Metabolic Phenotyping Identifies Responders to Sustainable Healthy Diets: Analysis from the MyPlanetDiet Randomized Controlled Trial

Background Understanding inter-individual variation is critical for developing strategies to improve health. Metabolic phenotypes impact how individuals respond to dietary change. While literature has examined response to dietary interventions between metabolic phenotypes, none have yet considered sustainable healthy diets. Objective Exploratory, hypothesis-generating, post-hoc secondary analysis from the MyPlanetDiet randomised controlled trial is used to examine subgroup response to sustainable healthy diets. The aim is to identify metabolic phenotype subgroups (MPs) within the MyPlanetDiet sample population and examine if MPs responded differently to dietary change. Methods The MyPlanetDiet study was a 12-week randomised controlled trial testing the impacts of more sustainable healthy diet advice relative to advice from food-based dietary guidelines. Dietary assessments, fasting serum samples and anthropometry were collected at baseline and endpoint. k -means cluster analysis of baseline serum metabolic health markers identified MPs. Metabolomic profiles, dietary intakes and changes in metabolic health over time were compared between MPs using partial least squares discriminant analysis (PLS-DA) and analysis of covariance (ANCOVA). Results Three distinct MPs were identified with significantly different metabolomic profiles despite comparable baseline dietary intakes. MP1 was categorised as the most metabolically unhealthy, MP2 was mostly female with the lowest anthropometry and high total cholesterol, and MP3 was the youngest with the healthiest blood lipid concentrations. There were significant time*group interactions for total cholesterol (p=0.001), HDL cholesterol (p<0.001), triglycerides (p=0.002) and C-reactive protein (p=0.043) between MPs. Significant decreases in total cholesterol within MP1 (5.92±0.88mmol/L to 5.74±0.83mmol/L, p=0.006) and MP2 (5.81±0.86mmol/L to 5.64±0.94mmol/L, p=0.018) were observed, but no significant difference was observed in MP3 (p=0.146). Increased intakes of fruit and vegetables and whole grains were associated with decreased total cholesterol in MP1 and MP2 (all p<0.05). Conclusion Metabolic phenotyping shows inter-individual variation in metabolic response to dietary change. Metabolic phenotyping could be a valuable tool for informing personalised nutrition strategies to improve diet and metabolic health. Clinical Trial Registry Number Registered with ClinicalTrials.gov on February 21, 2022, Registration ID: NCT05253547 (https://clinicaltrials.gov/study/NCT05253547)