Use of a combined transient elastography and biochemical strategy to determine liver fibrosis in pediatric intestinal failure.

Background & Aims: Non-invasive monitoring of intestinal failure (IF) associated liver disease is an ongoing challenge in children with IF. Our objective was to develop a combined algorithm of clinical, transient elastography (TE) and biochemical parameters to identify liver fibrosis in this population.
Methods: A retrospective cohort study of IF patients followed by our intestinal rehabilitation program between November 2015 to October 2019. Patients with a liver biopsy and TE were included. Demographic and liver function tests were collected. Fibrosis on liver biopsies was graded using the modified Scheuer score. Decision tree based algorithms classified low (F0-F1) versus high (F2-F4) fibrosis scores based on a combination of TE, biochemical and demographic parameters, using 6-fold classification error, sensitivity and specificity cross-validation (CV) scores.
Results: 42 patients (74% male, median age 7.6 (4.6; 42.7) months) were evaluated. Median length of PN therapy was 182 (121; 556) days. High fibrosis was present in 40.5% with a median TE of 12.1 (6.7; 12.9) kPa in high fibrosis children. An algorithm, based on cut-off values for TE of 11.3 kPa and AST of 40 U/L, and grouping of the underlying etiology resulted in a correct classification of 88.1% of the pathology scores; with sensitivity 0.82 (95% CI 0.57; 0.96), specificity 0.92 (95% CI 0.74; 0.99), positive predictive value 0.88 (95% CI 0.64; 0.96) and negative predictive value 0.88 (95% CI 0.73; 0.96). The CV classification error was 28.6%, CV sensitivity 72.2% and CV specificity 75.5%.
Conclusions: This algorithm shows promising results that could simplify non-invasive monitoring of liver fibrosis in children with IF. Validation in additional IF cohorts is needed.
Competing Interests: Conflict of Interest PW has received research funding support from Takeda, VectivBio and Baxter Pharmaceuticals, YA has received not restricted research funding and consultant fees from Takeda, Zealand Pharma and IPSEN, the other authors have no conflict of interest relevant to this study.
(Copyright © 2022 Elsevier Ltd and European Society for Clinical Nutrition and Metabolism. All rights reserved.)