The problem is, that although the difference in MACE between ABL and placebo in the low GFR patients is impressive, it is not certain it was due to the ABL and not to something else (because the low GFR patients were not identified at baseline and then randomised as a subgroup to the two treatments). The TOTAL patients were randomised to the treatments without reference to GFR. That was a weakness in the design of the BETonMACE.