Genetic Influence of Oily Fish Intake on Age-Related Macular Degeneration Risk: A Two-Sample Mendelian Randomization Analysis

Abstract

Purpose: Emerging research indicates a link between the intake of fatty fish and age-related macular degeneration (AMD). However, observational studies fall short in establishing a direct causal link between oily fish intake and AMD. We wanted to determine whether causal association lies between oily fish intake and age-related macular degeneration (AMD) risk in human beings.

Methods: This two-sample mendelian randomization (MR) study used the MR method to probe the genetic causality in the relationship between oily fish intake and AMD. The genome-wide association study (GWAS) data for AMD were acquired from a Finnish database, whereas the data on fish oil intake came from the UK Biobank. The analysis used several approaches such as inverse-variance weighted (IVW), MR Egger, weighted median, simple mode, and weighted mode MR. In addition, the Cochran's Q test was used to evaluate heterogeneity in the MR data. The MR-Egger intercept and MR–pleiotropy residual sum and outlier (MR-PRESSO) tests were used to assess the presence of horizontal pleiotropy. A leave-one-out sensitivity analysis was conducted to determine the reliability of the association.

Results: The IVW method revealed that the intake of oily fish is an independent risk factor for AMD (P = 0.034). It also suggested a minimal likelihood of horizontal pleiotropy affecting the causality (P > 0.05), with no substantial heterogeneity detected in the genetic variants (P > 0.05). The leave-one-out analysis confirmed the reliability and stability of this correlation.

Conclusions: This research used a two-sample MR analysis to provide evidence of a genetic causal relationship between the eating of oily fish and AMD. This discovery held potential significance in AMD prevention.

Introduction

Age-related macular degeneration (AMD) represents a leading cause of blindness among the elderly, with projections indicating an escalation to 288 million AMD patients globally by 2040.¹ AMD can cause severe vision loss and vision distortion and manifests in three stages: early, intermediate, and advanced.² Advanced AMD can cause severe vision loss and vision distortion, including two forms: advanced dry AMD (geographic atrophy [GA]) and advanced wet AMD (mainly includes neovascular AMD [NV-AMD]), with NV-AMD arising amid degenerative changes.^3,4 There is no treatment currently available for GA, recent advancements in NV-AMD treatment notwithstanding, and the condition continues to be a significant contributor to visual impairment. The pathogenesis of AMD is multifaceted, encompassing age, genetic predisposition, smoking, and dietary factors as primary risk elements.^5,6 

Recent studies increasingly indicate a potential link between diet, particularly oily fish consumption, and a lower AMD risk.^7–9 Oily fish, which contains high levels of ω-3 long-chain unsaturated fatty acids and polyunsaturated fatty acids (PUFA), might potentially protect against AMD by virtue of its anti-inflammatory and antioxidant characteristics.^10,11 Nonetheless, existing research, predominantly observational, fails to adequately address confounding variables and the issue of reverse causation. 

To address these limitations, Our work used Mendelian randomization (MR), using genetic variants as instrumental factors to examine the potential causal association between the diet of oily fish and AMD. MR is a method that addresses the difficulties posed by unmeasured confounding variables and reverse causation, which are common in observational epidemiology. MR use genetic differences to overcome these obstacles,^12,13 such as single nucleotide polymorphisms (SNPs), related to a specific exposure (here, oily fish consumption). These genetic markers serve as proxies to evaluate the impact of this exposure on the health outcome of interest (AMD, in this context). The goal of this study is to furnish more definitive evidence on the connection between oily fish consumption and AMD risk, thereby informing potential nutritional interventions for AMD prevention across all stages of the disease. 

Methods

Study Design and Data Sources

This investigation uses a two-sample MR approach to assess the causal relationship between oily fish intake and AMD, encompassing both its dry and wet forms. Unlike single-sample MR, two-sample MR uses independent data from distinct genome-wide association studies (GWAS), enhancing effectiveness and reliability. In this context, oily fish intake is the exposure variable, and AMD is the outcome variable. Instrumental variables (IVs) are chosen to be SNPs. Our study adheres to the three fundamental assumptions of two-sample MR: (1) The chosen IVs are strongly associated with the exposure variable (P < 5 × 10⁻⁸, F statistic > 10); (2) They are not associated with any factors that might possibly bias the exposure findings; and (3) Disregarding all other factors, the outcome is only influenced by the exposure variable. 

The UK Biobank has GWAS summary data on oily fish consumption, including 460,443 cases, whereas the FinnGen Consortium provides data specifically on AMD (https://r7.finngen.fi/), including 3763 cases and 205,359 controls. All participants have European ancestry. 

Selection of Instrumental Variables

In MR analysis, strict commitment to the principles of relevance, independence, and exclusion is of utmost importance. Consequently, IVs undergo stringent screening. Initially, SNPs with strong associations to the exposure variable (P < 5 × 10⁻⁸) are selected, excluding those with an F-value < 10 to mitigate weak IV bias. The F-value is calculated using the formula: F = R² × (N − 2)/(1 – R²), where R² represents the variance proportion in the exposure variable explained by each IV, calculated as R² = 2 × EAF × (1 − EAF) × Beta². Beta represents the value of the effect allele (EAF). Furthermore, to counteract potential bias from high linkage disequilibrium among SNPs, clumping is performed (r² < 0.001, physical distance window = 10,000 kb), ensuring IV independence. SNPs linked to confounding factors and the outcome variable are excluded using Phenoscanner V2, focusing on confounders such as diabetes, cardiovascular diseases, and unhealthy lifestyle habits. This step guarantees that IVs impact the outcome exclusively via the exposure variable. Figure 1 illustrates permissible direct connections (there are straight lines and arrows connecting points A and B) versus impermissible indirect connections (there are dashed lines and arrows connecting points C and D). Additionally, datasets for the exposure and outcome variables are harmonized, excluding palindromic SNPs with allele frequencies that fall within the intermediate range. 

Figure 1.

 

The procedure of MR analysis. The SNPs that showed a significant association with oily fish consumption were considered as IVs in this figure. Only the chosen SNPs have the potential to influence the outcomes via exposure (shown by the solid line and arrows “A” and “B”). None of the instrumental variables (IVs) should have a direct impact on the outcomes or be influenced by confounding factors (represented by the dashed line and arrows from “C” and “D” are prohibited).

Statistical Analysis

Five methodologies were used to examine the genetic correlation between the ingestion of oily fish and AMD: The statistical methods used include MR-Egger regression, weighted median, inverse-variance weighted (IVW), simple mode, and weighted mode. The IVW method, presuming all SNPs are valid, is considered the primary analysis method because of its accuracy in effect estimation. The statistical power of the investigation was computed via the website https://shiny.cnsgenomics.com/mRnd/, which uses the non-central parameter method. To do this computation, it is necessary to determine the overall variance of the exposure variable that is accounted for by all IVs, R², as previously defined. To verify the results’ reliability, several tests were conducted. Cochrane's Q test assessed heterogeneity in MR findings, complemented by funnel plots examining symmetry. The MR-Egger intercept test and MR-PRESSO global test identified the presence of multicollinearity among the genetic factors, with MR-PRESSO also identifying outliers and providing adjusted effect estimations after their removal. A stepwise removal test evaluated the sensitivity of the results, assessing the impact of each SNP removal. The statistical analyses were performed using the TwoSampleMR package in R software (version 4.2.0), with a significance level set at P < 0.05. 

Results

Selection of Instrumental Variables

In the process of selecting instrumental variables (IVs), 63 SNPs were identified as preliminary candidates. These SNPs were closely associated with the exposure factor (P < 5 × 10⁻⁸, F-value > 10) and exhibited independence (r² < 0.001, physical distance = 10,000 kb), with the lowest F-value recorded at 45.61. The Table provides comprehensive details of these F-values. Despite using Phenoscanner V2 to screen out SNPs related to outcomes and confounding factors (P < 1 × 10⁻⁵), no SNPs were excluded at this phase. After aligning the exposure and outcome data, 63 SNPs were shortlisted for MR analysis. After that, there were two SNPs with allele frequencies that are moderate and exhibit palindromic patterns, rs34555420 and rs11986122, that were removed, resulting in a final count of 61 SNPs for the MR analysis. Association of genetic instruments with oil fish were shown in the Supplementary Table. 

Table.

 

MR Results of Oily Fish Intake on Risk of AMD

 

MR Analysis

The primary method used in this study for evaluating the genetic link between oily fish intake and the AMD technique used the random-effects IVW approach. This analysis revealed that oily fish intake is a significant independent risk factor for AMD (P = 0.034, odds ratio [95% confidence interval], 0.58 [0.35–0.96]) as depicted in Figure 2. Figure 2A indicates the study of “Oily fish intake” and its MR effect size on “AMD (whether dry or wet).” These markers represent different locations of genetic variations that might be linked to AMD risk. The chart illustrates the role of multiple SNPs in the relationship between oily fish intake and AMD risk. By analyzing the effect sizes of these genetic variations, the study identifies potential protective SNPs, providing a basis for future prevention and treatment strategies. The Figure 2B studies the “SNP effect on AMD (whether dry or wet)” with a specific identifier (id) possibly related to a dataset. Negative values indicate a protective effect, suggesting a reduced risk of AMD. The figure demonstrates the relationship between SNPs, oily fish intake, and the risk of AMD using multiple MR methods. Some SNPs show a significant negative effect size, suggesting a protective role in AMD risk reduction. The use of different MR methods provides a comprehensive view of the potential causal relationships and ensures robustness in the findings. The weighted median method corroborated the genetic causal connection between oily fish intake and AMD. The results of the five MR analysis methods utilized in this study are summarized in the Table. 

Figure 2.

 

The association between oily fish intake and AMD. (A) Forest plot of causal effects of oily fish intake on AMD. (B) Scatter plot of causal effects of oily fish intake on AMD. The slope of the line represents the causal effect of each method.

The Cochrane Q test, used to identify heterogeneity, did not reveal any substantial heterogeneity between the consumption of oily fish and AMD (P = 0.837), as confirmed by the symmetric distribution of SNPs in funnel plots (Fig. 3A). Both the MR-Egger intercept and the MR-PRESSO global tests indicated an absence of horizontal pleiotropy in this association (P = 0.881). Although IVW method showed significant association, MR-Egger method failed to show significance, mainly because MR-Egger method was more conservative in pleiotropic adjustment and had lower statistical power. No outliers were detected in the MR-PRESSO analysis. In addition, a sensitivity analysis was conducted where particular SNPs were removed (Fig. 3B), which confirmed the strength and dependability of our MR analysis results. 

Figure 3.

 

The effect size for oily fish intake on AMD. (A) The funnel plot demonstrated symmetrical distribution of the SNPs, suggesting the absence of heterogeneity in the connection. (B) The leave-one-out test demonstrated that the outcome remained unaffected by individual influential SNP, therefore indicating the stability of this connection.

These differences emphasize the importance of using multiple methods of analysis when interpreting causal relationships. By combining the results of the IVW and MR-Egger methods, the researchers were able to arrive at more comprehensive and reliable conclusions, leading to a better understanding of the potential impact of oily fish intake on AMD risk. 

Discussion

This study used a two-sample MR analysis of comprehensive GWAS data to examine the causal connection between the intake of oily fish and AMD. The differences in different analytical methods highlight the importance of using multiple analytical methods when interpreting causality, allowing researchers to draw more comprehensive and reliable conclusions and thus better understand the potential impact of oily fish intake on AMD risk. Our findings indicate a notable correlation between increased oily fish consumption and a decreased risk of AMD, thus reinforcing the genetic basis for the preventive potential of oily fish in AMD. 

Emerging research underscores the connection between dietary patterns, specifically oily fish consumption, and reduced AMD risk.^14–21 In a study by Ulańczyk et al.²² involving 330 AMD patients and 121 controls,⁸ an investigation into dietary habits revealed the benefits of oily fish for AMD patients, potentially delaying disease progression and preserving visual function. Another study by Cristina Augood and colleagues, involving 105 neovascular AMD (NV-AMD) patients and 2170 controls, indicated that consuming cholesterol-rich fish at least once weekly, compared to less frequently, halved the incidence of NV-AMD. Furthermore, a retrospective analysis of a 7756-participant cohort from an age-related eye disease control trial,²³ which examined diets via a questionnaire, suggested that adherence to a Mediterranean diet, particularly fish consumption, could mitigate AMD risk and progression. 

However, the causal link between oily fish consumption and AMD has not been definitively established by observational studies, with concerns about confounding factors and reverse causation. GWAS has proven effective in understanding complex diseases, moving beyond single-gene associations to identify gene clusters, thereby affirming prior research and offering new insights.²⁴ Our study, based on comprehensive GWAS data, lends genetic support to the causal role of oily fish consumption in AMD prevention. 

The pathogenesis of AMD remains incompletely understood, although inflammation and oxidative stress are considered key factors in its development.^25–28 Omega-3 long-chain polyunsaturated fatty acids, including docosahexaenoic acid (DHA) and eicosapentaenoic acid (EPA), obtained mainly from oily fish, have a vital function in reducing retinal inflammation. Furthermore, omega-3 fatty acids are implicated in modulating angiogenesis, a process integral to the development of wet AMD, wherein abnormal neovascularization significantly contributes to vision loss. These fatty acids are believed to suppress such pathological angiogenesis, offering potential in slowing or preventing AMD progression. DHA and EPA are instrumental in regulating retinal inflammatory responses and oxidative stress, enhancing cellular resistance to oxidative stress through modifications in membrane fluidity and lipid composition, thereby potentially diminishing AMD risk.^29–32 

In addition, PUFA, such as N-3 PUFA and N-6 PUFA, are the primary active ingredients in oily fish, including DHA and EPA, and it is essential to investigate their specific role in AMD risk. Numerous studies have underscored the protective effects of PUFA on retinal health. Previous research supports the hypothesis that higher plasma levels of PUFA are associated with a reduced risk of advanced AMD.^18,31,33 The SNPs selected for our MR analysis are strongly associated with oily fish intake and may also be related to PUFA levels, given the high content of these fatty acids in oily fish. 

By integrating these findings, our study not only suggests a potential genetic causal relationship between oily fish intake and reduced AMD risk but also highlights the critical role of omega-3 fatty acids and PUFA in this protective effect. Future research should aim to isolate the effects of omega-3 fatty acids and PUFA from other components of oily fish to better understand their specific contribution to AMD prevention. 

In conclusion, oily fish consumption plays a beneficial role in AMD risk reduction. These findings lay a theoretical foundation for AMD prevention through dietary means and guide future research directions. 

This study boasts several strengths. It is, to our knowledge, the first to use large-scale GWAS data in exploring the causal link between oily fish consumption and AMD. The two-sample MR approach effectively tackles the inherent constraints of observational research, such as reverse causality and confounding variables. Additionally, the study meticulously selected instrumental variables for MR analysis, enhancing result accuracy. Diverse techniques were used to assess sensitivity, horizontal pleiotropy, and heterogeneity, all affirming the stable and robust association between oily fish consumption and AMD. 

Nevertheless, the study has limitations. Its racial scope is confined to participants of European descent in the GWAS, questioning the applicability of results to other populations. Although attempts have been made to account for genetic pleiotropy, confounding factors like education, personality, and nutrition might still bias the results. Moreover, the MR analysis depends on the existing GWAS meta-analysis data, precluding stratified analyses by country, race, or age group. Hence, the observed AMD effects may not extend to other specific demographic groups. 

Conclusions

Our study suggests a significant causal connection between higher oily fish consumption and reduced AMD risk, offering vital genetic evidence for AMD prevention through dietary measures, particularly increased oily fish intake. Future research should target more diverse populations to affirm this relationship's universality and further explore the protective mechanisms of oily fish and various types of PUFA against AMD. 

Acknowledgments

Supported by grants from the National Natural Science Foundation of China (81874491). The sponsor or funding organization had no role in the design or conduct of this research. 

Disclosure: M. Gao, None; J. Zhou, None; J. Zhao, None; Z. Liu, None; X. Luo, None; C. Yang, None; X. Yu, None; M. Tang, None; J. Zhu, None; X. Yan, None