433 - Risk of Fatal Infection and Malignancy Related to Use of Anti-Tumor Necrosis Factor-Alpha Biologics by Rheumatoid Arthritis Patients

Sunday, November 6, 2011: 9:00 AM-6:00 PM
Hall F2 - Poster Hall (McCormick Place West)
Veena Thyagarajan1, Heather Norman2, Kimberly Alexander3, Pavel Napalkov3 and Cheryl Enger1, 1OptumInsight, Ann Arbor, MI, 2OptumInsight, Waltham, MA, 3Genentech, Inc., South San Francisco, CA
Presentation Number: 433

Background/Purpose: Rheumatoid arthritis (RA) may be treated with biologics which inhibit the pro-inflammatory cytokine tumor necrosis factor alpha (aTNF). The pro-inflammatory and apoptotic properties of TNF play a role in controlling infection and reducing malignant cell formation. This study estimated the risks of fatal infection and fatal malignancy among RA patients treated with aTNF biologics.

Method: A retrospective cohort study of RA patients who initiated treatment with adalimumab (ADA), etanercept (ETN), or infliximab (INF) (including aTNF naïve and aTNF switchers) from January 2000 to December 2008 was conducted using the Normative Health Information (NHI) database, an administrative database of a large US healthcare insurer. Patients were followed for the occurrence of fatal infection and fatal malignancy. aTNF therapy was identified in the NHI database by procedure claims for physician-administered drugs and by pharmacy dispensings for patient-administered drugs. Current exposure for physician-administered drugs was defined according to prescribing guidelines for RA plus 30 days and by recorded days supply in pharmacy records for patient-administered drugs plus 30 days. Study outcomes were identified by external linkage to the National Death Index database. Unadjusted incidence rates (IRs) per 1,000 person-years (PY) for study outcomes were calculated for each aTNF.  Intent-to-treat analysis was the primary analysis for fatal malignancy where exposure status is based on the aTNF on which the person entered the cohort, regardless of whether the patient discontinued the drug or switched to another aTNF during follow-up.  Time-on-drug analysis was the primary analysis for fatal infections where exposure status is based on current exposure with exposure censored upon treatment discontinuation or switch to a different aTNF. The association between baseline covariates and fatal infection were estimated with unadjusted ORs and 95% CIs for each covariate.

Result: 7,734 patients initiated ADA, ETN, or INF (13,296 PY for fatal malignancy; 10,710 PY current exposure for fatal infections). 21 fatal malignancies were identified. The IR for fatal malignancy was 1.24 (95% CI: 0.42-2.96) among ADA initiators, 1.57 (95% CI: 0.78-2.87) among ETN initiators, and 1.84 (95% CI: 0.87-3.48) among INF initiators. 12 fatal infections, including 2 fatal opportunistic infections, were identified. The IR for fatal infection was 0.78 (95% CI: 0.15-2.49) during current ADA exposure, 0.88 (95% CI: 0.30-2.10) during current ETN exposure, 0.83 (95% CI: 0.23-2.21) during current INF exposure, and 1.16 (95% CI: 0.32-3.09) during no current drug exposure. An increased unadjusted OR of fatal infection was observed with baseline chronic lung disease (OR: 13.62; 95% CI: 2.98-51.16), cardiovascular disease (OR: 10.12; 95% CI: 2.22-37.96), and steroidal estrogen use (OR: 4.79; 95% CI: 1.52-15.12).

Conclusion: The occurrence of fatal malignancy and fatal infection was rare. Risks for fatal malignancy and fatal infections were similar across aTNFs. The small number of outcomes limits inference of these risk estimates and prevented inclusion of baseline covariates into an adjusted model.

Keywords: anti-TNF therapy, infection, malignancy and rheumatoid arthritis (RA)

Disclosure: V. Thyagarajan, OptumInsight, 3 ; H. Norman, OptumInsight, 3 ; K. Alexander, Genentech, Inc., 3, Genentech, Inc., 1 ; P. Napalkov, Genentech, Inc., 3, Genentech, Inc., 1 ; C. Enger, OptumInsight, 3 .