To Örebro University

BACKGROUND: Studies have shown an association between use of tumour necrosis factor-α (TNFα) inhibitors and increased risk of serious infection in adult inflammatory bowel disease (IBD). However, data on this topic for paediatric patients are scarce and inconclusive. The aim of this study was to investigate whether there is an association between the use of TNFα inhibitors and the risk of serious infection in children with IBD.

METHODS: In this nationwide Danish cohort study, we searched health registers (from Jan 1, 2007, to Dec 31, 2016) to identify episodes of children and adolescents (<18 years) with at least two recorded IBD diagnoses in specialist care. We categorised follow-up time in mutually exclusive episodes of incident TNFα inhibitor use or no TNFα inhibitor use from specialist care records. We used Cox proportional hazards models to estimate hazard ratios (HRs), adjusting using propensity score weighting for demographic characteristics, comorbidities, treatment history, health-care use, and indicators of disease severity. The primary outcome, incident serious infection, was defined as infection requiring a stay in hospital and was identified through hospital records.

FINDINGS: Among 2817 paediatric patients with IBD, we identified 618 episodes of incident TNFα inhibitor use and 2925 episodes of no TNFα inhibitor use. In the cohort of exposed and not exposed episodes that was propensity-score weighted, 53·9% were of male sex, the mean age was 15·1 (SD 1·7) years, 69·9% had Crohn's disease, and 30·1% had ulcerative colitis or IBD-unclassified; median follow-up was 1·4 years (IQR 0·4-3·0). The weighted incidence of serious infection was 54·6 events per 1000 patient-years for the TNFα inhibitor episodes and 61·9 events per 1000 patient-years for the no-use episodes. The weighted HR of serious infection associated with TNFα inhibitor use was 0·81 (95% CI 0·54-1·21).

INTERPRETATION: There was no significant association between use of TNFα inhibitors and the risk of serious infection in children with IBD, and, based on the upper bound of the confidence interval, a relatively small risk increase seems unlikely, contrary to previous findings in adults. Observational data such as these can support paediatric clinical practice.