Pooled across ages, melatonin cut about 7 minutes off falling asleep. Split by age, adults whose insomnia arrives on its own showed nothing significant.
The older pooling mixed adults and children without separating them, so its average may be describing the children in it. In the newer one, the adult comorbid group rests on a single study out of 24, and only four of those 24 covered insomnia alongside another condition at all. The two reviews draw on overlapping trials, so they are not independent checks on each other. Neither names which other conditions those four trials covered.
The open question is who melatonin's average benefit belongs to. Pooled across all ages it helped; split by age, adults whose insomnia arrives on its own showed nothing significant, and the benefit looks to belong to children.