Multi-omic predictors of BMI trajectories in adults with MASLD undergoing personalized dietary treatment

IF 0.3 Q4 NUTRITION & DIETETICS

CAHIERS DE NUTRITION ET DE DIETETIQUE Pub Date : 2026-08-01 Epub Date: 2026-09-24 DOI:10.1016/j.cnd.2026.08.081

Carolynne Martins Teixeira, Itziar Abete, Fermin I. Milagro, José I. Herrero, Mariana Elorz, Josep A. Tur, Jose Alfredo Martínez, Helen Hermana Miranda Hermsdorff, M. Angeles Zulet

{"title":"Multi-omic predictors of BMI trajectories in adults with MASLD undergoing personalized dietary treatment","authors":"Carolynne Martins Teixeira, Itziar Abete, Fermin I. Milagro, José I. Herrero, Mariana Elorz, Josep A. Tur, Jose Alfredo Martínez, Helen Hermana Miranda Hermsdorff, M. Angeles Zulet","doi":"10.1016/j.cnd.2026.08.081","DOIUrl":null,"url":null,"abstract":"<div><h3>Objective</h3><div>To identify Body Mass Index (BMI) trajectories over 24 months and evaluate the ability of multi-omic biomarkers to discriminate individuals according to dietary intervention response patterns.</div></div><div>To identify Body Mass Index (BMI) trajectories over 24 months and evaluate the ability of multi-omic biomarkers to discriminate individuals according to dietary intervention response patterns.</div><div><h3>Methods</h3><div>Data were obtained from the Fatty Liver in Obesity (FLiO) randomized controlled trial, involving 98 adults with overweight/obesity and diagnosed with Metabolic Dysfunction-Associated Steatotic Liver Disease (MASLD). Participants underwent a 30% caloric restriction and were randomized to one of two dietary interventions (FLiO or American Heart Association [AHA] diet). BMI trajectories over 24 months were identified using Latent Class Growth Modeling (LCGM) to classify participants according to response patterns. A Polygenic Risk Score (PRS) was calculated based on 35 SNPs across 11 candidate genes by summing risk alleles. Gut microbiota composition was analyzed by shotgun sequencing. Genera significantly associated with BMI trajectories in logistic regression analyses were selected as predictors for microbiota-based models. Predictive performance was assessed using the Area Under the Receiver Operating Characteristic Curve (AUC), comparing favorable versus unfavorable trajectory patterns. The FLiO study was approved by the Ethics Committee of the University of Navarra (54/2015) and registered with Clinical Trials (FLiO: Fatty Liver on Obesity study; <span><span>NCT03183193</span><svg><path></path></svg></span>).</div></div><div>Data were obtained from the Fatty Liver in Obesity (FLiO) randomized controlled trial, involving 98 adults with overweight/obesity and diagnosed with Metabolic Dysfunction-Associated Steatotic Liver Disease (MASLD). Participants underwent a 30% caloric restriction and were randomized to one of two dietary interventions (FLiO or American Heart Association [AHA] diet). BMI trajectories over 24 months were identified using Latent Class Growth Modeling (LCGM) to classify participants according to response patterns. A Polygenic Risk Score (PRS) was calculated based on 35 SNPs across 11 candidate genes by summing risk alleles. Gut microbiota composition was analyzed by shotgun sequencing. Genera significantly associated with BMI trajectories in logistic regression analyses were selected as predictors for microbiota-based models. Predictive performance was assessed using the Area Under the Receiver Operating Characteristic Curve (AUC), comparing favorable versus unfavorable trajectory patterns. The FLiO study was approved by the Ethics Committee of the University of Navarra (54/2015) and registered with Clinical Trials (FLiO: Fatty Liver on Obesity study; <span><span>NCT03183193</span><svg><path></path></svg></span>).</div><div><h3>Results</h3><div>LCGM identified three distinct BMI trajectory patterns. Two trajectories were characterized by a marked BMI reduction at 6 months followed by better long-term weight maintenance, whereas the third one showed a less favorable response. The mean PRS was 22.25±7.25 points. At the taxonomic level, the genera Faecalibacterium, Blautia, and Bilophila were significantly associated with trajectory patterns. The microbiota-based model showed strong discriminative ability (AUC: 0.89; 95%CI: 0.70–1.00), outperforming the isolated PRS model (AUC: 0.50; 95%CI: 0.20–0.82). The multi-omic model achieved similar performance (AUC: 0.88; 95%CI: 0.72–1.00), highlighting the predominant role of modifiable biological factors over genetic susceptibility in predicting clinical response.</div></div><div>LCGM identified three distinct BMI trajectory patterns. Two trajectories were characterized by a marked BMI reduction at 6 months followed by better long-term weight maintenance, whereas the third one showed a less favorable response. The mean PRS was 22.25±7.25 points. At the taxonomic level, the genera Faecalibacterium, Blautia, and Bilophila were significantly associated with trajectory patterns. The microbiota-based model showed strong discriminative ability (AUC: 0.89; 95%CI: 0.70–1.00), outperforming the isolated PRS model (AUC: 0.50; 95%CI: 0.20–0.82). The multi-omic model achieved similar performance (AUC: 0.88; 95%CI: 0.72–1.00), highlighting the predominant role of modifiable biological factors over genetic susceptibility in predicting clinical response.</div><div><h3>Conclusion</h3><div>Gut microbiota composition emerged as a key predictor of BMI trajectories, suggesting that modifiable biomarkers may provide greater precision for identifying good and poor responders to dietary interventions.</div></div><div>Gut microbiota composition emerged as a key predictor of BMI trajectories, suggesting that modifiable biomarkers may provide greater precision for identifying good and poor responders to dietary interventions.</div>","PeriodicalId":43722,"journal":{"name":"CAHIERS DE NUTRITION ET DE DIETETIQUE","volume":"61 ","pages":"Page e44"},"PeriodicalIF":0.3000,"publicationDate":"2026-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"CAHIERS DE NUTRITION ET DE DIETETIQUE","FirstCategoryId":"1085","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S0007996026001574","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"2026/9/24 0:00:00","PubModel":"Epub","JCR":"Q4","JCRName":"NUTRITION & DIETETICS","Score":null,"Total":0}

引用次数: 0

Abstract

Objective

To identify Body Mass Index (BMI) trajectories over 24 months and evaluate the ability of multi-omic biomarkers to discriminate individuals according to dietary intervention response patterns.

Methods

Data were obtained from the Fatty Liver in Obesity (FLiO) randomized controlled trial, involving 98 adults with overweight/obesity and diagnosed with Metabolic Dysfunction-Associated Steatotic Liver Disease (MASLD). Participants underwent a 30% caloric restriction and were randomized to one of two dietary interventions (FLiO or American Heart Association [AHA] diet). BMI trajectories over 24 months were identified using Latent Class Growth Modeling (LCGM) to classify participants according to response patterns. A Polygenic Risk Score (PRS) was calculated based on 35 SNPs across 11 candidate genes by summing risk alleles. Gut microbiota composition was analyzed by shotgun sequencing. Genera significantly associated with BMI trajectories in logistic regression analyses were selected as predictors for microbiota-based models. Predictive performance was assessed using the Area Under the Receiver Operating Characteristic Curve (AUC), comparing favorable versus unfavorable trajectory patterns. The FLiO study was approved by the Ethics Committee of the University of Navarra (54/2015) and registered with Clinical Trials (FLiO: Fatty Liver on Obesity study; NCT03183193).

Results

LCGM identified three distinct BMI trajectory patterns. Two trajectories were characterized by a marked BMI reduction at 6 months followed by better long-term weight maintenance, whereas the third one showed a less favorable response. The mean PRS was 22.25 ± 7.25 points. At the taxonomic level, the genera Faecalibacterium, Blautia, and Bilophila were significantly associated with trajectory patterns. The microbiota-based model showed strong discriminative ability (AUC: 0.89; 95%CI: 0.70–1.00), outperforming the isolated PRS model (AUC: 0.50; 95%CI: 0.20–0.82). The multi-omic model achieved similar performance (AUC: 0.88; 95%CI: 0.72–1.00), highlighting the predominant role of modifiable biological factors over genetic susceptibility in predicting clinical response.

Conclusion

Gut microbiota composition emerged as a key predictor of BMI trajectories, suggesting that modifiable biomarkers may provide greater precision for identifying good and poor responders to dietary interventions.

查看原文本刊更多论文

接受个性化饮食治疗的成人MASLD患者BMI轨迹的多组学预测因子

目的确定24个月内的体重指数（BMI）轨迹，并评估多组学生物标志物根据饮食干预反应模式区分个体的能力。确定24个月的体重指数（BMI）轨迹，并评估多组学生物标志物根据饮食干预反应模式区分个体的能力。方法数据来自肥胖脂肪肝（FLiO）随机对照试验，涉及98名超重/肥胖并诊断为代谢功能障碍相关脂肪变性肝病（MASLD）的成年人。参与者接受30%的热量限制，并随机分为两组饮食干预（FLiO或美国心脏协会饮食）。使用潜在类别增长模型（LCGM）确定24个月以上的BMI轨迹，根据反应模式对参与者进行分类。多基因风险评分（PRS）基于11个候选基因的35个snp，通过风险等位基因的累加计算得到。采用散弹枪测序法分析肠道菌群组成。在逻辑回归分析中，选择与BMI轨迹显著相关的属作为微生物群模型的预测因子。预测性能通过接受者工作特征曲线下面积（AUC）进行评估，比较有利和不利的轨迹模式。FLiO研究已获得纳瓦拉大学伦理委员会批准（54/2015），并注册临床试验（FLiO: Fatty Liver on Obesity study; NCT03183193）。数据来自肥胖脂肪肝（FLiO）随机对照试验，涉及98名超重/肥胖并诊断为代谢功能障碍相关脂肪变性肝病（MASLD）的成年人。参与者接受30%的热量限制，并随机分为两组饮食干预（FLiO或美国心脏协会饮食）。使用潜在类别增长模型（LCGM）确定24个月以上的BMI轨迹，根据反应模式对参与者进行分类。多基因风险评分（PRS）基于11个候选基因的35个snp，通过风险等位基因的累加计算得到。采用散弹枪测序法分析肠道菌群组成。在逻辑回归分析中，选择与BMI轨迹显著相关的属作为微生物群模型的预测因子。预测性能通过接受者工作特征曲线下面积（AUC）进行评估，比较有利和不利的轨迹模式。FLiO研究已获得纳瓦拉大学伦理委员会批准（54/2015），并注册临床试验（FLiO: Fatty Liver on Obesity study; NCT03183193）。结果slcgm鉴定出三种不同的BMI轨迹模式。两种轨迹的特点是在6个月时BMI显著降低，随后长期体重保持得更好，而第三种轨迹的反应则不太好。平均PRS为22.25±7.25分。在分类水平上，Faecalibacterium、Blautia和Bilophila属与轨迹模式显著相关。基于微生物群的模型具有较强的判别能力（AUC: 0.89; 95%CI: 0.70 ~ 1.00），优于分离的PRS模型（AUC: 0.50; 95%CI: 0.20 ~ 0.82）。多组学模型取得了类似的效果（AUC: 0.88; 95%CI: 0.72-1.00），突出了可改变的生物因素在预测临床反应方面比遗传易感性更重要。LCGM确定了三种不同的BMI轨迹模式。两种轨迹的特点是在6个月时BMI显著降低，随后长期体重保持得更好，而第三种轨迹的反应则不太好。平均PRS为22.25±7.25分。在分类水平上，Faecalibacterium、Blautia和Bilophila属与轨迹模式显著相关。基于微生物群的模型具有较强的判别能力（AUC: 0.89; 95%CI: 0.70 ~ 1.00），优于分离的PRS模型（AUC: 0.50; 95%CI: 0.20 ~ 0.82）。多组学模型取得了类似的效果（AUC: 0.88; 95%CI: 0.72-1.00），突出了可改变的生物因素在预测临床反应方面比遗传易感性更重要。结论肠道菌群组成是BMI轨迹的关键预测指标，表明可修改的生物标志物可能为识别饮食干预的良好和不良反应提供更高的精度。肠道微生物群组成成为BMI轨迹的关键预测指标，这表明可修改的生物标志物可能为识别对饮食干预的良好和不良反应提供更高的精度。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

CAHIERS DE NUTRITION ET DE DIETETIQUE NUTRITION & DIETETICS-

CiteScore

0.60

自引率

33.30%

发文量

审稿时长

65 days

期刊介绍： Les Cahiers de nutrition et de diététique sont le lieu de communication reconnu pour les médecins spécialistes de la nutrition, les diététiciens, les professionnels et les chercheurs de agro-alimentaire, les enseignants et les étudiants. Les Cahiers de nutrition et de diététique abordent les sujets se rapportant à la biologie générale, à la médecine et à la nutrition, aux aliments, aux comportements alimentaires, à la diététique pratique, ainsi que des thèmes variés : activité physique, apports nutritionnels conseillés, dénutrition, alcool, obésité de enfant, aliments à effet santé, plaisir alimentaire, restauration collective, goûters et collations, allergies alimentaires...