Cancer Incidence and Outcome for Patients With Rheumatoid Arthritis: A Long-term Population Study in Western Australia

Objective Our aim was to compare the incidence of malignancy and its effect on mortality between hospitalized patients with rheumatoid arthritis (RA) and controls.

Methods We conducted a population-level observational study of patients with RA (International Classification of Diseases, 9th revision, Clinical Modification [ICD-9-CM] code 714 and International Statistical Classification of Diseases and Related Health Problems, 10th revision, Australian Modification [ICD-10-AM] codes M05-M06) in the Hospital Morbidity Data Collection (HMDC) in Western Australia (WA) between 1985 and 2015, as well as nonexposed hospitalized controls matched on sex, age, and year of index admission. HMDC data were linked to the WA Cancer Registry and the WA Death Registry data, and cancer incidence rates (CIRs) per 1000 person-years, incidence rate ratios (IRR) with 95% CIs, and Kaplan Meier survival were estimated.

Results Among 14,041 patients with RA (67.56% female, median age 65.1 years) and 33,785 controls (65.16% female, median age 65.3 years), preexisting cancer in patients with RA was less prevalent than in controls (7.6% vs 14.2%; P < 0.01). In participants without prior cancer, the overall post index CIR was lower in those with RA (CIR 19.68 vs 24.77; IRR 0.79, 95% CI 0.76-0.83) and stable over 3 study decades. CIR was higher in patients with RA for lung (CIR 1.17, 95% CI 1.04-1.34) and hematological cancer (CIR 1.21, 95% CI 1.03-1.43) but lower for most other cancer types. Overall median survival was lower for patients with RA than controls (3.3 vs 5.3 years; P < 0.001) with increased mortality rates observed for most cancer subtypes.

Conclusion Overall CIR in patients with RA was consistently lower over time than in matched controls. CIR was only increased for lung and hematological cancer. Despite the overall lower CIR, post cancer mortality was higher for patients with RA in most cancer subtypes.