Personalized Nutrition by Prediction of Glycemic Responses.
Level 2 - randomized trial
Blinded randomized controlled trial component alongside prospective cohort derivation and validation
PubMed 26590418 · doi:10.1016/j.cell.2015.11.001
What was done
Researchers tracked week-long continuous glucose levels in an 800-person cohort across 46,898 meals. Using blood parameters, dietary habits, anthropometrics, physical activity, and gut microbiota, they trained a machine-learning algorithm to predict personalized postprandial glycemic responses. The model was validated in an independent 100-person cohort. Finally, a blinded randomized controlled dietary intervention based on the algorithm was conducted to evaluate postprandial glucose responses and gut microbiota changes.
What was found
The authors observed high interpersonal variability in glycemic response to identical meals. The machine-learning algorithm accurately predicted personalized postprandial glycemic responses in both the discovery cohort and the 100-person validation cohort. In the blinded randomized controlled trial, the algorithm-based personalized diet significantly lowered postprandial glycemic responses and induced consistent alterations in gut microbiota configuration. The abstract reports no numerical values, effect sizes, or p-values.
Why it matters
This work demonstrates that universal dietary advice overlooks substantial individual glycemic variation, providing proof-of-concept for personalized, algorithm-driven nutrition to manage blood glucose.
Limits
The abstract provides no numerical data, effect sizes, or statistical confidence intervals. The primary monitoring period was short (one week), the specific sample size of the randomized intervention sub-group was not stated, and long-term metabolic outcomes were not evaluated.
Cited by
- context In the Personalized Nutrition Project and subsequent interventional trials, personalized dietary modifications led to improvements in HDL cholesterol.
- partial Over 100,000 individuals have been tested on thousands of food components and food additives through the Personalized Nutrition pipeline.
- supports Dietary meal timing, including the timing of dinner the previous night, serves as a predictive feature for postprandial glycemic response algorithms.
- supports In a 2015 study of 1,000 individuals, machine learning algorithms integrating continuous glucose monitoring and microbiome data accurately predicted personalized glycemic responses to food.
- context For some individuals, eating tomatoes alone causes a sharp rise in blood glucose, whereas combining tomatoes with white bread reduces the glycemic response.