Abstract

Purpose This study aimed at quantifying and ranking the effects of different foods or food groups on weight loss. Methods We searched PubMed, Scopus, Cochrane Central Register of Controlled Trials, and Embase to April 2021. We included randomized trials evaluating the comparative effects of two or more food groups, or compared a food group against a control group (usual diet, no intervention) for weight loss in adults. We conducted random-effects network meta-analysis with Bayesian framework to estimate mean difference MD and 95% credible interval CrI of the effect of food groups on weight loss. Results 152 RCTs with 9669 participants were eligible. Increased consumption of fish (MD - 0.85 kg, 95% CrI - 1.66, - 0.02; GRADE = low), whole grains (MD - 0.44 kg, 95% CrI - 0.88, 0.0; GRADE = very low), and nuts (MD - 0.37 kg, 95% CI - 0.72, - 0.01; GRADE = low) demonstrated trivial weight loss, well below minimal clinically important threshold (3.9 kg), when compared with the control group. Interventions with other food groups led to no weight loss when compared with either the control group or other food groups. The certainty of the evidence was rated low to very low with the point estimates for all comparisons less than 1 kg. None of the food groups showed an important reduction in body weight when restricted to studies conducted in participants with overweight or obesity. Conclusions Interventions with a single food or food group resulted in no or trivial weight loss, especially in individuals with overweight or obesity. Further trials on single foods or food groups for weight loss should be highly discouraged.