Flares in patients with systemic lupus erythematosus

Abstract Objective SLE is characterized by relapses and remissions. We aimed to describe the frequency, type and time to flare in a cohort of SLE patients. Methods SLE patients with one or more ‘A’ or ‘B’ BILAG-2004 systems meeting flare criteria (‘new’ or ‘worse’ items) and requiring an increase in immunosuppression were recruited from nine UK centres and assessed at baseline and monthly for 9 months. Subsequent flares were defined as: severe (any ‘A’ irrespective of number of ‘B’ flares), moderate (two or more ‘B’ without any ‘A’ flares) and mild (one ‘B’). Results Of the 100 patients, 94% were female, 61% White Caucasians, mean age (s.d.) was 40.7 years (12.7) and mean disease duration (s.d.) was 9.3 years (8.1). A total of 195 flares re-occurred in 76 patients over 781 monthly assessments (flare rate of 0.25/patient-month). There were 37 severe flares, 32 moderate flares and 126 mild flares. By 1 month, 22% had a mild/moderate/severe flare and 22% had a severe flare by 7 months. The median time to any ‘A’ or ‘B’ flare was 4 months. Severe/moderate flares tended to be in the system(s) affected at baseline, whereas mild flares could affect any system. Conclusion . In a population with active SLE we observed an ongoing rate of flares from early in the follow-up period with moderate–severe flares being due to an inability to fully control the disease. This real-world population study demonstrates the limitations of current treatments and provides a useful reference population from which to inform future clinical trial design.


Introduction
SLE is a major autoimmune multi-system rheumatic disease that is most common in women during the childbearing years [1]. It can affect any organ system and the disease varies in its clinical manifestations and severity between individuals. For most patients the disease is characterized by unpredictable relapses (flares) and remissions [2]. In recent years, mortality rates have improved [3] but there is still no cure, and flares of disease, infection and damage all continue to contribute to excess morbidity and mortality [4].
Flare is defined by International Consensus as 'a measurable increase in disease activity in one or more organ systems involving new or worse clinical signs and symptoms and/or laboratory measurements. It must be considered clinically significant by the assessor and usually there would be at least consideration of change or an increase in treatment' [5]. Flares are assessed using various validated disease activity measures. However, there is no standardized definition of a measurable increase in disease activity. The three main disease activity indices currently used in clinical trials are the Safety of Estrogen in Lupus Erythematosus National Assessment Trial-SLEDAI (SELENA-SLEDAI) [6,7], SLEDAI-2K [8] and the BILAG-2004 index [9]. For the SLEDAI and its derivatives, items are scored numerically and a numerical increase of at least three compared with the previous assessment constitutes a flare [5,10,11]. For the BILAG-2004 index, the presence of an item that is 'worse' or 'new' constitutes a flare [12][13][14].
Flares have been associated with more hospitalizations [15] and more organ and system damage, which in turn can lead to poorer prognosis and increased mortality [16][17][18]. In addition, the prolonged use of CS in the presence of persistence of disease or during flares can contribute to damage [19]. Flares, damage and prolonged use of CS can contribute to poor health-related quality of life [20][21][22][23]. Therefore, flare prevention is an important treatment goal in patient management.
Understanding the pattern of flares in SLE patients would be informative not only in the day to day management of these patients, but also in the interpretation of clinical trials of new medications where frequency and type of flares are included as outcome measures. This study describes the flare rates and types of flares in a prospective observational multicentre study of patients with active SLE after treatment of a flare.

Methods
The study was granted Multicentre Research Ethics Committee (MREC 02/5/035) approval and participants from the collaborating centres gave written informed consent. The collaborating rheumatology units were UK centres with an interest in SLE as part of the BILAG: Bangor, Birmingham (two centres), Blackburn, University College London, Nottingham, Manchester, Doncaster and Sheffield.
The inclusion and exclusion criteria of the patients recruited, and the demographic and clinical data collected have been detailed previously in the longitudinal study to determine the sensitivity to change of the LupusQoL [24]. Patients were eligible to be included in the study if they had a flare of SLE requiring specific treatment. For this study, flare was defined as a significant increase in disease activity resulting in a BILAG-2004 index 'A' or 'B' score based on manifestation(s) that are 'new' or 'worse' [9,13,14,25]. In addition, the flare definition for this study required patients to have an increase in therapy defined as one or more of the following: an increase of oral prednisolone to 20 mg/day, introduction of MTX, parenteral methylprednisolone, and/or other immunosuppressive therapy (e.g. CYC, rituximab). These patients were followed up monthly for 9 months and the BILAG-2004 disease activity index was assessed at each time point. For the purposes of this study, subsequent flares were defined as: severe ('A' flare/s irrespective of number of 'B' flares), moderate (two or more 'B' flares without any 'A' flares) and mild (one 'B' flare). We calculated the total numerical BILAG-2004 score at baseline where A ¼ 12, B ¼ 8, C ¼ 1 for each system [25].

Statistical methods
Patient data were summarized using the following descriptive statistics; means (S.D.), medians (interquartile ranges) and/or frequency counts. Flare rates were expressed as the number of flares per patient-month. Time to flare was also estimated using the Kaplan-Meier method.

Discussion
This multicentre prospective observational study captured the frequency of subsequent flare in a population entering the study at the time of moderate or severe flare. Using the BILAG-2004 index, we have described the pattern of flares in patients with SLE who were treated for severe or moderate flares over a 9-month period. These patients were followed monthly and were treated with various medications. We found that 76% of patients had a subsequent flare and the flare rate was 0.25/patient-month, of which 19% were severe, 16% moderate and the rest were mild flares. The flares were predominantly in the original organ system, demonstrating the inadequacy of current treatment to control disease and to prevent further flares that increase the risk of future organ damage.
There are only a small number of studies in the literature exploring incidence of flare in SLE as an outcome. These studies are difficult to compare with our study or with each other for the following reasons; different study populations and various methodological differences including study design (observational/interventional), duration of study, outcome measures employed and flare definitions. The challenges of making cross study comparisons are illustrated in supplementary Table S1, available at Rheumatology online, in studies that have used the BILAG index as a disease activity measure and a similar definition of flare [26][27][28][29][30][31].
Our study presents real-world data on patterns of flares and flare frequency in SLE patients in the UK after treatment of a severe or moderate flare with conventional therapy [32]. Thus, the patients included in this study are typical of those that meet the eligibility criteria (inclusion criteria of moderate-severe flares) for an interventional study. The frequency (monthly) of patient review has generated a wealth of data not only on the natural history of subsequent flares after standard treatment (flare rates), but also on the types (severity and system involvement) of    n pat : number of patients; n flare : number of flares; n system : number of times system affected with flare; %: percentage of total number of systems affected with flare; n/a: not applicable.