A Systematic Review on Zinc for the Prevention and Treatment of Age-Related Macular Degeneration

Abstract

Purpose.: The objective of this systematic review was to examine the evidence on zinc intake from foods and supplements in the primary prevention and treatment of AMD.

Methods.: Randomized controlled trials (RCTs), prospective cohort, retrospective cohort, and case-control studies that investigated zinc intake from foods and/or supplements, and AMD in men and women with a mean age of 50 years or older were included. Medline and Cochrane Central were searched from inception to February 2012 and November 2012, respectively. Data extraction and quality appraisal were done on all eligible studies.

Results.: Ten studies were included: four RCTs, four prospective cohort, and two retrospective cohort studies. Age-related Eye Disease Study (AREDS) showed zinc treatment to significantly reduce the risk of progression to advanced AMD. The risk of visual acuity loss was of similar magnitude, but not statistically significant. Two RCTs reported statistically significant increases in visual acuity in early AMD patients and one RCT showed no effect of zinc treatment on visual acuity in advanced AMD patients. Results from six cohort studies on associations between zinc intake and incidence of AMD were inconsistent.

Conclusions.: Current evidence on zinc intake for the prevention of AMD is inconclusive. Based on the strength of AREDS, we can conclude that zinc treatment may be effective in preventing progression to advanced AMD. Zinc supplementation alone may not be sufficient to produce clinically meaningful changes in visual acuity.

Introduction

AMD is the leading cause of blindness among the elderly in industrialized countries. ^1,2 There is currently no cure for AMD and no treatment that can reverse the damage caused by AMD. The greatest risk factor for AMD is age, with prevalence growing dramatically with increasing age. Nearly 24% of Americans over the age of 80 have intermediate signs of AMD and 12% have advanced AMD, whereas the respective prevalence among people 40 to 49 years old are 2% and 0.1%. ² Smoking has consistently been reported to be a risk factor for AMD. ³ Early AMD is clinically characterized by yellow deposits known as drusen and changes in pigmentation of the retina, which generally results in loss of central vision. Late AMD develops when there is growth of new blood vessels that bleed into the subretinal space (wet, exudative or neovascular AMD) or when the macula atrophies (geographic atrophy or “dry” type). In the later stages there may be distortion of vision or complete loss of visual function, particularly central vision. ⁴  

The pathogenesis of AMD is not completely understood. However, chemical and light-induced oxidative damage to the photoreceptors is thought to be involved. The retina is susceptible to oxidative damage because of high oxygen consumption, high concentrations of readily oxidizable polyunsaturated fatty acids, and constant exposure to visible light. Zinc has been proposed to have a role in AMD prevention because of its structural role in antioxidant enzymes. ^5,6 Zinc is found in high concentrations in regions of the retina that are affected by AMD. ⁷ Zinc in the retina and RPE is also believed to interact with taurine and vitamin A, modify photoreceptor plasma membranes, regulate the light-rhodopsin reaction, and modulate synaptic transmission. ⁸ Retinal zinc content has been shown to decline with age. ⁹ The purpose of this systematic review was to examine the evidence on (1) dietary zinc from food and/or supplements in the primary prevention of AMD and (2) the efficacy of zinc supplementation in the treatment of early and late stage AMD. 

Methods

Literature Search

The present systematic review on zinc and AMD was part of a larger systematic review that focused on the role of specific nutrients in the prevention and treatment of age-related eye diseases. The nutrients of interest included vitamins C and E, omega-3 fatty acids, α-linolenic acid, docosahexaenoic acid, eicosapentaenoic acid, carotenoids, lutein, zeaxanthin, β-carotene, B-vitamins (B1/thiamine, B2/riboflavin, B3/niacin, B6/pyridoxine, folate, and B12/cyanocobalamin), zinc, and vitamin D. Also of interest were fruits and vegetables, which included citrus fruits, dark green leafy vegetables, cruciferous vegetables, fish, and eggs. The eye diseases of interest were AMD, cataract, and glaucoma. An electronic search of Medline (from inception to the second week of February 2012) and Cochrane Central Register of Controlled Trials (from inception to November 2012) was conducted through Ovid (http://gateway.ovid.com). The search strategy combined MeSH or search terms for nutrients of interest (e.g., zinc and antioxidants), and outcomes of interest (i.e., AMD, cataract, and glaucoma) for the larger systematic review. The complete search strategy is listed in Supplementary Material S1. The methods and results pertaining to only zinc and AMD are reported in this systematic review in accordance with the Preferred Reporting Items for Systematic Reviews and Meta Analyses (PRISMA) guidelines. ¹¹  

Study Eligibility Criteria

Randomized controlled trials (RCTs) and prospective cohort, retrospective cohort, case-control, and cross-sectional studies that evaluated the effects of zinc from foods and supplements on AMD in men and women with a mean age of 50 years or older were included. Studies that reported total antioxidant status were excluded since dosage or intake levels of zinc alone were not defined. Also excluded were AMD-related studies that used in vitro, cell culture, and animal models and publications that were not in English. The outcomes of interest were incidence of early, intermediate, or advanced AMD, and also AMD-related drusen and retinal pigment abnormalities (hypo/hyper-pigmentation). For RCTs on early, intermediate and/or advanced AMD the outcomes of interest were (1) best corrected visual acuity (BCVA), (2) progression to advanced AMD, and (3) geographic atrophy (GA). 

All abstracts identified through literature search were screened by three investigators (RV, MC, and EJJ) based on the above mentioned screening criteria with a low threshold to exclude irrelevant abstracts. The first 100 abstracts were screened by all three investigators and the remaining abstracts were double screened. Rejected abstracts were reviewed by all investigators to ensure that all articles that merited inclusion were not left out. Full text articles of the included abstracts were retrieved. Additional eligibility criteria that were of importance in AMD pathogenesis were defined at this time. Studies were excluded if data were not adjusted for age and smoking status (with the exception of RCTs, assuming that randomization minimizes confounding), or if age and smoking status were not matched between groups in case-control studies. Cohort studies were included only if the follow up period was at least 3 and 5 years for subjects aged greater than 65 years and greater than 40 years, respectively. Cross-sectional studies were excluded since they do not clearly distinguish cause and effect from simple association. They are also likely to suffer from recall bias due to the nature of dietary assessment methods. Additionally, RCTs and case-control studies with a sample size of less than 10, RCTs without a concurrent control group, and studies that looked at zinc effects in only carriers or noncarriers of a specific genotype were excluded. If multiple publications were available for the same study, only the one with the longest follow up period was included. Rejected full text articles were examined only once, unless the articles were equivocal for inclusion or exclusion. In that event, the article in question was examined again by a different reviewer and a consensus was reached after discussion with the first reviewer. The main reason for rejection was recorded. 

Data Extraction and Quality Appraisal

A single investigator (RV) extracted data for all studies that qualified the full text screening. All quantitative data were checked and verified by a second investigator (MC). All questionable aspects of the included studies were discussed by the three investigators to ensure accuracy of data extraction. A standardized data extraction form was used to record key data elements, such as study design, study inclusion/exclusion criteria, subject characteristics (number enrolled and analyzed, age, sex, and race), baseline disease status (diabetes or cardiovascular disease), intervention (dose, source, and duration), adverse effects in case of RCTs, primary and secondary endpoints with definition, and study results. Both adjusted and unadjusted (RCTs only) results were extracted. For adjusted data all confounders/covariates were recorded. When results for multiple models were reported, only data for the fully adjusted model with all covariates was extracted. 

Quality appraisals of all included studies are described in the Supplementary Tables S2 and S3. The methodological quality of studies was assessed using the Cochrane Risk of Bias tool for RCTs¹⁰ and the Ottawa-Newcastle scale for observational studies¹² with some additional quality items that are specific to this systematic review. Quality grading was outcome-specific, such that a given study with two outcomes would be graded differently for both the outcomes if the primary outcome was reported well but the analysis on the secondary outcome was incomplete. Studies of different study designs were graded separately within the context of their study design. For each included study, one reviewer rated study quality, which was confirmed by at least one other reviewer. Overall risk of bias assessment (i.e., low, medium, or high risk of bias) was performed based on the ratings of all quality items. Disagreements were resolved by consensus. 

Data Synthesis

Meta-analysis could not be performed due to heterogeneity in zinc supplementations used in the RCTs. For observational studies, the data required to perform dose-response meta-analysis were insufficient. Thus, qualitative syntheses were done separately for RCTs and observational studies to analyze the strength of the evidence on zinc intake for the prevention and treatment of AMD. It was a consensus amongst all investigators that data from AMD subgroups (early versus intermediate versus advanced/late AMD) should be evaluated separately because the effects of zinc may differ between subgroups. Therefore, the results of observational studies were organized by associating risk of early, intermediate, or advanced AMD with categorical exposures (e.g., quintiles of total zinc intake), and were plotted for examining dose-response relationships. The plots allow the reader to appreciate the direction of the estimated effects, even when the choice of the reference category is different across studies. 

For graphical presentation of the results from RCTs, we computed the effect size (mean difference, MD) of the net changes in BCVA scores by comparing zinc treatment group with controls. The graph includes only RCTs that reported sufficient data on both mean net change and standard error (SE) or SD of the net change. A net change is the difference between the change from baseline in the zinc supplementation group and the change from baseline in the control group. Data on the mean net changes in BCVA score and their SEs or SDs were either extracted or estimated from the reported data. The mean net change was estimated by subtracting the mean change in BCVA score from baseline in the zinc supplement group from the mean change in BCVA score from baseline in the control group. The SE or SD of the mean net change was estimated from the SEs or SDs of the changes of both zinc supplement and control groups, assuming these to be independent of each other. The SEs or SDs of the zinc supplementation and control changes were calculated from the standard error of the baseline and final values assuming a correlation of 0.5. Analyses were performed using Stata, version 11 (StataCorp, College Station, TX). 

Results

Search Results

As shown in the PRISMA flowchart 1361 abstracts (1308 through Medline and 53 through Cochrane Central) were identified (Fig. 1). Of the 1103 abstracts that were excluded, three abstracts were common to both the Medline and Cochrane Central searches. Full text articles of the remaining 258 abstracts were screened. Of these, 234 articles were excluded as they focused on nutrients other than zinc and outcomes other than AMD. Full texts of 24 studies that focused on zinc and AMD were screened. Of these, 13 studies were rejected based on the additional eligibility criteria. Of the 11 studies that were included, four were prospective cohort studies, two were retrospective cohort studies, and four were RCTs (one RCT had an additional publication). ^13–22 The study specifics including study design and subject characteristics are summarized in Table 1 and the overall results are summarized in Table 2. 

Figure 1

 

PRISMA flowchart describing the study selection process.

Table 1

 

Description of Prospective Cohort Studies, Retrospective Cohort Studies, and Randomized Controlled Trials That Evaluated Zinc for the Prevention and Treatment of AMD

Author Year (Reference)	Trial or Cohort Name (Enrollment y)	No. of Subjects Enrolled (% Men, % Women)	No. of Subjects Analyzed	Mean Age, y (Range)	Intervention/ Follow-Up Duration	Ascertainment of AMD	Ascertainment of Zinc Intake	Confounders Adjusted for	Risk of Bias
Prospective cohort studies
VandenLangenberg 199813	NFEDS - BDES (1988–1990)	2152 (44.5% men, 55.5% women)*	1709	NR (43–84)*	5 y	Fundus photography graded by WARMGS	Block-National Cancer Institute Health Habits and History Questionnaire	Age, sex, total calorie, pack-y smoked, beer intake (g/wk), history of cardiovascular disease, and history of diabetes	Low
Cho 200114	NHS (1984) HPFS (1986)	NHS, 42,735† HPFS, 29,705† (59% women, 41% men)	NHS, 66,572 HPFS, 37,636	NHS, 56.2 HPFS, 60.4 (50–74)	NHS, 10 y HPFS, 8 y	Self-report followed by completion of a questionnaire by ophthalmologist for signs of AMD	Validated semiquantitative FFQ	2-y time period, age, smoking, blood cholesterol, blood pressure, BMI, physical activity, alcohol intake, total calorie and lutein/zeaxanthin intake, postmenopausal hormone use (women), and profession (men)	Low
VanLeeuwen 200515	Rotterdam Study (1990–1993)	5836 (no AMD = 40.4% male 59.6% female AMD = 42.7% male 57.3% female)*	4170	No AMD = 66.4 AMD = 68.2 (≥55)*	8 y (mean) 0.3–13.9 y (range)	Fundus photography graded by International Classification and Grading system	Validated semiquantitative FFQ	Age, sex, BMI, smoking status, pack-y of smoking, systolic blood pressure, atherosclerosis composite score, serum total cholesterol, and alcohol intake	Low
Tan 200816	BMES (1992–1994)	3654 (43% men, 57% women)	1952‡	64 (≥49)§	10 y	Fundus photography graded by WARMGS	Validated semiquantitative FFQ	Age, sex, smoking, white cell count, family history of AMD, and job prestige‖	Medium
Retrospective cohort studies
Mares-Perlman 199617	NFEDS - BDES (1988–1990)	2152 (45% men, 55% women)*	1968	61 (43–84)*	10 y	Fundus photography graded by WARMGS	Validated Health Habits and History Questionnaire	Age and sex (Smoking, ethanol [from beer], time spent outdoors in the summer, serum cholesterol levels, high density lipoprotein cholesterol levels, and diabetes mellitus did not alter OR).	Low
Morris 200718	NVP NHS (1993–1995)	398 (100% women)	398 (596 eyes)	62 (53–74)	13–15 y prior (1980) during the past 13–15 y (1980–1990)	Fundus photography graded by WARMGS	Validated semiquantitative FFQ	Age, energy intake, smoking, high blood pressure history, sun exposure, BMI, and intakes of fat and wine	Medium
Randomized controlled trials
Newsome 198819	None	174 (63% women, 37% men)*	151	T = 67.6 P = 68.3 (42–89)*	24 mo	Fundus photography graded by two independent blinded observers	Serum zinc concentration before and after intervention	T and P groups were similar at baseline with regard to age, smoking status, cardiovascular disease, hypertension, and visual characteristics	Low
Stur 199620	None (Mar 1990– Jun 1992)	112 (T = 41.1% men and 58.9% women) P = 44.6% men and 55.4% women)	92	T = 72.3 P = 70.6 (NR)	24 mo	Fundus photography and fluorescein angiography graded using WARMGS	Serum zinc concentration before and after intervention	Age, sex, smoking, hypertension, AMD grading, and visual function tests	Medium
AREDS report 8 200121	AREDS (Nov 1992– Jan 1998)	3640 (44% men, 56% women)	3597 (visual acuity) 3609 (AMD progression)	NR (55–80)	6.3 y	Fundus photography followed by AREDS classification	Serum zinc concentration at baseline, y 1 and y 5	Age, race, sex, AMD category, and smoking status	Low
AREDS report 26 200923	AREDS	181 (57% women)	68 (available to study treatment effects)	69.7 (55–80)	6.3 y	Fundus photography followed by AREDS classification	Serum zinc concentration at baseline, y 1 and y 5	Age, sex, smoking, and BMI	Low
Newsome 200822	None	80 (T = 18.9% men and 81.1% women P = 21.6% men and 78.4% women)*	74	T = 72.1 P = 73.3 (NR)*	6 mo	Dilated slit lamp biomicroscopy and fluorescein angiography	Not assessed	T and P groups were similar at baseline with regard to age and sex. No data on smoking status.	Low

 NVP, Nutrition and Vision Project; NFEDS, Nutritional Factors in Eye Disease Study; NR, not reported; BMI, body mass index; T, treatment; P, placebo.

*  Sex and age data is for number of subjects analyzed at baseline.

†  This number represents the women and men included at baseline. The authors report that subjects were added to the analyses as they reached 50 years of age.

‡  ‘n' analyzed is not reported clearly. Methods state 1952 returned for 10-year follow up. Tables state the ‘n' at risk as 2035 for late AMD and 2040 for neovascular AMD.

§  Data is for the 2083 subjects who completed Food Frequency Questionnaire.

‖  List of covariates was taken from the statistical methods section. The zinc data described in the results section does not state if relative risk was adjusted or unadjusted.

Table 2

 

Summary of Results of Prospective Cohort Studies, Retrospective Cohort Studies, and Randomized Controlled Trials That Evaluated Zinc for the Prevention and Treatment of AMD

Author Year (Reference)	Trial/Cohort Name Design)	Definition of Outcomes	n Analyzed	Association or Effect*	Risk of Bias
VandenLangenberg 199813	BDES (PC)	Specific AMD lesions: Large drusen and pigmentary abnormalities Early AMD: Soft indistinct drusen or any type of drusen associated with increased RP or depigmentation	114 cases (1361 at risk large drusen) 63 cases (1405 at risk for pigmentary abnormalities) 103 cases (1279 at risk for early AMD)	Zn intake was not associated with risk of large drusen Zn intake (total) was associated with reduced risk of pigmentary abnormalities Zn intake was not associated with early AMD	Low
Cho 200114	NHS (PC) HPFS (PC)	Any AMD: AMD associated visual acuity loss of 20/30 or worse in at least one eye Early AMD: Drusen and/or RPE changes Late AMD: Presence of geographic atrophy, RPE detachment, CNV membrane, and/or disciform scar	NHS 230 AMD cases (66,572 at risk)† 117 early AMD cases 113 late AMD cases HPFS 154 AMD cases (37,636 at risk)† 78 early AMD cases 75 late AMD cases	Zn intake was not associated with: Risk of any AMD Early AMD Late AMD	Low
vanLeeuwen 200515	Rotterdam study (PC)	Incidence of any AMD (early or late): Early AMD: Large (>63 μm), soft, distinct drusen with pigment irregularities or indistinct or reticular drusen with or without pigment irregularities Late AMD: Central and noncentral geographic atrophy, CNV, or a combination of both	560 AMD cases (4170 at risk)	Zn intake (diet) was associated with reduced risk of any AMD	Low
Tan 200816	BMES (PC)	Early AMD: Large (>125 μm) indistinct soft or reticular drusen, or combined large distinct soft drusen and retinal pigmentary abnormalities without late AMD signs Late AMD: Neovascular AMD or GA as described in the international classification	59 late AMD (2035 at risk) 38 neovascular AMD (2040 at risk)	Zn intake (total) was associated with reduced risk of early AMD and any AMD (early or late AMD)	Medium
Mares-Perlman 199617	BDES (RC)	Early AMD: Soft indistinct or reticular drusen or presence of any drusen type with RPE degeneration or increased RP in the macula Late AMD: RPE detachment or serous detachment of the sensory retina, subretinal or subretinal epithelium pigment hemorrhage and/or subretinal fibrous scars	314 early AMD and 30 late AMD cases (1968 at risk)	Zn intake (diet) was associated with reduced risk of early AMD Zn intake was not associated with risk of late AMD	Low
Morris 200718	NVP NHS (RC)	AMD-related drusen: Large drusen (≥125 μm), soft drusen appearance (i.e., solid, thick appearance, non-uniform density, or fuzzy edges) and drusen extending over an area of ≥125 μm that contained intermediate drusen (63–125 μm) but no large or soft drusen Pigmentary abnormalities: Hyperpigmentation or depigmentation of the RPE	289 eyes (596 eyes at risk for early AMD)	Zn intake was not associated with risk of AMD-related drusen Zn intake (13–15 y prior) was marginally associated with increased risk of pigmentary abnormalities	Medium
Newsome 1988+	None (RCT)	BCVA: Visual acuity measured using ETDRS charts Macular changes: Fundus photographs for changes from baseline in visible pigment, drusen, or atrophy	151	1°: BCVA ↑ 2°: Macular changes ↑ (Zn treated eyes more stable and significantly less visible drusen)	Low
Stur 199620	None (RCT)	BCVA: Measured using Bailey Lovie charts #4 and #5 Contrast sensitivity: Measured at five different spatial frequencies Color discrimination: Measured using Farnsworth-Munsell 100 hue test Retinal grating acuity: Measured using Lotmar Visometer Macular changes: Fundus photographs and fluorescein angiography	92	1°: BCVA ↔ 1°: Contrast sensitivity ↔ 1°: Color discrimination ↔ 1°: Retinal grating acuity ↔ 2°: Macular changes ↔	Medium
Newsome 200822	None (RCT)	BCVA: Visual acuity measured using ETDRS charts Contrast sensitivity: Measured using a Pelli-Robson chart in approximately 85 cd/m2 lighting Photorecovery time: Measured using a device that delivered a reproducible bleaching flash to the macula. Time to recover from the flash induced glare	74	1°: BCVA ↑ 1°: Contrast sensitivity ↑ 1°: Photorecovery time ↑	Low
AREDS report 821	RCT	Visual acuity loss: A decrease in BCVA score from baseline of 15 or more letters in the study eye	3597 (visual acuity) 3609 (progression to advanced AMD)	1°: Risk of loss of visual acuity ↔ 1°: Reduced risk of progression to advanced AMD ↑ (AMD categories 1, 2, 3 and 4)‡	Low
AREDS report 821	RCT	Advanced AMD: Photocoagulation or other treatment for CNV or photographic documentation of any of the following: GA involving the center of the macula, non-drusenoid RPE detachment, serous or hemorrhagic retinal detachment , hemorrhage under the retina or RPE and/or sub-retinal fibrosis	2549 (visual acuity) 2556 (progression to advanced AMD)	1°: Risk of loss of visual acuity ↔ 1°: Reduced risk of progression to advanced AMD ↑ 2°: Reduced risk of development of central GA ↔ (AMD categories 3 and 4)‡	Low
AREDS report 2623	RCT	GA progression: Change is GA area in subjects with either (1) cumulative GA ≥ 0.5 disc areas within 1500 μm of the fovea in 1 or both eye at baseline or (2) no GA or cumulative GA area ≥ 0.5 disc areas at baseline, with the subject developing a GA of 0.5 disc areas or larger within 1500 μm of the fovea in both eyes at follow up visits	68 (48 with unilateral GA 20 with bilateral GA, area ≥ 0.5 disc area)	2°: GA progression ↔ (AMD categories 3 and 4)‡	Low

 PC, prospective cohort; RC, retrospective cohort; RP, retinal pigment.

 ↑ = benefit/stable with statistical significance, ↔ = no statistically significant difference between comparison groups, ↓ = harm with statistical significance.

*  For cohort studies zinc intake represents total zinc intake from foods and supplements. For RCT the effect is of zinc supplementation.

†  Represents women and men who contributed to the analyses as they reached 50 years of age; 42,735 women and 29,705 men were included at baseline.

‡  AREDS category 1: Small <63-μm drusen, area < 125-μm diameter circle (∼5–15 small drusen) with no pigment abnormalities. Second eye same as first.

 Category 2: Small <63-μm drusen or intermediate 63 to 125-μm drusen (area ≥ 125-μm diameter circle) with pigment abnormalities present or absent but geographic atrophy absent or no drusen if pigmentary abnormalities present. Second eye same as first or category 1.

 Category 3a: intermediate 63 to 125 μm or large drusen ≥ 125 μm or neither if noncentral geographic atrophy presents, with or without pigmentary abnormalities, central geographic atrophy absent. Drusen area ≥ 360-μm diameter circle if soft indistinct drusen present or > 656-μm diameter circle if soft indistinct drusen absent. Second eye same as first or category 1 or 2.

 Category 3b: First eye same as category 3a. Second eye visual acuity < 20/32 not due to AMD or uniocular disqualifying disorder is present.

 Category 4a: First eye same as category 1, 2, or 3a. Second eye advanced AMD.

 Category 4b: First eye same as category 1, 2, or 3a. Second eye visual acuity < 20/32 due to AMD, but advanced AMD absent.

Prospective and Retrospective Cohort Studies

Six (four prospective and two retrospective) cohort studies recruited participants between the years 1984 and 1995. The Nurses' Health Study (NHS) cohort was examined both prospectively and retrospectively by two different research groups. ^14,18 The retrospective study examined a subgroup of the NHS cohort who were residents of Boston. ¹⁸ There is a possibility that some subjects from the Boston NHS cohort were also part of the larger prospective cohort. Similarly, the Beaver Dam Eye Study (BDES) cohort was also examined both prospectively and retrospectively. ^13,17 The populations in the remaining two prospective studies were Blue Mountains Eye Study (BMES, Australia) and Rotterdam Study (The Netherlands) cohorts. ^15,16  

Four studies assessed dietary zinc intake using a food frequency questionnaire, ^14–16,18 and two studies used the Health Habits and History questionnaire, ^13,17 both of which are validated dietary assessment tools. Four studies evaluated the incidence of early AMD, ^13,14,16,17 two of which also looked at late AMD, ^14,17 and two also looked at any AMD (early and/or late AMD). ^14,16 One retrospective study only evaluated risk of early AMD related drusen and pigmentary abnormalities, ¹⁸ while another study only evaluated the risk of early and late AMD combined. ¹⁵ Of the six studies, four used the Wisconsin Age-related Maculopathy Grading system (WARMGS) to characterize AMD, ^13,16–18 one used the International Classification and Grading system, ¹⁵ and one ascertained AMD by self-report followed by completion of a questionnaire by the ophthalmologist. ¹⁴  

The follow up period ranged from 5 to 10 years, which was long enough for AMD outcomes to occur. The follow up rate was at least 80% in five out of the six cohort studies. ^{13–15,17,18} Four studies were rated to have a low risk of bias ^13–15,17 and two studies a medium risk of bias. ^16,18 Morris et al. used each eye as a unit of observation instead of each subject, but the authors used appropriate statistical analysis that adjusted for this so the significance of the results is not overestimated; however, this study was still rated to have a medium risk of bias due to unclear reporting of blinding of outcome assessors. ¹⁸ In the Tan et al. study the number of subjects followed up was not consistently reported throughout the manuscript resulting in a medium risk of bias rating. ¹⁶  

Randomized Controlled Trials

Four randomized double-blinded, placebo-controlled trials met our eligibility criteria. ^19–22 Of these the strength of the Age-related Eye Disease Study (AREDS) that looked at antioxidant nutrients with and without zinc for the treatment of AMD supersedes the other three RCTs. AREDS was a large scale, multicenter trial with a sample size of 3640 adults and average follow up of 6.3 years. ²¹ The number of subjects enrolled in all the other three RCTs combined was 317. AREDS had an additional publication that evaluated the effect of zinc treatment on progression of geographic atrophy area in a smaller subset of subjects. ²³ All studies were conducted in the United States (US) with the exception of the Stur et al. ²⁰ study, which was conducted in Austria. 

Zinc supplementation was administered orally in all four RCTs. The types/forms of zinc used differed among studies: two RCTs used zinc sulfate, ^19,20 one used zinc monocysteine, ²² and AREDS used zinc oxide. ²¹ Zinc supplementation was accompanied with cupric oxide in AREDS to prevent potential anemia. The two RCTs that used zinc sulfate administered the highest daily dose of zinc, 200 mg per day, and had an intervention period of 2 years. ^19,20 The daily dose of zinc was 50 mg in the RCT that used zinc monocysteine with an intervention period of 6 months, ²² and 80 mg in the AREDS trial that used zinc oxide with an intervention period of 6.3 years. ²⁰  

Two RCTs enrolled subjects with early AMD, ^19,22 and one enrolled subjects with advanced AMD ²⁰ as defined in Table 2. AREDS enrolled subjects with both early (one or both eyes) and advanced (one eye) AMD. ²¹ The primary outcome of interest in all four RCTs was BCVA, measured using Early Treatment Diabetic Retinopathy Study (ETDRS) charts in three studies ^19,21,22 and using Bailey Lovie charts (Numbers 4 and 5; National Vision Research Institute, Victoria, Australia)in the Stur et al. study. ²⁰ The BCVA score, which was expressed in logMAR units in the Stur et al. ²⁰ study, was converted to ETDRS letters using a formula described by Gregori et al. ²⁴ in order to compare BCVA outcomes in the three RCTs. AREDS defined the risk of loss of visual acuity as a decrease in BCVA score from baseline of 15 or more letters in the study eye, which is recognized as a clinically significant improvement since it represents a doubling of the visual angle. ²¹ AREDS also included progression to advanced AMD as a primary outcome. ²¹ Stur et al. ²⁰ also included contrast sensitivity, measured with a VCTS 6500 Chart (Vistech Consultants, Dayton, OH), color discrimination, measured using Farnsworth-Munsell 100 hue test, and retinal grating acuity, measured using Lotmar Visometer (Haag-Streit, Bern, Switzerland), as primary outcomes; while Newsome et al. ²² included contrast sensitivity, measured using a Pelli Robson Chart and photorecovery time, measured using a device that delivered a reproducible bleaching flash to the macula. Two RCTs included macular changes as a secondary outcome. ^19,20 AREDS reported GA and progression of GA (measured as change in GA area), which are regarded as clinically significant, as secondary outcomes. ^21,23  

Three of the four RCTs were rated low risk of bias. One RCT was rated medium risk of bias due to potential for attrition bias. ²⁰ Compliance assessed by counting the remainder of supplements during visits was greater than or equal to 80% for the Newsome et al. ²² study and the Stur et al. ²⁰ study and greater than or equal to 75% for the AREDS. All four studies had a follow up rate of at least 80% and the loss to follow up was also balanced between the treatment groups. 

Zinc Intake and Any AMD

Three prospective cohort studies reported associations between zinc intake (total = foods + supplements and foods alone) and any AMD, which included incidence of early and/or late AMD combined (Figs. 2, 3). ^14–16 In the Rotterdam study cohort, zinc intake from foods alone (mean intake 9.67 mg/d) was inversely associated with incidence of any AMD. ¹⁵ Majority of the AMD cases in this study were early AMD. Also, in the BMES cohort those in the top decile of total zinc intake (> 15.8 mg/d) were less likely to develop any AMD. ¹⁶ One study on the NHS and HPFS cohorts reported no associations. ¹⁴  

Figure 2

 

Summary of results from cohort studies that examined the relationship between total zinc intake from foods and supplements and the risk of early, late, and any AMD. ^aNumber of AMD cases is a count for number of eyes; ^bAMD outcome is classified as pigmentary abnormalities; ^cAMD outcome is classified as AMD-related drusen; ^dzinc intake is measured in mg/1000 kcal; and ^emedian zinc intakes for pooled data were calculated using median intakes provided for men and women weighted by number of men and women for each AMD outcome. ROB, Risk of Bias; L, Low; M, Medium.

Figure 3

 

Summary of results from cohort studies that examined the relationship between zinc from foods alone and the risk of early, late, and any AMD. ^aNumber of AMD cases is a count for number of eyes; ^bAMD outcome is classified as pigmentary abnormalities; ^cAMD outcome is classified as AMD-related drusen; ^dzinc intake is measured in mg/1000 kcal; and ^emedian zinc intakes for pooled data were calculated using median intakes provided for men and women weighted by number of men and women for each AMD outcome. ROB, Risk of Bias; L, Low; M, Medium.

AREDS showed that, among subjects with early, intermediate, and advanced AMD combined (AREDS categories 2, 3, and 4), zinc treatment did not significantly reduce the risk of visual acuity loss compared with placebo (odds ratio [OR] = 0.82, 99% confidence interval [CI] 0.63–1.08, P = 0.07); however, a significant reduction in the risk of visual acuity loss (OR = 0.77, 99% CI 0.58–1.03, P = 0.02) was observed when zinc was administered with antioxidant supplement. ²¹ In this group of AREDS subjects zinc treatment was found to reduce the risk of progression (or further progression) to advanced AMD (OR = 0.71, 99% CI 0.51–0.98, P = 0.005). ²¹  

Zinc Intake and Early AMD

Four cohort (three prospective and one retrospective) studies investigated the association between zinc intake (total = foods plus supplements and foods alone) and early AMD (Figs. 2, 3). ^13,14,16,17 Of these, two studies showed no association between dietary and total zinc intake and incidence of early AMD. ^13,14 Amongst these, in one study total zinc intake in the past (15 years prior to AMD onset) was inversely related to pigmentary abnormalities, but not large drusen (data not shown). ¹³ In BMES cohort, those in the top decile of total zinc intake (>15.8 mg/d) were less likely to develop early AMD. Zinc from foods alone did not have the same effect. ¹⁶ In the BDES population, retrospective data showed that high intake of zinc from foods alone (5.3–16.4 mg/d) was associated with lower odds of early AMD. ¹⁷ In this study total zinc intake in the highest versus lowest quintile was related to lower odds for increased retinal pigment. ¹⁷ In the NHS-Boston cohort that investigated only early AMD-related outcomes, total zinc intake was found to be associated with greater incidence of pigmentary abnormalities in the year that represented diet prior to AMD onset. ¹⁸ Zinc intake was not associated with AMD-related drusen in this cohort. 

Two RCTs on early AMD subjects reported a statistically significant increase in BCVA scores with zinc treatment compared with placebo (Fig. 4). ^19,22 The eyes of subjects with early AMD either remained stable or had decreased accumulation of visible drusen and also less atrophy. ¹⁹ Subjects with early AMD also showed improvement in contrast sensitivity, and photostress recovery times within 6 months of treatment with 50 mg/d of zinc (Table 2). ²² In AREDS subjects without GA at baseline, zinc treatment had no effect on the development of GA, (OR = 1.13, 99% CI 0.74–1.74), not necessarily involving the center of the macula. ²¹ AREDS did not evaluate AMD progression among subjects with early AMD due to occurrence of very few advanced AMD events. 

Figure 4

 

Summary of results of randomized controlled trials that evaluated the effect of zinc supplementation on changes in BCVA scores (as number of ETDRS letters). *BCVA scores were measured in logMAR units using Bailey Lovie charts no. 4 and no. 5 (National Vision Research Institute, Australia) in Stur et al. study. ²⁰ The scores were converted to ETDRS letter scores using this formula: ETDRS letters = 85 − 50 × logMAR. ²⁴ ROB, Risk of Bias; L, Low; M, Medium.

Zinc Intake and Late AMD

Two cohort studies (one prospective and one retrospective) investigated the association between zinc intake (total = foods plus supplements and foods alone) and incidence of late AMD (Figs. 2, 3). ^14,17 No significant associations were observed in both the cohorts. However, in the retrospective analysis of the BDES cohort the OR was found to increase with increasing quintile for zinc intake. ¹⁷  

In one RCT on advanced AMD subjects no effect of zinc supplementation on BCVA score was observed. ²⁰ Also, no effects were observed on secondary outcomes, which included contrast sensitivity, color discrimination, retinal grating acuity, and morphological changes in the macula in this study. AREDS showed that among subjects with intermediate and advanced AMD (AREDS categories 3 and 4) zinc treatment significantly lowered the risk of progression to advanced AMD (OR = 0.70, 99% CI 0.50–0.97, P = 0.005), but did not significantly lower risk of visual acuity loss (OR = 0.82, 99% CI 0.61–1.09, P = 0.07). Also, zinc treatment did not significantly lower the risk of development of GA (OR = 0.76, 99% CI 0.46–1.27, P = 0.17, unadjusted analysis) in the central macula, although it was in the direction of a beneficial effect. ²¹ Also, in this group of AREDS subjects zinc treatment was suggestive of a beneficial effect on development of neovascular AMD (OR = 0.73, 99% CI 0.51–1.04, P = 0.02, unadjusted analysis). ²¹ No effect was observed on change in digital GA area with zinc in subjects with unilateral and bilateral GA area greater than or equal to 0.5 disc area. ²³  

Adverse Events in RCTs

Adverse events that were reported may or may not be associated with zinc supplementation. In two RCTs subjects in both the placebo and zinc treatment group had gastrointestinal side effects. ^20,21 In the Stur et al. study, 14 patients (nine in the treatment group and five in the placebo group) developed choroidal neovascularization (CNV) in the study eye. ²⁰ In the zinc treatment arm of the Newsome et al. study, one subject complained of aggravation of preexisting peptic ulcer and one subject had gastrointestinal distress, which was relieved when zinc was taken with food. ¹⁹ In AREDS, excess self-reported anemia was observed but serum hematocrit was normal. ²¹ Also, genitourinary hospitalizations (e.g., unspecified urinary tract infections and prostatic hyperplasia in men and stress incontinence in women) were more frequently reported in the zinc arm of AREDS. 

Discussion

To our knowledge, this is the first systematic review that focuses on the independent effects of zinc for the prevention and treatment of AMD. Previous systematic reviews have evaluated the effect of zinc in combination with other antioxidant nutrients. ^25,26 Based on the six cohort studies no conclusions can be made on the associations between dietary zinc intake (from foods and supplements) and the incidence of early, late, and any AMD since the results were inconsistent. The strongest evidence for zinc in the treatment of AMD comes from AREDS, a large scale nationwide multicenter clinical trial on AMD. AREDS showed that zinc supplementation alone reduced the risk of progression to advanced AMD in those with intermediate AMD and those with advanced AMD in one eye. Treatment with zinc alone did not reduce the risk of loss of visual acuity in AREDS subjects. Based on observation alone, data from the other three RCTs suggest that zinc supplementation may improve visual acuity in early AMD patients, but not in patients with advanced AMD. 

Even though the two RCTs reported statistically significant improvements in visual acuity, the actual mean change in BCVA score was only three to five letters, which is not considered to be a clinically significant change. In AREDS visual acuity loss was defined as a decrease in BCVA score from baseline of 15 or more letters in the study eye (equivalent to doubling or more of the initial visual angle, e.g., 20/20 to 20/40 or worse). ²¹ The probable reason for the marginally significant effect on the risk of loss of visual acuity in AREDS could be the more stringent and clinically meaningful criteria for assessment of visual acuity. A study in retinitis pigmentosa patients showed the overall interobserver and intraobserver variability in visual acuity measures to be 1.3 to 2.3 letters, which is very close to the three to five letter change observed in the two RCTs. ²⁷ However, it is noteworthy that in the Newsome et al. ²² study, three subjects in the zinc treated group had a gain of greater than or equal to 10 letters (two lines), also considered clinically meaningful, while visual acuity declined by greater than or equal to 15 letters in the placebo group. ²⁸ Although the net improvements in visual acuity scores were smaller than what is considered clinically significant, these small changes may still be of importance, because it is unknown whether longer intervention duration in these trials may have yielded more clinically meaningful improvements in visual acuity. The risk of visual acuity loss was significantly lowered in AREDS subject who received zinc in combination with antioxidants. ²¹  

Another primary outcome that AREDS evaluated was progression to advanced AMD, which cannot be evaluated in short term clinical trials as AMD progresses slowly. Zinc supplementation in AREDS categories 3 and 4 subjects probably increased zinc content of the RPE, thus, preventing progression to advanced AMD. Additionally, only those AREDS subjects who were thought to be at risk of vision loss from AMD based on presence of drusen or RPE changes were part of the factorial design that evaluated zinc (with or without antioxidants). ²⁹ AREDS subjects who received zinc supplementation were at more advanced stages of AMD compared with the other two RCTs on early AMD. It appears that zinc supplementation is beneficial in preventing AMD progression in those subjects who are at a high risk of advanced AMD. Zinc, in combination with antioxidants, had more significant effects on lowering risk of progression and the accompanying visual acuity loss. 

One possible explanation for the differences in visual acuity results between the two RCTs on early AMD subjects could be the form of zinc supplementation, zinc monocysteine versus zinc sulfate. ^19,22 Cysteine, like zinc, has relatively small number of dietary sources and is important for antioxidant activity in the eye. Cysteine can be metabolized to antioxidant glutathione in the RPE, which may have contributed to the observed positive effects on visual acuity. ³⁰ Zinc supplementation also resulted in a statistically significant improvement in contrast sensitivity and photostress recovery times in these studies. ^19,22 Zinc-treated eyes also had significantly less drusen. However, the results from these secondary outcomes only provide circumstantial evidence and are not accepted by the Food and Drug Administration as endpoints for clinical studies. 

The reported beneficial associations in cohort studies were between zinc intakes and early/any AMD. In one study association was with total zinc and in two studies with zinc from foods alone. ^15–17 The RDA for zinc is 11 mg/d and 8 mg/d for males and females, respectively. ³¹ The observed effects were at levels of greater than or equal to15.8 mg/d of total zinc and 9.67 mg/d (mean) and 5.3 to 16.4 mg/d (range) of zinc from foods alone, which were all close to the RDA for zinc. Only Mares-Perlman et al. reported a range of zinc intake, 16.4 mg/d being the upper limit of the highest quintile group. ¹⁵ All other studies reported only mean or median intake levels, which were all lower than 50 mg/d. ^13–16,18 The maximum intake levels of zinc in these studies is unclear. It is not surprising that no association was reported between zinc intake and late AMD, given that the dietary intake levels of zinc were much lower than the doses of zinc used in RCTs. Also, it may be too late in the disease stage for a dietary intervention to have an effect. It is also important to keep in mind that study populations differ between cohort studies and RCTs. 

There are several limitations in our systematic review. First, we only searched for and included published studies because the data are peer reviewed. There is an underlying suspicion of publication bias against studies having either null or negative outcomes. ³² Second, our systematic review is limited by potential language bias as only English publications were included. There are also limited studies on the effects of zinc alone on AMD. Zinc has been more widely studied in combination with other antioxidant nutrients. Future research should focus on evaluating different forms of zinc supplementation (zinc oxide versus zinc sulfate versus zinc monocysteine) to determine the form that is most bioavailable. A highly bioavailable form of zinc may be more efficacious at biological doses. One study found that zinc gluconate had significantly higher bioavailability compared with inorganic zinc oxide in humans. ³³ Preventing the need to supplement at pharmacologic levels could reduce incidence of gastrointestinal discomfort and anemia, which are likely to occur with high levels of zinc supplementation. The ongoing AREDS 2 trial is evaluating zinc at doses lower than in the original AREDS formulation. ³⁴ Additionally, the differences in study populations, dietary intakes and possible interactions with other nutrients limit the evaluation of nutrients in isolation. 

In conclusion, zinc treatment can be effective in preventing progression to advanced AMD, which is critical in preventing complete vision loss in patients with early, intermediate, and advanced AMD. Zinc in combination with antioxidants, rather than zinc alone, can be used to achieve clinically and statistically significant improvements in visual acuity in AMD patients. 

Evidence thus far on zinc intake for the prevention of AMD is insufficient to draw any overall conclusions. 

Supplementary Materials

pdf S1 

pdf S2 

pdf S3 

Acknowledgments

The authors thank Samantha Berger, BS, and Eric Mboggo, MD, MS, for their administrative support. Mei Chung contributed her efforts to this study without receiving funding or salary support. 

Support by grants from the United States Department of Agriculture 1959051000-073-025 and Bausch & Lomb, Inc. 

Disclosure: R. Vishwanathan, None; M. Chung, None; E.J. Johnson, Bausch & Lomb, Inc. (F)