Morbid liver manifestations are intrinsically bound to metabolic syndrome and nutrient intake based on a machine-learning cluster analysis.

Mico, VictorSan-Cristobal, RodrigoMartin, RobertoMartinez-Gonzalez, Miguel AngelSalas-Salvado, JordiCorella, DoloresFito, MontserratAlonso-Gomez, Angel MWärnberg, JuliaVioque, JesusRomaguera, DoraLopez-Miranda, JoseEstruch, RamonTinahones, Francisco JLapetra, JoseSerra-Majem, J LuisBueno-Cavanillas, AuroraTur, Josep AMartin Sanchez, VicentePinto, XavierDelgado-Rodriguez, MiguelMatia-Martin, PilarVidal, JosepVazquez, ClotildeGarcia-Arellano, AnaPertusa-Martinez, SalvadorChaplin, AliceGarcia-Rios, AntonioMuñoz Bravo, CarlosSchröder, HelmutBabio, NancySorli, Jose VGonzalez, Jose IMartinez-Urbistondo, DiegoToledo, EstefaniaBullón, VanessaRuiz-Canela, MiguelPortillo, Maria Puy-Macias-Gonzalez, ManuelPerez-Diaz-Del-Campo, NuriaGarcia-Gavilan, JesusDaimiel, LidiaMartinez, J Alfredo2023-05-032023-05-032022-07-26Micó V, San-Cristobal R, Martín R, Martínez-González MÁ, Salas-Salvadó J, Corella D, et al. Morbid liver manifestations are intrinsically bound to metabolic syndrome and nutrient intake based on a machine-learning cluster analysis. Front Endocrinol (Lausanne). 2022 Sep 6;13:9369561664-2392http://hdl.handle.net/10668/20567Metabolic syndrome (MetS) is one of the most important medical problems around the world. Identification of patient´s singular characteristic could help to reduce the clinical impact and facilitate individualized management. This study aimed to categorize MetS patients using phenotypical and clinical variables habitually collected during health check-ups of individuals considered to have high cardiovascular risk. The selected markers to categorize MetS participants included anthropometric variables as well as clinical data, biochemical parameters and prescribed pharmacological treatment. An exploratory factor analysis was carried out with a subsequent hierarchical cluster analysis using the z-scores from factor analysis. The first step identified three different factors. The first was determined by hypercholesterolemia and associated treatments, the second factor exhibited glycemic disorders and accompanying treatments and the third factor was characterized by hepatic enzymes. Subsequently four clusters of patients were identified, where cluster 1 was characterized by glucose disorders and treatments, cluster 2 presented mild MetS, cluster 3 presented exacerbated levels of hepatic enzymes and cluster 4 highlighted cholesterol and its associated treatments Interestingly, the liver status related cluster was characterized by higher protein consumption and cluster 4 with low polyunsaturated fatty acid intake. This research emphasized the potential clinical relevance of hepatic impairments in addition to MetS traditional characterization for precision and personalized management of MetS patients.enAttribution 4.0 Internationalhttp://creativecommons.org/licenses/by/4.0/BiomarkersClusterDyslipidemiaGlucose disordersHepatic enzymesMetabolic syndromeBlood glucoseCholesterolCluster analysisDietary proteinsEatingFatty acids, unsaturatedHumansLiverMachine learningMetabolic syndromeMorbid liver manifestations are intrinsically bound to metabolic syndrome and nutrient intake based on a machine-learning cluster analysis.research article36147576open accessAnálisis por conglomeradosAprendizaje automáticoColesterolGlucemiaHígadoIngestión de alimentosProteínas en la dietaSíndrome metabólicoÁcidos grasos insaturados10.3389/fendo.2022.936956PMC9487178https://www.frontiersin.org/articles/10.3389/fendo.2022.936956/pdfhttps://www.ncbi.nlm.nih.gov/pmc/articles/PMC9487178/pdf