Post-Hemodialysis Flow-Dependent Hepatic Function Impairment in Individuals With End Stage Kidney Disease and Chronic Inflammation.

Introduction: The liver plays an important role to prevent translocation of gut-derived toxins from the portal to the systemic circulation. Chronic inflammation is common in patients receiving hemodialysis, and increased gut permeability to microbial material has been implicated in its pathogenesis. This study sought to establish if flow-dependent hepatic function was impaired in chronically inflamed individuals treated with hemodialysis.

Methods: Fifty adults receiving outpatient hemodialysis were recruited. Subjects with known liver or gastrointestinal disease, acute inflammation, and hemodynamic instability during hemodialysis were excluded. Participants were divided into two groups (n = 25): individuals with chronic inflammation (defined as a median high-sensitivity C-reactive protein (hs-CRP) ≥ 5 mg/dL over the preceding 3 months) with no apparent cause and a noninflamed group. Flow-dependent hepatic function (defined as a composite of hepatic perfusion, hepatocyte clearance and biliary excretion) was assessed following hemodialysis by indocyanine green clearance to derive: (1) indocyanine green-plasma disappearance rate and (2) indocyanine green-retention after 15 min. Serum beta-D-glucan levels pre- and post-hemodialysis were measured as surrogate markers of gastrointestinal permeability.

Findings: Indocyanine green-plasma disappearance rate was reduced in the inflamed group versus the noninflamed group (19.4 (8.7)%/min vs. 23.8 (14.4)%/min; p = 0.02). Indocyanine green-retention after 15 min was higher in the inflamed group (5.4 (6.8)% vs. 2.9 (5.0)%; p = 0.02). Noninvasive hepatic fibrosis and steatosis assessments were similar in both groups. Pre-hemodialysis beta-D-glucan levels were similar (63 (42) pg/ml vs. 49 (11) pg/ml; p = 0.13), whereas post-hemodialysis beta-D-glucan levels were higher in the inflamed group (82 (48) pg/ml vs. 58 (27) pg/ml; p < 0.001), and in those with flow-dependent hepatic impairment (72 (45) vs. 55 (32) pg/ml; p = 0.004). In linear regression analysis, indocyanine green-retention after 15 min and post-hemodialysis beta-D-glucan levels were independent predictors of median hs-CRP, explaining 21% of the variation.

Discussion: Individuals with otherwise unexplained inflammation had impaired hepatic function post-hemodialysis and higher post-hemodialysis beta-D-glucan levels. These findings are compatible with the notion that impaired hepatic gut-derived toxin removal propagates chronic inflammation in hemodialysis.