RESUMO
Background: Scientists have been investigating efficient interventions to prevent and manage obesity. This network meta-analysis (NMA) compared the effect of different diets [moderate macronutrients (MMs), low fat/high carbohydrate (LFHC), high fat/low carbohydrate (HFLC), and usual diet (UD)] on weight, body mass index (BMI), and waist circumference (WC) changes at ≥12 months. Methods: We searched Medline, Embase, PubMed databases, and the Cochrane Library. We systematically assessed randomized controlled trials (RCTs) evaluating dietary interventions on adults (mean BMI ≥ 25 kg/m2) receiving active dietary counseling for ≥12 months. We pooled the data using a random-effect NMA. We assessed the quality of the included RCTs using the Cochrane risk of bias (ROB) tool. Results: We included 36 trials, 14 of which compared HFLC with MM diets. Compared with UD, all diets were associated with a significant weight loss (WL) at ≥12 months, HFLC [mean difference in kg (95% CI): -5.5 (-7.6; -3.4)], LFHC [-5.0 (-7.1; -2.9)] and MM [-4.7 (-6.8; -2.7)]. HFLC, compared with MM diet, was associated with a slightly higher WL (of -0.77 kg) and drop in BMI (of -0.36 kg/m2), while no significant difference was detected in other dietary comparisons. WC was lower with all diets compared to UD, with no significant difference across specific diets. There was no significant interaction of the results with the pre-specified sub-groups. The ROB was moderate to high, mostly related to unclear allocation concealment, high dropout rate and unclear or lack of blinding of participants, providers, and outcome assessors. Conclusion: Dietary interventions extending over ≥12 months are superior to UD in inducing weight, BMI and WC loss. HFLC might be associated with a slightly higher WL compared with MM diets. Systematic Trial Registration: https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=103116, PROSPERO (CRD42018103116).