Plant-Based Meals Generate Lower Ammonia and Have a Unique Metabolomic Signature Compared to Meat-Based Meals Despite Similar Baseline Microbiome in Patients With Cirrhosis: A Randomized Clinical Trial

Introduction: Interactions between the gut microbiome and diet could affect ammoniagenesis in cirrhosis & hepatic encephalopathy (HE) but the impact of dietary preferences on metabolomics in cirrhosis & HE is unclear. As most Western populations follow meat-based diets, we aimed to determine the impact of substituting a single meat-based meal with an equicaloric & equal protein-containing plant-based (vegan/vegetarian) alternative on ammonia & metabolomics in outpatients with cirrhosis on a meat-based diet. Methods: Outpatients with cirrhosis with/without prior HE on a stable Western meat-based diet were randomized 1:1:1 into 3 groups (Figure 1A). Patients were given an equicaloric burger with 20g protein (meat substitute for vegan [V], bean for vegetarian [VG], and pork/beef patty for meat [M]). Blood for metabolomics (liquid chromatography-mass spectrometry) & ammonia was drawn at baseline & hourly for 3 hours post-meal while patients were fasting under observation. Baseline stool microbiome characteristics, ammonia, & metabolite levels were compared among groups. Results: We enrolled 30 men (10 per group), and 5 per group had a prior history of HE. There were no baseline differences between these groups (Table 1A). Stool microbiome: Similar across groups without changes in α/β-diversity or individual bacteria (Figure 1B) Ammonia: Increased above baseline in M but not V or VG (Figure 1C) Metabolomics: Significant differences were found in >3000 metabolites of which 106 metabolites between M & V, 71 between M & VG, and 45 between V & VG were significant on repeated measures analysis of variance. The M group had higher phospholipid/ceramide, methionine, & 5-hydroxylysine levels than the V group. The M group had higher long-chain fatty acid, sphingolipid, acylcarnitine metabolites, & phosphatidylcholine levels than the VG group, but lower lysophosphatidylcholine than V or VG. The V group had higher N-acyl amino acid levels than M or VG. (Table 1B). Conclusion: Substitution of just 1 meat-based meal with plant-based alternatives may improve metabolite signatures associated with HE in outpatients with cirrhosis, despite having similar microbiome profiles at baseline. High methionine, high PC, and low LPC have been associated with HE, and high lysine with hyperammonemia. N-acyl amino acids increased in the non-meat meals, may have anti-inflammatory/neuroprotective effects and are downregulated in HE. Plant-based meals could be considered in patients with cirrhosis who usually follow a Western meat-based diet.Figure 1.: A. Randomization Flowchart. B. Baseline stool microbiome overlap among groups. C. Serum ammonia levels over time for each group. Table 1. - A) Baseline Patient Characteristics - no significant differences among groups. B) Analysis Results - classes of metabolites that differed between meat/vegetarian and meat/vegan groups 1A. Baseline Characteristic Meat (n=10) Vegetarian (n=10) Vegan (n=10) P-Value Age 66.2±2.7 66.1±7.8 70.5±3.1 0.15 Male Sex 10 10 10 1.0 MELD-Na 10.1±4.5 9.1±3.0 8.9±2.9 0.46 Etiology (HCV/Alcohol/ NASH/Others) 4/4/2/0 2/5/3/0 2/5/3/0 0.8 Prior HE 5 5 5 1.0 Lactulose 5 5 5 1.0 Rifaximin 4 4 4 0.59 Proton pump inhibitor 7 8 7 0.94 Prior ascites 5 4 4 0.81 1B. Metabolomic Pathway Metabolite (Meat/Vegetarian Fold Change) Metabolite (Meat/Vegan Fold Change) Higher in Meat vs Other Alpha amino acids 5-hydroxylysine (3.17), thioproline (2.67) Methionine (1.48), 5-hydroxylysine (3.08) Lipids: phosphatidyl-ethanolamines 1-(1-enyl-palmitoyl)-2-oleoyl-glycerophosphoethanolamine (1.7) 1-(1-enyl-palmitoyl)-2-oleoyl-glycerophosphoethanolamine (1.64), 1-(1-enyl-palmitoyl)-2-linoleoyl-glycerophosphoethanolamine (1.7), 1-(1-enyl-stearoyl)-2-oleoyl-glycerophosphoethanolamine (1.85), 1-(1-enyl-stearoyl)-2-linoleoyl-glycerophosphoethanolamine (1.85), 1-(1-enyl-stearoyl)-2-docosapentaenoyl-glycerophosphoethanolamine (1.8) Lipids: glyco-sphingolipid 1-glycosyl-N-(2-hydroxynervonoyl)-sphingosine (1.98) Glycosyl-N-palmitoyl-sphingosine (1.99), glycosyl-N-behenoyl-sphingosine (1.74), glycosyl-N-nervonoyl-sphingosine (2.07), glycosyl-N-(2-hydroxynervonoyl)-sphingosine (2.07), glycosyl ceramide (1.7), glycosyl ceramide (2.82), glycosyl ceramide (2.37) Lipids: sphingolipid Palmitoyl-sphingosine-phosphoethanolamine (1.35), glycosyl-N-(2-hydroxynervonoyl)-sphingosine (1.98) Palmitoyl sphingomyelin (1.29), palmitoyl-sphingosine-phosphoethanolamine (1.53), sphingadienine (2.65), N-palmitoyl-sphingosine (1.37) Lipids: phosphatidyl-choline 1-palmitoyl-2-stearoyl-glycerophosphocholine (1.46), 1-margaroyl-2-linoleoyl-glycerophosphocholine (1.4), 1-stearoyl-2-oleoyl-glycerophosphocholine (1.42), 1-(1-enyl-palmitoyl)-2-oleoyl-glycerophosphocholine (1.63), 1-palmityl-2-stearoyl-glycerophosphocholine (1.61), 1-margaroyl-2-oleoyl-glycerophosphocholine (1.47), 1-nonadecenoyl-glycerophosphocholine (1.96) 1-palmityl-2-stearoyl-glycerophosphocholine (1.64) Lipids: long chain fatty acids Pentadecanoate (1.7), nonadecanoate (1.64), octadecanedioate (2.88), eicosanedioate (1.88), 2-hydroxypalmitate (1.59), 2-hydroxystearate (1.37) Gutarate (2.47) Lowerin Meat vs Other N-acyl amino acids N-acetylisoleucine (0.56), 2-oxoarginine (0.54), N-delta-acetylornithine (0.4) N-formylanthranilic acid (0.59), N-acetylleucine (0.43), N-acetylisoleucine (0.56), tigloylglycine (0.54), propionylglycine (0.43)