Abstract 1343: Rheumatoid Arthritis, Interstitial Lung Disease, Mortality, and Anti-TNF Therapy: Results from the BSR Biologics Register (BSRBR)
Authors:
W. G. Dixon, K. D. Watson, M. Lunt, K. L. Hyrich
Background:
There are a variety of pulmonary manifestations associated with RA, with clinical severity ranging from asymptomatic to life-threatening. Symptomatic RA-associated interstitial lung disease (ILD) is uncommon, but when present is associated with increased mortality. Whether RA patients with ILD who are treated with TNF inhibitors are at an increased risk of mortality compared those not treated with these agents has not been investigated.
Methods
The aims of this investigation were to estimate the impact of symptomatic RA-associated ILD on mortality and to explore whether TNF inhibitor exposure is associated with an increase in mortality. Subjects were RA patients participating in the BSR Biologics Register compared to a group of RA patients with active disease receiving non-biologic DMARDs. Cause of death was identified using ICD.10 codes up to October 31, 2006 from the United Kingdom Office of National Statistics. The prevalence and mortality from baseline physician-reported ILD and the association with treatment was explored.
Results
There were 9,847 TNF inhibitor treated patient compared to 2,906 non-TNF inhibitor treated patients. The mean age of TNF inhibitor treated patients was lower than non-TNF inhibitor treated patients (68 ± 10 vs. 62 ± 11, respectively). RA-associated ILD was reported at baseline in 337 patients, 55 (1.9%) of the non-TNF inhibitor group and 282 (2.9%) of the TNF inhibitor treated group. The odds of baseline ILD was 3.3-fold greater in patients with extra-articular manifestations of RA than those without (95% CI 2.7 – 4.2), 2-fold greater in those with rheumatoid factor (RF) seropositivity (95% CI 1.5 – 2.5), and higher in smokers. After adjusting for age, gender, RF, extra-articular disease, and smoking, the risk of death was 2.73 times greater in patients with RA-ILD compared to those without (95% CI 2.02 – 3.69).
There were four deaths among the 55 non-TNF inhibitor treated RA-ILD patients (7.3%) over a median follow-up time of 1.1 years and 49 deaths among the 282 TNF inhibitor treated RA-ILD patients (17.4%) over a median follow-up time of 2.63 years. RA-ILD was listed as the cause of death in 12 (4.3%) of the TNF inhibitor treated RA-ILD deaths, but none of the non-TNF inhibitor treated RA-ILD deaths. Considering follow-up time and censoring, the risk of overall mortality was 2.11 times greater in RA-ILD patients treated with TNF inhibitors vs. RA-ILD patients receiving non-TNF inhibitor DMARDs, after adjusting for age and gender.
Conclusions
Mortality is increased in RA-ILD patients, and may be disproportionately higher in TNF inhibitor patients with ILD. These findings suggest that TNF inhibitor use could increase the risk of mortality in patients with RA-associated ILD.
Editorial Comment
These data confirm that symptomatic RA-associated ILD is associated with an alarmingly high mortality rate (approximately 15% of the affected patients in about 2 years of follow-up) and are largely convincing that TNF inhibitors may contribute to mortality in RA-ILD patients. The mechanism of this increase in mortality related to TNF inhibition is not immediately clear from this data. Possible contributors include increased risk of pulmonary infection or even a direct detrimental effect of TNF inhibition on the damaged pulmonary parenchyma. Issues of selection may also explain some of the difference, as the reasons practitioners chose to use TNF inhibitors in RA-ILD patients may be related in some way to their overall mortality risk. Accounting for these decisions in some way, which is difficult to do in a study of this type, would address the issue of selection bias, particularly since a randomized trial of TNF inhibitors in RA-ILD patients is unlikely to be conducted.
