Abstract
Purpose of review
In this narrative review complemented by a novel meta-analysis, we critically analyzed current scientific evidence from RCTs and cohort studies regarding the impact of non-nutritive sweeteners (NNS) on cardiometabolic health, and assessed the interplay with the gut microbiome as a potential mechanistic pathway. We focused on the question of direct physiological effects of NNS, rather than the additional effects of energy displacement by NNS, to inform future research and the development of dietary and clinical guidelines.
Recent findings
Cohort studies assessing NNS from all dietary sources suggest that total NNS and each commonly used NNS are associated with higher risk of type 2 diabetes, and that total intake and specific agents are associated with certain cardiovascular disease outcomes. These findings are consistent with prior evidence from cohorts focusing on NNS in beverages. Such observational evidence may be confounded by reverse causation: people at higher cardiometabolic risk choosing to use NNS. However, our new meta-analysis of RCTs with non-caloric comparators and a recent RCT on glycemia outcomes with human-to-mice microbiota transplant suggest that NNS have harmful effects on glucose-insulin homeostasis including fasting insulin, HbA1c, and glucose area under the curve during oral glucose tolerance test (OGTT), potentially mediated by effects on the composition and functional potential of the gut microbiome.
Summary
The summed evidence supports potential long-term risk of cardiometabolic diseases associated with NNS intake and short-term harmful effects of NNS on glycemia. Future clinical trials of physiologic effects and molecular mechanisms will strengthen interpretations and causal inference. Given potential for harm, caution is warranted for the use of NNS.
Summary
This narrative review of randomised controlled trials (RCTs) and prospective cohort studies, alongside a meta-analysis of RCTs, suggests that consumption of low/no calorie sweeteners (LNCS) may adversely affect glucose-insulin regulation compared with non-caloric controls such as water or placebo. However, the meta-analysis findings should be interpreted with caution as their clinical significance is uncertain, while evidence from observational studies may be influenced by reverse causation and residual confounding.
The meta-analysis of RCTs indicated that LNCS intake increased fasting insulin concentrations (mean difference 6.80 pmol/L; 16 RCTs, n=888) and slightly increased HbA1c (mean difference 0.07%; 11 RCTs, ≥2 months, n=659). A trend toward reduced insulin sensitivity was reported but it was not statistically significant, and a reduction in blood pressure that was initially identified became non-significant after sensitivity analysis. No significant effects were observed for other measures inclusing fasting glucose, blood lipids, or body mass index. The authors interpret these findings as suggesting a potential adverse effect of LNCS on aspects of glucose–insulin homeostasis under conditions where sugar replacement benefits are not the case.
However, several points should be considered when interpreting these results. The observed change in HbA1c was very small (0.07%) and only marginally above the null value, raising questions about its clinical relevance despite statistical significance. Likewise, fasting insulin is a surrogate biomarker without established thresholds for clinically meaningful change in isolation, particularly in the absense of consistent differences in insulin sensitivity or other glycaemic outcomes, limiting direct interpretation for health outcomes. Finally, the review’s focus on LNCS compared with non-caloric controls does not reflect typical real-world use, where LNCS primarily function as replacements for sugar and contribute to reduced energy and added sugar intake. Evidence from substitution studies confirming benefits of LNCS use as sugar substitutes remains key when informing dietary recommendations (Ayoub-Charette et al, 2025).