Article Text

Download PDFPDF

OP0002 No difference in the risk of malignancy in tocilizumab versus tnf inhibitor initiators in patients with rheumatoid arthritis: a multi-database cohort study
  1. S.C. Kim1,
  2. A. Pawar1,
  3. R.J. Desai1,
  4. S. Gale2,
  5. M. Klearman2,
  6. K. Sarsour2,
  7. D.H. Solomon1,
  8. S. Schneeweiss1
  1. 1Brigham and Women’s Hospital, Boston
  2. 2Genentech, South San Francisco, USA


Background Few studies have conducted head-to-head comparisons of malignancy risk between different types of biologics in rheumatoid arthritis (RA).

Objectives To examine the rate of incident malignancy excluding non-melanoma skin cancer (NMSC) in RA patients newly treated with tocilizumab (TCZ) versus TNF inhibitors (TNFi).

Methods We conducted a cohort study using data from 3 U.S. healthcare claims databases (2010–2015) – Medicare, IMS PharMetrics Plus or Truven MarketScan. Adults aged ≥18 years with RA who newly started TCZ or a TNFi after failing a different TNFi, abatacept or tofacitinib were included. The primary outcome was incident malignancy excluding NMSC based on 2 diagnosis codes within 2 months (specificity >98%). The 10 most frequently occurring cancers, leukaemia and human papilloma virus-related cancer and all-cause mortality were analysed as individual secondary endpoints. (table 1). For the primary as-treated analysis, follow-up time started the day after cohort entry and ended on treatment discontinuation, outcome occurrence, disenrollment, death, or the end of study period. To control for >60 potential confounders, TCZ starters were propensity score (PS)-matched to TNFi starters with a variable ratio of 1:3 within each database. Hazard ratios (HR) from the 3 PS-matched cohorts were combined by a fixed-effects model.

Results We included a total of 10,393 TCZ initiators PS-matched to 26,357 TNFi initiators. A total of 118 malignancies occurred in TCZ starters and 322 in TNFi starters across the three databases. The IR of malignancy per 100 person-years ranged from 0.81 (IMS) to 2.18 (Medicare) in TCZ and from 0.98 (MarketScan) to 2.16 (Medicare) in TNFi. The risk of incident malignancy was similar between the two groups across all three databases (table 1), with a combined HR of 0.92 (95% CI: 0.74 to 1.14) in TCZ versus TNFi. Secondary analyses by cancer subtype and all-cause mortality showed similar results.

Abstract OP0002 – Table 1

Combined HR (95% confidence interval) of incident malignancy: a 1:3 variable ratio PS matched analysis comparing TCZ to TNFi.

Conclusions This large multi-database cohort study found no difference in the risk of malignancy excluding NMSC in patients with RA who newly start TCZ versus TNFi.

Disclosure of Interest S. Kim Grant/research support from: Roche, Pfizer, Bristol-Myers Squibb, A. Pawar : None declared, R. Desai: None declared, S. Gale Employee of: Genentech, M. Klearman Employee of: Genentech, K. Sarsour Employee of: Genentech, D. Solomon Grant/research support from: Roche, Pfizer, Lilly, Amgen, CORRONA, S. Schneeweiss Grant/research support from: Genentech/Roche, Boehringer Ingelheim, Consultant for: Aetion, WHISCON, LLC

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.