Association between lipoprotein(a) and ischemic brain lesions in patients with atrial fibrillation: insight from the Swiss-AF cohort

Introduction Lipoprotein(a) (Lp(a)) exhibits both proatherosclerotic and prothrombotic properties and could be a prognostic biomarker in patients with atrial fibrillation (AF). However, data is scarce on the relation between Lp(a) and stroke in AF patients. Purpose We assessed the association between Lp(a) serum concentration and the presence of ischemic brain lesions in patients with AF. Methods Data were derived from the Swiss-AF cohort (N=2,415), a multicenter prospective observational cohort of patients aged ≥65 years with established AF enrolled in 14 centers in Switzerland. We included participants with Lp(a) serum concentration and brain magnetic resonance imaging (MRI) at baseline. Multivariable logistic regression was used, adjusting for a pre-specified set of confounders. Baseline Lp(a) was treated as a continuous variable and categorized in quartiles. MRI-detected brain infarcts at baseline (with or without prior clinical stroke) were classified as large non-cortical or cortical infarcts (likely embolic) or small non-cortical infarcts (lacunar infarcts). Prespecified subgroup analysis included notably age category (≥75 years), sex, and AF type. Results Among 1’649 patients with baseline Lp(a) and MRI (72.6±8.4 years, 72.6% male, 54.7% non-paroxysmal AF), 608 (37.1%) exhibited ischemic brain lesions at baseline MRI. Median Lp(a) serum concentration was 16.5 nmol/L (IQR 6.7-47.1). Higher Lp(a) was associated with prevalent ischemic brain lesions, with a dose-response pattern across quartiles (p for trend=0.001), driven by upper quartiles: adjusted odds ratio (OR) for highest versus lowest quartile, 1.63; 95% confidence interval (CI), 1.19-2.24 (Figure, Panel A). The association appeared driven by likely embolic infarcts (OR per 100 nmol/L, 1.22; 95% CI, 1.00-1.48) rather than lacunar infarcts (OR per 100 nmol/L, 1.00; 95% CI, 0.82-1.22). There was significant effect modification by AF type (p for interaction=0.002): per 100 nmol/L increase, the adjusted OR was 1.61 (95% CI, 1.25-2.08) in non-paroxysmal AF versus 0.83 (95% CI, 0.62-1.13) in paroxysmal AF (Figure, Panel B). Conclusion In patients with AF, higher Lp(a) was associated with increased prevalence of ischemic brain lesions on MRI, especially in those with non-paroxysmal AF. Lp(a) might be a therapeutic target in this population.