Disclaimer I only have access to the abstract. It seems unreal to me that you could explain a bmi jump from 25 to 31 in six months with a fruit diet. Their randomization doesn’t pass the sniff test.

IMHO, again without having access to the article, the most likely explanation is they have poorly matched trial and control groups with small (80) number of participants. That they didn’t report changes at all and just the final values raises a red flag.

The other major caveat is this was presumably done with Iranians and I have mostly no idea what their baseline diet is.