Take-Home Message: An analysis of independent data published in AREDS Report 38 shows that the response of patients with moderate to advanced age-related macular degeneration (AMD) to treatment with zinc or the Age-Related Eye Disease Study (AREDS) formulation varied substantially based on genotype.
Nashville, TN—An analysis of independent data published in the Age-Related Eye Disease Study (AREDS) Report 38 shows the response of patients with moderate to advanced age-related macular degeneration (AMD) to treatment with the AREDS formulation varied substantially based on the genotype groups of AREDS patients.
Physicians should be aware of the different responses to treatment with the AREDS formulation or zinc because of the potential for harm for patients with specific genotypes, noted Carl C. Awh, MD, a private-practice retina specialist in Nashville, TN.
Dr. reviewed his published analyses suggesting that the response to AREDS formulation is influenced by genetic risk alleles and presented a new analysis of data from the recently published AREDS Report 38. Although AREDS Report 38 concluded that there was no significant interaction between genotype and treatment with the AREDS formulation, he showed that the data in AREDS Report 38 is actually consistent with his findings, with zinc promoting advanced AMD in some groups.
A genetic analysis of 989 AREDS patients, published by Dr. Awh in November 2013, found that the risk ratios for AMD progression associated with complement factor H (CFH) and age-related maculopathy susceptibility 2 (ARMS2) risk alleles differed greatly based on the assigned AREDS treatment group.
“That the AMD progression risk associated with CFH and ARMS2 genetic risk alleles was significantly different for patients in different AREDS treatment group was the major finding of this analysis,” he said. “We predicted an adverse response to the AREDS formulation or zinc in patients with two CFH and no ARMS2 risk alleles.”
Dr. Awh explained that the biologic features of CFH and ARMS2 genetic risk factors are consistent with an interaction with zinc, which binds to and inactivates CFH, causing an increase in complement activation. Drusen contain complement cascade proteins and zinc, both of which are considered signs of local inflammation in the pathogenesis of AMD.
The second risk allele, ARMS2, localizes in the mitochondria, potentially affecting the interaction of antioxidants and free radicals, Dr. Awh said.
ARMS2 risk alleles may also disrupt superoxide dismutase 2 causing increased oxidative damage in retinal pigment epithelial cells.
CFH and ARMS2 risk
In a recent second study, Dr. Awh and colleagues performed an outcomes analysis to evaluate the effect of combined CFH and ARMS2 risk on the treatment response in AREDS patients. The investigators combined the nine possible genotype combinations of CFH and ARMS2 risk allele number into four larger genotype groups based on relatively higher or lower frequency of the risk alleles to create groups large enough for a statistically significant outcomes analysis.
Dr. Awh presented the 7-year outcomes of progression to advanced AMD in the four genotype groups, highlighting those with high ARMS2/low CFH risk and those with high CFH/no ARMS2 risk.
“The former did extremely well if treated with the AREDS formulation or zinc,” he said. “However, the latter group had the opposite response.
“Treatment with the AREDS formulation in that group was associated with a 40% progression rate to advanced AMD over 7 years, compared with 17% in those receiving placebo,” Dr. Awh said. “This represents a 135% increase in AMD progression for patients in this genotype group if treated with the AREDS formulation versus placebo.”
A criticism of these findings was forwarded by Emily Chew, MD, principal investigator of the AREDS, who pointed out that because the genotype groups were selected based on outcomes, a selection bias was introduced into the study performed by Dr. Awh and colleagues.
Dr. Awh clarified that this was not how the genotype groups were defined.
“This criticism is unjustified,” he said. “The projected outcomes we reported in our first paper for individuals with each of the 9 genotype combinations were based on the presumption that the risk associated with CFH and ARMS2 risk genotypes acted independently.
“We didn’t attempt to determine the actual outcomes for each of the 9 genotype groups, because the numbers of patients in some of the groups were too small, so it is incorrect to state that we created genotype groups based on known outcomes,” Dr. Awh continued.
The second paper reported actual progression rates in a smaller number of CFH and ARMS2 genotype group combinations, which demonstrated a significant interaction between these genetic risk factors and with treatment.
“There are only a few ways to combine 9 groups into a smaller number,” he said. “We chose to create four groups based on the frequency of the risk alleles in the AREDS population.”
AREDS Report 38, an alternative statistical analysis of 1,237 AREDS patients published by the AREDS group, failed to identify an interaction between CFH and ARMS2 genotypes and treatment with AREDS supplements.
The authors also concluded that, “The AREDS supplement reduced the rate of AMD progression across all genotype groups.”
Evaluating the AREDS 38 conclusions
Dr. Awh discussed figure 1A of the AREDS Report 38, which illustrated the overall benefit of the AREDS formulation. He pointed out that the figure was excerpted from AREDS Report 8 from 2001 and was not derived from data in the current report.
Figure 1B of the report summarized the analysis of the combined effect of the CFH and ARMS2 genotypes on treatment with the AREDS supplement components, reporting on 27 subgroups. This analysis expectedly showed no significant interaction between the treatment effect and genotype in any subgroup.
Dr. Awh noted, “The analysis was underpowered and incapable of finding an interaction, with only 1,237 patients and 27 independent subgroups, and with some groups containing virtually no patients.”
Outcomes in the table also did not demonstrate superiority of the AREDS formation over placebo in any of the 27 subgroups, because the analysis was underpowered, he noted.
“There is no proof in this analysis that the AREDS supplement reduced the rate of AMD progression across all genotype groups, or for any genotype group,” he said.
Insertion of a figure from the AREDS Report 8, a study in which no genetic analysis was performed, was misleading, he noted.
“These are inconsistent conclusions drawn from an underpowered statistical model,” Dr. Awh said. “They designed a study incapable of demonstrating a genetic interaction, conclude that there is no interaction, and then make an unsupported claim that the AREDS formulation was effective in all genotype groups.”
An alternative analysis of AREDS 38 data
Despite the analytical shortcomings, Dr. Awh that AREDS Report 38 provides an important source of independent genotyping and outcomes data.
Dr. Awh and colleagues performed an outcomes analysis of the data from AREDS Report 38 using the four CFH and ARMS2 genotype groups defined in their recent publication.
“Using the data clearly presented in Figure 1B of the report, we can identify which patients are in each genotype group, the numbers of patients in each group, and the 7-year outcomes of the patients in each group,” he said.
When Dr. Awh and colleagues re-analyzed the AREDS 38 data they found that the high ARMS2 and low CFH risk group, having an odds ratio of 0.456 for AMD progression “benefits tremendously” from the AREDS formulation compared with placebo. When they evaluated the high CFH and no ARMS2 risk group, an odds ratio for progression of 2.14 suggested an adverse response to the AREDS formulation, also consistent with the Awh analysis.
“The differences in outcome between these 2 groups were highly statistically significant, with a p value of less than 0.0005.”
Dr. Awh pointed out that this group of 1,237 AREDS patients was larger than the 989 patients used in his own analysis and represented almost half of the 2,516 patients upon whom the AREDS recommendations were based.
Dr. Awh summarized, “We found that AREDS patients in different genotype groups had dramatically different responses to the AREDS formulation. Analysis of independent genotyping and outcomes data from AREDS Report 38 supports a relationship between genotype group and treatment response. The identified interactions are biologically plausible.
“The potential for harm from zinc or the zinc-containing AREDS formulation in patients with two CFH and no ARMS2 risk alleles should be considered by clinicians and researchers,” he said. “We do not claim to have provided the last work on this interaction, but there is enough evidence to suggest caution and further careful investigation.”
• Awh CC, Lane AM, Hawken S, Zanke B, Kim IK: CFH and ARMS2 genetic polymorphisms predict response to antioxidants and zinc in patients with age-related macular degeneration. Ophthalmology. 2013;120:2317-2323.
• Awh CC, Hawken S, Zanke BW: Treatment response to antioxidants and zinc based on CFH and ARMS2 genetic risk allele number in the Age-Related Eye Disease Study. Ophthalmology. 2015;122:162-169.
• Chew EY, Klein ML, Clemons TE, et al: No clinically significant association between CFH and ARMS2 genotypes and response to nutritional supplements: AREDS report number 38. Ophthalmology. 2015;121:2173-2180.