Skip to main content
Download PDF

Association between thyroid dysfunction and diabetic retinopathy: a two-sample bidirectional Mendelian randomization study

Diabetology & Metabolic Syndrome volume 16, Article number: 297 (2024) Cite this article

Abstract

Objectives

To assess the association between thyroid dysfunction and diabetic retinopathy (DR), a two-sample bidirectional Mendelian randomization (MR) study utilizing the Genome-wide Association Study (GWAS) database was conducted to investigate the causal relationship between these two variables.

Methods

In this study, GWAS of 48,328,151 single nucleotide polymorphisms(SNP) in the European population from the IEU open GWAS database were utilized as genetic tools for investigating thyroid dysfunction. The total sample size for the study on hyperthyroidism was 460,499 (case group: 3557; control group: 456,942). The total sample size for hypothyroidism was 410,141 (case group: 30,155; control group: 37,986). In addition, the data on DR were extracted from the FinnGen Biobank, comprising a total sample size of 319,046 individuals (10,413 cases and 308,633 controls). For the forward MR analysis, hyperthyroidism and hypothyroidism were considered as exposures with DR as the outcome. Reverse MR analysis was conducted using DR as exposure and hyperthyroidism and hypothyroidism as outcomes. Methods: The main analytical approach employed inverse variance weighting(IVW), supplemented by MR-Egger, Weighted mode method, weighted median, and Simple mode. Cochran's Q test, MR-PRESSO, MR-Egger and leave-one-out analysis were used to evaluate the sensitivity and pleiotropy.

Results

Two-sample bidirectional MR analysis revealed a significant association between the presence of hyperthyroidism and hypothyroidism and an increased risk of DR in the forward MR analysis (IVW: OR = 1.29, 95% [CI] = 1.12–1.49, P < 0.001; OR = 1.17, 95% CI = 1.10–1.25, P < 0.001). In the reverse MR analysis, DR was found to be associated with an elevated risk of developing hyperthyroidism and hypothyroidism (IVW: OR = 1.56, 95% CI 1.38–1.76, P < 0.001; OR = 1.41, 95% CI 1.25–1.59, P < 0.001). Furthermore, most supplementary MR methods also demonstrated statistically significant differences and exhibited effect sizes consistent with those obtained from IVW. The sensitivity analysis confirmed the relative reliability of our causal findings.

Conclusions

Our findings provide genetic evidence supporting a bidirectional causal relationship between thyroid function and DR.

Introduction

Diabetic retinopathy (DR) is a common complication of diabetes and considered the most serious microvascular complication. Its prevalence increases with the duration of diabetes [1]. Reports indicate that DR is the most common complication among individuals aged 20–65 years and is the leading cause of vision loss in adults [2, 3]. The condition is attributed to chronic hyperglycemia resulting from abnormal glucose metabolism due to prolonged exposure of the retina to high glucose levels during diabetes. The main pathological features include retinal inflammation, vascular leakage, and neovascularization. Current treatment options for DR involve anti-vascular endothelial growth factor drugs, retinal laser photocoagulation, and vitrectomy; however, these treatments have not shown significant clinical improvement in patients’ vision [4].

The main forms of thyroid dysfunction are hyperthyroidism and hypothyroidism. Clinically, thyroid hormones, including serum free triiodothyronine (FT3), free thyroxine (FT4), and serum thyroid-stimulating hormone (TSH), are measured to assess thyroid function. While research on the relationship between thyroid function and diabetic microvascular complications deepens, the association between thyroid function and diabetic retinopathy (DR) remains unclear. Some studies suggest that thyroid hormones (TH) can be an independent risk factor for diabetic retinopathy, aiding in predicting its development and guiding clinical management [5]. However, in a cross-sectional survey study found no significant relationship between TH levels and the development of DR in contrast to its apparent opposite findings [6, 7]. Similarly, conflicting findings exist, as demonstrated in various clinical observational studies across different populations. For instance, a cross-sectional survey in southern India revealed a higher prevalence of thyroid dysfunction, particularly subclinical hypothyroidism, among patients with type 2 diabetes, which was linked to an increased risk of diabetic retinopathy [8]. Conversely, a clinical retrospective trial in the Caucasus found no significant relationship between TSH levels or subclinical hypothyroidism and diabetic retinopathy [9]. These discrepancies in findings may be influenced by sample size or confounding variables, highlighting the need for larger multi-center randomized controlled trials to validate these results. It is important to note that conducting RCT research poses challenges such as resource constraints, time limitations, and ethical considerations.

Mendelian randomization studies (MR) utilize large-scale genome-wide association (GWAS) data to address limitations of traditional observational experiments and RCTs. By using genetic variation as an instrumental variable (IV), MR can help infer potential causal relationships between exposure factors and outcomes [10]. Genetic variation is inherent and not influenced by confounding social or environmental factors, minimizing bias [11]. MR methods are commonly employed to validate causal relationships identified in observational studies. For instance, Jiang et al. utilized MR to investigate the bidirectional causal link between type 1 diabetes and pulmonary tuberculosis, revealing that type 1 diabetes (T1DM) and high-density lipoprotein (HDL-C) levels are risk factors for pulmonary tuberculosis (PTB). This suggests that managing T1 DM and improving HDL-C levels could play a role in reducing PTB risk, offering valuable insights for clinical practice [12]. According to current relevant research results, the relationship between thyroid dysfunction and DR is unclear. Different studies have produced conflicting results.Moreover,traditional studies have only shown correlation, not causation. Therefore, this study aims to use MR analysis to further explore the relationship between thyroid dysfunction (including hyperthyroidism and hypothyroidism) and DR, investigating the potential bidirectional causal relationship.

Materials and methods

Study design and data sources

In this study, different GWAS data were used to study the causal relationship between thyroid dysfunction and DR by two-sample bidirectional MR study. All study populations were individuals of European ancestry, which can reduce the risk of potential confounding bias due to ethnic differences. The MR analysis method should meet three basic assumptions: ① there is a robust strong correlation between instrumental variables and exposure factors (correlation assumption), ② instrumental variables are independent of the confounding factors that affect the exposure-outcome relationship (independence assumption), ③ genetic variation can only affect outcomes through exposure factors, but cannot affect outcomes through other means (exclusivity assumption) [13]. The flow chart of the specific MR design is shown in Fig. 1.

Fig. 1
figure 1

Directed acyclic diagram of MR Framework to explore the causal relationship between thyroid dysfunction and Diabetic retinopathy. Directed acyclic graph of MR Framework to explore the causal relationship between Diabetic retinopathy and thyroid dysfunction SNP: single nucleotide polymorphism. 1 Assumption of correlation, 2 Assumption of independence, 3 Assumption of exclusivity

Full size image

The single nucleotide polymorphisms (SNP) associated with DR were obtained from the FinnGen biobank database R9 (https://www.finngen.fi/en) including 319,046 European cases (10,413 cases, 308,633 control cases). Genetic association data with hyperthyroidism and hypothyroidism were obtained from the IEU Open GWAS database (https://gwas.mrcieu.ac.uk/). The data set for hyperthyroidism is: ebi-a-GCST90018860, with a total sample size of 460,499 Europeans (3557 cases, 456,942 controls); the data set for hypothyroidism is: ebi-a-GCST90018862, with a total sample size 410,141 Europeans (30,155 cases, 379,986 controls). Detailed information about the phenotype is provided in Table 1.

Table 1 Details of the GWAS included in the Mendelian randomization
Full size table

As this study is based on published publicly available GWAS data, each included GWAS study received ethical approval from their respective institutional review boards, and therefore no additional ethical approval or informed consent was required.

SNPs selection

We assigned SNPs from the GWAS database as IV [14]. To construct genetic IVs, first, we identified SNPs significantly associated with exposure according to strict criteria (P < 5 × 10 −8), while performing SNP removal in linkage disequilibrium (r2 < 0.001, kb = 10,000). We then screened and excluded SNPs associated with confounders using the PhenoScanner database [15, 16]. Body mass index, smoking, or alcohol consumption may affect the risk of hyperthyroidism or hypothyroidism, and hypertension is one of the key risk factors for DR [17,18,19]. Then, we extracted the corresponding SNPs for the outcomes from the GWAS summary statistics and coordinated the two sets of SNPs described above, removing palindromic SNPs with intermediate allele frequencies [20]. Finally, to avoid weak IV bias, we used the F-statistic, with an F-score greater than 10 indicating a lower risk of weak IV bias [21]. The calculation formula is F = (N-2) * R^2/(1-R^2), where R^2 represents the exposure variance explained by the genetic tool (determined by the effect allele frequency (EAF) and the genetic effect of the exposure decision), N represents the sample size [22]. The screening process of instrumental variables is shown in Fig. 2.

Fig. 2
figure 2

A Instrumental variable selection for forward Mendelian randomization analysis. B Instrumental variable selection for reverse Mendelian randomization analysis

Full size image

Statistical analysis

In this study, the inverse variance weighting method (IVW) was used as the main research method, and the MR-Egger, weighted mode, weighted median (WM) and simple mode were used as supplementary methods. IVW is a MR method for meta-summarizing the effects of multiple sites when analyzing multiple SNPs. The premise of the application of IVW is that all SNPs are effective instrumental variables and are completely independent of each other, and have strong causality detection capabilities [23, 24]. Sensitivity analyses, including tests for heterogeneity and level pleiotropy, using Cochran’s Q values to test for heterogeneity in thyroid dysfunction and DR, P < 0.05 were considered to indicate heterogeneity [25]. If heterogeneity is present, we identify and remove heterogeneous IVs from the analysis using the MR-PRESSO. If heterogeneity persists, a random-effects IVW model is used, otherwise a fixed-effect IVW model is used [26]. Furthermore, we assessed the presence of horizontal pleiotropy using the MR-Egger intercept test, with P < 0.05 considered to indicate the presence of horizontal pleiotropy [27]. We finally performed a leave-one-out analysis to visualize the impact of removing individual SNPs on the overall results [28]. And use forest plots, scatter plots, and funnel plots to perform visual analysis and display of corresponding data. In the reverse MR analysis, we used the same analysis method described above.

All of our statistical studies were done under RStudio software and MR analyses were performed using the R packages "TwoSampleMR" and "MR-PRESSO".

Results

Instrumental variable results

After a series of strict screening, in the forward MR analysis, The IVs used in our study were carefully selected based on their significance (P < 5 × 10–8) and independence (r2 < 0.001, kb = 10,000), with a total of 20 and 90 biomarkers chosen respectively. And weak instrumental bias (F > 10) was eliminated for all SNPs. The specific SNP site information is shown in the supplementary Table 1–2. In the reverse MR analysis, we also followed the above strict criteria and screened 8 SNPs as IVs for diabetic retinopathy, as shown in Table 3 of the supplementary table.

Causal association of thyroid dysfunction with DR via forward MR

When hyperthyroidism and hypothyroidism are used as exposure factors, after removing relevant palindromes and heterozygous SNPs, we can obtain 7 and 66 SNPs respectively for final analysis. As shown in Fig. 3, the IVW method showed that the presence of hyperthyroidism and hypothyroidism was associated with an increased risk of diabetic retinopathy (IVW: OR = 1.29, 95% CI 1.12–1.49, P < 0.001; OR = 1.17, 95% CI 1.10–1.25, P < 0.001). When hyperthyroidism was used as an exposure factor, except for the WM, which showed that it was not associated with DR risk (p = 0.063), all the other supplementation methods showed that hyperthyroidism and hypothyroidism were significantly associated with an increased risk of DR. A visual scatter plot of the relationship between hyperthyroidism, hypothyroidism exposure and DR is shown in Fig. 4. The forest diagram is also shown in Fig. 4. In addition, the MR–Egger intercept method (P = 0.093; P = 0.0501) did not detect pleiotropy at the genetic level. However, there was significant heterogeneity, therefore, we chose a random-effects model for final analysis. The funnel plot and leave-one-out plot of the above forward MR analysis are shown in Supplementary Fig. 1–2.

Fig. 3
figure 3

MR estimates for the causal effect of Diabetic retinopathy on thyroid dysfunction. SNP number of single-nucleotide polymorphism, OR odds ratio, CI confidence interval

Full size image
Fig. 4
figure 4

Scatter plot and forest plot of the causal relationship between thyroid dysfunction and Diabetic retinopathy(DR). A Scatter plot of causality between hyperthyroidism and DR. B the susceptibility of hyperthyroidism to the risk of DR; Red dots indicate the combined causal estimates for all SNPs using both MR-Egger and IVW methods. C Scatter plot of causality between hypothyroidism and DR. D the susceptibility of hypothyroidism to the risk of DR; Red dots indicate the combined causal estimates for all SNPs using both MR-Egger and IVW methods

Full size image

Causal association of DR with thyroid dysfunction via reverse MR

When using DR as an exposure factor and removing related palindromes and heterozygous SNPs, we can obtain 5 and 4 SNPs respectively for final analysis. As shown in Fig. 5, the IVW method showed that diabetic retinopathy was associated with an increased risk of hyperthyroidism and hypothyroidism (IVW: OR = 1.56, 95% CI 1.38–1.76, P < 0.001; OR = 1.41, 95% CI 1.25–1.59, P < 0.001). Among the other supplementation methods, except for the MR-Egger method of DR as an exposure factor for hyperthyroidism, which showed no correlation (p = 0.315), the rest showed that DR was associated with an increased risk of hyperthyroidism and hypothyroidism. The visual scatter plot of the relationship between DR exposure and the effects of hyperthyroidism and hypothyroidism is shown in Fig. 6, and the forest diagram is shown in Fig. 6. Furthermore, the MR–Egger intercept method (P = 0.641; P = 0.858) did not detect pleiotropy at the genetic level. However, there was significant heterogeneity in the final analysis of DR and hypothyroidism, so we selected the random effects model for the final analysis, and the fixed effects model for DR and hyperthyroidism. The funnel plot and leave-one-out plot of the above reverse MR analysis are shown in Supplementary Figs. 3–4.

Fig. 5
figure 5

MR estimates for the causal effect of thyroid function on Diabetic retinopathy. nSNP number of single-nucleotide polymorphism; OR odds ratio, CI confidence interval

Full size image
Fig. 6
figure 6

Scatter plot and forest plot of the causal relationship between Diabetic retinopathy (DR) and thyroid dysfunction. A Scatter plot of causality between DR and hyperthyroidism. B the susceptibility of DR to the risk of hyperthyroidism; Red dots indicate the combined causal estimates for all SNPs using both MR-Egger and IVW methods. C Scatter plot of causality between DR and hypothyroidism. D the susceptibility of DR to the risk of hypothyroidism; Red dots indicate the combined causal estimates for all SNPs using both MR-Egger and IVW methods

Full size image

Discussion

This study is the first to investigate a bidirectional causal relationship between thyroid dysfunction and DR using MR analysis. The findings suggest that thyroid dysfunction and DR have a mutual causal relationship, where abnormal thyroid function is linked to a higher risk of DR, and DR is also linked to a higher risk of abnormal thyroid function.

In contrast to conventional observational epidemiological studies, this research minimized the influence of potential confounding variables and reverse causation, providing a more precise examination of the causal link between thyroid dysfunction and diabetic retinopathy (DR). While previous findings have indicated a positive association between thyroid dysfunction and DR, they do not definitively establish a causal connection. A meta-analysis suggested that subclinical hypothyroidism (SCH) could serve as a significant risk factor for DR, with a notable association observed between DR and SCH (OR = 2.13, 95% CI = 1.41–3.23, P < 0.001) [29].Another meta-analysis demonstrated a 2.13-fold increased risk of diabetic retinopathy in diabetic patients with exposure to subclinical hypothyroidism [30]. Furthermore, a study utilizing Mendelian randomization analysis revealed a positive association between hypothyroidism and severe nonproliferative diabetic retinopathy as well as proliferative diabetic retinopathy (β = 8.427943, se = 2.142493, P = 8.36E-05 and β = 3.100939, se = 0.74956, P = 3.52E-05) [31]. Additionally, a case–control study identified a higher risk of subclinical hypothyroidism in type 2 diabetic patients with proliferative diabetic retinopathy [32]. A cross-sectional survey conducted in North India also indicated a high prevalence of subclinical hypothyroidism in type 2 diabetic patients, with retinopathy being a significant contributing factor. These findings align with the results obtained in our Mendelian randomization analysis.

Our study focuses on investigating the association between thyroid dysfunction and diabetic retinopathy (DR), encompassing both hyperthyroidism and hypothyroidism. Currently, the majority of traditional observational studies have explored the link between hypothyroidism and DR, with limited research on hyperthyroidism and DR. While our initial findings using MR analysis have indicated a positive correlation between hyperthyroidism and the development of DR, further extensive observational and clinical studies are required to validate this relationship in future research.

Based on traditional epidemiology and current MR findings, further exploration is needed to understand the mechanisms linking abnormal thyroid function to an increased risk of DR. This association is closely tied to factors such as dysregulation of thyroid hormone homeostasis, inflammation, elevated VEGF expression, and abnormal lipid profiles. Disruption of thyroid hormone homeostasis may contribute to a higher incidence of DR. Thyroid hormones play a crucial role in energy homeostasis, metabolic rate regulation, and are linked to the prognosis of acute and chronic diseases [33]. They impact processes like lipogenesis, lipolysis, gluconeogenesis, glucose handling, insulin resistance, and other key biochemical pathways related to energy distribution in the body. Additionally, thyroid hormones influence the structure and development of the retina, with the hypothalamic-pituitary-thyroid axis influencing retinal vascular density [28, 34, 35]. Studies in rats have shown that induced hypothyroidism leads to reduced levels of the anti-aging enzyme 2 (SIRT-2) protein in the ganglion cell layer of the retina, along with decreased eye size and thinning of retinal layers [35, 36]. Research has also demonstrated the involvement of thyroid homeostasis in changes in optic protein expression in diabetic rats and humans, emphasizing the importance of normal thyroid function for optic protein expression and optic cone cell viability in both rodent and human retinas [37].Chronic inflammation of the retina plays a crucial role in the progression of Diabetic Retinopathy (DR). Czarnywojtek et al. demonstrated that serum CRP levels are elevated in patients with thyroid disorders, including both hyperthyroidism and hypothyroidism [38]. Previous studies have indicated increased VEGF expression in the early stages of DR, with observational data revealing higher VEGF levels in DR patients with subclinical hypothyroidism (SCH) compared to those with normal thyroid function [39]. Moreover, disruptions in thyroid function have been associated with changes in lipid profiles [40], potentially exacerbating the development of DR.

Altered thyroid function, even within the normal range, has been shown to be linked to chronic complications in individuals with type 2 diabetes. A cross-sectional study conducted in Shaanxi Province, China, revealed a negative correlation between normal range FT3 levels and diabetic retinopathy (DR) [41]. Another study indicated that higher TSH levels within the normal range were associated with an increased risk of DR in patients with well-controlled glycemic status [7]. The primary pathological changes in DR are attributed to high glucose-induced mitochondrial apoptosis leading to the loss of peripapillary cells (PCs) [42]. Lin et al. identified functional TSH receptors in peripapillary cells and demonstrated that TSH exacerbates DR by regulating glucose-induced PC loss through TSH receptor (TSHR)-dependent mitochondrial apoptosis. Furthermore, elevated TSH levels have been linked to stenosis of small retinal arteries [43], alterations in optic protein expression, and early retinal changes.

Given the confirmed causal relationship between thyroid hormone (TH) levels and diabetic retinopathy (DR) through the MR analysis, it suggests that TH levels could serve as a predictive factor for DR development. This finding offers a potential clinical foundation for the prevention and management of DR. Therefore, it might be beneficial to consider incorporating routine thyroid function screening in DR patients to potentially discover novel strategies for preventing or tailoring the treatment of DR in clinical settings. Early intervention in DR could significantly benefit more patients. While current MR studies support the causal link between abnormal thyroid function and DR, future interventional studies are necessary to uncover the underlying functional mechanisms. Moreover, extensive prospective trials are still required to validate the prognosis of DR patients with thyroid abnormalities and to establish personalized treatment approaches.

This study is limited to analyzing a dataset from the European population, and further research is necessary to determine if the findings can be generalized to other ethnic groups. The study identified heterogeneity, possibly stemming from variations in SNP data sources, experimental conditions, detection methods, and population inclusion. Notably, thyroid dysfunction was more common in women, yet the study did not stratify the double samples by sex and age to analyze the association between thyroid dysfunction and DR across different demographic groups. While the results were reliable due to stringent instrumental variable screening, the low number of SNPs used raises the need for more detailed and confirmatory studies, as well as in-depth mechanistic investigations in future research.

Conclusion

Our study shows the bidirectional causal relationship between thyroid dysfunction and DR based on MR analysis, which may have a certain impact on management and prevention decisions of clinical public health diseases. However, in the future we still need more basic, large-scale RCT and other studies to further elucidate the mechanism of the association between thyroid dysfunction and DR.

Availability of data and materials

No datasets were generated or analysed during the current study.

References

  1. Li S-Y, Zhao N, Wei D, Pu N, Hao X-N, Huang J-M, Peng G-H, Tao Y. Ferroptosis in the ageing retina: a malevolent fire of diabetic retinopathy. Ageing Res Rev. 2024;93: 102142. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/j.arr.2023.102142.

    Article  CAS  PubMed  Google Scholar 

  2. Wykoff CC, Khurana RN, Nguyen QD, Kelly SP, Lum F, Hall R, Abbass IM, Abolian AM, Stoilov I, To TM, et al. Risk of blindness among patients with diabetes and newly diagnosed diabetic retinopathy. Diabetes Care. 2021;44:748–56. https://doiorg.publicaciones.saludcastillayleon.es/10.2337/dc20-0413.

    Article  PubMed  PubMed Central  Google Scholar 

  3. Levine SR, Sapieha P, Dutta S, Sun JK, Gardner TW. It is time for a moonshot to find “Cures” for diabetic retinal disease. Prog Retin Eye Res. 2022;90: 101051. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/j.preteyeres.2022.101051.

    Article  CAS  PubMed  Google Scholar 

  4. Wang W, Lo ACY. Diabetic retinopathy: pathophysiology and treatments. Int J Mol Sci. 2018;19:1816. https://doiorg.publicaciones.saludcastillayleon.es/10.3390/ijms19061816.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  5. Zhao X, Sun J, Xu X, Xin S, Zhang X. The effect of central and peripheral thyroid resistance indices on diabetic retinopathy: a study of hospitalized euthyroid patients with T2DM in China. Ann Med. 2023;55:2249017. https://doiorg.publicaciones.saludcastillayleon.es/10.1080/07853890.2023.2249017.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  6. Qi Q, Zhang Q-M, Li C-J, Dong R-N, Li J-J, Shi J-Y, Yu D-M, Zhang J-Y. Association of thyroid-stimulating hormone levels with microvascular complications in Type 2 diabetes patients. Med Sci Monit. 2017;23:2715–20. https://doiorg.publicaciones.saludcastillayleon.es/10.12659/msm.902006.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  7. Wang Y, Lu J, Ni J, Wang M, Shen Y, Lu W, Zhu W, Bao Y, Zhou J. Association of thyroid stimulating hormone and time in range with risk of diabetic retinopathy in euthyroid type 2 diabetes. Diabetes Metab Res Rev. 2023;39: e3639. https://doiorg.publicaciones.saludcastillayleon.es/10.1002/dmrr.3639.

    Article  CAS  PubMed  Google Scholar 

  8. Reddy N, Pradeep TVS, Tirupati S, Sarathi V, Kumar D. Thyroid dysfunction and its association with microvascular complications in patients with type 2 diabetes mellitus in south India. Diabetes Metab Syndr. 2020;14:615–7. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/j.dsx.2020.05.005.

    Article  PubMed  Google Scholar 

  9. Ramis JN, Artigas CF, Santiago MAA, Mañes FJV, Canonge RS, Comas LM. Is there a relationship between TSH levels and diabetic retinopathy in the Caucasian population? Diabetes Res Clin Pract. 2012;97:e45-47. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/j.diabres.2012.05.015.

    Article  CAS  PubMed  Google Scholar 

  10. Grant AJ, Burgess S. A Bayesian approach to Mendelian randomization using summary statistics in the univariable and multivariable settings with correlated pleiotropy. Am J Hum Genet. 2024;111:165–80. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/j.ajhg.2023.12.002.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  11. Ebrahim S, Davey SG. Mendelian randomization: can genetic epidemiology help redress the failures of observational epidemiology? Hum Genet. 2008;123:15–33. https://doiorg.publicaciones.saludcastillayleon.es/10.1007/s00439-007-0448-6.

    Article  PubMed  Google Scholar 

  12. Jiang Y, Zhang W, Wei M, Yin D, Tang Y, Jia W, Wang C, Guo J, Li A, Gong Y. Associations between type 1 diabetes and pulmonary tuberculosis: a bidirectional mendelian randomization study. Diabetol Metab Syndr. 2024;16:60. https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s13098-024-01296-x.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  13. Emdin CA, Khera AV, Kathiresan S. Mendelian randomization. JAMA. 2017;318:1925–6. https://doiorg.publicaciones.saludcastillayleon.es/10.1001/jama.2017.17219.

    Article  PubMed  Google Scholar 

  14. Lawlor DA, Harbord RM, Sterne JAC, Timpson N, Davey SG. Mendelian randomization: using genes as instruments for making causal inferences in epidemiology. Stat Med. 2008;27:1133–63. https://doiorg.publicaciones.saludcastillayleon.es/10.1002/sim.3034.

    Article  PubMed  Google Scholar 

  15. Kamat MA, Blackshaw JA, Young R, Surendran P, Burgess S, Danesh J, Butterworth AS, Staley JR. PhenoScanner V2: an expanded tool for searching human genotype-phenotype associations. Bioinformatics. 2019;35:4851–3. https://doiorg.publicaciones.saludcastillayleon.es/10.1093/bioinformatics/btz469.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  16. Staley JR, Blackshaw J, Kamat MA, Ellis S, Surendran P, Sun BB, Paul DS, Freitag D, Burgess S, Danesh J, et al. PhenoScanner: a database of human genotype-phenotype associations. Bioinformatics. 2016;32:3207–9. https://doiorg.publicaciones.saludcastillayleon.es/10.1093/bioinformatics/btw373.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  17. Andersen SL, Olsen J, Wu CS, Laurberg P. Smoking reduces the risk of hypothyroidism and increases the risk of hyperthyroidism: evidence from 450,842 mothers giving birth in Denmark. Clin Endocrinol (Oxf). 2014;80:307–14. https://doiorg.publicaciones.saludcastillayleon.es/10.1111/cen.12279.

    Article  PubMed  Google Scholar 

  18. Chen X, Wang J-J, Yu L, Wang H-Y, Sun H. The association between BMI, smoking, drinking and thyroid disease: a cross-sectional study in Wuhan. China BMC Endocr Disord. 2021;21:184. https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12902-021-00852-0.

    Article  CAS  PubMed  Google Scholar 

  19. Klein R, Sharrett AR, Klein BEK, Moss SE, Folsom AR, Wong TY, Brancati FL, Hubbard LD, Couper D. The association of atherosclerosis, vascular risk factors, and retinopathy in adults with diabetes: the atherosclerosis risk in communities study. Ophthalmology. 2002;109:1225–34. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/s0161-6420(02)01074-6.

    Article  PubMed  Google Scholar 

  20. Wang Z, Chen M, Wei Y-Z, Zhuo C-G, Xu H-F, Li W-D, Ma L. The causal relationship between sleep traits and the risk of schizophrenia: a two-sample bidirectional Mendelian randomization study. BMC Psychiatry. 2022;22:399. https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12888-022-03946-8.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  21. Pierce BL, Ahsan H, Vanderweele TJ. Power and instrument strength requirements for Mendelian randomization studies using multiple genetic variants. Int J Epidemiol. 2011;40:740–52. https://doiorg.publicaciones.saludcastillayleon.es/10.1093/ije/dyq151.

    Article  PubMed  Google Scholar 

  22. Zhang Y, Xiong Y, Shen S, Yang J, Wang W, Wu T, Chen L, Yu Q, Zuo H, Wang X, et al. Causal association between tea consumption and kidney function: a mendelian randomization study. Front Nutr. 2022;9: 801591. https://doiorg.publicaciones.saludcastillayleon.es/10.3389/fnut.2022.801591.

    Article  PubMed  PubMed Central  Google Scholar 

  23. Slob EAW, Burgess S. A comparison of robust Mendelian randomization methods using summary data. Genet Epidemiol. 2020;44:313–29. https://doiorg.publicaciones.saludcastillayleon.es/10.1002/gepi.22295.

    Article  PubMed  PubMed Central  Google Scholar 

  24. Burgess S, Davey Smith G, Davies NM, Dudbridge F, Gill D, Glymour MM, Hartwig FP, Kutalik Z, Holmes MV, Minelli C, et al. Guidelines for performing Mendelian randomization investigations: update for summer 2023. Wellcome Open Res. 2019;4:186. https://doiorg.publicaciones.saludcastillayleon.es/10.12688/wellcomeopenres.15555.3.

    Article  PubMed  Google Scholar 

  25. Greco MFD, Minelli C, Sheehan NA, Thompson JR. Detecting pleiotropy in Mendelian randomisation studies with summary data and a continuous outcome. Stat Med. 2015;34:2926–40. https://doiorg.publicaciones.saludcastillayleon.es/10.1002/sim.6522.

    Article  Google Scholar 

  26. Hemani G, Bowden J, Davey SG. Evaluating the potential role of pleiotropy in Mendelian randomization studies. Hum Mol Genet. 2018;27:R195–208. https://doiorg.publicaciones.saludcastillayleon.es/10.1093/hmg/ddy163.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  27. Hartwig FP, Davey Smith G, Bowden J. Robust inference in summary data Mendelian randomization via the zero modal pleiotropy assumption. Int J Epidemiol. 2017;46:1985–98. https://doiorg.publicaciones.saludcastillayleon.es/10.1093/ije/dyx102.

    Article  PubMed  PubMed Central  Google Scholar 

  28. Mendoza A, Hollenberg AN. New insights into thyroid hormone action. Pharmacol Ther. 2017;173:135–45. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/j.pharmthera.2017.02.012.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  29. Wu J, Yue S, Geng J, Liu L, Teng W, Liu L, Chen L. Relationship between diabetic retinopathy and subclinical hypothyroidism: a meta-analysis. Sci Rep. 2015;5:12212. https://doiorg.publicaciones.saludcastillayleon.es/10.1038/srep12212.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  30. Yang J-K, Liu W, Shi J, Li Y-B. An association between subclinical hypothyroidism and sight-threatening diabetic retinopathy in type 2 diabetic patients. Diabetes Care. 2010;33:1018–20. https://doiorg.publicaciones.saludcastillayleon.es/10.2337/dc09-1784.

    Article  PubMed  PubMed Central  Google Scholar 

  31. Fang T, Deng X, Wang J, Han F, Liu X, Liu Y, Sun B, Chen L. The effect of hypothyroidism on the risk of diabetes and its microvascular complications: a Mendelian randomization study. Front Endocrinol (Lausanne). 2023;14:1288284. https://doiorg.publicaciones.saludcastillayleon.es/10.3389/fendo.2023.1288284.

    Article  PubMed  Google Scholar 

  32. Yang G-R, Yang J-K, Zhang L, An Y-H, Lu J-K. Association between subclinical hypothyroidism and proliferative diabetic retinopathy in type 2 diabetic patients: a case-control study. Tohoku J Exp Med. 2010;222:303–10. https://doiorg.publicaciones.saludcastillayleon.es/10.1620/tjem.222.303.

    Article  PubMed  Google Scholar 

  33. Iwen KA, Schröder E, Brabant G. Thyroid hormones and the metabolic syndrome. Eur Thyroid J. 2013;2:83–92. https://doiorg.publicaciones.saludcastillayleon.es/10.1159/000351249.

    Article  PubMed  PubMed Central  Google Scholar 

  34. Martino E, Seo H, Lernmark A, Refetoff S. Ontogenetic patterns of thyrotropin-releasing hormone-like material in rat hypothalamus, pancreas, and retina: selective effect of light deprivation. Proc Natl Acad Sci U S A. 1980;77:4345–8. https://doiorg.publicaciones.saludcastillayleon.es/10.1073/pnas.77.7.4345.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  35. Kocaturk T, Ergin K, Cesur G, Evlicoglu GE, Cakmak H. The effect of methimazole-induced postnatal hypothyroidism on the retinal maturation and on the Sirtuin 2 level. Cutan Ocul Toxicol. 2016;35:36–40. https://doiorg.publicaciones.saludcastillayleon.es/10.3109/15569527.2015.1007509.

    Article  CAS  PubMed  Google Scholar 

  36. Pinazo-Durán MD, Pons-Vázquez S, Gallego-Pinazo R, Galbis Estrada C, Zanón-Moreno V, Vila Bou V, Sanz SP. Thyroid hormone deficiency disrupts rat eye neurodevelopment. Brain Res. 2011;1392:16–26. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/j.brainres.2011.04.005.

    Article  CAS  PubMed  Google Scholar 

  37. Énzsöly A, Hajdú RI, Turóczi Z, Szalai I, Tátrai E, Pálya F, Nagy ZZ, Mátyás C, Oláh A, Radovits T, et al. The predictive role of thyroid hormone levels for early diabetic retinal changes in experimental rat and human diabetes. Invest Ophthalmol Vis Sci. 2021;62:20. https://doiorg.publicaciones.saludcastillayleon.es/10.1167/iovs.62.6.20.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  38. Czarnywojtek A, Owecki M, Zgorzalewicz-Stachowiak M, Woliński K, Szczepanek-Parulska E, Budny B, Florek E, Waligórska-Stachura J, Miechowicz I, Bączyk M, et al. The role of serum C-reactive protein measured by high-sensitive method in thyroid disease. Arch Immunol Ther Exp (Warsz). 2014;62:501–9. https://doiorg.publicaciones.saludcastillayleon.es/10.1007/s00005-014-0282-1.

    Article  CAS  PubMed  Google Scholar 

  39. El-Sehrawy AA, Elkhamisy EM, Badawi AE, Elshahawy HA, Elsayed E, Mohammed NT, El-Eshmawy MM. Subclinical hypothyroidism in patients with diabetic retinopathy: role of vascular endothelial growth factor. Endocr Metab Immune Disord Drug Targets. 2022;22:502–9. https://doiorg.publicaciones.saludcastillayleon.es/10.2174/1871530321666210809151258.

    Article  CAS  PubMed  Google Scholar 

  40. Rizos CV, Elisaf MS, Liberopoulos EN. Effects of thyroid dysfunction on lipid profile. Open Cardiovasc Med J. 2011;5:76–84. https://doiorg.publicaciones.saludcastillayleon.es/10.2174/1874192401105010076.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  41. Zou J, Li Z, Tian F, Zhang Y, Xu C, Zhai J, Shi M, Wu G, Zhang Z, Yang C, et al. Association between normal thyroid hormones and diabetic retinopathy in patients with type 2 diabetes. Biomed Res Int. 2020;2020:8161797. https://doiorg.publicaciones.saludcastillayleon.es/10.1155/2020/8161797.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  42. Lin D, Qin R, Guo L. Thyroid stimulating hormone aggravates diabetic retinopathy through the mitochondrial apoptotic pathway. J Cell Physiol. 2022;237:868–80. https://doiorg.publicaciones.saludcastillayleon.es/10.1002/jcp.30563.

    Article  CAS  PubMed  Google Scholar 

  43. Ittermann T, Dörr M, Völzke H, Tost F, Lehmphul I, Köhrle J, Jürgens C. High serum thyrotropin levels are associated with retinal arteriolar narrowing in the general population. Thyroid. 2014;24:1473–8. https://doiorg.publicaciones.saludcastillayleon.es/10.1089/thy.2014.0190.

    Article  CAS  PubMed  Google Scholar 

Download references

Acknowledgements

We would like to thank the IEU Open GWAS Project Database (https://gwas.mrcieu.ac.uk/) and the FinnGen Biobank Database (https://www.finngen.fi/en) for providing GWAS data.

Funding

This work was supported by the Jiangxi Provincial Administration of Traditional Chinese Medicine Key Research Laboratory on the Fundamentals of Chinese Medicine Evidence (Gan TCM Science and Education Word [2022] No.8–4),the Jiangxi Province Traditional Chinese Medicine Young Chinese Backbone Talent Training Project (Gan TCM Comprehensive Word [2020] No.9) and Jiangxi University of Traditional Chinese Medicine School-level Science and Technology Innovation Team (CXTD22016).

Author information

Authors and Affiliations

  1. School of Chinese Medicine, Jiangxi University of Chinese Medicine, Nanchang, 330000, Jiangxi, China

    Jiali Chen, Fenfen Zhang, Wanyu Pan & Shaomin Cheng

  2. Department of Respiratory and Critical Care Medicine, Jiangxi Provincial Hospital of Integrated Traditional Chinese and Western Medicine, Nanchang, 330000, Jiangxi, China

    Jianghao Xiong

Authors
  1. Jiali Chen
    View author publications

    You can also search for this author inPubMed Google Scholar

  2. Jianghao Xiong
    View author publications

    You can also search for this author inPubMed Google Scholar

  3. Fenfen Zhang
    View author publications

    You can also search for this author inPubMed Google Scholar

  4. Wanyu Pan
    View author publications

    You can also search for this author inPubMed Google Scholar

  5. Shaomin Cheng
    View author publications

    You can also search for this author inPubMed Google Scholar

Contributions

C.J, C.S conceived and designed the manuscript. C.J, X.J. and C.S. wrote the manuscript. C.J, Z.F. and P.W. collected and analyzed the references. X.J, Z.F. and P.W. prepared figures and tables. C.J, X.J. and C.S. checked, proofread, and polished the manuscript. All authors reviewed the manuscript.

Corresponding author

Correspondence to Shaomin Cheng.

Ethics declarations

Ethics approval and consent to participate

Our study is based on published publicly available GWAS data, and each GWAS study included has received ethical approval from the respective institutional review board, so no additional ethical approval or informed consent is required.

Competing interests

The authors declare no competing interests.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

13098_2024_1552_MOESM1_ESM.doc

Supplementary material 1. Table 1 Characteristics of SNPs associated with hyperthyroidism. Table 2 | Characteristics of SNPs associated with hypothyroidism. Table 3 | Characteristics of SNPs associated with Diabetic retinopathy. Figure 1 The Leave-one-out sensitivity analysis for causal effect of thyroid function on DR.hyperthyroidism on DR,hypothyroidism on RA. Figure 2 The funnel plot of individual SNP effects of thyroid function on DR.hyperthyroidism on DR,hypothyroidism on RA. Figure 3 The Leave-one-out sensitivity analysis for causal effect of DR on thyroid function.DR on hyperthyroidism,DR on hypothyroidism. Figure 4 The funnel plot of individual SNP effects of DR on thyroid function.DR on hyperthyroidism,DR on hypothyroidism

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Chen, J., Xiong, J., Zhang, F. et al. Association between thyroid dysfunction and diabetic retinopathy: a two-sample bidirectional Mendelian randomization study. Diabetol Metab Syndr 16, 297 (2024). https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s13098-024-01552-0

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s13098-024-01552-0

Keywords

Diabetology & Metabolic Syndrome

ISSN: 1758-5996

Contact us