We evaluated whether atrial natriuretic peptide (ANP) can identify latent PAF in ESUS, compared with brain natriuretic peptide (BNP).
METHODS: Prospective single-center cohort of ESUS patients admitted within 24 h without admission AF. Admission ANP/BNP were assayed; PAF was screened by 14-day continuous ECG. Patients were grouped as PAF-detected or ESUS. Logistic regression (LAD, ANP, BNP) and ROC assessed diagnostic performance.
RESULTS: Of 567 patients with cerebral infarction, 111 met the criteria for ESUS, and 64 were included after exclusions. PAF was detected in 30 patients, and 34 remained classified as ESUS. Baseline comparisons showed significant differences in age, ANP, BNP, and LAD. In multivariable logistic regression analysis, ANP was independently associated with PAF detection (p = 0.015), whereas BNP and LAD were not. ROC analysis showed an AUC of 0.901 (95% CI, 0.818-0.984) for ANP and 0.892 (95% CI, 0.811-0.973) for BNP, with no significant difference between them (p = 0.821). ANP demonstrated clinically useful performance with a cutoff value of 63.9 pg/mL (sensitivity 94.1%, specificity 83.3%).
CONCLUSIONS: ANP was identified as an independent predictor of PAF detection in patients with ESUS, with strong ROC performance. Given its atrial origin and very short half-life, ANP may serve as an atrium-specific biomarker reflecting acute atrial load, potentially enabling earlier identification of latent PAF in the acute phase.
PMID:42303893 | DOI:10.1007/s10072-026-09180-4