Table 1.

Characteristics of the eight studies included in this review evaluating the association between the Mediterranean diet and AMD.

Article Country Study Name and Design	Period of Data Collection Sample Size Age and Sex	Exposure and Outcome Assessments	Outcome and Compared Variables	Adjusted Confounders	OR or HR (95% CI) and p-Value	Study Quality	Notes (See Main Text for Further Comments)
› Mares et al. 2011 [36] › USA › CAREDS: cross-sectional nested in WHIOS (prospective)	› CAREDS baseline: 2001–2004; WHIOS baseline: 1994–1998 › 1313 › 55–74 › F	› Validated, semiquantitative FFQ at WHIOS baseline (122 items) › aMED score (0–9) › aMED quartiles: Q1 = 0–1; Q2 = 2–3; Q3 = 4–5; Q4 = 6–9 › Fundus stereoscopic photography › AMD grading based on a modified Wisconsin grading classification	› Early AMD in at least one eye (n = 187) › aMED Q4 (n = 53) vs. aMED Q1 (n = 490)	(a) Model 1: age, pack-years smoked, history of diabetes, AMD, CVD and HRT, and iris colour (b) Model 2: further adjustment for physical activity	› Model 1: OR = 0.34 (0.08–0.98) p = 0.046 › Model 2: OR = 0.44 (0.10–1.27) p = 0.23	High (8)	› Selected participantshad intakes of lutein plus zeaxanthin that were above the 78th and below the 28th percentiles › aMED Q4: small sample size › Evaluation of the diet using the HEI showed similar results
› Merle et al. 2015 [48] › USA › Prospective cohort within AREDS (RCT)	› 13 years (enrolment 1992–1998) › 2525 › 55–80 at baseline › M and F	› Validated, self-administered, semiquantitative FFQ (90 items) at AREDS baseline › aMED score (0–9) › aMED tertiles: T1 = low (0–3), T2 = medium (4–5), T3 = high (6–9) › Retinal stereoscopic images › AMD grading at baseline based on the CARMS system	› Progression to advanced AMD (n = 1028) › aMED T3 (n = 676) vs. aMED T1 (n = 852)	› Model 1: age, sex, AREDS treatment, AMD grade at baseline, TEI › Model 2: further adjustment for education, smoking history, BMI, supplement use, and 10 genetic variants (SNPs)	› Model 1: HR = 0.74 (0.61–0.90) p = 0.005 › Model 2: HR = 0.74 (0.61–0.91) p = 0.007	High (7)	› Evaluation of the interaction between aMED score and genetic variations on risk of AMD (10 SNPs analysed in 7 different genes) › Fish and vegetable consumption was associated with lower odds of progression
› Hogg et al. 2017 [45] › Europe (Norway, Estonia, UK, France, Italy, Greece, Spain) › Cross-sectional, within EUREYE study (cross-sectional study with retrospective and current exposure measurements)	› 2001–2002 › 4753 › Mean age = 73.2 ± 5.6 years › M and F	› Semiquantitative FFQ (130 goods) tailored to each country › MDS (0–9) from Martinez-Gonzalez et al. 2004 › MDS score quartiles: Q1 = ≤ 4, Q2 = 5, Q3 = 6, and Q4 = > 6 › Full eye examination and stereoscopic colour fundus digital photography › AMD graded according to the ICS for age-related maculopathy	› Presence of AMD: early (n = 2333), large drusen (n = 641), GA (n = 49), nvAMD (n = 109); control (n = 2262) › Q4 (n = 199) vs. Q1 (n = 787)	› Model 1: unadjusted › Model 2: age, sex, country, education, smoking, drinking, history of CVD, aspirin consumption, and diabetes	› Model 1: Early AMD OR = 0.94 (0.85–1.03) p = 0.4 Large drusen OR = 0.79 (0.65–0.97) p = 0.05 nvAMD OR 0.52 (0.29–0.93) p = 0.03 › Model 2: Early AMD OR = 0.96 (0.83–1.11) p = 0.9 Large drusen OR = 0.80 (0.65–0.98) p = 0.1 nvAMD OR = 0.53 (0.27–1.04) p = 0.01	High (9)	› No association between MDS and prevalence of GA
› Nunes et al. 2018 [46] › Portugal › Nested case-control study within the “Epidemiologic Study of the Prevalence of Age-Related Macular Degeneration in Portugal: The Coimbra Eye Study” (cross-sectional) [54]	› 2012–2014 (Coimbra study = 2009–2011) › 1992 › >55 years › M and F	› Validated FFQ (86 items) › mediSCORE (0–9); high adherence = ≥6 › Complete ophthalmological examination and digital mydriatic colour fundus photography › AMD graded according to the ICS for age-related maculopathy (as in Hogg et al. 2016)	› AMD: case group = 768 (control = 1224, age and sex-matched) › High mediSCORE vs. prevalence of AMD	› Age, sex, BMI, abdominal perimeter, physical activity, smoking status, diabetes, and hypertension	› OR = 0.73 (0.58–0.93) p = 0.009	High (7)	› Food group analysis: higher consumption of vegetables reduced odds of AMD onset by 36% (OR = 0.63 (0.52–0.76), p < 0.001), and higher intake of nuts and fruits lowered odds by 21% (OR = 0.78, (0.65–0.94), p = 0.010) › Cases were significantly older
› Raimundo et al. 2018 [47] › Portugal › Nested case-control study within the “Epidemiologic Study of the Prevalence of Age-Related Macular Degeneration in Portugal: The Coimbra Eye Study” (cross-sectional)	› 2012–2014 (Coimbra study = 2009–2011) › 883 › >55 years › M and F	› Same as Nunes et al. 2018	› AMD: case group = 434 (control = 449, age and sex-matched) › High mediSCORE vs. prevalence of AMD	› Age, sex, smoking, calories consumption	› OR = 0.62 (0.38–0.97) p = 0.041	High (7)	› Physical activity and fruit consumption were higher in controls (p = 0.012 and p = 0.029, respectively) › Consumption of 150 g fruit lowered odds by 10% (OR = 0.90 (0.82–0.98; p = 0.028)
› Merle et al. 2019 [49] › Europe › Prospective cohort study of the Rotterdam Study I (RS-I) and Antioxydants, Lipides Essentiels, Nutrition et maladies Oculaires (Alienor) study populations, part of the EYE-RISK project	RS-I › 21 years (1990–2011, mean follow-up time 9.9 y) › 4446 › ≥ 55 years › M and F Alienor › 6 years (2006–2012, mean follow-up time 4.1) › 550 › ≥73 years › M and F	› RS-I: 170-item validated semiquantitative FFQ at baseline › Alienor: 40-item validated FFQ at baseline and a 24 h dietary recall › mediSCORE (0–9) › Three groups: low (0–3), medium (4–5), high (6–9) › Ophthalmologic examinations and fundus photographs › AMD graded based on the Wisconsin Age-Related System (RS-I) and the ICS (Alienor)	› Progression to advanced AMD (n = 155;RS-I = 117; Alienor = 38) with subtype analysis › mediSCORE high (RS-I n = 947; Alienor n = 143) vs. mediSCORE low (RS-I, n = 1376; Alienor, n = 171)	› Model 1: unadjusted › Model 2: age, sex, AMD grade at baseline (no or early AMD), TEI, education, BMI, smoking, multivitamin or mineral supplement use, diabetes, and hypercholesterolemia	› Model 1: RS-I, HR = 0.56 (0.33–0.96) p = 0.036; Alienor, ns; Combined, HR = 0.53 (0.33–0.84) p = 0.009) › Model 2: RS-I and Alienor alone = ns Combined, HR = 0.59 (0.37–0.95) p = 0.04 › No association with nvAMD › GA → RS-I, HR = 0.41 (0.16–1.03) p = 0.046; Alienor, ns; Combined, HR = 0.42 (0.20–0.90) p = 0.04	High (RS-I = 8; Alienor = 7)	› Association remain after adjustment for two AMD-related SNPs › No single Med diet component was associated with the incidence of advanced AMD
› Keenan et al. 2020 [44] › USA › Retrospective analysis of two RCTs: Age-Related Eye Disease Study (AREDS) and AREDS2	› 13 years (median follow-up 10.2 years), enrolment AREDS 1992–1998; AREDS2 2006–2008 › 7756 (13,204 eyes) › 71 ± 6.6 years › 56.5% F	› AREDS: 90-item, validated, semiquantitative FFQ at baseline AREDS2: 131-item, validated semiquantitative FFQ at baseline › aMED score (modified), ranging from 9 to 36 in main analysis, with assessment using quartile ranks (see main text for details), and from 0 to 9 in sensitivity analyses with assessment using sex-specific medians › Population divided in tertiles: T1 = low, T2 = medium, T3 = high › Eye examinations and colour fundus photographs › AMD graded based on the Wisconsin Age-Related System	› Progression to advanced AMD (AREDS, n = 2273; AREDS2, n = 2763), with subtype analysis › T3 (AREDS, n = 1349; AREDS2, n = 1224) and T2 (AREDS, n = 1436; AREDS2, n = 1101) vs. T1 (AREDS, n = 1470; AREDS2, n = 1286)	› Treatment assignment, age, sex, smoking, TEI, BMI (for AREDS only), and correlation between eyes › In combined AREDS/AREDS2 analyses, adjustment was also made for the cohort	› Combined cohort: Advanced AMD HRs = T2: 0.87 (0.80–0.94) p = 0.001; T3: 0.78 (0.71–0.85) p < 0.0001 › Subtypes › GA HRs = T2: 0.80 (0.71–0.90) p = 0.0002; T3: 0.71 (0.63–0.80) p < 0.0001 › nvAMD HRs = T2: 0.90(0.80–1.01) p = 0.08; T3: 0.84 (0.75–0.95) p = 0.005 › Large drusen HR = 0.79 (0.68–0.93) p = 0.004	High (8)	› Analysis of interaction between aMED and genotype: in AREDS, protective effect was present only in subject with one particular protective allele › Sensitivity analyses: results showed similar pattern but were partially attenuated › Analysis of individual components of the Med diet showed that higher fish consumption was inversely associated with AMD progression
› Merle et al. 2020 [50] › USA › Prospective cohort within AREDS (RCT)	› 13 years (enrolment from 1992 to 1998) › 1838 › 55–80 (at baseline) › M and F	› Validated, self-administered, 90-item, semiquantitative FFQ at baseline › aMED (0–9) › Two groups: low aMED (0–3) or medium-high aMED (4–9) › Complete eye examination and retinal stereoscopic colour images › Maximal drusen size graded in a ordinal scale as detailed in the figure legend	› Drusen size progression (n = 587), defined as an eye advancing at least two grades during the study period (from grade 0 to 2, or grade 1 to 3, or grade 2 to 4) › Medium-high aMED vs. low aMED	› Age, sex, education, smoking, BMI, AREDS treatment, multivitamin supplement use, TEI, genetic variants, and maximum drusen size category at baseline in each eye	› HR = 0.83 (0.68–0.99) p = 0.049	High (8)	› Drusen = major hallmark of AMD

Characteristics of the eight studies included in this review evaluating the association between the Mediterranean diet and age-related macular degeneration. Abbreviations: AMD: Age-related Macular Degeneration; aMED: alternative or alternate Mediterranean diet score; AREDS/AREDS2: Age-Related Eye Disease Study; BMI: body mass index; CAREDS: Carotenoids in Age-Related Eye Disease Study; CARMS: Clinical Age-Related Maculopathy Staging; CVD: cardiovascular disease; EUREYE: European Eye (study) [55]; FFQ: Food Frequency Questionnaire; GA: geographic atrophy; HR: Hazard Ratio; HEI: Healthy Eating Index; HRT: hormone replacement therapy; ICS: International Classification System; MDS: Mediterranean diet score; mediSCORE: Mediterranean score; ns: not significant; nvAMD: neovascular AMD; OR: Odds Ratio; RCT: randomised controlled trial; SNPs: single nucleotide polymorphisms; TEI: total energy intake; WARMGS: Wisconsin Age-Related Maculopathy Grading System; WHIOS: Women’s Health Initiative Observational Study [56]. Tertile: a statistical value of a data set representing one-third of a given population; quartile: a statistical value of a data set representing 25% of a given population. AMD grading systems employed in the studies: WARMGS: early AMD = absence of signs of advanced AMD and the presence of (1) soft indistinct or reticular drusen or (2) hard distinct or soft distinct drusen with pigmentary abnormalities. Late AMD = presence of either (1) geographic atrophy or (2) exudative AMD. Exudative AMD is defined as the presence of any of the following exudative lesions: pigment epithelial detachment or age-related retinal detachment, subretinal haemorrhage, subretinal scar (subretinal fibrous scar), or prior laser treatment for exudative AMD [57,58]. CARMS (Clinical Age-Related Maculopathy Staging) = no AMD (few small drusen, <63 μm, grade 1); early AMD (drusen within 63–124 μm, grade 2); intermediate AMD (large drusen ≥ 125 μm grade 3); GA (geographic atrophy, both central and noncentral, grade 4); neovascular disease (hemorrhagic retinal detachment, haemorrhage under the retina or retinal pigment epithelium, subretinal fibrosis, grade 5) [48]. ICS: grade 0 = macula free of drusen or pigmentary irregularities or with hard drusen (<63 μm) only; early AMD is subdivided as follows: grade 1 = soft distinct drusen (≥63 μm) or pigmentary abnormalities; grade 2 = soft indistinct drusen (≥125 μm) or reticular drusen only or soft distinct drusen (≥63 μm) with pigmentary abnormalities; grade 3 = soft indistinct drusen (≥125 μm) or reticular drusen with pigmentary abnormalities; advanced AMD = grade 4 = presence of nvAMD (presence of serous or hemorrhagic retinal or retinal pigment epithelial detachment, subretinal neovascular membrane, periretinal fibrous scar) or GA (well-demarcated area of retinal pigment atrophy with visible choroidal vessels). Large drusen (≥125 μm) in any grade of early AMD also is categorized as a separate outcome [59,60]. In Mares et al. [36], the classification was based on a modified WARMGS and was as follows: early AMD = presence of either (1) large drusen (≥1 large drusen (≥125 μm) or extensive intermediate drusen (area ≥ 360 μm when soft indistinct drusen are present or ≥650 μm when soft indistinct drusen are absent)) or (2) pigmentary abnormalities of the retinal pigment epithelium (an increase or decrease in pigmentation accompanied with ≥1 drusen (≥63 μm)). This manuscript only considered early AMD. In Merle et al. [50] the classification was as followed: 0 = no drusen or questionable drusen; 1 = small drusen (<63 μm); 2 = intermediate drusen (63–124 μm); 3 = large drusen (125–249 μm); 4 = very large drusen (≥250 μm).