If I could add my 2 cents, the p value for placebo shouldn't really matter seeing as the reduction in AHI was insignifiant (exactly what you'd expect with a placebo). Did this occur by chance? Well, seeing as nothing of significance actually occured, I expect this is not very meaningful to the outcome of the trial results.
Thoughts on your comment
@hubbo27. Perhaps the p values for the control were higher because placebo was only given to each participant for a single 7 day period, whereas the drug was administered at low, med, or high doses over the remianing 3 weeks. That could explain the difference.