OBJECTIVE: To investigate the direct and indirect relationships between statin use, low-density lipoprotein cholesterol (LDL-C) levels, and intracerebral hemorrhage (ICH), providing new insights into this complex scientific question. METHODS: In this cohort study, UK Biobank data from 2006 to 2010 were used to construct Structural Equation Models of statin use, LDL-C, and ICH, including 414,253 participants with LDL-C data. Published Genome-Wide Association Studies data were used for drug-target Mendelian Randomization analysis. RESULTS: The study included 414,253 participants, comprising 225,454 women (54.4%) with a mean age of 56.07 (8.11) years. During a median follow-up of 14.01 years, 2973 patients experienced ICH. Structural Equation Modelling showed the indirect effect (path a∗b) of statin on ICH was 0.003 (P <
0.001), the direct effect (path c') was -0.001 (P = 0.568), the total effect (path c) was 0.002 (P = 0.391), and the mediation proportion of LDL-C (a∗b/c) was 150.0%. Mendelian Randomization showed a negative association between LDL-C levels and ICH (β: -0.663, SE: 0.229, P = 0.004), with no causal relationship between statin use and ICH (β: -1.454, SE: 3.133, P = 0.643). Drug-targeted Mendelian Randomization revealed LDL-C levels, predicted by variants in or near HMGCR, PCSK9, CETP, ABCG8/5, and LAP, were negatively associated with ICH risk. CONCLUSIONS: This study confirmed that statins increase the risk of ICH primarily through their LDL-C-lowering effects, rather than the direct effects of the statins themselves. LDL-C is negatively associated with ICH, an association not confined to the effects of the HMGCR loci. This advance provides evidence for the controversy between statin use, LDL-C levels, and ICH risk.