OP0131 ANIFROLUMAB EFFECTS ON RASH AND ARTHRITIS IN PATIENTS WITH SLE AND IMPACT OF INTERFERON SIGNAL IN POOLED DATA FROM PHASE 3 TRIALS

Abstract

Background:Treatment with the type I interferon (IFN) receptor antibody anifrolumab was associated with clinical improvements in mucocutaneous and musculoskeletal disease activity in patients with systemic lupus erythematosus (SLE) in the phase 2 MUSE trial (NCT01438489) and phase 3 TULIP trials.1–4 Because rash and arthritis are the most common manifestations of SLE, the effect of anifrolumab on these symptoms can be examined in biomarker-defined subsets, as previously reported for the MUSE trial.2Objectives:To evaluate the effect of anifrolumab on rash and arthritis in patients with SLE, and the impact of IFN gene signature (IFNGS) on treatment response, using disease measures of different stringency in pooled data from the phase 3 TULIP trials.Methods:TULIP-1 (NCT02446912) and TULIP-2 (NCT02446899) were placebo-controlled, 52-week trials of intravenous anifrolumab administered every 4 weeks in patients with moderate to severe SLE.3,4 In this post hoc analysis, outcomes of rash and arthritis were evaluated using mucocutaneous and musculoskeletal domains of the SLE Disease Activity Index 2000 (SLEDAI-2K) and the British Isles Lupus Assessment Group (BILAG) index. In addition, the modified Cutaneous Lupus Erythematosus Disease Area and Severity Index (mCLASI) score was used to evaluate rash, and tender and swollen joint counts were used to assess arthritis.Results:360 patients received anifrolumab 300 mg (IFNGS test–high, n=298; IFNGS test–low, n=62) and 366 patients were given placebo (IFNGS test–high, n=302; IFNGS test–low, n=64). Change from baseline to Week 52 compared with placebo was measured by outcomes, ordered by their stringency. More anifrolumab-treated patients achieved rash improvement using SLEDAI-2K (complete resolution: difference 13.5%, nominal P<0.001), BILAG (at least 1 severity grade lowering: difference 15.5%, nominal P<0.001), and mCLASI (≥50% improvement, if baseline score >0: difference 15.6%, nominal P<0.001). Results were comparable in the IFNGS test–high subset (SLEDAI-2K: difference 17.0%, nominal P<0.001, BILAG: difference 16.1%, nominal P<0.001; mCLASI: difference 18.1%, nominal P<0.001). There was a trend toward anifrolumab-associated rash improvement in IFNGS test–low patients using BILAG (Figure). More patients receiving anifrolumab had SLEDAI-2K–defined resolution in arthritis (difference 8.2%, nominal P=0.029), BILAG severity lessening (difference 11.8%, nominal P=0.002), and ≥50% decrease in tender/swollen joint counts, when ≥6 at baseline (difference 12.6%, nominal P=0.016). Results were comparable in the IFNGS test–high subset (SLEDAI-2K: difference 11.7%, nominal P=0.005; BILAG: difference 12.9%, nominal P=0.003; joint counts: difference 11.3%, nominal P=0.054). In IFNGS test–low patients, there was a trend toward anifrolumab-associated arthritis improvement when measured using BILAG, and the effect of anifrolumab on the number of swollen/tender joint counts was similar to the IFNGS test–high group, although the IFNGS test–low sample size in this analysis was very small (Figure).Conclusion:In pooled data from the TULIP trials, anifrolumab treatment was associated with improvements in rash and arthritis using measures of different stringency. The SLEDAI-2K findings were largely driven by the subset of patients who were IFNGS test–high. However, using measures that were more sensitive to change, despite small sample sizes, IFNGS test–low patients may also have benefit.