Assessment of Intestinal Failure Associated Liver Disease according to different diagnostic criteria

Background & aims: Intestinal failure associated liver disease (IFALD) has been defined using numerous criteria; however the clinical relevance of these criteria has never been compared. We therefore aimed to evaluate the prevalence, incidence, evolution of IFALD diagnosed by different criteria and to assess any clinical features that may be associated with its occurrence. Methods: A cross sectional (CS) and retrospective study were carried out on adults on home parenteral nutrition (HPN) for chronic intestinal failure (CIF) managed at a single center. Inclusion criteria at CS: age ≥18 years, benign disease. Collected data included: patient demographics, CIF and HPN characteristics, episodes of central venous catheter related bloodstream infection (CRBSI). IFALD was diagnosed by 9 criteria based on liver function tests and liver ultrasound (US) imaging. IFALD diagnoses were categorized as steatosis (2 criteria), cholestasis (3 criteria) or fibrosis (2 criteria) and unclassified (2 criteria). Prevalence was assessed at CS and at starting HPN (baseline, BS). Evolution was assessed as change of IFALD between BS and CS. Incidence was calculated as patients who developed IFALD from BS to CS. Results: A total of 113 patients were included. At CS, IFALD prevalence range in each diagnostic categories was: cholestasis 5–15%; steatosis 17–43%; fibrosis 10–20%; unclassified 7–38%. A 28.5% of patients did not have IFALD according to any criteria. Two cholestasis criteria and one fibrosis criterion were significantly (P < 0.05) associated with a short bowel syndrome as the pathophysiological mechanism of CIF, HPN requirement and the number of CRBSI episodes. At BS, IFALD prevalence range was: cholestasis 13–40%; steatosis 27–90%; fibrosis 2–5%; unclassified 8–75%. The incidence range of IFALD was: cholestasis 0–7%; steatosis 0–39%; fibrosis 7–18%; unclassified 4–9%. IFALD steatosis diagnosed by US was the most frequent diagnosis at both CS prevalence and incidence assessments. Notably, IFALD criteria normalized in various percentages (2–70%), depending on the diagnostic categories, between BS and CS. Conclusions: This is the first study to systematically demonstrate that the frequency of IFALD varies greatly depending on diagnostic criteria used, confirming the need for a consensus definition to be used between different national and international IF units. IFALD can be present at HPN initiation but may resolve thereafter; further work is required to evaluate the factors associated with improvement.