Causal effects of serum sex hormone binding protein levels on the risk of amyotrophic lateral sclerosis: a mendelian randomization study

Introduction

Because of extended lifespans, the prevalence of age-related neurodegenerative disorders is increasing (1). These disorders impair individuals’ memory, cognition, mood, and movement, but currently none are curable; existing treatments can only manage the symptoms or delay disease progression. Thus, viable biomarkers of aging-related neurodegenerative diseases, especially in peripheral blood, are crucial for the early warning, diagnosis, and treatment of these diseases (2).

Sex hormones have been reported to play an important role in human brain development, showing neuroprotective effects by preserving neural function and promoting neuronal survival (3). Sex-hormone binding globulin (SHBG) is a hepatically secreted binding protein for sex hormones in plasma that prevents hormones from binding to intracellular androgen or estrogen receptors (4). Thus, it is considered the major factor controlling the balance between biologically active testosterone and estradiol (4). Emerging evidence indicates that peripheral SHBG levels might be an effective indicator of the occurrence or progression of neurodegenerative disorders (5). However, inconsistencies and uncertainties exist and causality remains unclear. Higher SHBG has been found to be associated with worse cognitive performance and an increased risk of developing Alzheimer’s disease (AD) and all-cause dementia (6). Recently, by integrating information from two databases, the Chinese Alzheimer’s Biomarker and LifestylE (CABLE) study and Alzheimer’s Disease Neuroimaging Initiative (ADNI) cohort, our research team found that plasma SHBG could be a predictive biomarker for AD progression (5). However, the effects of SHBG on other neurodegenerative diseases have been poorly investigated. A previous Polish study revealed no difference in SHBG levels between Parkinson’s disease (PD) patients and healthy subjects (7). In addition, a subtle relationship between SHBG and amyotrophic lateral sclerosis (ALS) has been revealed, which might be explained by the possible involvement of testosterone in ALS causation (8). However, the exploration of other neurological disorders remains controversial.

In addition, existing studies are limited to observational designs, which are subject to confounding bias and reverse causation (9). Although epidemiological studies have adjusted for confounding factors observed in study participants, confounding bias is inevitable (10). Reverse causality bias arises if preclinical states that lead to outcomes also affect their risk factors (10). People consciously reduce their exposure to risk factors after acquiring an illness. Mendelian randomization (MR) is a technique that allows the examination of causal relationships (10). This method minimizes confounding bias because genetic variants are randomly allocated during conception. Reverse causality bias is also precluded because the genotypes are not affected by the disease. MR assesses lifelong exposure to health-related outcomes; thus, it can reveal potential causal associations (11). It is becoming increasingly viable, as data from numerous large genome-wide association studies (GWAS) over the past decade are now publicly available. The MR approach has been employed to uncover the causal effects of many risk factors regarding the incidence of neurodegenerative diseases (12-15).

We conducted a two-sample MR study to explore the causal effects of genetically predicted SHBG levels in the serum on common neurodegenerative diseases such as AD, including maternal/paternal family history of AD, as well as PD, ALS, multiple sclerosis (MS), dementia with Lewy bodies (DLB), and frontotemporal dementia (FTD). We present the following article in accordance with the STROBE-MR reporting checklist (available at https://atm.amegroups.com/article/view/10.21037/atm-22-1156/rc).

Methods

Study design and instrument identification

Two-sample MR analysis is a genetic instrumental variable (IV) analysis based on summary-level data with single nucleotide polymorphisms (SNPs) as instruments for risk factors. This method has been widely used. The MR approach is based on three assumptions as follows: (I) the genetic variants are significantly associated with SHBG levels in serum; (II) the IVs (namely, SNPs) have no association with confounding factors; and (III) the risks of outcomes (the six neurodegenerative diseases) are influenced only by exposure (serum SHBG), not by other pathways (16) (Figure 1). This study analyzed publicly available summary level data from large GWASs. Informed consents and ethical approvals were obtained for the original studies. The study was conducted in accordance with the Declaration of Helsinki (as revised in 2013).

Figure 1 Scheme diagram of the Mendelian randomization design. The MR approach should satisfy three assumptions. 1: the genetic variants significantly associate with the exposures; in this study, the two exposures are serum SHBG levels and neurological disorders; 2: the IVs (namely, SNPs) have no association with confounding factors; 3: the risk of outcome (neurological disorders) is influenced only by the exposure (serum SHBG), not by other pathways. MR, Mendelian randomization; SHBG, sex-hormone binding globulin; SNP, single nucleotide polymorphism.

Significant SNPs (P<5×10⁻⁸) for serum SHBG levels were identified in a GWAS meta-analysis that included 363,228 individuals of European ancestry drawn from the UK Biobank (UKB) (17). The UKB was established in 2006 and enrolled participants from the entire UK population, with a recruitment age of 45–69 years. This study evaluated the genetic basis of 35 blood and urine laboratory measurements and identified 1,857 loci associated with at least one trait. Serum SHBG level was one of the 35 variables. Serum SHBG levels (nmol/L) were detected using the chemiluminescent two-step sandwich immunoassay method in Beckman Coulter DXI 800 (Beckman Coulter, UK), Ltd. The good or acceptable distribution rate was 95%. Further detailed information is provided in Table S1.

Outcome databases

The outcomes of AD, PD, ALS, MS, DLB, and FTD were all clinically diagnosed or autopsy-diagnosed. The genetic variants associated with AD were extracted from the following GWAS summary statistics: (I) the International Genomics of Alzheimer’s Project (IGAP) GWAS Stage 1 result (N=21,982 cases, 41,944 controls) (18); (II) maternal family history of AD (N=27,696 cases, 260,980 controls); and (III) paternal family history of AD (N=14,338 cases, 245,941 controls) (19). Late-onset AD was either autopsy-confirmed or clinically confirmed. Maternal and paternal family histories of AD were self-reported (Table S1). The data of maternal and paternal family history of AD were extracted from the same dataset (UKB) as SHBG; thus, there was considerable overlap between the exposure and outcome samples. We only considered this part of the analysis as supplementary to support our main findings.

A recently published PD GWAS meta-analysis, which included three sources of data (three previously published GWAS studies, 13 new datasets, and proxy-case data from the UKB), was used as the PD GWAS source (N=37,688 cases, 18,618 UKB proxy-cases, and 1.4 million controls) (20). Summary statistics for ALS were obtained from a large GWAS involving 80,610 participants of European descent (20,806 ALS cases and 59,804 controls) (21). Patients were diagnosed with probable or definite ALS according to the EI Escorial criteria (22). The MS GWAS data leveraged genotype data from 47,429 MS cases and 68,374 controls of European descent from the International Multiple Sclerosis Genetics Consortium (23). DLB data were obtained from whole-genome sequencing of a cohort of 2,981 patients diagnosed with DLB and 4,391 neurologically healthy individuals (24). Participants were recruited from 44 institutions/consortia and diagnosed according to the established consensus criteria. FTD data were obtained from a two-stage GWAS with samples from 3,526 clinical FTD patients and 9,402 healthy controls (25). To reduce genetic heterogeneity, all the participants were of European ancestry (Table S1).

Instrument selection

We estimated the overall effect of serum SHBG on multiple neurodegenerative diseases by combining the effects of genome-wide significant SNPs (P<5×10⁻⁸) from the GWAS, which were then clumped based on the European 1000 Genomes panel to a stringent LD threshold (R²<0.00001) and then a default LD threshold (R²<0.001). Several SNPs were further excluded to eliminate the genetic bias produced by the palindrome with intermediate allele frequencies. Eventually, 131 SNPs were included in late-onset AD, 130 SNPs for maternal AD, 130 SNPs for paternal AD, 130 SNPs for PD, 131 SNPs for ALS, 125 SNPs for MS, 122 SNPs for DLB, and 109 SNPs for FTD. There were no SNPs with F-statistics <10. The screening process is shown in Figure 2, and the included SNPs are shown in Tables S2-S9.

Figure 2 A flow diagram of the process in this MR analysis. SNP, single nucleotide polymorphism; GWAS, genome-wide association studies; IV, instrumental variable; LD, linkage disequilibrium; AD, Alzheimer’s disease; PD, Parkinson’s disease; ALS, amyotrophic lateral sclerosis; MS, multiple sclerosis; DLB, dementia with Lewy bodies; FTD, frontotemporal dementia; MR, Mendelian randomization; SHBG, sex hormone binding globulin.

Statistical analyses

Causal effects were estimated using the random-effects maximum likelihood estimation method. We applied four complementary methods [inverse variance weighted (IVW), MR-Egger, weighted median, and weighted mode], which provided different assumptions regarding horizontal pleiotropy (26). The IVW method was performed as our primary method, which combined the Wald ratio estimates of the causal effects obtained from different SNPs. The intercept was assumed to be zero and associated with a weighted regression of SNP-exposure effects with SNP-outcome effects (27). MR-Egger regression is not constrained to have a slope of zero; therefore, its causal estimate represents a genotype-outcome dose-response relationship that takes pleiotropic effects into account (26,28). The weighted median approach is defined as the median of a weighted empirical density function of the ratio estimates, giving more weight to more precise IVs. The estimate is consistent even when up to 50% of the information comes from invalid or weak instruments (10). Results were presented as odds ratios (ORs) and 95% CIs. The MR-Egger intercept, MR-PRESSO global test, and Cochran Q statistics were used to test for the presence of heterogeneity or directional pleiotropy (29). The leave-one-SNP-out analysis was performed by systematic removal of genetic instruments from MR analysis to identify influential outliers. F-statistics were used to measure the strength of the genetic instruments in IVW (30). The F-statistics were >10, indicating that the instrument strength was sufficient for MR analysis and less likely to be influenced by weak instrument bias (31).

Statistical significance of the above analyses was set at a 2-sided P value of <0.05. Statistical analyses were conducted using R (version 3.6.3), and MR analyses were conducted using “TwoSampleMR”.

Results

Genetically determined serum SHBG levels and AD risk

No obvious significant causal association between genetically determined serum SHBG levels and the risk of late-onset AD was found (OR_IVW =0.948, 95% CI: 0.854–1.052, P=0.315; SNPs =131; Figure 3, Figure S1), whereas the results of the sensitivity analysis using the weighted median method were suggestive of protective effects with an OR of 0.862 (95% CI: 0.740–1.004, P=0.057). Additional analysis applying an R²<0.001 showed similar results (Figures S2,S3). There was evidence of heterogeneity in the causal effect estimates (P for MR-Egger =4.71E-05, P for IVW =5.37E-05; Table 1). The MR-PRESSO global test revealed horizontal pleiotropic effects (P<0.001). However, no significant outliers were observed. Similar null associations between maternal and paternal AD were identified using 130 IVs. No evidence for heterogeneity of effect sizes (Cochran Q statistic, P>0.05) or pleiotropy (intercept: 2.39E-04, P=0.959; P for MR-PRESSO global test =0.953) was found for paternal AD, while evidence of heterogeneity existed in the causal effect estimates for maternal AD (P for MR Egger =0.028, P for IVW =0.031; P for MR-PRESSO global test =0.044). There was no distortion in the leave-one-out and single-SNP plots, suggesting that no single SNP was driving the observed effect in any of the analyses (Figures S4-S6). The F-statistics for the three association pairs were 162.19, 153.85, and 153.85.

Figure 3 Mendelian randomization results of serum SHBG level and neurodegenerative diseases. Genome-wide significantly associated (P<5×10⁻⁸) independent (linkage disequilibrium R²=0.00001, clumping distance =10,000 kb) SNPs were used as instruments. Bold fonts represent significant results. SNP, single nucleotide polymorphism; OR, odds ratio; MR, Mendelian randomization; AD, Alzheimer’s disease; PD, Parkinson’s disease; ALS, amyotrophic lateral sclerosis; MS, multiple sclerosis; DLB, dementia with Lewy bodies; FTD, frontotemporal dementia; SHBG, sex-hormone binding globulin.

Genetically determined serum SHBG and other neurodegenerative diseases

The IVW method failed to support a causal relationship between serum SHBG and PD, with an OR of 1.020 (95% CI: 0.891–1.169, P=0.772; SNPs =130; Figure 3, Figure S1). This effect was supported by the weighted mode method (OR =0.831, 95% CI: 0.683–1.010, P=0.065). Additional analyses (LD R²<0.001) using MR-Egger (OR =795, 95% CI: 0.638–0.990, P=0.042; SNPs =193) and weighted mode (OR =0.827, 95% CI: 0.700–0.977, P=0.027; SNPs =193) methods showed a supportive association (Figures S2,S3). Evidence for the heterogeneity of effect sizes (P for MR Egger =3.20E-05, P for IVW =2.96E-05; Table 1) was revealed. Nonetheless, horizontal pleiotropic effects were absent (intercept: 0.005, P=0.309; P for the MR-PRESSO global test =0.202). We did not find a single genetic variant of SHBG that influenced the association in the leave-one-out analysis (Figure S7). The F-statistic value was 153.85.

Interestingly, the results showed a statistically significant causal effect of genetically predicted serum SHBG levels on ALS (OR_IVW =1.113, 95% CI: 1.019–1.215, P=0.017; Figure 3, Figure S1). The causal effect was confirmed by sensitivity analyses including MR-Egger (OR_MR-Egger =1.229, 95% CI: 1.049–1.441, P=0.012), weighted median (OR_{weighted median} =1.231, 95% CI: 1.077–1.406, P=0.002), and weighted mode (OR_{weighted mode} =1.235, 95% CI: 1.067–1.431, P=0.005) methods (Figures S2,S3). Almost no evidence of heterogeneity of effect sizes (P for MR Egger =0.055, P for IVW =0.046; Table 1) or directional pleiotropy (intercept: −0.004, P=0.144; P for MR-PRESSO global test =0.051) was found. However, rs9892297 significantly drove the overall direction (Figure S8). The F statistics of 166.54 indicated that the association was less likely to be affected by weak instrumental bias.

No causal effects of SHBG on MS (125 SNPs), DLB (122 SNPs), or FTD (109 SNPs) were observed (Figure 3, Figure S1). Additional analyses supported these findings (Figures S2,S3). Cochran’s Q statistics (P for MR Egger =5.13E-131, P for IVW =2.53E-131) indicated notable heterogeneity across instrument SNP effects for MS (Table 1). The MR-PRESSO global test indicated pleiotropy (P<0.001). Although several outliers were identified, the overall null causal effect remained unchanged. Directional pleiotropy by Egger analysis was found for DLB (intercept: −0.018, P=0.014). No heterogeneity or pleiotropy was detected in the FTD analysis (all P>0.05). No single SNP was found to drive the above associations (Figures S9-S11). The F-statistics for the three association pairs were 166.54, 146.97, and 158.91, respectively.

Discussion

To the best of our knowledge, this is the first two-sample MR study to examine the causal associations between serum SHBG levels and several common neurodegenerative diseases. We revealed that genetically predicted serum SHBG levels were associated with the risk of developing ALS but did not provide solid MR evidence to support the causal effects on AD (including maternal and paternal family history of AD), PD, MS, DLB, or FTD. These results should be interpreted with caution, given that some of our results may be driven by genetic pleiotropy and heterogeneity. However, the consistency of our results across MR methods strengthens our inference of causality. Overall, these results will help interpret the results from current observational studies and indicate the direction of future application of hormone replacement therapy (HRT).

Our MR study revealed the deleterious effects of SHBG on ALS. Few observational studies have explored the association between SHBG and ALS. ALS is a progressive neurodegenerative disorder characterized by the involvement of both upper and lower motor neurons (32). It has been postulated that testosterone may play a role in ALS (8). In patients susceptible to ALS, there is possibly a sort of “testosterone resistance” at the level of the blood-brain barrier (BBB) commencing from birth (8). In these patients, testosterone at low levels can penetrate the BBB and enter the central neural axis. Then, 5 α-reductase in the anterior pituitary converts testosterone into dihydrotestosterone (DHT). DHT deficiency can lead to motor neuron death, ultimately leading to ALS. SHBG is a major factor controlling the balance between biologically active sex hormones (4). Higher SHBG levels are associated with lower levels of biologically active free testosterone. Obviously, with advancing age, the circadian excursion in free non-SHBG-binding testosterone declines, resulting in a decrease in free testosterone available for intracerebral transport. Thus, susceptibility to ALS increases in individuals with BBB “testosterone resistance” (4). This is probably the reason for the increased incidence of ALS with age. To our knowledge, this is the first MR study to use genetic instruments, showing that genetically determined serum SHBG levels are causally associated with ALS. Our study sheds light on the causal relationship between SHBG and ALS, and highlights the potential of SHBG as a biomarker for ALS.

Accumulating observational studies have revealed significant associations between SHBG and AD risk (6) and pathologies (5,33). Nevertheless, the present MR analysis failed to support causality. We acknowledge several reasons for this finding. First, previous observational studies may have been affected by reverse causality and confounding bias (9). Increased SHBG levels are a surrogate marker for other known risk factors for dementia, especially age (34). Furthermore, many previous studies have demonstrated that high SHBG levels are associated with smoking, alcohol intake, lower physical activity, hyperinsulinemia, and metabolic syndrome (35,36), all of which have been reported to be related to AD development. Thus, it may be useful to consider SHBG as an early indicator of AD rather than as a direct modifier of AD risk. Second, the relationship between SHBG and AD may be complex and nonlinear, and could not be fully explored in our study. Thus, SHBG may exert stronger effects on later disease development. The association between SHBG and these diseases also largely depends on heterogeneous features (e.g., age, sex, and HRT). Therefore, further detailed studies are warranted. Third, the strength of current genetic variants may have contributed to discrepancies between phenotypic and MR associations, and the results may require updating as new genetic discoveries become available. Although previous observational studies have highlighted the association between serum SHBG and incident AD, it was not supported in the MR analysis.

Apart from AD, other null results should be interpreted with caution. The prevalence of PD is 1.5 times higher in men than in women (37). This evident sex difference in the occurrence of the disease suggests that sex hormones may alter an individual's susceptibility to the disease (38). However, a previous Polish study revealed no difference in SHBG levels between patients with PD (N=36) and healthy controls (N=69) (7). Therefore, further research is required to explore the effects of sex hormones and SHBG on PD. To the best of our knowledge, few studies have investigated the effects of SHBG on MS, DLB, and FTD, and these associations warrant further investigation.

There are significant sex differences in the onset and progression of neurodegenerative diseases (39,40). Estrogens have been reported to exert neuroprotective effects through their action on cognate nuclear and membrane receptors (39). In the brain, sex hormones might affect a variety of signaling pathways, including catecholaminergic and acetylcholine pathways, which regulate cognition, motor, emotion, and other functions (41). Correspondingly, sex hormone receptors can induce various signaling cascades that mediate sex differences in neurodegenerative disorders (42). Furthermore, the preventive effects of HRT on cognitive impairment and dementia have been extensively studied (43). Unfortunately, these trials were not successful. Previous randomized controlled trials (RCTs) tended to focus on hormones (44,45) and rarely considered hormone transporters such as SHBG. Our MR research systematically interpreted the relationship between SHBG and neurodegenerative diseases, extending traditional observational associations to gene-mediated causality. It may be crucial for SHBG to be more widely included as a measure, along with other blood biomarkers, in hormone therapy studies or clinical trials.

Our study had several limitations that merit consideration. First, to avoid horizontal pleiotropy, a general challenge for MR, we used MR-PRESSO and MR-Egger regressions to estimate the extent to which heterogeneity and pleiotropy may bias the reported results. However, the possibility of heterogeneity and pleiotropy cannot be ruled out. Second, we cannot exclude the possibility of inflating the type 1 error rate because there were overlaps between exposure GWAS and outcome GWASs, especially for maternal and paternal family histories of AD. Third, it is important to recognize that MR measures the cumulative effect of lifelong exposure to genetic variants related to serum SHBG levels, unlike studies of the effects of discrete clinical interventions in adult life. Therefore, these MR results should not be extrapolated to determine the effect of SHBG on outcomes at a particular time period. Finally, the serum SHBG data were extracted from the UK biobank, a more educated, less deprived cohort, whose age range was 40–70 years, and thus might have poor representativeness of the UK general population. Moreover, since the majority of participants were of European ancestry, the results of this study are not necessarily applicable to other ethnicities.

Conclusions

Our findings highlight the important role of serum SHBG levels in neurodegenerative diseases, particularly ALS. Since the evaluation of sex hormones and SHBG in peripheral blood can be readily performed, they have the potential to be useful screening biomarkers for aging-related diseases. Further research is required regarding to study the role of SHBG, especially because of the increasing utilization of hormone therapy.

Acknowledgments

We would like to thank Editage (https://www.editage.cn/) for English language editing. This work was made possible by the generous sharing of GWAS summary statistics. We gratefully acknowledge the authors and participants of all GWAS from which we used summary statistics data. We thank Kunkle et al., Marioni et al., Nalls et al., Nicolas et al., International Multiple Sclerosis Genetics Consortium, Chia et al., and Ferrari et al. for the late-onset AD, familial AD, PD, ALS, MS, DLB, and FTD GWAS summary results data. We also thank Sinnott-Armstrong et al. for the SHBG GWAS dataset. IGAP was made possible by the generous participation of the control subjects, the patients, and their families. The i-Select chips were funded by the French National Foundation on AD and related disorders. EADI was supported by the LABEX DISTALZ grant, Inserm, Institut Pasteur de Lille, Université de Lille 2, and the Lille University Hospital. GERAD was supported by the Medical Research Council (Grant No. 503480), Alzheimer’s Research UK (Grant No. 503176), the Wellcome Trust (Grant No. 082604/2/07/Z), and German Federal Ministry of Education and Research (BMBF): Competence Network Dementia (CND) grant No. 01GI0102, 01GI0711, 01GI0420. CHARGE was partly supported by the NIH/NIA grant R01 AG033193 and the NIAAG081220 and AGES contract N01-AG-12100, the NHLBI grant R01 HL105756, the Icelandic Heart Association, and the Erasmus Medical Center and Erasmus University. ADGC was supported by the NIH/NIA grants: U01 AG032984, U24 AG021886, U01AG016976, and the Alzheimer’s Association grant ADGC-10-196728. The investigators within above consortiums contributed to the design and implementation of the GWAS datasets and/or provided data but did not participate in analysis or writing of this report.

Funding: This study was supported by grants from the National Natural Science Foundation of China (Nos. 82071201, 91849126, 82071997), Shanghai Municipal Science and Technology Major Project (No. 2018SHZDZX01), Research Start-up Fund of Huashan Hospital (No. 2022QD002), Excellence 2025 Talent Cultivation Program at Fudan University (No. 3030277001), and ZHANGJIANG LAB, Tianqiao and Chrissy Chen Institute, and the State Key Laboratory of Neurobiology and Frontiers Center for Brain Science of Ministry of Education, Fudan University.

Footnote

Reporting Checklist: The authors have completed the STROBE-MR reporting checklist. Available at https://atm.amegroups.com/article/view/10.21037/atm-22-1156/rc

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://atm.amegroups.com/article/view/10.21037/atm-22-1156/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The current analyses based on publicly available summary data and therefore does not require ethical approval. Original studies have been approved by ethic committees and written informed consent was obtained from study participants or caregivers. The study was conducted in accordance with the Declaration of Helsinki (as revised in 2013).

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Heemels MT. Neurodegenerative diseases. Nature 2016;539:179. [Crossref] [PubMed]
2. Turner MR. The early biomarker challenge in neurodegenerative disorders. J Neurol Neurosurg Psychiatry 2019;90:1190-1. [Crossref] [PubMed]
3. Veiga S, Melcangi RC, Doncarlos LL, et al. Sex hormones and brain aging. Exp Gerontol 2004;39:1623-31. [Crossref] [PubMed]
4. Rosner W. The functions of corticosteroid-binding globulin and sex hormone-binding globulin: recent advances. Endocr Rev 1990;11:80-91. [Crossref] [PubMed]
5. Xu W, Su BJ, Shen XN, et al. Plasma sex hormone-binding globulin predicts neurodegeneration and clinical progression in prodromal Alzheimer’s disease. Aging (Albany NY) 2020;12:14528-41. [Crossref] [PubMed]
6. Muller M, Schupf N, Manly JJ, et al. Sex hormone binding globulin and incident Alzheimer’s disease in elderly men and women. Neurobiol Aging 2010;31:1758-1765. [Crossref] [PubMed]
7. Nitkowska M, Tomasiuk R, Czyżyk M, et al. Prolactin and sex hormones levels in males with Parkinson’s disease. Acta Neurol Scand 2015;131:411-6. [Crossref] [PubMed]
8. Sawal N, Kaur J, Kaur K, Gombar S. Dihydrotestosterone in Amyotrophic lateral sclerosis-The missing link? Brain Behav 2020;10:e01645. [Crossref] [PubMed]
9. Xu J, Xia LL, Song N, et al. Testosterone, Estradiol, and Sex Hormone-Binding Globulin in Alzheimer’s Disease: A Meta-Analysis. Curr Alzheimer Res 2016;13:215-22. [Crossref] [PubMed]
10. Davies NM, Holmes MV, Davey Smith G. Reading Mendelian randomisation studies: a guide, glossary, and checklist for clinicians. BMJ 2018;362:k601. [Crossref] [PubMed]
11. Walker VM, Davey Smith G, Davies NM, et al. Mendelian randomization: a novel approach for the prediction of adverse drug events and drug repurposing opportunities. Int J Epidemiol 2017;46:2078-89. [Crossref] [PubMed]
12. Rasooly D, Patel CJ. Conducting a Reproducible Mendelian Randomization Analysis Using the R Analytic Statistical Environment. Curr Protoc Hum Genet 2019;101:e82. [Crossref] [PubMed]
13. Domenighetti C, Sugier PE, Ashok Kumar Sreelatha A, et al. Dairy Intake and Parkinson’s Disease: A Mendelian Randomization Study. Mov Disord 2022;37:857-64. [Crossref] [PubMed]
14. Di H, Zhu Y, Xia W, et al. Bidirectional Mendelian randomization to explore the causal relationships between Sleep traits, Parkinson's disease and Amyotrophic lateral sclerosis. Sleep Med 2022;96:42-9. [Crossref] [PubMed]
15. Yang Y, Musco H, Simpson-Yap S, et al. Investigating the shared genetic architecture between multiple sclerosis and inflammatory bowel diseases. Nat Commun 2021;12:5641. [Crossref] [PubMed]
16. Davey Smith G, Hemani G. Mendelian randomization: genetic anchors for causal inference in epidemiological studies. Hum Mol Genet 2014;23:R89-98. [Crossref] [PubMed]
17. Sinnott-Armstrong N, Tanigawa Y, Amar D, et al. Genetics of 35 blood and urine biomarkers in the UK Biobank. Nat Genet 2021;53:185-94. [Crossref] [PubMed]
18. Kunkle BW, Grenier-Boley B, Sims R, et al. Genetic meta-analysis of diagnosed Alzheimer’s disease identifies new risk loci and implicates Aβ, tau, immunity and lipid processing. Nat Genet 2019;51:414-30. Erratum in: Nat Genet 2019;51:1423-4. [Crossref] [PubMed]
19. Marioni RE, Harris SE, Zhang Q, et al. GWAS on family history of Alzheimer's disease. Transl Psychiatry 2018;8:99. [Crossref] [PubMed]
20. Nalls MA, Blauwendraat C, Vallerga CL, et al. Identification of novel risk loci, causal insights, and heritable risk for Parkinson’s disease: a meta-analysis of genome-wide association studies. Lancet Neurol 2019;18:1091-102. [Crossref] [PubMed]
21. Nicolas A, Kenna KP, Renton AE, et al. Genome-wide Analyses Identify KIF5A as a Novel ALS Gene. Neuron 2018;97:1268-1283.e6. [Crossref] [PubMed]
22. Brooks BR. El Escorial World Federation of Neurology criteria for the diagnosis of amyotrophic lateral sclerosis. Subcommittee on Motor Neuron Diseases/Amyotrophic Lateral Sclerosis of the World Federation of Neurology Research Group on Neuromuscular Diseases and the El Escorial “Clinical limits of amyotrophic lateral sclerosis” workshop contributors. J Neurol Sci 1994;124:96-107. [Crossref] [PubMed]
23. International Multiple Sclerosis Genetics Consortium. Multiple sclerosis genomic map implicates peripheral immune cells and microglia in susceptibility. Science 2019;365:eaav7188. [Crossref] [PubMed]
24. Chia R, Sabir MS, Bandres-Ciga S, et al. Genome sequencing analysis identifies new loci associated with Lewy body dementia and provides insights into its genetic architecture. Nat Genet 2021;53:294-303. [Crossref] [PubMed]
25. Ferrari R, Hernandez DG, Nalls MA, et al. Frontotemporal dementia and its subtypes: a genome-wide association study. Lancet Neurol 2014;13:686-99. [Crossref] [PubMed]
26. Bowden J, Davey Smith G, Burgess S. Mendelian randomization with invalid instruments: effect estimation and bias detection through Egger regression. Int J Epidemiol 2015;44:512-25. [Crossref] [PubMed]
27. Burgess S, Butterworth A, Thompson SG. Mendelian randomization analysis with multiple genetic variants using summarized data. Genet Epidemiol 2013;37:658-65. [Crossref] [PubMed]
28. Hemani G, Bowden J, Davey Smith G. Evaluating the potential role of pleiotropy in Mendelian randomization studies. Hum Mol Genet 2018;27:R195-208. [Crossref] [PubMed]
29. Bowden J, Spiller W, Del Greco M F, et al. Improving the visualization, interpretation and analysis of two-sample summary data Mendelian randomization via the Radial plot and Radial regression. Int J Epidemiol 2018;47:1264-78. [Crossref] [PubMed]
30. Pichler I, Del Greco M F, Gögele M, et al. Serum iron levels and the risk of Parkinson disease: a Mendelian randomization study. PLoS Med 2013;10:e1001462. [Crossref] [PubMed]
31. Burgess S, Thompson SGCRP CHD Genetics Collaboration. Avoiding bias from weak instruments in Mendelian randomization studies. Int J Epidemiol 2011;40:755-64. [Crossref] [PubMed]
32. Chiò A, Logroscino G, Hardiman O, et al. Prognostic factors in ALS: A critical review. Amyotroph Lateral Scler 2009;10:310-23. [Crossref] [PubMed]
33. Gillett MJ, Martins RN, Clarnette RM, et al. Relationship between testosterone, sex hormone binding globulin and plasma amyloid beta peptide 40 in older men with subjective memory loss or dementia. J Alzheimers Dis 2003;5:267-9. [Crossref] [PubMed]
34. Diver MJ, Imtiaz KE, Ahmad AM, et al. Diurnal rhythms of serum total, free and bioavailable testosterone and of SHBG in middle-aged men compared with those in young men. Clin Endocrinol (Oxf) 2003;58:710-7. [Crossref] [PubMed]
35. Muller M, den Tonkelaar I, Thijssen JH, et al. Endogenous sex hormones in men aged 40-80 years. Eur J Endocrinol 2003;149:583-9. [Crossref] [PubMed]
36. Muller M, Grobbee DE, den Tonkelaar I, et al. Endogenous sex hormones and metabolic syndrome in aging men. J Clin Endocrinol Metab 2005;90:2618-23. [Crossref] [PubMed]
37. Mayeux R, Denaro J, Hemenegildo N, et al. A population-based investigation of Parkinson’s disease with and without dementia. Relationship to age and gender. Arch Neurol 1992;49:492-7. [Crossref] [PubMed]
38. Haaxma CA, Bloem BR, Borm GF, et al. Gender differences in Parkinson’s disease. J Neurol Neurosurg Psychiatry 2007;78:819-24. [Crossref] [PubMed]
39. Vegeto E, Villa A, Della Torre S, et al. The Role of Sex and Sex Hormones in Neurodegenerative Diseases. Endocr Rev 2020;41:bnz005. [Crossref] [PubMed]
40. Nebel RA, Aggarwal NT, Barnes LL, et al. Understanding the impact of sex and gender in Alzheimer’s disease: A call to action. Alzheimers Dement 2018;14:1171-83. [Crossref] [PubMed]
41. Velez LM, Van C, Moore T, et al. Genetic variation of putative myokine signaling is dominated by biological sex and sex hormones. Elife 2022;11:76887. [Crossref] [PubMed]
42. Sumien N, Cunningham JT, Davis DL, et al. Neurodegenerative Disease: Roles for Sex, Hormones, and Oxidative Stress. Endocrinology 2021;162:bqab185. [Crossref] [PubMed]
43. Savolainen-Peltonen H, Rahkola-Soisalo P, Hoti F, et al. Use of postmenopausal hormone therapy and risk of Alzheimer’s disease in Finland: nationwide case-control study. BMJ 2019;364:l665. [Crossref] [PubMed]
44. Zhou HH, Yu Z, Luo L, et al. The effect of hormone replacement therapy on cognitive function in healthy postmenopausal women: a meta-analysis of 23 randomized controlled trials. Psychogeriatrics 2021;21:926-38. [Crossref] [PubMed]
45. Manson JE, Chlebowski RT, Stefanick ML, et al. Menopausal hormone therapy and health outcomes during the intervention and extended poststopping phases of the Women’s Health Initiative randomized trials. JAMA 2013;310:1353-68. [Crossref] [PubMed]