Arthritis & Rheumatism, Volume 62,
November 2010 Abstract Supplement
Abstracts of the American College of
Rheumatology/Association of Rheumatology Health Professionals
Annual Scientific Meeting
Atlanta, Georgia November 6-11, 2010.
Analysis of the Level of Agreement between Different Measures of Disease Flare in an Exploratory Study of Abatacept in Systemic Lupus Erythematosus (SLE).
Gordon3, Caroline, Becker1, Jean-Claude, Kelly1, Sheila, Peng2, Yun, Kinaszczuk1, Michael, Merrill4, Joan T.
This Phase II exploratory trial of abatacept used BILAG A/B flare (adjudicated per protocol) as the primary endpoint1. Post-hoc analyses examined other BILAG and physician- defined flare measures.
This was a double-blind, Phase II study in which SLE patients with active polyarthritis, serositis and/or discoid lesions were randomized to abatacept (~10 mg/kg) or placebo (PLC) for 1 year. Prednisone (30 mg/day or equivalent) was given for 1 month, then tapered by protocol. The primary endpoint was new adjudicated BILAG A/B flare (classic BILAG index)2 over 1 year after the start of steroid taper. Post-hoc assessments included BILAG A only flare (adjudicated), or physician-defined flare in response to the question 'Since the last visit, does the patient exhibit symptoms of an acute SLE flare?' (yes/no). Treatment modifications were recorded on visits when physician-defined flares occurred. Analyses are based on data for all randomized and treated patients.
118 abatacept and 57 PLC patients were assessed. Proportions of patients with >=1 BILAG A/B flare, BILAG A only flare or physician-assessed flare, respectively, were 79.7, 40.7 and 63.6% for abatacept versus 82.5, 54.4 and 82.5% for PLC. In the abatacept group, physician's assessment agreed with BILAG A/B flare in 71/165 (43%) cases: when the flare definition was restricted to BILAG A only, agreement occurred in 23/33 (70%) cases (Table). Most BILAG-defined flares that physicians did not rate as flares were B events (84/94 in abatacept-treated patients). 45.8% of physician-assessed flares occurred when no BILAG A/B flare was recorded. In instances where physicians recorded flare, BILAG A/B flare coincided with treatment modification the majority of the time (49/71 [69.0%]). Results in the PLC group were similar to abatacept.
Cross-tabulation of the presence/absence of flare according to BILAG and physician-defined measures
These data demonstrate some discordance between BILAG adjudication- and physician-determined flares. This may be due to: a) BILAG flare rules being based on an item worsening after improving at 2 prior visits (to avoid recording non-significant fluctuations in disease severity), whereas physicians may have recorded worsening irrespective of prior visits; b) physicians recording flare when items were the same over 4 prior weeks (not just when new, or worse after improving); c) the classic BILAG index may fail to capture some flare features detected by physicians. Some BILAG A/B flares occurred when no physician-defined flares were recorded; BILAG flares were more likely to be confirmed by the physician when restricted to BILAG A, suggesting increased ability of BILAG A to define events that physicians considered significant. These data provide valuable insight into issues encountered when assessing flare in clinical trials of SLE, and may help inform future trial design.
To cite this abstract, please use the following information:
Gordon, Caroline, Becker, Jean-Claude, Kelly, Sheila, Peng, Yun, Kinaszczuk, Michael, Merrill, Joan T.; Analysis of the Level of Agreement between Different Measures of Disease Flare in an Exploratory Study of Abatacept in Systemic Lupus Erythematosus (SLE). [abstract]. Arthritis Rheum 2010;62 Suppl 10 :1843