Abstract 1344: Are Cancers Occurring in RA Patients Treated with TNF Antagonists Particularly Aggressive?
Authors
P. Raaschou, J. Askling, F. Granath, M. Fored
Background
While a strong signal for a direct causal association between TNF inhibitor use and most cancers has not been detected (except possibly for non-melanoma skin cancers), meta analyses identifying an increased risk of malignancy risk in patients treated for short time periods in clinical trials could suggest that lead time to malignancy may be reduced in TNF inhibitor users. Malignancies with shortened lead times tend to be more aggressive, yet whether malignancies in TNF inhibitor treated patients are more aggressive than in those treated with non-biologic DMARDs has not been heretofore investigated.
Methods
Patients enrolled in the Swedish Biologics Register comprised the TNF inhibitor treated group (n = 6,299) while patients not treated with TNF inhibitors were drawn from inpatient and outpatient health records from Swedish health registries and from the Swedish Early Rheumatoid Arthritis Cohort (n = 66,471). Among these groups, patients with first malignancies were identified through linkage to the Swedish Cancer Registry. All cause mortality after cancer diagnosis in TNF inhibitor treated vs. untreated patients was determined by further linkage to the Swedish Death Registry up to July 2006 and compared using Cox regression.
Results
Among the TNF inhibitor treated RA patients, 170 solid tumors were identified resulting in 63 deaths by the endpoint date (36%). Among the non-TNF inhibitor treated RA patients, 3,347 solid tumors were identified resulting in 1,792 deaths by the endpoint date (53.5%). Using Cox regression to account for person-time and censoring, the adjusted relative risk of death for TNF inhibitor treated patients was not significantly different from non-TNF inhibitor treated patients (RR 0.8 (95% CI 0.5 – 1.2)). There was no difference in mortality in analyses stratified by gender, age at the time of cancer diagnosis, or by specific cancer subtypes (i.e. breast, lung, colorectal, prostate, or non-melanoma skin cancer
Conclusions
Mortality after cancer diagnosis does not appear to be increased in those with prior exposure to TNF inhibitors
Editorial Comment
The question of whether TNF inhibitors are associated with an increased incidence of cancer has been studied across a variety of populations; however, the question of whether mortality is increased after cancer diagnosis has not been addressed. This investigation goes a long way to demonstrating that mortality is not accelerated in RA patients who develop cancers after exposure to TNF inhibitors. This study was accomplished by linking several well-developed and highly complete national registries together. This strategy has many advantages, including completeness of data and large numbers of patients. However, one possible disadvantage is that important data at the individual patient level may not be available. For this study, it is not clear whether the TNF inhibitor treated and un-treated groups are comparable in other factors related to survival after cancer diagnosis other than their differences in exposure to TNF inhibitors. However, because the magnitude of difference in mortality favored the TNF inhibitor group for all solid tumors and by subtype, unbalanced factors would have had to be of such magnitude as to push the relative effect to the opposite side of the null, increasing the likelihood that these results represent truth.
