This is important because it is not uncommon to find interventions (including drugs) that benefit people who do not have a given condition that do nothing for people who have the condition.
To take a ridiculous but hopefully memorable example: steroids will not help a person with no arms lift more weight, even though they may help other people a lot. In the case of metabolic diseases, sufferers are generally missing important metabolic pathways that the rest of us have, so the example isn't completely irrelevant.
In the present case, I think it's perfectly reasonable that exercise will help (although what I think is perfectly reasonable and what is true have nothing to do with each other) and as the Atlantic articles says, other studies have shown this, but I just wanted to make a pedantic point about that particular pediatric paper.