It would be interesting to see comparisons to tea and decaf coffee, correcting for the amount of caffeine in those beverages. They seem to be assuming caffeine is the causal agent here, in which case there should be detectable parallels with other caffeinated beverages.