Sixteen trials met the inclusion criteria, of which 13 trials (3,907 participants) were included in the analyses. Follow-up times ranged from six months to five years. The authors identified a number of potential quality issues, including poor reporting of randomisation methods, and lack of blinding, particularly in the lifestyle intervention trials.
In direct treatment comparisons, lifestyle interventions were found to increase the odds of metabolic syndrome reversal (OR 3.81, 95% CI 2.47 to 5.88; eight trials; Ι²=57%), as were pharmaceutical interventions (OR 1.59, 95% CI 1.04 to 2.45; five trials; Ι²=69%) and anti-diabetic drugs specifically (OR 2.56, 95% CI 1.07 to 6.13; three trials; Ι²=78%). There was no evidence of publication bias.
The results of the multiple treatment comparison analysis agreed with the direct comparisons, and found that lifestyle interventions had the highest probability of being the best choice (87.4%). When broken down into specific types of intervention, diet combined with exercise had the highest probability of being the best choice (33.8%), followed by anti-obesity drugs with lifestyle advice (31.4%).