Abstract Background Glaucoma is a leading cause of irreversible blindness, influenced by systemic and lifestyle factors. This study investigates the causal relationships between dietary habits, sleep traits, amino acids, metabolites, and inflammatory factors with glaucoma subtypes using Mendelian randomization (MR) and validates findings through cross-sectional analysis. Methods MR analysis assessed the causal effects of 226 dietary factors, 11 sleep traits, 20 amino acids, 1400 metabolites, and 91 inflammatory factors on five glaucoma subtypes (NTG, POAG, PACG, NVG, XFG). Mediation MR analysis explored the role of amino acids and inflammatory factors in these relationships. Validation was conducted using NHANES cross-sectional data. Results High-fat, high-calorie diets increased glaucoma risk, while antioxidant-rich foods and better sleep quality reduced it. Key mediators included proline, tyrosine, IL-1 A, and PDL1. NHANES data confirmed lower intake of vitamins A and C, higher water consumption among glaucoma patients, and significant sleep-related associations. Conclusion Our findings highlight the role of balanced diets and optimized sleep patterns in glaucoma prevention and management. This study provides evidence for targeted lifestyle interventions focusing on metabolic and inflammatory pathways to mitigate glaucoma risk. Supplementary Information The online version contains supplementary material available at 10.1186/s12986-025-00967-4. Keywords: Glaucoma, Mendelian randomization, Dietary habits, Sleep traits, Amino acids, Inflammatory factors, Metabolomics, NHANES, Lifestyle interventions, Neuroprotection. Introduction Glaucoma is a group of heterogeneous conditions characterized by progressive retinal ganglion cell (RGC) loss and optic nerve damage, often associated with elevated intraocular pressure (IOP), leading to irreversible blindness if untreated [[30]1]. Globally, nearly 95 million people are affected by glaucoma, with a rising prevalence as populations age, highlighting the urgent need for early detection and effective management [[31]2]. Glaucoma encompasses several phenotypes, primarily distinguished by anterior chamber angle anatomy, with primarily open-angle glaucoma (POAG) being more common but primarily angle-closure glaucoma (PACG) contributing disproportionately to blindness [[32]3]. Globally, approximately 65 million people have POAG, while 30 million have PACG, the latter accounting for nearly half of all glaucoma-related blindness [[33]4, [34]5]. Normal tension glaucoma (NTG) is a common form of POAG characterized by RGC death and progressive visual field loss despite records of intraocular pressures (IOP) within the normal range (< 21 mm Hg) [[35]6, [36]7]. Neovascular glaucoma (NVG) is a severe abnormal blood vessel growth on the iris and anterior chamber angle, typically secondary to ischemic retinal conditions [[37]8]. Exfoliation glaucoma (XFG) is a secondary open-angle glaucoma caused by the deposition of fibrillar aggregates in the anterior eye segment, linked to exfoliation syndrome (XFS) and characterized by disrupted cellular processes such as protein aggregation and stress [[38]9]. Glaucoma is a progressive optic neuropathy and a leading cause of irreversible blindness, involving not only elevated intraocular pressure but also systemic factors such as oxidative stress, metabolic dysregulation, inflammation, and vascular dysfunction [[39]9–[40]14]. In recent years, lifestyle factors such as diet and sleep patterns have received increasing attention for their role in diseases, particularly glaucoma [[41]15–[42]17]. These behavioral factors can directly affect metabolic health and may also indirectly influence optic nerve function by modulating inflammatory states and neuroprotective mechanisms [[43]18, [44]19]. For example, high-protein diets have been shown to regulate amino acid metabolism, with amino acid metabolic pathways (such as the kynurenine pathway) closely linked to neuroprotection and oxidative stress [[45]20, [46]21]. Conversely, irregular sleep patterns may exacerbate optic nerve damage through chronic inflammatory pathways [[47]22, [48]23]. Some Researchers reported that Sleep disturbances, particularly conditions like obstructive sleep apnea, are associated with an increased risk of optic nerve damage and may exacerbate the progression of glaucoma through mechanisms such as hypoxia, inflammation, and vascular dysregulation [[49]23]. However, the precise mechanisms by which these factors influence the onset and progression of glaucoma remain unclear. Mendelian randomization (MR) is a robust analytical method that simulates a randomized controlled trial to infer causal relationships between genetic variants and intermediate phenotypes [[50]24]. By leveraging the random allocation of genetic variants during gamete formation, MR minimizes the effects of reverse causation and confounding bias [[51]25]. This method has proven effective in investigating the causal links between glaucoma and lifestyles like diets and sleep features. The UK Biobank (UKB) provides a robust platform for such research, offering detailed genotype and phenotype data for over 480,000 unrelated individuals, including laboratory results, data on 226 dietary habits, and 11 sleep features [[52]26]. This study adopts an innovative approach by integrating MR analysis, mediation analysis, and cross-sectional data to comprehensively explore the causal relationships between dietary factors, sleep traits, amino acids, metabolites, inflammatory factors, and glaucoma subtypes. These relationships, particularly their underlying mechanisms, have not been thoroughly investigated in previous studies. By employing systematic MR, we evaluated the causal effects of 226 dietary factors, 11 sleep traits, 1400 metabolites, 91 inflammatory factors, and 20 amino acids on various glaucoma subtypes. The primary objective of this study is to unravel the causal links between dietary factors, sleep characteristics, and glaucoma subtypes. Additionally, we aimed to determine whether amino acids and inflammatory factors serve as potential mediators in these relationships and identify critical metabolic pathways through metabolite enrichment analysis. To validate our findings, we utilized data from the National Health and Nutrition Examination Survey (NHANES), investigating the associations between identified dietary patterns, sleep traits, and glaucoma to confirm their clinical significance further. Methods and materials Study design Figure [53]1 shows the overview diagram of our study design. This study was conducted in three distinct phases. In the first phase, we applied MR analysis to explore the causal relationships between 226 dietary factors, 11 sleep traits, 1400 metabolites, 91 inflammatory factors, and 20 amino acids and various glaucoma subtypes, including NTG, POAG, PACG, NVG, and XFG. To ensure the robustness of the MR analysis, three essential criteria were satisfied: (1) the genetic variants used as instrumental variables (IVs) must be significantly associated with the exposure; (2) the selected IVs must not be correlated with confounding factors influencing both the exposure and the outcome; and (3) horizontal pleiotropy must be absent, meaning the IVs should influence glaucoma only through the specified exposure mechanism [[54]27]. In the second phase, we employed a mediation analysis to explore the mediation effect between diets, sleep features and glaucoma subtypes mediated by inflammatory factors and amino acids. Finally, an observational study using data from NHANES was performed to identify significant associations between diets, sleep traits and glaucoma subtypes. Fig. 1. [55]Fig. 1 [56]Open in a new tab Study design Study population Data sources for 226 dietary factors, 11 sleep traits, 1400 metabolites, 91 inflammatory factors, and 20 amino acids in MR The UK Biobank is a large-scale prospective cohort study that recruited more than half a million individuals, aged 40 to 69, from 22 different assessment centers across the United Kingdom from 2006 to 2010 [[57]28]. To minimize the impact of ancestry-related confounding factors, we focused our analyses exclusively on unrelated individuals of European heritage. Our study involved a thorough examination and detailed fine mapping of loci linked to 226 diets within this cohort. Data on 11 sleep traits were primarily sourced from the UK Biobank, which includes Genome-Wide Association Study (GWAS) analyses of self-reported and objective sleep traits among individuals of European ancestry, supplemented by FinnGen data on obstructive sleep apnea and nocturnal oxygen saturation phenotypes. The metabolomic GWAS data, derived from a study on genetic influences on metabolite profiles, analyzed 1,091 metabolites and 309 ratios [[58]29]. GWAS data for 91 inflammatory factors were sourced from a protein quantitative trait locus study involving 14,824 European-ancestry participants, accessible via the GWAS Catalog [[59]30]. Summary GWAS data for 20 amino acids were obtained from a sizeable cross-platform metabolite study involving up to 86,507 participants [[60]31]. In this study, five glaucoma subtypes were utilized as outcome measures. The glaucoma GWAS was obtained from FinnGen Release 12 (R12), a genotype dataset derived from Finnish biobanks encompassing over 500,000 individuals. Data sources and study population of NHANES analysis The NHANES survey is an ongoing, cross-sectional study to assess non-institutionalized U.S. civilians’ health and nutritional status. Utilizing a multistage probability sampling approach, NHANES ensures a nationally representative dataset. Data collection involves structured interviews, clinical assessments, and laboratory analyses, with the National Center for Health Statistics (CDC) methodologies detailed. To improve the representativeness of the results, statistical analyses incorporate survey weights, stratification, and clustering adjustments. The study received ethical approval, and all participants provided written informed consent [[61]32]. This study used publicly available NHANES data from the 2005/2006 and 2007/2008 cycles to investigate the association between diets, sleep features and glaucoma. Initially, 20,498 participants aged 40 years and older were identified. After excluding individuals without necessary data or information, a final cohort of 6505 participants (471 glaucoma patients and 6034 controls) was included. Glaucoma diagnoses were based on standardized criteria assessed by experienced professionals. This robust dataset provides a reliable foundation for exploring associations between diets, sleep traits and glaucoma. Although we adjusted for key demographic factors in NHANES analysis, potential residual confounding factors, such as medication use and comorbid conditions, were not accounted for due to data limitations. Future studies should consider sensitivity analyses, such as excluding participants with major comorbidities, to further confirm the robustness of these associations. Statistical analysis Instrument selection of MR We employed a rigorous strategy for selecting instrumental variables, starting with a p-value threshold of > 5E − 06 to identify suitable single-nucleotide variants (SNVs) as instruments. We selected this threshold (as opposed to the conventional 5E − 08 genome-wide significance level) to obtain sufficient SNP instruments for exposures like diet and sleep, which have relatively few significant variants. This approach is precedent in other MR studies that use p < 5E − 06 when genome-wide significant SNPs are limited, balancing the need for more instruments with an acceptable level of instrument strength. To minimize linkage disequilibrium and ensure the inclusion of independent variants in the MR analysis, we applied a clumping procedure, eliminating variants linked to the top-associated SNPs (r² threshold = 0.001, window size = 10,000 kb). Two-sample MR analysis for Glaucoma subtypes A two-sample MR, using the inverse variance weighted (IVW) method with a random-effects model, was performed via the TwoSampleMR package (v0.5.6, [62]https://mrcieu.github.io/TwoSampleMR/) to assess the effects of exposures on glaucoma subtypes. Data harmonization and analysis were also performed using this package, with the MR analysis executed via the mr() function. Assessment of heterogeneity was carried out using the mr_heterogeneity() function, where a heterogeneity p-value (Q_pval) < 0.05 indicated significant variability among instrumental variables. The presence of directional pleiotropy was assessed using the MR-Egger intercept test, where a significant deviation of the intercept from zero (P < 0.05) indicated potential pleiotropy. To ensure the robustness of the selected instrumental variables, we calculated the F-statistic (F > 10), confirming a strong correlation between SNPs and the exposures. Additionally, MR-Egger regression was employed to further evaluate pleiotropic effects, with a P-value < 0.05 suggesting the possibility of pleiotropy among the instrumental variables. Due to the exploratory nature and the large number of tested hypotheses, multiple testing correction methods, such as the false discovery rate (FDR), were not strictly implemented. This approach aligns with our aim of generating hypotheses rather than definitive conclusions. No significant horizontal pleiotropy was detected for the reported causal associations (MR-Egger intercept p > 0.05; MR-PRESSO global test p > 0.05), supporting the validity of the instruments [[63]33–[64]36]. We used Steiger filtering to verify causal directionality explicitly (Table [65]S14). A leave-one-out sensitivity analysis was also conducted to assess the influence of individual SNPs, reinforcing the reliability of the inferred causal relationship between exposures and glaucoma risk. Additional MR methods, including MR-Egger, weighted median, and weighted mode—were employed alongside IVW to validate the robustness of findings. Moreover, reverse MR analysis was conducted for exposures demonstrating causal associations, ensuring the absence of bidirectional relationships between exposures and outcomes. Mediation analysis This study employed MR to assess the mediating role of specific intermediates between diets and glaucoma. The analysis was conducted in two stages. First, MR was used to estimate the causal association (β₁) between exposures and potential mediators, determining whether the exposures significantly influenced these intermediates. In the second stage, MR was applied to evaluate the effect of the identified mediators (β₂) on the risk of the outcome, establishing their potential contribution to disease development. The mediation proportion was calculated by determining the ratio of the indirect effect (β₁× β₂) to the overall causal impact of the exposures on the outcome, providing insight into the extent to which mediators contributed to the observed associations [[66]37]. This approach gave a more comprehensive understanding of the biological mechanisms underlying the exposure-outcome relationship [[67]38]. Enrichment analysis of metabolites To gain deeper insights into the metabolic pathways associated with different glaucoma subtypes, we conducted an enrichment analysis on metabolites that demonstrated significant associations (P < 0.05). Identified metabolites were systematically mapped to a curated metabolic pathway database to explore their functional relevance [[68]39]. The enrichment ratio was determined by calculating the proportion of significantly associated metabolites within each pathway relative to the total number of metabolites reported in the reference set [[69]40]. To assess the statistical significance of pathway enrichment, we employed a hypergeometric test, which considered both the global metabolite reference set and the specific subset of metabolites linked to glaucoma. This approach allowed us to identify metabolic pathways that were disproportionately represented among the associated metabolites, providing potential insights into the underlying biochemical mechanisms contributing to glaucoma pathogenesis [[70]41]. Statistical analysis in cross-sectional analysis In this study, clinical data were analyzed using EmpowerStats software. Baseline characteristics of the study population were statistically summarized for both the glaucoma and control groups. Continuous variables were expressed as means ± standard deviations (SD). The associations between dietary intake, sleep characteristics, and glaucoma were assessed, with statistical significance set at P < 0.05. A weighting approach was applied to minimize data variability. We performed additional sensitivity analyses using NHANES data. Specifically, we conducted analyses using three progressively adjusted models. Model 1 was the unadjusted crude model; Model 2 was adjusted for fundamental demographic factors (age, sex, and race); and Model 3 was fully adjusted for additional critical confounders, including body mass index (BMI), education level, hypertension, diabetes, cholesterol, LDL-C, HbA1c, albumin, apolipoproteins, and insulin. Results The causal association between diets and Glaucoma The causal effect between 226 diets and glaucoma (NTG, NVG, PACG, POAG and XFG) was assessed by MR. Of the 226 diets, the IVW method found that a series of diets were significantly associated (P < 0.05) with glaucoma subtypes (Fig. [71]2A ~ E, Table [72]S1 ~ [73]5). In NTG, higher genetically predicted intake of certain energy-dense foods (e.g., cheese, chocolate, and bread) was associated with increased NTG risk, suggesting that diets rich in fats and sugars may contribute to NTG. Conversely, greater consumption of nutrient-rich foods like fruits (e.g., grapes) and whole grains (e.g., high-fiber cereals) showed protective associations, indicating that dietary balance and antioxidant intake may be beneficial. For PACG, diets high in animal products (meats and eggs) showed significant positive causal associations, whereas higher intake of vegetables and plant-based foods appeared protective. The MR analysis found that increased consumption of refined carbohydrates (e.g., naan bread) significantly raised POAG risk in POAG. NVG showed positive causal links with several diet factors, including bananas, sweetened foods, red wine, and high-glycemic index items, implying metabolic susceptibility in NVG. Across subtypes, a common theme was that high-calorie, high-fat dietary patterns tended to increase glaucoma risk, while diets abundant in vitamins and antioxidants were protective. Fig. 2. [74]Fig. 2 [75]Open in a new tab A: The causal effects between Dietary Intake and NTG. B: The causal effects between Dietary Intake and PACG. C: The causal effects between Dietary Intake and POAG. D: The causal effects between Dietary Intake and NVG. E: The causal effects between Dietary Intake and XFG The causal association of sleep with Glaucoma After MR, we identified six sleep traits with significant causal associations across different glaucoma subtypes. Figure [76]3 summarizes the MR results for sleep traits. We observed that specific sleep characteristics had protective effects. For example, genetically longer daytime napping and a tendency toward daytime sleepiness were associated with a lower risk of POAG. This suggests that the ability to take restorative naps or perhaps a phenotype of compensating sleep may have neuroprotective effects in POAG. More sleep episodes (fragmented sleep with multiple bouts) were. Fig. 3. [77]Fig. 3 [78]Open in a new tab The causal effects between Sleep and Glaucoma protective for PACG, which might indicate that distributed sleep or napping reduces stress on ocular perfusion. Interestingly, longer sleep duration was estimated to have a protective effect in NTG and NVG, and higher sleep efficiency (a measure of sleep quality) was protective against XFG. These findings support that better sleep quality and adequate sleep duration can benefit ocular health. However, we did not find significant MR evidence linking insomnia or chronotype to glaucoma, suggesting that sleep quantity/quality may matter more than sleep timing preference. The causal association between amino acids, metabolites and Glaucoma We assessed the causal relationship between amino acids, metabolites and glaucoma subtypes. This analysis identified six significant causal amino acids with glaucoma (Fig. [79]4A). Glutamate (OR = 11.954), Asparagine (OR = 3.627), and Proline (OR = 3.476) were significantly associated with NTG, while Aspartate (OR = 8.773) showed a significant association with NVG. Tryptophan (OR = 0.422) was significantly linked to POAG, and Tyrosine (OR = 0.535) was significantly associated with PACG. Subsequently, we identified 56 metabolites with a significant causal relationship with NTG (Table [80]S8), 44 metabolites with NVG (Table [81]S9), 40 metabolites with PACG (Table [82]S10), 88 metabolites with POAG (Table [83]S11), and 59 metabolites with XFG (Table [84]S12). Fig. 4. [85]Fig. 4 [86]Open in a new tab A: The causal effects between amino acids and glaucoma. B: The enrichment of significant metabolites in POAG. C: The enrichment of significant metabolites in PACG. D: The Enrichment of significant metabolites in NTG Enrichment analysis was performed on the identified metabolites to investigate further potential metabolic pathways associated with NTG, PACG, and POAG (Fig. [87]4B–D, Table [88]S15 - [89]S17). In POAG, the most prominently enriched pathways included alanine, aspartate, glutamate metabolism, and tyrosine metabolism, followed by pathways linked to sulfur metabolism and caffeine metabolism (Fig. [90]4B). MR analysis revealed a significant causal association between tyrosine and PACG, suggesting that tyrosine may be closely linked to PACG risk. Several metabolic pathways were significantly enriched in POAG, including arginine biosynthesis, alanine, aspartate, glutamate, glutathione, nitrogen, and caffeine metabolism. Enrichment was observed in the bile acid biosynthesis and methyl histidine metabolism pathways in NTG. POAG and PACG exhibited significant enrichment in the alanine, aspartate, and glutamate metabolism and caffeine metabolism pathways, suggesting their potential regulatory roles in different glaucoma subtypes. These pathways may serve as promising therapeutic targets for future interventions. Causal effect of inflammatory factors with Glaucoma in MR We used MR analysis to investigate the causal relationships between 91 inflammatory factors and glaucoma subtypes (Table [91]S7). We identified significant causal relationships between various inflammatory factors and different glaucoma subtypes. Specifically, 4 inflammatory factors (2B4, CX3CL1, CCL28, CXCL11) were associated with NTG; 8 inflammatory factors (CCL19, CDCP1, CSF1, S100-A12, IL-10, MMP-10, TWEAK, uPA) were linked to NVG; 7 inflammatory factors (eIF4E-BP1, IL-6, SCF, FGF19, TNFRSF9, IL-24, TWEAK) showed significant associations with PACG; and 11 inflammatory factors (MMP-10, IL-18, IL-10RA, CXCL1, IL-1 A, TNFRSF9, 2B4, MMP-1, CXCL11, PDL1, CXCL9) were related to POAG. Additionally, six inflammatory factors (TGF-A, MCP-3, IL-17 C, MCP-2, FGF21, IL-1 A) demonstrated significant causal relationships with XFG (Fig. [92]5). Fig. 5. [93]Fig. 5 [94]Open in a new tab The causal effects between inflammatory factors and glaucoma subtypes. Abbreviations: IL-1 A (interleukin-1α); PDL1 (programmed death-ligand 1); 2B4 (CD244, also known as SLAMF4); CX3CL1 (C-X3-C motif chemokine ligand 1, i.e., fractalkine); CCL28 (C-C motif chemokine ligand 28); CXCL11 (C-X-C motif chemokine ligand 11); CCL19 (C-C motif chemokine ligand 19); CDCP1 (CUB domain-containing protein 1); CSF1 (colony-stimulating factor 1); S100-A12 (S100 calcium-binding protein A12); IL-10 (interleukin-10); MMP-10 (matrix metalloproteinase-10); TWEAK (TNF-related weak inducer of apoptosis, TNFSF12); uPA (urokinase plasminogen activator); eIF4E-BP1 (eukaryotic initiation factor 4E binding protein 1); IL-6 (interleukin-6); SCF (stem cell factor, also known as KIT ligand); FGF19 (fibroblast growth factor 19); TNFRSF9 (tumor necrosis factor receptor superfamily member 9, CD137); IL-24 (interleukin-24); IL-18 (interleukin-18); IL-10RA (interleukin-10 receptor subunit alpha); CXCL1 (C-X-C motif chemokine ligand 1); MMP-1 (matrix metalloproteinase-1); CXCL9 (C-X-C motif chemokine ligand 9); TGF-A (transforming growth factor alpha); MCP-3 (monocyte chemoattractant protein-3, CCL7); IL-17 C (interleukin-17 C); MCP-2 (monocyte chemoattractant protein-2, CCL8); FGF21 (fibroblast growth factor 21) The mediation effect in MR We used MR mediation analysis to investigate the role of amino acid levels as mediators between dietary intake and glaucoma and inflammatory factors as mediators between sleep traits and glaucoma, along with the magnitude of these mediation effects. We used MR-based mediation analysis to investigate the role of specific intermediaries. Our findings revealed several notable mediated pathways. Sleep efficiency was associated with a lower risk of XFG, potentially mediated by higher IL-1 A levels, while daytime napping was associated with a lower risk of POAG, possibly via increased PDL1. In the mediation relationship between dietary habits and amino acid levels, soft cheese intake and alcohol intake frequency were associated with higher NTG risk, with elevated proline levels partially mediating these associations. Additionally, avocado intake was associated with reduced PACG risk, potentially via increased tyrosine levels, and complex cheese intake was linked to lower POAG risk, possibly through effects on tryptophan. These examples suggest that a portion of the impact of diet and sleep on glaucoma is conveyed through specific biochemical pathways (including amino acids such as proline, tyrosine, and tryptophan, and inflammatory mediators such as IL-1 A and PDL1). ( Table [95]1) (In Table [96]1, “Step 1” refers to the MR effect of the exposure on the mediator (β₁) and “Step 2” refers to the effect of the mediator on the outcome (β₂); Beta1 and Beta2 are the estimated effect sizes for these steps, and the mediation proportion indicates the percentage of the exposure’s total effect on glaucoma risk mediated through that pathway). Table 1. The mediation effect of amino and inflammations between diets, sleep and glaucoma exposure mediator outcome Step1 Step2 Total Mediation Effect Mediation Proportion Beta1 Pval Beta2 Pval Beta Pval Sleep efficiency IL-1 A XFG 0.470 0.015 -0.177 0.050 -0.965 0.010 -0.083 8.61% daytime napping PDL1 POAG 0.415 0.004 -0.130 0.034 -1.705 0.047 -0.054 3.17% Soft cheese intake Proline NTG 0.452 0.025 1.246 0.049 1.607 0.045 0.563 35.04% Alcohol intake frequency Proline NTG 0.158 0.044 1.246 0.049 0.221 0.041 0.197 89.30% Avocado intake Tyrosine PACG 0.306 0.048 -0.625 0.036 -1.249 0.016 -0.191 15.28% Hard cheese intake Tryptophan POAG -0.040 0.027 -0.886 0.007 0.373 0.017 0.036 9.6% [97]Open in a new tab Baseline characteristics of cross-sectional research This study included 471 glaucoma patients and 6,034 controls (Tables [98]2 and [99]3, Table [100]S13, Table [101]S18). Table [102]2 presents group differences in age, gender, race, and education levels. Expanding on previous MR studies, we examined the associations between dietary intake (Table [103]2), sleep patterns (Table [104]3), and glaucoma. Among key demographic factors, age, race, education level, and blood pressure were significantly linked to glaucoma, with blood pressure showing the strongest association. Hypertension was more prevalent in glaucoma patients (63.3%) compared to controls (45.5%). We adjusted for these demographic factors in the NHANES logistic regression analyses to account for confounding. Table 2. Weighted characteristics of the study population based on Glaucoma and diets Phenotype Glaucoma Control P-Value N 471 6034 AGE 69.0 ± 11.1 59.5 ± 12.7 < 0.001 GENDER 0.607 Male 229 (48.6%) 3008 (49.9%) Female 242 (51.4%) 3026 (50.1%) RACE < 0.001 Mexican American 52 (11.0%) 953 (15.8%) Other Hispanic 37 (7.9%) 437 (7.2%) Non-Hispanic White 233 (49.5%) 3168 (52.5%) Non-Hispanic Black 135 (28.7%) 1260 (20.9%) Other Race 14 (3.0%) 216 (3.6%) EDUCATION < 0.001 Less Than 9th Grade 101 (21.4%) 912 (15.1%) 9-11th Grade 80 (17.0%) 933 (15.5%) High School 119 (25.3%) 1466 (24.3%) College or AA degree 112 (23.8%) 1496 (24.8%) College Graduate or above 59 (12.5%) 1221 (20.2%) BP < 0.001 Yes 298 (63.3%) 2714 (45.0%) No 173 (36.7%) 3320 (55.0%) DM 0.067 Yes 23 (4.9%) 427 (7.1%) No 439 (93.2%) 5539 (91.8%) BMI 29.2 ± 6.1 29.1 ± 6.5 0.957 WAIST 101.8 ± 14.2 100.5 ± 14.7 0.019 ENERGY 131.9 ± 162.9 129.3 ± 182.1 0.485 PROTEIN 4.5 ± 8.5 4.8 ± 10.0 0.607 CARBON 16.9 ± 22.7 16.3 ± 25.3 0.605 FIBER 0.8 ± 1.4 0.8 ± 1.8 0.504 SUGAR 8.4 ± 14.0 7.9 ± 15.6 0.541 TFAT 4.9 ± 7.7 4.9 ± 9.3 0.987 ISFAT 1.6 ± 2.6 1.7 ± 3.4 0.861 IMFAT 1.7 ± 2.9 1.8 ± 3.7 0.713 IPFAT 1.1 ± 2.4 1.0 ± 2.4 0.568 CHOLESTEROL 12.9 ± 38.0 16.5 ± 52.7 0.140 RET 27.4 ± 61.8 26.2 ± 70.1 0.731 ACAR 4.4 ± 29.2 14.7 ± 186.3 0.228 BCAR 72.9 ± 390.2 90.6 ± 770.3 0.620 VITA 27.7 ± 58.9 36.8 ± 98.5 0.049 VITB1 0.1 ± 0.2 0.1 ± 0.2 0.441 VITB2 0.1 ± 0.2 0.1 ± 0.2 0.439 VITB6 0.1 ± 0.2 0.1 ± 0.3 0.579 VITB12 0.3 ± 0.7 0.3 ± 0.9 0.940 VITC 4.0 ± 13.7 6.4 ± 24.5 0.038 VITE 0.4 ± 0.7 0.4 ± 1.0 0.773 VITK 5.1 ± 27.0 4.8 ± 27.4 0.813 IFOLA 22.3 ± 60.1 23.9 ± 76.9 0.657 CALCIUM 63.7 ± 107.3 57.9 ± 120.7 0.314 MAGN 15.6 ± 21.0 15.8 ± 23.2 0.905 PHOS 76.2 ± 105.5 77.9 ± 130.0 0.784 IRON 0.9 ± 2.3 1.0 ± 2.2 0.493 ZINC 0.6 ± 1.2 0.7 ± 1.9 0.428 COPP 0.1 ± 0.1 0.1 ± 0.1 0.526 SODI 189.5 ± 356.1 198.1 ± 394.2 0.647 POTA 150.3 ± 202.3 150.6 ± 221.4 0.975 SELE 6.0 ± 12.4 6.4 ± 15.0 0.564 CAFF 5.0 ± 27.0 4.9 ± 28.8 0.975 WATER 183.9 ± 384.6 143.9 ± 287.3 0.005 [105]Open in a new tab Table 3. The association between sleep features and glaucoma Phenotype Glaucoma Control P-Value N 471 6034 Sleep onset latency(min) 24.6 ± 21.0 22.5 ± 20.0 0.032 Sleep duration(hours/day) < 0.001 1 1 (0.2%) 8 (0.1%) 2 3 (0.6%) 24 (0.4%) 3 7 (1.5%) 86 (1.4%) 4 27 (5.7%) 243 (4.0%) 5 58 (12.3%) 613 (10.2%) 6 103 (21.9%) 1421 (23.5%) 7 99 (21.0%) 1616 (26.8%) 8 113 (24.0%) 1584 (26.3%) 9 33 (7.0%) 289 (4.8%) 10 16 (3.4%) 102 (1.7%) 11 1 (0.2%) 25 (0.4%) 12 or over 10 (2.1%) 23 (0.4%) Insomnia 0.260 No 460 (97.7%) 5935 (98.4%) Yes 11 (2.3%) 99 (1.6%) Wake up during night 0.310 Never 165 (35.0%) 2158 (35.8%) Rarely 71 (15.1%) 1117 (18.5%) Sometimes 126 (26.8%) 1445 (23.9%) Often 64 (13.6%) 791 (13.1%) Always 45 (9.6%) 523 (8.7%) Wake up early 0.183 Never 195 (41.4%) 2601 (43.1%) Rarely 70 (14.9%) 1081 (17.9%) Sometimes 110 (23.4%) 1180 (19.6%) Often 56 (11.9%) 713 (11.8%) Always 40 (8.5%) 459 (7.6%) Unrested during the day 0.176 Never 194 (41.2%) 2201 (36.5%) Rarely 70 (14.9%) 974 (16.1%) Sometimes 101 (21.4%) 1550 (25.7%) Often 64 (13.6%) 782 (13.0%) Always 42 (8.9%) 527 (8.7%) Overly sleepy during day 0.106 Never 195 (41.4%) 2328 (38.6%) Rarely 81 (17.2%) 1216 (20.2%) Sometimes 106 (22.5%) 1498 (24.8%) Often 52 (11.0%) 654 (10.8%) Always 37 (7.9%) 338 (5.6%) [106]Open in a new tab The study assessed 33 dietary and nutritional factors, including energy intake, macronutrients (protein, carbohydrates, fats), micronutrients (vitamins and minerals), caffeine, and water intake. Significant associations were identified for vitamin A, C, and water intake. Vitamin A and C intake were notably lower in glaucoma patients than controls, whereas water intake was significantly higher. No other dietary factors showed significant associations with glaucoma. We evaluated the relationships between glaucoma and sleep traits, including sleep onset latency, sleep duration, insomnia, nighttime awakenings, early waking, daytime fatigue, and excessive sleepiness. Significant associations were found between sleep onset latency and sleep duration. Glaucoma patients had a more prolonged sleep onset latency (24.6 min vs. 22.5 min in controls). Additionally, a higher proportion of glaucoma patients slept more than 12 h. No significant differences were observed for other sleep parameters. Discussion This study employed multi-omics MR and cross-sectional analyses to identify significant causal relationships between various dietary habits, sleep traits, and different types of glaucoma. Additionally, we observed that these factors could regulate glaucoma risk by modulating amino acid levels and inflammatory factors. We found that cheese, chocolate, and bread may have a positive causal relationship with the onset of NTG, while diets rich in meat and eggs showed a significant positive causal association with PACG. In POAG, naan bread was observed to increase the disease risk significantly. For NVG, foods such as bananas, chocolate raisins, sweeteners, lamb, juice, red wine, and biscuits significantly increased the risk of glaucoma. Among sleep traits, daytime napping and daytime sleepiness were found to reduce the risk of POAG. Sleep episodes provided a protective effect against PACG. Long sleep may have a protective role in both NTG and NVG, while sleep efficiency was observed to offer a protective effect against XFG. Additionally, through mediation analysis, we discovered that dietary and sleep habits influence different glaucoma subtypes by regulating amino acids and inflammatory factors. Finally, using population data from the NHANES database, we found that vitamin A and vitamin C can reduce the risk of glaucoma, while high water intake increases the risk. Moreover, sleeping for more than 12 h was also associated with an increased risk of glaucoma. In this study, we identified several dietary factors with both positive and negative causal relationships to NTG. Foods such as non-oily fish, hard cheese, chocolate bars, naan bread, and alcohol intake frequency showed a positive causal association, suggesting that excessive energy-dense and processed foods may increase NTG risk. In contrast, moderate consumption of the same or similar foods, such as grapes, cereals, and cereal bars, demonstrated protective effects, highlighting the importance of intake quantity, dietary balance, and nutrient composition in NTG development. Previous studies have shown that fruits and vegetables with antioxidant properties (e.g., kale, collard greens, carrots, and peaches) protect against POAG in Black women. Similarly, dietary fat may influence IOP through pathways involving endogenous n-6 prostaglandins, and fat-free diets and parenteral nutrition have been shown to reduce IOP [[107]42–[108]44]. Therefore, a low-fat, high-grain diet combined with antioxidant-rich fruits and vegetables may help prevent neurodegenerative diseases such as glaucoma [[109]45]. In addition, our analysis revealed that foods such as naan bread, hard cheese, lamb, mango, dark chocolate, whole egg, and energy-dense processed items were associated with POAG and PACG risk, suggesting that high-fat, high-calorie diets may contribute to glaucoma progression through metabolic dysfunction or inflammatory pathways. This is consistent with evidence that diets high in fat can exacerbate retinal oxidative stress and neurodegeneration [[110]46]. Conversely, the intake of antioxidant-rich foods (e.g., fruits, vegetables, and avocado), fiber-rich cereals, and nutrient-dense items (e.g., milk, bran cereal, and dairy smoothies) demonstrated protective effects, likely through their anti-inflammatory and neuroprotective properties. These findings highlight the potential role of a balanced diet—low in saturated fats and processed foods while rich in antioxidants and fiber—to mitigate glaucoma risk. Furthermore, mediation analysis provided additional insights into the underlying mechanisms. The observed increase in proline levels due to soft cheese intake and alcohol intake frequency, which significantly elevated the risk of NTG, suggests a possible role in proline-associated oxidative stress in retinal ganglion cell damage. In contrast, the protective effects of avocado intake on PACG through elevated tyrosine levels and complex cheese intake on POAG via tryptophan may reflect the neuroprotective properties of these amino acids, particularly in regulating neurotransmitter synthesis and reducing neuroinflammation. For example, tyrosine is a precursor to catecholamine neurotransmitters such as dopamine, which support retinal ganglion cell function; treatments that enhance dopaminergic activity (e.g., citicoline) have shown neuroprotective effects in glaucoma [[111]47]. These findings underscore the importance of dietary interventions targeting amino acid pathways as a promising approach for glaucoma prevention and management. Furthermore, our metabolite pathway enrichment analysis identified several metabolic pathways associated with glaucoma. Similar integrative metabolomics studies have successfully applied pathway enrichment to interpret disease-associated metabolite changes, supporting the reliability of our pathway-level results. At the same time, further studies are needed to validate these associations and explore the underlying mechanisms. Our findings highlight the intricate relationship between sleep traits and glaucoma subtypes, suggesting that sleep duration and quality play a significant role in glaucoma risk modulation through immune and inflammatory pathways. The observed negative causal associations of daytime napping and daytime sleepiness with POAG suggest that short, restorative naps or compensatory sleep episodes may help regulate systemic and ocular homeostasis, potentially alleviating stress-related pathways associated with glaucomatous damage. Similarly, the protective effect of sleep efficiency on XFG and the number of sleep episodes on PACG underscores the importance of stable and efficient sleep patterns in maintaining neurovascular integrity, as sleep disturbances are often linked to oxidative stress and dysregulation of ocular perfusion pressure. Our mediation analysis further deepened this understanding by identifying specific inflammatory mediators. Sleep efficiency reduced XFG risk by promoting IL-1 A, a cytokine involved in tissue repair and immune regulation, which may mitigate trabecular meshwork dysfunction and extracellular matrix remodeling, processes central to XFG pathology. Indeed, IL-1α has been shown to induce matrix metalloproteinases in trabecular meshwork cells and increase aqueous humor outflow, potentially alleviating the outflow resistance characteristic of XFG [[112]48]. Conversely, the association between daytime napping and reduced POAG risk through elevated PDL1 levels suggests an immunomodulatory mechanism, as PDL1 is known to regulate cellular stress and promote ocular immune privilege, potentially protecting RGCs from apoptosis. PD-L1 expressed on ocular cells can inhibit pro-inflammatory T-cell activity, reinforcing immune privilege, and modulation of the PD-1/PD-L1 pathway has been reported to influence RGC apoptosis in glaucoma models [[113]49]. However, our NHANES cross-sectional analysis showed that extremely long sleep durations (> 12 h) were associated with higher glaucoma risk, seemingly contradicting the MR finding that longer sleep is protective. This discrepancy might be due to confounding or reverse causation in the observational data (for instance, poor health leading to prolonged sleep) [[114]50], whereas the MR approach captures the effect of genetically influenced sleep traits with less bias. These results indicate that sleep-related interventions targeting immune modulation and inflammation could offer potential glaucoma prevention and management strategies. In previous studies, data on sleep quality and sleep architecture in glaucoma patients have been limited and inconsistent [[115]51]. Some researchers used the Epworth Sleepiness Scale (ESS) to assess daytime sleepiness and conduct polysomnography, finding that glaucoma patients had significantly higher average ESS scores than control groups [[116]52, [117]53]. Other studies have reported that glaucoma patients exhibited reduced considerably sleep efficiency and lower average minimum oxygen saturation, as well as significantly higher wake after sleep onset and an increased burden of periodic limb movements associated with arousals [[118]54]. Further research is needed to explore the therapeutic implications. Our study highlights the significant role of dietary habits and sleep traits in the development and progression of different glaucoma subtypes, offering critical clinical implications for glaucoma prevention and management. We found that high-fat, high-calorie diets may increase glaucoma risk through metabolic and inflammatory pathways. At the same time, antioxidant-rich fruits, vegetables, fiber-rich cereals, and nutrient-dense foods provide protective effects by mitigating oxidative stress and neurodegeneration. Notably, glaucoma patients had significantly lower vitamin A and C intake and higher water consumption, which may exacerbate intraocular pressure fluctuations and disease progression. It should be noted that this observation comes from cross-sectional data; therefore, the lower vitamin intake in glaucoma patients could be influenced by other health or lifestyle differences (residual confounding), and we cannot infer a direct causal protective effect of vitamins without longitudinal evidence. Sleep traits such as daytime napping, sleep efficiency, and sleep duration were associated with reduced glaucoma risk, potentially through immune regulation and oxidative stress pathways. For instance, promoting IL-1 A and PDL1 may contribute to tissue repair and immune homeostasis. Our findings suggest that glaucoma patients could benefit from targeted lifestyle interventions, including increasing vitamin A and C intake, moderating water consumption, and optimizing sleep quality. Integrating dietary and sleep assessments into glaucoma care may provide a personalized strategy to prevent disease progression and improve patient outcomes. Moreover, translating improvements in sleep efficiency into clinical practice could involve structured sleep hygiene education or targeted interventions addressing underlying sleep disorders, thus potentially providing practical strategies for glaucoma management. Limitations The study has several limitations: (1) Our analysis relied on GWAS data primarily derived from European populations, so undetected population stratification or ancestry-related biases may have affected our results, and this also limits the generalizability of our findings to other ethnicities or populations worldwide. (2) In the NHANES database, glaucoma cases were not classified into specific subtypes, which may obscure subtype-specific associations and reduce the granularity of our findings. (3) Although key demographic covariates were adjusted, residual confounding factors, including medication use, physical activity, and comorbidities, might still exist. Future analyses should implement sensitivity analyses by excluding participants with significant comorbidities or explicitly adjusting for medication use. (4)While efforts were made to address horizontal pleiotropy in the MR analysis, its influence on the results cannot be completely ruled out. (5) A significant limitation lies in the reliance on self-reported dietary habits and sleep traits, which may introduce recall bias and measurement inaccuracies. (6) Finally, while we identified associations between dietary intake, sleep patterns, amino acid pathways, and inflammatory markers, the underlying mechanisms remain speculative and require validation through longitudinal studies and experimental research. We also acknowledge that our analysis primarily focused on amino acids and inflammatory mediators; however, other potential pathways such as oxidative stress and vascular dysfunction may play crucial roles in the pathogenesis of glaucoma and potentially explain some of the observed lifestyle effects. Particularly notable is the complex relationship between sleep traits and glaucoma subtypes revealed in our study. Specifically, our MR analysis demonstrated that daytime napping and higher sleep efficiency significantly reduced the risks of POAG and XFG, respectively. These protective effects may be mediated through inflammatory and immune-regulatory pathways. For example, increased PDL1 associated with daytime napping might enhance ocular immune privilege, thus preventing retinal ganglion cell apoptosis, while elevated IL-1 A levels linked to improved sleep efficiency could alleviate trabecular meshwork dysfunction typical of XFG. However, we observed contradictory results in our NHANES cross-sectional analysis, indicating that prolonged sleep duration (≥ 12 h) significantly increased glaucoma risk. This discrepancy must be interpreted cautiously. The associations observed in the NHANES data may be influenced by reverse causation or residual confounding factors, such as poorer general health status or medication usage, leading individuals to sleep longer durations. Conversely, MR analyses are less prone to these confounders and reverse causation issues because they rely on genetically determined sleep characteristics. Future studies involving multi-center cohorts, larger sample sizes, and more diverse populations are necessary to validate these findings and improve their generalizability. Furthermore, the sample sizes for NVG and XFG were relatively small in both GWAS and NHANES datasets, limiting the statistical power for these subtypes. Therefore, findings related to these glaucoma subtypes should be interpreted with caution and validated using larger datasets in future research. Additionally, given the explicitly exploratory nature of our study, we tested multiple hypotheses without strictly applying FDR correction, which increases the likelihood of false-positive findings. Although we have highlighted consistency with previous literature to support the plausibility of many associations, some identified relationships could still be incidental, and independent validation in future studies is warranted. Conclusion This study employed MR analysis to investigate the causal impact of dietary habits, sleep traits, amino acids, and inflammatory markers on glaucoma subtypes, with findings validated using NHANES data. We identified that high-fat, high-calorie diets increased glaucoma risk, while antioxidant-rich foods and better sleep quality showed protective effects. Specific sleep traits, such as daytime napping and sleep efficiency, influenced glaucoma risk through pathways involving IL-1 A and PDL1, highlighting the role of immune and inflammatory regulation. These results offer new insights for glaucoma prevention and management through targeted lifestyle interventions, though further studies are needed to confirm these findings across diverse populations. Electronic supplementary material Below is the link to the electronic supplementary material. [119]Supplementary Material 1^ (11KB, doc) [120]Supplementary Material 2^ (18.5KB, doc) [121]Supplementary Material 3^ (4.9MB, xlsx) [122]Supplementary Material 4^ (5.1MB, xlsx) Acknowledgements