They weren't. It looks like the lowest in the study was ~20%, which is about the upper bound on what could be considered "low carb".