They probably needed more time for various types of data, but I doubt they needed more time to release the ADL mean change. This is because the odds ratios they reported are produced by the same data as the ADL mean change, however the latter is much simpler to compute. IMO, the ADL change for the treated group is either worse than placebo or the difference to placebo is not statistically significant, even when using one-tail p-values. This is for the whole sample of treated patients. Maybe in the 50mg subsample the picture will be completely different.