Abstract
Polyol sweeteners in sugar-free beverages can cause osmotic diarrhea after acute bolus intake. Reference points, mostly no-observed-adverse-effect levels, depend on tested doses and lack uncertainty intervals. A systematic search yielded eight datasets covering six polyols; six datasets (five polyols) formed the primary analysis, with dose-escalation sorbitol data analyzed exploratorily. Bayesian model averaging at 10% extra risk yielded adult lower credible bounds on benchmark doses (BMDLs) of 0.400, 0.358, 0.162, 0.132, and 0.087 g/kg body weight for erythritol, maltitol, xylitol, isomalt, and lactitol, respectively; the pediatric erythritol BMDL was 0.552 g/kg body weight. Lowering the benchmark response to 5% reduced BMDLs by 44.4–53.7% without changing adult rankings. Using consumption and body-weight data, BMDLs were converted into concentrations. For a 20 kg child consuming 0.9 L once (uncertainty factor 1), reference concentrations were 1.9–12.2 g/L, inversely proportional to volume. Differences in matrices, endpoints, background incidences, and dose designs preclude definitive intrinsic potency ranking. Most estimates required low-dose extrapolation; except for erythritol, child scenarios used adult data without pediatric dose–response evidence. These concentrations are not safety or regulatory limits, formulation recommendations, individual no-effect thresholds, or concentrations validated to protect all children. Values derived from solid matrices require validation in beverages.