Iconoclast,

Cityslicker asked the same question previously and Bear gave the same (correct) answer after I gave the wrong answer!  But here is what needs to be reconciled:

How do you get the following results from two analyses of the same dataset:

10% of patients (CKD) get a MACE reduction of about 50% whereas the other 90% get a minimal benefit, and also get

50% of patients (low LDL) get a MACE reduction of about 40% (almost as good as the CKD result but in 5 times as many patients) whereas the other 50% get no benefit?

One way is to have the CKD patients disproportionately present in the low LDL group but this does not appear to be the case.  Another way is to have a synergistic effect between CKD and low LDL such that the CKD patients in the low LDL group did better than the CKD patients in the high LDL group.  In other words, if there aren't more CKD patients in the low LDL group than in the high LDL group, the ones that are in the low LDL group must have done better than the ones in the high LDL group.

Anyone have another explanation?

Jupe