Abstract Disclosure: K. Bennett: None. T. Garner: None. D. Chantzichristos: Speaker; Self; Sanofi, Takeda. P. Jansson: None. A. Stevens: None. G. Johannsson: Consulting Fee; Self; Novo Nordisk, Shire, AstraZeneca. Speaker; Self; Novo Nordisk, Pfizer, Inc.. Introduction: Treatment of Adrenal Insufficiency (AI) by glucocorticoid (GC) replacement therapy is mainly guided by clinical judgement, as there is no biomarker of GC action to individualise its dosage. Hypergraphs are networks modelling higher order associations (HOAs), which are interactions between more than two ‘omic elements. The aim of this study was to use hypergraphs to identify serum markers that reflect the action of GCs in adipose tissue (AT). Methods: In a randomised, cross-over trial, 10 patients with primary AI received saline i.v. (GC withdrawal), or circadian infusion of hydrocortisone (GC exposure), with > 2 weeks apart. After 26 hours (8AM), AT biopsy and PBMCs were collected for transcriptomic analysis, and abdominal subcutaneous microdialysis and PBMCs were collected for metabolomic profiling using LC/MS. Hypergraph models were generated to integrate serum and AT transcriptomics and metabolomics, identifying clusters of putatively causally linked metabolites which could represent serum markers of downstream GC action in AT. Hypergraph structure was assessed in AT and serum using entropy to infer associations within pre-defined metabolic pathways. Differences in entropy between the interventions was measured using Bayes Factor. Univariate ROC curve analysis was used to calculate AUCs for the serum markers. Results: In AT dialysate, 22 metabolites were identified as significantly different between GC exposure and withdrawal states. These included metabolites found in lipolysis and proteolysis pathways, such as acyl carnitines, lysophosphatidylcholines, fatty acids, and peptide derivatives. Nine metabolites were significantly different between the interventions in the serum. DEGs in AT (n=2048) were found to be enriched in metabolic pathways. Hypergraphs were used to refine clusters in the serum based on their HOAs with AT differentially expressed genes and metabolites. Tryptophan, Kynurenine, and Tyrosine clustered together, demonstrating a relationship between their expression in serum, and the metabolomic and transcriptomic profile of AT. These HOAs were shared with cortisol in the exposure intervention, but not in GC withdrawal, suggesting GC dependent regulation. All 3 metabolites had AUC-ROCs > 0.8, with Kynurenine performing the best at 0.86. Tissue specific differences were seen in the entropy of metabolic pathways between GC exposure and withdrawal, with similar changes seen in lipid and proteolysis pathways, but no difference seen in serum in carbohydrate and amino acid biosynthesis. Conclusions: During near physiological exposure, we identified changes to hypergraph structure and metabolite concentrations in lipid, protein, and carbohydrate pathways in AT and serum, compared to GC withdrawal. Hypergraph integration of multi-omic data identified tryptophan, kynurenine, and tyrosine as potential serum markers reflecting GC action in AT. Presentation: 6/3/2024