Background/Purpose Most studies of the safety profile of TNF inhibitors (TNFi) – in particular in relation to cancer risks – have been performed in patients with rheumatoid arthritis. Today, however, TNFi are widely used in patients with axial spondyloarthritis (AxSpA) and psoriatic arthritis (PsA), i.e., in populations with different age/sex distributions and with potentially different underlying risks for cancer. The aim of this study was to assess risks of overall and site-specific cancers in patients with AxSpA and PsA treated (vs. not) with TNFi, and to compare these risks to the risk in the general population. To do this we used pooled data from two Scandinavian national biologics registers and other population-based data-sources.

Methods By linking data from the Swedish (ARTIS) and Danish (DANBIO) biologics registers, we assembled a cohort of 8,156 (ARTIS=4,901 and DANBIO=3,255; total person-years=29,011) patients with AxSpA (57%) and PsA (43%) that started a TNFi 2001-2010. From the Swedish National Patient Register we identified a national TNFi-naïve AxSpA/PsA comparator cohort (n=24,058, person-years=112,714), and a Swedish age- and sex matched general population comparator cohort (n=103,380, person-years= 535,345).  The first invasive cancer for each subject was identified through linkage with the nationwide Swedish and Danish Cancer Registers (2001-2010). Subjects with previous cancers were excluded in all three cohorts. Age- and sex-standardized incidence rates and relative risks (RR) were calculated for cancer overall, and for six specific cancer sites (lung-, colorectal-, female breast-, prostate cancer, malignant melanoma, and lymphoma).

Results In total we identified 129 cancers among the TNFi treated patients, 744 in the

TNFi –naïve cohort, and 3,259 in the general population comparator cohort. TNFi-naïve patients were not at increased cancer risk vs. the general population (RR=1.06), but displayed a lower risk for colorectal cancer (RR=0.70). Based on 129 incident cancers, TNFi treated patients did not have a higher risk of cancer (vs. TNFi-naïve patients), RR=0.80 (95% CI 0.66-0.98).  Point estimates below 1 were noted for lung-, colorectal- and prostate cancer though only the latter reached statistical significance (Table).

Conclusion Based on these nationwide clinical data from Sweden and Denmark, TNFi treatment in patients with AxSpA and PsA was not associated with any significantly increased risks of cancer overall, nor for the six most common cancer types. The tendency towards decreased risks for lung, and prostate cancer following TNFi treatment may reflect selection related to pre-treatment screening. Furthermore, we observed a decreased risk of colorectal cancer in AxSpA/PsA.

¹Age- and sex-standardized relative risks (RR) with 95% confidence intervals (Cl)