Background Even if randomized trials EXPLORER and LUNAR failed to prove the superiority of rituximab versus placebo in patients with systemic lupus erythematosus, several encouraging indications as refractory lupus nephritis and new clinical trials renewed interest for this molecule.
We hypothetized that SLE response criteria used in EXPLORER were not sensible enough to show rituximab efficacy and that new response criteria could show a significant difference.
Objectives Our objective was to reanalyze EXPLORER trial's raw data using the newly described SLE response criteria.
Methods We proceeded to a pre-specified re-analyze of EXPLORER trial's raw data. The patients included in EXPLORER study had active SLE disease defined by a British Isles Lupus Assessment Group (BILAG) score A or 2 BILAG B despite immunosuppressive regimen. Renal and neurological SLE were excluded. Patients were randomized through a 2/1 ratio to receive either two 1gr rituximab infusions (day 0 and 15) repeated at month 6 or a placebo. Standard SLE treatment and other immunosuppressant were continued. Patients received in a stratified manner prednisone ranging from 0.5 to 1.0 mg/kg depending on disease severity at inclusion. The original efficacy criterion was a composite clinical score using BILAG at week 52.
In our new analysis, rituximab efficacy was assessed at week 52 using 4 criteria: SRI-4 (Systemic lupus erythematosus Responder Index) with and without a concomitant oral prednisone (OP) tapering objective of <10mg at month 6 (SRI-4 with and without OP tapering), Lupus Low Disease Activity Score (LLDAS) and BILAG-based Combined Lupus Assessment (BICLA).
Results Data from all 257 patients were available. There was 234 women (91%) with a mean age of 40,3 years among which 177 (69%) received hydroxychloroquine.
At week 52, SRI-4 response rate was 27,2% in the rituximab group vs 22,7% in the placebo group (p=0.43); SRI-4 with OP tapering was 16% in the rituximab group vs 13.6% in the placebo group (p=0.62); LLDAS was 16% in the rituximab group vs 12.5% in the placebo group (p=0.46) and BICLA was 15.4% in the rituximab group vs 15.9% in the placebo group (p=0.91).
Subgroup analyses demonstrated a trend for better efficacy of rituximab compared to placebo in the subgroup of patients co-treated with methotrexate: SRI-4 of 30.6% in the rituximab group vs 12% in the placebo group (n=74, p=0.08). This trend was not found in the subgroup co-treated with azathioprine: SRI-4 of 26.7% in the rituximab group vs 30.6% in the placebo group (p=0.68), nor in the subgroup co-treated with mycophenolic acid: SRI-4 of 23.1% in the rituximab group vs 21.6% in the placebo group (p=0.86). In the subgroup of patients with a BILAG A/B in hematological system or vasculitis at baseline, there was a significantly higher SRI-4 response rate with rituximab: 28,6% vs 5,3% in the hematological group (n=61, p=0,047) and 39,3% vs 0% in the vasculitis group (n=38, p=0,037).
Cumulative dose of steroids at week 52 were not statistically different: 4223mg in the rituximab group vs 4390mg in the placebo group (p=0,65).
Conclusions Our study confirms the results from the original EXPLORER study. These results might be partly related to the study's design, notably the high daily oral prednisone used. Our work suggests efficacy in subgroups with vasculitis and hematological involvement.
Disclosure of Interest None declared