Actually, I am a bit unimpressed by the body weight data. You can see clearly in their individual data that there are a ton of errors in readings, and while the

Actually, I am a bit unimpressed by the body weight data. You can see clearly in their individual data that there are a ton of errors in readings, and while the SEM is low by the final day, I'm not sure you can make much of that information when so many of the individuals had daily swings >1kg. It does look like OGTT changes are dependent on weight change (post-hoc weight-gaining group did not see the same benefit as the total cohort). To me, the diastolic BP, baseline fasting insulin, and OGTT AUC are the real stars here.