Abstract Body: Objectives
Spontaneous intracerebral hemorrhage (ICH) is caused by the rupture of small arterioles due to cerebral small vessel disease (SVD), commonly with hypertensive arteriolosclerosis (HTA) or cerebral amyloid angiopathy (CAA). Several genes associated with SVD or ICH have been reported in studies of European populations, but it is unclear whether these genes are associated with ICH in the Japanese cohort. The aim of this study was to assess genetic variants associated with SVD in patients with ICH.

Methods
We analyzed 920 patients with primary ICH and genetic data available from an ongoing, prospective, cohort study. ICH was diagnosed according to CLASsification system for ICH (CLAS-ICH) by stroke neurologists, and mixed-SVD was classified as HTA. We focused on the 27 genes previously identified in SVD (APOE, CETP, PMF1, SLC25A44, TRHDE, APP, NOTCH3, COL4A1, COL4A2, PARVB, FOXC1, FOXF2, TREX1, ZCCHC14, WDR12, ICA1L, GCH1, SH2B3, ACOX1, UNC13D, TRIM65, TRIM47, WBP2, PLEKHG1, EFEMP1, VCAN, HTRA1) to investigate their association with the ICH phenotypes (ICH subtypes [HTA-ICH vs. CAA-ICH], hematoma location [non-lobar vs. lobar], modified Rankin Scale (mRS) at 90d [0-4 vs. 5-6]). Variants with minor allele frequency ≥0.05 were analyzed. Trend test was used to assess the association, and the significance level was defined as false discovery rate (FDR) <0.1. The reference allele was defined as a base allele of odds ratio (OR).


Results
Characteristics of 920 patients with ICH are mean age 71 ± 13 years, 59% men, HTA-ICH vs. CAA-ICH:88% vs. 10%, non-lobar vs. lobar: 74% vs. 26%, and mRS 0-4 vs. 5-6: 78% vs. 22%. We assessed the 667 variants in 27 genes. In CAA-ICH vs HTA-ICH, two variants on APP (amyloid-β protein precursor) gene showed significance. One was rs2830008 (NM_000484:c.865+2696A>G, OR=2.67, p=5.9×10^-6, FDR=0.039) and the other was rs73163761 (NM_000484:c.866-7085C>T, OR=2.59, p=1.9×10^-4, FDR=0.064). These 2 variants were in linkage equilibrium (r²=0.93). The alternative allele frequencies of rs2830008: T>C and rs73163761: G>A were 0.182 vs. 0.077 and 0.166 vs. 0.071 in CAA-ICH vs. HTA-ICH, respectively, with the HTA-ICH frequencies aligning with those of the general population (0.081, 0.072, respectively). No significant association was observed for other phenotypes.

Conclusion
We found that 2 variants on APP genes were associated with CAA-ICH in the Japanese ICH cohort, confirming the link to amyloid β pathology.