Date: Monday, November 9, 2015
Session Type: ACR Poster Session B
Session Time: 9:00AM-11:00AM
Background/Purpose : The current treatment paradigm in rheumatoid arthritis (RA) is to attempt to decrease, when clinically feasible, concomitant use of OGCs after their use as part of the initial therapeutic approach. This study examined OGC-sparing effects of tocilizumab (TCZ) and RA biologics used second-line (BDM2) in RA patients (pts) treated with combination therapies and monotherapies.
Methods: This was a retrospective cohort study performed using the Truven MarketScan administrative claims database. RA pts were identified as those with an RA diagnosis (ICD-9-CM: 714.xx excluding 714.3x) followed by confirmatory RA diagnosis in 7-365 days or confirmatory DMARD procedure or prescription any time after the initial RA diagnosis. Pts had to be ≥18 y at the time of RA index date and continuously enrolled around the RA index date. Pts were then split into 2 treatment groups: TCZ treated and BDM2 treated, not including TCZ. Cohort demographic characteristics and treatment history were analyzed. OGC use and dosing were investigated during several time frames with respect to index therapy initiation. OGC doses were calculated using a prednisone-equivalent dose for each medication. A multilevel random coefficient model was used to test for statistically significant OGC-sparing effects of TCZ compared with BDM2 over time while adjusting for baseline OGC use, comorbidities, and demographics.
Results: The analysis population comprised 10,402 pts (2604 in TCZ initiator cohort, 8209 in BDM2 comparator cohort). Mean age at RA diagnosis was similar between the cohorts (51.0 vs 51.2 y). More women were observed in the TCZ cohort (81.2% vs 77.1%). Mean baseline Charlson Comorbidity Index (CCI) scores and individual comorbidities were higher in the TCZ cohort (2.6 vs 2.3). Mean baseline cumulative OGC dose was higher for TCZ than BDM2 (1988 mg vs 1616 mg). In multivariate results, there was a significant reduction in OGC use in both cohorts over time (b=0.177 mg/d; p<0.000). There was a significant time × cohort interaction such that the TCZ cohort had an additional reduction of 0.167 mg/d (p<0.0001). In the TCZ cohort, mean OGC dose decreased over time from 4.6 mg/d in months 1-3 to 1.2 mg/d in months 27-30, a 74% reduction. In the BDM2 cohort, mean OGC use decreased from 3.3 mg/d to 1.4 mg/d, a 58% reduction in the same period. Factors significantly associated with greater OGC use over time included baseline comorbidities as measured by the CCI (b=0.85, p<0.0001), female sex (b=0.99, p<0.000), and baseline OGC dose (b=0.78 mg/d, p<0.0001). Similar results were observed when we restricted the analysis to pts on monotherapy though the sample size was smaller.
Conclusion: In this large cohort of RA pts, initiators of TCZ had greater mean starting and cumulative baseline OGC doses and higher mean CCIs than initiators of other RA biologics previously treated with bDMARDs, implying that historically they might have been more difficult to treat. We found statistically significant OGC-sparing effects in RA pts treated with TCZ and BDM2. TCZ use was associated with significantly quicker and relatively more pronounced taper of OGC than BDM2. The clinical implications of such OGC-sparing effects warrant further investigation.
To cite this abstract in AMA style:Arnieri B, Sarsour K, Oliveri D, Pethö-Schramm A, Shah A, Quartey G. Comparison of Oral Glucocorticoid (OGC)-Sparing Effects in Tocilizumab and Other Biologic Dmards Using Multilevel Models in an Administrative Health Care Claims Database [abstract]. Arthritis Rheumatol. 2015; 67 (suppl 10). https://acrabstracts.org/abstract/comparison-of-oral-glucocorticoid-ogc-sparing-effects-in-tocilizumab-and-other-biologic-dmards-using-multilevel-models-in-an-administrative-health-care-claims-database/. Accessed May 8, 2021.
« Back to 2015 ACR/ARHP Annual Meeting
ACR Meeting Abstracts - https://acrabstracts.org/abstract/comparison-of-oral-glucocorticoid-ogc-sparing-effects-in-tocilizumab-and-other-biologic-dmards-using-multilevel-models-in-an-administrative-health-care-claims-database/