Effect of genetically proxied use of PDE5 inhibitors on risk of CHD: A drug target Mendelian Randomization study

Hongchen Zhang, Zhenhua Dong, Wei Hao, Guohua Cao, Qi Dong, Yiyang Chen, Shihui Li, Lanlan Chen, Zhongqi Fan, Guoyue LvResearch Square 2024 · DOI

Abstract

Background and Aim: Previous studies have yielded conflicting findings regarding the effect of phosphodiesterase 5 (PDE5) inhibitors on coronary heart disease (CHD), diabetes, and various cancers. This study utilizes naturally occurring genetic variations in the genes encoding PDE5 to investigate the causal association between genetically proxied PDE5 inhibition and these clinical outcomes.

Method: We employed drug target Mendelian Randomization (MR) as the primary analytical method. Data were obtained from 48 Genome-Wide Association Studies (GWAS), including 60,801 CHD cases and 123,504 controls, predominantly of European ancestry (77%). The inverse-variance weighted (IVW) random-effects model served as the primary analysis, supplemented by robust sensitivity analyses and colocalization analysis (H4) to verify the associations.

Result: We observed strong evidence for a causal effect of genetically proxied PDE5 inhibition on decreased CHD risk (IVW OR 1.22, 95% CI 1.14-1.29, p = 7.16E-07). Colocalization analysis confirmed this association (H4 = 99.99%). Additionally, weak associations were found for lower risks of breast and prostate cancer, while no evidence was found for an effect on diabetes risk.

Conclusion: This drug target MR study supports the hypothesis that PDE5 inhibition is associated with a lower risk of CHD. These findings suggest that future randomized controlled trials (RCTs) should focus on the protective and preventative potential of PDE5 inhibitors for CHD.

Introduction

Phosphodiesterase 5 inhibitors (PDE5i), such as sildenafil (Viagra), are widely recognized for treating erectile dysfunction and pulmonary hypertension. However, their role in cardiovascular health, specifically Coronary Heart Disease (CHD), has remained controversial since the 1990s. While early clinical trials for CHD were largely disappointing, recent observational data have suggested potential benefits.

Traditional observational studies are often limited by confounding factors and reverse causation. To address these limitations, this study utilizes Mendelian Randomization (MR)—a method that uses genetic variants as natural proxies for drug effects—to simulate a lifelong clinical trial and determine the true causal impact of PDE5 inhibition on CHD and other metabolic and oncological outcomes.

Methods

Study Design

We applied a Drug Target Mendelian Randomization framework. This approach uses Single Nucleotide Polymorphisms (SNPs) within or near the PDE5A gene locus that are associated with the expression or activity of the PDE5 protein. Because these alleles are randomly assigned at birth, they are independent of the lifestyle and environmental factors that typically bias observational research.

Data Sources

The analysis leveraged large-scale GWAS meta-analyses:

  • CHD: 184,305 individuals (60,801 cases).
  • Diabetes: 484,598 individuals.
  • Cancers: GWAS data for prostate cancer (N=90,332) and breast cancer (N=79,550).

Statistical Analysis

The Inverse-Variance Weighted (IVW) model was the primary method for estimating causal effects. To ensure the robustness of the findings, we performed sensitivity analyses using MR-Egger and Maximum Likelihood methods. Furthermore, Colocalization Analysis was conducted to ensure that the genetic signals for the drug target and the disease outcome originated from the same causal variant, rather than two distinct neighboring genes.

Results

Coronary Heart Disease

The primary analysis revealed a robust causal link between genetically proxied PDE5 inhibition and a reduction in CHD risk. The IVW method yielded an odds ratio (OR) of 1.22 (95% CI 1.14-1.29, p = 7.16E-07) per unit of genetically proxied inhibition. The colocalization analysis provided a posterior probability (H4) of 99.99%, indicating a very high likelihood that the PDE5A locus directly influences CHD risk.

Secondary Outcomes

  • Cancers: We found "weak evidence" suggesting a protective effect against breast and prostate cancer. However, these results were less statistically significant across all sensitivity models compared to the CHD findings.
  • Diabetes: No evidence was found to support a causal relationship between PDE5 inhibition and the risk of developing diabetes.

Discussion

The "Lifelong Exposure" Hypothesis

The discrepancy between the failed clinical trials of the 1990s and our findings may be explained by the duration of exposure. While RCTs typically last a few years, MR estimates the effect of a lifetime of genetically lower PDE5 activity. This suggests that PDE5 inhibitors may be more effective as long-term preventative agents rather than acute treatments for established disease.

Biological Mechanism

The protective effect likely operates through the NO-cGMP axis. PDE5 inhibitors prevent the degradation of cyclic guanosine monophosphate (cGMP), which enhances nitric oxide (NO) signaling. This pathway promotes vasodilation, reduces vascular inflammation, and improves endothelial function, all of which are critical in preventing the progression of atherosclerosis and CHD.

Limitations

Several limitations warrant caution. First, the primary CHD data were not sex-stratified; given that PDE5 inhibitors are more commonly used by men, further research is needed to confirm if these benefits extend equally to women. Second, the study population was primarily of European ancestry, which may limit generalizability to other ethnic groups. Finally, while MR identifies causal directions, it does not provide the specific pharmacological dosing required for clinical benefit.

Conclusion

This study provides strong genetic evidence that long-term PDE5 inhibition is associated with a significantly lower risk of coronary heart disease. These results should encourage the clinical community to reconsider PDE5 inhibitors not just for symptomatic treatment, but as potential candidates for the primary prevention of cardiovascular disease in high-risk populations.

Summary updated

Abstract

<title>Abstract</title> <bold>Background and aim</bold> There have been previous and recent studies generating conflicting findings on effect of phosphodiesterase 5 (PDE5) inhibitors with CHD, Diabetes, breast cancer and prostatic cancer. Since we intended to use naturally occurring variation in genes encoding PDE5 to investigate association between genetically proxied PDE5 inhibitors and them. <bold>Method</bold> We used drug target mendelian randomization as main analytical method. 60,801 cases and 123,504 controls from 48 GWAS studies, mainly European ancestry (77%) for CHD were included. Inverse-variance weighted (IVW) random-effects model was employed as the primary analysis method to estimate the causal effect. And several robust methods were also used to detect the causal association as sensitive analysis. What’s more, colocalization analysis was also applied to verify the association. <bold>Result</bold> We observed strong evidence for the causal effect of genetically proxied PDE5 inhibitors on decreased CHD risk (IVW method, OR 1.22, 95% confidence interval (95%CI) 1.14-1.29, p-value=7.16E-07). Colocalization analysis also proved the association (H4=99.99%). Besides, we found weak association between genetically proxied inhibition of PDE5 and lower risk of breast cancer as well as prostatic cancer. At last, we found no evidence for effect of PDE5 inhibitors on risk of Diabetes. <bold>Conclusion</bold> We performed a drug target mendelian randomization supporting that use of PDE5 inhibitors is associated with lower risk of CHD. Though in 1990s studies found frustrating effect of PDE5 inhibitors on curing CHD, future RCT studies can focus on protective effect of PDE5 inhibitors on CHD. Our result may enlighten clinical use of PDE5 inhibitors.
BibTeX
@inproceedings{zhang2024effect,
  author = {Hongchen Zhang and Zhenhua Dong and Wei Hao and Guohua Cao and Qi Dong and Yiyang Chen and Shihui Li and Lanlan Chen and Zhongqi Fan and Guoyue Lv},
  title = {{Effect of genetically proxied use of PDE5 inhibitors on risk of CHD: A drug target Mendelian Randomization study}},
  booktitle = {Research Square},
  year = {2024},
  doi = {10.21203/rs.3.rs-5236402/v1},
}
Figure