Background/Purpose: The immune system plays a critical protective role in cancer (CA) development. Perturbations in immune signaling, including cytokine dysregulation, may disrupt this homeostatic balance. Recently, IL-1β inhibition reduced lung cancer incidence and mortality in a large RCT. We have previously shown serum cytokines are predictive of CA mortality in RA. The goal of the current study was to examine whether circulating cytokines are associated with CA incidence.

Methods: We linked the Veterans Affairs Rheumatoid Arthritis (VARA) registry with the VA Central Cancer Registry (VACCR) and the National Death Index (NDI). VACCR captures >90% of CAs occurring in US Veterans. VARA participants with available cytokine data and without a history of CA prior to VARA enrollment were included in analyses. Serum cytokines and chemokines (CKs) were measured with a 17-plex bead based assay using banked serum collected at VARA enrollment. CK score, an overall measure of CK concentrations, was calculated from individual analytes. Associations of CK with incident CA (identified in VACCR and NDI) were assessed using multivariable Cox regression models adjusting for age, sex, race, smoking status, DAS28, MD-HAQ, RF titer, methotrexate, and biologic use. CK score and individual analytes were analyzed as log-transformed continuous values and in quartiles. Additional analyses were stratified by smoking status (current vs. former/never).

Results: We studied 1,216 US Veterans with RA (mean age 63 yrs, 89% male, 78% anti-CCP positive) with available CK data, all without a pre-enrollment history of cancer. During 10,034 pt-yrs of follow-up, 159 incident CAs occurred with a median time to cancer of 4.7 yrs (IQR 2.7-6.9). Lung CA was the most frequent site (n=41), followed by prostate (n=26), lymphoproliferative (n=23), and skin (n=15). Log-transformed CK score was associated with an increased risk of incident CA (aHR 1.45, 95% CI 1.27-1.67). Higher concentrations of 13 analytes were associated with increased CA risk (Figure). The highest quartile of CK score was associated with a >2-fold increased risk of CA (aHR 2.40; 95% CI 1.21-4.76). The highest quartiles of 8 analytes were significantly associated with CA risk, with the strongest effect sizes for IL-2 (aHR 2.33; 95% CI 1.48-3.67), IL-17 (aHR 2.49; 95% CI 1.63-3.82), and MIP-1B (aHR 2.32; 95% CI 1.24-4.35). In stratified analyses, CK score was significantly associated with incident CA in both current and former/never smokers, though greater in current smokers.

Conclusion: Serum CK concentrations are predictive of future CA in RA patients, even after adjustment for RA disease activity and smoking. In addition to suggesting a potential role of CK profiling in modeling CA risk in RA, these results suggest that disease control and “normalization” of systemic CK disturbances could be important strategies in CA prevention in RA.