Sir,

We would like to thank Drs Blann and Lip for their interest in our paper.1 We believe that the association of atherosclerosis and age-related macular degeneration (AMD) is a complex one,2 it is likely that they share a common mechanism but local factor might dictate the final disease process. It is possible that they share similar risk factors. The suggestion of a cross-sectional analysis of the two groups, one with AMD but without atherosclerosis and one with atherosclerosis but no AMD, suffers selection bias and it is not practically possible to say who are without atherosclerosis, at least in an eye clinic. In our study design, we included consecutive patients and normal controls, using a medical questionnaire to correct, for potential alternations in MMP levels. As the method of acquisition is identical, it reduces bias and also allowed us to study the interaction of atherosclerosis and AMD.

The rest of the concern of Drs Blann and Lip was on our statistical analysis. The mere reason that there are so many different statistical tests is that no one can really agree on the best test to be used in different circumstances. Although I do agree with their comments, the threefold increase of MMP-9 levels in AMD patients will be significant, no matter how you want to analyse the data.

As for the specific comment, the statistical analysis of the clinical data was summarised in Table 1 in the published paper. This is a pilot study as the number of patients was small, as we have mentioned in the original paper, and we need a larger study (which is underway) to confirm our findings and the current data set has allowed us to perform more accurate study size calculation.