Background/Purpose: Patients with rheumatoid arthritis (RA) are at a small, but significantly increased risk of cancer compared with the general population. This risk varies by cancer type, with highest risk observed in lymphoma and lung cancer. Findings relative to prostate cancer are conflicted (standardized incidence ratio [SIR] 1.15, 95% CI 0.98-1.34, in recent meta-analysis). Recognizing the male predominance and large number of cases available in the Veterans Health System, we estimated the risk of prostate cancer in a national population of male US Veterans with RA.

Methods:

We identified male patients with RA using an algorithm that required ≥2 diagnostic codes ≥30 days apart, rheumatologist diagnosis, and either a DMARD prescription or positive RF or anti-CCP within the VA Corporate Data Warehouse (1/2000-4/2018). We excluded individuals with diagnostic codes for psoriatic arthritis or ankylosing spondylitis, and those with prostate cancer or prostatectomy prior to the index date. Prostate cancer and prostate cancer death were identified from the VA Central Cancer Registry (VACCR; captures up to 90% of cancers in the VA) and the National Death Index. Prostate cancer incidence rates (IR; per 1,000 pt-yrs) and 95% CIs were calculated from the index date (first date RA diagnosis by algorithm). SIR and 95% CI were calculated from age- and race-matched rates in the Surveillance, Epidemiology, and End Results Program (SEER). Multivariable Cox regression models assessed the association of seropositivity (RF or anti-CCP positive) with prostate cancer risk adjusting for age, race, smoking status, and Agent Orange exposure (reported risk factor for prostate cancer).

Results: RA patients included in analyses (n=50,870) had a mean (SD) age of 64 (11) years, 74% were white, 11% were black, 75% were ever smokers, and most were positive for either RF (67%) or anti-CCP (66%). Over 361,419 pt-yrs of follow-up, there were 1,243 incident prostate cancers (IR 3.4; 95% CI 3.3-3.6 per 1,000 pt-yrs). Similar rates were observed in seropositive (IR 3.5, 95% CI 3.3-3.7) and seronegative (IR 3.2, 95% CI 2.8-3.6) RA patients. Relative to SEER rates, SIR of prostate cancer was 0.89 (95% CI 0.85-0.95). Stratified by seropositivity, prostate rates were lower than those in SEER for seronegative patients (SIR 0.81, 95% CI 0.71-0.92) but similar for seropositive patients (SIR 0.94, 95% CI 0.88-1.01). In multivariable models, age, black race, and Agent Orange were independently associated with prostate cancer risk, while seropositivity was marginally associated (HR 1.15, 95% CI 0.97-1.36).

Conclusion: In one of the largest studies of its kind to date, we estimated prostate cancer rates to be slightly lower in RA than in the general population. However, this small difference may be accounted for by cases not captured within VACCR (estimated to approach 10%). Autoantibody status does not appear to portend substantial risk, however, additional investigation will be needed to identify whether other disease-related factors are associated with prostate cancer risk in this population.

« Back to 2018 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/risk-of-prostate-cancer-in-us-veterans-with-rheumatoid-arthritis/