Age-Related Eye Disease Study 2 (AREDS2) – an Analysis

What follows is an analysis of the results of the AREDS2 randomized clinical trial “Lutein Zeaxanthin and Omega-3 Fatty Acids for Age-Related Macular Degeneration” (AREDS2) recently published on the JAMA website.

AREDS 2 has been long-awaited. AREDS1 demonstrated decreased progression of AMD with the administration of high dose antioxidants and minerals. There were problems with AREDS1 though; beta-carotene was used whereas lutein and zeaxanthin were not. High dose zinc appeared to increase hospitalizations for GI and GU disorders. The omega-3s DHA and EPA were omitted. In response to these shortcomings, AREDS 2 was established. Now the results are in. Unfortunately the trial raises more questions than it answers.

Here are the issues.

  • Although it is called a “placebo controlled” trial, there is no placebo arm. Everyone in the placebo group was offered the AREDS1 formula as a base. Nearly everyone took it. This confounds results.
  • The use of prior supplementation with omega-3s and carotenoids was not controlled for. As baseline blood levels for lutein, zeaxanthin, DHA, and EPA were higher than normal, the participants likely had a high use of prior supplementation. This confounds results.
  • The omega-3s chosen were 350 mg DHA and 650 mg EPA, yet DHA is known to be a necessary structural element in the macula whereas EPA is thought to be the precursor for signaling molecules that might be important for the eye. DHA is a known quantity; EPA has hypothetical value.
  • The study was enormously complex: Two tiers of randomization occurred in this 2×2 factorial trial. Patient groups included:  (1). Placebo – actually on AREDS1 formula. (2). AREDS1 + EPA/DHA. (3). AREDS1 + lutein/zeaxanthin. (4). AREDS1 + EPA/DHA/lutein/zeaxanthin. (5). AREDS1 + 25 mg Zn (not 80 mg). (6). AREDS1 – beta carotene. (7). AREDS1 with 25 mg Zn and no beta-carotene. Such a complex study design created a great deal of difficulty in understanding the findings. In the authors’ own words, “complicated design involving a secondary randomization which may have affected our ability to evaluate the role of adding lutein+zeaxanthin and DHA+EPA to the AREDS formula”. This unfortunately is the crux of the trial. Introducing doubt in the primary objective of the study is “study suicide” in my view.
  • 30% of participants opted not to undergo the secondary randomization and among those who did decide to move forward, women and highly educated participants did so at a significantly higher rate. This confounds the results.
  • 13% of participants stopped their supplements during the trial. As this was an Intention to Treat study, they remained in the trial even though they were “off drug”.
  • 14% of participants admitted to taking lutein/zeaxanthin/DHA/EPA during the trial even though they were not supposed to do so. This confounded the results.
  • Only 84% of participants took > 75% of their pills.
  • A small portion of patients had blood tests of studied supplements. There was a 200% increase in lutein/zeaxanthin and a 105% increase in EPA, but because of the low dose of DHA there was only a 35% increase in its level.
  • Beta-carotene lowered the effect of lutein, proving the negative interaction between these two carotenoids. This was unexpected by the authors. They stated, “In this analysis we assumed there would be little interaction between the various nutrients used”. To me, this statement is disturbing; this is a nutrient trial. Interactions between nutrients are inevitable and must be considered in such a trial design. This is one of the fatal flaws of the trial.
  • Beta-carotene increased lung cancer. This was previously understood but reconfirmed. In fact, 91% of lung cancer cases occurred in patients taking the standard AREDS formula. Though the patients were former smokers, they were not current smokers. In my view this negates the use of the standard AREDS formula in anyone with a history of smoking.
  • Low dose zinc and the absence of beta-carotene did not diminish efficacy of the formula. This demonstrates that low dose Zn and no beta carotene are preferred.
  • Lutein/Zeaxanthin did demonstrate less AMD progression compared to the AREDS1 formula.

In sum, AREDS2 was a wonderful concept that was poorly implemented. The take-home message from the trial: Do not use beta-carotene; use lutein and zeaxanthin; low dose zinc is optimal; efficacy of omega-s remains unanswered. Interestingly, the May, 2013 issue of JAMA Ophthalmology included the LUTEGA trial in which DHA and lutein/zeaxanthin were found to be beneficial in AMD. LUTEGA used DHA in a dose 3x that of EPA, the relationship that probably should have been used in AREDS2. One final note, meso-zeaxanthin was not studied in either trial. Based upon our current understanding of pathophysiology and biology meso-zeaxanthin should be included in an AMD formulation.

Learn more about the highest quality vitamins, minerals, and omega-3’s – created by a leading preventive cardiologist at

Tags: , , , , , , , , ,

One Comment

  1. Zac Denning June 10, 2013 at 6:56 pm #

    Thank you for the excellent discussion of the issues with the AREDS 2 study design – and for posting the new recommendations. Following is a link to a good overview of the findings, as well as the new National Eye Institute recommendations:

    The Bar was Set Very High for AREDS 2 Findings
    It should be noted that because of this design, any intervention nutrient had to perform at least 25% better then the antioxidant + zinc formula in the original AREDS study, in order to achieve statistical significance and be considered a valid finding. This was such a huge threshold that it was not surprising that there were no primary study findings. Since primary findings are often the only findings listed in the study abstract – this fooled many in the media into thinking that there were not important findings at all.

    Was the Primary AREDS 2 Research Question the Right One?
    I think it’s also important to point out a few stats from the secondary findings. As was described in the presentation of the findings by Emily Chew, adding lutein / zeaxanthin to the original AREDS formula (with beta-carotene) – part of the primary study design – likely wasn’t the right research question. As you point out, beta-carotene interfered too much with lutein & zeaxanthin uptake. So it’s the secondary study endpoints (aka ‘research questions’) that were essentially the most tellling. These secondary questions were probably the right questions to ask.

    Some Important Findings from AREDS 2 (secondary endpoints):

    When AREDS 2 researchers looked at lutein / zeaxanthin takers, taking an AREDS formula WITHOUT beta-carotene, they saw an 18% decreased risk of AMD progression, compared to those taking just the original AREDS formula WITH beta-carotene.

    Those with lutein / zeaxanthin deficient diets experienced a 26% decreased risk of AMD progression when they supplemented with lutein / zeaxanthin. It’s likely that a large proportion of the American population would fall into this lutein / zeaxanthin deficient category, unlike the unusually well-nourished AREDS 2 study population.

    Further benefits were seen for lutein & zeaxanthin in reducing cataract progression.

Post Comment