Work loss in patients with rheumatoid arthritis treated with abatacept, rituximab, tocilizumab or TNF inhibitors: a nationwide direct drug-to-drug comparison

Abstract

Objective: To compare work loss after starting tumour necrosis factor inhibitors (TNFi), rituximab, abatacept or tocilizumab in patients with rheumatoid arthritis (RA).

Methods: We used data from the Swedish Rheumatology Quality Register to identify patients aged 19-62 years who were treated with TNFi (n=15 093), rituximab (n=2123), abatacept (n=1877) or tocilizumab (n=1720) between 2007 and 2020. Data on work loss (0-365 days per year) from sick leave and disability pension were retrieved from linkage to the Social Insurance Agency. Patients in the different treatment arms were balanced regarding baseline covariates using inverse probability weighting (IPTW).

Results: Work loss increased for patients with RA until drug treatment initiation, reached a peak in the month after treatment initiation and then levelled off. Following IPTW, at 3 years before starting the treatment, there were no statistically significant differences in the mean annual adjusted work loss days between rituximab, abatacept or tocilizumab vs TNFi (mean difference vs TNFi: rituximab 1.1 days, 95% CI -4.5 to 6.7; abatacept 3.3, 95% CI -2.6 to 9.2; tocilizumab 1.2, 95% CI -4.9 to 7.3). At 3 years after starting the treatment (latest January 2021), there were also no statistically significant differences in the mean annual adjusted work loss days (mean difference: rituximab -4.8 days, 95% CI -11.3 to 1.7; abatacept 5.3, 95% CI -1.8 to 12.3; tocilizumab -0.6, 95% CI -7.7 to 6.5).

Conclusions: Taking channelling into account, patients with RA treated with TNFi, rituximab, abatacept or tocilizumab had similar trajectories of work loss from sick leave and disability pension until treatment initiation, and similar trend breaks and plateau 3 years thereafter.

Keywords: Biological Therapy; Economics; Rheumatoid Arthritis.

Figure 1. Mean inverse probability weighted monthly work loss days from 3 years before to 3 years after the treatment initiation for patients with rheumatoid arthritis receiving tumour necrosis factor inhibitors, rituximab, abatacept or tocilizumab. Numbers below x-axis are the number of observations for groups of drugs at different points in time. Observations are weighted by age, sex, education (three levels), year of treatment initiation, number of work loss days prior to treatment initiation, any work loss days (yes/no) prior to the treatment initiation and number of previous biologic drug treatments (zero, one, two, three or more).

Figure 2. Distribution of mean monthly work loss days. Mean monthly work loss days for treatment episodes with tumour necrosis factor inhibitors (top left), rituximab (top right), abatacept (bottom left) and tocilizumab (bottom right). Solid line is weighted mean monthly work loss days. Bars show weighted percentage of treatment episodes by category of mean monthly work loss days (maximum 30 days).

Figure 3. Incidence of work loss (top) and incidence of remission of work loss (bottom). In the incidence cohort (top), an event is a month with 15 or more work loss days. In the remission cohort (bottom), an event is a month with 15 or less work loss days. Cumulative incidence functions are inverse probability weighted by age, sex, education (three levels), year of treatment initiation, number of work loss days prior to treatment initiation, any work loss days (yes/no) prior to treatment initiation and number of previous biologic drug treatments (zero, one, two, three or more).