Objective: Selection of flare as the primary outcome variable in systemic lupus erythematosus (SLE) clinical trials fails to capture patients with persistently active disease (PAD). We sought to elucidate the frequency and determinants of flare and PAD.
Methods: Prospectively collected data from the Toronto Lupus Cohort were used to determine the incidence of flare and PAD in 2004 and 2005. Flare was defined as an increase in SLE Disease Activity Index 2000 update (SLEDAI-2K) score of >/=4 from the previous visit. PAD was defined as a SLEDAI-2K score of >/=4, excluding serology alone, on >/=2 consecutive visits. Data from 1, 2, and 3 years prior were used to model flare and PAD in 2004. Model properties were tested for prediction of flare and PAD in 2005.
Results: One-third of the patients had >/=1 flare, whereas nearly half experienced PAD in a given year. Nearly 60% of the patients had episodes of flare or PAD per year. At least 25% of patients had PAD without achieving the definition of flare. In the best-fitting model, predictors of PAD in 2004 were SLEDAI-2K score at the start of the outcome interval and prior cutaneous or musculoskeletal disease activity. This model gave 79% correct prediction of PAD in 2005. In contrast, flare prediction models performed poorly.
Conclusion: Persistent activity is a common disease state in SLE and should be an outcome variable in SLE clinical trials. Our PAD prediction model may aid prognostication and selection of patients for inclusion in clinical trials.