Reply

Dr Odaibo correctly identifies that the relationship between high-dose zinc supplements (as in the Age-Related Eye Disease Study [AREDS] formulation), genotype, and age-related macular degeneration (AMD) progression risk is an important public health issue. He argues that a dense sampling of a population is likely to result in a more accurate representation than a sparse sampling. A more familiar approach to this issue is the concept of statistical significance. How large a sample is needed to reduce the risk of a false representation of the sampled population to an acceptable level? We found the difference in response to zinc-containing treatments in AREDS patients with high-risk CFH/low-risk ARMS2 versus low-risk CFH/high-risk ARMS2 under a Cox model to be significant to P < 0.00099, meaning that the chance that the observed relationship is owing to chance alone is <1 in 1000.