A Systematic Review and Meta-Analysis of Changes in Body Weight in Clinical Trials of Vegetarian Diets

Author(s): 

Barnard ND, Levin SM, Yokoyama Y

Journal: 

Journal of the Academy of Nutrition and Dietetics

Year Published: 

2015

Category: 

Study Design: 

Abstract: 

In observational studies, vegetarians generally have lower body weights compared with omnivores. However, weight changes that occur when vegetarian diets are prescribed have not been well quantified. We estimated the effect on body weight when vegetarian diets are prescribed. We searched PubMed, EMBASE, and the Cochrane Central Register of Controlled Trials for articles through December 31, 2013. Additional articles were identified from reference lists. We included intervention trials in which participants were adults, interventions included vegetarian diets of ≥4 weeks' duration without energy intake limitations, and effects on body weight were reported. Two investigators independently extracted data using predetermined fields. Estimates of body weight change, comparing intervention groups to untreated control groups, were derived using a random effects model to estimate the weighted mean difference. To quantify effects on body weight of baseline weight, sex, age, study duration, study goals, type of diet, and study authorship, additional analyses examined within-group changes for all studies reporting variance data. We identified 15 trials (17 intervention groups), of which 4 included untreated controls. Prescription of vegetarian diets was associated with a mean weight change of -3.4 kg (95% CI -4.4 to -2.4; P<0.001) in an intention-to-treat analysis and -4.6 kg (95% CI -5.4 to -3.8; P<0.001) in a completer analysis (omitting missing post-intervention values). Greater weight loss was reported in studies with higher baseline weights, smaller proportions of female participants, older participants, or longer durations, and in studies in which weight loss was a goal. Using baseline data for missing values, I2 equaled 52.3 (P=0.10), indicating moderate heterogeneity. When missing data were omitted, I2 equaled 0 (P=0.65), indicating low heterogeneity. Studies are relatively few, with variable quality. The prescription of vegetarian diets reduces mean body weight, suggesting potential value for prevention and management of weight-related conditions.