Association of Dietary Nitrate and a Mediterranean Diet With Age-Related Macular Degeneration Among US Adults: The Age-Related Eye Disease Study (AREDS) and AREDS2

Key Points

Question  Is dietary nitrate intake associated with age-related macular degeneration (AMD) progression?

Findings  In this cohort study, a post hoc analysis of 2 randomized clinical trials found that increased dietary nitrate intake was associated with a decreased risk of late AMD. When dietary patterns were adjusted to include a Mediterranean diet, dietary nitrate intake was no longer associated with AMD progression independently.

Meaning  These findings suggest that dietary nitrate may be a modifiable risk factor for AMD progression; however, much of the association of nitrate intake was confounded by plant-based dietary patterns.

Importance  Low dietary nitrate intake has previously been suggested to be a risk factor for age-related macular degeneration (AMD) progression; however, this finding has not been replicated in other cohorts or adjusted for dietary patterns.

Objective  To determine whether there is an association between dietary nitrate intake and AMD progression.

Design, Setting, and Participants  This cohort study analyzed data from the prospective Age-Related Eye Disease Study (AREDS) and AREDS2 randomized clinical trial cohorts and their extended follow-up studies, which were conducted in multicenter outpatient retinal practices. Participants in both trials had non–late AMD in at least 1 eye. Data were analyzed from March 1, 2020, to September 30, 2022.

Exposure  Dietary nitrate intake.

Main Outcomes and Measures  Association between dietary nitrate intake and development of late AMD (neovascular AMD [nAMD] or geographic atrophy [GA]) or large drusen. The interactions of dietary patterns, with common at-risk single-nucleotide polymorphisms, were also assessed.

Results  In the combined AREDS/AREDS2 cohort of 7788 participants (4288 AREDS participants and 3610 AREDS2 participants [110 of whom participated in both studies]), there were 13 511 eligible eyes. The combined cohort comprised 4396 women (56%) and 3392 men (44%), and the combined mean (SD) age was 71.1 (6.6) years. Dietary nitrate intake was associated with a decreased risk of progression to late AMD in the combined AREDS/AREDS2 cohort (hazard ratio [HR], 0.77 [95% CI, 0.69-0.86] for quartile 4 vs quartile 1 of intake) and a decreased risk of GA (HR, 0.71 [95% CI, 0.61-0.83]) and nAMD (HR, 0.85 [95% CI, 0.73-0.99]). In AREDS, increased nitrate intake (quartile 4 vs quartile 1) was associated with a decreased risk of late AMD (HR, 0.77 [95% CI, 0.65-0.91]) and GA (HR, 0.80 [95% CI, 0.65-0.97]) but not nAMD; in AREDS2, there was no association between nitrate intake (quartile 4 vs quartile 1) and late AMD (HR, 0.90 [95% CI, 0.80-1.02]) or nAMD (HR, 0.93 [95% CI, 0.78-1.11]). There was a correlation between Mediterranean dietary patterns and dietary nitrate intake (r = 0.52, P < .001).

Conclusions and Relevance  The findings of this cohort study suggest that dietary nitrate intake was associated with lower AMD risk. However, this association disappeared after adjusting for Mediterranean dietary patterns. These results are subject to potential bias and are hypothesis-generating in nature; therefore, they are insufficient to support new clinical recommendations. Previously described associations between dietary nitrate intake and AMD may in fact represent overall dietary patterns. Further research is needed before dietary nitrate intake can be recommended as a therapy for AMD.

Age-related macular degeneration (AMD) is a leading cause of vision loss in high-income nations.¹ Although the exact pathogenesis remains unknown, alterations in blood flow and inflammation have been implicated as possible mechanisms in AMD pathogenesis.² Both nitric oxide (NO) and nitric oxide synthase (NOS) are involved in these processes, and altered levels of NO and NOS have been reported in cases of AMD.³ This raises the possibility that dietary intake of nitrates, which are the precursors for NO, may influence AMD development or progression.

Dietary nitrate is obtained predominantly from vegetables, particularly leafy green vegetables.⁴ Leafy green vegetable intake has previously been associated with lower risk of AMD.⁵ This finding is believed to relate to lutein and its isomer zeaxanthin in these vegetables, as these molecules selectively accumulate in the macula and have antioxidant and anti-inflammatory properties.^5,6 Given the potential role of NO in mediating inflammation and vascular function,³ it is possible that nitrate intake contributes to the benefit associated with leafy green vegetable intake. Recent research has found an association between narrower retinal microvasculature caliber and lower dietary nitrate intake, suggesting a possible mechanism by which nitrate intake influences retinal vasculature and AMD development.⁷ There is also evidence that increased dietary intake of nitrates may be associated with reduced risks of early AMD⁸; however, this result has not been replicated in other late-AMD cohorts. It remains unclear whether such findings may be confounded by the effect of other vegetable substances that are known to be protective against AMD (eg, lutein).^5,6,9

We assessed the association between dietary nitrate intake and the progression to intermediate and late AMD in a large prospective study comprising 2 randomized clinical trial (RCT) cohorts. We also explored potential confounding and interactions between dietary nitrate intake and lutein intake.

This cohort study analyzed the Age-Related Eye Disease Study (AREDS; 1992 to 2005) and AREDS2 (2006 to 2012), 2 large, multicenter randomized clinical trials^10,11 that enrolled 4757 participants at 11 and 4203 participants at 82 retinal clinics, respectively, across the US. Institutional review board approval was obtained at each site; no incentives or compensation were provided for AREDS participants, while travel funds were provided for AREDS2 participants. All participants provided written informed consent. The research adhered to the Declaration of Helsinki, and only AREDS2 complied with the Health Insurance Portability and Accountability Act. The present study followed the STROBE reporting guideline for cohort studies.

Participants in AREDS were randomly assigned to placebo, zinc, antioxidants, or a combination of zinc and antioxidants for 5 years. AREDS2 participants were randomly assigned to an AREDS supplement that reduced AMD risk (1) with lutein and zeaxanthin, (2) with docosahexaenoic acid and eicosapentaenoic acid, (3) with a combination of both of these substance combinations, or (4) alone for 5 years. In both trials, participants underwent baseline and yearly eye examinations including color fundus photography (CFP), and photographs were centrally graded by the University of Wisconsin Reading Center.¹² After study closeout at 5 years, epidemiologic follow-up was conducted among 3549 of the 4203 surviving AREDS participants for an additional 5 years¹³; in AREDS2, 3882 participants underwent follow-up via telephone for an additional 5 years.¹⁴

Progression to late AMD was the development of late AMD, which was defined as either the presence of neovascular AMD (nAMD) on CFP or treatment of nAMD, and the presence of geographic atrophy (GA; either central or noncentral) on CFP. Progression to large drusen was defined as the presence of large drusen as graded on CFP in an eye that did not have either large drusen or late AMD at baseline. For the AREDS2, in eyes without late AMD at the end of the 5-year follow-up, progression to late AMD (not subtyped) at 10 years was defined as noted in the eMethods in Supplement 1.

In both AREDS and AREDS2, validated food frequency questionnaires (FFQs) were administered to all participants at randomization. The AREDS FFQ consisted of 90 items in a semiquantitative modified Block FFQ¹⁵ and the AREDS2 FFQ consisted of 131 items in a semiquantitative nature.^16,17 Both FFQs asked participants to report how often a specific food or beverage was consumed during the preceding year (eMethods in Supplement 1).

For both AREDS and AREDS2, a subset of participants (2899 in AREDS and 1826 in AREDS2) underwent genotype analysis. For each participant, the AMD genetic risk score (GRS), a weighted risk score for developing late AMD, was calculated.¹⁸ Single-nucleotide variants (SNVs) were analyzed via a custom Illumina HumanCoreExome array (Illumina Inc). Four SNVs at 3 loci were also selected for their increased attributable risk of late AMD: ARMS2 rs10490924, CFH rs10922109 and rs1061170, and C3 rs2230199.

Eyes without late AMD at baseline participants with at least 2 study visits were analyzed. For the combined AREDS and AREDS2 cohort, multivariate Cox proportional hazards regression analyses were performed for the outcomes of progression to late AMD, GA, and nAMD, according to the energy-adjusted nitrate quartile of intake (quartile 1 as the reference). P values for trends were calculated using the energy-adjusted nitrate intake as a continuous variable; Cox proportional hazards regression was tested in all cases. In the lone situation where the assumption was not met (for the combined cohort for nAMD), stratified Cox proportional hazards regression was performed instead.¹⁹ The unit of analysis was the eye, including adjustments for age, sex, smoking status, body mass index (for AREDS only), and total caloric intake. Race and ethnicity were not analyzed as covariates, as all of the participants self-reported as White. Correlation between eyes was made in SAS, version 9.4 (SAS Institute Inc) by using the robust sandwich estimate for the covariance matrix in the Wald tests.²⁰ These analyses were repeated for the individual AREDS and AREDS2 cohorts separately.

To determine whether genotype modified associations between nitrate intake and AMD, Cox proportional hazards regression analyses were repeated, including the interaction term with nitrate intake (treated continuously) and genotype. For the nutrient-genotype combinations with P values for interaction of .01 or less, the regression analyses were performed separately for each level of the genetic characteristic. Associations of nitrate intake to progression to late AMD or development of progression to large drusen (in AREDS only) were evaluated with adjustment for lutein and zeaxanthin intake (treated continuously). Data were analyzed from March 1, 2020, to September 30, 2022.

In the combined AREDS/AREDS2 cohort of 7788 participants (4288 AREDS participants and 3610 AREDS2 participants [110 of whom participated in both studies]), there were 13 511 eligible eyes at baseline. The combined cohort comprised 4396 women (56%) and 3392 men (44%), and the combined mean (SD) age was 71.1 (6.6) years. Additional participant characteristics, including AMD severity and dietary nitrate intake at baseline, are provided in eTable 1 in Supplement 1.

By final follow-up (median, 10.2 y [IQR, 6.7-11.0 y]), 4575 eyes (33.9%), 2227 eyes (16.8%), and 2086 eyes (15.4%) had progressed to late AMD, GA, or nAMD, respectively. Individually, AREDS comprised 7822 eyes (4288 participants) and AREDS2 comprised 5896 eyes (3610 participants). In the analysis of progression to large drusen in the AREDS, there were 5068 eligible eyes (2981 participants), with 1375 eyes (27.1%) progressing to large drusen by final follow-up.

In the combined cohort, statistically significant linear associations were observed for quartile 4 of dietary intake vs quartile 1. The hazard ratios (HRs) for late AMD, GA, and nAMD were 0.77 (95% CI, 0.69-0.86), 0.71 (95% CI, 0.61-0.83), and 0.85 (95% CI, 0.73-0.99), respectively (Figure 1A).

Results for the AREDS cohort alone were similar to those for the combined cohort (Figure 1B). In the AREDS cohort, increased nitrate intake (quartile 4 vs quartile 1) was associated with a decreased risk of late AMD (HR, 0.77 [95% CI, 0.65-0.91]) and GA (HR, 0.80 [95% CI, 0.65-0.97]) but not nAMD. In contrast, in the AREDS2 cohort, the reduced risk for progression to late AMD did not have a linear trend (P for trend = .26); rather, lower progression to overall late AMD was observed among participants whose nitrate intake increased in quartile 3 vs quartile 1 (HR, 0.88 [95% CI, 0.78-1.00]) (Figure 1C). However, the P for trend was not associated with dietary nitrate intake for GA (P = .11) or nAMD (P = .99) (Figure 1C).

For progression to large drusen, there was no linear trend with increasing nitrate intake in the AREDS cohort (P = .37) (eFigure in Supplement 1). There appeared to be a threshold at quartile 3, with no further decline in risk of progression to large drusen beyond a nitrate intake of approximately 139 mg/d (the median for quartile 3).

We explored the correlation of dietary nitrate intake with individual vitamins and minerals which might confound, explain, or modify relationships of nitrate to late AMD (eTable 2 in Supplement 1). Some nutrients were so highly correlated (r > 0.6) with dietary nitrate that it was not possible to reliably disentangle the associations of these to AMD outcomes. These included lutein, zeaxanthin, folate, beta-carotene, and vitamins A, C, and E.

The correlation of lutein and zeaxanthin intake with nitrate intake was r = 0.48 (eTable 2 in Supplement 1). Increased nitrate intake (quartiles 2-4) was associated with a reduced risk of progression to late AMD even after adjusting for lutein and zeaxanthin intake (Figure 2). Quartile 4 of nitrate intake was also associated with a reduced risk of both GA and nAMD in the combined cohort after adjusting for lutein and zeaxanthin intake (Figure 2).

The intake of many other vitamins, minerals, dietary fiber (soluble and insoluble) and healthy fats (including eicosapentaenoic acid and docosahexaenoic acid) were also correlated with dietary nitrate intake (Pearson r = 0.1-0.5) (eTable 2 in Supplement 1). This finding suggests that dietary nitrate was correlated with an overall nutrient-rich, plant-based diet, potentially explaining the role of nitrates in progression to late AMD. We thus further explored relationships of nitrate intake to adherence to a Mediterranean diet pattern.

In general, individuals with high nitrate intakes were likely to have diets that were denser in vitamins, minerals, and healthy fats. Consistent with this observation, the Pearson correlation coefficient between nitrates (mg/d) and diet index in the combined cohort was r = 0.52 (P < .001). When the Mediterranean diet index tertiles were added to the model, the associations of dietary nitrate with late AMD were attenuated and were no longer statistically significant for nAMD but remained significant for late AMD and GA (Figure 3), suggesting that dietary nitrate is not an entirely independent estimator of late AMD in these cohorts and may be a marker of other components in a fruit and vegetable–rich dietary pattern. In both the AREDS and AREDS2 cohorts, after adjusting for diet index tertiles, there were no protective associations between nitrate intake and late AMD.

In the combined cohort regression model, there was an interaction between continuous values of nitrate intake and Mediterranean diet scores in relation to late AMD (P for interaction = .04), such that a protective nitrate association was observed among individuals in the lower tertile for adherence to a Mediterranean diet but not among those with better adherence. The HRs for quartile 4 vs quartile 1 were 0.72 (95% CI, 0.57-0.92; P = .008), 0.86 (95% CI, 0.69-1.07; P = .17), and 1.07 (95% CI, 0.86-1.34; P = .54) for diet index tertiles 1, 2 and 3, respectively (Table 1).

In the AREDS cohort, the protective associations of nitrates were observed to be most prominent among participants with lower intakes of lutein and zeaxanthin. The interaction term between continuous nitrate intake and intake of lutein and zeaxanthin was not significant. No association was observed between nitrate intake and risk of late AMD, GA, or nAMD in either AREDS or AREDS2 when evaluated by tertile of nitrate intake. In the combined cohort, a protective association of dietary nitrate was observed for overall late AMD (HR, 0.81 [95% CI, 0.71-0.93]) and GA (HR, 0.70 [95% CI, 0.58-0.85]) in the highest tertile.

In the AREDS2 cohort, the interaction term between continuous levels of nitrate intake and diet score was not significant. There were no associations between nitrate intake and late AMD, GA, or nAMD subtypes among participants in any tertile for diet index score.

The associations of nitrate intake with progression to late AMD and large drusen were examined by number of AMD risk variants and GRS. Interactions with P ≤ .01 underwent Cox proportional hazards regression analysis for each genotype level (Table 2). For the combined cohort, the protective associations between nitrate intake and progression to overall late AMD were strongest among participants with no risk alleles for CFH variant rs10922109 and among participants with 1 risk allele for GA. For the GRS, protective associations of nitrates with overall late AMD and AMD subtypes were most prominent among participants with medium levels of GRS. In both AREDS and AREDS2, nitrate associations were not statistically different by genotype.

Dietary nitrates have been suggested as possible contributors to the cardioprotective effect of fruit and vegetable–rich diets^21,22 but have not been studied in relation to AMD until recently. In the AREDS/AREDS2 combined cohort, dietary nitrate intake was associated with a reduced risk of progression to late AMD, including its subtypes. These results support and expand on similar results from the Blue Mountain Eye Study (BMES), which first identified a reduced risk of incident early AMD with increased nitrate intake.⁸

Similar to the BMES results, there appeared to be a threshold for the association of nitrate intake to progression to large drusen in the AREDS cohort. In BMES, early AMD was not associated with increasing risk beyond the third quartile.⁸ In the present study, increasing nitrate intake beyond quartile 3 in AREDS (nitrate median of 139 mg/d) did not further lower risk for large drusen. These results suggest that that it might not be necessary to consume nitrate levels of 140 mg/d or greater to prevent early AMD.

A large number of vitamins, minerals, and healthy fats, many of which do not share the same food sources of nitrates, were correlated with nitrate intake in these cohorts. A Mediterranean diet pattern, which was associated with lower risk of progression to late AMD in the AREDS cohort, is rich in many of these nutrients and was highly correlated with nitrate intake.²³ Healthy diet patterns such as the Mediterranean diet are commonly dense in a wide variety of vitamins, minerals, fatty acids, and protective bioactive molecules, which may have contributed to the protective associations observed for dietary nitrates in the present study. These include folate, pyridoxine hydrochloride (vitamin B₆), and cyanocobalamin (vitamin B₁₂). In the BMES, elevated serum homocysteine, and folate and/or vitamin B₁₂ deficiencies predicted increased risk of incident AMD.^24,25 Consequently, isolating the effect of these food components and other individual dietary components in observational studies can be difficult due to the correlation between intake of multiple nutrients. In addition, dietary patterns such as the Mediterranean diet, which appear to be protective against late AMD,²³ may reflect the effect of healthier overall lifestyles (or to unknown or unmeasured aspects of healthy lifestyles) rather than individual components, which reduce progression to late AMD.

The associations between nitrate intake and both late AMD phenotypes were more pronounced in the AREDS cohort than in the AREDS2 cohort. These differences might reflect the narrower distribution of dietary nitrate in the AREDS2 cohort. In the combined cohort, the highest nitrate intake quartile compared to individuals with nitrate intakes in the lowest quartile had a 20% lower risk for progression to overall late AMD, 22% lower risk for progression to GA, and 39% lower risk for progression to nAMD. Because dietary nitrate was associated with all stages of AMD, it is possible that the cumulative lower risk over a period that is longer than the follow-up of the individuals in these cohorts could be substantially lower.

The progression to large drusen in the AREDS cohort was found to be significant for the third quartile of nitrate intake, similar to a secondary analysis of the BMES.⁸ However, the risk of progression to large drusen in BMES may not be sufficient to detect an association with the highest quartile, given the low incidence of AMD in the cohort. Taken together, the results of our study and this previous investigation suggest that increased nitrate intake may be an additional modifiable risk factor for intermediate AMD that warrants additional exploration.

The strengths of this study include the large, well-defined participant population with an extensive, prospective follow-up. In addition, all data were collected in a standardized fashion, and centralized reading center grading was conducted.

The study limitations include possible unmeasured confounding factors, such as physical activity, or variations in growing or storage conditions, which can affect vegetable nitrate content,²⁶ as well as differences in variables between cohorts, such as ethnicity. Other potential known sources of nitrates, including well water and nitrate-containing medications such as nitroglycerin, were also unable to be assessed, which may have resulted in imprecise measurements. There are also known limitations to the accuracy of FFQs as a measurement of diet, and variations in the FFQ used between the AREDS and AREDS2 may have contributed to the differences in results between these studies. Adjustment of the major analyses for genetic risk was conducted in only a subset of participants. Enhanced imaging techniques, such as optical coherence tomography, were only performed in a small subset of the AREDS2 participants, and advanced psychophysical assessments were also not undertaken. These analyses did not take into account potential modifications of diets because of concomitant diseases such as diabetes or cardiovascular disease; additionally, they were post hoc in nature, such that any conclusions should be considered as hypothesis-generating rather than supporting new clinical recommendations. Finally, these results may not be generalizable to other populations.

The findings of this cohort study suggest that nitrate intake was associated with a decreased risk of progression to late AMD, including both late AMD subtypes. This association persisted even after accounting for lutein and zeaxanthin intake as a potential confounder and was strongest among individuals with lower adherence to a fruit and vegetable–rich diet pattern. A protective association was also seen against the development of large drusen, suggesting that the benefit of dietary nitrates may extend across a wide spectrum of AMD severities; however, no evidence was seen to support recommending additional nitrate supplementation. These results are from post hoc analyses and are therefore hypothesis-generating in nature. Much of the outcome associated with nitrate intake can be attributed to plant-based dietary patterns in general, such as a Mediterranean diet. Further research, including potential randomized clinical trials, may be warranted to further assess the role of dietary nitrate in reducing the risk of AMD progression.

Accepted for Publication: October 17, 2022.

Published Online: December 22, 2022. doi:10.1001/jamaophthalmol.2022.5404

Corresponding Author: Emily Y. Chew, MD, Division of Epidemiology and Clinical Applications, National Eye Institute, National Institutes of Health, Bldg 10 Center Dr, MSC 120, 410, CRC, Room 3-2531, Bethesda, MD 20892 (echew@nei.nih.gov).

Author Contributions: Dr Chew had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Agrón, Keenan, Chew.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Broadhead, Agrón, Peprah, Mares, Chew.

Critical revision of the manuscript for important intellectual content: Broadhead, Agrón, Keenan, Lawler, Mares, Chew.

Statistical analysis: Agrón, Peprah.

Obtained funding: Chew.

Administrative, technical, or material support: Chew.

Supervision: Broadhead, Chew.

Conflict of Interest Disclosures: None reported.

Funding/Support: This research was supported by the Intramural Research Program of the National Eye Institute and by contracts from the National Eye Institute (contract NOI-EY-0-2127 for AREDS and contract HHS-N-260-2005-00007-C and ADB contract N01-EY-5-0007 for AREDS2), which received funding from the Office of Dietary Supplements; the National Center for Complementary and Alternative Medicine; the National Institute on Aging; the National Heart, Lung, and Blood Institute; and the National Institute of Neurological Disorders and Stroke.

Role of the Funder/Sponsor: The sponsor and funding organization participated in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, and approval of the manuscript; and the decision to submit the manuscript for publication.

Group Information: The AREDS/AREDS2 Investigators are listed in Supplement 2.

Meeting Presentations: Preliminary analysis of this work was presented at The Association for Research and Vision in Ophthalmology Annual Meeting; May 1 to 7, 2021 (virtual).

Data Sharing Statement: See Supplement 3.