Predicting weight loss success on a new Nordic diet: an untargeted multi-platform metabolomics and machine learning approach

• DOI: 10.3389/fnut.2023.1191944
• Publication Date: 2023-08-01T09:39:51Z
• Authors: Kristina Pigsborg; Valdemar Stentoft...; Samuel Demharter; Mona Adnan Aldubayan; Alessia Trimigno; Bekzod Khakimov; Søren Balling Eng...; Arne Astrup; Mads Fiil Hjorth; Lars Ove Dragsted; Faidon Magkos
• Keywords: Weight change

Results from randomized controlled trials indicate that no single diet performs better than other for all people living with obesity. Regardless of the plan, there is always large inter-individual variability in weight changes, some individuals losing and not or even gaining weight. This raises possibility that, different individuals, optimal successful loss may differ. The current study utilized machine learning to build a predictive model subjects overweight obesity on New Nordic Diet (NND).Ninety-one consumed an NND ad libitum 26 weeks. Based their loss, were classified as responders (weight ≥5%, n = 46) non-responders <2%, 24). We used clinical baseline data combined urine plasma untargeted metabolomics two analytical platforms, resulting set including 2,766 features, employed symbolic regression (QLattice) develop success.There differences parameters at between non-responders, except age (47 ± 13 vs. 39 11 years, respectively, p 0.009). final contained adipic acid argininic (both metabolites found lower levels responders) generalized training (AUC 0.88) test 0.81). Responders also able maintain 4.3% 12 month follow-up period.We identified containing predict likelihood achieving clinically significant NND. work demonstrates models based multi-platform approach can be optimize precision dietary treatment