Health, not weight loss, focused programmes versus conventional weight loss programmes for cardiovascular risk factors: A systematic review and meta-analysis

Abstract

Background: Obesity is a cardiovascular disease risk factor. Conventional weight loss (CWL) programmes focus on weight loss, however 'health, not weight loss, focused' (HNWL) programmes concentrate on improved health and well-being, irrespective of weight loss. What are the differences in CVD risk outcomes between these programmes? Aim: To conduct a systematic review and meta-analysis to compare the effects of HNWL with CWL programmes on cardiovascular disease risk factors. Methods: We searched CENTRAL, MEDLINE, EMBASE, PsycINFO, CINAHL, ASSIA, clinical trial registers, commercial websites and reference lists for randomised controlled trials comparing the two programmes (initially searched up to August 2015 and searched updated to 5 April 2019). We used the Mantel-Haneszel fixed-effect model to pool results. Sub-group and sensitivity analyses that accounted for variations in length of follow-up, enhanced programmes and risk of bias dealt with heterogeneity. Results: Eight randomised controlled trials of 20,242 potential studies were included. Improvements in total cholesterol-HDL ratio (mean difference-0.21 mmol/L, 95% confidence interval [-3.91, 3.50]) and weight loss (-0.28 kg [-2.00, 1.44]) favoured HNWL compared to CWL programmes in the long term (53-104 week follow-up), whereas improvements in systolic (-1.14 mmHg, [-5.84, 3.56]) and diastolic (-0.15 mmHg, [-3.64, 3.34]) blood pressure favoured CWL programmes. These differences did not reach statistical significance. Statistically significant improvements in body satisfaction (-4.30 [-8.32,-0.28]) and restrained eating behaviour (-4.30 [-6.77,-1.83]) favoured HNWL over CWL programmes. Conclusions: We found no long-term significant differences in improved CVD risk factors; however, body satisfaction and restrained eating behaviour improved more with HNWL compared to CWL programmes. Yet firm conclusions cannot be drawn from small studies with high losses to follow-up and data sometimes arising from a single small study.