Using a Metabotype Framework to Deliver Personalized Nutrition Improves Dietary Quality and Metabolic Health Parameters: A 12‐Week Randomized Controlled Trial

Abstract

Scope Effective strategies for tailoring dietary advice to individuals are urgently needed. The effectiveness of personalized nutrition advice delivered using a metabotype framework in improving dietary quality and metabolic health biomarkers compared to population‐level advice is investigated. Materials and results A 12‐week parallel randomized controlled trial is performed with 107 healthy adults. Individuals in the personalized group are classified into metabotypes using four markers (triacylglycerol, high‐density lipoprotein [HDL]‐cholesterol, total cholesterol [TC], and glucose) and received dietary advice from decision tree algorithms containing metabotypes characteristics and individual traits. Individuals in the control group received generic dietary advice based on national guidelines. The personalized approach results in higher dietary quality assessed by the Alternate Mediterranean Diet Score (effect size [95% confidence interval, CI], 0.77 [0.07, 1.48], 12% versus 3% increase) and significantly lower concentrations of triacylglycerol (−0.17 [−0.28, −0.06] log10 mmol L−1), TC (−0.42 [−0.74, −0.10] mmol L−1), low‐density lipoprotein (LDL)‐cholesterol (−0.34, [−0.60, −0.09] mmol L−1), and lower triacylglycerol‐glucose index (−0.40, [−0.67, −0.13]). Sixteen phosphatidylcholines and six lysophosphatidylcholines, predominately with chain lengths of 30–36 carbons, are lower in the personalized group. Conclusions Personalized nutrition advice delivered using the metabotype framework is effective to improve dietary quality, which could result in reduced CVD risk, and metabolic heath biomarkers.

Human Nutrition Research in the Data Era: Results of 11 Reports on the Effects of a Multiple-Micronutrient-Intervention Study

“…Our report contributes to an emerging trend of precision nutrition, using physiological measures to predict response to diet. Others have used postprandial glucose levels [ 6 ], k-means groups based on plasma triacylglycerol, total cholesterol, HDL-cholesterol, and glucose [ 41 ], and response to meal challenges (rev in [ 13 ]) to predict responses to dietary changes. The microbiome is also likely to contribute to individual responses to nutrition (rev.…”

“…Virtually all plasma metabolites measured in the studies reviewed here were similar to the pattern observed for LDL-cholesterol ( Figure 2 ); that is, there was a “continuous” distribution of the measured metabolite across the population. As a step toward personalized nutrition, we and others are defining “metabolic groups”—or metabotypes [ 95 ] (and rev in [ 79 ])—as based on some a priori criteria [ 41 ] or through discovery-based approaches using unsupervised algorithms. Metabotypes group individuals with similar metabolic, genomic, or proteomic similarities.…”

Distinct patterns of personalised dietary advice delivered by a metabotype framework similarly improve dietary quality and metabolic health parameters: secondary analysis of a randomised controlled trial

“…The present paper reports secondary analyses carried out on data from participants that received personalised nutrition advice using a metabotype framework as part of a 12-week single-blind parallel RCT (ISRCTN15305840) ( 22 ). Throughout the study, participants attended three visits (baseline, week 4, and week 12) and received reports with dietary advice following baseline and week 4 visits.…”

“…Participants received personalised nutrition advice based on a metabotype framework previously published ( 23 , 24 ). The advice consisted of two reports containing three dietary goals each that were delivered following baseline and week 4 visits ( 22 ). Following week 4, participants were asked to follow both reports at the same time.…”

“…Recently, we demonstrated with a 12-week randomised controlled trial (RCT) that personalised dietary advice based on metabotypes for a generally healthy population produced greater improvements in dietary quality and lipid profile (reductions in total cholesterol, LDL-C, triacylglycerol and glycerophosphocholines) compared to population-level nutrition advice ( 22 ). The framework applied comprises the classification of individuals into metabotypes using metabolic health markers and the delivery of dietary advice through a decision tree approach ( 23 , 24 ), which results in a set of food or food group-related messages that are personalised for each individual.…”