Lipid-lowering and antihypertensive drugs on aortic disease risk: insights from Mendelian randomization analysis and real-world pharmacovigilance data

Abstract

Objective

To assess the impact of lipid-lowering drugs (LLDs) and antihypertensive drugs on the risk of aortic diseases.

Methods

Mendelian randomization was utilized to analyse data from 500 000 participants in the UK Biobank to evaluate the effects of statins, PCSK9 inhibitors (PCSK9i), β-blockers, and calcium channel blockers on the risks of thoracic aortic aneurysm, abdominal aortic aneurysm, and aortic dissection (AD) using genetic variants as proxies. Real-world pharmacovigilance data from the FAERS (FDA Adverse Event Reporting System) database were used.

Results

PCSK9i and statins significantly reduced the risks of aortic aneurysms and AD, respectively. Furthermore, the two LLDs reduced the risk of aortic diseases through certain metabolites. Meanwhile, real-world pharmacovigilance reports also indicated a low incidence of aortic diseases with PCSK9i and statin treatment.

Conclusion

LLDs, particularly statins and PCSK9i, significantly protect against aortic diseases, providing a scientific basis for preventing and treating aortic diseases.

Graphical Abstract

Schematic overview of the causal analysis framework used to investigate the effects of lipid-lowering and antihypertensive drugs on aortic diseases.

Open in new tabDownload slide

Abbreviations

Abbreviations
 
• AA

  aortic aneurysm

 
• AAA

  abdominal aortic aneurysm

 
• ABs

  α-blockers

 
• ACEIs

  angiotensin-converting enzyme inhibitors

 
• AD

  aortic dissection

 
• ARBs

  angiotensin II receptor blockers

 
• AE

  adverse events

 
• BBs

  β-blockers

 
• CCBs

  calcium channel blockers

 
• CVD

  cardiovascular disease

 
• FAERS

  FDA Adverse Event Reporting System

 
• FDA

  Food and Drug Administration

 
• GWAS

  Genome-Wide Association Studies

 
• HDL-C

  HDL cholesterol

 
• IVW

  inverse variance weighted

 
• LDL-C

  LDL cholesterol

 
• MR

  Mendelian randomization

 
• OR

  odds ratio

 
• PCSK9i

  PCSK9 inhibitors

 
• ROR

  reporting odds ratio

 
• SNPs

  single-nucleotide polymorphisms

 
• TAA

  thoracic aortic aneurysm

 
• VLDL-C

  very low-density lipoprotein cholesterol

Introduction

Cardiovascular diseases (CVDs), particularly aortic diseases, are the leading cause of death worldwide.¹ Aortic diseases such as thoracic aortic aneurysm (TAA), abdominal aortic aneurysm (AAA), and aortic dissection (AD) represent significant subcategories of CVD with severe health implications.^2–4 Dyslipidaemia and hypertension are the major risk factors associated with these conditions.⁵ Lipid-lowering drugs (LLDs) and antihypertensive drugs (AHTDs) are usually recommended, and their roles in reducing cardiovascular events have been extensively researched.^6,7

LLDs and AHTDs have garnered substantial attention for their potential in preventing and treating aortic diseases.^8–10 However, despite substantial evidence confirming the significant role of these medications in reducing cardiovascular events, their impact on aortic diseases remains elusive.^11,12 Therefore, exploring their precise effects on TAA, AAA, and AD is imperative.

Several studies have explored the relationship between lipid levels and aortic diseases. Allara et al.¹³ found that lowering LDL cholesterol (LDL-C) prevents ADs. Higher HDL cholesterol (HDL-C) levels reduced aortic aneurysm (AA) risk, while elevated triglyceride levels increased it.¹⁴ PCSK9 inhibitors (PCSK9i) and statins both lower LDL-C, but PCSK9i were less effective at reducing very low-density lipoprotein cholesterol (VLDL-C).¹⁵

Hypertension is a well-established risk factor for aortic diseases. AHTDs such as α-blockers (ABs) and β-blockers (BBs) can control blood pressure and stress on aortic wall to reduce cardiovascular risk.^16,17 Calcium channel blockers (CCBs), angiotensin-converting enzyme inhibitors (ACEIs), and angiotensin II receptor blockers (ARBs) can lead to vasodilation and reduced blood pressure.^18,19 Kwok and Schooling²⁰ suggested that PCSK9i and AHTDs may extend the male lifespan, emphasizing potential benefits in the comprehensive management of aortic diseases.

This study integrates Mendelian randomization (MR) analysis with real-world pharmacovigilance data from the FDA Adverse Event Reporting System (FAERS) database. MR provides a robust framework for causal inference by leveraging genetic variants as proxies for drug targets, while the FAERS offers comprehensive real-world evidence of adverse drug events. By combining these complementary methodologies, this study aims to fill a critical knowledge gap in aortic disease research, providing both genetic and clinical validation of the effects of antihypertensive and lipid-lowering therapies. This novel approach represents a significant advancement in understanding the interplay between therapeutic interventions and aortic disease risk, offering a framework that can be applied to other complex diseases and drug classes.

Methods

The information about databases and sources used in this study is shown in Supplementary material online, Table S2. All systems are accessible to users. Due to the nature of the data sources, informed consent from the original patients was not required for this study. The human-related datasets used in this study were provided by third parties and are available for open use in any research project, with appropriate ethical approval obtained for each study where the data were utilized. Approximately 2 000 000 participants, primarily of European ancestry, were recruited from the UK Biobank (https://www.ukbiobank.ac.uk/), the Genome-Wide Association Studies (GWAS) Catalog (https://www.ebi.ac.uk/gwas/home), and FinnGen (https://www.finngen.fi/en). Additionally, 13 038 441 real-world pharmacovigilance reports were retrieved from the FAERS database, with the majority originating from the United States, Germany, the United Kingdom (UK), and the Netherlands.

See supplementary material online, Appendix E1 for details.

Results

Validation of genetic instruments for drugs

For LLDs, 5 and 11 genetic variants in the HMGCR and PCSK9 genes, respectively, were selected as proxies for statins and PCSK9i. For AHTDs, 12, 6, 24, 5, and 3 genetic variants were selected as proxies for ABs, BBs, CCBs, ACEIs, and ARBs, respectively (see details in Supplementary material online, Table S1).

MR analysis for drug effects on aortic diseases

MR analysis results revealed that different classes of cardiovascular drugs had varying effects on the risk of TAA, AAA, and AD. Unlike AHTDs, LLDs showed exceptional effectiveness in aortic diseases. Statins significantly negatively correlated with the risk of TAA, AAA, and AD [inverse-variance weighted (IVW), odds ratio (OR) = 0.527, P < 0.001; IVW, OR = 0.228, P < 0.001; and IVW, OR = 0.248, P = 0.012, respectively]. Although PCSK9i did not show a statistically significant difference in the risk of AD, they significantly reduced the risk of TAA and AAA (IVW, OR = 0.805, P = 0.002; and IVW, OR = 0.595, P < 0.001, respectively) (Figure 1 and Table 1). No statistically significant heterogeneity and pleiotropy were present (Tables 2 and 3).

In this study, additional independent datasets were utilized to assess the relationship between LLDs and AHTDs and aortic disease risk. Due to the unavailability of other AD data, only TAA and AAA were reassessed. PCSK9i maintained a low-risk relationship with TAA and AAA (IVW, OR = 0.562, P = 0.039; and IVW, OR = 0.718, P = 0.030, respectively) but statins were only significantly negatively correlated with AAA risk (IVW, OR = 0.427, P < 0.001), with no statistical difference in TAA. Furthermore, CCBs were not statistically significantly associated with TAA risk. No statistically significant heterogeneity and pleiotropy were present (see Supplementary material online, Figures 1 and 2; Tables 1–3).

MR analysis of circulating metabolites and aortic diseases

MR and IVW methods were applied to assess the relationships between different metabolite levels and various single-nucleotide polymorphisms (SNPs) in aortic disease populations to further explore whether the effects of LLDs in the body are mediated through metabolites (see Supplementary material online, Table S6).

Fifty-nine metabolites were associated with TAA risk (Figure 2; see Supplementary material online, Table S7). 4-Guanidinobutanoate (OR = 1.197, P = 0.044), octadecadienoate (18:2) (OR = 1.196, P = 0.009), palmitoyl-sphingosine-1-phosphate (d18:1/16:0) (OR = 1.098, P = 0.041), and 1-linoleoyl-glycerol (18:2) (OR = 1.071, P = 0.044) levels were significantly positively associated with the risk of TAA. Conversely, 5-hydroxyindole (OR = 0.972, P = 0.022), hexadecanedioate (C16-DC) (OR = 0.872, P = 0.042), tetradecanedioate (OR = 0.862, P = 0.017), and sulfo-3-methylthiopropyl cysteine (OR = 0.841, P = 0.018) levels were inversely associated with TAA risk, suggesting a protective effect. No statistically significant heterogeneity or pleiotropy was observed (see Supplementary material online, Table S8).

Moreover, significant associations were found between specific circulating metabolites and AAA risk. Fifty-eight metabolites had

 

 

significant associations (Figure 3; see Supplementary material online, Table S9). Benzoate (OR = 1.113, P = 0.017), 5-hydroxyindole (OR = 1.144, P = 0.015), and linoelaidylcarnitine (OR = 1.134, P = 0.035) levels were significantly positively associated with an increased risk of AAA. No statistically significant heterogeneity or pleiotropy was found (see Supplementary material online, Table S10).

Seventy-three circulating metabolites were significantly associated with AD risk (Figure 4; see Supplementary material online, Table S11). Tyrosine (OR = 1.420, P = 0.003), 1-methyl-4-imidazoleacetate (OR = 1.220, P = 0.046), hexadecanedioate (C16-DC) (OR = 2.742, P = 0.007), γ-glutamyltyrosine (OR = 1.411, P = 0.018), and androstenediol (3α, 17α) monosulfate (3) (OR = 0.712, P = 0.025) levels were significantly positively associated with the risk of AD. No statistically significant heterogeneity or pleiotropy was detected (see Supplementary material online, Table S12).

Mediation effect analysis

In the mediation analysis (Figure 5 and Table 4), PCSK9i, lignoceroyl sphingomyelin (d18:1/24:0) exhibited the highest mediation effect on TAA, with 57.4% of the mediation effects of PCSK9i. Campesterol and deoxycholic acid 12-sulfate levels also showed substantial mediation effects on TAA, with 30.2% and 16.7% mediation effects of PCSK9i, respectively.

For statins, the adenosine 5′-monophosphate (AMP) to urate ratio had a 24% mediation effect on AD, suggesting a strong influence through purine metabolism pathways. Tetradecanedioate (C14-DC) levels mediate 17.5% of statin effects on AAA, indicating the relevance of fatty acid metabolism in statin action.

Characteristics of real-world drug reports

The FAERS database was utilized to explore the relationship between LLDs and aortic diseases. The FAERS database received 14 931 458 reports between the first quarter of 2015 and the first quarter of 2024. After excluding duplicates and outliers based on the data cleaning criteria of the FAERS, 13 038 441 reports were retained. A total of 152 284 reports for PCSK9i, 675 511 for statins, and 1 838 035 or 2 140 233 for control group drugs, with at least one PCSK9i, statin, and control group drug as a suspect or concomitant medication, respectively, were identified (Figure 6). Over 39% of patients were 65 years or older, with fewer males than females taking PCSK9i (39.2% vs. 53.4%). Most reports came from healthcare professionals, with the United States leading in report numbers across all groups (Table 5).

Disproportionality analysis of LLDs and aortic diseases

The proportion of reports identifying AA as an adverse event (AE) was significantly lower in the PCSK9i group than in the control group, with a reporting odds ratio (ROR) of 0.631 (P < 0.05) and an IC₀₂₅ of −2.298 (Figure 7 and Table 6). This indicates an abnormally low incidence of AA with PCSK9i compared with control group drugs, consistent with results from MR analyses. Diabetes is often accompanied with dyslipidaemia, increasing their likelihood of taking LLDs. Since diabetes is a negative factor for AA, this may lead to inaccuracies in the results. Intriguingly, the PCSK9i group still showed a considerably low incidence of AA (ROR = 0.569, P < 0.01) even after excluding reports from patients with diabetes. This disproportionality was consistently high in the female gender and age >60 subgroups (Figure 7 and Table 6).

MR analysis showed that statins had the same protective effect in AAA as PCSK9i but not in TAA. However, the FAERS database does not distinguish between TAA and AAA but groups them as AA. PCSK9i are often used as an alternative for inadequate response to statins. A comparison of AA incidence between PCSK9i and statins revealed that the PCSK9i group showed a lower AA incidence (ROR = 0.631, P < 0.001) than the statins group. This consistency was maintained in gender, age >60, and no diabetes subgroups (Figure 7 and Table 6). Collectively, these results demonstrate that PCSK9i exhibit superior efficacy in reducing AA incidence compared with other drugs (Figure 7 and Table 6).

Unlike AA, MR analysis revealed a significant negative correlation between statins and the risk of AD (Figure 1). The proportion of reports identifying AD as an AE was significantly lower in the statin than in the control group, with an ROR of 0.620 (P < 0.05) and an IC₀₂₅ of −2.127. This indicates an abnormally low AD incidence with statins compared with control group drugs, confirming previous findings. Subgroup analyses also revealed that statins maintained a lower AD incidence under conditions of male gender, age >60, and no diabetes subgroups (Figure 8 and Table 7), highlighting the exceptional performance of statins in reducing AD risk.

Discussion

Surgery is the preferred treatment for aortic diseases but it is not effective in all patients. Therefore, new interventions should be developed. Hypertension, hyperlipaemia, smoking, gender, and age are crucial factors to consider when assessing aortic disease risk in a patient. The underlying mechanisms of aortic diseases likely involve an interplay between genetic predispositions and these acquired risk factors.²¹ Studies have shown that LLDs can prevent AAs and aortic valve stenosis.^22–24 However, reports on the relationship between LLDs and the risk of various aortic diseases remain limited and inconclusive.

This study employed multiple MR methods to evaluate the effects of cardiovascular drugs, including LLDs (mediated by statins or PCSK9i) and blood pressure-lowering agents (mediated by ABs, BBs, CCBs, ACEIs, or ARBs), on the risk of aortic diseases. Results showed that PCSK9i exhibited significant protective effects in reducing the risk of AA. Additionally, statins significantly reduced AD risk.

The distinct origins of smooth muscle cells in the thoracic aorta (neural crest and somitic mesoderm) compared with those in the abdominal aorta (splanchnic mesoderm) may underlie the differing pathogenic mechanisms between AD and AAA. Additionally, AD exhibits notable differences from AAA in terms of population prevalence, patterns of inheritance, and the specific genes associated with predisposition.²⁵ Statin ameliorates endothelial dysfunction via up-regulation of endothelial nitric oxide synthase (eNOS) and endothelium-derived nitric oxide production, inhibition of Rho prenylation, and other antioxidant effects.²⁶ Endothelial dysfunction plays an important role in the pathogenesis of vascular remodelling, including thoracic AD.²⁷ PCSK9i primarily influence cholesterol metabolism by regulating the degradation of LDL receptors.¹⁵ Lowering systemic inflammation and slowing atherosclerosis progression may protect against AA development. PCSK9i, often used when conventional LLDs are ineffective, have stronger lipid-lowering effects and more effectively reduce arterial wall inflammation.²⁸ AD may be characterized by a tear in the intima, primarily induced by arteriosclerosis and abnormal matrix metalloproteinases (MMP) activity. Unlike PCSK9i, statins also improve endothelial function and provide antioxidant benefits.^29,30 Endothelial cells form the arterial intima, so statins may lower AD risk by improving endothelial function. This may explain the differing effects of PCSK9i and statins on these two diseases.¹⁶ Real-world drug reports also indicated that except in the male gender and age 40–60 subgroups, PCSK9i displayed a lower rate of AA reports in other subgroups than in the control or the statin group. Compared with the control group, statins maintained a lower reporting ratio in all subgroups, except for the female gender and age 40–60 subgroups, in reports with AD as the major disease.

The use of genetic proxies enhances causal inferences by minimizing confounding factors. Unlike traditional observational studies, MR provides more reliable insights. Real-world drug data further highlight the roles of PCSK9i and statins in reducing aortic diseases, supporting future personalized medicine and prevention strategies.³¹ Nonetheless, this study has some limitations including use of genetic tools that explain only a small proportion of phenotypes,³² focus on cardiovascular drugs, excluding other non-drug factors that affect aortic disease risk,³³ and possibility of other confounding genetic pathways.^31,34 Although real-world drug reports were used to enhance the accuracy of the results, these reports were from subjects with a history of medication use and diseases other than the target disease. The effect of all potential confounding factors could not be excluded in the analyses.

Study limitations

This study had some limitations that need to be considered in the interpretation of the findings. First, it was an observational study based on multiple sources; therefore, reverse causality might exist. However, the study rigorously adjusted for confounding factors and validated the association through MR analysis, thereby addressing this issue to the best extent possible. Second, we acknowledge that the FAERS database inherently includes subjects with comorbidities unrelated to the target disease. To minimize confounding, we carefully excluded subjects with conditions that significantly influence aortic disease risk. However, it remains challenging to identify a population exclusively affected by the target disease, given the nature of real-world pharmacovigilance data. Third, there was a chance of misclassification of antihypertensive agents used in the UK Biobank data during follow-up because the antihypertensive agent use was only evaluated once at baseline. Fourth, our study included the largest set of real-world data for PCSK9i spanning nearly a decade; the inherent limitations of the FAERS database, including its spontaneous reporting nature and the relatively short post-marketing period of PCSK9i, must be acknowledged. These factors may limit the ability to detect rare or long-term AEs associated with these medications. To further validate these findings, long-term cohort studies and clinical trials with extended follow-up periods are needed to better assess the safety profile and clinical impact of PCSK9i on aortic disease risk.

Conclusions

In summary, this study highlights the potential value of genetic information in the prevention and treatment of aortic diseases. Specifically, statins and PCSK9i offer new perspectives in the development of new drugs and the design of precision medicine strategies. Elevated levels of lipoprotein(a) [Lp(a)] are causally associated with an increased risk of AAs, as well as AD, through pathways involving pro-atherogenic, pro-inflammatory, and pro-thrombotic mechanisms.^35,36 This study would provide a roadmap for continuing investigations into the connections between more risk factors, emerging drugs, and their impact on aortic disease risk.

Acknowledgements

The acknowledgments to GWAS consortia were described in detail in Supplementary material online, Tables S1 and S2. This study conforms to the principles outlined in the Declaration of Helsinki.

Funding

This work was supported by the National Natural Science Foundation of China (grant numbers 82060095 and 82460103), the Key Science and Technology Innovation Project of Jiangxi Provincial Health Commission (grant number 2024ZD005), and the Second Affiliated Hospital of Nanchang University Funding Program (2023efyA02).

Conflict of interest: none declared.

Data availability

All data used in this study are publicly available, and the source of the data is described in the main text.

Author contributions

H.N. and W. Zhao conceived the original conception, designed the experiment plan, carried out data analysis, and drafted the manuscript. Q.W revised the manuscript. W. Zhou reviewed and edited the manuscript. All authors have read and agreed to the published version of the manuscript.

References

Author notes