Risk of Non-Hodgkin Lymphoma Following Treatment of Inflammatory Conditions with Tumor Necrosis Factor-Alpha Inhibitors

Purpose

Tumor necrosis factor-alpha (TNF-α) inhibitors are an increasingly common biologic treatment for moderate to severe inflammatory conditions such as psoriasis and rheumatoid arthritis. A limited number of studies and case reports of aggressive lymphoma and other malignancies in children and adolescent patients prompted a U.S. Food and Drug Administration black box warning for all TNF-α inhibitors. Although the two types of TNF-α inhibitors, anti-TNF monoclonal antibodies and TNF fusion protein, have similar clinical efficacy, their pharmacologic activity and modulation of immune pathways are different. Our objective was to examine and compare the incidence of non-Hodgkin lymphoma (NHL) among patients with inflammatory conditions treated with anti-TNF antibodies and TNF fusion protein.

Patients and Methods

We conducted a retrospective cohort study of new users of TNF-α inhibitors between 2009 and 2013 in the Truven Health MarketScan Research Database. Patients were included if they were ages 15+ years and had ≥12 months of continuous enrollment prior treatment initiation. Exclusion criteria included presence of any cancer, HIV or stem cell transplant in the year prior to first TNF-α inhibitor use.

Using longitudinal pharmacy claims data, we measured continuous use of anti-TNF antibodies (infliximab, adalimumab, golimumab, certolizumab), TNF fusion protein (etanercept) and other immunosuppressive medications. NHL cases were identified using a validated algorithm with administrative claims and ICD-9 diagnosis codes. These data were also used to identify diagnoses of inflammatory conditions and calculate Charlson comorbidity index scores. NHL incidence rates per 100,000 person-years (PY) and 95% confidence intervals (CI) were calculated and compared to age-standardized expected rates in Surveillance, Epidemiology and End Results Program registries from the same time period and geographic regions with stratification by type of TNF-α inhibitor, gender and age group (15-39, 40-64, 65+ years). Standardized incidence ratios (SIR) and exact 95% CI were calculated using Poisson regression. NHL risk with use of anti-TNF antibodies was compared to etanercept use in multivariable Cox proportional hazards models. We estimated hazard ratios (HR) and 95% CI with adjustment for inflammatory conditions and concurrent immunosuppressive medications.

Results

In a cohort of 118,050 TNF-α inhibitor users, 85,327 (72%) used anti-TNF antibodies and 42,406 (36%) used TNF fusion protein alone or consecutively after switching TNF-α inhibitor type. The most prevalent indications for TNF-α inhibitors were rheumatoid arthritis (47%), followed by inflammatory bowel disease (23%), psoriasis (21%), psoriatic arthritis (15%) and ankylosing spondylitis (6%). During the 212,479 PY of follow-up, a total of 194 TNF-α inhibitor users developed NHL; and the crude NHL incidence rate (91 per 100,000 PY) was greater than expected (28 per 100,000 PY, age-standardized). Compared to non-cases, TNF-α inhibitor users that developed NHL were older (median age 59 vs. 48 years) and had more concurrent use of corticosteroids (77% vs. 56%) and methotrexate (48% vs. 37%).

The overall age-standardized incidence ratio for NHL was 3.7 (95% CI 3.1-4.2) for TNF-α inhibitor users. Observed SIR was even higher in adolescent and young adult patients (15-39 years: SIR=7.3, 95% CI 4.5-11.3). Compared to etanercept users, patients treated with anti-TNF antibodies had greater risk of NHL (HR=1.5, 95% CI 1.1-2.0). This increased risk with anti-TNF antibodies was present in female users (HR=1.8, 95% CI 1.2-2.8) but less apparent in male users (HR=1.2, 95% CI 0.8-1.8).

Conclusions

In this large sample of patients treated with TNF-α inhibitors, we found higher incidence of NHL than expected in a similarly aged population and greater lymphoma risk with anti-TNF antibodies vs. etanercept. Further research with long-term follow up and clinical information on duration and severity of inflammatory conditions are needed to confirm these findings. Treatment of moderate to severe inflammatory conditions with TNF-α inhibitors is chronic and the potential lifetime exposure to these and other immunomodulating drugs is high, particularly for younger patients. Thus, determining the comparative safety of lymphoma among different types of TNF-α inhibitors has potentially significant implications and warrants continued evaluation.