AREDS Eye Supplements

Helpful or Harmful?

Patients with intermediate dry AMD have historically been treated with a fixed formulation of high-dose eye supplements and zinc (AREDS) as standard of care since the pivotal placebo controlled AREDS [8] study was published by the National Eye Institute (NEI) in 2001. The study confirmed that AREDS reduced progression to Wet AMD. This formulation was modified slightly following a second AREDS2 study published in 2013 by the NEI [10].
AREDS has been shown to reduce AMD progression risk for vision loss by 25% for patients with intermediate Dry AMD. However, it was not known who would benefit and who would not. Researchers [5,9] analyzed outcomes of these AREDS compounds on 989 AREDS study patients who donated their DNA and were followed for 7 years. The AREDS study is the only placebo-controlled study with patient DNA that is suitable for this analysis. The data demonstrated that even though most patients benefitted from AREDS, for 13% of patients with a specific genetic profile (2 high-risk CFH alleles and 0 ARMS2 risk alleles) the standard AREDS formula was detrimental and accelerated vision loss much faster than placebo.

In 2014 Dr. Chew et al [12] at the NEI published their own analysis of the AREDS study DNA. Their conclusions were contrary; they concluded that ‘AREDS supplements reduced the rate of AMD progression in all genotypes’. Their data however as graphed below does not support that conclusion. When compared to placebo, the AREDS formula was beneficial in 69% of patients. This beneficial effect is better than expected. The expected benefit was  a 25% reduction in risk for progression to advanced AMD with vision loss.
For patients with genetic profiles C1A1, C0A1, C0A0 and C1A2, the efficacy of AREDS was much better than a 25% reduction in risk for vision loss (Wet AMD). As you can see, the hazard ratios above compared to placebo showed much fewer patients progressing on AREDS. In fact these four genetic profiles offer a 60% to 80% reduction in risk.
For 31% of patients, however, in the C2A0 and the C1A0 genetic groups, the AREDS treatment was less effective than placebo. More patients progressed to vision loss on AREDS. The C2A0 genotype had the worst outcomes with a 78% increased progression to advanced AMD with vision loss vs. placebo. Some argue whether that result is statistically significant, others suggest it doesn’t matter, there is no benefit for AREDS in these patients and the trend is towards harm.

In 2016 an additional independent study also evaluated the role of genetic variants in modifying the relationship between supplementation and progression to advanced AMD among 4124 eyes (2317 subjects with a genetic specimen), This study published in the British Journal of Ophthalmology demonstrated a statistically significant relationship between genotype and the efficacy of the AREDS formulation[11]. Some patients benefitted, while others did not.

Approximately two out of every three patients had a reduced risk for vision loss on AREDS that ranged from a 35% to 85% improved outcome vs. placebo. For example, the ‘Low/Low’ group did the best with a risk of only 0.15, which is a statistically significant benefit vs. placebo (1.0). This is a much better result than expected from AREDS. The average AREDS benefit is a 25% reduced risk for AMD progression but this is a blended result for all patients — some patients who benefit and some patients who don’t.

The use of AREDS to prevent the progression of AMD in all patients is not proven for all patient genotypes. It seems to help most patients, but it is not effective and may trend toward harm in others. We recommend caution for the indiscriminate use of AREDS. Testing is available through eye care professionals across the USA and Canada.


Is the design of these genetic studies retrospective?

The AREDS study was a prospective placebo controlled study. The decision to collect and analyze DNA was in the original AREDS study protocol and the DNA was collected prospectively. AREDS DNA has repeatedly been analyzed prospectively by Drs Awh and Kim, by Dr. Chew et al and now by Dr. Seddon et al [9,11, 12]. One can’t analyze the DNA until the clinical outcomes are observed, but the analysis of the DNA is done prospectively. The FDA accepts these study designs as prospective genetic analyses; they are not retrospective genetic analyses.

These gene studies focus on one cohort from the AREDS study. Shouldn’t there be another study in a different patient cohort?

It is valid scientific and regulatory practice to repeat claims made on the basis of a single study cohort. The FDA, for example, does not accept a medical claim without data supporting the claim from at least two independent cohorts. There is one significant exception to that principle. There is never a requirement to repeat a study claiming toxicity. Phase 1 drug safety studies are not repeated if the compound is not safe. Calls to repeat an AREDS study asking patients with the suspect genotype to enroll under informed consent might not meet any regulatory or ethical standard now.

Should companies promoting AREDS provide a warning to doctors and patients?

There are FDA defined regulatory processes for making medical claims on a compound intended for the treatment of a disease. People promoting nutritional products are not permitted to make medical claims (ie treatment of a disease). In this instance, however it is the NEI and the AAO that are promoting AREDS for the treatment of AMD. There is no FDA approval for their claim. The NEI/NIH have been sharing the ownership of the AREDS/AREDS2 commercial patent and they have been sharing the revenues from the sale of AREDS/AREDS2 products with the manufacturers. Are they not both AREDS/AREDS2 commercial partners promoting the compounds for the treatment of intermediate AMD? FDA inquiry might provide an appropriate answer to this question including the need for appropriate warnings and approved medical claims.

Dr. Chew et al claim that their study did not demonstrate a statistically significant interaction between the genes and the AREDS supplements. Should doctors and patients continue to use the AREDS/AREDS2 products regardless of genetic profile?

Dr. Chew’s analysis in AREDS #38 reported a 78% increase in progression to advanced AMD with vision loss for patients with the suspect genotype[12]. Her statistical model suggested that this was not statistically significant. Other biostatisticians claim it is significant and there is a deleterious effect. 78% more patients with this genetic profile lost vision while on AREDS vs. placebo. There does not seem to be any data published, including AREDS report #38, that shows the treatment is safe in patients with 2 high-risk CFH alleles and 0 high-risk ARMS2 alleles. It all shows negative outcomes. The frequency of this genetic profile is between 13% and 19% of patients. Until proven safe and effective, the body of literature suggests that the compound is detrimental to these patients. The concept of ‘Do No Harm’ might be considered.