NO, the CRUISE endpoint only requires 0.31 gram change for Efficacy
Cecause the SFP trial had recruted 300 patients, assuming 80% of patients complete the trial, that is 120 on each arm. Let's further assume that the standard deviation for Hgb remains the same 1.1 - 1.2 as the SFP Phase IIb trial and the Venofer (IV iron) Phase III trial (NCT00236938). With SD = 1.2, and N = 120, you only need 0.31 g/dL for efficacy.
First. data from those who do not complete the trial will for the most part be entered into the calculations. They plan to use the last few weeks (or 3 readings...don't quite remember) prior to termination as the read out for those patients. Also, if ESA is to remain constant; 37% more ESA was required in the placebo group in the shorter PII trial to maintain pseudonormal levels of Hbg; and normal Hbg levels were predefined to a fairly tight range of 9.5-11.5, then over a 12 month trial I would expect that far greater than 20% will be culled; dare I project 30-40% +/-???
You might be right that for the placebo arm where there might be more drop-outs due to hemoglobin dropping lower than 9.5 g/dL. However, the CRUISE "Primary Outcome Measures" has specified that data of the last one-sixth of the treatment period for those drop-outs will be included if that 1/6th data are more than two measures. So no, I don't expect 30-40% of those enrolled will be excluded from analysis. In the WORST scenario, assuming 40% patients being excluded from analysis from the placebo arm, that's 90 and 120 patients for placebo and treatment arms respectively. Taking standard deviation 1.2, the difference required to claim efficacy is ... ONLY 0.36 g/dL.