Abstract
Objective
Prior observational studies have reported that levels of sex hormones constitute a risk factor for the fracture. The aim of this study was to ascertain whether there is a causal relationship between the levels of sex hormones and the risk of fracture through Mendelian randomization (MR).
Methods
Single‐nucleotide polymorphisms (SNPs) associated with two indicators of sex hormone levels, circulating sex hormone‐binding globulin (SHBG) and bioavailable testosterone levels, as exposures were selected from a large genome‐wide association study (GWAS) from UK Biobank. The summary statistics for 11 different types of fracture as outcomes from the FinnGen consortium. This study employed the two‐sample MR approach. For the main analysis, the inverse‐variance‐weighted (IVW) method was utilized. To assess the heterogeneity of MR results, the IVW method and MR–Egger method were utilized. To evaluate potential pleiotropy, MR–Egger regression was conducted. Additionally, a leave‐one‐SNP‐out test was performed to assess the robustness of MR results to the exclusion of any individual SNP.
Results
The MR analyses demonstrated a conspicuous impact of SHBG on the risk of pathological fracture with osteoporosis (OP). We found that an increase of one standard deviation (SD) in SHBG correspondingly increased the risk of pathological fracture with OP [odds ratio (OR) 2.42, 95% confidence interval (CI), 1.52–3.85; p = 1.93 × 10−4]. The bioavailable testosterone showed the negative casual genetic associations with fractures of foot and forearm. An increase of one SD in the genetically predetermined bioavailable testosterone was associated with a reduction of 37% in the risk of fracture of foot (OR 0.63, 95% Cl 0.49 to 0.81; p = 3.37 × 10−4), as well as a 39% decrease in the risk of fracture of forearm (OR 0.61, 95% Cl 0.50 to 0.76; p = 5.40 × 10−6).
Conclusions
Our study confirms that individuals experiencing elevated SHBG concentrations showed a major causal effect on pathological fracture with OP. High bioavailable testosterone levels play an important role in preventing the fractures of foot and forearm. Although increasing bioavailable testosterone and decreasing SHBG levels had no casual effect on most fractures in the general population, they are likely to have the most clinically relevant effect on certain fracture risk reduction.
Keywords: Fracture, Mendelian Randomization, Sex Hormone‐Binding Globulin, Testosterone
Our analysis utilized a two‐sample Mendelian randomization approach, which leverages genetic variants as instrumental variables, to circumvent potential confounding factors and reverse causation commonly encountered in conventional epidemiologic studies.
It is assumed that the instrumental variables are valid if it meet all of the following criteria:
Strongly associated with sex hormone levels (sex hormone‐binding globulin, bioavailable testosterone).
Not associated with confounding factors.
Risk association with 11 types of fractures were only detected via sex hormone‐binding globulin or bioavailable testosterone.
Introduction
Fractures, one of the prime encumbrances concerning musculoskeletal disorders, are a serious complication of osteoporosis (OP) that can have significant consequences for patients and healthcare systems. 1 Therefore, prevention of fractures stands as a paramount objective in public health. Osteoporotic fractures due to an imbalance in bone remodeling can occur at any bone site, but are most common in the vertebrae in the absence of major trauma, especially for postmenopausal women over the age of 50. 2 When bone resorption surpasses bone formation, there is a resultant decrease in bone mass, thereby increasing the susceptibility to fractures. 3
Sex hormones play a crucial role in bone remodeling. 4 Androgen deficiency is a common secondary cause of osteoporosis in men, leading to decreased bone mineral density (BMD) and increased fracture risk. 5 Recent guidelines have demonstrated testosterone replacement therapy significantly increases BMD in the lumbar spine and hip in men with hypogonadism. 6 Additionally, sex‐hormone binding globulin (SHBG), a glycoprotein found in plasma, exhibits a strong binding affinity for sex steroids such as testosterone. 7 Notably, an inverse relationship between serum SHBG levels and bone mineral density has been observed, with higher SHBG levels serving as an indicator for the occurrence of osteoporotic fractures. 8 Low serum androstenedione/SHBG ratio have been associated with an increased risk of fracture, particularly in postmenopausal women. 9 The fraction of testosterone circulating in the bloodstream that remains unattached to SHBG constitutes bioavailable testosterone, which encompasses free testosterone and testosterone bound to albumin. Bioavailable testosterone is deemed to be the biologically active form of testosterone, capable of penetrating cellular receptors and eliciting their effects. 10 Elderly males who suffer from diminished bioavailable testosterone face a heightened likelihood of nonvertebral fracture. 11 In conclusion, the most robust correlation emerged when SHBG and bioavailable testosterone were considered collectively.
Although the heritability of fracture risk has been suggested to be moderate. 12 Accumulating studies have revealed the crucial causal function of low BMD in the pathophysiology of fracture risk. 13 There is substantial evidence to suggest that sex hormones play a significant role as a major regulator of BMD in bone metabolism. 14 , 15 Throughout the pubertal phase, testosterone collaborates with estradiol to vigorously stimulate the expansion of both periosteal bone layers and trabecular bone, thereby underpinning the surge in adolescent growth and the culmination of peak bone mass. 16 As individuals reach the zenith of their bone mass potential, testosterone's influence endures, upholding BMD and fortitude by tempering the pace of bone remodeling and upholding a harmonious equilibrium between the processes of bone resorption and formation. 17 However, the potential impact of sex hormones on fracture risk remains unclear. A novel research methodology is required to further investigate this issue.
The Mendelian randomization (MR) approach has gained widespread use in recent years as a means of determining the potential causal effects of exposures such as sex hormone levels on clinical outcomes such as fractures. 18 MR uses single nucleotide polymorphisms (SNPs) as instrumental variables to mitigate the impact of confounding factors and reverse causation, which can lead to biases in traditional observational studies. 19 Previous two‐sample MR studies have explored the associations of SHBG and bioavailable testosterone with arthritis, 20 with breast cancer, 21 aneurysmal subarachnoid hemorrhage. 22 However, there no studies have established the role of sex hormone levels in fracture risk adequately.
We used large‐scale genome‐wide association study (GWAS) datasets to estimate the causal association between sex hormone levels and fractures. This study may help to reveal the genetic characteristics and conduct a more comprehensive investigation into the risk of fractures.
Methods
Study Design
To explore the impact of sex hormone levels on the risk of bone fractures, we employed a two‐sample MR methodology. Initially, genetic instruments used to approximate SHBG and bioavailable testosterone were acquired from GWAS summary statistics (sample 1). These instruments were then amalgamated with genetic association effect estimates for the risk of bone fractures obtained from published GWAS results (sample 2).The detailed overview of study design was shown in Figure 1.The characteristic of exposure and outcome was shown in Table 1.
FIGURE 1.
The detailed overview of study design. SHBG, Sex hormone‐binding globulin; SNPs, Single‐nucleotide polymorphisms; IVW, inverse‐variance weighted; MR, Mendelian randomization.
TABLE 1.
The characteristics of the exposure and outcome datasets.
| Phenotype | Used as | Ancestry | GWAS | Sample size | Number of SNPs | SNPs as instrumental variables |
|---|---|---|---|---|---|---|
| SHBG | Exposure | European | Ruth et al. 21 | Males: 180,094; Females: 188,908 | 16,585,865 | 250 |
| Bioavailable testosterone | Exposure | European | Ruth et al. 21 | Males: 178,782; Females: 188,507 | 16,585,744 | 92 |
| Pathological fracture with OP | Outcome | European | FinnGen consortium | Case: 1433; Controls: 261,098 | 20,166,569 | NA |
| Fracture of lumbar spine and pelvis | Outcome | European | FinnGen consortium | Case: 5364; Controls: 331,508 | 20,169,405 | NA |
| Fracture of rib(s), sternum and thoracic spine | Outcome | European | FinnGen consortium | Case: 7845; Controls: 329,501 | 20,169,422 | NA |
| Fracture of neck | Outcome | European | FinnGen consortium | Case: 1288; Controls: 336,380 | 20,169,450 | NA |
| Fracture of skull and facial bones | Outcome | European | FinnGen consortium | Case: 6051; Controls: 300,857 | 20,168,454 | NA |
| Fracture of femur | Outcome | European | FinnGen consortium | Case: 7671; Controls: 329,501 | 20,169,341 | NA |
| Fracture of lower leg including ankle | Outcome | European | FinnGen consortium | Case: 17,690; Controls: 295,290 | 20,169,413 | NA |
| Fracture of foot except ankle | Outcome | European | FinnGen consortium | Case: 6548; Controls: 320,287 | 20,169,155 | NA |
| Fracture of shoulder and upper arm | Outcome | European | FinnGen consortium | Case: 10,273; Controls: 314,458 | 20,169,004 | NA |
| Fracture of forearm | Outcome | European | FinnGen consortium | Case: 17,201; Controls: 319,704 | 20,169,143 | NA |
| Fracture at wrist and hand level | Outcome | European | FinnGen consortium | Case: 9950; Controls: 307,057 | 20,168,817 | NA |
Abbreviations: GWAS, Genome‐wide association study; NA, Not appliable; SHBG, Sex hormone‐binding globulin; SNPs, Single nucleotide polymorphisms.
Genetic Associations with Sex Hormone Levels
We selected genetic instruments associated with SHBG and bioavailable testosterone from the largest GWAS data. SNPs prognosticating the levels of SHBG (male: 180,094; female: 188,908) and bioavailable testosterone (male: 178,872; female: 188,507) were acquired from publicly accessible summary statistics given by Ruth et al., utilizing the UK Biobank data. 23 The threshold for SNP selection was p < 5 × 10−8. To avoid issues with multi‐collinearity, we made the deliberate decision to incorporate independent SNPs in our analysis. This was achieved by performing linkage disequilibrium (LD) clumping with an r2 < 0.001, which resulted in only the SNP that exhibited the strongest association with the hormone within a 10,000 kb window being considered for further analysis. Then we calculated F statistics for the instrumental variables associated with exposure, all SNPs were greater than 10, which was similar to the previous study, implying that the likelihood of bias due to weak instrumental variables as negligible. 21 , 24 , 25 In cases where an instrumental SNP for the exposure was not available in the outcome dataset, we opted to replace it with a suitable proxy SNP (with an R2 value greater than 0.8 in the European 1000 Genomes Project reference panel, using LDlink [https://ldlink.nci.nih.gov/]) or remove it altogether if no such proxy was available. Finally, we identified 250 and 92 SNPs as instrumental variables to proxy for an standard deviation (SD) increase in SHBG and bioavailable testosterone, respectively.
The concentrations of testosterone and SHBG were assessed in units of nmol/L using a one‐step competitive analysis and two‐step sandwich immunoassay analysis. 23 In the analysis, the genotyping chip, age at baseline, mins since blood draw, time of blood draw, menopause, operation status, and 10 principal components derived from genetics were utilized as covariates to adjust for population stratification. Additionally, for SHBG, body mass index (BMI) was incorporated as a covariate. Genetic loci identified from the BMI‐adjusted analyses were incorporated, along with corresponding effect estimates from the BMI‐unadjusted analyses, in order to ameliorate potential collider bias. 26
Genetic Associations with Fracture
The latest large‐scale GWAS and meta‐analyses of bone fractures in European populations were procured from the eighth release of the FinnGen consortium. There were 1433 cases of pathological fracture with OP (261,098 controls), 5364 cases of fracture of lumbar spine and pelvis (331,508 controls), 7845 cases of fracture of rib(s), sternum, and thoracic spine (329,501 controls), 1288 cases of fracture of neck (336,380 controls), 6051 cases of fracture of skull and facial bones (300,857 controls), 7671 cases of fracture of femur (329,501 controls), 17,690 cases of fracture of lower leg including ankle (295,290 controls), 6548 cases of fracture of foot except ankle (320,287 controls), 10,273 cases of fracture of shoulder and upper arm (314,458 controls), 17,201 cases of fracture of forearm (319,704 controls), and 9950 cases of fracture at wrist and hand level (307,057 controls). The entirety of the 11 subcategories of bone fractures were classified under the code ST19 in the Tenth Revision of the International Classification of Diseases (ICD‐10). Comprehensive data pertaining to the participants, genotyping, imputation, and quality control procedures can be accessed through the FinnGen website (https://www.fifinngen.fifi/en). All studies contributing data for analyses were approved by relevant ethics committees. Moreover, all participants provided written informed consent prior to the study. The SNPs related to all kinds of fracture when exposure is SHBG, and bioavailable testosterone were respectively shown in Tables S1 and S2.
Statistical Analysis
The two‐sample MR approach was used to identify the potential causal link of SHBG and bioavailable testosterone concentrations with fracture by applying the inverse‐variance weighted (IVW) analysis. 27 Moreover, the MR–Egger and weighted median methods were calculated to further identify the potential causal association between exposures and outcomes.
The heterogeneity of MR analysis results was evaluated through IVW and MR–Egger methods. IVW analysis was performed using Cochran's Q statistic, while MR–Egger analysis was performed using Rucker's Q statistic. 28 MR–Egger regression was also used to detect the pleiotropy of MR results, with an intercept p value greater than 0.05 indicating pleiotropy deficiency. 29 Furthermore, a leave‐one‐SNP out analysis was conducted to explore the possibility that a single SNP was driving the causal association between exposures and outcomes. 30
For MR analyses, upon adjustment for multiple comparisons, the level of statistical significance in this investigation was set at p < 4.17 × 10−3, derived from dividing 0.05 by the 22 tests. A p < 0.05, but above the threshold of Bonferroni correction significance, was considered a suggestive causal association. All statistical analyses were conducted using R version 4.1.0 (Lucent Technologies; New Jersey, USA) using the “TwoSampleMR” package (version 0.4.26).
Results
Causal Association between SHBG with Fracture
In the IVW models, as shown in Figure 2, we found that a SD increase in SHBG increased the risk of pathological fracture with OP [(OR) 2.42, 95% confidence interval (CI) 1.52–3.85), p = 1.93 × 10−4]. We observed a potential causal association between genetically predicted circulating SHBG concentrations and risk of fracture of femur (OR 1.28, 95% CI 1.05–1.57, p = 0.01), fracture of low leg including ankle (OR 1.16, 95% CI 1.00–1.33, p = 0.04) and fracture of forearm (OR 1.23, 95% CI 1.06–1.42, p = 7.13 × 10−3). There was no causal association between SHBG levels and risk of the other types of fractures, as shown in Figure 2. We found potential evidence of association between SHBG and the risk of fracture of femur (OR 1.50, 95% CI 1.11–2.02, p = 0.01) and forearm (OR 1.27, 95% CI 1.02–1.59, p = 0.03). Further potential association between SHBG and the risk of fracture of femur was found using the weighted median approach (OR 1.44, 95% CI 1.03–2.00, p = 0.03) (Table 2). The scatter plot was shown in Figures S1–S3.
FIGURE 2.
Forest plot of the MR results between SHBG levels and 11 types of fractures based on the IVW method. MR, Mendelian randomization; SHBG, Sex hormone‐binding globulin; IVW, Inverse‐variance weighted; OR, Odds ratio; CI, Confidence interval.
TABLE 2.
The MR analysis of sex hormones and 11 fractures subtypes using the MR–Egger and weighted median methods.
| Exposure | Outcomes | SNP (n) | MR–Egger | Weighted median | ||
|---|---|---|---|---|---|---|
| OR (95%Cl) | p value | OR (95%Cl) | p value | |||
| SHBG | Pathological fracture with osteoporosis | 250 | 1.85 (0.92, 3.69) | 0.08 | 2.03 (0.93, 4.45) | 0.07 |
| Fracture of lumbar spine and pelvis | 1.37 (0.94, 2.01) | 0.10 | 0.93 (0.65, 1.35) | 0.73 | ||
| Fracture of ribs, sternum and thoracic spine | 1.12 (0.85, 1.48) | 0.42 | 1.06 (0.77, 1.46) | 0.71 | ||
| Fracture of neck | 1.49 (0.77, 2.89) | 0.24 | 1.31 (0.62, 2.79) | 0.49 | ||
| Fracture of skull and facial bones | 0.87 (0.62, 1.22) | 0.41 | 0.81 (0.56, 1.18) | 0.28 | ||
| Fracture of femur | 1.50 (1.11, 2.02) | 0.01 | 1.44 (1.03, 2.00) | 0.03 | ||
| Fracture of lower leg, including ankle | 1.10 (0.89, 1.36) | 0.37 | 1.04 (0.84, 1.29) | 0.72 | ||
| Fracture of foot, except ankle | 0.99 (0.73, 1.35) | 0.95 | 0.79 (0.57, 1.09) | 0.17 | ||
| Fracture of shoulder and upper arm | 1.06 (0.81, 1.40) | 0.67 | 1.00 (0.74, 1.34) | 0.99 | ||
| Fracture of forearm | 1.27 (1.02, 1.59) | 0.03 | 1.06 (0.85, 1.33) | 0.60 | ||
| Fracture at wrist and hand level | 1.06 (0.82, 1.38) | 0.65 | 0.88 (0.67, 1.16) | 0.37 | ||
| BT | Pathological fracture with osteoporosis | 92 | 0.29 (0.09, 0.89) | 0.03 | 0.38 (0.16, 0.91) | 0.03 |
| Fracture of lumbar spine and pelvis | 0.73 (0.38, 1.42) | 0.36 | 0.94 (0.60, 1.48) | 0.79 | ||
| Fracture of ribs, sternum and thoracic spine | 0.83 (0.50, 1.37) | 0.47 | 0.83 (0.59, 1.18) | 0.31 | ||
| Fracture of neck | 0.49 (0.16, 1.52) | 0.22 | 0.79 (0.32, 1.97) | 0.59 | ||
| Fracture of skull and facial bones | 1.13 (0.65, 1.97) | 0.67 | 1.17 (0.75, 1.82) | 0.52 | ||
| Fracture of femur | 0.83 (0.50, 1.37) | 0.47 | 0.83 (0.58, 1.20) | 0.31 | ||
| Fracture of lower leg, including ankle | 1.16 (0.83, 1.62) | 0.39 | 0.91 (0.70, 1.18) | 0.46 | ||
| Fracture of foot, except ankle | 0.77 (0.46, 1.29) | 0.32 | 0.72 (0.49, 1.06) | 0.09 | ||
| Fracture of shoulder and upper arm | 0.73 (0.44, 1.21) | 0.23 | 0.83 (0.59, 1.16) | 0.29 | ||
| Fracture of forearm | 0.47 (0.31, 0.73) | 1.09E‐03 | 0.52 (0.40, 0.68) | 9.03E‐07 | ||
| Fracture at wrist and hand level | 0.76 (0.46, 1.25) | 0.28 | 0.79 (0.55, 1.13) | 0.21 |
Abbreviation: MR, Mendelian randomization.
Causal Association between Bioavailable Testosteronewith Fracture
Using an IVW approach, we found that a one standard deviation increase in bioavailable testosterone was associated with a decreased risk of fracture of foot except ankle and forearm (OR 0.63, 95% CI 0.49–0.81, p = 3.37 × 10−4 and OR 0.61, 95% CI 0.50–0.76, p = 5.40 × 10−6 respectively), as shown in Figure 3. With regards to pathological fracture with OP, we only found suggestive evidence of associations with bioavailable testosterone (OR 0.46, 95% CI 0.26–0.81, p = 7.13 × 10−3). We found that an SD increase in bioavailable testosterone decreased the risk of fracture of femur (OR 0.47, 95% CI 0.31–0.73, p = 1.09 × 10−3) using the MR–Egger approach. Further negative association was also found using the weighted median approach (OR 0.52, 95% CI 0.40–0.68, p = 9.03 × 10−7). We found potential evidence of association between bioavailable testosterone and the risk of pathological fracture with OP (OR 0.29, 95% CI 0.09–0.89, p = 0.03) using the MR–Egger approach (Table 2). Further potential association between SHBG and the risk of pathological fracture with OP was found using the weighted median approach (OR 0.38, 95% CI 0.16–0.91, p = 0.03) (Table 1). The scatter plot was shown in Figures S4–S6.
FIGURE 3.
Forest plot of the MR results between bioavailable testosterone levels and 11 types of fractures based on the IVW method. MR, Mendelian randomization; IVW, Inverse‐variance weighted; OR, Odds ratio; CI, Confidence interval.
Sensitivity Analysis
The heterogeneity analysis to verify the reliability of IVW results was shown in Table 2. Using the IVW test, we observed no heterogeneity in the MR analysis outcomes pertaining to the association between SHBG levels and the risk of pathological fracture with OP, fractures in various anatomical sites, ribs, sternum and thoracic spine, neck, femur, foot except ankle, and at wrist and hand level. There was heterogeneity in MR analysis results between SHBG and the risk of fracture of lumbar spine and pelvis, skull and facial bones, lower leg including ankle, shoulder and upper arm, and forearm. In the MR–Egger test, we noticed no heterogeneity in the MR analysis outcomes pertaining to the association between bioavailable testosterone levels and the risk of pathological fracture with OP, fractures in ribs, sternum and thoracic spine, neck, skull and facial bones, femur, lower leg including ankle, foot except ankle. There was heterogeneity in MR analysis results between SHBG and the risk of fracture of lumbar spine and pelvis, shoulder and upper arm, forearm, and at wrist and hand level. The funnel plot was shown in Figures S7–S12.
In addition, we performed an MR–Egger intercept test to assess the presence of directional pleiotropy for overall types of fracture risk, but we did not detect any evidence of it (p > 0.05) in Table 3. Using the leave‐one‐out method for both the initial and verified analyses, we found that none of the independent variables had an abnormal influence on the overall results in in Figures S13–S21.
TABLE 3.
The heterogeneity and horizontal pleiotropy tests in the MR analysis.
| Exposure | Outcome | Heterogeneity test | Horizontal pleiotropy test | ||||
|---|---|---|---|---|---|---|---|
| MR–Egger | IVW | ||||||
| Cochran's Q | p value | Cochran's Q | p value | Egger‐intercept | p value | ||
| SHBG | Pathological fracture with osteoporosis | 227.32 | 0.05 | 228.57 | 0.05 | 5.68E‐03 | 0.30 |
| Fracture of lumbar spine and pelvis | 258.76 | 1.10E‐03 | 261.32 | 9.05E‐04 | −4.17E‐03 | 0.17 | |
| Fracture of ribs, sternum and thoracic spine | 196.37 | 0.40 | 196.43 | 0.42 | −5.36E‐04 | 0.81 | |
| Fracture of neck | 184.06 | 0.63 | 184.15 | 0.65 | −1.57E‐03 | 0.77 | |
| Fracture of skull and facial bones | 234.90 | 0.02 | 235.21 | 0.02 | 1.36E‐03 | 0.62 | |
| Fracture of femur | 219.85 | 0.08 | 222.01 | 0.07 | −3.27E‐03 | 0.17 | |
| Fracture of lower leg, including ankle | 251.94 | 2.39E‐03 | 252.44 | 2.60E‐03 | 1.04E‐03 | 0.54 | |
| Fracture of foot, except ankle | 208.92 | 0.19 | 210.11 | 0.19 | 2.59E‐03 | 0.30 | |
| Fracture of shoulder and upper arm | 251.79 | 2.44E‐03 | 252.46 | 2.60E‐03 | 1.55E‐03 | 0.48 | |
| Fracture of forearm | 269.55 | 1.56E‐04 | 269.84 | 1.80E‐04 | −8.09E‐04 | 0.65 | |
| Fracture at wrist and hand level | 220.52 | 0.07 | 220.77 | 0.08 | −9.68E‐04 | 0.64 | |
| BT | Pathological fracture with osteoporosis | 61.40 | 0.32 | 62.38 | 0.32 | 1.21E‐02 | 0.34 |
| Fracture of lumbar spine and pelvis | 79.09 | 0.03 | 79.13 | 0.03 | 1.16E‐03 | 0.88 | |
| Fracture of ribs, sternum and thoracic spine | 65.80 | 0.20 | 66.04 | 0.22 | 2.56E‐03 | 0.65 | |
| Fracture of neck | 54.12 | 0.58 | 56.95 | 0.51 | 2.14E‐02 | 0.10 | |
| Fracture of skull and facial bones | 63.83 | 0.25 | 63.96 | 0.28 | 2.12E‐03 | 0.74 | |
| Fracture of femur | 65.80 | 0.20 | 66.04 | 0.22 | 2.56E‐03 | 0.65 | |
| Fracture of lower leg, including ankle | 64.03 | 0.24 | 67.67 | 0.18 | −6.73E‐03 | 0.08 | |
| Fracture of foot, except ankle | 58.98 | 0.40 | 59.78 | 0.41 | −5.10E‐03 | 0.38 | |
| Fracture of shoulder and upper arm | 86.36 | 7.29E‐03 | 86.76 | 8.56E‐03 | 2.90E‐03 | 0.61 | |
| Fracture of forearm | 101.29 | 2.77E‐04 | 104.53 | 1.75E‐04 | 6.42E‐03 | 0.18 | |
| Fracture at wrist and hand level | 85.58 | 8.48E‐03 | 86.03 | 9.85E‐03 | 3.10E‐03 | 0.59 | |
Abbreviation: MR, Mendelian randomization.
Discussion
Our findings indicate that an increase of one SD in SHBG was concomitantly associated with a 1.42‐fold elevation in the risk of pathological fracture with OP, whereas elevating the genetically predetermined bioavailable testosterone by one SD was linked to a 37% decline in the susceptibility to foot fractures, along with a 39% decrease in the likelihood of forearm fractures.
SHBG and Fractures
A prior MR study indicated strong associations between SHBG and various diseases, particularly fracture (primarily in females). 31 In an investigation of a group of elderly men over time, researchers observed that elevated SHBG levels persisted as a robust and autonomous risk factor for the risk of hip fracture, exhibiting an OR of 1.76. 32 Another observational study also suggested that a progressive elevation in serum SHBG over time retained its association with both overall fracture incidence (β = 0.060) and hip fracture incidence (β = 0.041), while no such relationship was observed with nonvertebral fracture incidence. 33 These results were consistent with our MR results. An elevation of one SD in SHBG was concurrently linked to a 1.42‐fold increase in the susceptibility to pathological fracture with OP, which most commonly occurs in the hip. 34 SHBG exhibited a positive correlation with the likelihood of hip fracture. 35 Contrary to our expectations, we did not observe any significant association between SHBG and the risk of pelvis fracture.
A possible reason for this is that elevated SHBG levels diminish the amount of unbound testosterone, thus weakening the protective influence of testosterone on bone tissue. Although a distinctive gender‐specific pattern in SHBG levels with age was observed, with men exhibiting a linear increase and women showing a U‐shaped trend, the SHBG level among elderly individuals was all high. 36 A recent comprehensive review also revealed elevated SHBG concentrations indicated a higher likelihood of bone fractures in elderly patients. 37 Future investigations should clarify the pathways through which SHBG influences the activity of osteoblast and osteoclast, thereby affects bone mass and skeletal health.
Bioavailable Testosterone and Fractures
Our discoveries indicate that elevated levels of bioavailable testosterone are causally linked with a diminished risk of fractures of forearm and foot except ankle. Also, we encountered potential corroboration of a causal link between bioavailable testosterone levels and the risk of pathological fracture with OP. This corroborates the idea that bioavailable testosterone has a beneficial effect on bone health. In observational studies about the effects of testosterone supplementation on frail elderly men, testosterone levels show a weak but consistent correlation with BMD and bone quality. 38 Elderly and weak men who underwent testosterone replacement therapy enhanced their testosterone levels and BMD exhibited a 1.4% augmentation at thefemoral neck and a 3.2% enhancement at the lumbar spine (p = 0.005). 39 This risk is amplified when bioavailable testosterone levels are low in the presence of high SHBG levels.
The potential reason was that testosterone can foster the differentiation and function of osteoblasts, and impede the activation and proliferation of osteoclasts. 40 In fact, the interplay of testosterone and fractures entails a multifaceted issue that encompasses various factors, such as age, sex, BMD, and OP. Ordinarily, diminished levels of testosterone exhibited the highest susceptibility to hip fracture (adjusted hazard ratio: 1.8, 95% CI, 0.8–3.8) in elderly men. 41 Testosterone deficiency in females correlates with diminished BMD and heightened susceptibility to OP. 42 The incidence of fractures was lower in Danish women with polycystic ovary syndrome (10.3/1000 patient years) who had elevated levels of testosterone compared to controls (13.6/1000 patient years), the adjusted ORs for all fractures were 0.76 (95% CI, 0.71 to 0.80), for major osteoporotic fractures they were 0.82 (95% CI, 0.74 to 0.92), and for craniofacial fractures they were 0.57 (95% CI, 0.47 to 0.70). 43
Strengths and Limitations
To the best of our knowledge, this is the first MR study investigating the potential causal relationships between sex hormones and the risk of all types of fractures. MR studies have several advantages over traditional observational studies. One of the key strengths of our study design is that it reduces the potential for confounding factors, as alleles are assigned randomly at conception and are generally not influenced by confounders. 44 Furthermore, since allele assignment occurs before the outcome (fracture), the risk of bias due to reverse causation is minimized. 13 These findings provide valuable insights for designing future clinical trials aimed at reducing fracture risk.
In addition, we utilized recently published comprehensive GWAS data and conducted multiple sensitivity analyses to ascertain the stability and reliability of our results.
There are some limitations in our study. Firstly, some of the relevant exposure SNPs are not available in the outcome GWAS dataset, even after searching for potential proxy SNPs. As a result, a significant number of exposure SNPs could not be included in our MR analyses. Although the lack of data for some relevant exposure SNPs in the outcome GWAS reduce our statistical power to detect small effects, we are still able to include a reasonable number of SNPs and conduct robust MR analyses. Secondarily, the MR results obtained using the IVW method were inconsistent with those obtained from the MR–Egger and weighted median analyses. This discrepancy could be attributed to potential differences in the validity of all the SNPs used in the different methods. Another limitation was that our study exclusively included individuals of European ancestry. Further studies incorporating individuals from different races are necessary to achieve more conclusive results. In addition, confounding factors such as age, gender, and other environmental factors can potentially influence the results of MR analyses. Lastly, it's worth noting that MR estimates the effects over a lifetime, rather than acute effects. Lifelong exposure usually has a greater impact on an outcome compared to short‐term exposure, since the cumulative effects of most exposures on the associated outcome accumulate over time. 45 Despite the limitations, this study offers new and valuable insights into the genetic causal relationship between iron homeostasis and OA, from a genetic standpoint. Thus, it can serve as a valuable reference for future investigations in this field.
Conclusion
To conclude, our two‐sample MR study provides compelling evidence of a causal relationship between elevated SHBG levels and a reduced risk of pathological fracture with OP. Conversely, lower levels of bioavailable testosterone were found to be associated with an increased risk of forearm and foot fractures except for the ankle. These findings may have implications for the prevention and treatment of fracture, particularly in elderly individuals who are at increased risk of bone loss due to elevated SHBG levels and decreased bioavailable testosterone levels. Further research is needed to confirm our findings and to explore the potential mechanisms underlying the relationship between sex hormone levels and fracture risk.
Author Contributions
Conception and design: KS and YM. Administrative support: BS. Provision of study materials: YW, YZ, and JX. Collection and assembly of data: ML, LW. Data analysis and interpretation: KS, YM, YW, and YZ. Manuscript writing: KS and YM. All authors contributed to the article and approved the submitted version.
Conflict of Interest Statement
The authors declare that they have no competing interests.
Authorship Declaration
All authors listed meet the authorship criteria according to the latest guidelines of the International Committee of Medical Journal Editors. All authors are in agreement with the manuscript.
Disclosure
All authors declared no financial support or relationships that may pose a conflict of interest.
Consent for Publication
All authors provided consent for publication.
Ethics Approval and Consent to Participate
The source of the data was a publicly available database, and no human participants were involved; hence, ethical parameters are not applicable.
Supporting information
Figure S1. The scatter plot of SNPs associated with SHBG and their risk of fractures. (A) Pathological fracture with OP; (B) Fracture of lumbar spine and pelvis; (C) Fracture of ribs, sternum, and thoracic spine; (D) Fracture of neck. SHBG: Sex hormone‐binding globulin; OP: Osteoporosis.
Figure S2. The scatter plot of SNPs associated with SHBG and their risk of fractures. (A) Fracture of skull and facial bones; (B) Fracture of femur; (C) Fracture of lower leg including ankle; (D) Fracture of foot except ankle. SHBG: Sex hormone‐binding globulin.
Figure S3. The scatter plot of SNPs associated with SHBG and their risk of fractures. (A) Fracture of shoulder and upper arm; (B) Fracture of forearm; (C) Fracture at wrist and hand level. SHBG: Sex hormone‐binding globulin.
Figure S4. The scatter plot of SNPs associated with bioavailable testosterone and their risk of fractures. (A) Pathological fracture with OP; (B) Fracture of lumbar spine and pelvis; (C) Fracture of ribs, sternum, and thoracic spine; (D) Fracture of neck. OP: osteoporosis.
Figure S5. The scatter plot of SNPs associated with bioavailable testosterone and their risk of fractures. (A) Fracture of skull and facial bones; (B) Fracture of femur; (C) Fracture of lower leg including ankle; (D) Fracture of foot except ankle.
Figure S6. The scatter plot of SNPs associated with bioavailable testosterone and their risk of fractures. (A) Fracture of shoulder and upper arm; (B) Fracture of forearm; (C) Fracture at wrist and hand level.
Figure S7. The funnel plot of SNPs associated with SHBG and their risk of fractures. (A) Pathological fracture with OP; (B) Fracture of lumbar spine and pelvis; (C) Fracture of ribs, sternum, and thoracic spine; (D) Fracture of neck. SHBG: Sex hormone‐binding globulin; OP: Osteoporosis.
Figure S8. The funnel plot of SNPs associated with SHBG and their risk of fractures. (A) Fracture of skull and facial bones; (B) Fracture of femur; (C) Fracture of lower leg including ankle; (D) Fracture of foot except ankle. SHBG: Sex hormone‐binding globulin.
Figure S9. The funnel plot of SNPs associated with SHBG and their risk of fractures. (A) Fracture of shoulder and upper arm; (B) Fracture of forearm; (C) Fracture at wrist and hand level. SHBG: Sex hormone‐binding globulin.
Figure S10. The funnel plot of SNPs associated with bioavailable testosterone and their risk of fractures. (A) Pathological fracture with OP; (B) Fracture of lumbar spine and pelvis; (C) Fracture of ribs, sternum, and thoracic spine; (D) Fracture of neck. OP: osteoporosis.
Figure S11. The funnel plot of SNPs associated with bioavailable testosterone and their risk of fractures. (A) Fracture of skull and facial bones; (B) Fracture of femur; (C) Fracture of lower leg including ankle; (D) Fracture of foot except ankle.
Figure S12. The funnel plot of SNPs associated with bioavailable testosterone and their risk of fractures. (A) Fracture of shoulder and upper arm; (B) Fracture of forearm; (C) Fracture at wrist and hand level.
Figure S13. The leave‐one‐out test plot of the causal association between SHBG and the risk of fractures. (A) Pathological fracture with OP; (B) Fracture of lumbar spine and pelvis. SHBG: Sex hormone‐binding globulin; OP: Osteoporosis.
Figure S14. The leave‐one‐out test plot of the causal association between SHBG and the risk of fractures. (A) Fracture of ribs, sternum, and thoracic spine; (B) Fracture of neck. SHBG: Sex hormone‐binding globulin.
Figure S15. The leave‐one‐out test plot of the causal association between SHBG and the risk of fractures. (A) Fracture of skull and facial bones; (B) Fracture of femur. SHBG: Sex hormone‐binding globulin.
Figure S16. The leave‐one‐out test plot of the causal association between SHBG and the risk of fractures. (A) Fracture of lower leg including ankle; (B) Fracture of foot except ankle. SHBG: Sex hormone‐binding globulin.
Figure S17. The leave‐one‐out test plot of the causal association between SHBG and the risk of fractures. (A) Fracture of shoulder and upper arm; (B) Fracture of forearm. SHBG: Sex hormone‐binding globulin.
Figure S18. The leave‐one‐out test plot of the causal association between SHBG and the risk of fracture at wrist and hand level. SHBG: Sex hormone‐binding globulin.
Figure S19. The leave‐one‐out test plot of the causal association between bioavailable testosterone and the risk of fractures. (A) Pathological fracture with OP; (B) Fracture of lumbar spine and pelvis (C) Fracture of ribs, sternum and thoracic spine (D) Fracture of neck. OP: Osteoporosis.
Figure S20. The leave‐one‐out test plot of the causal association between bioavailable testosterone and the risk of fractures. (A) Fracture of skull and facial bones; (B) Fracture of femur; (C) Fracture of lower leg including ankle; (D) Fracture of foot except ankle.
Figure S21. The funnel plot of SNPs associated with bioavailable testosterone and their risk of fractures. (A) Fracture of shoulder and upper arm; (B) Fracture of forearm; (C) Fracture at wrist and hand levels.
Table S1. The SNPs related to the outcomes when exposure is SHBG.
Table S2. The SNPs related to the outcomes when exposure is bioavailable testosterone.
Acknowledgments
This study was funded by National Natural Science Foundation of China (No. 81974347), China Postdoctoral Science Foundation (No. 2021M702351), Medical Science and Technology Project of Health Commission of Sichuan Provincial (No. 21PJ040), and Science and Technology Department of Sichuan Province (No. 2023YFS0096).
Kaibo Sun and Yue Ming contributed equally to this paper as first authors.
Data Availability Statement
The original contributions presented in the study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding author.
References
- 1. Johnston CB, Dagar M. Osteoporosis in older adults. Med Clin North Am. 2020;104(5):873–884. [DOI] [PubMed] [Google Scholar]
- 2. Anthamatten A, Parish A. Clinical update on osteoporosis. J Midwifery Womens Health. 2019;64(3):265–275. [DOI] [PubMed] [Google Scholar]
- 3. Hadjidakis DJ, Androulakis II. Bone remodeling. Ann N Y Acad Sci. 2006;1092:385–396. [DOI] [PubMed] [Google Scholar]
- 4. Laurent MR, Dedeyne L, Dupont J, Mellaerts B, Dejaeger M, Gielen E. Age‐related bone loss and sarcopenia in men. Maturitas. 2019;122:51–56. [DOI] [PubMed] [Google Scholar]
- 5. Ebeling PR. Androgens and osteoporosis. Curr Opin Endocrinol Diabetes Obes. 2010;17(3):284–292. [DOI] [PubMed] [Google Scholar]
- 6. Minhas S, Bettocchi C, Boeri L, Capogrosso P, Carvalho J, Cilesiz NC, et al. European Association of Urology guidelines on male sexual and reproductive health: 2021 update on male infertility. Eur Urol. 2021;80(5):603–620. [DOI] [PubMed] [Google Scholar]
- 7. Schneider HP. Androgens and antiandrogens. Ann N Y Acad Sci. 2003;997:292–306. [DOI] [PubMed] [Google Scholar]
- 8. Hoppé E, Bouvard B, Royer M, Audran M, Legrand E. Sex hormone‐binding globulin in osteoporosis. Joint Bone Spine. 2010;77(4):306–312. [DOI] [PubMed] [Google Scholar]
- 9. Moberg L, Nilsson PM, Samsioe G, Borgfeldt C. Low androstenedione/sex hormone binding globulin ratio increases fracture risk in postmenopausal women. The Women's health in the Lund area study. Maturitas. 2013;75(3):270–275. [DOI] [PubMed] [Google Scholar]
- 10. Shea JL, Wong PY, Chen Y. Free testosterone: clinical utility and important analytical aspects of measurement. Adv Clin Chem. 2014;63:59–84. [DOI] [PubMed] [Google Scholar]
- 11. LeBlanc ES, Nielson CM, Marshall LM, Lapidus JA, Barrett‐Connor E, Ensrud KE, et al. The effects of serum testosterone, estradiol, and sex hormone binding globulin levels on fracture risk in older men. J Clin Endocrinol Metab. 2009;94(9):3337–3346. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Andrew T, Antioniades L, Scurrah KJ, MacGregor AJ, Spector TD. Risk of wrist fracture in women is heritable and is influenced by genes that are largely independent of those influencing BMD. J Bone Miner Res. 2005;20(1):67–74. [DOI] [PubMed] [Google Scholar]
- 13. Trajanoska K, Morris JA, Oei L, Zheng HF, Evans DM, Kiel DP, et al. Assessment of the genetic and clinical determinants of fracture risk: genome wide association and mendelian randomisation study. BMJ. 2018;362:k3225. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Kenkre JS, Bassett J. The bone remodelling cycle. Ann Clin Biochem. 2018;55(3):308–327. [DOI] [PubMed] [Google Scholar]
- 15. Cauley JA. Estrogen and bone health in men and women. Steroids. 2015;99:11–15. [DOI] [PubMed] [Google Scholar]
- 16. Almeida M, Laurent MR, Dubois V, Claessens F, O'Brien CA, Bouillon R, et al. Estrogens and androgens in skeletal physiology and pathophysiology. Physiol Rev. 2017;97(1):135–187. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Porcelli T, Maffezzoni F, Pezzaioli LC, Delbarba A, Cappelli C, Ferlin A. Management of endocrine disease: male osteoporosis: diagnosis and management‐should the treatment and the target be the same as for female osteoporosis? Eur J Endocrinol. 2020;183(3):R75–r93. [DOI] [PubMed] [Google Scholar]
- 18. Emdin CA, Khera AV, Kathiresan S. Mendelian randomization. Jama. 2017;318(19):1925–1926. [DOI] [PubMed] [Google Scholar]
- 19. Sekula P, del Greco M F, Pattaro C, Köttgen A. Mendelian randomization as an approach to assess causality using observational data. J Am Soc Nephrol. 2016;27(11):3253–3265. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Qu Z, Huang J, Yang F, Hong J, Wang W, Yan S. Sex hormone‐binding globulin and arthritis: a mendelian randomization study. Arthritis Res Ther. 2020;22(1):118. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Nounu A, Kar SP, Relton CL, Richmond RC. Sex steroid hormones and risk of breast cancer: a two‐sample mendelian randomization study. Breast Cancer Res. 2022;24(1):66. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Molenberg R, Thio CHL, Aalbers MW, Uyttenboogaart M, ISGC Intracranial Aneurysm Working Group* , Larsson SC, et al. Sex hormones and risk of aneurysmal subarachnoid hemorrhage: a mendelian randomization study. Stroke. 2022;53(9):2870–2875. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Ruth KS et al. Using human genetics to understand the disease impacts of testosterone in men and women. Nat Med. 2020;26(2):252–258. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Zheng J, Baird D, Borges MC, Bowden J, Hemani G, Haycock P, et al. Recent developments in mendelian randomization studies. Curr Epidemiol Rep. 2017;4(4):330–345. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Burgess S, Thompson SG. Avoiding bias from weak instruments in mendelian randomization studies. Int J Epidemiol. 2011;40(3):755–764. [DOI] [PubMed] [Google Scholar]
- 26. Day FR, Loh PR, Scott RA, Ong KK, Perry JRB. A robust example of collider bias in a genetic association study. Am J Hum Genet. 2016;98(2):392–393. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Lawlor DA. Commentary: two‐sample mendelian randomization: opportunities and challenges. Int J Epidemiol. 2016;45(3):908–915. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Linden AB, Clarke R, Hammami I, Hopewell JC, Guo Y, Whiteley WN, et al. Genetic associations of adult height with risk of cardioembolic and other subtypes of ischemic stroke: a mendelian randomization study in multiple ancestries. PLoS Med. 2022;19(4):e1003967. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Burgess S, Thompson SG. Interpreting findings from mendelian randomization using the MR‐Egger method. Eur J Epidemiol. 2017;32(5):377–389. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Bowden J, Holmes MV. Meta‐analysis and mendelian randomization: a review. Res Synth Methods. 2019;10(4):486–496. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Yuan S, Wang L, Sun J, Yu L, Zhou X, Yang J, et al. Genetically predicted sex hormone levels and health outcomes: phenome‐wide mendelian randomization investigation. Int J Epidemiol. 2022;51(6):1931–1942. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Lee JS, LaCroix AZ, Wu LL, Cauley JA, Jackson RD, Kooperberg C, et al. Associations of serum sex hormone‐binding globulin and sex hormone concentrations with hip fracture risk in postmenopausal women. J Clin Endocrinol Metab. 2008;93(5):1796–1803. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Hsu B, Seibel MJ, Cumming RG, Blyth FM, Naganathan V, Bleicher K, et al. Progressive temporal change in serum SHBG, but not in serum testosterone or estradiol, is associated with bone loss and incident fractures in older men: the Concord health and ageing in men project. J Bone Miner Res. 2016;31(12):2115–2122. [DOI] [PubMed] [Google Scholar]
- 34. Boonen S, Dejaeger E, Vanderschueren D, Venken K, Bogaerts A, Verschueren S, et al. Osteoporosis and osteoporotic fracture occurrence and prevention in the elderly: a geriatric perspective. Best Pract Res Clin Endocrinol Metab. 2008;22(5):765–785. [DOI] [PubMed] [Google Scholar]
- 35. Rosenberg EA, Bůžková P, Fink HA, Robbins JA, Shores MM, Matsumoto AM, et al. Testosterone, dihydrotestosterone, bone density, and hip fracture risk among older men: the cardiovascular health study. Metabolism. 2021;114:154399. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Aribas E, Kavousi M, Laven JSE, Ikram MA, Roeters van Lennep JE. Aging, cardiovascular risk, and SHBG levels in men and women from the general population. J Clin Endocrinol Metab. 2021;106(10):2890–2900. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Hidayat K, Du X, Shi BM. Sex hormone‐binding globulin and risk of fracture in older adults: systematic review and meta‐analysis of observational studies. Osteoporos Int. 2018;29(10):2171–2180. [DOI] [PubMed] [Google Scholar]
- 38. Spitzer M, Huang G, Basaria S, Travison TG, Bhasin S. Risks and benefits of testosterone therapy in older men. Nat Rev Endocrinol. 2013;9(7):414–424. [DOI] [PubMed] [Google Scholar]
- 39. Kenny AM, Kleppinger A, Annis K, Rathier M, Browner B, Judge JO, et al. Effects of transdermal testosterone on bone and muscle in older men with low bioavailable testosterone levels, low bone mass, and physical frailty. J Am Geriatr Soc. 2010;58(6):1134–1143. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Shigehara K, Izumi K, Kadono Y, Mizokami A. Testosterone and bone health in men: a narrative review. J Clin Med. 2021;10(3):530. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41. Amin S, Zhang Y, Felson DT, Sawin CT, Hannan MT, Wilson PWF, et al. Estradiol, testosterone, and the risk for hip fractures in elderly men from the Framingham study. Am J Med. 2006;119(5):426–433. [DOI] [PubMed] [Google Scholar]
- 42. Jassal SK, Barrett‐Connor E, Edelstein SL. Low bioavailable testosterone levels predict future height loss in postmenopausal women. J Bone Miner Res. 1995;10(4):650–654. [DOI] [PubMed] [Google Scholar]
- 43. Rubin KH, Glintborg D, Nybo M, Andersen M, Abrahamsen B. Fracture risk is decreased in women with polycystic ovary syndrome: a register‐based and population‐based cohort study. J Bone Miner Res. 2016;31(4):709–717. [DOI] [PubMed] [Google Scholar]
- 44. Davey Smith G, Hemani G. Mendelian randomization: genetic anchors for causal inference in epidemiological studies. Hum Mol Genet. 2014;23(R1):R89–R98. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45. Ference BA. How to use mendelian randomization to anticipate the results of randomized trials. Eur Heart J. 2018;39(5):360–362. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Figure S1. The scatter plot of SNPs associated with SHBG and their risk of fractures. (A) Pathological fracture with OP; (B) Fracture of lumbar spine and pelvis; (C) Fracture of ribs, sternum, and thoracic spine; (D) Fracture of neck. SHBG: Sex hormone‐binding globulin; OP: Osteoporosis.
Figure S2. The scatter plot of SNPs associated with SHBG and their risk of fractures. (A) Fracture of skull and facial bones; (B) Fracture of femur; (C) Fracture of lower leg including ankle; (D) Fracture of foot except ankle. SHBG: Sex hormone‐binding globulin.
Figure S3. The scatter plot of SNPs associated with SHBG and their risk of fractures. (A) Fracture of shoulder and upper arm; (B) Fracture of forearm; (C) Fracture at wrist and hand level. SHBG: Sex hormone‐binding globulin.
Figure S4. The scatter plot of SNPs associated with bioavailable testosterone and their risk of fractures. (A) Pathological fracture with OP; (B) Fracture of lumbar spine and pelvis; (C) Fracture of ribs, sternum, and thoracic spine; (D) Fracture of neck. OP: osteoporosis.
Figure S5. The scatter plot of SNPs associated with bioavailable testosterone and their risk of fractures. (A) Fracture of skull and facial bones; (B) Fracture of femur; (C) Fracture of lower leg including ankle; (D) Fracture of foot except ankle.
Figure S6. The scatter plot of SNPs associated with bioavailable testosterone and their risk of fractures. (A) Fracture of shoulder and upper arm; (B) Fracture of forearm; (C) Fracture at wrist and hand level.
Figure S7. The funnel plot of SNPs associated with SHBG and their risk of fractures. (A) Pathological fracture with OP; (B) Fracture of lumbar spine and pelvis; (C) Fracture of ribs, sternum, and thoracic spine; (D) Fracture of neck. SHBG: Sex hormone‐binding globulin; OP: Osteoporosis.
Figure S8. The funnel plot of SNPs associated with SHBG and their risk of fractures. (A) Fracture of skull and facial bones; (B) Fracture of femur; (C) Fracture of lower leg including ankle; (D) Fracture of foot except ankle. SHBG: Sex hormone‐binding globulin.
Figure S9. The funnel plot of SNPs associated with SHBG and their risk of fractures. (A) Fracture of shoulder and upper arm; (B) Fracture of forearm; (C) Fracture at wrist and hand level. SHBG: Sex hormone‐binding globulin.
Figure S10. The funnel plot of SNPs associated with bioavailable testosterone and their risk of fractures. (A) Pathological fracture with OP; (B) Fracture of lumbar spine and pelvis; (C) Fracture of ribs, sternum, and thoracic spine; (D) Fracture of neck. OP: osteoporosis.
Figure S11. The funnel plot of SNPs associated with bioavailable testosterone and their risk of fractures. (A) Fracture of skull and facial bones; (B) Fracture of femur; (C) Fracture of lower leg including ankle; (D) Fracture of foot except ankle.
Figure S12. The funnel plot of SNPs associated with bioavailable testosterone and their risk of fractures. (A) Fracture of shoulder and upper arm; (B) Fracture of forearm; (C) Fracture at wrist and hand level.
Figure S13. The leave‐one‐out test plot of the causal association between SHBG and the risk of fractures. (A) Pathological fracture with OP; (B) Fracture of lumbar spine and pelvis. SHBG: Sex hormone‐binding globulin; OP: Osteoporosis.
Figure S14. The leave‐one‐out test plot of the causal association between SHBG and the risk of fractures. (A) Fracture of ribs, sternum, and thoracic spine; (B) Fracture of neck. SHBG: Sex hormone‐binding globulin.
Figure S15. The leave‐one‐out test plot of the causal association between SHBG and the risk of fractures. (A) Fracture of skull and facial bones; (B) Fracture of femur. SHBG: Sex hormone‐binding globulin.
Figure S16. The leave‐one‐out test plot of the causal association between SHBG and the risk of fractures. (A) Fracture of lower leg including ankle; (B) Fracture of foot except ankle. SHBG: Sex hormone‐binding globulin.
Figure S17. The leave‐one‐out test plot of the causal association between SHBG and the risk of fractures. (A) Fracture of shoulder and upper arm; (B) Fracture of forearm. SHBG: Sex hormone‐binding globulin.
Figure S18. The leave‐one‐out test plot of the causal association between SHBG and the risk of fracture at wrist and hand level. SHBG: Sex hormone‐binding globulin.
Figure S19. The leave‐one‐out test plot of the causal association between bioavailable testosterone and the risk of fractures. (A) Pathological fracture with OP; (B) Fracture of lumbar spine and pelvis (C) Fracture of ribs, sternum and thoracic spine (D) Fracture of neck. OP: Osteoporosis.
Figure S20. The leave‐one‐out test plot of the causal association between bioavailable testosterone and the risk of fractures. (A) Fracture of skull and facial bones; (B) Fracture of femur; (C) Fracture of lower leg including ankle; (D) Fracture of foot except ankle.
Figure S21. The funnel plot of SNPs associated with bioavailable testosterone and their risk of fractures. (A) Fracture of shoulder and upper arm; (B) Fracture of forearm; (C) Fracture at wrist and hand levels.
Table S1. The SNPs related to the outcomes when exposure is SHBG.
Table S2. The SNPs related to the outcomes when exposure is bioavailable testosterone.
Data Availability Statement
The original contributions presented in the study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding author.