Skip to main content
NCBI home page
Medicine logoLink to Medicine
. 2025 Aug 8;104(32):e43838. doi: 10.1097/MD.0000000000043838

Statins and hemorrhagic stroke: New insight from a Mendelian randomization study

Xiaoxia Gu a, Xinchao Du a, Zhiwei Yao b, Shengyuan Gu a,*
PMCID: PMC12338289  PMID: 40797393

Abstract

The relationship between statin usage and hemorrhagic stroke (HS) has been reported in clinical studies, but the link was contradictory and causality from statin usage to HS remained unclarified. We aimed to investigate statin’s causal effects of blood and brain on HS risk with the Summary-data-based Mendelian randomization (SMR) method. We used 3-hydroxy-3-methylglutaryl-coenzyme A reductase (HMGCR) gene expression-associated single-nucleotide polymorphisms (SNPs) to proxy the pleiotropic effects of statin. Comparison analysis was performed to explore the difference between blood and the brain in the pattern of HMGCR affecting HS and HMGCR gene expression. SMR analysis found that a higher blood HMGCR expression decreased the risk of subarachnoid hemorrhage (SAH) (OR = 0.64, 95% CI = 0.43–0.95, SMR-P = .027), while brain HMGCR gene expression increased the risk of SAH at a boundary significance (OR = 1.11, 95% CI = 1.00–1.23, SMR-P = .049). Comparison analysis showed a negative relationship between blood and brain in the pattern of HMGCR affecting HS and HMGCR gene expression. This MR study suggested that HMGCR expression in the brain increased HS risk. The controversial results from the blood data were due to the difference in HMGCR gene expression between the brain and blood. Lipophilic statins might benefit patients more from the lower HS risk than the hydrophilic ones.

Keywords: causality, FinnGen, hemorrhagic stroke, Mendelian randomization, statins

1. Introduction

Statins, 3-hydroxy-3-methylglutaryl-coenzyme A (HMG-CoA) reductase inhibitors, a “cornerstone” drug for lowering low-density lipoprotein cholesterol (LDL-C) – a well-known risk factor with a positive “dose-response” effect on atherosclerotic cardiovascular disease (ASCVD),[1,2] was recommended as a prescription by the US Preventive Services Task Force (USPSTF) for the primary prevention of cardiovascular disease (CVD).[3] Statins are used by more than 200 million people worldwide[4] and by 1 in 4 Americans > 40 years of age in the US.[5]

However, regarding the side effects of statins, concerns were raised about the potentially increasing risk of myopathy, hepatotoxicity, diabetes, cataracts, dementia, cancer, etc, which would probably limit reaching a lower LDL-C target by statin therapy.[6] In terms of this topic, the American Heart Association (AHA) published a scientific statement in 2019 that provided a thorough review of statin side effects and safety, concluding that there was a positive relationship between statin and rhabdomyolysis, serious hepatotoxicity, newly diagnosed diabetes mellitus but with a low incidence (<0.1%, 0.001% and 0.2%, respectively). Of note, the writing group found that the incidence of hemorrhagic stroke (HS) was possibly increased by statin use in patients with cerebrovascular disease but with unstable evidence.[7] Nevertheless, the AHA published another scientific statement in 2023 especially to evaluate whether the LDL-C lowering drugs exert a harmful effect on the brain, including cognitive impairment, dementia and HS. Based on the contemporary evidence, the AHA writing group concluded that lipid-lowering treatment had no relationship with cognitive impairment and dementia, and statin therapy to patients without a history of cerebrovascular disease did not increase the risk of HS statistically. Yet, contemporary studies remain controversial on the results of whether statins increase the risk of HS among patients with a history of cerebrovascular disease, which calls for additional focused studies.[8]

Mendelian randomization (MR) study, an effective method to detect a causal relation between exposure and outcome by using genetic variants as the instrument, can conquer the potential biases that often inevitably exist in observational studies, such as confounders or reverse causation.[9,10] Genome-wide association studies with large sample sizes for various traits make the MR analysis more and more popular in assessing the causal relation between different phenotypes.[11] For example, blood LDL-C levels have been proven to decrease the risk of HS with the MR analysis,[12] which follows the epidemiological studies.[13–16]

The publicly available databases of genome-wide association studies make it possible for us to utilize MR to determine the relationship between statin and HS. As usual, we can choose the single-nucleotide polymorphisms (SNPs) of blood LDL-C levels around the 3-hydroxy-3-methylglutaryl-coenzyme A reductase (HMGCR) gene region to represent the statin drug to detect a causal relationship between statin and HS. But, from our perspective, it is not the perfect study design, because of 2 reasons below: (i) Gene expression-associated SNPs of HMGCR are more powerful, reliable and explicable in representing statins’ pleiotropic effects than LDL-C-associated SNPs of HMGCR, because the total pleiotropic effects of statin, including reducing LDL-C levels, alleviating endothelial dysfunction, preventing the formation of reactive oxygen species, etc are all from HMGCR expression suppression (Fig. S1, Supplemental Digital Content, https://links.lww.com/MD/P690).[17] Thus, gene expression-associated SNPs of HMGCR are just the ideal proxy for statin pleiotropic effects when studying the side effects of statin with the MR method. (ii) Would the pleiotropic effects of HMGCR on HS differ across the blood-brain barrier (BBB)? If so, what should we do to “capture” it? Thus, based on the knowledge above, we aim to apply Summary-data-based Mendelian randomization (SMR) method, which is established specially to test for the causal association between the expression level of a gene and a complex trait from expression quantitative trait loci (eQTL) and GWAS studies,[18] to detect the causal relation between statin and HS and to assess the difference in the HMGCR-pleiotropic effects on HS between in blood and brain, using HMGCR eQTLs in blood and brain to index statin drugs in blood and brain respectively.

2. Methods

Following the re-analysis of publicly available data, no additional ethics approval is required.

2.1. Study design

The study design is shown in Figure 1. First, we implemented SMR analysis to detect the association of HMGCR gene expression with HS in blood and brain respectively, using blood HMGCR cis-eQTLs from eQTLGen Consortium,[19] brain HMGCR cis-eQTLs from GTEx Consortium V7,[20] and HS GWAS data from FinnGen study.[21] Second, sensitivity analyses, including F-statistic assessment, positive control analysis (HMGCR as exposure and blood LDL-C as outcome), reverse causality test, linkage disequilibrium test, and multi-SNP-based SMR (SMR-multi) test, made the primary findings more robust. Third, we made a replication analysis replacing FinnGen GWAS data with UK Biobank GWAS data (http://www.nealelab.is/uk-biobank/). Finally, we conducted a comparison analysis to explore the correlation of the HMGCR expression effects on HS in blood with that in brain, and the relationship between blood HMGCR expression level and brain HMGCR expression level.

Figure 1.

Figure 1.

Open in a new tab

Study design for estimating the effect of blood and brain HMGCR gene expression on HS. HMGCR = 3-hydroxy-3-methylglutaryl-coenzyme A reductase, HS = hemorrhagic stroke, ICH = intracerebral hemorrhage, NICH = nontraumatic intracranial hemorrhage, SAH = subarachnoid hemorrhage.

2.2. Blood and brain HMGCR gene eQTL

We used eQTLs of the HMGCR gene – the statin-inhibited target gene, as the proxy of the total pleiotropic effects of statin. Blood HMGCR cis-eQTLs were retrieved from the eQTLGen Consortium,[19] while brain HMGCR cis-eQTLs were obtained from GTEx Consortium V7.[20] The source details are shown in Table S1, Supplemental Digital Content, https://links.lww.com/MD/P691. The inclusion criteria for potential eligible eQTL SNPs included: (i) minor allele frequency of SNPs > 1%; (ii) at a genome-wide significance level (P < 5 × 10−8) – if there were no SNPs reaching the threshold, the P-value was relaxed to < 1 × 10−5. Finally, 921 cis-eQTL SNPs in blood and 1 cis-eQTL SNP in the Brain–Spinal cord (cervical c-1) met the criteria (SNP details are presented in Table S2, Supplemental Digital Content, https://links.lww.com/MD/P691).

2.3. GWAS for HS

The GWAS summary data for HS, including intracerebral hemorrhage (ICH), nontraumatic intracranial hemorrhage (NICH), and subarachnoid hemorrhage (SAH), were retrieved from the FinnGen study[21] and the UK Biobank GWAS database (http://www.nealelab.is/uk-biobank/), as the discovery and replication cohort, respectively. Details of the GWAS data are summarized in Table S1 (Supplemental Digital Content, https://links.lww.com/MD/P691).

2.4. Statistical analyses

2.4.1. Primary SMR analysis

In this study, we performed analyses with SMR software (version 1.3.1), which is an effective tool to identify whether a certain gene is associated with a complex disease through changes in its expression levels, to detect the SMR effect size between HMGCR and ICH, NICH and SAH, in blood and brain respectively. The SMR software is publicly available online (https://yanglab.westlake.edu.cn/software/smr/#Overview).

2.4.2. Sensitivity analysis

2.4.2.1. Instrument assessment

F-statistic of ≤ 10 was used to hint at potential weak instrument bias in this MR analysis.[22] We performed a positive control MR analysis (TwoSampleMR package of version 0.5.7 in R), using blood and brain HMGCR gene expression respectively as the exposure and blood LDL-C level as the outcome (GWAS data from Global Lipids Genetics Consortium[23]), to validate the included eQTL SNPs’ appropriateness for representing the HMGCR expression levels.

2.4.2.2. Reverse causality and linkage disequilibrium detection

As SMR that is based on the analysis of a single SNP cannot identify causality from pleiotropy,[18] the MR Steiger method using TwoSampleMR package of version 0.5.7 in R was used to orient the causal relationship between HMGCR expression and HS.[24] HEIDI test provided by SMR tools (version 1.3.1) can distinguish causality from linkage disequilibrium when the primary SMR passes the significant P-value.[18]

2.4.2.3. SMR-multi test and replication analysis

The SMR-multi analysis established by Wu et al,[25] and replication analysis using UK Biobank GWAS data of HS, could enhance confidence in the primary SMR significant results.

2.4.3. Comparison analysis for the different scenarios – in blood and in brain

We hypothesized that there would be a different pattern of HMGCR affecting HS and a different model of HMGCR expression across the BBB – in blood and in brain. Therefore, the linear regression method (stats package of version 4.3.1 in R) was applied to detect the relationship between the HMGCR expression effects on HS of blood and that of brain – using the effect estimate from the primary SMR analysis, including 6 sample pairs (3 pairs from FinnGen outcomes and 3 pairs from UK Biobank outcomes), and the Spearman correlation analysis was performed to assess the correlation between the blood HMGCR gene expression and the brain HMGCR gene expression – using the original eQTL study effect estimate of the 922 SNPs meeting the above inclusion criteria, as the proxy. For instance, rs6453133, an SNP with a significant P-value of blood HMGCR gene expression – but nonsignificant of the brain, was retrieved for its effect estimate (β value) in blood and brain respectively and was taken as 1 sample pair after being harmonized to the same effect allele and other allele. Finally, 882 sample pairs were successfully retrieved from the 922 SNPs – 922 SNPs were all available in blood eQTL data, but 40 SNPs of the 922 SNPs were not found in brain eQTL data (Table S3, Supplemental Digital Content, https://links.lww.com/MD/P691).

2.4.4. Correction for multiple analyses

To account for multiple testing, a Bonferroni-corrected p threshold of <.0166 (α = 0.05/3 outcomes) was regarded as a piece of strong evidence for results, and .0166 ≤ P < .05 was considered to be a piece of suggestive evidence. A P-value in HEIDI test < .05 was taken as the evidence for linkage disequilibrium, which was more conservative than the Bonferroni-corrected HEIDI P threshold of <.0166.[18]

3. Results

3.1. Appraising the effect of HMGCR gene expression on HS

Suggestive evidence was obtained from the discovery cohort (FinnGen study) that blood HMGCR gene expression decreased the risk of SAH (OR = 0.64, 95% CI = 0.43–0.95, SMR-P = .027, SMR-multi-P = .038). Surprisingly, brain HMGCR gene expression increased the risk of SAH at a boundary significance (OR = 1.11, 95% CI = 1.00–1.23, SMR-P = .049) (Fig. 2 and Table S4, Supplemental Digital Content, https://links.lww.com/MD/P691). In the replication cohort (UK Biobank study), blood HMGCR gene expression decreased the risk of NICH (OR = 0.999, 95% CI = 0.998–0.999, SMR-P = .035, SMR-multi-P = .005), while no evidence was found to support brain HMGCR having a causal effect on HS (Fig. 3 and Table S4, Supplemental Digital Content, https://links.lww.com/MD/P691).

Figure 2.

Figure 2.

Open in a new tab

SMR association between HMGCR gene expression and FinnGen HS outcomes. HS = hemorrhagic stroke, pval = P-value, pval-multi = P-value of SMR-multi analysis, NA = not applicable to SMR-multi analysis, SMR = summary-data-based Mendelian randomization.

Figure 3.

Figure 3.

Open in a new tab

SMR association between HMGCR gene expression and UK Biobank HS outcomes. HMGCR = 3-hydroxy-3-methylglutaryl-coenzyme A reductase, SMR = summary-data-based Mendelian randomization.

3.2. Sensitivity analysis

All SNPs included in this study passed the test of F-statistic > 10, which can avoid bias from weak instruments in MR analysis (Table S2, Supplemental Digital Content, https://links.lww.com/MD/P691). Positive control analysis (Table S5, Supplemental Digital Content, https://links.lww.com/MD/P691) showed a significant result of blood HMGCR reducing blood LDL-C levels (OR = 1.48, 95% CI = 1.39–1.57, P < .001), which meant that the included blood eQTL SNPs represented blood HMGCR expression levels well. It was expected that the effect of brain HMGCR on blood LDL-C levels was nonsignificant (OR = 1.00, 95% CI = 0.98–1.03, P = .702), highlighting the importance of detecting the tissue-specific difference between blood and brain when discussing the topic of HMGCR affecting HS. All the SMR/SMR-multi analyses with strong/suggestive evidence have passed the Steiger and HEIDI tests (Table S4, Supplemental Digital Content, https://links.lww.com/MD/P691), signifying that our SMR results were reliable of the causal relation detection against the reverse causality and linkage disequilibrium.

3.3. Comparison of blood with the brain in the pattern of HMGCR affecting HS and HMGCR expression

The linear regression analysis showed a significantly negative relationship between the SMR effect estimates of HMGCR affecting HS in blood and that in brain (Intercept = −0.01, 95% CI = −0.03 to 0.01, P = .348; Slope = −0.23, 95% CI = −0.33 to −0.12, P = .004) (Fig. 4 and Tables S6 and S7, Supplemental Digital Content, https://links.lww.com/MD/P691). A significantly negative relationship from the Spearman correlation analysis was found between blood and brain HMGCR expression level (Spearman correlation = −0.36, P < .001) (Table S8, Supplemental Digital Content, https://links.lww.com/MD/P691), which would explain why the SMR effect estimates of HMGCR affecting HS from blood and brain tissue were “poles apart.” Specifically, gene expression differs in different tissues while the SNPs were relatively fixed across all the tissues – the same SNP in different tissues could associate with a certain gene expression in an opposite direction; when we use SNPs as gene-expression instruments for an MR causal analysis, we will, of course, obtain discrepant conclusions from different tissue-derived eQTL GWAS data.

Figure 4.

Figure 4.

Open in a new tab

The linear regression analysis of the SMR effect estimates of HMGCR affecting HS between blood and brain. HMGCR = 3-hydroxy-3-methylglutaryl-coenzyme A reductase, HS = hemorrhagic stroke, SMR = summary-data-based Mendelian randomization.

4. Discussion

To the best of our knowledge, this is the first SMR study powering concentration on the difference between blood and brain to assess the pleiotropic effects of HMGCR on HS. Our SMR study provided a statistically significant inverse association between the effect of HMGCR expression on HS in blood and that in brain. What is more, we explained this “contradictory” association by exploring the tissue-specific HMGCR-expression differences, concluding that HMGCR-expression in the brain increased HS incidence and that the peripheral circulation HMGCR-expression decreased the risk of HS was meaningful but “superficial.” Our study prioritized the central nervous system, not peripheral circulation, mattering in the issue of the HMGCR’s effects on the brain. Statin concentration in the brain would benefit patients by reducing the incidence of HS.

4.1. The biological pleiotropic effects of statin drugs

Statins exert their biological effects, including lowering cholesterol,[17] anti-thrombosis,[26] anti-inflammation,[26–28] etc, via downregulation of HMGCR. Concretely speaking, inhibition of HMGCR modulates the “mevalonate pathway,” by reducing the production from HMG-CoA to mevalonate. The reduction of subsequent isoprenoid substance – farnesyl pyrophosphate (FPP), lowers cholesterol through decreasing squalene levels and reduces the generation of geranylgeranyl pyrophosphate (GGPP).[29] FPP and GGPP, as the important lipid attachments and membrane anchors, affect the posttranslational modification of several proteins (i.e. small GTPases of the Ras, and Rho families).[30] These proteins are involved in several key cell signaling pathways of cytokine production, reactive oxygen species generation, etc, and are the intermediates to exert pleiotropic effects of statin drugs independent of lowering cholesterol.[17] (Fig. S1, Supplemental Digital Content, https://links.lww.com/MD/P690).

4.2. BBB and cholesterol metabolism in the brain

The BBB, which has a complex tight junction system with tightly packed cells in the brain capillaries[31] and the continuous basement membrane around the brain endothelial cells,[32] separates the brain from other organs, such as in cholesterol metabolism. The tight junction system of BBB naturally limits lipids’ transport from blood to the brain by vesicular mechanisms,[31,32] thus the main source of cholesterol in the brain largely relies on auto-synthesis, with few dependencies on dietary intake or hepatic synthesis. Cholesterol synthesis within the brain is similar to that of the liver, starting with HMGCR.[33] Although the brain holds about 25% of the whole content of cholesterol in the body, the synthesis rate within the brain is much lower (<2%) than that in the liver.[34] Hydroxylation by cholesterol 24-hydroxylase is the main pathway for brain cholesterol metabolism.[35] To summarize, brain HMGCR is crucial for brain cholesterol auto-synthesis to meet the structural and functional demands for cholesterol because brain cholesterol is fairly independent of blood LDL-C – the main form of cholesterol in systemic circulation – due to BBB.

Thus, based on the presence of BBB that differs brain from blood in cholesterol and, as an assumption, in other biologically active molecules made by the HMGCR, it is vitally important to explore the difference between blood and brain when studying the effects of HMGCR gene expression on HS.

4.3. Effect of HMGCR gene expression on HS differs in blood and brain

This SMR analysis gave an opposite effect of HMGCR expression on HS in blood and brain – blood HMGCR decreased HS while brain HMGCR increased the risk, which seemed inconceivable. We made further analyses to explain the confusing phenomenon. First, MR/SMR analysis uses βzx and βzy – the effect size (β) of SNP (z) on exposure (x) and outcome (y) respectively – to detect the causal relationship from × to y – βxy = βzy/ βzx.[18] The direction of βxy relies on both βzy and βzx. Specifically, suppose that the brain HMGCR gene has an inverse expression compared with the blood. In that case, the eQTL GWAS database will publish the opposite βzx in the brain against the blood, leading to an opposite MR result when different tissue-eQTL SNPs are used as exposure. Second, comparison analyses indeed confirmed this inference – the blood gene expression level had a significantly inverse relationship with the brain.

More specifically, when using HMGCR SNPs in blood, we got the opposite SMR results, of which the statistical significance was right but the direction was “wrong” or “superficial.” Because of the inverse gene expression in the brain and blood, if we adopt the peripheral circulation biomarkers such as blood LDL-C, blood HMGCR, blood LDL-C associated SNPs, or blood HMGCR associated SNPs as the research object, we will obtain a meaningful but “superficial” result – HMGCR or LDL-C will decrease the incidence of HS. This “paradox” could be found in other epidemiological and MR studies[12–16] which have proved that blood LDL-C levels lowered the risk of HS.

4.4. A never-ending debate about statin and HS

Though the lipophilic properties vary a lot from different statins, leading to different rates of serum statin delivery to the brain,[36] long-term usage of the quite hydrophilic statins (e.g. pravastatin) will still increase the statin concentrations in animal brains,[37] which implies that statins will affect the brain in a straightway through HMGCR gene expression. In addition to lowering cholesterol, inhibiting HMGCR brings various pleiotropic biological effects, such as antithrombotic and fibrinolytic properties that might increase the risk of HS,[26] and anti-inflammatory properties that might decrease the incidence of hemorrhagic transformation after ischemic stroke[38] or even improve the outcome after HS.[27,28] As the opposite opinions from different reviews,[26–28,38] numerous clinical and meta-analysis studies have also drawn different conclusions on whether statin usage increases the risk of HS. Of note, most of these clinical studies are retrospective, nonrandomized or observational, of which the results are susceptible to confounding bias. There seems to be a never-ending debate on the clinical implications of lipid-lowering therapy on HS, just as David Gaist and Magdy Selim said in their editorial article.[39]

From our viewpoint, investigating the effects of brain HMGCR expression on the brain itself is the key to letting the “endless debate” calm down. Based on the analyses of this SMR study, we concluded that HMGCR increased the risk of HS, supporting that statin drugs will benefit patients by reducing the incidence of HS.

4.5. Strength

There are several strengths in this SMR study. First, we used HMGCR gene expression eQTL SNPs as the proxy for the pleiotropic effects of statin, making the SMR results more powerful, reliable and explicable than using SNPs of other phenotypes such as LDL-C. Second, we conducted the tissue-specific analyses vital to the gene expression SMR analysis, due to the difference of gene expression in different tissues. Third, we explored the difference between blood and brain in the gene expression level, explaining the “conflicting” results. Fourth, our study cautions that MR analysis of gene expression has a character of tissue-specificity – you will get different or even contradictory MR results from diverse tissues, but only one is real about the truth. Fifth, our findings provided new insight into statin pleiotropic effects on HS and might influence recommendations for statin administration.

4.6. Limitations

This study also has several limitations. First, the sample size of brain HMGCR eQTL GWAS was relatively small. We scanned several eQTL databases such as GTEx Consortium V6-V8, BrainMeta Consortium V1-V2 and PsychENCODE Consortium, and we found only one brain HMGCR cis-eQTL meeting the inclusion criteria of our study, located in the Brain–Spinal cord (cervical c-1) from GTEx Consortium V7. The 2 factors might limit the statistical power, and account for the boundary P-value in the primary SMR analysis of the brain HMGCR gene expression’s effect on HS. We predict that the results of brain HMGCR increasing HS will be more solid using eQTL GWAS with larger sample sizes in the future. Second, due to the scarce available eQTL GWAS data meeting this study’s criterion, we only replace the outcome GWAS data to perform a replication analysis. Third, the GWAS data used in this study was all/predominantly obtained from European population ancestry, which guaranteed the reliability of our MR results but, on the other hand, restricted the extrapolation of our findings to other races.

5. Conclusion

In conclusion, this SMR study suggested that HMGCR expression in the Brain–Spinal cord (cervical c-1) increased the risk of HS. The “controversial” results from the blood eQTL data were due to the difference in HMGCR gene expression between the brain and blood. Addressing the tissue-specific effect on the relationship between statin and HS is the key to obtaining a robust, reliable and explicable result. Lipophilic statins, which could reach the central system from blood across BBB more easily than hydrophilic statins, might benefit patients more from the lower HS risk than the hydrophilic ones. Our study appeals to future clinical trials for a subgroup design comparing the effects of lipophilic statins with that of hydrophilic ones on the HS risk.

Acknowledgments

We want to acknowledge the participants and investigators of the FinnGen and UK Biobank studies. We also would like to thank the participants and investigators of eQTLGen, GTEx and Global Lipids Genetics Consortium.

Author contributions

Conceptualization: Xiaoxia Gu, Shengyuan Gu.

Data curation: Xinchao Du.

Formal analysis: Xiaoxia Gu, Zhiwei Yao, Shengyuan Gu.

Funding acquisition: Zhiwei Yao.

Methodology: Xinchao Du, Zhiwei Yao, Shengyuan Gu.

Project administration: Shengyuan Gu.

Resources: Xiaoxia Gu, Xinchao Du.

Validation: Zhiwei Yao.

Visualization: Xiaoxia Gu, Xinchao Du.

Writing – original draft: Xiaoxia Gu.

Writing – review & editing: Xinchao Du, Shengyuan Gu.

Supplementary Material

graphic file with name medi-104-e43838-s001.jpg

Open in a new tab

Abbreviations:

AHA
American Heart Association
ASCVD
atherosclerotic cardiovascular disease
BBB
blood-brain barrier
CVD
cardiovascular disease
eQTL
expression quantitative trait loci
FPP
farnesyl pyrophosphate
GGPP
geranylgeranyl pyrophosphate
GWAS
genome-wide association study
HMG-CoA
3-hydroxy-3-methylglutaryl-coenzyme A
HMGCR
3-hydroxy-3-methylglutaryl-coenzyme A reductase
HS
hemorrhagic stroke
ICH
intracerebral hemorrhage3
LDL-C
low-density lipoprotein cholesterol
MAF
minor allele frequency
MR
Mendelian randomization
NICH
nontraumatic intracranial hemorrhage
ROS
reactive oxygen species
SAH
subarachnoid hemorrhage
SMR
summary-data-based Mendelian randomization
SNPs
single-nucleotide polymorphisms
USPSTF
US Preventive Services Task Force

This work was supported by the Shandong Province medical health science and technology project (Grant No. 202307010965), the Science and Technology Innovation Plan of Medical Staff in Shandong Province (Grant No. SDYWZGKCJHLH2023096), the Tai’an Science and Technology Innovation Development Project (Grant No. 2023NS456), and the project supported by Shandong Provincial Natural Science Foundation (ZR2024QH113).

The funders of this study had no role in study design, data collection, data analysis, data interpretation, or the writing of the report.

The authors have no conflicts of interest to disclose.

The datasets generated during and/or analyzed during the current study are publicly available.

Supplemental Digital Content is available for this article.

How to cite this article: Gu X, Du X, Yao Z, Gu S. Statins and hemorrhagic stroke: New insight from a Mendelian randomization study. Medicine 2025;104:32(e43838).

XG, XD, ZY, SG contributed equally to this work.

Contributor Information

Xiaoxia Gu, Email: gootionyuan@126.com.

Xinchao Du, Email: duthingchao@126.com.

Zhiwei Yao, Email: aspiring_yao@126.com.

References

  • [1].Braunwald E. Cholesterol: the race to the bottom. Eur Heart J. 2021;42:4612–3. [DOI] [PubMed] [Google Scholar]
  • [2].Mach F, Baigent C, Catapano AL, et al. ; ESC Scientific Document Group. 2019 ESC/EAS Guidelines for the management of dyslipidaemias: lipid modification to reduce cardiovascular risk. Eur Heart J. 2020;41:111–88. [DOI] [PubMed] [Google Scholar]
  • [3].Mangione CM, Barry MJ, Nicholson WK, et al. ; US Preventive Services Task Force. Statin use for the primary prevention of cardiovascular disease in adults: US preventive services task force recommendation statement. JAMA. 2022;328:746–53. [DOI] [PubMed] [Google Scholar]
  • [4].Blaha MJ, Martin SS. How do statins work?: changing paradigms with implications for statin allocation. J Am Coll Cardiol. 2013;62:2392–4. [DOI] [PubMed] [Google Scholar]
  • [5].Gu Q, Paulose-Ram R, Burt VL, Kit BK. Prescription cholesterol-lowering medication use in adults aged 40 and over: United States, 2003-2012. NCHS Data Brief. 2014;177:1–8. [PubMed] [Google Scholar]
  • [6].Desai CS, Martin SS, Blumenthal RS. Non-cardiovascular effects associated with statins. Bmj. 2014;349:g3743. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [7].Newman CB, Preiss D, Tobert JA, et al. ; American Heart Association Clinical Lipidology, Lipoprotein, Metabolism and Thrombosis Committee, a Joint Committee of the Council on Atherosclerosis, Thrombosis and Vascular Biology and Council on Lifestyle and Cardiometabolic Health; Council on Cardiovascular Disease in the Young; Council on Clinical Cardiology; and Stroke Council. Statin Safety and associated adverse events: a scientific statement from the American Heart Association. Arterioscler Thromb Vasc Biol. 2019;39:e38–81. [DOI] [PubMed] [Google Scholar]
  • [8].Goldstein LB, Toth PP, Dearborn-Tomazos JL, et al. ; American Heart Association Council on Arteriosclerosis, Thrombosis and Vascular Biology; Council on Cardiovascular and Stroke Nursing; Council on Peripheral Vascular Disease; and Stroke Council. Aggressive LDL-C Lowering and the brain: impact on risk for dementia and hemorrhagic stroke: a scientific statement from the American Heart Association. Arterioscler Thromb Vasc Biol. 2023;43:e404–42. [DOI] [PubMed] [Google Scholar]
  • [9].Richmond RC, Smith GD. Mendelian Randomization: concepts and Scope. Cold Spring Harb Perspect Med. 2022;12:a040501. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [10].Smith GD, Ebrahim S. “Mendelian randomization”: can genetic epidemiology contribute to understanding environmental determinants of disease? Int J Epidemiol. 2003;32:1–22. [DOI] [PubMed] [Google Scholar]
  • [11].Porcu E, Rüeger S, Lepik K, Santoni FA, Reymond A, Kutalik Z; eQTLGen Consortium. Mendelian randomization integrating GWAS and eQTL data reveals genetic determinants of complex and clinical traits. Nat Commun. 2019;10:3300. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [12].Falcone GJ, Kirsch E, Acosta JN, et al. ; International Stroke Genetics Consortium. Genetically elevated LDL associates with lower risk of intracerebral hemorrhage. Ann Neurol. 2020;88:56–66. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [13].Ma C, Gurol ME, Huang Z, et al. Low-density lipoprotein cholesterol and risk of intracerebral hemorrhage: a prospective study. Neurology. 2019;93:e445–57. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [14].Pezzini A, Grassi M, Iacoviello L, et al. ; Multicenter Study on Cerebral Haemorrhage in Italy (MUCH-Italy) Investigators. Serum cholesterol levels, HMG-CoA reductase inhibitors and the risk of intracerebral haemorrhage. The Multicenter Study on Cerebral Haemorrhage in Italy (MUCH-Italy). J Neurol Neurosurg Psychiatry. 2016;87:924–9. [DOI] [PubMed] [Google Scholar]
  • [15].Rist PM, Buring JE, Ridker PM, Kase CS, Kurth T, Rexrode KM. Lipid levels and the risk of hemorrhagic stroke among women. Neurology. 2019;92:e2286–94. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [16].Zhang X, Patel A, Horibe H, et al. ; Asia Pacific Cohort Studies Collaboration. Cholesterol, coronary heart disease, and stroke in the Asia Pacific region. Int J Epidemiol. 2003;32:563–72. [DOI] [PubMed] [Google Scholar]
  • [17].Bahrami A, Bo S, Jamialahmadi T, Sahebkar A. Effects of 3-hydroxy-3-methylglutaryl coenzyme A reductase inhibitors on ageing: molecular mechanisms. Ageing Res Rev. 2020;58:101024. [DOI] [PubMed] [Google Scholar]
  • [18].Zhu Z, Zhang F, Hu H, et al. Integration of summary data from GWAS and eQTL studies predicts complex trait gene targets. Nat Genet. 2016;48:481–7. [DOI] [PubMed] [Google Scholar]
  • [19].Võsa U, Claringbould A, Westra HJ, et al. ; BIOS Consortium. Large-scale cis- and trans-eQTL analyses identify thousands of genetic loci and polygenic scores that regulate blood gene expression. Nat Genet. 2021;53:1300–10. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [20].Battle A, Brown CD, Engelhardt BE, Montgomery SB; GTEx Consortium. Genetic effects on gene expression across human tissues. Nature. 2017;550:204–13. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [21].Kurki MI, Karjalainen J, Palta P, et al. ; FinnGen. FinnGen provides genetic insights from a well-phenotyped isolated population. Nature. 2023;613:508–18. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [22].Burgess S, Thompson SG; CRP CHD Genetics Collaboration. Avoiding bias from weak instruments in Mendelian randomization studies. Int J Epidemiol. 2011;40:755–64. [DOI] [PubMed] [Google Scholar]
  • [23].Willer CJ, Schmidt EM, Sengupta S, et al. ; Global Lipids Genetics Consortium. Discovery and refinement of loci associated with lipid levels. Nat Genet. 2013;45:1274–83. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [24].Hemani G, Tilling K, Smith GD. Orienting the causal relationship between imprecisely measured traits using GWAS summary data. PLoS Genet. 2017;13:e1007081. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [25].Wu Y, Zeng J, Zhang F, et al. Integrative analysis of omics summary data reveals putative mechanisms underlying complex traits. Nat Commun. 2018;9:918. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [26].Violi F, Calvieri C, Ferro D, Pignatelli P. Statins as antithrombotic drugs. Circulation. 2013;127:251–7. [DOI] [PubMed] [Google Scholar]
  • [27].Chen CJ, Ding D, Ironside N, et al. Statins for neuroprotection in spontaneous intracerebral hemorrhage. Neurology. 2019;93:1056–66. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [28].Xue M, Yong VW. Neuroinflammation in intracerebral haemorrhage: immunotherapies with potential for translation. Lancet Neurol. 2020;19:1023–32. [DOI] [PubMed] [Google Scholar]
  • [29].Istvan ES, Deisenhofer J. Structural mechanism for statin inhibition of HMG-CoA reductase. Science. 2001;292:1160–4. [DOI] [PubMed] [Google Scholar]
  • [30].Zhang FL, Casey PJ. Protein prenylation: molecular mechanisms and functional consequences. Annu Rev Biochem. 1996;65:241–69. [DOI] [PubMed] [Google Scholar]
  • [31].Abbott NJ, Patabendige AA, Dolman DE, Yusof SR, Begley DJ. Structure and function of the blood-brain barrier. Neurobiol Dis. 2010;37:13–25. [DOI] [PubMed] [Google Scholar]
  • [32].de Boer AG, Gaillard PJ. Blood-brain barrier dysfunction and recovery. J Neural Transm (Vienna). 2006;113:455–62. [DOI] [PubMed] [Google Scholar]
  • [33].Nieweg K, Schaller H, Pfrieger FW. Marked differences in cholesterol synthesis between neurons and glial cells from postnatal rats. J Neurochem. 2009;109:125–34. [DOI] [PubMed] [Google Scholar]
  • [34].Mitsche MA, McDonald JG, Hobbs HH, Cohen JC. Flux analysis of cholesterol biosynthesis in vivo reveals multiple tissue and cell-type specific pathways. Elife. 2015;4:e07999. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [35].Björkhem I, Lütjohann D, Breuer O, Sakinis A, Wennmalm A. Importance of a novel oxidative mechanism for elimination of brain cholesterol. Turnover of cholesterol and 24(S)-hydroxycholesterol in rat brain as measured with 18O2 techniques in vivo and in vitro. J Biol Chem. 1997;272:30178–84. [DOI] [PubMed] [Google Scholar]
  • [36].Climent E, Benaiges D, Pedro-Botet J. Hydrophilic or Lipophilic Statins? Front Cardiovasc Med. 2021;8:687585. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [37].Johnson-Anuna LN, Eckert GP, Keller JH, et al. Chronic administration of statins alters multiple gene expression patterns in mouse cerebral cortex. J Pharmacol Exp Ther. 2005;312:786–93. [DOI] [PubMed] [Google Scholar]
  • [38].Ma G, Pan Z, Kong L, Du G. Neuroinflammation in hemorrhagic transformation after tissue plasminogen activator thrombolysis: Potential mechanisms, targets, therapeutic drugs and biomarkers. Int Immunopharmacol. 2021;90:107216. [DOI] [PubMed] [Google Scholar]
  • [39].Gaist D, Selim M. Lipid-lowering therapy and hemorrhagic stroke: a never-ending debate of clinical implications. have we found the answer? J Am Heart Assoc. 2024;13:e034099. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

medi-104-e43838-s002.pdf (4MB, pdf)

Articles from Medicine are provided here courtesy of Wolters Kluwer Health

RESOURCES