Association of atrial fibrillation and chronic kidney disease with cerebral microbleeds
N Gisler, M Mutti, S Aeschbacher, F Mahfoud, L H Bonati, T Sinnecker, N Rodondi, T Reichlin, G Moschovitis, G Conte, A Mueller, J H Beer, D Conen, M Kuehne, P KrisaiAbstract
Background
Atrial fibrillation (AF) and chronic kidney disease (CKD) are associated with cerebral microbleeds (CMB). However, their combined impact on CMBs remains underinvestigated.
Purpose
To quantify the associations of CKD, AF and their combination with MRI-detected CMBs.
Methods
We analysed 2,712 patients from the prospective, multicenter Swiss-AF cohort with standardized cranial MRI. The primary outcome was the presence of any CMB. Secondary outcomes were CMB localization (non-lobar vs. lobar) and CMB burden (≥3 CMBs). Patients were stratified into four groups based on their medical history: AF-/CKD-, AF-/CKD+, AF+/CKD-, and AF+/CKD+. We used multivariable logistic regression models for the associations of AF and CKD with CMBs, adjusted for age, sex, hypertension, diabetes, history of stroke/TIA, and anticoagulation.
Results
Among 2,712 patients (mean age 73.1 years, 30.8% female), the overall prevalence of cerebral microbleeds was 24.5%. Stratified by group, the prevalence was 26.2% in AF-/CKD-, 33.6% in AF-/CKD+, 21.2% in AF+/CKD- and 29.6% in AF+/CKD+ (Fig. Panel A). For anatomical distribution the highest prevalence for non-lobar (22.4%) and lobar (23.8%) CMBs were observed in AF-/CKD+. While single microbleeds (1 CMB) constituted most findings across all cohorts, the proportion of patients with a severe burden (≥3 CMBs) also peaked in AF-/CKD+ with 9.1%.
In adjusted regression models for any CMB with AF-/CKD- as the reference group, the OR (95% CI) was 0.68 (0.48-0.95; p=0.025) for AF+/CKD-, 1.14 (0.77-1.69; p=0.503) for AF-/CKD+ and 0.87 (0.58-1.32; p=0.518) for AF+/CKD+ (Fig. Panel B). Results were consistent in analyses stratified for CMB location. In separate analyses for any CMB, the OR (95% CI) was 0.69 (0.50-0.96; p=0.028) for AF and 1.23 (0.96-1.58; p=0.100) for CKD with no interaction between the two diseases (p for interaction=0.729). In adjusted models for CMB burden (≥3), the OR (95% CI) was 0.81 (0.49-1.35; p=0.428) for AF+/CKD-, 1.44 (0.71-2.75; p=0.298) for AF-/CKD+ and 0.94 (0.49-1.74; p=0.852) for AF+/CKD+.
Conclusion
Patients with CKD had a higher prevalence of CMBs compared to patients with AF, with no additive effect of the two diseases. The distribution and burden of CMBs were similar between AF and CKD patients. After adjustment for vascular risk factors and anticoagulation, only AF but not CKD was associated with CMBs.Visual Abstract Association AF x CKD with CMB