Intensity and duration of lifestyle interventions for long-term weight loss and association with mortality: A meta-analysis of randomised trials

Abstract

Objectives To evaluate the importance of the frequency and duration of lifestyle interventions for achieving weight loss over ≥1 year and associations with all-cause mortality. Design Meta-analysis of randomised trials using Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines and RevMan software version 5·2 (The Nordic Cochrane Centre, The Cochrane Collaboration, Copenhagen). Data sources MEDLINE, CENTRAL, Google and Science Direct databases alongside reference lists of appropriate articles and meta-analyses. Eligibility criteria Randomised studies published in English-language journals from 1980 to June 2018 that assessed lifestyle compared with control interventions on weight loss and that included ≥100 subjects and reported weight change and mortality for ≥1 year. Data extraction and synthesis Two independent reviewers extracted data and assessed risk of bias. Data were pooled using the generic inverse-variance method and expressed as mean differences (MDs) with 95% CI and OR with 95% CI as appropriate. Heterogeneity was assessed (Cochran Q statistic) and quantified (I 2 statistic). The Grading of Recommendations Assessment, Development, and Evaluation score was used to assess the certainty of the evidence. Results 31 randomised trials with a total of 20 816 overweight or obese participants were included. 70% of participants had cardiometabolic risk factors. Body weight was lower for lifestyle intervention compared with the control at 1 year (3.63 kg, 95% CI 2.58 to 4.67) and at 3 years (2.45 kg, 95% CI 1.17 to 3.73). Weight loss at 1 year was greater in studies with >28 compared with ≤28 interventions per year (4.50 kg, 95% CI 3.03, 5.97 vs 2.38, 95% CI 0.78 to 3.98 kg, p=0.001). In all studies, there were 593 deaths (∼0.3%/year). The ORs for mortality for weight loss interventions compared with the controls was 0.86 (95% CI 0.73 to 1.02), p=0.09. Conclusion In predominantly healthy populations with risk factors, there is a dose response with number of lifestyle interventions and weight loss. Frequent and sustained interventions are needed to achieve a clinically significant 5% weight loss. There was insufficient evidence to reliably evaluate the benefits in persons with known cardiovascular disease or cancer.