Background Most randomized controlled clinical trials (RCTs) in systemic lupus erythematosus (SLE) include stratification factors to assure a balanced allocation of subgroups that might respond differently to therapeutic interventions. Strata could include baseline disease activity, baseline corticosteroid dosage, or race. Despite such stratification, geographic differences appear to impact responses in some studies. Sifalimumab and anifrolumab are fully human, IgG1 κ monoclonal antibodies in Phase IIb clinical development for SLE. Sifalimumab binds to and neutralizes the majority of IFN-α subtypes (Phase II study completed1). Anifrolumab binds to and neutralizes the type I IFN receptor and thus prevents signaling by any of the type I IFNs (Phase II study ongoing).
Objectives To assess geographic differences as potential confounders of efficacy analysis, we compared demographic data, baseline clinical characteristics, and standard of care (SOC) across geographic regions of 2 worldwide sifalimumab and anifrolumab Phase IIb SLE RCTs.
Methods This post-hoc analysis compared combined baseline data from the 2 RCTs between pre-specified geographic regions with an expected high placebo plus SOC response (Region 1 [R1]: Central America, South America, Eastern Europe, Asia) and low placebo plus SOC response (Region 2 [R2]: North America, Western Europe, South Africa). Eligibility criteria, similar for the 2 studies, resulted in enrollment of patients with moderate to severe SLE. Patients with active lupus nephritis or severe neuropsychiatric SLE were excluded.
Results Of the 736 randomized patients, 507 (68.9%) were from R1, and 229 (31.1%) were from R2. There were no clinically meaningful differences between regions in mean scores of global measures of disease activity (SLEDAI-2K: 11.3 vs 10.8; BILAG-2004 composite: 19.3 vs 19.2; physician global assessment: 1.79 vs 1.83). However, differences were seen in mean values of several demographic characteristics: patients from R1 were younger (38.0 vs 43.0 years), had a lower BMI (25.2 vs 28.1 kg.m2), a shorter duration of SLE (81.8 vs 131.2 months), and a lower SLICC/ACR damage index score (0.5 vs 1.1). In addition, higher percentages of R1 patients had elevated double-stranded DNA antibodies (85.0% vs 67.3%) and hypocomplementemia (C3: 44.8% vs 33.6%; C4: 29.0% vs 18.3%). Antimalarial use during the study was similar, but more R1 patients were treated with azathioprine (28.2%.vs 12.2%) and corticosteroids (94.1% vs 65.1%) and at higher corticosteroid dosages (≥10 mg/day: 66.9% vs 36.7%). In contrast, fewer patients received mycophenolate in R1 (5.1% vs 21.0%).
Conclusions SLE patients enrolled in RCTs from different geographic regions had notable differences in some demographic and baseline clinical characteristics as well as prescribed SOC medications. These differences may impact the analysis of the treatment response with SOC and/or investigational drug. Therefore, an imbalance in patients from regions with expected high or low SOC response should be considered in the feasibility and statistical considerations in SLE RCTs.
Khamashta M et al. Arthritis Rheumatol. 2014;66:S10 (Abstract L4)
Acknowledgements Funded: MedImmune. Editorial Assistance: Mark Hughes, PhD, QXV Communications, UK
Disclosure of Interest R. Furie Consultant for: Medimmune, M. Khamashta Grant/research support from: Bayer, Consultant for: INOVA diagnostics, Medimmune, GSK, UCB, L. Wang Employee of: MedImmune, J. Drappa Employee of: MedImmune, W. Greth Shareholder of: AstraZeneca, Employee of: MedImmune/AstraZeneca, G. Illei Shareholder of: AstraZeneca, Employee of: MedImmune