P435 Liver injury associated with anti-tumour necrosis factor therapy in paediatric irritable bowel disease patients
A. Ricciuto*, B. M. Kamath, T. Walters, P. Church, S. C. Ling, K. Frost, A. Griffiths
Hospital for Sick Children, Division of Gastroenterology, Hepatology and Nutrition, Toronto, Canada
The frequency of hepatotoxicity in children treated for irritable bowel disease (IBD) with anti-tumour necrosis factor (TNF) α has never been examined. The aim of this study is to characterise liver enzyme abnormalities and drug-induced liver injury (DILI) in a large cohort of paediatric IBD patients on anti-TNF α.
All IBD patients treated with anti-TNF α at SickKids Hospital (2000–2015) were linked electronically with laboratory data to identify ALT elevations (≥ 80 U/L). Patient characteristics and natural history of liver enzyme elevations were recorded; likelihood of causality was assessed using the Roussel Uclaf Causality Assessment Method.
Of 672 paediatric IBD patients treated with anti-TNF α, 65 (10%) developed 67 episodes of ALT elevation ≥ 80 U/L (25% female, 89% Crohn’s disease [CD], 85% infliximab [IFX]/15% adalimumab). Median peak ALT was 117 U/L. The median time to ALT elevation was 5 weeks and the median duration of ALT elevation was 6.2 weeks. ALT normalised in 85% of cases. 40/65 received a concomitant immunomodulator. Further, 15/672 patients (2%) had persistent ALT elevation (≥ 80 days) and at least a ‘possible’ causal relationship with anti-TNF therapy. All had CD, and 12/15 received IFX. In all patients, the pattern of injury was hepatocellular. Moreover, 10/14 had ≥ 1 positive autoantibody, and 4/15 underwent liver biopsy. The first, with peak ALT 401, met criteria for ‘definite’ autoimmune hepatitis (AIH). IFX cessation was followed by liver enzyme normalisation over 15 weeks. Causality in the second patient, with peak ALT 205 and GGT 102, is unclear. Focal periductal fibrosis and portal inflammation were seen on biopsy, and mildly dilated ducts on imaging. However, liver enzymes normalised after IFX discontinuation and rose upon re-exposure, and ANA titre fluctuated with IFX exposure, suggesting the changes may have been drug-induced. Anti-TNF therapy was discontinued in 2 additional patients, the first due to elevated transaminases and concurrent pancytopenia, with high titre ANA, and the second because of poor clinical response. ALT normalised after IFX discontinuation in both. The 2 other patients to undergo biopsy displayed non-specific findings (portal inflammation, pericholangitis, cholangiolar proliferation, and focal periductal fibrosis), with normal cholangiograms. Both have remained on anti-TNF α and failed to normalise ALT. Of the remaining patients who continued on anti-TNF α, ALT remains elevated in 5/9 after follow-up of 36.7 to 78 weeks.
Significant ALT elevations occurred in 10% of children receiving anti-TNF therapy for IBD. The majority were transient but a small subset experienced marked and persistent elevations. Triggering of immune-mediated phenomena, such as AIH, can occur, in which case prompt anti-TNF α cessation is indicated.