Article Text

Download PDFPDF

  1. Sebastian Unizony1,
  2. Min Bao2,
  3. Jian Han2,
  4. Yves Luder3,
  5. Paris Sidiropoulos2,
  6. Jinglan Pei2,
  7. John H. Stone1
  1. 1Massachusetts General Hospital Rheumatology Unit, Harvard Medical School, Boston, United States of America
  2. 2Genentech, South San Francisco, United States of America
  3. 3F. Hoffmann-La Roche, Basel, Switzerland


Background Risk factors for treatment failure in patients with giant cell arteritis (GCA) are poorly understood.

Objectives To identify predictors of treatment failure in GCA patients receiving tocilizumab (TCZ) or placebo (PBO) in combination with prednisone in a randomized controlled trial.1

Methods Two hundred fifty GCA patients received weekly or every-other-week TCZ plus a 26-week prednisone taper (TCZ+pred) or PBO plus a 26- or 52-week prednisone taper (PBO+pred). Patients who achieved and maintained clinical remission (CR) from week 12 to week 52 while adhering to the protocol prednisone taper were classified as responders. CR, adjudicated by investigators, was defined as the absence of disease flare (GCA signs or symptoms, and/or ESR elevation attributable to GCA that required further treatment [eg, rescue prednisone]) regardless of CRP level. Treatment failure was defined as failure to achieve CR by week 12 or occurrence of flare between weeks 12 and 52. Both TCZ groups and both PBO groups were combined for this analysis. Potential predictors investigated included baseline demographics, disease- and treatment-related factors, and health-related quality of life (HRQOL) measures. Univariate and multivariate analyses were performed.

Results Overall, 45% (113/250) of patients were responders: 27% (27/101) in the PBO+pred groups and 58% (86/149) in the TCZ+pred groups. In contrast, 44% (111/250) of patients experienced treatment failure: 66% (67/101) in the PBO+pred group and 30% (44/149) in the TCZ+pred group. The other 10% (26/250) of patients were nonresponders for reasons other than treatment failure: 7 in the PBO+pred group and 19 in the TCZ+pred group. In univariate analysis, female sex and lower baseline SF-36 Physical Component Summary (PCS), Mental Component Summary, and FACIT-Fatigue scores were associated with treatment failure among PBO+pred–treated patients, whereas higher patient global assessment of disease activity scores and lower SF-36 PCS, FACIT-Fatigue, and EQ-5D scores were associated with treatment failure among TCZ+pred–treated patients (Figure 1). Among TCZ+pred–treated patients, no treatment response difference according to sex was observed. Age, previous relapse, starting prednisone dose, and GCA clinical features (cranial or polymyalgia rheumatica symptoms) were not associated with treatment failure in either group based on univariate analysis. Multivariate logistic regression demonstrated that PBO+pred treatment, female sex, worse FACIT-Fatigue scores at baseline, and historical CRP >2.5 mg/dL all independently increased the risk for treatment failure (Figure 2).

Conclusion Female GCA patients responded particularly poorly if treated with prednisone alone according to univariate analysis. Female sex, impaired HRQOL at baseline, history of elevated CRP, and treatment with prednisone alone are independent risk factors for treatment failure in GCA. These factors may be considered when determining which treatment would be best for a particular patient.

References [1] Stone JH, et al. N Engl J Med. 2017;377:317-328.

Disclosure of Interests Sebastian Unizony Grant/research support from: F. Hoffmann-La Roche, Genentech, Consultant for: Kiniksa, Sanofi, GSK, Min Bao Shareholder of: F. Hoffmann-La Roche Ltd., Employee of: Genentech, Jian Han Shareholder of: F. Hoffmann-La Roche Ltd., Employee of: Genentech, Yves Luder Shareholder of: F. Hoffmann-La Roche, Employee of: F. Hoffmann-La Roche, Paris Sidiropoulos Employee of: Genentech, Jinglan Pei Employee of: Genentech, John H. Stone Grant/research support from: F. Hoffmann-La Roche, Genentech, Xencor, Consultant for: Chugain, F. Hoffmann-La Roche, Genentech, Xencor

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.