Sex hormone-binding globulin exerts sex-related causal effects on lower extremity varicose veins: evidence from gender-stratified Mendelian randomization

Abstract

Background: The association between serum sex hormones and lower extremity varicose veins has been reported in observational studies. However, it is unclear whether the association reflects a causal relationship. Besides, serum sex hormone-binding globulin (SHBG) has been rarely studied in lower extremity varicose veins. Here, we aim to investigate the association between serum levels of SHBG, testosterone, and estradiol and the risk of lower extremity varicose veins using Mendelian randomization (MR).

Methods: We obtained genome-wide association study summary statistics for serum SHBG levels with 369,002 European participants, serum testosterone levels with 424,907 European participants, serum estradiol levels with 361,194 European participants, and lower extremity varicose veins with 207,055 European participants. First, a univariable MR was performed to identify the causality from SHBG and sex hormone levels to lower extremity varicose veins with several sensitivity analyses being performed. Then, a multivariable MR (MVMR) was performed to further assess whether the causal effects were independent. Finally, we performed a gender-stratified MR to understand the role of genders on lower extremity varicose veins.

Results: Genetically predicted higher serum SHBG levels significantly increased the risk of lower extremity varicose veins in the univariable MR analysis (OR=1.39; 95% CI: 1.13-1.70; P=1.58×10^-3). Sensitivity analyses and MVMR (OR=1.50; 95% CI:1.13-1.99; P=5.61×10^-3) verified the robustness of the causal relationships. Gender-stratified MR revealed that higher serum SHBG levels were associated with lower extremity varicose veins in both sexes. However, the OR of serum SHBG levels on lower extremity varicose veins risk in females (OR=1.51; 95% CI: 1.23-1.87; P=1.00×10^-4) was greater than in males (OR=1.26; 95% CI: 1.04-1.54; P=1.86×10^-2).

Conclusions: Serum SHBG levels are positively related to lower extremity varicose veins risk in both sexes, especially in females. This may partly explain the higher prevalence of varicose vines among females.

Keywords: gender-stratified Mendelian randomization; lower extremity varicose veins; sex difference; sex hormone; sex hormone-binding globulin.

Figure 1

The overview of the study design. (A) Flow chart of the univariable MR, MVMR, and gender-stratified MR. First of all, we performed a univariable MR analysis to explore whether genetically predicted serum levels of SHBG, testosterone, and estradiol (exposures) were associated with the risk of lower extremity varicose veins (outcome). Then, for exposures that survived the univariable MR, we conducted a multivariable MR analysis to estimate the independent causal effect of these exposures on lower extremity varicose veins. Finally, for exposures that survived the above screening criteria, we further perform a gender-stratified MR analysis to understand the roles of genders on lower extremity varicose veins using gender-stratified GWAS datasets. (B) The schematic chart of the three basic principles of MR: (I) IVs are associated with the exposure; (II) IVs are independent of potential confounders; (III) IVs are associated with the outcome through the studied exposure only. SHBG, sex hormone-binding globulin; MR, Mendelian randomization; MVMR, multivariable Mendelian randomization; IVs, instrumental variables.

Figure 2

Results of the univariable MR analysis. MR, Mendelian randomization; SHBG, sex hormone-binding globulin; N. IVs, the number of IVs; WM, weighted median; RE-IVW, random-effects inverse variance weighted. *: P of statistical significance (<0.017) after Bonferroni correction.

Figure 3

Results of the MVMR analysis. MVMR, multivariable Mendelian randomization; SHBG, sex hormone-binding globulin; N. IVs, the number of IVs. *: P of statistical significance (<0.017) after Bonferroni correction.

Figure 4

Results of the gender-stratified MR analysis. MR, Mendelian randomization; SHBG, sex hormone-binding globulin; N. IVs, the number of IVs. WM, weighted median; RE-IVW, random-effects inverse variance weighted. *: P of statistical significance (<0.005) after Bonferroni correction.