My understanding of the need for two studies was that the first...

My understanding of the need for two studies was that the first had to demonstrate statistical significance of efficacy and the second has to confirm (though not necessarily statistically significantly) - can anyone confirm or correct this?

If so, the FDA re-analysis still shows one significant study (302) but the second study (301) cannot be considered confirmatory after accounting for the impact of differential loss to follow-up (in the same direction, but not strong because of the large early losses in 301). If a reanalysis is possible using the original intention to treat (rather than the modified) the continuous data (FEV1) is likely to show a more positive outcome for 301 than the dichotomous outcome the FDA were forced to use. If these FEV1 data are not available for a substantial proportion of the patients, it is possible that PXS may need one more RCTs to confirm or reject the result of 302.

There is another possible problem. The FDA report suggests that study 302 may have been inadequately blinded. Interestingly I can only find mention of this possibility in the Division Memorandum (as part of a record of the 10 December pre-NDA meeting) and not in the Clinical or Statistical briefings. It is possible, therefore, that this is not currently considered a problem. Any views?