Two questions it would be nice to have answers to:

1. Was the standard deviation derived from people with normal scores or from patients with low scores?

2. Given that CVD rates were lower in the placebo group than expected, did whatever account for that also improve cognitive function in the placebo group?

And remember, even 99 per cent power does not give an absolute guarantee of detecting an effect. There's always that one per cent!

Let's hope!