King's College London

Research portal

Environmental risk factors and nonpharmacological and nonsurgical interventions for obesity: An umbrella review of meta-analyses of cohort studies and randomized controlled trials

Research output: Contribution to journalReview article

Marco Solmi, Cristiano A Köhler, Brendon Stubbs, Ai Koyanagi, Beatrice Bortolato, Francesco Monaco, Davy Vancampfort, Myrela O Machado, Michael Maes, Ioanna Tzoulaki, Joseph Firth, John P A Ioannidis, André F Carvalho

Original languageEnglish
Article numbere12982
JournalEuropean Journal of Clinical Investigation
Early online date13 Jul 2018
Publication statusE-pub ahead of print - 13 Jul 2018

King's Authors


BACKGROUND: Multiple environmental factors have been implicated in obesity, and multiple interventions, besides drugs and surgery, have been assessed in obese patients. Results are scattered across many studies and meta-analyses, and they often mix obese and overweight individuals.

MATERIALS AND METHODS: PubMed and Cochrane Database of Systematic Reviews were searched through 21 January 2017 for meta-analyses of cohort studies assessing environmental risk factors for obesity, and randomized controlled trials investigating nonpharmacological and nonsurgical therapeutic interventions for obesity. We excluded data on overweight participants. Evidence from observational studies was graded according to criteria that included the statistical significance of the random-effects summary estimate and of the largest study in a meta-analysis, the number of obesity cases, heterogeneity between studies, 95% prediction intervals, small-study effects and excess significance. The evidence of intervention studies for obesity was assessed with the GRADE framework.

RESULTS: Fifty-four articles met eligibility criteria, including 26 meta-analyses of environmental risk factors (166 studies) and 46 meta-analyses of nondrug, nonsurgical interventions (206 trials). In adults, the only risk factor with convincing evidence was depression, and childhood obesity, adolescent obesity, childhood abuse and short sleep duration had highly suggestive evidence. Infancy weight gain during the first year of life, depression and low maternal education had convincing evidence for association with paediatric obesity. All interventions had low or very-low-quality evidence with one exception of moderate-quality evidence for one comparison (no differences in efficacy between brief lifestyle primary care interventions and other interventions for paediatric obesity). Summary effect sizes were mostly small across compared interventions (maximum 5.1 kg in adults and 1.78 kg in children) and even these estimates may be inflated.

CONCLUSIONS: Depression, obesity in earlier age groups, short sleep duration, childhood abuse and low maternal education have the strongest support among proposed risk factors for obesity. Furthermore, there is no high-quality evidence to recommend treating obesity with a specific nonpharmacological and nonsurgical intervention among many available, and whatever benefits in terms of magnitude of weight loss appear small.

View graph of relations

© 2018 King's College London | Strand | London WC2R 2LS | England | United Kingdom | Tel +44 (0)20 7836 5454