Nine randomised controlled trials (n=692) provided data for 13 aerobic exercise groups and seven non-randomised controlled trials (n=263) provided data for eight aerobic exercise groups. Six groups in four studies had metabolic-related disorders.

The calculated measures of exercise MET h/w ranged from 5.9 to 47.1. The change in %VF/w ranged from -6.062 to 0.078. Overall, there was a small non-significant negative correlation between MET h/w and change in %VF/w (r=-0.28, p=0.225). Groups that excluded people with metabolic-related disorders had a statistically significant negative correlation between MET h/w and change in %VF/w (r=-0.75, p=0.01). Further analysis of patients without metabolic-related disorders divided into three groups by MET h/w showed no evidence of a difference between groups. However, each exercise group was statistically significantly higher than the control group (p<0.05). The group with the greatest amount of METs h/w showed the greatest overall reduction in change in %VF/w (median approximately -1.4%).

For subgroup analyses, significant correlations were seen for women only for all data (r=-0.89, p=0.007) and short-term intervention groups without metabolic-related disorders (r=-0.81, p=0.027). For analyses of change in percentage weight per week, this showed a significant correlation with METs h/w (r=-0.79) in all data as well as those without metabolic-related disorders (r=-0.870). There was also a strong positive correlation between change in percentage weight per week and change in %VF/w in both groups (r=0.93 and r=0.64).