A propensity score-matched (PSM) analysis of organ damage in patients with systemic lupus erythematosus (SLE) from the pooled bliss long-term extension (LTE) trials versus the toronto lupus cohort (TLC)

Background: A pooled analysis of the open-label BLISS LTE studies (BEL112233/BEL112234) reported low levels of organ damage accrual (measured by Systemic Lupus International Collaborating Clinics [SLICC]/American College of Rheumatology Damage Index [SDI]) in patients who received belimumab (BEL) plus standard therapy (SoC) over a 5 year period. However, the LTE studies had no SoC arm. This post-hoc study (206347) used PSM to match BLISS LTE patients to TLC patients to define a SoC treatment comparison cohort. Objectives: To assess damage accrual in patients with SLE treated with BEL plus SoC compared with PSM patients from the TLC treated with SoC alone. Methods: This analysis compared the mean SDI change from baseline (over 5 years), time to SDI event (on all patients with ≥1 year follow-up), and magnitude of year-to-year SDI change (over 5 years), from baseline to Year 5 in patients treated with BEL plus SoC (pooled United States [US] and non-US data from the BLISS LTE studies), and SoC alone. Patients in the LTE and TLC were 1:1 PSM based on 16 clinical variables with a propensity score calliper ±20%. Regression augmented inverse propensity score weighting (IPSW) tested the robustness of the PSM results. Results: For the 5 year analysis, 181 LTE patients were matched to 181 TLC patients (mean bias 3.8%) from a larger pool of 973 patients (BLISS LTE n=592; TLC n=381). Time-to-event PSM resulted in 323 LTE and 323 TLC patients (mean bias 3.7%) from a larger pool of 1541 patients (BLISS LTE n=949; TLC n=592). The mean SDI score change from baseline in the BEL group was 0.265 (95% confidence interval [CI]: 0.180, 0.350) compared with 0.718 (95% CI: 0.547, 0.889) in the SoC group, resulting in a BEL treatment effect of –0.453 fewer SDI units (95% CI: –0.646,–0.260; p<0.001) over 5 years compared with SoC alone. The IPSW model produced similar results (–0.374, 95% CI: −0.512,–0.236; p<0.001). Patients treated with BEL were 60% less likely to progress to a higher SDI score over any given year of follow–up compared with SoC patients (hazard ratio 0.397, 95% CI: 0.275, 0.572; p<0.001). A patient receiving BEL has a 3.1% annual probability of organ damage progression compared with a 7.5% annual probability with SoC. Among the 646 time–to–event matched patients, there were 49 increases in SDI over the first 5 years in the BEL group and 102 in the SoC group. Of these, 4.1% (n=2/49) of the BEL group had an SDI increase ≥2 compared with 25.5% (n=26/102) of the SoC group. Therefore, for patients who experienced any increase, the likelihood of experiencing a≥2 point SDI increase was 6–times greater in the SoC group (25.5/4.1=6.22; p=0.002). Conclusions: This PSM analysis demonstrates that BEL plus SoC reduces, and slows the rate of organ damage progression and reduces the magnitude of progression compared with SoC alone. Acknowledgements: Study funded by GSK. Emma Hargreaves, MA, of Fishawack Indicia Ltd, UK, provided editorial assistance funded by GSK. Disclosure of Interest: M. Urowitz Grant/research support from: GSK, Consultant for: GSK, R. Ohsfeldt Employee of: GSK contractors with Medical Decision Modelling Inc., R. Wielage Employee of: GSK contractors with Medical Decision Modelling Inc., J. Dever Employee of: GSK contractors with Medical Decision Modelling Inc., M. Zakerifar Employee of: GSK contractors with Medical Decision Modelling Inc., Y. Asukai Shareholder of: GSK, Employee of: GSK, S. Ramachandran Shareholder of: GSK, Employee of: GSK, A. Joshi Shareholder of: GSK, Employee of: GSK DOI: 10.1136/annrheumdis-2018-eular.3319 Citation: Ann Rheum Dis, volume 77, supplement Suppl, year 2018, page A177Session: Present and future treatments for SLE, Sjögren’s and APS