You don't need to disprove an underpowered study. You can just default to ignoring it. Especially in a field as notorious for replication issues as fitness and nutrition.