358. RISK FACTORS FOR TREATMENT FAILURE IN PATIENTS WITH GIANT CELL ARTERITIS TREATED WITH TOCILIZUMAB PLUS PREDNISONE VERSUS PREDNISONE ALONE

Abstract

Background: Risk factors for treatment failure in patients with GCA are poorly understood. The interleukin‐6 receptor‐α inhibitor tocilizumab (TCZ) was approved for treatment of giant cell arteritis (GCA) following a randomized, placebo‐controlled trial (Stone et al. N Engl J Med. 2017;377:317). We aimed to identify predictors of treatment failure in GCA patients treated with TCZ+prednisone versus placebo (PBO)+prednisone in post hoc exploratory analysis of data from this trial. SkMip etoth Moadins: C Poanttientts received either weekly or every‐other‐week TCZ plus a 26‐week prednisone taper (TCZ+prednisone) or PBO plus a 26‐ or a 52‐week prednisone taper (PBO+prednisone). Both TCZ groups and both PBO groups were combined for this analysis. Primary endpoint was sustained remission at week 52. Treatment failure was defined as failure to achieve remission by week 12 or as flare between weeks 12 and 52 after remission by week 12. Variables analyzed as potential predictors of treatment failure included baseline demographics, disease‐ and treatment‐related factors, and health‐related quality‐of‐life (HRQoL) measures. Statistical significance was assessed by chi‐square tests for categorical variables and 2‐sample T tests for continuous variables. Results: Among 250 patients in the intent‐to‐treat population, 98 (39%) were in sustained remission at week 52. These responders included 16% (16/101) of the PBO+prednisone groups and 55% (82/149) of the TCZ+prednisone groups. In contrast, 46% of all patients (115/250) were treatment failures. These included 70% (71/101) of the PBO+prednisone group and 30% (44/149) of the TCZ+prednisone group. Among PBO‐treated patients, female sex (p = 0.003), longer disease duration (p = 0.031), and lower baseline SF‐36 physical component summary (PCS) scores (p = 0.026) were associated with treatment failure (Table 1). Female sex was not a predictor of treatment failure in the TCZ groups. Rather, higher patient global assessment of disease activity scores (p = 0.001) and lower SF‐36 PCS (p = 0.005), FACIT‐Fatigue (p < 0.001), and EQ‐5D scores (p = 0.005) predicted treatment failure among TCZ‐treated patients. Age, disease onset (newly diagnosed/relapsing), prednisone dose, and clinical features (polymyalgia rheumatica or cranial symptoms) were not associated with treatment failure in either group. Conclusion: Women with GCA responded poorly when treated with prednisone alone compared to men. Treatment with TCZ resolved treatment disparity between the sexes. Impaired HRQoL at baseline was an important predictor of treatment failure among TCZ‐treated patients. FACIT, Functional Assessment of Chronic Illness Therapy; MCS, Mental Component Summary; PCS, Physical Component Summary; PMR, polymyalgia rheumatica; PtGA, patient's global assessment of disease activity; SD, standard deviation; VAS, visual analog scale. 14 patients in the PBO+prednisone group and 23 patients in the TCZ+prednisone group did not meet the primary end point for reasons other than treatment failure. Percentages are based on number of responders/treatment failures (n).