M. Harrison1, C. Cummins2, G. Horneff3, T. Southwood2, D. Mines1. 1PFIZER, Collegeville, PA, United States; 2University of Birmingham-Children's Hospital, Birmingham, United Kingdom; 3Asklepios Clinic, Sankt Augustin, Germany

Background: Previous studies have suggested a link between the use of TNF inhibitors and malignancy in adults. Recent reports of cancer in children exposed to these agents raise the question of a similar association in the pediatric population.
Objectives: To compare the risk of incident malignancy in pediatric patients with JIA who were treated with etanercept to that risk in the general population
Methods: This retrospective cohort analysis combined data from 3 prospective JIA biologics registries in the UK, Germany, and US. Analyses included all registry patients who received a first dose of etanercept before age 18 years. All cases of cancer diagnosed prior to age 22 were included (per FDA guidance). Follow-up time began with the first dose of etanercept and ended at the earliest of: age 22, data lock date, date of death, or date of cancer diagnosis, regardless of whether etanercept was discontinued, i.e., once exposed, always at-risk. We estimated relative risk using standardized incidence ratios (SIR) defined as the observed divided by expected number of events. For reference rates, we used country-specific cancer statistics for the general population stratified by age and gender. To calculate the expected number, we applied these rates to the corresponding follow-up time from each registry. A sensitivity analysis excluded the first 90 days after etanercept initiation to reduce the chance of misclassifying prevalent (pre-existing) cancers as incident (de novo).
Results: As of August 2008, 1721 patients were enrolled in the 3 registries. Of these, 1641(95%) qualified for the primary analysis with a mean age at registry entry of 11.5±4.1 years; 1134 (69%) were female. Mean follow-up time was 2.3±1.7 years. A total of 3 cancers were reported, including one case of lymphoma. One case that occurred in an adult JIA patient who received the first dose of etanercept after age 18 was excluded from the analyses. One case was diagnosed <90 days after the initial etanercept dose and was excluded from the sensitivity analysis. Using the primary case definition, the incidence rate of cancer in this pediatric JIA population exposed to etanercept was 52/100,000 person-years (PYs) (95% CI 6-188).

AnalysisNFollow-up (PYs)Observed cancersExpected cancersSIR (95% CI)
Primary1641385220.543.7 (0.5-13.4)
Sensitivity1553345910.492.1 (0.1-11.5)

Conclusion: The the observed frequency of cancer in JIA patients treated with etanercept was higher than that expected for the general population but not statistically significant. Very few events were observed, leading to wide confidence intervals. This limits the interpretation of the findings, as does uncertainty about the underlying cancer risk in biologics-naïve JIA patients.
Disclosure of Interest: M. Harrison Employee of: PFIZER, C. Cummins Grant/Research Support from: Wyeth, G. Horneff Grant/Research Support from: Wyeth, Pfizer, T. Southwood Grant/Research Support from: Wyeth, D. Mines Employee of: PFIZER

Citation: Ann Rheum Dis 2010;69(Suppl3):147

Session: From biology to classification and targeted therapy in Juvenile Idiopathic Arthritis


Close Window