Abstract
Objective
Outcome measures used in SLE randomised controlled trials (RCTs) have impeded trial success and drug approval. In an ongoing global project to develop a novel outcome measure for SLE RCTs, a ‘core set’ of domains with which to determine treatment response was chosen via consensus. We investigated the impact of restricting disease activity assessment to this core set on responder classification in RCTs.
Methods
Using pooled baseline and week 52 data from two phase III SLE RCTs (Belimumab in Subjects with Systemic Lupus Erythematosus (BLISS)-52, NCT00424476; BLISS-76, NCT00410384), we divided items from the British Isles Lupus Assessment Group (BILAG) disease activity index into core and non-core sets. The core set included 30 of 86 BILAG items across five domains: mucocutaneous (rash, alopecia, mucosal ulcers), haematological (thrombocytopenia, haemolytic anaemia), arthritis, serositis and nephritis. We analysed baseline disease activity and treatment response at week 52.
Results
We analysed 1526 patients. At baseline, the core set captured the majority of disease activity captured using the full BILAG (1426/1526 (93.5%) of patients). Cutaneous vasculitis, non-haemolytic anaemia, leucopenia, dyspnoea and fatigue were the most common non-core set manifestations present. At week 52, about 1278/1426 (89.5%) patients demonstrated concordant treatment responses measured with the core set only versus all BILAG items. Discordance was again driven by cutaneous vasculitis, non-haemolytic anaemia, leucopenia and fatigue.
Conclusions
Limiting the classification of treatment response to a core set of domains has the potential to simplify SLE RCT endpoints without a significant negative impact on patient inclusion or responder classification.
Keywords: Clinical Trial; Lupus Erythematosus, Systemic; Outcome Assessment, Health Care; Severity of Illness Index
WHAT IS ALREADY KNOWN ON THIS TOPIC
Disease activity measures used in legacy SLE clinical trial endpoints capture an extensive list of lupus manifestations, including many that are rare or difficult to measure precisely.
As part of the development of a novel outcome measure for lupus clinical trials, a smaller ‘core set’ of disease activity domains will be measured to determine treatment response, based on their importance and suitability for the clinical trial context; however, the impact of restricting disease activity assessment to this core set in a randomised controlled trial (RCT) setting is unknown.
WHAT THIS STUDY ADDS
The majority of disease activity and patients enrolled in contemporary SLE RCTs may be captured by a smaller core set of lupus manifestations than traditionally measured.
Treatment responses were overall comparably captured using a core set of lupus domains compared with all British Isles Lupus Assessment Group activity, with discordance driven by cutaneous vasculitis, non-haemolytic anaemia, leucopenia and fatigue.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
Limiting the classification of treatment response to a core set of disease activity domains has the potential to simplify SLE RCT endpoints without a significant negative impact on patient inclusion or responder classification.
These results will inform the ongoing development of novel endpoints for SLE RCTs.
Introduction
The assessment of disease activity and treatment response in SLE (lupus) randomised controlled trials (RCTs) is complicated by the heterogeneous nature of the disease. Manifestations of SLE are diverse, can affect any organ in the body and can occur in any combination.1 Further, disease course is highly variable,2 and there are no reliable biomarkers that universally correlate to fluctuations in disease activity.3 4 This has challenged the interpretability of primary efficacy endpoints used in SLE RCTs, which have been identified as contributors to recurrent failures to demonstrate drug efficacy.5,8
Disease activity measures such as the British Isles Lupus Assessment Group (BILAG) index9 10 and the SLE Disease Activity Index (SLEDAI)11,13 underpin the legacy treatment response measures used in SLE clinical trials, the BILAG-based Composite Lupus Assessment (BICLA)14 and the SLEDAI-based SLE Responder Index (SRI).15 The designs of these indices attempt to capture the broad spectrum of disease features that can potentially manifest in active SLE; however, they are limited in their ability to assess individual disease features in depth. This lack of granularity has made it difficult to interpret RCT results in SLE and has impeded drug approval.5 16 Moreover, we previously reported that many organ domains of SLE activity are rarely present in clinical trial populations,17 a finding also reported in a large real-world cohort.18 These considerations lead to the concept of a limited set of core domains having greater utility in the setting of a clinical trial.
The Treatment Response Measure for Systemic Lupus Erythematosus (TRM-SLE) project is an ongoing international collaboration between clinicians, patients and industry partners aiming to develop a novel clinical outcome assessment to be used in an SLE RCT context to measure response to treatment.19 TRM-SLE proposes a multidomain instrument based on prioritising detailed measurement of a succinct ‘core set’ of disease activity domains: mucocutaneous (rash, alopecia, mucosal ulcers), haematological (thrombocytopenia, haemolytic anaemia), arthritis, serositis and nephritis. These domains were selected via a modified Delphi consensus study involving both clinician and patient experts, based on importance to patients, appropriateness in capturing active disease, typical representation in SLE RCT cohorts and measurability in a clinical trial setting.20 However, by prioritising detailed measurement of select disease manifestations over breadth, clinically important disease features indicating SLE disease activity could be overlooked, meaning responses in some aspects of disease might not be captured using this core set.
The aim of this study was to investigate whether limiting disease activity assessment to a ‘core set’ of organ manifestations captures patient inclusion and treatment responses in SLE RCTs comparably to existing outcome measures that assess the full breadth of disease activity.
Methods
Study population
We conducted an analysis of pooled data from two phase III SLE RCTs comparing belimumab to placebo on a background of standard of care SLE therapy: BLISS-52 (NCT00424476, n=865)21 and BLISS-76 (NCT00410384, n=819).22 Data were accessed via the Clinical Study Data Request consortium following protocol approval from GlaxoSmithKline (Uxbridge, UK).
As described in the original publications, entry criteria for the BLISS studies included (a) age ≥18, (b) SLE as per 1997 American College of Rheumatology Classification Criteria, (c) Safety of Estrogens in Lupus Erythematosus National Assessment–SLE Disease Activity Index (SELENA-SLEDAI) ≥6, (d) positive serology and (e) stable non-biologic medication regime for at least 30 days prior to enrolment. Exclusion criteria included severe active lupus nephritis or neuropsychiatric lupus. In our study, the pooled population was restricted to patients who also had baseline BILAG A or BILAG B disease activity (n=1526), more closely matching recent RCT enrolment approaches (see online supplemental figure 1).
The BLISS trials used three key disease activity measures: the BILAG index,10 the SELENA-SLEDAI11 and the Physician Global Assessment (PGA).23 The primary endpoint of both BLISS trials was a composite of these three assessments: the SRI-4, defined as ≥4 point reduction from baseline SLEDAI score; no new BILAG A or no more than one new BILAG B score; and no worsening of PGA by ≥0.3 points.24 For the purpose of the current study, we analysed data at two time points: week 0 (baseline) and week 52 (time of primary endpoint assessment).
As the definitions and thresholds for domain activity in the TRM-SLE instrument were not captured in the BLISS trials, we used BILAG index data to act as placeholder definitions for the TRM core set disease activity domains. The BILAG was chosen to define core versus non-core set activity and responses, due to its extensive coverage of SLE disease features including all core set activity. In contrast, the SLEDAI does not capture haemolytic anaemia, and the PGA only captures disease activity at a global level, so were not suitable for this purpose. BILAG has recently been shown in post hoc analysis to have adequately captured treatment effects in the BLISS studies.25 The version of the BILAG index used in the BLISS trials comprises 86 items separated into eight domains: general, mucocutaneous, neurological, musculoskeletal, cardiorespiratory, vasculitis, renal and haematological.10 BILAG domains were considered active if they were graded A (severe activity) or B (moderate activity). 30 BILAG items spanning five BILAG organ domains were allocated to represent the TRM core set, based on the manifestations meeting consensus for inclusion in TRM-SLE in a previous Delphi consensus process which resulted in domain selection (rash, alopecia, mucosal ulcers, thrombocytopenia, haemolytic anaemia, arthritis, serositis and nephritis).20 The remaining 56 BILAG items were categorised as non-core set manifestations. This meant four BILAG domains included both TRM core and non-core set items (mucocutaneous, musculoskeletal, cardiorespiratory, haematological); three BILAG domains included non-core set items only (neurological, vasculitis, general); and one BILAG domain included core set items only (renal) (see online supplemental table 1). Core and non-core set domain grades were assigned according to the BILAG appendix and glossary.10 For BILAG domains containing a mix of core set and non-core set items, this included generating separate BILAG grades for core and non-core set domain activities. Specific item allocation is included in the online supplemental table 1.
We first compared the baseline prevalence of core set and non-core set activities at both domain and patient levels. We then assessed treatment response at week 52, defined according to disease activity improvement. At a domain level, improvement was defined as a step-down from baseline BILAG A to B, C or D, or a step-down from BILAG B to C or D at week 52. Improvement at the patient level was defined according to the following categories: non-responder (no domains active at baseline improved by week 52), incomplete responder (at least one, but not all domains active at baseline, improved by week 52) and complete responder (all domains active at baseline improved by week 52) (figure 1). To determine whether use of the TRM-SLE core set impacted on discriminant validity, we also examined complete responders versus non-responders for the 10 mg/kg dose of belimumab compared with placebo.
Figure 1. Illustration of responder classification possibilities in a single patient. BILAG, British Isles Lupus Assessment Group.
Statistical analysis
All statistical analyses were performed using Stata V.18.0 (College Station, Texas, USA). Patient characteristics, baseline disease activity and responder status at week 52 were summarised as either frequencies and percentages (n, %) for categorical variables or means with corresponding SDs for continuous variables. For the analysis of discriminant validity, univariable logistic regression analysis was used to calculate ORs for inferential analysis of responder and non-responder groups. ORs and p values were reported with 95% CIs.
Patient and public involvement
Patients or the public were not involved in the design, or conduct, or reporting, or dissemination plans of our research.
Results
Baseline demographics and disease activity
Data from 1526 patients were studied. Baseline demographics and disease activity are summarised in the online supplemental table 2. Mean (SD) age was 37.99 (11.6) years and patients were predominantly female (93.9%) and of white ethnicity (48.8%). At enrolment, 990 (64.9%) patients were on antimalarial therapy, 746 (48.9%) on immunosuppressants and 1312 (86.0%) on glucocorticoids. Mean (SD) SELENA-SLEDAI was 9.98 (3.7), with 832 (54.5%) patients having a SELENA-SLEDAI ≥10. Mean (SD) PGA score was 1.46 (0.5), with 1338 (87.7%) patients having a PGA >1. As per our study inclusion criteria, all patients analysed had at least one organ domain with BILAG A or B activity. At baseline, BILAG A activity was present in 263 (17.2%) patients, while 1263 (82.8%) patients had BILAG B activity only. Regarding the number of active domains per patient at baseline, 561 (36.8%) of patients had disease activity in one organ domain only, 668 (43.8%) had two active domains and 297 (19.5%) had disease activity in three or more domains (see online supplemental table 3). Musculoskeletal (66.0%) and mucocutaneous (64.9%) disease activity was most common, both as a single active domain and in combination (see online supplemental table 3).
Baseline disease activity stratified by core set status
The majority of baseline disease activity in the study population was captured by the TRM-SLE core set. Musculoskeletal and mucocutaneous disease activity was almost entirely (>99%) captured when measuring core set disease features only (figure 2). Renal disease activity was completely captured when measuring core set features only. In contrast, in the cardiorespiratory and haematological domains, non-core set items constituted the majority of baseline disease activity (figure 2). Non-core set disease activity was responsible for 73.8% of cardiorespiratory domain activity, of which 97.8% (44/45) was due to dyspnoea. Similarly, 86.3% of activity in the haematological domain was non-core set, of which 83.3% (195/234) was due to non-haemolytic anaemia (table 1). Three domains (general, vasculitis and neurological) were entirely missing from the core set. In both the general and vasculitis domains, disease activity was driven by a single prominent disease feature: 97.8% (179/183) of general domain activity was attributed to fatigue/malaise/lethargy, and 99.3% (150/151) of vasculitis domain activity was attributed to cutaneous vasculitis (table 1). Neurological disease activity was distributed among multiple disease manifestations; however, the sample with neurological disease activity was overall very small (n=21). Notably, BILAG A grade disease activity was almost entirely captured by the core set. Non-core set BILAG A activity contributed to less than 0.6% of activity in each domain, with the exception of vasculitis (49 patients, 3.2% with major cutaneous vasculitis) (table 1).
Figure 2. Organ domain-level baseline disease activity, stratified by core and non-core sets.
Table 1. Baseline disease activity contributing to BILAG A or B organ domain activity not captured by core set (n=1526).
| Disease activity not captured by core set | Patients with baseline involvement (n, %) |
|---|---|
| Musculoskeletal (n=7) | |
| Myositis* | 7 (0.5) |
| Mucocutaneous (n=13) | |
| Severe panniculitis* Angio-oedema* Subcutaneous nodules Perniotic skin lesions |
1 (0.1) 3 (0.2) 7 (0.5) 2 (0.1) |
| Cardiorespiratory (n=45) | |
| Dyspnoea Progressive chest X-ray changes (lungs or heart) Cardiac arrhythmias including tachycardia Cardiac failure* Pulmonary function fall by >20%* Cytohistological evidence of inflammatory lung disease* |
44 (2.9) 3 (0.2) 2 (0.1) 1 (0.1) 0 (0) 0 (0) |
| Haematological (n=234) | |
| Anaemia without associated haemolysis Leucopenia (total white cell count 1.0–2.5×109/L) Both non-haemolytic anaemia and leucopenia |
195 (12.8) 35 (2.3) 14 (1.0) |
| General (n=183) | |
| Fever*† Weight loss Lymphadenopathy/splenomegaly Fatigue/malaise/lethargy Anorexia/nausea/vomiting |
39 (2.6) 6 (0.4) 47 (3.1) 179 (11.7) 126 (8.3) |
| Neurological (n=21) | |
| Impaired level of consciousness (same or improving) Acute psychosis/delirium/confusion (same or improving) Peripheral or cranial neuropathy (worse or new)* Headaches—severe unremitting (worse or new) Organic depressive illness (worse or new) Organic brain syndrome (worse or new)* All other‡ |
1 (0.1) 4 (0.3) 4 (0.3) 4 (0.3) 9 (0.3) 1 (0.1) 0 (0) |
| Vasculitis (n=151) | |
| Major cutaneous vasculitis including ulcers* Major abdominal crisis due to vasculitis* Thromboembolism excluding stroke (recurrent* or 1st episode) Minor cutaneous vasculitis (nailfold, digital, purpura, urticaria) Superficial phlebitis |
49 (3.2) 0 (0) 0 (0) 150 (9.8) 3 (0.2) |
Item contributing to BILAG A grade activity.
General domain BILAG A requires fever and two other items.
Refers to remaining non-core set category A and B manifestations per Hay et al10: grand mal seizure, stroke or stroke syndrome, aseptic meningitis, mononeuritis multiplex, ascending or transverse myelitis, chorea, cerebellar ataxia, disc swelling or cytoid bodies.
BILAG, British Isles Lupus Assessment Group.
Baseline disease activity was completely captured by the core set in 1426/1526 (93.5%) patients. Core set activity occurred in a single domain only in 713 (46.7%) patients in the study population, while 645 (42.3%) had baseline activity in two core set domains, and less than 5% had activity in at least three core set domains (see online supplemental table 3). Only 100 (6.6%) patients had solely non-core set BILAG A or B disease activity at baseline, and thus would have been excluded from the evaluation of response if improvement assessment was limited to the core set. Of these 100 patients, 84% were only active in one non-core set domain at baseline. The most frequently active domains in this group were haematological (46%), vasculitis (29%) and general (20%).
Disease activity improvement at week 52
We compared response as defined by core set features only (core set response) to response defined by measurement of all 86 BILAG items (BILAG response) (table 2). This analysis excluded the 100 patients who had no core set activity at baseline. Demographic and clinical characteristics of these excluded patients are shown alongside the total study population in the online supplemental table 2; patients within this excluded subset were more likely to be of non-white race and more likely to have enrolled in the study on an immunosuppressant and with BILAG 1B activity. At a patient level, BILAG response rates were as follows: 678 (47.5%) improved in all active domains (complete responders), 424 (29.7%) improved in some but not all active domains (incomplete responders) and 324 (22.7%) did not demonstrate any improvement from baseline activity (non-responders). Response defined by core set features only was as follows: 754 (52.9%) complete responders, 276 (19.4%) incomplete responders and 396 (27.8%) non-responders. Thus, 1278 (89.5%) of patients had concordant responder status and 148 (10.4%) participants had discordant responder status when classified based on measurement of core set features only versus the full set of BILAG disease features (table 2).
Table 2. Treatment response at week 52 defined by TRM core set domains compared with BILAG (n=1426).
| TRM core set | Total (n) |
||||
|---|---|---|---|---|---|
| Non-responder | Incomplete responder | Complete responder | |||
| BILAG, n (%) | Non-responder | 324 (100%) | 0 | 0 | 324 |
| Incomplete responder | 72 (16.98%) (95% CI 13.53% to 20.90%) |
276 (65.09%) (95% CI 60.34% to 69.63%) |
76 (17.92%) (95% CI 14.39% to 21.91%) |
424 | |
| Complete responder | 0 | 0 | 678 (100%) | 678 | |
Non-responder defined as patients who did not improve in any organ domain active at baseline. Incomplete responder defined as patients who improved in some organ domains active at baseline but not all. Complete responder defined as patients who improved in all organ domains active at baseline.
BILAG, British Isles Lupus Assessment Group; TRM, treatment response measure.
We further investigated this population of patients with discordant core set and BILAG responder status (table 3). Within this group, 72 (48.6%) patients were non-responders based on the core set but had improvement in one or more non-core set domains at week 52. Of these patients, the most common non-core set domains that improved were vasculitis (predominantly cutaneous vasculitis), haematological (non-haemolytic anaemia and/or leucopenia) and general (predominantly fatigue/malaise/lethargy). Conversely, 76 (51.4%) patients were complete responders based on the core set but did not improve in one or more non-core set domains active at baseline. Of these patients, no improvement in haematological activity (non-haemolytic anaemia and/or leucopenia) was the most common finding (80.3%).
Table 3. Baseline non-core set domain activity in patients with week 52 incomplete BILAG improvement, stratified by core set non-responder and core set complete responder.
| Non-core set domains and items | Non-core set domain responses in core set non-responders* (n=72) | Non-core set domain non-responses in core set complete responders† (n=76) |
|---|---|---|
| Musculoskeletal (myositis) | 0 (0) | 0 (0) |
| Mucocutaneous (angio-oedema, panniculitis, nodules, perniotic skin lesions) | 1 (1.4) | 4 (5.3) |
| Cardiorespiratory (dyspnoea, other non-serositis features) | 7 (9.7) | 1 (1.3) |
| Haematological (non-haemolytic anaemia, leucopenia) | 19 (26.4) | 61 (80.3) |
| Neurological (neuropsychiatric lupus symptoms) | 5 (7.0) | 3 (4.0) |
| Vasculitis (cutaneous vasculitis) | 23 (32.0) | 3 (4.0) |
| General (fatigue/malaise/lethargy) | 30 (41.7) | 5 (6.6) |
Values in this column denote non-core set domains which improved at week 52.
Values in this column denote non-core set domains which did not improve at week 52.
BILAG, British Isles Lupus Assessment Group.
Finally, we analysed differences in rates of complete response versus non-response for the 10 mg/kg dose of belimumab compared with placebo, using the two cohorts. Importantly, this analysis addresses BILAG grade improvement and not other components of non-responder imputation such as medication rule violations. The analysis shows that the TRM core set had at least equivalent discriminant validity (p=0.042) compared with BILAG (p=0.058) (see table 4).
Table 4. Complete responder versus non-responder rates between placebo and belimumab 10 mg/kg treatment arms.
| Placebo group (n=486) |
Belimumab 10 mg/kg (n=471) |
Univariable logistic regression OR (95% CI), P value |
|
|---|---|---|---|
| TRM core set | |||
| Non-responder | 150 (30.9) | 123 (26.1) | 1.00 |
| Responder | 237 (48.8) | 264 (56.1) | 1.36 (1.01 to 1.83), 0.042 |
| BILAG | |||
| Non-responder | 121 (24.9) | 99 (21.0) | 1.00 |
| Responder | 210 (43.2) | 241 (51.2) | 1.35 (0.99 to 1.83), 0.058 |
BILAG, British Isles Lupus Assessment Group; TRM, treatment response measure.
Discussion
SLE is a heterogeneous disease that is characteristically associated with a broad spectrum of manifestations, ranging in frequency from almost ubiquitous to extremely rare. Despite this, there is a paucity of literature that examines the trajectory of treatment responses between different manifestations in patients with multiorgan involvement at baseline. Multiple studies of both registry and clinical trial datasets indicate that the vast majority of disease activity in SLE is captured in a relatively small number of domains,17 18 giving rise to the concept that measuring disease activity and responses to treatment in the setting of clinical trials could be simplified without losing sensitivity. The novel TRM-SLE outcome assessment, which is being developed by an international task force,19 has arrived at a core set of domains for measurement in the setting of SLE trials via a comprehensive consensus process.20 This has the potential to simplify trial outcome measurement, allow detailed assessment of the active manifestations most relevant for clinical trial populations and thus improve the interpretability of SLE clinical trials. However, determining whether limiting measurement to this core set of domain risks losing significant clinical meaning, as well as sensitivity and specificity, is required. Here, we used the BLISS trial datasets to evaluate the effect of limiting measurement to the core set of domains included in the TRM-SLE outcome measure development compared with the more comprehensive set of domains described by the BILAG index.
The results from this study indicate that the vast majority of clinically important disease activity at baseline in the BLISS trials was captured by a limited core set of organ domains. Musculoskeletal and mucocutaneous disease activity was comprehensively captured in the core set, with only rare events in <1% of patients falling outside the core set of domains in these two organ systems. Severe disease activity (BILAG A) was rarely missed by the core set, with the notable exception of major cutaneous vasculitis. At the patient level, the core set captured all but 7% of patients at baseline; this minority would thus not be eligible for assessment of treatment response and would be excluded from a trial in which eligibility was defined by the TRM core set only. The majority of this patient group was active in a single domain, mostly the non-core set haematological or general domains, and had overall less severe disease at baseline.
Importantly, most baseline disease activity features not captured by the core set are unsuitable for assessing treatment response in SLE. Common non-core set features including non-haemolytic anaemia, dyspnoea and fatigue/malaise/lethargy were considered in the TRM-SLE domain selection Delphi process, but were judged to be inappropriate for measuring disease activity improvement and thus did not meet consensus for inclusion in the final instrument.20 An important exception is the omission of cutaneous vasculitis, which was present at baseline in 13% of the BLISS population. While acknowledged to be important and appropriate for measurement in an SLE RCT, vasculitis was not included in the TRM-SLE core set due to the lack of an available outcome measure suitable for assessing treatment response. Our findings therefore indicate that development of methods to measure cutaneous vasculitis in the setting of SLE is an area of research need, so that vasculitis disease activity can be captured in a future iteration of the TRM-SLE outcome measure.
Clearly, the utility of any outcome measure in a trial rests on its ability to detect clinically meaningful changes in response to therapy. Treatment response at week 52 measured using only the core set was highly concordant with response measured using all 86 BILAG disease features. Furthermore, the ability of the core set to discriminate between the belimumab 10 mg/kg treatment arm and placebo group was comparable using both approaches. Discordance in response status between the core set and BILAG definitions was driven predominantly by non-haemolytic anaemia and/or leucopenia, predominantly in the circumstance that core set domains improved, but non-haemolytic anaemia and/or leucopenia did not. Non-haemolytic anaemia in SLE is most commonly anaemia of chronic disease,26 which often persists despite clinical resolution of a flare, and so may be less appropriate to be used to quantify treatment response in an RCT context. Other causes of non-haemolytic anaemia in patients with SLE include iron deficiency, bleeding or renal anaemia, none of which reflect haematological involvement directly attributable to SLE. Furthermore, resolving leucopenia, although understood to reflect haematological disease improvement, has shown minimal impact on overall patient outcomes27,29 and is often confounded by other factors such as medications or infection. Interestingly, fatigue was also a common cause of discordance; while undoubtedly important in SLE, fatigue is an unreliable measure of SLE disease activity states as it does not reliably respond to immunomodulatory therapy.30 Importantly, however, cutaneous vasculitis was also a source of discordance, in that response to therapy would not be captured by the TRM-SLE core set; of note, however, this would have resulted in responder misclassification (ie, core set response but vasculitis non-response) in only three patients and would have failed to detect response in 23 (table 3).
There were some limitations to this study. As this study was undertaken concurrently with the instrument development phase of TRM-SLE, analysis of core and non-core domain activity and treatment response was based on the BILAG index rather than the continuous variable-based metrics being developed in TRM-SLE. Nevertheless, the aim of this study was to assess the impact of limiting the domains in which disease activity is measured, a goal for which we consider the BILAG index data appropriate. Second, we acknowledge the selection bias created by the trial inclusion criteria, which excluded severe neuropsychiatric lupus and lupus nephritis. Third, the BLISS trials used the original BILAG index,10 which contained disease features that were removed in the 2004 revision as they were considered too ambiguous or poor indicators of active disease.9 Most notably, vasculitis does not exist in the BILAG-2004 as an individual domain and instead select disease features are redistributed throughout other organ domains. The BILAG-2004 also includes gastrointestinal and ophthalmic lupus which would have both been classified as non-core set domains and may have provided additional insight regarding the comprehensiveness of the core set, although manifestations of these domains are rare in the context of clinical trials.17
In conclusion, this study shows that measuring fewer disease activity domains can adequately capture the vast majority of patient-level disease activity and treatment response in an SLE RCT. Despite this, we identified cutaneous vasculitis as a domain that may require further research to derive a validated method of measurement. By limiting the number of domains assessed, it is more feasible to measure each individual domain in greater detail, for example, using continuous variables,3 which may be simpler and more reliable. Results from this study support the results of the Delphi expert consensus process that resulted in the core set of domains that will be measured in the TRM-SLE outcome measure.
Supplementary material
Acknowledgements
The authors thank GlaxoSmithKline for providing data from the BLISS-52 (NCT00424476) and BLISS-76 (NCT00410384) trials through the Clinical Study Data Request consortium, and all patients who participated in the trials.
Footnotes
Funding: The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Provenance and peer review: Not commissioned; externally peer reviewed.
Patient consent for publication: Not applicable.
Ethics approval: This study involves human participants and was approved by the Monash University Human Research Ethics Committee (MUHREC Project ID 41946). Participants gave informed consent to participate in the study before taking part.
Data availability free text: This publication is based on research using data from data contributor GlaxoSmithKline that have been made available through Vivli, Inc. Vivli has not contributed to or approved, and is not in any way responsible for, the contents of this publication. Data access is subject to approval by GlaxoSmithKline (Uxbridge, UK) via the Clinical Study Data Request consortium.
Patient and public involvement: Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.
Data availability statement
Data may be obtained from a third party and are not publicly available.
References
- 1.Ameer MA, Chaudhry H, Mushtaq J, et al. An Overview of Systemic Lupus Erythematosus (SLE) Pathogenesis. Classification, and Management Cureus. 2022;14 doi: 10.7759/cureus.30330. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Tselios K, Gladman DD, Touma Z, et al. Disease course patterns in systemic lupus erythematosus. Lupus (Los Angel) 2019;28:114–22. doi: 10.1177/0961203318817132. [DOI] [PubMed] [Google Scholar]
- 3.Connelly K, Kandane-Rathnayake R, Hoi A, et al. Associations of improvement in laboratory tests with clinical outcomes in patients with active systemic lupus erythematosus: a multinational longitudinal cohort study. Lancet Rheumatol. 2022;4:e831–41. doi: 10.1016/S2665-9913(22)00307-1. [DOI] [PubMed] [Google Scholar]
- 4.Gladman DD, Urowitz MB, Keystone EC. Serologically active clinically quiescent systemic lupus erythematosus: a discordance between clinical and serologic features. Am J Med. 1979;66:210–5. doi: 10.1016/0002-9343(79)90529-1. [DOI] [PubMed] [Google Scholar]
- 5.Connelly K, Golder V, Kandane-Rathnayake R, et al. Clinician-reported outcome measures in lupus trials: a problem worth solving. Lancet Rheumatol. 2021;3:e595–603. doi: 10.1016/S2665-9913(21)00119-3. [DOI] [PubMed] [Google Scholar]
- 6.Lorenzo-Vizcaya A, Isenberg DA. Clinical trials in systemic lupus erythematosus: the dilemma-Why have phase III trials failed to confirm the promising results of phase II trials? Ann Rheum Dis. 2023;82:169–74. doi: 10.1136/ard-2022-222839. [DOI] [PubMed] [Google Scholar]
- 7.Gatto M, Saccon F, Zen M, et al. Success and failure of biological treatment in systemic lupus erythematosus: A critical analysis. J Autoimmun. 2016;74:94–105. doi: 10.1016/j.jaut.2016.06.014. [DOI] [PubMed] [Google Scholar]
- 8.Peña Y, Tse K, Hanrahan LM, et al. Establishing Consensus Understanding of the Barriers to Drug Development in Lupus. Ther Innov Regul Sci. 2020;54:1159–65. doi: 10.1007/s43441-020-00134-2. [DOI] [PubMed] [Google Scholar]
- 9.Isenberg DA, Rahman A, Allen E, et al. BILAG 2004. Development and initial validation of an updated version of the British Isles Lupus Assessment Group’s disease activity index for patients with systemic lupus erythematosus. Rheumatology (Sunnyvale) 2005;44:902–6. doi: 10.1093/rheumatology/keh624. [DOI] [PubMed] [Google Scholar]
- 10.Hay EM, Bacon PA, Gordon C, et al. The BILAG index: a reliable and valid instrument for measuring clinical disease activity in systemic lupus erythematosus. QJM: An International Journal of Medicine. 1993;86:447–58. [PubMed] [Google Scholar]
- 11.Petri M, Kim MY, Kalunian KC, et al. Combined oral contraceptives in women with systemic lupus erythematosus. N Engl J Med. 2005;353:2550–8. doi: 10.1056/NEJMoa051135. [DOI] [PubMed] [Google Scholar]
- 12.Dafna DG, Dominique I, Murray BU. Systemic lupus erythematosus disease activity index 2000. J Rheumatol. 2002;29:288. [PubMed] [Google Scholar]
- 13.Bombardier C, Gladman DD, Urowitz MB, et al. Derivation of the SLEDAI. A disease activity index for lupus patients. The Committee on Prognosis Studies in SLE. Arthritis Rheum. 1992;35:630–40. doi: 10.1002/art.1780350606. [DOI] [PubMed] [Google Scholar]
- 14.Wallace DJ, Kalunian K, Petri MA, et al. Efficacy and safety of epratuzumab in patients with moderate/severe active systemic lupus erythematosus: results from EMBLEM, a phase IIb, randomised, double-blind, placebo-controlled, multicentre study. Ann Rheum Dis. 2014;73:183–90. doi: 10.1136/annrheumdis-2012-202760. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Furie RA, Petri MA, Wallace DJ, et al. Novel evidence‐based systemic lupus erythematosus responder index. Arthritis & Rheumatism . 2009;61:1143–51. doi: 10.1002/art.24698. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Mahieu MA, Strand V, Simon LS, et al. A critical review of clinical trials in systemic lupus erythematosus. Lupus (Los Angel) 2016;25:1122–40. doi: 10.1177/0961203316652492. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Connelly K, Vettivel J, Golder V, et al. Measurement of specific organ domains in lupus randomized controlled trials: a scoping review. Rheumatology (Oxford) 2022;61:1341–53. doi: 10.1093/rheumatology/keab777. [DOI] [PubMed] [Google Scholar]
- 18.Carter LM, Gordon C, Yee C-S, et al. Easy-BILAG: a new tool for simplified recording of SLE disease activity using BILAG-2004 index. Rheumatology (Oxford) 2022;61:4006–15. doi: 10.1093/rheumatology/keab883. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Connelly K, Eades LE, Koelmeyer R, et al. Towards a novel clinical outcome assessment for systemic lupus erythematosus: first outcomes of an international taskforce. Nat Rev Rheumatol. 2023;19:592–602. doi: 10.1038/s41584-023-00993-7. [DOI] [PubMed] [Google Scholar]
- 20.Connelly K, Koelmeyer R, Ayton D, et al. Domains for inclusion in a novel Treatment Response Measure for Systemic Lupus Erythematosus (TRM-SLE): results of a modified Delphi study. Lupus Sci Med . 2025;12:e001484. doi: 10.1136/lupus-2024-001484. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Navarra SV, Guzmán RM, Gallacher AE, et al. Efficacy and safety of belimumab in patients with active systemic lupus erythematosus: a randomised, placebo-controlled, phase 3 trial. The Lancet. 2011;377:721–31. doi: 10.1016/S0140-6736(10)61354-2. [DOI] [PubMed] [Google Scholar]
- 22.Furie R, Petri M, Zamani O, et al. A phase III, randomized, placebo-controlled study of belimumab, a monoclonal antibody that inhibits B lymphocyte stimulator, in patients with systemic lupus erythematosus. Arthritis Rheum. 2011;63:3918–30. doi: 10.1002/art.30613. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Liang MH, Socher SA, Larson MG, et al. Reliability and validity of six systems for the clinical assessment of disease activity in systemic lupus erythematosus. Arthritis Rheum. 1989;32:1107–18. doi: 10.1002/anr.1780320909. [DOI] [PubMed] [Google Scholar]
- 24.Luijten K, Tekstra J, Bijlsma JWJ, et al. The Systemic Lupus Erythematosus Responder Index (SRI); a new SLE disease activity assessment. Autoimmun Rev. 2012;11:326–9. doi: 10.1016/j.autrev.2011.06.011. [DOI] [PubMed] [Google Scholar]
- 25.Parodis I, Lindblom J, Palazzo L, et al. Outcomes of patients with systemic lupus erythematosus treated with belimumab: a post hoc efficacy analysis of five phase III clinical trials by British Isles Lupus Assessment Group-based Combined Lupus Assessment criteria. RMD Open. 2025;11:e005444. doi: 10.1136/rmdopen-2025-005444. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Giannouli S, Voulgarelis M, Ziakas PD, et al. Anaemia in systemic lupus erythematosus: from pathophysiology to clinical assessment. Ann Rheum Dis. 2006;65:144–8. doi: 10.1136/ard.2005.041673. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Carli L, Tani C, Vagnani S, et al. Leukopenia, lymphopenia, and neutropenia in systemic lupus erythematosus: Prevalence and clinical impact--A systematic literature review. Semin Arthritis Rheum. 2015;45:190–4. doi: 10.1016/j.semarthrit.2015.05.009. [DOI] [PubMed] [Google Scholar]
- 28.Fayyaz A, Igoe A, Kurien BT, et al. Haematological manifestations of lupus. Lupus Sci Med . 2015;2:e000078. doi: 10.1136/lupus-2014-000078. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Martínez-Baños D, Crispín JC, Lazo-Langner A, et al. Moderate and severe neutropenia in patients with systemic lupus erythematosus. Rheumatology (Oxford) 2006;45:994–8. doi: 10.1093/rheumatology/kel016. [DOI] [PubMed] [Google Scholar]
- 30.Eudy AM, Rogers JL, Corneli A, et al. Intermittent and Persistent Type 2 lupus: patient perspectives on two distinct patterns of Type 2 SLE symptoms. Lupus Sci Med . 2022;9:e000705. doi: 10.1136/lupus-2022-000705. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Data may be obtained from a third party and are not publicly available.