Causal cascade of direct and indirect effects of anifrolumab on patient-reported outcomes: structural equation modelling of two Phase 3 trials

Abstract

Abstract Objectives SLE significantly impairs health-related quality of life (HRQoL). In this post hoc analysis, structural equation modelling was used to examine the ‘causal cascade’ of interaction between anifrolumab, disease activity and patient-reported outcomes (PROs) in pooled data from the phase 3 TULIP-1 and TULIP-2 trials. Methods Data were pooled from the TULIP-1 (n = 364) and TULIP-2 (n = 362) randomized, placebo-controlled, 52-week trials of intravenous anifrolumab (300 mg every 4 weeks for 48 weeks). We evaluated changes from baseline to week 24 and week 52 in four clinical (BICLA, BILAG-2004, SLEDAI-2K and changes in glucocorticoid dosage) and six PRO measures (SF-36, FACIT-F, EQ-5D, LupusQoL, PHQ-8 and pain NRS) in our hypothesized model of interactions. Results Our hypothesized model had an acceptable fit to the pooled TULIP trial data. At week 24, significant paths revealed that when compared with placebo, anifrolumab treatment improved disease activity as measured by BICLA, BILAG-2004, SLEDAI-2K and changes to glucocorticoid dosage. In turn, these clinical measures reduced pain, which improved fatigue, physical functioning, mood/emotions and HRQoL. When the model incorporated number of glucocorticoid tapers as the measure of change in glucocorticoid dosage, treatment effects of anifrolumab on glucocorticoid tapers were not retained at week 52. However, at week 52 treatment indirectly improved HRQoL through its direct effects on BICLA. Conclusions Anifrolumab is associated with significant patient-reported improvements in aspects of HRQoL including pain, fatigue, mood and physical function. These benefits are from the direct effect of anifrolumab treatment on disease activity and reduction in glucocorticoid dosage.