Novel insights into the genetics of intracerebral hemorrhage.

Isolated observations suggest a role for genetic risk factors in intracerebral hemorrhage (ICH). However, no systematic evaluation on the role of DNA variation in ICH has been attempted to date. We performed a large-scale genetic association study of 2189 ICH cases and 4041 controls. We included 1104 ICH cases attributable to cerebral amyloid angiopathy (CAA-ICH) and 1085 hypertensive ICH cases (H-ICH). Recently developed methods were used to determine heritability from unrelated ICH cases and controls. We also analyzed clinical ICH phenotypes: these included admission ICH volume (quantified on computed tomography scan), hematoma expansion (on follow-up computed tomography scan), and functional independence and mortality at 90 days post-ICH. We tested single variants across the entire genome (at genome-wide significance threshold of P<5.0×10). We also tested 2 genetic risk scores, summarizing the effect of known hypertension and Alzheimer disease risk loci. We estimated heritability of 45% for ICH, 70% for the CAA-ICH subset, and 35% for the H-ICH subset. Apolipoprotein E ε2/ε4 were associated with CAA-ICH at genome-wide significance levels (odds ratio [OR]=1.82; P=6.6×10 and OR=2.20; P=2.4×10, respectively). ε2 was also associated with larger admission CAA-ICH volume (hematoma size increase=5.3 cc per allele copy; P=0.004) and with CAA-ICH hematoma expansion (OR=2.72; P=0.009). We also found associations between apolipoprotein E ε2 and functional independence (OR=0.68; P=0.009) and mortality (OR=1.57; P=0.021) after CAA-ICH. A genetic risk score summarizing the effects of all known Alzheimer disease loci was associated with CAA-ICH (OR=1.22 for each of the 11 incorporated variants; P=2.2×10). This was largely mediated by one variant in the CR1 gene (OR=1.61; P=8.0×10), which was also associated with CAA-ICH recurrence (hazard ratio=1.35; P=0.024). A genetic risk score summarizing the effects of all known hypertension loci was associated with the risk of H-ICH (OR=1.18 of each of 40 incorporated variants; P=0.001) but not CAA-ICH (P=0.34). We provided the first quantitative assessment of ICH heritability. We also identified aolipoprotein E ε2/ε4 as risk factors for CAA-ICH at genome-wide level. ε2 is also associated with CAA-ICH volume and outcome. Another Alzheimer disease–related locus, CR1, also influences CAAICH incidence and recurrence. Conversely, hypertension loci play a role in H-ICH risk but not in CAA-ICH. Our results confirm the suspected clinical and etiopathogenetic heterogeneity of ICH: amyloid-related processes influence incidence and clinical course of CAA-ICH, whereas hypertension genetics play a major role in H-ICH.