Effect of subcutaneous tocilizumab treatment on work/housework status in biologic-naïve rheumatoid arthritis patients using inverse probability of treatment weighting: FIRST ACT-SC study

The decrease in work productivity caused by rheumatoid arthritis (RA), whether it be for paid work or housework, has been gaining increasing attention [1‐3]. Participation in work-related activities was added as one of the overarching principles of the primary goals of RA treatment [4]. It is estimated that, 6 months after the onset of RA, patients experience functional decline secondary to joint symptoms caused by joint inflammation and cartilage destruction. This not only affects activities of daily living, such as home activities, recreation, and social relations, but also results in reduced work productivity among house workers (HWs) and paid workers (PWs) [5‐10]. Furthermore, it has been reported that, in Japanese patients with RA, work productivity and activity impairment are strongly correlated with the extent of physical disability and quality of life (QOL) [11].

The assessment methods for work productivity status are absenteeism (the decrease in number of actual working days by disease), presenteeism (the loss in the demonstration of the subject’s original working ability by disease activity), and overall work impairment (OWI; the sum of absenteeism and presenteeism). The Work Productivity and Activity Impairment Questionnaire (WPAI) is one of the recommended assessment methods for work productivity in RA patients [12, 13].

Recent advancements in the understanding of the molecular and cellular mechanisms of RA have led to the identification of novel targets and the development of effective biologic agents, such as tumor necrosis factor inhibitors [14] and anti-interleukin (IL)-6 receptor antibodies [15]. Tocilizumab (TCZ) is an anti-IL-6 receptor antibody that blocks the IL-6 receptor and inhibits the binding between IL-6 and its receptor. TCZ (in solution for intravenous administration) was approved for the treatment of RA in Japan in April 2008, in Europe in 2009, and in the US in 2010. Additionally, TCZ by subcutaneous injection (TCZ-SC) was approved in Japan in March 2013; thus, there are now two formulations available for RA patients.

No clinical studies have reported on the efficacy of TCZ in improving work productivity, either for paid work or housework, among RA patients. Therefore, in this study, we evaluated the effect of TCZ-SC based on improvements in work productivity and activity impairment among biologic-naive Japanese HWs and PWs with RA in a real-world clinical practice.

The present study is the first to assess the effect of TCZ-SC and/or csDMARDs on WPAI in PWs and HWs among Japanese patients with RA. We did not identify any significant difference between PWs treated with TCZ-SC and/or csDMARDs in terms of the change in OWI according to the WPAI at 52 weeks. However, we did observe an improvement in OWI from baseline in both treatment groups, meaning that RA treatment intervention was effective in decreasing disease activity, improving function, and improving overall QOL. These findings are consistent with a previous, large-scale study that evaluated the effects of adalimumab on WPAI in Japanese RA patients [5]. Previous studies of etanercept plus methotrexate in Latin America [25] and Asia [26] showed similar improvements in patient-reported outcomes, including WPAI. A previous study comparing baricitinib with placebo and adalimumab reported statistically significant improvements in absenteeism (P ≤ 0.05), presenteeism (P ≤ 0.001), and work productivity loss (P ≤ 0.001) with baricitinib compared with placebo; however, improvements compared with adalimumab were not statistically significant at week 52 [27].

Despite the result in OWI in the present study, improvement in the percentage of activity impairment in the overall population and among HWs was significantly better in the TCZ-SC group compared with the csDMARDs-alone group at week 52. This difference between treatment groups remained significant after adjusting for baseline characteristics using the IPTW method. This finding shows that, compared with the csDMARDs-alone treatment, TCZ-SC treatment resulted in improvement in disease activity (CDAI, SDAI, and DAS28) and significant improvement in QOL (EQ-5D, HAQ-DI, and K6). A recent, 48-week, observational study on adalimumab in Japan, focusing on work-related outcomes, showed that, compared with baseline, adalimumab treatment significantly improved measures of productivity loss due to absenteeism, presenteeism, OWI, and activity impairment in RA patients in all employment types, including PWs and HWs (P <  0.01) [28]. A fairly recent observational study in the US focused on work and activity impairment in employed moderate to severe RA patients and showed that etanercept led to significant reductions in overall work and activity impairment (P ≤ 0.0001) [1]. Furthermore, results of past studies have shown that total body inflammation and fatigue can be suppressed by inhibiting IL-6 [29]. In the present study, both clinical indexes and daily life (activity impairment) tended to improve in HWs and even in PWs. Although these changes did not reach statistical significance, these tendencies aligned with the results reported in previous studies. We consider that improvement in fatigue affected the improvement in activity impairment (daily life).

In the present study, according to DAS28-ESR, significantly more patients in the overall population, as well as PWs and HWs, treated with TCZ-SC achieved remission at week 52 (P <  0.0001 for all) compared with those receiving csDMARDs alone. However, by CDAI and SDAI, only those in the overall population and HWs treated with TCZ-SC achieved remission at week 52 (P <  0.0001). A study by Radner et al. that assessed the benefit of remission over low disease activity in RA showed that patients who achieved remission achieved better function, health-related QOL, and productivity [30].

Regarding the main differences in baseline characteristics between groups by type of work performed, HWs had greater disease severity at baseline than PWs. Similar findings were reported previously [5]. Although a tendency for improvement was observed in terms of other work productivity indices and QOL measures, no differences within the groups were observed for PWs, and a remarkable improvement in QOL measures was observed among HWs. The greater disease severity among HWs may be the reason this population experienced significantly greater improvements in overall QOL measures at week 52 compared with PWs.

An improvement in work productivity/activity impairment could be determined by differences in the treatment received as shown by the results of the presented analysis. Additionally, differences in disease severity, duration, treatment, and working conditions between PWs and HWs at baseline might have also contributed to this result. Similar conclusions were drawn in the study by Takeuchi et al. [5]. Furthermore, differences in the mechanical load on the affected joints between PWs and HWs could also be attributable to the differences in work productivity/activity impairment between PWs and HWs receiving TCZ-SC in addition to csDMARDs compared with those receiving csDMARDs alone.

Regarding the exploratory results, among PWs factors related to symptom improvement differed by study drug according to the results for HAQ-DI. These results indicated that TCZ-SC administration to patients with high HAQ-DI at baseline might result in greater improvement in work productivity and activity impairment.

The planned method for the primary analysis was changed from propensity score matching to IPTW to include all applicable patients in the analysis set because the planned sample size with a balanced number of patients was insufficient for propensity score matching. Propensity score matching allowed for easy calculation of the sample size and interpretation of the results; however, the feasibility of this method depended on whether the data of enrolled patients regarding the size of the matched sample and patient background were well balanced. For this reason, although we initially attempted to use propensity score matching, we decided to change the method after half of the patients were enrolled. In terms of bias, we consider that this change in the analysis method was acceptable because it was based on baseline data and not postbaseline data.

As part of the sensitivity analysis, we changed the adjustment method for estimating the propensity score, performed adjusted analysis with propensity score matching, used linear regression models adjusting for clinically significant factors, and performed subgroup analysis by unbalanced factors. These results were consistent with the primary results, and the robustness of the primary analysis was confirmed.

The AEs reported in the present study were in line with those previously reported for TCZ-SC in a real-world setting [31, 32]. Thus, no new safety concerns were raised, and TCZ-SC was considered a safe treatment option for Japanese RA patients.

This study had several limitations. First, the primary adjustment method for confounding was changed from propensity score matching to IPTW based on baseline data and not postbaseline data. However, we cannot deny the possibility of residual confounding effects related to the observational study design. Second, this study was conducted in a real-world clinical practice setting, and no specific criteria to initiate treatment with TCZ were applied at any of the 82 participating centers. Additionally, the dose in the TCZ-SC ± csDMARDs group was prescribed by the treating physician according to the prescribing information in the package insert; thus, we had no control over the doses prescribed. This may have affected the lower limit data of WPAI by causing a “floor effect”; in other words, patients with a low score may not have been able to show sufficient improvement. Third, we did not specifically collect information on whether a treat-to-target approach was used. However, the physicians who participated in this study were internal medicine specialists with extensive experience in the management of RA. They assessed patients at every visit (every 1–3 months). Therefore, we consider that all the patients were managed according to a treat-to-target approach. Fourth, PWs are exposed to compelling power (force) depending on the nature of the work they perform; thus, we could hardly confirm the difference between groups in terms of medical treatment. Conversely, HWs can determine their work activity level at will; thus, differences between groups in terms of the effect of medical treatment on QOL and activity level occurred easily. Fifth, WPAI is not adequate to evaluate the productivity of HWs as it was developed specifically for PWs. WPAI indexes, other than activity impairment, cannot be calculated in HWs. The other indexes are considered valid for all patients regardless of whether they are PWs or HWs. Finally, we hypothesized that the observed differences in QOL outcomes for HWs compared with PWs were related to a worst disease status in this subpopulation at baseline. Given the lack of treatment blinding, we cannot rule out other possible reasons for these outcomes such as the psychological bias and the potential emotional component related to initiating a novel treatment (e.g., biologics) that may be judged more effective than the conventional treatments. This may have affected the objectivity of the responses to the questionnaire measurements for QOL in HWs. Nevertheless, all the subjective components of CDAI and SDAI and physician's global assessment of disease activity, as well as the objective parameters, such as C-reactive peptide (CRP) and patient's global assessment of disease activity, improved with TCZ-SC in addition to csDMARDs.

Despite the lack of differences in OWI between groups at week 52, the overall group (particularly HWs) receiving TCZ-SC in addition to csDMARDs showed significant improvements in activity impairment, disease activity, and QOL compared with individuals receiving csDMARDs alone. The safety of TCZ-SC was acceptable for the treatment of Japanese RA patients in a real-world clinical practice. This study may help promote the evaluation of work productivity improvements in HWs and PWs by RA treatment.