P667 Safety of anti-TNF treatment in liver transplant recipients – a meta-analysis

Background: Primary Sclerosing Cholangitis (PSC) patients with refractory inflammatory bowel disease (IBD) after liver transplantation (LT) pose a dilemma for treating physicians, as little is known about the risk of serious infection when combining anti-TNF therapy with immunosuppression for prevention of rejection. Our aim was to investigate the infection risk in this patient group by systematic review and meta-analysis of the available data. Methods: A literature search was conducted for full papers and conference proceedings through September 2015 regarding liver transplant recipients and anti-TNF therapy. All studies were appraised using the adapted Newcastle-Ottawa Scale (NOS), which contains 9 criteria for cohort studies and is adapted to 6 criteria for case series and case reports. Two reviewers (MWvM and PWJM) independently extracted study and control-patient data (age, duration of follow up, number of all infections, number of serious infections, time since transplant). As additional control population, PSC-IBD patients from the LUMC LT cohort were used. Poisson regression was used to compare serious infections (according to ICH-definition) per patient year follow up between the anti-TNF and control group, correcting for mean time since transplant. Results: Initially, 465 articles and abstracts were identified, of which 8 were included. These 8 studies contained 53 post-LT patients on anti-TNF therapy and 23 post-LT control patients not on anti-TNF therapy. None of the studies scored less than 75% of the NOS quality criteria. From the LUMC LT cohort, 41 PSC-patients with PSC-IBD but without anti-TNF therapy were included as control population. Serious infection rates differed from 0 to 0.38 serious infections per patient year in the anti-TNF therapy group, and 0.04 to 0.24 in the control group. The overall infection rate for TNF-exposed patients was 0.12, compared to 0.15 in the control patients, resulting in a rate ratio of 0.80 (95% CI: 0.17–3.97, p=0.80). Age at time of transplant was not associated with the rate ratio for serious infections, whereas the time since transplantation was. Although correcting for time since transplant causes the infection rates in the anti-TNF-group to be higher than in the control group (0.13 vs 0.12 serious infections per patient year) the rate ratio remained non-significant (1.1, p=0.82). Conclusions: No significant increase in the serious infection rate was observed in LT-recipients with PSC-IBD during exposure to anti-TNF therapy. However, the wide confidence intervals of these results show that more data is needed to provide a definitive conclusion on the safety of anti-TNF therapy in these patients.