Causal Effects of Dietary Macronutrients and Micronutrients on Myopia and Other Major Ocular Disorders: A Mendelian Randomization Study

Summary

This protocol describes a method for investigating the causal effects of dietary nutrients on ocular diseases using Mendelian randomization analysis. This approach enables the identification of potential nutritional risk and protective factors for major eye disorders.

Abstract

Myopia and other major ocular disorders, such as cataracts, glaucoma, and age-related macular degeneration (AMD), pose growing public health concerns worldwide. While nutrition is believed to influence ocular development and degeneration, causal evidence remains limited. A two-sample Mendelian randomization (MR) analysis was performed using large-scale genome-wide association study (GWAS) summary data to evaluate the causal effects of dietary macronutrient and micronutrient intake on the risk of myopia and other ocular diseases. The inverse variance-weighted (IVW) method was used as the primary analysis, supported by weighted median, MR-Egger, and MR-PRESSO for sensitivity testing. To account for multiple comparisons, false discovery rate (FDR) correction was applied. After FDR adjustment, higher genetically predicted vitamin C levels remained significantly associated with an increased risk of AMD (odds ratio [OR] = 1.324, 95% confidence interval [CI]: 1.133–1.546). Several additional associations were observed at nominal significance levels (P < 0.05), including higher relative dietary fat intake with increased myopia risk (OR = 1.033, 95% CI: 1.003–1.063), sugar intake with reduced cataract risk (OR = 0.746, 95% CI: 0.561–0.993), iron with lower myopia risk (OR = 0.995, 95% CI: 0.991–0.999), vitamin D with increased cataract risk (OR = 1.256, 95% CI: 1.062–1.487), folic acid with reduced cataract risk (OR = 0.843, 95% CI: 0.724–0.981), and phosphorus with reduced glaucoma risk (OR = 0.727, 95% CI: 0.554–0.953). No evidence of horizontal pleiotropy or heterogeneity was observed. These findings suggest a potential causal role of certain nutrients in ocular disease risk, although most associations did not remain significant after multiple testing correction and should therefore be interpreted with caution. Further studies are needed to confirm these findings and explore underlying biological mechanisms.

Introduction

Visual impairment caused by ocular disorders represents a growing global public health challenge, with significant socioeconomic and quality-of-life implications^1,2,3. Among these conditions, refractive error has emerged as a leading cause of vision loss, especially in East and Southeast Asia, where its prevalence among young adults reaches 80–90%, and high myopia affects up to 20% of the population⁴. More recent epidemiological evidence suggests that myopia prevalence continues to increase globally, and newer studies have highlighted the contribution of modern lifestyle changes, particularly increased digital screen exposure and reduced outdoor activity, to this trend⁵. Projections estimate that by 2050, nearly half of the global population could be myopic, with approximately 10% experiencing high myopia and its associated vision-threatening complications⁴. Beyond myopia, other major ocular disorders, such as age-related macular degeneration (AMD), cataracts, and glaucoma, also contribute substantially to global vision loss^6,7,8. Given the absence of universally effective interventions to halt vision loss progression, the identification of modifiable environmental and lifestyle factors is crucial for developing effective prevention strategies.

These conditions share complex etiologies involving both genetic predispositions and environmental exposures. Emerging evidence suggests that nutrition may play a critical role in ocular development, retinal metabolism, oxidative stress modulation, and the maintenance of structural integrity across various eye tissues^9,10,11. Several micronutrients, including vitamin A, vitamin D, zinc, and iron, have been implicated in maintaining ocular health by supporting photoreceptor function, vascular regulation, and antioxidant defense^12,13,14,15. For example, vitamin D has been suggested to influence ocular growth and inflammatory pathways, iron plays an essential role in retinal metabolism and oxidative stress regulation, and zinc participates in antioxidant defense and retinal pigment epithelium function. In addition, dietary macronutrients such as fats and carbohydrates may affect ocular health through systemic metabolic pathways, including lipid metabolism, insulin signaling, and inflammatory responses^16,17. However, most previous studies investigating these nutrients have relied on observational designs, which are susceptible to confounding and reverse causality, making it difficult to determine whether these associations are causal.

However, findings from observational studies examining the relationship between dietary intake and ocular diseases have been inconsistent, with some reporting significant associations and others finding null results^17,18,19. These discrepancies likely reflect limitations such as small sample sizes, residual confounding, recall bias in dietary assessments, and the potential for reverse causality. To address these gaps, Mendelian randomization (MR) offers a powerful epidemiologic tool by using genetic variants as instrumental variables to infer causality between exposures and outcomes²⁰. Because genetic variants are randomly assorted at conception and largely unaffected by environmental or behavioral confounders, MR reduces bias and allows for more credible causal inference than traditional observational designs²¹. Recent large-scale genome-wide association studies (GWAS) have identified genetic instruments for a wide range of dietary macronutrients and micronutrients^22,23, as well as for ocular phenotypes including myopia, AMD, cataracts, and glaucoma, enabling a comprehensive MR-based assessment of diet-ocular disorder relationships.

Therefore, the primary aim of this study was to systematically evaluate the potential causal effects of dietary macronutrient and micronutrient intake on the risk of major ocular diseases using a two-sample Mendelian randomization framework. Specifically, the study investigated whether genetically predicted levels of dietary macronutrients (such as fats and carbohydrates) and micronutrients (including vitamins and minerals) were associated with the risk of myopia, cataracts, glaucoma, and age-related macular degeneration (AMD). The findings provide novel evidence to inform future dietary recommendations and preventive strategies targeting eye health.

Protocol

The Mendelian Randomization (MR) analyses utilized data from a publicly accessible GWAS (https://gwas.mrcieu.ac.uk/), which had already obtained the necessary ethical and informed consent approvals during the initial data collection. Consequently, no further ethical or informed consent was required for the study. The software and the databases used are listed in the Table of Materials.

1. Exposure data

GWAS summary information on relative dietary macronutrient intake of fat, protein, carbohydrates, and sugar was obtained from over 235,000 individuals of European ancestries²³. Dietary intake of the previous day (UKB) or habitual (all other cohorts) was captured using comprehensive food item questionnaires. The relative intake of each macronutrient was corrected by the total energy intake, allowing for nonlinear effects. Following stringent quality control procedures, the final aggregated sample size is 235,391 for sugar and 268,922 for fat, protein, and carbohydrate²³. The basic characteristics of the participants included in the GWAS of relative dietary macronutrient intake are summarized in Table 1.

Genetic association data for micronutrients were sourced from a comprehensive review of the existing literature and the availability of robust GWASs in European populations. Briefly, GWAS summary statistics were included for seven essential minerals, serum calcium, copper, iron, magnesium, phosphorus, selenium (measured in blood or toenail samples), and zinc^{24,25,26,27,28,29,30,31}, as well as seven vitamins, including 25-hydroxyvitamin D, vitamin A1 (retinol), vitamin B6, vitamin B9 (folic acid), vitamin B12, vitamin C, and vitamin E^{32,33,34,35,36,37}. The basic characteristics of the participants included in the GWAS of micronutrients are summarized in Table 1.

2. Outcome data

The outcome data were obtained from the IEU GWAS Project (https://gwas.mrcieu.ac.uk/), developed at the MRC Integrative Epidemiology Unit at the University of Bristol. Specifically, myopia was examined as the outcome using data from source: ukb-b-6353 (n = 460,536), cataracts from source: ebi-a-GCST90018814 (n = 491,877), glaucoma from source: ebi-a-GCST90013865 (n = 406,927), and early age-related macular degeneration from source: ebi-a-GCST010723 (n = 105,248). The basic characteristics of the participants included in the GWAS of myopia and other major ocular disorders are summarized in Table 2. The definitions of ocular disease phenotypes were based on the original GWAS datasets. Myopia (ukb-b-6353) was derived from the UK Biobank and defined based on self-reported use of glasses or contact lenses mainly for distance viewing (short-sightedness). Cataract and glaucoma outcomes were obtained from large GWAS meta-analyses based on clinically diagnosed cases identified through medical records or hospital episode statistics. Early age-related macular degeneration (AMD) was defined according to standardized diagnostic criteria used in the original GWAS consortium studies. Differences in phenotype definitions across studies may introduce some heterogeneity and should be considered when interpreting the results.

3. Selection of instrumental variables

Single-nucleotide polymorphisms (SNPs) associated with exposures at genome-wide significance (P < 5 × 10⁻⁸) were selected as instrumental variables. Independent SNPs were selected as instrumental variables with an r2 threshold of < 0.001, based on their distance within 10,000 kb in the 1000G European data, using the “TwoSampleMR” R package. The strength of the selected instruments was assessed using the F statistic³⁸. An F statistic greater than 10 is generally considered indicative of strong instrumental variables³⁹. All analyses were harmonized to ensure alignment of effect alleles between exposure and outcome datasets.

4. Mendelian randomization estimates

To estimate the causal relationship between dietary macronutrients and micronutrients and myopia and other major ocular disorders, the “TwoSampleMR” R package was applied for MR analysis. Genetic variants associated with each of the dietary macronutrients and micronutrients were used as instrumental variables to examine their causal relationship with myopia and other major ocular disorders. MR analysis relies on three key assumptions: (1) the relevance assumption, whereby the genetic variants are strongly associated with the exposure; (2) the independence assumption, whereby the variants are independent of confounders; and (3) the exclusion restriction assumption, whereby the variants influence the outcome only through the exposure of interest. Multiple MR methods were implemented, including inverse variance weighted (IVW), weighted median, weighted mode, and MR-Egger, with IVW serving as the primary approach and the other methods used to assess the robustness of the findings. The IVW method provides the most precise causal estimate when all instrumental variables are valid or when horizontal pleiotropy is balanced across variants. The weighted median estimator can generate consistent estimates even if up to 50% of the genetic instruments are invalid. The MR-Egger regression method allows for the detection of directional pleiotropy by estimating an intercept term that reflects the average pleiotropic effect across genetic variants. Initially, causal estimates were calculated using fixed effects IVW models, and random effects IVW models were employed when significant heterogeneity was detected (p < 0.05). To account for multiple comparisons across multiple exposures and outcomes, false discovery rate (FDR) correction using the Benjamini–Hochberg method was applied. Statistical significance was defined as FDR-adjusted P < 0.05, while P < 0.05 was considered nominally significant. All analyses were performed in R (version 4.1.2). To clarify the analytical framework, a schematic representation of the MR design and assumptions is shown in Figure 1.

5. Heterogeneity and sensitivity analyses

To evaluate the robustness of the MR results, heterogeneity tests were conducted to detect potential inconsistencies among the instrumental variables. Heterogeneity was assessed using the Cochran Q statistic, where a P value of < 0.05 indicated significant heterogeneity⁴⁰. Sensitivity analyses were further performed to assess potential pleiotropy, where genetic variants may influence the outcome through biological pathways other than the exposure of interest. Directional pleiotropy was evaluated using the MR-Egger regression intercept test⁴¹. In addition, the Mendelian Randomization Pleiotropy RESidual Sum and Outlier (MR-PRESSO) method was applied to identify and remove potential pleiotropic outliers (p < 0.05) and to provide corrected causal estimates when necessary⁴¹.

Results

The significant causal effects of dietary macronutrient and micronutrient intake on myopia and other major ocular disorders, as identified by the IVW method, are summarized in Table 3 and presented in Figure 2. To account for multiple comparisons across multiple exposures and outcomes, FDR correction was applied. After FDR adjustment, only the association between genetically predicted vitamin C levels and increased risk of early AMD remained statistically significant (OR = 1...

Discussion

In this study, the potential causal effects of dietary macronutrient and micronutrient intake on the risk of myopia and other major ocular disorders, including cataracts, glaucoma, and AMD, were systematically evaluated. After correcting for multiple comparisons using the false discovery rate (FDR) method, only the association between genetically predicted vitamin C levels and increased risk of AMD remained statistically significant. Several additional associations were observed at nominal significance levels, including ...

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
IEU GWAS database	MRC Integrative Epidemiology Unit, University of Bristol	www.gwas.mrcieu.ac.uk	Source of GWAS summary statistics
MR-PRESSO R package	GitHub (MR-PRESSO developers)	www.github.com/rondolab/MR-PRESSO	Used for detecting pleiotropy and outliers
R software (version 4.1.2)	The R Foundation	www.r-project.org	Used for all statistical analyses and Mendelian randomization
TwoSampleMR R package	MRC Integrative Epidemiology Unit	www.mrcieu.github.io/TwoSampleMR	Used for Mendelian randomization analysis