Carotenoid Status and Psychological Impact of Presymptomatic Macular Degeneration Genetic Risk Assessment: The MAGENTA Randomized Trial

Abstract

Purpose

Genetic testing for age-related macular degeneration (AMD) risk reliably offers insight into an individual’s susceptibility for future visual loss; however, an American Academy of Ophthalmology expert panel in 2012 discouraged routine AMD genetic risk testing because there was no evidence that such knowledge would alter an individual’s clinical course. To address this knowledge gap, we investigated whether disclosure of AMD genetic risk to presymptomatic individuals would encourage healthier lifestyle adoption that could reduce the incidence of AMD later in life.

Design

The Moran AMD Genetic Testing Assessment (MAGENTA) trial is a single-site, prospective, controlled clinical study (NCT05265624) that randomized 80 presymptomatic subjects in a 3:1 ratio to immediate or 1-year deferred disclosure groups. We stratified participants into high-, intermediate-, and low-risk groups by Mendelian randomization.

Subjects

We enrolled Whites aged 18 to 64 years with no clinical signs of AMD.

Methods

As a biomarker of healthy nutritional status, participants’ skin, plasma, and macular carotenoid concentrations were measured using resonance Raman and reflectance spectroscopies, high-performance liquid chromatography, and autofluorescence imaging, respectively. Nutritional and emotional status were assessed with validated surveys.

Main Outcome Measures

We looked for changes in skin, plasma, and macular carotenoids as biomarkers of healthier lifestyle adoption and for the impact of AMD genetic risk disclosure on participants’ psychological well-being.

Results

Of the 80 participants, 94% had a family history of AMD, and the AMD genetic risk distribution was 36% high, 24% intermediate, and 40% low. We found no statistically significant difference in skin, plasma, and macular carotenoids between study groups at 12 months relative to baseline (P > 0.05 for all comparisons). Participants’ interest and compliance were high, as shown by the 95% subject study completion rate, and there was no evidence of worsening stress following AMD genetic risk disclosure to the participants.

Conclusions

While AMD genetic risk disclosure did not significantly impact nutritional biomarker levels over 12 months, there were no adverse psychological effects, and the subjects generally felt knowledge of their AMD risk was valuable. Our findings could guide future presymptomatic AMD genetic testing trials with extended biomarker assessments in larger and more diverse populations.

Financial Disclosure(s)

Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.

The clinical management of age-related macular degeneration (AMD) imposes a significant social and financial impact on the U.S. population.¹ As public awareness of AMD grows and the fear of blindness becomes increasingly widespread, there is considerable value in developing and implementing strategies aimed at preventing or delaying AMD onset. Over the past 2 decades, researchers have shown that genetic variants in complement factor H (CFH), age-related maculopathy susceptibility protein 2/high-temperature requirement factor A1 (ARMS2/HTRA1), and other loci are key risk factors for AMD.2, 3, 4, 5, 6 Concurrently, epidemiological and clinical studies have demonstrated that dietary modification or nutritional supplementation with antioxidant vitamins, minerals, and macular carotenoids, as outlined in the Age-related Eye Disease Study 2 trial, can slow the progression of this blinding condition.7, 8, 9, 10, 11, 12, 13, 14, 15 These findings suggest that lifestyle modifications in presymptomatic individuals informed by early identification of genetic risk could serve as an effective strategy to mitigate the likelihood of future vision loss from AMD. Although large-scale, long-term clinical studies to prove this hypothesis have not been conducted, multiple epidemiological studies support the value of what is generally considered a healthy lifestyle.

Positive lifestyle changes could include achieving a healthy weight, exercising, smoking cessation, and ingesting carotenoid-rich diets (i.e., dark green leafy vegetables or orange/yellow fruits and vegetables).^6,7,16 All of these lifestyle modifications increase the body’s lutein, zeaxanthin, and meso-zeaxanthin levels, xanthophyll carotenoids collectively known as the macular pigment (MP).¹⁷ Macular pigment carotenoids are not synthesized de novo in humans and must originate from the diet. They are uniquely localized in the fovea, the center of the macula responsible for distinct spatial vision, and are important in enhancing visual performance and maintaining retinal health via their blue light filtering, antioxidant, and anti-inflammatory properties.18, 19, 20, 21, 22, 23

Presymptomatic genetic testing for AMD offers insights into individual risk profiles but may cause anxiety, depression, and other types of distress, partly because there is still no cure for the disease. Hence, an American Academy of Ophthalmology (AAO) expert panel in 2012 discouraged routine AMD genetic testing, not because genetic testing is unreliable, but because there is no evidence from adequately powered prospective clinical trials to ascertain the benefit of such genetic testing and whether such knowledge will inform positive lifestyle changes.^24,25 Despite the AAO’s position on routine AMD genetic testing, many individuals, particularly those with AMD-affected relatives, are eager to know their risk and still patronize commercially available AMD genetic testing services. This suggests that presymptomatic individuals are worried about the future and are willing to make informed lifestyle decisions that could delay AMD onset.

Therefore, we conducted this trial to investigate whether knowledge of AMD genetic risk will inform the adoption of a healthier lifestyle that could reduce the incidence of AMD later in life. Additionally, we explored the psychological impact of AMD genetic risk disclosure on individuals’ well-being. We hypothesized that individuals informed of a high risk of AMD are more likely to make sustained positive lifestyle changes relative to individuals with low risk or in the deferred disclosure group. We investigated this hypothesis through the Moran AMD Genetic Testing Assessment (MAGENTA) clinical trial.

Methods

We have previously published a detailed report on the design and procedures for the MAGENTA trial.²⁶ The MAGENTA trial is a single-site, prospective, double-masked, randomized clinical trial conducted at the John A. Moran Eye Center of the University of Utah, Salt Lake City, Utah, United States. The first participant’s enrollment was in August 2022, and the last participant completed their final study visit in July 2024. The MAGENTA trial is registered with ClinicalTrials.gov at https://clinicaltrials.gov/ct2/show/NCT05265624.

To comply with the International Conference on Harmonization Guidance E6 for the conduct of human research, a trained study staff member ensured that all participants provided written informed consent prior to study participation. The University of Utah’s Institutional Review Board approved the MAGENTA study. The MAGENTA trial complied with the tenets of the Declaration of Helsinki and the International Conference on Harmonization Harmonized Tripartite Guidance for Good Clinical Practice (ICH-GCP E6 [R1]) and fully adhered to the code of ethics regarding participant enrollment, study assessment, and data protection. A Data Safety and Monitoring Board, consisting of an ophthalmologist and a social worker not involved in the study, reviewed the study’s progress and adverse events every 6 months. We did not change the study methods after study initiation.

In this trial, we included participants of White descent aged 18 to 64 years, with a preference for those with a positive family history of AMD in order to enhance the percentage of high-risk genotypes in the study population. Participants were excluded if they had a personal history of AMD, prior AMD genetic testing, non-White ethnicity, employment at the Moran Eye Center or other eye care practices, or planned cataract surgery during the study period. Additional exclusions included major psychiatric disorders and eye diseases affecting MP assessment, such as Stargardt disease, macular telangiectasia type 2, and albinism. Qualified ophthalmologists with retinal subspecialties (P.S.B. and M.E.H.) determined participants’ eligibility following a thorough review of fundus images.

Once study eligibility was confirmed, a study staff member ensured that the participant’s baseline assessments, including systemic (plasma and skin) and ocular carotenoid levels, nutritional and behavioral surveys, and anthropometric measurements, were carried out. A month after the baseline study visit, when participants’ AMD genetic testing results were available, the subjects were then randomly assigned to the immediate or deferred disclosure groups in a 3:1 ratio using a computer-generated random sequence. The subjects were further categorized into high-, medium-, and low-risk groups by Mendelian randomization using results from a genetic testing battery done in Dr Hageman’s laboratory.²⁶ The deferred disclosure group did not know their AMD genetic risk until their final study visit (12 months from study enrollment).

All participants, clinicians, and staff members involved in the trial were blinded to the group assignment, except for the trained genetic counselor who performed the randomization and advised participants on proven ways to decrease risk of AMD and to improve their carotenoid status. Thus, all participants received their AMD genetic risk disclosure during an individual, in-person counseling session conducted by the trained genetic counselor at the John A. Moran Eye Center. The counselor, who was not involved in study assessments, followed a standardized disclosure procedure to ensure consistent delivery of information across participants. During each session, the counselor used 2 visual handouts developed by the Moran Eye Center: (1) an AMD fact sheet summarizing modifiable lifestyle risk factors and evidence-based prevention strategies, and (2) a genetic testing information sheet explaining the meaning of genetic variants associated with AMD, their implications for risk, and contact information for follow-up questions. Participants received printed copies of both materials, which the counselor reviewed verbally, answered questions about, and confirmed comprehension before concluding the session. This structured disclosure process provided a uniform educational experience across participants and facilitated understanding of both genetic and behavioral risk information.

Repeated assessments of participants’ carotenoid status in the plasma, skin, and eye; nutritional and behavioral surveys; and anthropometric measures were performed at subsequent study visits. Participants’ skin carotenoid status, dietary, and behavioral assessments were assessed at baseline and months 1, 3, 6, 9, and 12. In contrast, MP, plasma carotenoid assessment, and height and weight measurements for body mass index (BMI) were obtained at baseline, month 6, and month 12. We regularly contacted participants via phone calls, text messages, and emails to ensure adherence to the study regimen.²⁶

Study Outcomes and Comparisons

The predeclared primary outcome for the MAGENTA trial was the change in skin carotenoid levels measured by resonance Raman spectroscopy (RRS) from baseline to month 12. We compared skin carotenoid status in high AMD risk in the immediate disclosure group to the deferred disclosure group. Also, we compared the immediate disclosure of all AMD risk groups relative to deferred disclosure. Our secondary outcome was plasma and ocular carotenoid status changes, compared by AMD genetic risk disclosure groups. Additionally, we evaluated the impact of AMD genetic risk disclosure on psychological health via standardized surveys. An exploratory analysis, which will be addressed in a future publication, examined whether AMD-associated genetic risk factors affect systemic and ocular carotenoid status.

Demographic, Dietary, and Psychological Surveys

Participants provided demographic and lifestyle data, including contact information (i.e., name, age, phone number, and residential and email addresses), educational level, occupation, ethnicity, health history, smoking habits (history and frequency), alcohol intake (average consumption per week and frequency), and height and weight for BMI calculation. We used the lutein and zeaxanthin quantitative food frequency questionnaire to determine participants’ dietary intake of carotenoid-rich foods. The lutein and zeaxanthin quantitative food frequency questionnaire is validated and captures about 90% of lutein and zeaxanthin consumed in the United States using the National Health and Nutrition Examination Survey data.²⁷ Participants also completed the Hospital Anxiety and Depression Scale (HADS) and the Impact of Event Scale (IES).^28,29 The HADS assessed anxiety and depression in nonpsychiatric individuals through 2 subscales, HADS-A for anxiety and HADS-D for depression, each consisting of 7 questions scored from 0 to 3. Scores range from 0 to 21, with higher scores indicating distress and scores between 11 and 21 suggesting the need for intervention.³⁰ The IES measured distress resulting from AMD genetic risk disclosure using 15 self-reported items that evaluated intrusive thoughts and avoidance behaviors. The IES score was from 0 to 5, with a total score of 26 or more indicating moderate to severe stress.²⁹ Both HADS and IES have been validated for clinical assessment.³¹

AMD Risk Genetic Testing

We determined participants’ AMD risk using genotypes at 2 key loci, CFH-CFHR5 and ARMS2/HTRA1, which account for most genetic AMD risk.³² Haplotypes (rs800292, rs1061170, rs12144939) determined AMD risk at the CFH-CFHR5 locus, whereas at the ARMS2/HTRA1 locus, the number of risk alleles at rs10490924 assessed risk. Diplotype combinations at these loci were ranked by increasing AMD odds ratios and divided into low-, intermediate-, and high-risk groups. In the general population, the low-risk group represents 50% of participants, with another 25% in the intermediate and 22% in the high-risk categories, and 3% consisting of haplotypes too rare to establish AMD risk accurately.

Skin Carotenoid Measurement

We used RRS and pressure-mediated reflectance spectroscopy (RS) to measure skin carotenoids.³³ While RRS provides more specific assessments of carotenoid status than RS, it is bulkier and more expensive. In contrast, RS is portable, affordable, and easier to operate. Resonance Raman spectroscopy and RS are validated noninvasive devices that serve as biomarkers for fruit and vegetable consumption and correlate significantly with plasma carotenoids.33, 34, 35, 36, 37 Resonance Raman spectroscopy uses a harmless blue laser light (488 nm) to excite carotenoids in the participant’s palm. The device then collects back-scattered light, filters out Rayleigh-scattered light, and a cooled spectrograph analyzes the resulting fluorescence and Raman-shifted light. The peak intensity at about 1525 cm^–1 represents C=C vibrations.³⁸ For RS measurement, the index finger is illuminated with broadband white light (350–850 nm), and the diffusely reflected light is analyzed in real-time. Participants press their fingers against a convex lens, temporarily squeezing out blood to reduce the impact of interfering chromophores on reflection spectra. A connected laptop regulates light exposure and data acquisition and displays the skin carotenoid score, ranging from 0 to 800.^33,39 Resonance Raman spectroscopy and RS take 3 readings each, and the average is used for statistical analysis. The devices are calibrated daily before assessments.

Plasma Carotenoid Assessment

We analyzed blood samples for their carotenoid status using high-performance liquid chromatography.⁴⁰ The participants’ blood samples were collected in 6 mL Vacutainer K₂EDTA tubes made by Becton, Dickinson and Company were allowed to sit at room temperature for 30 minutes, centrifuged, and plasma separated and stored at –80°C until analysis. We added 200 μL of ethanol with 0.1% butylated hydroxytoluene to 200 μL of plasma to extract carotenoids, followed by ethyl acetate and hexane extractions. We dried all the organic extracts using nitrogen gas. Afterward, the dried extracts were washed, resuspended in the high-performance liquid chromatography mobile phase (methanol: methyl tert-butyl ether [80:20, v/v]), and analyzed using an Agilent 1260 series high-performance liquid chromatography with diode array detection at 450 nm. We quantified carotenoid concentrations using standard curves obtained from authentic carotenoid standards, and their co-elution confirmed the identity of carotenoid peaks.

Macular Pigment Measurement

Before MP imaging, we used a drop each of 1% tropicamide and 2.5% phenylephrine, the standard of care for dilation in eye care practices, to achieve optimal dilation. Briefly, the dual-wavelength autofluorescence method on the Heidelberg Multicolor Spectralis (Heidelberg Engineering GmbH) measures the attenuation of lipofuscin autofluorescence by the MP.⁴¹ Autofluorescence images were collected by sequentially raster-scanning the macula using 488 nm (blue) and 514 nm (green) lasers. Macular pigment images were then created by digitally subtracting the green image from the blue one, using correction factors to account for the absorption spectrum of the macular carotenoid pigment. The images were analyzed with Heidelberg’s MP analysis software, setting the zero point at 9° eccentricity to maintain consistency with previous studies. For statistical analysis, MP measurements were recorded at 0.5°, 2°, and 9° eccentricities. This method is proven to be highly reliable, especially when measuring MP optical volume at 9° eccentricity (MPOV 9°).^42,43 One eye per patient was selected for statistical analysis. This was typically the subject’s nondominant eye to minimize visual discomfort postdilation.

Sample Size

We used our pilot study with 16 subjects below 60 years of age to determine our sample size. We randomized participants in a 3:1 ratio to the immediate and deferred disclosure groups and found a pattern that any disclosure of any degree of risk for AMD resulted in increased skin carotenoid levels compared to the deferred disclosure. Hence, with 80 participants (accounting for an attrition rate of 25%) and a 3:1 allocation ratio to immediate and deferred disclosure groups, there was 80% power to detect a 40.5% difference in our primary outcome.

Statistical Analysis

We fit a linear mixed-effects model to analyze the changes in the outcome of interest (i.e., skin, serum, and macular carotenoid status) over time, adjusting for baseline and accounting for interactions between time and the disclosure group/AMD risk combination. We included a random intercept for each participant to account for within-subject correlation. We computed estimated marginal means using the “emmeans” R package to facilitate the interpretation of the time-by-disclosure interaction via graphical visualization, showing estimates for all follow-up periods. We compared high-risk in the immediate disclosure group to deferred disclosure, any immediate disclosure to deferred disclosure, and within-group comparisons of immediate disclosure. All statistical analyses were executed using R and the “lme4” package for mixed-effects modeling. A two-sided alpha level of 0.05 was deemed statistically significant. Results were reported with estimated mean difference between groups in change from baseline at month 12, 95% confidence intervals (CIs), and associated P values for hypothesis testing.

To assess the level of agreement between skin carotenoid RRS and RS at baseline, a two-way mixed-effects model with absolute agreement was used to calculate the intraclass correlation coefficients (ICCs). This model accounted for random variability across subjects while treating the RRS and RS as fixed effects.

Results

Enrollment and Baseline Characteristics

Participant enrollment for the MAGENTA trial started in August 2022 and was completed in July 2023. Figure 1 shows the Consolidated Standards of Reporting Trials flow diagram for the MAGENTA trial. Of the 84 subjects who consented to this trial, four failed screening due to the presence of early AMD. Additionally, 1 participant withdrew from the study, while 3 were lost to follow-up due to personal reasons, such as time constraints, travel challenges, and family obligations. The demographics and baseline characteristics of the study participants at baseline stratified by disclosure and risk groups are presented in Tables 1 and 2, respectively.

Figure 1.

Moran AMD Genetic Assessment (MAGENTA) study Consolidated Standards of Reporting Trials flow diagram. AMD = age-related macular degeneration.

Table 1.

Summary of Participants’ Baseline Characteristics Stratified by Disclosure Groups

Variable	Immediate (N = 60)	Deferred (N = 20)	P Value
Age (yrs)	47.7 (10.6)	51.3 (10.3)	0.16∗
BMI (kg/m2)	29.7 (6.6)	28.6 (5.5)	0.69∗
Gender
Female	49 (81.7%)	14 (70.0%)	0.34†
Male	11 (18.3%)	6 (30.0%)	-
AMD family history
Negative	4 (6.7%)	1 (5.0%)	1.00†
Positive	56 (93.3%)	19 (95.0%)	-
Relative AMD risk
High	22 (36.7%)	7 (35.0%)	1.00†
Intermediate	14 (23.3%)	5 (25.0%)	-
Low	24 (40.0%)	8 (40.0%)	-
Alcohol intake
No	27 (45.0%)	9 (45.0%)	1.00†
Yes	33 (55.0%)	11 (55.0%)	-
Smoking habits
Never smoked	53 (88.3%)	18 (90.0%)	1.00†
Previously smoked	7 (11.7%)	2 (10.0%)	-
Current smokers	0 (0.0%)	0 (0.0%)	-
AREDS2 supplement use
No	56 (93.3%)	19 (95.0%)	1.00†
Yes	4 (6.7%)	1 (5.0%)	-
Dietary L+Z (mg/d)	4.9 (7.6)	4.8 (4.7)	0.82∗
RRS (RU)	27 865 (9089.6)	27 185 (7470.3)	0.85∗
RS (score)	269 (113.5)	261 (101.1)	0.92∗
Plasma carotenoids (ng/mL)
L+Z	264.9 (190.1)	227.4 (81.2)	0.97∗
Total carotenoids	623.9 (357.2)	541.8 (247.3)	0.48∗
MPOV	10 252 (4861.3)	9945 (4286.7)	0.79∗

AMD = age-related macular degeneration; AREDS2 = age-related eye disease study 2; BMI = body mass index; MPOV = macular pigment optical volume; RRS = resonance Raman spectroscopy; RS = reflectance spectroscopy; RU = Raman units.

^∗Exact Wilcoxon rank sum test.

^†Fisher exact test. Data are represented as mean (standard deviation) for continuous variables and number (percentage) for categorical variables.

Table 2.

Summary of Participants’ Baseline Characteristics Stratified by Risk Groups

Variable	Total (N = 80)	High (N = 29)	Intermediate (N = 19)	Low (N = 32)	P Value
Age (yrs)	48.6 (10.6)	49.3 (9.6)	49.2 (9.6)	47.7 (12.1)	0.92∗
BMI (kg/m2)	29.4 (6.3)	30.1 (6.5)	27.8 (7.0)	29.8 (5.8)	0.34∗
Gender
Female	63 (79%)	25 (86.2%)	14 (73.7%)	24 (75.0%)	0.48†
Male	17 (21%)	4 (13.8%)	5 (26.3%)	8 (25.0%)	-
AMD family history
Negative	5 (6.2%)	1 (3.4%)	0 (0.0%)	4 (12.5%)	0.28†
Positive	75 (94%)	28 (96.6%)	19 (100.0%)	28 (87.5%)	-
Disclosure group
Deferred	20 (25%)	7 (24.1%)	5 (26.3%)	8 (25.0%)	1.00†
Immediate	60 (75%)	22 (75.9%)	14 (73.7%)	24 (75.0%)	-
Alcohol intake
No	36 (45.0%)	16 (55.2%)	7 (36.8%)	13 (40.6%)	0.39†
Yes	44 (55.0%)	13 (44.8%)	12 (63.2%)	19 (59.4)	-
Smoking habits
Never smoked	71 (88.7%)	29 (100.0%)	14 (73.7%)	28 (87.5%)	1.00†
Previously smoked	9 (11.3%)	0 (0.0%)	5 (26.3%)	4 (12.5%)
Current smoker	0 (0.0%)	0 (0.0%)	0 (0.0%)	0 (0.0%)
AREDS2 supplement use
No	75 (93.8%)	27 (93.1%)	18 (94.7%)	30 (93.8%)	1.00†
Yes	5 (6.2%)	2 (6.9%)	1 (5.3%)	2 (6.2%)
Dietary L+Z (mg/d)	4.9 (6.9)	6.5 (10.6)	5.3 (4.7)	3.3 (2.8)	0.47∗
RRS (RU)	27 695 (8672)	29 225 (9622)	28 977 (8630)	25 547 (7538)	0.21∗
RS (score)	267 (110)	285 (122)	269 (89)	250 (109)	0.41∗
Plasma carotenoids (ng/mL)
L+Z	255 (169)	257 (218)	275 (157)	241 (125)	0.63∗
Total Carotenoids	603 (333)	622 (390)	654 (318)	555 (287)	0.55∗
MPOV	10 175 (4699)	10 478 (4237)	10 489 (4833)	9714 (5112)	0.67∗

AMD = age-related macular degeneration; AREDS2 = age-related eye disease study 2; BMI = body mass index; MPOV = macular pigment optical volume; RRS = resonance Raman spectroscopy; RS = reflectance spectroscopy; RU = Raman units.

^∗Kruskal-Wallis test.

^†Fisher exact test. Data are represented as mean (standard deviation) for continuous variables and number (percentage) for categorical variables.

Compliance and Adverse Events

The 76 participants who completed the final study visit had a high compliance rate, with only 1 subject in the immediate disclosure group missing the skin carotenoid assessments for month 3. Also, 3 subjects (2 in the immediate disclosure group and 1 in the deferred disclosure group) missed their month 1 RS skin carotenoid assessment due to an instrument malfunction. No significant adverse events were reported by the participants during the study.

Skin Carotenoid Status

We found no statistically significant differences in change in skin carotenoids between groups assessed by RRS at month 12 (and all other study time points) between high-risk subjects in the immediate disclosure and deferred disclosure groups (P = 0.79). Similarly, there was no significant change in RRS among participants with any immediate disclosure relative to deferred disclosure (P = 0.39; Fig 2) or within-group comparisons of participants in the immediate disclosure group (P > 0.05 for all comparisons). The findings are presented in Table 3.

Figure 2.

Skin carotenoid assessment with RRS was statistically insignificant between any immediate disclosure and deferred disclosure over the study duration. RRS = resonance Raman spectroscopy.

Table 3.

Linear Mixed Effects Model of Differences at Month 12 in Carotenoid Biomarker Outcomes among Study Groups

Group Comparison	Mean Difference (95% CI)	P Value
RRS (RU)
High-risk early disclosure with deferred disclosure	–398 (–3365 to 2569)	0.79
Any early disclosure with deferred disclosure	–1100 (–3621 to 1421)	0.39
Within the early disclosure
High-risk with low-risk	87 (–1361 to 1536)	0.90
High-risk with intermediate-risk	965 (–690 to 2622)	0.25
Intermediate-risk with low-risk	–878 (–2541 to 784)	0.30
RS (score)
High-risk early disclosure with deferred disclosure	–47 (–82 to –13)	0.007∗
Any early disclosure with deferred disclosure	–34 (–8 to 50)	0.022∗
Within the early disclosure
High-risk with low-risk	–8 (–25 to 8)	0.32
High-risk with intermediate-risk	–11 (–31 to 7)	0.22
Intermediate-risk with low-risk	3 (–15 to 22)	0.72
Plasma lutein + zeaxanthin (ng/mL)
High-risk early disclosure with deferred disclosure	–66 (–151 to 17)	0.12
Any early disclosure with deferred disclosure	–42 (–89 to 52)	0.23
Within the early disclosure
High-risk with low-risk	–30 (–70 to 8)	0.12
High-risk with intermediate-risk	–4 (–50 to 41)	0.83
Intermediate-risk with low-risk	–26 (–71 to 19)	0.25
Plasma total carotenoids (ng/mL)
High-risk early disclosure with deferred disclosure	–215 (–448 to 18.3)	0.07
Any early disclosure with deferred disclosure	–185 (–382 to 11)	0.06
Within the early disclosure
High-risk with low-risk	–30 (–141 to 79)	0.58
High-risk with intermediate-risk	–14 (–141 to 113)	0.83
Intermediate-risk with low-risk	–16 (–141 to 108)	0.79
MPOV
High-risk early disclosure with deferred disclosure	–577 (–1370 to 216)	0.15
Any early disclosure with deferred disclosure	108 (–619 to 733)	0.75
Within the early disclosure
High-risk with low-risk	–526 (–913 to –140)	0.008∗
High-risk with intermediate-risk	–501 (–945 to –57)	0.03∗
Intermediate-risk with low-risk	–25 (–472 to 422)	0.91
BMI (kg/m2)
High-risk early disclosure with deferred disclosure	0.75 (–0.057 to 1.56)	0.07
Any early disclosure with deferred disclosure	–0.40 (–0.78 to 0.59)	0.24
Within the early disclosure
High-risk with low-risk	0.24 (–0.15 to 0.62)	0.23
High-risk with intermediate-risk	0.28 (–0.17 to 0.73)	0.22
Intermediate-risk with low-risk	–0.04 (–0.50 to 0.41)	0.85
Dietary lutein + zeaxanthin intake (mg/d)
High-risk early disclosure with deferred disclosure	–1.309 (–3.85 to 1.23)	0.31
Any early disclosure with deferred disclosure	–1.178 (–3.29 to 0.935)	0.27
Within the early disclosure
High-risk with low-risk	–0.319 (–1.555 to 0.916)	0.61
High-risk with intermediate-risk	0.122 (–1.297 to 1.541)	0.86
Intermediate-risk with low-risk	–0.442 (–1.825 to 0.941)	0.53

BMI = body mass index; MPOV = macular pigment optical volume; RRS = resonance Raman spectroscopy; RS = reflectance spectroscopy.

^∗Statistical significance at P < 0.05.

Regarding skin carotenoids assessed with RS, the high-risk immediate disclosure group had significantly decreased scores relative to the deferred disclosure group over time (P = 0.007). However, we observed no significant differences between any immediate and deferred disclosure groups (P = 0.16; Fig 3). Likewise, comparing high, intermediate, and low risk within the immediate disclosure group showed no statistically significant changes over time (P > 0.05 for all comparisons).

Figure 3.

Skin carotenoid assessment with RS was not statistically significant between any immediate disclosure and deferred disclosure over the study duration. RS = reflectance spectroscopy.

Plasma Carotenoid Status

Table 3 provides the mean differences and significance level for plasma lutein + zeaxanthin (L+Z) and plasma total carotenoids between the study groups at month 12. We observed no statistically significant differences in mean plasma L+Z and plasma total carotenoids between the high-risk subgroup in the immediate disclosure relative to the deferred disclosure groups (P > 0.05). Also, comparisons between any immediate disclosure and deferred disclosure groups or among subgroups within the immediate disclosure group for serum L+Z and serum total carotenoids were not statistically significant (P > 0.05 for all comparisons).

Macular Pigment Status

Macular pigment findings for this study are shown in Table 3. Generally, high-risk participants in the immediate disclosure group had lower MPOV levels relative to low-risk or deferred disclosure groups over the study period, but MPOV did not differ significantly between the high-risk immediate disclosure and deferred disclosure groups (P = 0.15) or when comparing any immediate disclosure with deferred disclosure (P = 0.87) at month 12. However, within the immediate disclosure group, MPOV significantly decreased in the high-risk group compared to both the low-risk (P = 0.008) and intermediate-risk groups (P = 0.03), while no significant difference was observed between the intermediate- and low-risk groups at month 12 (P = 0.910).

BMI and Dietary L+Z Intake

Study participants’ BMI did not change significantly across study groups throughout the study duration. Participants in the high-risk immediate disclosure group had a nonsignificant increase in BMI relative to individuals in the deferred disclosure group over time (P = 0.07). Similarly, no significant differences were observed when comparing any early disclosure with deferred disclosure (P = 0.78) or within the early disclosure groups (P > 0.20 for all comparisons).

Dietary L+Z intake did not show statistically significant variations across study groups over the study period. The high-risk participants in the immediate disclosure group had an insignificant reduction in dietary L+Z intake compared to those in the deferred disclosure group (P = 0.31). Likewise, we found no significant differences when comparing any early disclosure with deferred disclosure (P = 0.27) or within the early disclosure groups (P > 0.50 for all comparisons). Thus, our results suggest that AMD genetic risk disclosure did not lead to meaningful changes in BMI or dietary carotenoid intake over the study period.

Psychological Impact

Table 4 presents the number of participants who experienced psychological impact at any study time point, stratified by AMD genetic risk groups and disclosure timing (deferred vs immediate). While psychological impact was observed across all AMD risk groups, we found no evidence of exacerbation of stress associated with either immediate or deferred AMD genetic risk disclosure, as there were no reports of study-associated anxiety or depression and no requests for counseling. This finding indicates that presymptomatic AMD genetic risk disclosure was psychologically safe and well-tolerated.

Table 4.

Counts of Participants Who Had Any Psychological Impact at Any Study Time Point Stratified by AMD Risk Groups

Risk Group	Deferred (N = 20)	Immediate (N = 60)
Anxiety
High	1 (5.0)	1 (1.6)
Intermediate	-	2 (3.3)
Low	1 (5.0)	3 (5.0)
Combined	2 (10.0)	6 (10.0)
Depression
High	1 (5.0)	-
Intermediate	-	1 (1.6)
Low	1 (5.0)	1 (1.6)
Combined	2 (10.0)	2 (3.3)
Impact of events
High	1 (5.0)	3 (5.0)
Intermediate	1 (5.0)	1 (1.6)
Low	-	5 (8.3)
Combined	2 (10.0)	9 (15.0)

Data presented as count (percentage, %).

Among participants in the immediate disclosure group (N = 60), anxiety was reported in 1 high-risk, 2 intermediate-risk, and 3 low-risk individuals, with a total of 6 (10%) participants experiencing anxiety. In the deferred disclosure group (N = 20), anxiety was observed in 1 high-risk and 1 low-risk participant, with a total of 2 individuals (10%) having anxiety.

Depressive symptoms were less prevalent across groups. In the immediate disclosure arm, 2 individuals (3.3%; 1 intermediate-risk and 1 low-risk group) experienced depression, whereas, in the deferred disclosure group, depression was observed in 2 participants (10%; 1 high-risk and 1 low-risk group).

An impact of event score of 26 or more, which is an indicator of psychological distress related to a significant experience, was reported by 9 participants (15%) in the immediate disclosure group, while in the deferred disclosure group, 2 participants (10%) reported distress.

Willingness to Pay for Genetic Testing

At the month 1 randomization visit, we explored participants’ willingness to pay out of pocket (up to $100) for AMD genetic testing if it became available in the future. As shown in Table 5, the majority of participants (87.5%) expressed willingness to pay, with 61.25% indicating they would pay $80 to $100. Smaller proportions were willing to pay $60 to $80 (12.5%) or $40 to $60 (13.75%), while only 12.5% reported they would not pay at all. These findings suggest strong perceived value for genetic testing among study participants.

Table 5.

Participants' Willingness to Pay for Genetic Testing Stratified by Disclosure Groups

Amount for Genetic Testing ($)	Total (N = 80)	Immediate (N = 60)	Deferred (N = 20)
0	10 (12.50)	8 (10.00)	2 (2.50)
40–60	11 (13.75)	5 (6.25)	6 (7.50)
60–80	10 (12.50)	7 (8.75)	3 (3.75)
80–100	49 (61.25)	40 (50.00)	9 (11.25)

Data presented as count (percentage, %).

Agreement between RRS and RS in Measuring Skin Carotenoids

We found a single-measure ICC of 0.84 (95% CI: 0.757–0.895), indicating strong agreement between individual skin carotenoid measurements obtained from RRS and RS. Furthermore, there was excellent reliability when measurements from both methods were averaged within a time period (average-measure ICC = 0.91; 95% CI: 0.862–0.944). These findings support the interchangeability and consistency of the 2 methods for assessing skin carotenoid levels in the studied population.

Discussion

The MAGENTA is the first prospective, randomized, controlled clinical trial to determine whether presymptomatic knowledge of AMD risk through genetic testing would stimulate a meaningful, healthier lifestyle change that could mitigate the onset of AMD in the future.²⁶ We conducted this study to provide empirical evidence for or against AAO’s 2012 stance regarding AMD genetic testing, which states: “avoid genetic testing for complex disorders like age-related macular degeneration…until specific treatment or surveillance strategies have been shown in 1 or more published clinical trials to have benefit….”^24,25 We observed no statistically significant differences in our predeclared primary and secondary study endpoints. However, we found that any disclosure of AMD genetic risk to presymptomatic participants was psychologically safe and well-embraced, as there were no reports of exacerbation of anxiety or depression. Also, there was high study compliance, and the majority (87.5%) of participants expressed willingness to pay out of pocket for affordable commercial AMD genetic testing.

Skin, plasma, and macular carotenoid concentrations are recognized as key biomarkers of fruit and vegetable intake, reflecting a healthy lifestyle.^36,37,44,45 However, our study found no statistically significant changes in skin, plasma, and macular carotenoid status over a year following AMD genetic risk disclosure, regardless of whether the information was provided immediately or deferred. These findings may, in part, reflect the limited sample size, which may have reduced our ability to detect subtle but meaningful changes, and also indicate the need for broader biomarker assessments to understand the impact of AMD genetic risk disclosure on lifestyle changes. There is substantial evidence to demonstrate that enhanced carotenoid status via supplementation for an extended duration significantly improves health and disease outcomes. For instance, the Age-related Eye Disease Study 2, which followed participants for over 10 years, found that individuals who received 10 mg of lutein and 2 mg of zeaxanthin lowered their chances of AMD progression by ∼20%.11, 12, 13 Similarly, a recent study from Seddon et al also indicates that diets rich in carotenoids significantly reduce the risk of AMD onset and progression.¹⁵ Consistent findings have been reported in several epidemiological studies.46, 47, 48, 49, 50, 51 Most of these studies provided supplements to subjects and were for a longer period; hence, the rise in carotenoid levels in the plasma, skin, and retina tissues. However, in this study, we only advised participants on lifestyle practices known to mitigate AMD onset/progression and to increase carotenoid levels without prescribing or providing any supplements, which might account for our study findings. Americans, on average, ingest 1 to 2 mg of L and Z per day, which is considerably lower than amounts provided in most commercial supplements.⁵² This could explain the wide variation observed in this study. It may be worthwhile for future studies to consider a longer study duration and to augment participants’ dietary intake with supplementation or intensive lifestyle coaching.

One finding worth highlighting is the fact that knowledge of AMD genetic risk did not exacerbate any distress among study participants. While immediate disclosure of AMD risk may lead to transient psychological effects, the overall impact remained relatively low across risk groups. With increased media reports of the rising incidence of AMD associated with the progressive aging of the American population, disclosing such information may induce anxiety and depression. However, our findings suggest that informing presymptomatic individuals of their AMD genetic risk is safe despite the potentially frightening nature of the disease.^24,25,53 Furthermore, most participants were appreciative and willing to commit at least $80 to $100 to know their genetic risk for AMD should it become more widely available in the future. This clearly shows that participants are keen to understand and know their risks, even though our results did not have enough evidence to prove that participants proactively adopted a healthier lifestyle.

The preferred practice patterns for AMD management usually emphasize lifestyle modifications such as eating a healthy diet, smoking cessation, BMI reduction, supplementation with antioxidants, vitamins, minerals, and physical exercise.^6,54, 55, 56 However, patient education and adherence to these recommendations remain suboptimal. Nearly 60% of patients either fail to take recommended oral supplements or use them incorrectly, and long-term goals like smoking cessation or dietary changes are rarely achieved.57, 58, 59 Thus, encouraging individuals to take actionable steps and change ingrained habits remains a complex issue. Although we envisaged that informing individuals of their high risk of eventual AMD would encourage positive behavioral and lifestyle changes, we observed insignificant changes in systemic and ocular biomarkers indicative of improved healthy lifestyles. This highlights the difficulty of sustaining long-term habit changes and underscores the need for extended studies to better understand the impact of AMD genetic risk disclosure on behavior.

Over the past decade, genetic testing has become significantly more affordable, accessible, and analytically valid due to advances in high-throughput genotyping and next-generation sequencing, particularly for AMD-associated variants such as CFH and ARMS2/HTRA1.³ The rise of both clinical and direct-to-consumer platforms has not only improved access to genetic information but also fostered a more genetically informed and receptive public.⁶⁰ In the MAGENTA trial, participants welcomed AMD genetic risk disclosure, which was safe, well-tolerated, and appreciated, reflecting broad public acceptance. However, the trial also demonstrated that genetic risk knowledge alone is often insufficient to induce sustained lifestyle changes, especially for late-onset diseases like AMD, where preventive actions such as dietary improvement or smoking cessation lack immediate outcomes and require ongoing reinforcement.^61,62 These findings underscore the limitations of demanding long-term behavioral proof as a benchmark for supporting genetic testing. Given current technological capabilities, increased public readiness, and supportive safety data, we urge the AAO to reconsider its restrictive stance and adopt a more progressive, patient-centered approach to routine AMD genetic testing.

Another key finding is the demonstration of a statistically significant strong agreement and reliability between RRS and RS in measuring skin carotenoid levels, as indicated by a high single-measure intraclass correlation coefficient (ICC = 0.84; 95% CI: 0.757–0.895) and even higher average-measure ICC (0.91; 95% CI: 0.862–0.944). These results suggest that both techniques produce consistent and comparable results in assessing skin carotenoid status within a free-living population.⁶³ The strong ICC values reported here are in line with prior research demonstrating the validity of noninvasive skin carotenoid measurements and extend previous work by confirming that RRS and RS can be used interchangeably for population-based monitoring.³³ Given the growing interest in skin carotenoid status as a biomarker for fruit and vegetable intake and systemic antioxidant capacity, these results support the broader application of both RRS and RS technologies in research and clinical applications.⁶³ Furthermore, this interchangeability opens up opportunities for resource-constrained settings to adopt RS, which is more feasible considering its portability, without compromising data quality, as was recently demonstrated in our Nepal vitamin A deficiency study.⁶⁴ Despite the strong agreement, slight methodological differences between the devices could still influence readings. Hence, future studies should evaluate the agreement across diverse demographic groups to validate the generalizability of these findings.

The strength of the MAGENTA trial lies in the fact that it is the first prospective, randomized, controlled trial to study whether disclosure of high genetic risk of progression to advanced AMD will promote sustained, quantifiable, positive lifestyle changes in clinically normal individuals.²⁶ This clinical trial provides evidence for the possible reconsideration of the AAO expert panel recommendation on routine AMD genetic testing. Additionally, it is worth highlighting that assessing skin carotenoids and MP as biomarkers of healthy lifestyle choices in response to AMD genetic disclosure has never been studied systematically in the past. Despite the success in recruiting 80 participants in a single-site study and the high study compliance, our study population was White because AMD genetic testing has been validated only for this specific population. The lack of ethnic diversity may limit the generalizability of our findings to more diverse populations. We hope to include diverse populations in future studies when their AMD genetic variants are established and validated. Due to logistical constraints, we could not undertake expanded measures of lifestyle modification and coaching, potential active nutritional interventions, or follow-up with individuals in the deferred disclosure group to determine any psychological impact. We plan to incorporate these features in a future, much larger, longer-term, multicenter phase III trial. Another limitation of our study is that participants were already consuming approximately twice the average American dietary intake of L+Z per day at baseline. This elevated baseline intake of L+Z may have limited our ability to detect meaningful changes in carotenoid status over time. Additionally, nearly all participants (94%) reported a family history of AMD, indicating prior awareness of their hereditary risk. This pre-existing knowledge may have reduced the behavioral or emotional response to AMD genetic risk disclosure, suggesting a possible underestimation compared with an unselected population without known familial risk. Future studies may benefit from incorporating lifestyle coaching and supplementation to assess the full impact of AMD genetic risk disclosure on modifiable lifestyle behaviors and could target nutritionally compromised populations.

In conclusion, study participants demonstrated a strong interest in learning their AMD genetic risk, and disclosing such information was safe and well-tolerated. Although our study endpoints did not reach statistical significance, as there was insufficient evidence that individuals proactively changed their lifestyle, our findings can guide future trials using extended biomarker assessments to evaluate the value of presymptomatic AMD genetic testing in larger and more diverse populations.

Manuscript no. XOPS-D-25-00654.