Association of metabolomic markers and response to nutritional support: A secondary analysis of the EFFORT trial using an untargeted metabolomics approach. Issue 9 (September 2021)

• Record Type: Journal Article
• Authors: Struja, Tristan; Wolski, Witold; Schapbach, Ralph; Mueller, Beat; Laczko, Endre; Schuetz, Philipp
• Is Part Of: Clinical nutrition. Volume 40:Issue 9(2021)
• Page Start: 5062
• Page End: 5070
• Publication Date: 2021-09
• Subjects: Metabolomics -- EFFORT -- Risk prediction -- Nutritional risk score; Critically ill -- Nutrition -- Periodicals; Diet therapy -- Periodicals; Parenteral feeding -- Periodicals; Enteral feeding -- Periodicals; Enteral Nutrition -- Periodicals; Parenteral Nutrition -- Periodicals; Metabolism -- Periodicals; Diétothérapie -- Périodiques; Alimentation parentérale -- Périodiques; Alimentation entérale -- Périodiques; Nutrition -- Périodiques; Diet therapy; Enteral feeding; Nutrition; Parenteral feeding; Electronic journals; Periodicals; Electronic journals; 615.854
• DOI: 10.1016/j.clnu.2021.07.029
• Languages: English
• ISSNs: 0261-5614

Abstract

Summary: Background & aims: The EFFORT trial reported a substantial risk reduction for adverse events and mortality in medical in-patients receiving a nutritional support intervention. With the use of an untargeted metabolomics approach, we investigated the prognostic and therapeutic potential of metabolomic markers to understand, whether there are distinct metabolic patterns associated with malnutrition risk as assessed by the Nutritional Risk screening (NRS 2002) score, the risk of 30-day mortality and the response to nutritional support, respectively. Methods: Out of the 2088 samples we randomly selected 120 blood samples drawn on day 1 after hospital admission and before treatment initiation. Samples were stratified by NRS 2002, treatment allocation (intervention vs. control), and mortality at 30 days, but not on the type of medical illness. We performed untargeted analysis by liquid chromatography mass spectrometry (LC-MS/MS). Results: We measured 1389 metabolites in 120 patients of which 81 (67.5%) survived until day 30. After filtering, 371 metabolites remained, and 200 were matched to one or more Human Metabolome Data Base (HMDB) entries. Between group analysis showed a slight distinction between the treatment groups for patients with a NRS 3, but not for those with NRS 4 and ≥ 5. C-statistic between those who died and survived at day 30 ranged from 0.49 (95% confidence interval 0.35–0.68) for a combination of 5 metabolites/predictors to 0.66 (95% confidence interval 0.53–0.79) for a combination of 100 metabolites. Pathway analysis found significant enrichment in the pathways for nitrogen, vitamin B3 (nicotinate and nicotinamide), leukotriene, and arachidonic acid metabolisms in nutritional support responders compared to non-responders. Conclusion: In our heterogenous population of medical inpatients with different illnesses and comorbidities, metabolomic markers showed little prognostic and therapeutic potential for better phenotyping malnutrition and response to nutritional therapy. Future studies should focus on more selected patient populations to understand whether a metabolomic approach can advance the nutritional care of patients.