Most likely because there were only 7 or 8 people with insomnia in the trial. That number is way too small to be granting AA upon, even though there was improvement on 100% of that very small number.
We don't know if those people were on both drugs or just 2-73.