Evaluation of the effect of anti-tumor necrosis factor agent use on rheumatoid arthritis work disability: The jury is still out

Objective. To examine the role of anti-tumor necrosis factor (anti-TNF) agents in predicting work disability in subjects with rheumatoid arthritis (RA). Methods. We studied 953 subjects with rheumatologist-diagnosed RA from a US cohort using a nested, matched, case-control approach. Subjects provided data on medication usage and employment every 6 months for 18 months, were employed at baseline, and were age <65 years at last followup. Cases were subjects who were not employed at followup (n = 231) and were matched similar to 3:1 by time of entry into the cohort to 722 controls who were employed at followup. Risk of any employment loss, or loss attributed to RA, at followup as predicted by use of an anti-TNF agent at baseline was computed using conditional logistic regression. Stratification on possible confounding factors and recursive partitioning analyses were also conducted. Results. Subjects' mean age was 51 years, 82% were female, 92% were white, and 72% had more than a high school education. Nearly half (48%) used an anti-TNF agent at baseline; characteristics of anti-TNF agent users were similar to nonusers. In the main analyses, anti-TNF use did not protect against any or RA-attributed employment loss (odds ratio [95% confidence interval] 1.1 [.0.7-1.6] versus 0.9 [0.5-1.5]). However, a protective effect was found for users with disease duration <11 years (odds ratio [95% confidence interval] 0.5 [0.2-0.9]). In recursive partitioning analyses, age, RA global severity, and functional limitation played a much greater role in determining employment loss than anti-TNF agent use. Conclusion. Anti-TNF agent use did not protect against work disability in the main analyses. In stratified analyses, their use was protective among subjects with shorter RA duration, whereas in nonparametric analyses, age and disease factors were the prominent predictors of work disability.