Zhang · Nutrients 2023 · systematic review and network meta-analysis · n=36 trials (472 participants)

The Effect of Non-Nutritive Sweetened Beverages on Postprandial Glycemic and Endocrine Responses: A Systematic Review and Network Meta-Analysis.

Cited 52 times in the scientific literature.

Level 1 - systematic review of randomized trials

Systematic review and network meta-analysis of clinical trials

PubMed 36839408 · doi:10.3390/nu15041050 · record verified 2026-08-29

What was done

A systematic review and network meta-analysis of acute, single-exposure randomized and non-randomized clinical trials in humans was conducted across MEDLINE, EMBASE, and The Cochrane Library through January 15, 2022. It examined non-nutritive sweetened (NNS) beverages versus water and sugar-sweetened beverages across three patterns: uncoupling (consumed alone), coupling (consumed with carbohydrates), and delayed coupling (preload before carbohydrates). The primary outcome was 2-hour incremental area under the curve (iAUC) for blood glucose, with secondary 2-hour iAUC outcomes for insulin, GLP-1, GIP, PYY, ghrelin, leptin, and glucagon.

What was found

Thirty-six trials involving 472 predominantly healthy participants were included. The abstract provides no quantitative point estimates or confidence intervals. In uncoupling interventions, single or blended NNS beverages had no effect on postprandial glucose, insulin, GLP-1, GIP, PYY, ghrelin, or glucagon, performing similarly to water (low to moderate confidence). Sugar-sweetened beverages increased postprandial glucose, insulin, GLP-1, and GIP. In coupling and delayed coupling interventions, NNS beverages showed no postprandial glycemic or endocrine differences compared to controls (low to moderate confidence).

Why it matters

This review demonstrates that acute consumption of non-nutritive sweeteners does not trigger postprandial glycemic or incretin perturbations, supporting their use as an acute metabolic substitute for sugar-sweetened beverages.

Limits

The abstract provides no exact numerical effect sizes, variance estimates, or confidence intervals. The evidence is derived from acute single-exposure studies with small trial cohorts (average ~13 participants per study) in mostly healthy populations, includes non-randomized designs, and was graded as having low to moderate confidence.

Cited by