Human postprandial responses to food and potential for precision nutrition.
Level 3 - non-randomized controlled study
Prospective cohort and postprandial challenge study with an independent validation cohort
PubMed 32528151 · doi:10.1038/s41591-020-0934-0
What was done
Investigators conducted the PREDICT 1 study in 1,002 twins and unrelated healthy adults in the United Kingdom, measuring postprandial metabolic responses to identical meals in both clinical and home settings. They evaluated the contribution of meal macronutrients, genetic variants, and individual factors (including the gut microbiome) to postprandial lipemia and glycemia. Findings were independently validated in a United States cohort (n = 100), and a machine-learning model was developed to predict individual metabolic responses to food intake.
What was found
Identical meals produced substantial inter-individual variability (population coefficient of variation) in postprandial blood triglyceride (103%), glucose (68%), and insulin (59%). Individual-specific factors such as the gut microbiome accounted for more variance in postprandial lipemia than meal macronutrient composition (7.1% vs 3.6%), whereas macronutrients had a larger influence than individual factors on postprandial glycemia (15.4% vs 6.0%). Genetic variants had modest predictive impact (9.5% for glucose, 0.8% for triglyceride, and 0.2% for C-peptide). The machine-learning model predicted postprandial triglyceride (r = 0.47) and glycemic (r = 0.77) responses.
Why it matters
This study demonstrates that individual biological factors, notably the gut microbiome, drive wide variations in postprandial metabolism beyond dietary composition alone, supporting the feasibility of machine-learning-guided personalized nutrition.
Limits
The study was conducted in healthy adult cohorts in the UK and US, which may not generalize to populations with established cardiometabolic disease or different demographic backgrounds. Predictive accuracy was considerably lower for postprandial triglycerides (r = 0.47) than for glucose (r = 0.77). Long-term clinical outcomes resulting from personalized dietary interventions based on these predictions were not evaluated.
Cited by
- supports The PREDICT trial demonstrated that gut microbiome and host data can be used to predict an individual's triglyceride levels.