Does cannabis use cause erectile dysfunction?

By Vijay Kumar MalesuReviewed by Susha Cheriyedath, M.Sc.Jun 7 2024

In a recent study published in the International Journal of Impotence Research, a group of researchers investigated the causal relationships between lifetime cannabis use (LCU), CU disorder (CUD), erectile dysfunction (ED), and sex hormone levels using Mendelian randomization (MR) analysis with data from genome-wide association studies (GWAS).

Study: The impact of cannabis use on erectile dysfunction and sex hormones: a Mendelian randomization analysis. Image Credit: Kitreel / Shutterstock

Background 

ED is a prevalent male sexual dysfunction characterized by the inability to maintain an erection, causing psychological and physical distress. The global prevalence of ED ranges from 37.2% to 48.6%, increasing with age, and is projected to affect 322 million men by 2025. ED affects self-confidence and relationships, underscoring the need for effective prevention and management strategies. With the legalization of cannabis, its impact on ED and sex hormones is under scrutiny. Evidence linking CU to ED is mixed, highlighting the need for further research to clarify causal relationships and guide prevention and treatment strategies.

About the study 

The present study performed a secondary analysis using publicly available GWAS data, employing two-sample MR to explore the causal relationship between CU, ED, and sex hormone levels. The instrumental variables (IVs) adhere to MR assumptions: strong correlation with exposure, no confounding influences, and outcome impact determined solely by exposure interaction.

Data sources include the Psychiatric Genomics Consortium and International Cannabis Consortium for CUD and LCU, as well as GWAS meta-analysis data from FinnGen Consortium and United Kingdom (UK) Biobank for ED phenotypes. The study focuses on European demographic data.

Single nucleotide polymorphism (SNP) selection criteria include genome-wide significance thresholds, linkage disequilibrium (LD) clumping, MR-Steiger test, and avoiding proxy SNPs. Primary MR analysis uses the Wald ratio test, inverse-variance-weighted (IVW) method, MR-Egger, and weighted median techniques. Additional methods like debiased IVW, robust adjusted profile score (RAPS), and contamination mixture (ConMix) ensure robustness.

Statistical power is calculated using R² and Burgess's online power calculator, with Bonferroni correction applied. Validation tests include Cochran's Q test for heterogeneity, MR-Egger regression for pleiotropy, MR Pleiotropy Residual Sum and Outlier (MR-PRESSO) test for outliers, and leave-one-out analysis. The MR-Linkage Disequilibrium Adjusted Population(Lap) method corrects biases associated with sample overlap and weak IVs.

Study results 

The MR analysis implemented a stringent selection process for IVs, incorporating between 3 to 12 IVs that accounted for genetic variance ranging from 0.59% to 5.36%. All IVs were confirmed using the MR-Steiger filter to satisfy the third assumption of MR, and each IV's F-statistic exceeded 10, with averages ranging from 192 to 1618, significantly minimizing bias from weak IVs. The rigorous application of MR-PRESSO eliminated outliers to prevent bias due to horizontal pleiotropy. 

In the MR analysis focusing on ED, no causal association was found between CUD and LCU with an increased incidence rate of ED, as confirmed by replication datasets and meta-analyses. Specifically, in the discovery dataset, the IVW analysis did not identify a significant causal relationship between genetically predicted CUD and LCU with the risk of ED (CUD: Odds Ratio (OR)=0.97, 95% Confidence Interval (CI) 0.87–1.10, P-value (P) = 0.66; LCU: OR = 1.13, 95% CI 0.84–1.50, P = 0.42). This finding was further validated by replication datasets and meta-analyses, consistently showing no significant associations across eight analytical methods.

Similarly, the IVW analysis did not reveal any causal association between genetically predicted CUD and LCU with levels of estradiol (E2) (CUD: β = 0.00, 95% CI 0.00–0.01, P = 0.37; LCU: β = 0.00, 95% CI −0.02–0.01, P = 0.62), bioavailable testosterone (BT) (CUD: β = 0.00, 95% CI −0.03–0.02, P = 0.90; LCU: β = 0.02, 95% CI −0.04–0.09, P = 0.46), follicle-stimulating hormone (FSH) (CUD: β = 0.01, 95% CI −0.18–0.20, P = 0.92; LCU: β = 0.01, 95% CI −0.44–0.47, P = 0.95), and luteinizing hormone (LH) (CUD: β = 0.01, 95% CI −0.18–0.21, P = 0.90; LCU: β = 0.13, 95% CI −0.22–0.49, P = 0.46). Other analytical methods provided consistent evidence. 

Sensitivity analyses confirmed the robustness of these MR findings. Cochran's Q test revealed no evidence of heterogeneity, and both MR-PRESSO and MR-Egger tests found no evidence of horizontal pleiotropy. Further analysis using MR-Lap indicated that the causal associations between LCU with the discovery dataset (OR = 1.06, 95% CI 0.91–1.23, P = 0.49) and the UK Biobank dataset (OR = 0.96, 95% CI 0.79–1.16, P = 0.67) were not biased by excessive sample overlap. Additionally, leave-one-out analysis indicated that any single SNP did not drive the results.

Conclusions 

To summarize, the findings do not support a causal association between CUD or LCU with increased ED risk or variations in sex hormones in European populations. Despite no direct causal link, lifestyle differences, mental health issues, and concurrent substance use may explain the association between CU and ED.