Background and purpose: Cerebral microbleeds (CMB) attributable to cerebral amyloid angiopathy generally occur in lobar regions, whereas those attributable to hypertensive vasculopathy are deep. Inflammation may be an underlying mechanism for CMB, with varying associations according to CMB location. Lipoprotein phospholipase-A2 (Lp-PLA2) is a circulating enzyme marker of vascular inflammation associated with risk of ischemic stroke and dementia. We hypothesized that higher Lp-PLA2 levels would be related to higher prevalence of CMB, with possible regional specificity.
Methods: Framingham Offspring participants aged 65 years or older with available Lp-PLA2 measures and brain magnetic resonance imaging were included. Logistic regression models were used to relate Lp-PLA2 activity and mass to presence of CMB, adjusted for age, sex, medication use (aspirin, anticoagulants, and statins), systolic blood pressure, APOE, current smoking, and diabetes.
Results: Eight-hundred nineteen participants (mean age, 73 years; 53% women) were included; 106 (13%) had CMB, 82 (10%) were lobar, and 27 (3%) were deep. We did not observe significant associations of CMB and LpPLA2 measures in multivariable adjusted analyses. However, there was a significant interaction between APOE genotype and Lp-PLA2 activity in their relation to presence of deep CMB (P interaction=0.01). Among persons with APOE ε3/ε3, the odds ratio for deep CMB was 0.95 (confidence interval, 0.59-1.53; P=0.83), whereas among those with at least 1 ε2 or ε4 allele, odds ratio was 3.46 (confidence interval, 1.43-8.36; P=0.006).
Conclusions: In our community-based sample of older adults, there was no significant association of Lp-PLA2 with total or lobar CMB. The association of higher levels of Lp-PLA2 activity with deep CMB among those with at least 1 APOE ε2 or ε4 allele merits replication.