Effects of non-nutritive sweeteners on the BMI of children and adolescents: a systematic review and meta-analysis of randomised controlled trials and prospective cohort studies

BackgroundConsidering the biological variation across subgroups during periods of growth, the role of non-nutritive sweeteners in weight-related outcomes among children and adolescents is unclear. We did a systematic review and meta-analysis to summarise the evidence on experimental and habitual consumption of non-nutritive sweeteners and prospective changes in BMI in paediatric populations.MethodsWe searched eligible (ie, lasting a minimum of 4 weeks) randomised controlled trials of the effect of non-nutritive sweeteners versus non-caloric or caloric comparators on BMI change and prospective cohort studies reporting multivariable-adjusted coefficients for non-nutritive sweetener intake and BMI in children (aged 2–9 years) and adolescents (aged 10–24 years). We generated pooled estimates using random effects meta-analysis and did secondary stratified analyses to explore heterogeneity by study-level and subgroup characteristics. We further evaluated the quality of the included evidence and classified industry-funded studies, or those whose authors were related to the food industry, as having potential conflicts of interest.FindingsFrom 2789 results, we included five randomised controlled trials (n=1498 participants; median follow-up 19·0 weeks IQR 13·0–37·5); three 60% with potential conflicts of interest), and eight prospective cohort studies (n=35 340 participants; median follow-up 2·5 years IQR 1·7–6·3; two 25% with potential conflicts of interest). Random allocation to intake of non-nutritive sweeteners (25–2400 mg/day, from food and beverages) suggested less BMI gain (standardised mean difference –0·42 kg/m2 95% CI –0·79 to –0·06; I2=89%) compared with intake of sugar from food and beverages. Stratified estimates were significant only in adolescents, participants with obesity at baseline, consumers of a mixture of non-nutritive sweeteners, longer trials, and trials not found to have potential conflicts of interest. No randomised controlled trials tested beverages containing non-nutritive sweeteners versus water. Prospective cohorts reported a non-significant association between consumption of beverages containing non-nutritive sweeteners and BMI gain (0·05 kg/m2 95% CI –0·02 to 0·12; I2=67%; per daily serving of 355 mL), which was accentuated for adolescents, boys, and cohorts with longer follow-ups. Removing studies with potential conflicts of interest attenuated the estimates. Evidence was predominantly classified as of low to moderate quality.InterpretationIntake of non-nutritive sweeteners versus sugar in randomised controlled trials resulted in less BMI gain in adolescents and participants with obesity. Better designed studies should contrast beverages containing non-nutritive sweeteners with water. Long-term prospective analyses with changes in repeated measures might clarify the effect of intake of non-nutritive sweeteners on BMI changes in childhood and adolescence.FundingNone.