The role of nutritional factors in transitioning between early, mid, and late stages of age-related macular degeneration: prospective longitudinal analysis

Abstract

Background

Transitions between different stages of age-related macular degeneration (AMD) are not completely captured by traditional survival models with an end point of advanced AMD.

Objectives

This study aimed to explore the transitions from early and intermediate AMD to higher non-advanced and advanced stages and determine the contributions of nutritional factors to these outcomes.

Methods

Eyes with early or intermediate AMD at baseline, classified according to the Age-Related Eye Disease Study severity scale, were included in this prospective longitudinal analysis. Foods and the biologically active nutrients associated with AMD [green leafy vegetables, fish, lutein/zeaxanthin (LZ), and ω-3 (n–3) fatty acids] were determined by a baseline food frequency questionnaire. Progression was defined as eyes transitioning to higher severity groups including non-advanced and advanced stages over 5 years, confirmed at 2 consecutive visits. Cox proportional hazards models for foods and nutrients were analyzed adjusting for demographics, lifestyle, baseline macular status, a family history of AMD, caloric intake, and genetic risk.

Results

Among 2697 eyes, 616 (23%) progressed to higher severity groups. In the food group model, higher intake of green leafy vegetables reduced incidence of transitions {hazard ratio [HR] (≥2.7 servings/wk compared with none): 0.75; 95% confidence interval [CI]: 0.59, 0.96; P = 0.02}. Higher fish intake was also protective [HR (≥ two 4-ounce servings/wk compared with <2): 0.79; 95% CI: 0.65, 0.95; P = 0.01]. In the nutrient model, LZ intake was protective [HR (≥2 mg/d compared with <2): 0.76; 95% CI: 0.60, 0.96; P = 0.02]. Higher intake of ω-3 fatty acids also tended to be beneficial [HR (≥0.7 g/wk compared with <0.7): 0.85; 95% CI: 0.71, 1.01; P = 0.06].

Conclusions

Increased consumption of green leafy vegetables, LZ, and fish nutritionally rich in ω-3 fatty acids during the initial stages of AMD may reduce rates of progression to higher severity of this debilitating disease.

This trial was registered at clinicaltrials.gov as NCT00594672.

Introduction

Age-related macular degeneration (AMD) is a multifactorial disease causing vision loss in an estimated 20 million people living in the United States, of which there are 18 million people affected with early and intermediate non-advanced AMD [1]. Across the globe, prevalence of AMD is expected to surge to 288 million in 2040 [2], and with the recent increase in life expectancies, this chronic neurodegenerative disease heavily impacts vision loss and quality of life in industrialized nations [3,4]. Therefore, prevention is of utmost importance.

AMD is usually initially diagnosed by the presence of drusen abnormalities under the retinal pigment epithelium (RPE) or retina. Even before noticeable visual acuity changes, other visual disturbances such as loss of dark adaptation owing to photoreceptor loss and RPE dysfunction [[5], [6], [7]] are often seen with earlier non-advanced stages of AMD. As the disease advances from early and intermediate non-advanced stages to later advanced stages of geographic atrophy (GA) and neovascularization (NV), individuals experience central vision loss, making it harder to see faces, read, or drive [4]. Because earlier non-advanced stages precede advanced end points of GA with limited therapeutic options and NV with a high burden of intravitreal injections, there is a growing need for prevention to decrease the rate of progression to both non-advanced and advanced stages of macular degeneration.

Previous studies have highlighted the contribution of factors that increase rates of progression of AMD over the life course, for example, age, smoking, BMI, exercise, education, diet, inflammatory biomarkers, baseline macular status, a family history of AMD, and genetic variants are significantly associated with AMD [[8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18], [19], [20], [21], [22], [23], [24], [25], [26], [27], [28], [29], [30]]. We have developed predictive models incorporating various non-genetic and genetic variables to distinguish between eyes likely to progress to advanced age-related macular degeneration (AAMD) and those that do not progress [[31], [32], [33], [34], [35], [36], [37]]. Our latest AMD calculator incorporates a family history and selective genetic variants to predict risk of AAMD, and AUC has reached 0.94 [38]. There is a paucity of information, however, on predictive factors in the earlier stages of disease.

In this study, we focus on transitions beyond the usual paradigm of progression to AAMD. Our primary objective was to investigate progression from early and intermediate AMD to higher non-advanced and advanced stages and analyze whether modifiable nutritional factors that have previously been associated with progression to AAMD [8,10,15,33] are also associated with these transitions, independent of other risk factors including genetic risk. Early disease prevention plays a pivotal role in slowing disease progression, thereby mitigating the transition to both non-advanced and advanced irreversible stages. This proactive approach might result in both improvement in personal quality of life and substantial mitigation of the socioeconomic burden related to AMD [4,39].

Methods

Study population

Deidentified data from the Age-Related Eye Disease Study (AREDS), a multicenter randomized clinical trial, was accessed from NIH Database of Genotype and Phenotypes through accession number phs000001.v3.p1. This trial was registered at clinicaltrials.gov as NCT00594672, enrolling participants from November 1992 to January 1998, with the clinical trial lasting until April 2001 and continued follow-up until December 2005 [40]. Research adhered to the tenets of the Declaration of Helsinki and was performed under approved institutional review board protocols of the participating sites.

Figure 1 displays the selection of subjects and eyes for the analytic data set. Patients with no genetic specimen, with incomplete phenotypic, genetic, nutritional, and family history data, and with baseline AMD severity group 1 (severity scale 1) were excluded. The analytic data set included eyes with scales 2–8 (non-advanced AMD) at baseline and complete genetic, phenotype, family history data, with ≥1 year follow-up and valid daily caloric intake (females: 600–3200 kcal; males: 600–4200 kcal), which yielded 2697 eyes obtained from 1757 subjects.

FIGURE 1.

Flowchart showing selection of subjects for this study who are at risk of progression to higher AMD severity groups from AREDS cohort (N = 2697 eyes). AMD, age-related macular degeneration; AREDS, Age-Related Eye Disease Study.

Nondietary covariates at baseline

Data on age (55–64, 65–74 and ≥75 years), sex (female/male), race, education (≤high school, >high school), and smoking status were collected at baseline and included as covariates. Race was reported as ‘White—not of Hispanic origin’, ‘Black—not of Hispanic origin’, ‘Hispanic’, ‘Asian’ or ‘Pacific Islander’, and ‘other’ and grouped for analysis as ‘non-Hispanic White’ or ‘other’. Smoking status was categorized as ‘Never’: those who never smoked cigarettes; ‘Past’: those who ever reported smoking 6 months or longer and were not smoking at baseline; and ‘Current’: those who reported smoking at baseline. BMI was not included because it may be in the causal pathway between dietary intake and AMD risk. AREDS treatment was defined as ‘Active’ for subjects in the antioxidants only, zinc only, or antioxidants–zinc group and ‘Placebo’ as subjects with placebo assignment. Multivitamin intake was defined as ‘No’ for those who never reported taking multivitamins or Centrum and ‘Yes’ for those who had taken a multivitamin supplement/Centrum in the past. Using the AREDS severity scale [41], non-advanced AMD was defined based on drusen area and size, increased pigment or depigmentation, and each eye was graded separately at baseline. AMD severity group 1 (scale 1, eyes with no AMD) was not included in our analytic data set. Scales 2 to 8 were then grouped into 2 categories to increase sample size within each group representing non-advanced levels of AMD risk—1) baseline AMD severity group 2: early AMD (scales 2–4) and 2) baseline AMD severity group 3: intermediate AMD (scales 5–8). Baseline AMD status of the fellow eye was categorized as non-advanced (scales 1–8) or AAMD with GA or NV (scales 9–12). A family history of AMD was self-reported as no family member affected with AMD, 1 family member affected with AMD, or ≥2 family members affected with AMD. A genetic risk score (GRS) was computed from β estimates of 12 loci from 9 genes shown to be most strongly associated with progression to AAMD, adjusted for age, sex, and race—complement factor H (CFH) Y402H (rs1061170); age-related maculopathy susceptibility 2/high-temperature requirement A serine peptidase 1 (ARMS2/HTRA1) A69S (rs10490924); CFH (rs1410996); CFH R1210C (rs121913059); complement component 3 (C3) R102G (rs2230199); C3 K155Q (rs147859257); RAD51 paralog B (RAD51B; rs8017304); transforming growth factor β receptor type 1 (TGFBR1; rs334353); ATP-binding cassette transporter (ABCA1; rs1883025); heat shock protein family H (Hsp110) member 1/β3-glucosyltransferase (HSPH1/B3GALTL; rs9542236); TIMP metallopeptidase inhibitor 3 (TIMP3; rs9621532); and solute carrier family 16 member 8 (SLC16A8; rs8135665) [38]. The follow-up time was restricted to 5 years to reduce misclassification due to changes in diet.

Exposure

Dietary data were obtained at baseline using a validated, self-administered, 90-item semiquantitative AREDS food frequency questionnaire based on the National Cancer Institute Health Habits and History Questionnaire (v2.1). Individuals were asked to report frequency of consumption of each food or beverage item on average, during the past year. The University of Minnesota Nutrition Coordinating Center Food Composition Database (v31) was used to estimate quantity of nutrient and caloric intake, and DietSys software (v3.0: Block Dietary Data Systems) was used to derive individual nutrient values for each food frequency questionnaire item. Nutrient estimates excluded any additional intake from oral supplementation. Weekly food consumption of green leafy vegetables [consisting of weekly ½ cup serving of spinach (raw or cooked), greens (cooked), mustard greens, turnip greens, collards] and fish intake (broiled or baked fish such as tuna, salmon, mackerel, trout) were categorized as sex-specific quintiles. Nutrient estimates of the active biological constituents of the foods: lutein/zeaxanthin (LZ) and ω-3 (n–3) fatty acids (FAs) were calorie adjusted separately for men and women and were then categorized as quintiles.

Outcome

We defined progression as an eye transitioning from baseline AMD severity group 2 (early AMD) to either group 3 (intermediate AMD) or group 4 (AAMD) or from a baseline AMD severity group 3 (intermediate AMD) to group 4 (AAMD). Progression was defined to include either of these transition stages and had to be confirmed at 2 consecutive visits, ≥6 months apart over 5 years (some examples are shown in Figure 2). Time since baseline (quantified by the visit number) was used as the time scale for this analysis. An eye was censored at the earliest time of reaching a positive outcome, at 5 years if an outcome was not reached, loss to follow-up, or death. For example, an eye with AMD severity group 2 at baseline, which transitioned to AMD severity group 3 at visits in both years 3 and 4 was considered to be an outcome at year 4 and follow-up was stopped at that point (Figure 2, patient A).

FIGURE 2.

Illustration of time at risk for progression to outcome and censoring. Patients A–E had AMD severity group 2 or 3 at baseline (year 0). Patient A with baseline severity group 2 and follow-up visits in year 2, 3, and 4 with AMD severity group for their eye under study graded as 2, 3, and 3, respectively, and thus was considered a progressor at year 4 (at the confirmation visit). Patient B (OD) with baseline AMD severity group 3 was considered as a non-progressor in year 4 (at the end of follow-up). Patient C (OD) with AMD severity group 2 at baseline, transitioned to AMD severity group 4 in year 2 and was graded as AMD severity group 3 in year 3. The eye transitioned through both the higher advanced and non-advanced severity stages and was confirmed as an outcome in year 3 at a non-advanced stage. Patient D (OS) with baseline AMD severity group 2, was considered a progressor as there were 2 consecutive visits with AMD severity group 3 in years 1 and 2. Patient E (OS) with baseline AMD severity group 2 was graded as AMD severity group 3 in year 3 with no transition confirmation in any of the consecutive visits and was considered as a non-progressor at year 5. Patient F (OS) with baseline AMD severity group 3 was considered a progressor as there were 2 consecutive visits with AMD severity group 4 in years 4 and 5. AMD, age-related macular degeneration; OD, right eye; OS, left eye.

Statistical analyses

We assessed transition to higher severity groups over 5 years using a Cox proportional hazards model, with the individual eye as the unit of analysis to account for inter-eye correlation (PROC PHREG of SAS with the covariate aggregate option). After initial assessment of individual food groups and nutrients, we further assessed the combined contributions of dietary factors on progression to higher severity groups and incorporated foods (green leafy vegetables, fish) and nutrients (LZ, ω-3 FAs) in 2 separate multivariate models, adjusting for baseline nondietary covariates [38]. In addition, we estimated direct adjusted cumulative incidence curves for 4 levels of adherence to dietary consumption of the food groups: 1) high adherence to both green leafy vegetables and fish; 2) high intake of green leafy vegetables only; 3) high intake of fish only; and 4) low adherence to both green leafy vegetables and fish, assuming mean levels of the other covariates. Statistical tests were performed as 2-tailed at an α level of 0.05. All statistical analyses were performed using SAS software (v9.4).

Results

AMD transitions

Table 1 displays the changes in AREDS severity scale from baseline to 5 years in our analytic data set of 2697 eyes. In eyes with baseline non-advanced AMD severity scales 2–8, while 17% of eyes progressed to AAMD, 37% of eyes transitioned to other higher non-advanced severity scales at 5 years, which would not be counted as outcomes in typical survival analyses using only AAMD as an end point. Nevertheless, an eye usually transitions through these stages before eventually progressing to AAMD. After confirmation of outcome across 2 consecutive visits, 268 eyes (10%) progressed to higher non-advanced stages and 348 eyes (13%) progressed to AAMD. Among the 616 progressing eyes (23%), 300 were in baseline severity group 2 and 316 were in baseline severity group 3 (Table 2). The mean baseline age was 69.4 years (SD: 5.2) and mean follow-up time was 4.7 years (SD: 0.8; range: 1–5 years). A cross-tabulation of individual AREDS severity scale at baseline compared with AREDS severity scale at 5 years helped visualize these transitions through stages of AMD (Supplemental Table 1). Because most of the transitions occurred in subjects with baseline severity scales 2–8, we restricted our analyses to these eyes and excluded eyes with severity scale 1 at baseline. Similar to our analytic data set, the full AREDS data set (Supplemental Tables 2 and 3) also had a higher percentage of eyes transitioning to a higher non-advanced stage than those reaching AAMD at 5 years (29% compared with 10%). The focus of this study was to study associations of risk and protective factors (particularly nutritional factors) with these transitions.

TABLE 1.

Change in AREDS severity scale from baseline to 5 years for the analytic data set (N = 2697 eyes).

AREDS severity scale at baseline	Decreased (%)1	Remained the same (%)1	Outcome not confirmed over 2 consecutive visits over a 5 year period (%)1		Confirmed as an outcome at second consecutive visit (%)1n = 616 (23%)		Total
			Transitioned to higher non-advanced stages	Progressed to AAMD	Transitioned to higher non-advanced stages	Progressed to AAMD
2	177 (28)	236 (37)	199 (31)	2 (0)	24 (4)	5 (1)	643
3	74 (28)	55 (21)	93 (35)	2 (1)	37 (14)	5 (2)	266
4	103 (21)	106 (21)	56 (11)	2 (0)	198 (40)	31 (6)	496
5	74 (24)	59 (19)	140 (45)	13 (4)	0 (0)	24 (8)	310
6	60 (15)	106 (26)	152 (37)	35 (9)	3 (1)	53 (13)	409
7	54 (13)	77 (19)	88 (21)	47 (11)	6 (1)	142 (34)	414
8	38 (24)	23 (14)	0 (0)	10 (6)	0 (0)	88 (55)	159
Total	580 (22)	662 (25)	728 (27)	111 (4)	268 (10)	348 (13)	2697

AAMD, advanced age-related macular degeneration; AREDS, Age-Related Eye Disease Study.

¹All percentages are row percentages.

TABLE 2.

Distribution of subjects and eyes according to baseline AMD severity group in the analytic dataset (N = 2697 eyes), according to progression status.

Baseline severity group	Subjects1 (n)	Eyes (n)	Non-progressing eyes2, n (%)	Progressing eyes2, n (%)
Group 2	1052	1405	1105 (79)	300 (21)
Group 3	885	1292	976 (75)	316 (25)
Total	17571	2697	2081 (77)	616 (23)

Abbreviation: AMD, age-related macular degeneration.

¹Only one eye was included for 817 subjects, while both eyes were included for 940 subjects in either baseline severity groups 2 or 3; therefore, since some subjects had one eye in group 2 and one eye in group 3 the number of subjects in the rows does not equal the total.

²Progression defined as eye in AMD severity group 2 or group 3 at baseline progressing to higher severity groups at 2 consecutive visits over a 5 year follow-up.

Associations of non-dietary covariates and transition to higher AMD severity groups

Analyses of non-dietary covariates by progression status are summarized in Table 3 [37]. Progression was more common among older subjects, those with ≤high school education, current smokers, if the fellow eye was advanced, and if other family members were affected with AMD. With every unit increase in GRS, there was an increased incidence of transitioning to higher severity groups [hazard ratio (HR): 2.40; 95% confidence interval (CI): 1.96, 2.93; P < 0.001], adjusting for nondietary covariates. Similar incidence for transitioning was also seen for GRS tertile 3 compared with that for tertile 1 (HR: 2.38; 95% CI: 1.89, 3.00; P < 0.001), adjusting for non-dietary covariates.

TABLE 3.

Associations between non-dietary covariates and AMD transitions from severity group 2 or group 3 to higher severity groups over 5 years (616 events/2697 eyes), adjusting for other covariates¹.

Variables	Non-progressing eyes (n = 2081), n (%)	Progressing eyes (n = 616), n (%)	Model 12, HR (95% CI)	P	Model 23, HR (95% CI)	P
Non-genetic factors
Age (years)
Mean (SD), range	69.1 (5.2), 55–81	70.5 (5.1), 56–81	1.05 (1.03, 1.06)	<0.001	1.05 (1.03, 1.07)	<0.001
55 to 64	75 (3.6)	17 (2.8)	1.00 (Ref)		1.00 (Ref)
65 to 74	1110 (53.3)	246 (39.9)	1.57 (1.22, 2.01)	<0.001	1.69 (1.30, 2.12)	<0.001
≥75	896 (43.1)	353 (57.3)	1.90 (1.44, 2.50)	<0.001	2.04 (1.52, 2.73)	<0.001
P-trend				<0.001		<0.001
Sex
Female	1179 (56.7)	332 (53.9)	1.00 (Ref)		1.00 (Ref)
Male	902 (43.3)	284 (46.1)	0.94 (0.81, 1.10)	0.43	0.97 (0.80, 1.16)	0.71
Race
Other	63 (3.0)	11 (1.8)	1.00 (Ref)		1.00 (Ref)
Non-Hispanic White	2018 (97.0)	605 (98.2)	1.45 (0.82, 2.59)	0.20	1.07 (0.53, 2.16)	0.85
Education
≤High school	693 (33.3)	238 (38.6)	1.00 (Ref)		1.00 (Ref)
>High School	1388 (66.7)	378 (61.4)	0.84 (0.71, 0.98)	0.03	0.90 (0.76, 1.06)	0.21
Smoking
Never	963 (46.3)	253 (41.0)	1.00 (Ref)		1.00 (Ref)
Past	1027 (49.3)	309 (50.2)	1.12 (0.95, 1.32)	0.17	1.10 (0.92, 1.32)	0.28
Current	91 (4.4)	54 (8.8)	2.13 (1.60, 2.83)	<0.001	1.67 (1.22, 2.30)	0.002
Caloric intake (kcal/d)
Mean (SD), range	1510.09 (567.56), 600.36–3994.63	1557.82 (574.07), 622.64–3869.29	1.19 (1.00, 1.41)4	0.05	1.08 (0.90, 1.30)4	0.42
AREDS treatment group
Placebo	491 (23.6)	149 (24.2)	1.00 (Ref)		1.00 (Ref)
Active5	1590 (76.4)	467 (75.8)	0.97 (0.81, 1.16)	0.74	0.95 (0.78, 1.15)	0.61
Multivitamin intake
No	1089 (52.3)	315 (51.1)	1.00 (Ref)		1.00 (Ref)
Yes	992 (47.7)	301 (48.9)	1.02 (0.87, 1.19)	0.83	0.98 (0.83, 1.15)	0.78
Fellow eye status at baseline
Non-advanced	1887 (90.7)	445 (72.2)	1.00 (Ref)		1.00 (Ref)
Advanced	194 (9.3)	171 (27.8)	2.90 (2.44, 3.45)	<0.001	2.37 (1.97, 2.87)	<0.001
AMD family history
No family member affected	1503 (72.2)	390 (63.3)	1.00 (Ref)		1.00 (Ref)
1 family member affected	401 (19.3)	146 (23.7)	1.39 (1.15, 1.68)	<0.001	1.18 (0.97, 1.45)	0.10
≥2 family members affected	177 (8.5)	80 (13.0)	1.60 (1.23, 2.01)	<0.001	1.38 (1.07, 1.78)	0.02
P-trend6				<0.001		0.006
Genetic risk score7
GRS_continuous
Mean (SD), range	1.60 (0.46), 0.26–3.11	1.84 (0.45), 0.28–2.99	2.83 (2.35, 3.41)	<0.001	2.40 (1.96, 2.93)	<0.001
GRS tertiles
GRS tertile 1	775 (37.2)	124 (20.1)	1.00 (Ref)		1.00 (Ref)
GRS tertile 2	718 (34.5)	182 (29.5)	1.53 (1.22, 1.92)	<0.001	1.39 (1.10, 1.77)	0.006
GRS tertile 3	588 (28.3)	310 (50.3)	2.87 (2.32, 3.55)	<0.001	2.38 (1.89, 3.00)	<0.001
P-trend8				<0.001		<0.001

Abbreviations: AMD, age-related macular degeneration; AREDS, Age-Related Eye Disease Study; CI, confidence interval; GRS, genetic risk score; HR, hazard ratio; SD, standard deviation.

¹Progression defined as eye in AMD severity group 2 or group 3 at baseline progressing to higher severity groups at 2 consecutive visits over a 5 year follow-up period.

²Model 1: Cox proportional hazard analyses adjusted for age groups and baseline AMD severity group.

³Model 2: multivariate Cox proportional hazard analyses adjusted for baseline AMD severity group in addition to all non-dietary and genetic covariates in the table.

⁴Relative risk normalized per 1340 kcal/d (average caloric intake of females and males combined).

⁵Active group consists of subjects treated by antioxidants only, zinc only, or a combination of antioxidants and zinc.

⁶Estimate indicates an increased incidence of progression with each additional family member affected with AMD.

⁷GRS calculated as sum of β coefficients multiplied by number of minor alleles for 12 genetic variants [37], after adjusting for age, sex, and race.

⁸Estimate indicates an increased incidence of progression with increase in GRS tertile.

Association between foods/nutrients and progression to higher AMD severity groups

Results for intake of individual food groups and nutrients by quintiles (mean and ranges), adjusting for other non-dietary covariates are provided in Table 4. Adjusting for age groups and baseline AMD severity group (Table 4, model 1), significant protective trends were seen with a higher consumption of green leafy vegetables, fish, LZ, and ω-3 FA compared to lower consumption (P-trends ranged from 0.005 to 0.03). After adjusting for other non-dietary covariates (Table 4, model 2), the protective effect of higher consumption of green leafy vegetables on transition to higher severity groups persisted [HR for weekly consumption of ≥2.7 servings compared with none (quintile 5 compared with quintile 1): 0.72; 95% CI: 0.56, 0.92; P-trend = 0.02]. Effects in a protective direction were also seen for intake of 2 servings of fish per week compared with less than once per month (HR quintile 4 compared with quintile 1: 0.80; 95% CI: 0.61, 1.05) and >2 servings of fish per week compared with less than once per month (HR quintile 5 compared with quintile 1: 0.81; 95% CI: 0.61, 1.07; P-trend = 0.04). Supporting these associations, there were effects in a protective direction for the biologically active components of these foods; high intake of LZ of >2 mg/d compared with <1 mg/d (HR quintile 5 compared with quintile 1: 0.72; 95% CI: 0.55, 0.95; P-trend = 0.07) and high ω-3 FA intake > 0.7 g/wk compared with <0.28 g/wk (HR quintile 5 compared with quintile 1: 0.82; 95% CI: 0.62, 1.07; P-trend = 0.08), although the trend in a protective direction for higher ω-3 FA intake was not statistically significant in this model.

TABLE 4.

Associations between dietary and nutritional factors and AMD transitions from severity group 2 or group 3 to higher severity groups over 5 years (616 events/2697 eyes), adjusting for other covariates¹.

Food group and nutrient quintiles2, mean intake (range)	Non-progressing eyes, n (%)	Progressing eyes, n (%)	Model 13, HR (95% CI)	P-trend4	Model 25, HR (95% CI)	P-trend4
Food groups
Green leafy vegetables (servings/wk)6
Quintile 1: 0.00 (0.00–0.00)	625 (30.0)	215 (34.9)	1.00 (Ref)	0.005	1.00 (Ref)	0.02
Quintile 2: 0.12 (0.12–0.12)	114 (5.5)	33 (5.4)	0.82 (0.57, 1.16)		0.87 (0.59, 1.27)
Quintile 3: 0.32 (0.23–0.53)	447 (21.5)	135 (21.9)	0.88 (0.72, 1.08)		0.87 (0.71, 1.11)
Quintile 4: 0.77 (0.53–1.17)	455 (21.9)	133 (21.6)	0.85 (0.69, 1.05)		0.88 (0.71, 1.10)
Quintile 5: 2.68 (1.05–21.0)	440 (21.1)	100 (16.2)	0.71 (0.56, 0.89)		0.72 (0.56, 0.92)
Fish (servings/wk)
Quintile 1: 0.24 (0.00–0.53)	388 (18.6)	129 (21.0)	1.00 (Ref)	0.008	1.00 (Ref)	0.04
Quintile 2: 0.75 (0.50–1.05)	427 (20.5)	140 (22.7)	0.99 (0.78, 1.25)		1.03 (0.81, 1.31)
Quintile 3: 1.27 (1.00–1.58)	402 (19.3)	140 (22.7)	1.03 (0.82, 1.30)		1.10 (0.87, 1.39)
Quintile 4: 1.95 (1.50–2.50)	430 (20.7)	100 (16.2)	0.74 (0.58, 0.96)		0.80 (0.61, 1.05)
Quintile 5: 3.88 (2.33–12.25)	434 (20.9)	107 (17.4)	0.79 (0.62, 1.01)		0.81 (0.61, 1.07)
Nutrients
Lutein/zeaxanthin (mg/d)
Quintile 1: 0.73 (0.24–0.99)	405 (19.5)	133 (21.6)	1.00 (Ref)	0.009	1.00 (Ref)	0.07
Quintile 2: 1.09 (0.89–1.29)	408 (19.6)	133 (21.6)	1.05 (0.83, 1.32)		0.94 (0.74, 1.21)
Quintile 3: 1.40 (1.18–1.65)	409 (19.7)	131 (21.3)	0.97 (0.77, 1.22)		1.00 (0.78, 1.27)
Quintile 4: 1.88 (1.54–2.32)	415 (20.0)	124 (20.1)	0.96 (0.76, 1.22)		1.00 (0.78, 1.30)
Quintile 5: 3.25 (2.17–12.01)	444 (21.3)	95 (15.4)	0.72 (0.55, 0.92)		0.72 (0.55, 0.95)
ω-3 fatty acids (g/wk)
Quintile 1: 0.14 (0.00–0.28)	404 (19.4)	134 (21.8)	1.00 (Ref)	0.03	1.00 (Ref)	0.08
Quintile 2: 0.35 (0.21–0.49)	413 (19.8)	127 (20.6)	0.95 (0.75, 1.20)		1.00 (0.78, 1.28)
Quintile 3: 0.49 (0.35–0.70)	405 (19.5)	136 (22.1)	1.03 (0.82, 1.29)		1.09 (0.86, 1.39)
Quintile 4: 0.70 (0.49–0.98)	428 (20.6)	111 (18.0)	0.81 (0.64, 1.04)		0.87 (0.67, 1.13)
Quintile 5: 1.82 (0.77–12.04)	431 (20.7)	108 (17.5)	0.80 (0.63, 1.02)		0.82 (0.62, 1.07)

Abbreviations: AMD, age-related macular degeneration; AREDS, Age-Related Eye Disease Study; CI, confidence interval; GRS, genetic risk score; HR, hazard ratio.

¹Progression defined as eye in AMD severity group 2 or group 3 at baseline progressing to higher severity groups at 2 consecutive visits over a 5-year follow-up period.

²Quintiles are defined as sex-specific food and nutrient quintiles. Therefore, some of the quintile boundaries may overlap.

³Model 1: Cox proportional hazard analyses adjusted for age and AMD baseline severity group.

⁴P-trend indicates increase in protective effect per quintile intake increase.

⁵Model 2: multivariate Cox proportional hazard analyses adjusted for age groups, sex, race, education, smoking status, caloric intake, multivitamin use, AREDS treatment group, fellow eyes status, AMD baseline severity group, family history of AMD and GRS (continuous).

⁶One medium serving of green leafy vegetables may include (but not limited to) ½ cup of spinach (raw or cooked), greens (cooked), mustard greens, turnip greens, collards.

Results of the multivariate Cox proportional hazards models for the joint contributions of both foods and nutrients to AMD transitions are displayed in Figure 3 and Table 5. In the multivariate food model jointly adjusted for green leafy vegetables and fish, the HR for consumption of ≥2.7 servings of green leafy vegetables per week compared with none was 0.75 (95% CI: 0.59, 0.96; P = 0.02), and the HR for ≥2 servings of fish per week compared with <2 was 0.79 (95% CI: 0.65, 0.95; P = 0.01), after adjusting for other variables. In the nutrient model, jointly adjusted for LZ and ω-3 FA, the HR for consuming ≥2 mg LZ/d compared with <2 was 0.76 (95% CI: 0.60, 0.96; P = 0.02), and the HR for ω-3 FA consumption ≥0.7 g/wk compared with <0.7 g was 0.85 (95% CI: 0.71, 1.01; P = 0.06), after adjusting for other variables. Furthermore, eyes of subjects who had higher intake for both dietary food groups had a 41% lower incidence rate than eyes of subjects with lower intake of both green leafy vegetables and fish (HR: 0.59; 95% CI: 0.44, 0.79; P < 0.001). Eyes of subjects with dietary nutrient intake of ≥2 mg/d of LZ and ≥0.7 g/wk ω-3 FAs compared with <2 mg/d LZ and <0.7 g/wk ω-3 FAs had a 36% lower incidence rate of progressing to higher severity groups (HR: 0.64; 95% CI: 0.49, 0.84; P = 0.001).

FIGURE 3.

Multivariate Cox proportional hazards analyses identifying nutritional factors related to AMD transitions from baseline severity group 2 or 3 to higher AMD severity groups over a 5 year follow-up period, adjusted for the other food group/nutrient in respective models, in addition to age groups, sex, race, education, smoking status, caloric intake, multivitamin use, AREDS treatment groups, fellow eye status, baseline AMD severity group, a family history of AMD, and GRS (continuous). One medium serving of green leafy vegetables may include (but not limited to) ½ cup of spinach (raw or cooked), greens (cooked), mustard greens, turnip greens, collards. AMD, age-related macular degeneration; AREDS, Age-Related Eye Disease Study; CI, confidence interval; GRS, genetic risk score; HR, hazard ratio.

TABLE 5.

Multivariate Cox proportional hazards analyses of associations between nutritional factors and AMD transitions (616 progressing eyes of 2697 eyes over 5 years of follow-up)¹.

Food group model	HR1 (95% CI); P	Nutrient model	HR2 (95% CI); P
Green leafy vegetables2 (≥2.7 servings/wk vs. none)	0.75 (0.59, 0.96); 0.02	Lutein/zeaxanthin (≥2 mg/d vs. <2)	0.76 (0.60, 0.96); 0.02
Fish (≥two 4 oz servings/wk vs. <2)	0.79 (0.65, 0.95); 0.01	ω-3 fatty acids (≥0.7 g/wk vs. <0.7)	0.85 (0.71, 1.01); 0.06
Contrasting adherence to diet with ≥2.7 servings/wk of green leafy vegetables and ≥two 4 oz servings of fish/wk compared with no green leafy vegetables and <2 servings of fish/wk3	0.59 (0.44, 0.79); <0.001	Contrasting adherence to dietary intake of ≥2 mg/d of LZ and ≥0.7 g/wk ω-3 fatty acids compared with <2 mg/d LZ and <0.7 g/wk ω-3 fatty acids3	0.64 (0.49, 0.84); 0.001

Abbreviations: AMD, age-related macular degeneration; AREDS, Age-Related Eye Disease Study; CI, confidence interval; GRS, genetic risk score; HR, hazard ratio; LZ, lutein/zeaxanthin.

¹Multivariate Cox proportional hazards analyses identifying nutritional factors related to AMD transitions from baseline severity group 2 or 3 to higher AMD severity groups, adjusted for the other food group/nutrient in respective models, in addition to age groups, sex, race, education, smoking status, caloric intake, multivitamin use, AREDS treatment groups, fellow eye status, baseline AMD severity group, a family history of AMD, and GRS (continuous).

²One medium serving of green leafy vegetables may include (but not limited to) ½ cup of spinach (raw or cooked), greens (cooked), mustard greens, turnip greens, collards.

³Contrasts estimated using linear combination of β estimates from each model.

Figure 4 displays directly adjusted cumulative incidence curves for the risk of transitioning to higher AMD severity, stratified by different levels of dietary intake of green leafy vegetables and/or fish. Results indicate a 22% (95% CI: 20%, 24%) risk of transitioning to higher severity in those with low adherence to consumption of both green leafy vegetables and fish compared with 13% (95% CI: 9%, 17%) risk for those with high adherence to both food groups at the end of 5 years of follow-up. Adherence to higher intake of only 1 of these food groups was associated with 18%–19% risk of transitioning. These results suggest 41% [(22%–13%)/22%] of outcomes could be prevented by sufficient intake of green leafy vegetables and fish over 5 years.

FIGURE 4.

Cumulative incidence curves for probability of eyes transitioning to higher AMD severity, stratified by adherence to green leafy vegetables and/or fish consumption. Results adjusted for age groups, sex, race, education, smoking status, caloric intake, multivitamin use, AREDS treatment groups, fellow eye status, baseline AMD severity group, a family history of AMD, and GRS (continuous). N indicates number of eyes. AMD, age-related macular degeneration; AREDS, Age-Related Eye Disease Study; GRS, genetic risk score.

Discussion

Main results

Results indicate beneficial associations between dietary intake of green leafy vegetables, fish, and nutrients LZ and ω-3 FAs and subsequent transitions from early or intermediate AMD to higher AMD severity groups. We found that moderate consumption of these individual nutritional components were each independently associated with a 15%–25% lower incidence rate of progressing to higher AMD severity, compared with no or low consumption. Furthermore, individuals consuming both ≥2.7 servings/wk of green leafy vegetables and ≥two 4 oz fish/wk had a lower incidence rate of progressing to higher AMD severity over 5 years compared with individuals with lower intake, independent of other risk factors. In addition, cumulative incidence of transitioning over a 5 year period was higher among individuals who did not adhere to these dietary recommendations and could be reduced by 41% among individuals who had high consumption of both food groups. Our study provides novel insights about progression from the earlier stages of AMD to higher AMD severity and the long-term impact of dietary behaviors on the course of a chronic disease, independent of demographics, lifestyle, baseline macular status, and genetic susceptibility.

Our approach builds upon rigorous statistical procedures in longitudinal studies where the eye is the unit of analysis and the Cox proportional hazards model is used to determine progression [42,43]. In addition, we present an alternative method of analysis for longitudinal AMD data, similar to methods used for longitudinal analyses of diabetic retinopathy severity [[44], [45], [46]]. Analogous to the AMD severity staging, the Early Treatment Diabetic Retinopathy Study scale accounts for different stages of the disease before progressing to end-stage advanced proliferative retinopathy [47]. In this study, we used the AMD severity groups to define outcomes to both mild and moderate non-advanced stages as well as advanced stages with GA and NV. Traditional survival models in previous prospective cohort studies analyzing predictive factors including nutritional factors from dietary intake in AMD have had a greater focus on progression to AAMD [11,33,40,[48], [49], [50], [51], [52], [53]], neglecting the significance of changes in earlier stages of AMD and the dynamic nature of AMD progression. To address this gap, our study used an innovative approach by incorporating changes in both eyes across early and intermediate stages of AMD and confirming these changes over 2 consecutive visits. This modeling approach allowed us to efficiently use ocular longitudinal data from all severity groups, especially because some eyes transition to non-advanced stages, but do not progress to advanced AMD during the study period. By recognizing the nuanced evolution of the disease, our model provides a more detailed understanding of the factors influencing AMD progression. This approach enhances the accuracy of predictions and contributes important insights for informed clinical decision-making, paving the way for more effective public health interventions. Our findings, therefore, offer a distinctive broader conceptualization of AMD progression and promote advancements in patient care.

Previous associations of dietary intake of green leafy vegetables and fish and AMD progression

The observed protective associations between nutritional factors and transitions between stages of AMD are driven mainly by the consumption of green leafy vegetables and fish rich in LZ and ω-3 FAs. Analysis of the individual food groups and dietary patterns incorporating green leafy vegetables and fish such as the Mediterranean diet have also identified protective effects on progression to advanced stages of AMD [11,33,40,[48], [49], [50], [51], [52], [53], [54]]. Insights from our previous analysis examining 2-step drusen size progression, genetic susceptibility, and diet quality revealed that a medium/high adherence to a healthful nutrient-rich diet, including fruits, vegetables, legumes, and fish, significantly mitigated the enlargement of drusen, a hallmark of AMD [36]. Drusen size is incorporated in the staging of non-advanced AMD in addition to RPE abnormalities. In this article, we broaden the focus on progression through non-advanced stages of AMD and highlight the targeted benefits of the foods and their active biologic components during these stages.

Dietary intake of LZ and AMD progression

Macular pigments in the fovea are abundant in lutein and zeaxanthin carotenoids. These pigments play a crucial role in safeguarding macular health [55]. The body does not naturally synthesize LZ, and hence, it is important to incorporate them into one’s diet through food or supplementation to support ocular health. The first study to show the beneficial effects of higher intake of lutein and zeaxanthin nutrients in the diet was a case–control study of exudative AMD published in 1994 [8]. Specifically, spinach intake of 5–6+ servings/wk was associated with an 86% reduced risk of advanced neovascular AMD (odds ratio: 0.14; 95% CI: 0.01, 1.20; P-trend < 0.001), and dietary LZ intake of 6 mg/d compared with <1 mg/d was associated with a 57% reduced risk (odds ratio: 0.43; 95% CI: 0.20, 0.70; P-trend < 0.001). Subsequent case–control and longitudinal studies also reported protective associations of dietary carotenoid intake on AAMD [53,[56], [57], [58], [59], [60], [61]] with results summarized in reviews [62]. The Nurses’ Health Study and Health Professionals Follow-Up study in 2015 found an inverse protective association of LZ intake on progression to AAMD but not intermediate AMD [60]. In our study, incidence of transitioning from early and intermediate AMD to higher severity groups decreased by 24% for a daily dietary LZ intake of ≥2 mg (median intake = 6 mg/d) across 5 years, after controlling for other factors including ω-3 FA.

Dietary intake of ω-3 FAs in early/intermediate AMD and progression to AAMD

ω-3 FAs are a class of polyunsaturated fatty acids, including the long-chain fatty acids EPA and DHA. In the retina, ω-3 FAs are essential for maintaining the integrity and fluidity of the photoreceptor membranes [63]. The first analyses to suggest a beneficial effect of ω-3 FAs on AMD were reported more than two decades ago [10,64,65]. Many studies have since also suggested a protective effect of ω-3 FAs on early and AAMD [11,15,26,40,48,49,[66], [67], [68], [69]]. However, literature is sparse concerning the effect of dietary ω-3 FAs on transitioning to higher non-advanced and advanced stages of AMD severity, adjusting for covariates included in this study. In our analysis, we observed a notable 15% reduction in the incidence of progression of AMD to more severe stages associated with a weekly dietary intake of 0.7 g of ω-3 FAs. This reduction persisted even after accounting for other covariates including LZ, although the protective trend did not reach statistical significance.

Supplementary intake of LZ and/or ω-3 FAs in early/intermediate stages and progression to AAMD

The AREDS2 randomized controlled trial analyzed the long-term effect of supplemental LZ and ω-3 FAs in addition to the original AREDS formulation on progression to AAMD in participants with bilateral or unilateral intermediate drusen. The primary study results at the 5 year closeout did not demonstrate efficacy of supplementation of LZ, ω-3 FAs, or a combination in addition to the original AREDS supplements [70]. However, when subjects were stratified by initial dietary LZ quintiles, there were significant effects within quintile 1 (low intake) for additional LZ supplementation compared with no LZ supplementation (HR: 0.74; P = 0.01), but not within quintiles 2–5, suggesting LZ supplementation might be protective only among subjects with low dietary LZ intake. After 10 years of follow-up in AREDS2 [71], an inverse association was found with LZ supplementation (HR: 0.88) but not with supplementation of ω-3 FAs (HR: 0.97) for progression to AAMD. The multivariate results from our 5 year prospective longitudinal analyses, including non-advanced and advanced events, show stronger protective effects for dietary intake of LZ and ω-3 FAs on slowing AMD progression. Our findings underscore the advantages of maintaining healthy dietary habits and highlight how even modest alterations in behavior can significantly affect transitions between AMD stages.

Strengths and limitations

The strengths of this study include incorporation of transitions through both earlier stages as well as development of AAMD. Because there is variability of assessment of AMD severity scales, another strength of the study was the requirement that transitions be confirmed at 2 consecutive visits (usually 6 months apart) to be counted as an end point, which reduced misclassification. Because similar results were observed for effects of the nutritional factors on progression from early to intermediate and intermediate to advanced AMD, combining longitudinal data from the two non-advanced stages of AMD increased the number of end points and increased power. One limitation of the study is that we examined dietary intake at baseline. However, restricting the follow-up period to 5 years reduced the possible effect of confounding by changes in diet over longer periods.

Conclusions

Exploring the transitions from early and intermediate AMD to higher severity stages and the contribution of diet and nutrition to these transitions enhances early prevention and holistic care. Results obtained when incorporating nutritional factors, after controlling for other non-genetic and genetic factors, provide actionable insights to enrich diets with foods such as dark green leafy vegetables (e.g.: spinach—raw or cooked, kale, collards, mustard greens, turnip greens), foods rich in LZ (e.g.: peas, summer squash, corn, pumpkin, brussels sprouts, broccoli, asparagus, lettuce, carrots, egg yolks) [72] and broiled/baked fatty fishes (e.g.: salmon, sardines, mackerel, tuna, and trout) rich in EPA/DHA [73]. Benefits of healthy nutrient-rich food choices include gaining antioxidant benefits and reduction of inflammatory biomarkers including those associated with aging and other chronic diseases such as C-reactive protein [13,18,74,75]. In addressing the challenges of a disease primarily affecting the elderly, dietary modifications tailored to the aging process are essential. Challenges associated with access to quality food products, safe food handling, nutrition counseling, food security, drug–nutrition interactions and growing disparities in diet quality also need to be considered in the elderly population [[76], [77], [78], [79]].

Clinical relevance

The recent concept of “food is medicine” integrates intake of nutrient-rich healthful foods to overall well-being [80], and increasing fruit and vegetable consumption has been shown to improve cardiometabolic health outcomes [81]. Physicians should emphasize the benefits of dietary and lifestyle modifications of patients during the earlier stages of AMD to slow progression of this chronic disease. Efforts should be made to instill and reinforce these habits, ensuring their sustainability throughout an individual’s lifetime. The implication for physicians is to adopt a more holistic approach in patient care, focusing not only on advanced stages but also on the earlier transitions. This study contributes to evidence improving ophthalmic practice by guiding eye care professionals to intervene proactively during the earlier stages of AMD, delaying progression to non-advanced and advanced stages, thus leading to better patient outcomes, preserving vision, and improving overall quality of life.

Author contributions

The authors’ responsibilities were as follows – JMS: designed the research, wrote the article, and had primary responsibility for final content; DD: performed statistical analyses and wrote the article; BR: designed the research, performed statistical analysis, and wrote the article; and all authors: have revised and read and approved the final manuscript.

Conflict of interest

JMS has equity interests in Gemini Therapeutics and Apellis Pharmaceuticals and has been a consultant for Laboratoires Thea. DD and BR declare no conflict of interest.

Funding

Supported in part by National Institutes of Health, National Eye Institute grants R01-EY011309, R01-EY028602; and R01-EY022445; the American Macular Degeneration Foundation; and the Macular Degeneration Center of Excellence, University of Massachusetts Chan Medical School, Department of Ophthalmology and Visual Sciences. The sponsor or funding organization had no role in the design or conduct of this research.

Data availability

Data described in the manuscript, code book, and analytic code will be made available upon reasonable request.

Appendix A. Supplementary data

The following is the Supplementary data to this article: