Vegetable-First Meal Sequence Suppresses Postprandial Glycaemic Excursions and Variability in Young Women With High-Risk Phenotype for Early-Onset Type 2 Diabetes: A 4-Arm Randomized Crossover Trial With Continuous Glucose Monitoring.
Level 2 - randomized trial
Four-arm randomized crossover trial in humans
PubMed 42608327 · doi:10.1111/dom.71227
What was done
Thirty-three healthy Korean women with a high-risk phenotype for early-onset type 2 diabetes completed a 4-arm randomized crossover trial. Participants consumed isocaloric test meals across 3-day repeated interventions under four conditions: (1) vegetable-protein-carbohydrate sequence (VPC) within 35 minutes; (2) carbohydrate-protein-vegetable sequence (CPV) within 35 minutes; (3) all components eaten together (TGR) within 35 minutes; and (4) meals consumed with no time limit (NTL). Postprandial glycaemic profiles and variability were monitored via continuous glucose monitoring for 3 hours after each meal.
What was found
Compared with CPV, TGR, and NTL, the vegetable-first sequence (VPC) significantly reduced the 0–1 hour and 0–2 hour glucose incremental area under the curve (iAUC) and 0–1 hour mean glucose (p < 0.05). VPC also lowered peak glucose and delayed time-to-peak compared to CPV (p < 0.0001). Time in range was highest for VPC, and mean amplitude of glycaemic excursion (MAGE) was significantly lower in VPC compared to all other conditions (p < 0.05). Exact numerical values for glucose metrics, iAUC, and MAGE were not reported in the abstract.
Why it matters
Demonstrates that meal sequencing (vegetables first, carbohydrates last) reliably dampens acute postprandial glucose spikes and variability even when repeated over consecutive days or compared to unconstrained eating. This supports meal order as a simple, non-pharmacological dietary tactic for glycaemic control in young women at elevated risk for type 2 diabetes.
Limits
The sample size was small (n = 33) and restricted entirely to young Korean women, limiting generalizability to other demographics, older adults, or individuals with established diabetes. The intervention was very short (3 days per arm), leaving long-term metabolic impacts, adherence feasibility, and hard clinical outcomes unmeasured. Absolute numerical means, effect sizes, and confidence intervals were omitted from the abstract.