Which is the best SLE activity index for clinical trials?

Abstract

Following the advent of molecular targeted drugs, a paradigm shift in treatment similar to that in rheumatoid arthritis has been expected in the treatment of systemic lupus erythematosus (SLE), but clinical trials for drugs that many specialists believed to be effective have failed repeatedly. The causes are not simple, but include the heterogeneity of SLE, inclusion criteria, lack of appropriate disease activity measures, and relapse criteria. This review outlines the disease activity indices used in SLE, discusses their advantages and disadvantages, and describes the ideal activity index.

Keywords:

• Activity index
• BILAG
• clinical trial
• SLEDAI
• systemic lupus erythematosus

Introduction

There are many disease activity indices of SLE [1,2], but most are complicated because they were created based on clinical trials and clinical studies. In daily clinical practice, activity is evaluated by comprehensively judging symptoms, organ damage, and activity markers such as complement and anti-DNA antibodies.

The disease activity indices are separated into global indices and organ-specific indices. The former representatives are SLEDAI (SLE disease activity index) [3–5], SLAM (Systemic Lupus Activity Measure) [6] and ECLAM (European Consensus Lupus Activity Measurement) [7], and the latter representatives are BILAG [8–10], CLASI [11], and renal activity score [12]. BILAG can also be used as a comprehensive measure. Each has its own strengths and weaknesses, and knowledge of their characteristics is necessary. In recent clinical trials, evaluation using SLEDAI and BILAG and their combinations, such as SRI [13] and BICLA [14], has been mainstream. Recent SLE clinical trials and measurement outcomes of primary endpoints are summarized in Table 1. None of the current indices can satisfactorily discriminate responders and non-responders. In this review, I focus mainly on global disease activity indices and discuss which activity index is best for successful SLE clinical trials.

Table 1. Recent clinical trials of molecular targeted drugs for SLE (non-organ specific).

Drug	Target	Outcome for primary endpoint	Primary endpoint satisfied?	References (author, year)	Phase/study name
Rituximab	CD20	BILAG	Failure	Merrill, 2010 [15]	II/III/EXPLORER
Abatacept	CD80/86	BILAG	Failure	Merrill, 2010 [16]	IIb
Atacicept	BAFF/APRIL	BILAG	Failure	Isenberg, 2015 [17]	IIb/APRIL-SLE
Belimumab	BAFF	SELENA-SLEDAI, SFI	Failure	Wallace, 2009 [18]	II
		SRI-4	Success	Navarra, 2011 [19]	III/BLISS-52
		SRI-4	Success	Furie, 2011 [20]	III/BLISS-76
Tabalumab	BAFF	SRI-5	Failure	Isenberg, 2016 [21]	III/ILLUMINATE1
		SRI-5	Success	Merrill, 2016 [22]	III/ILLUMINATE2
Epratuzumab	CD22	BICLA	Failure	Wallace, 2014 [14]	IIb/EMBLEM
		BICLA	Failure	Clowse, 2016 [23]	III/EMBODY1&2
PF-04236921	IL-6	SRI-4	Failure	Wallace, 2014 [24]	II
Rontalizumab	IFN-α	BILAG-2004	Failure	Kalunian, 2016 [25]	II/ROSE
Sifalimumab	IFN-α	SRI-4	Success	Khamashta, 2016 [26]	IIb
Anifrolumab	type I IFN-R	SRI-4	Failure	Furie, 2019 [27]	III/TULIP 1
		BICLA	Success	Furie, 2020 [28]	III/TULIP 2

SLEDAI

In addition to the original SLEDAI [3], there are improved versions such as SLEDAI-2K [4] and SELENA-SLEDAI [5]. At present, many clinical trials use SLEDAI-2K or SELENA-SLEDAI. The basis of the original SLEDAI is shown in Table 2. All three representative SLEDAIs have the same weighting for items and organ damage, but have different definitions for each item. Their comparison is shown in Table 3. In each case, judgment is made based on symptoms and findings over the last 10 days. The main differences are proteinuria and skin rash/stomatitis/hair loss. Proteinuria is scored when it is 0.5 g/day or more in any SLEDAI system, but the original SLEDAI and SELENA-SLEDAI can score new-onset proteinuria even when it is less than 0.5 g/day. In addition, skin rash, stomatitis (mucosal ulcer), and hair loss are supposed to be scored only when they are new or relapsed in the original SLEDAI, but they are scored even when they are not new or relapsed in SLEDAI-2K and SELENA-SLEDAI.

Table 2. Items and their weight in SLEDAI (original).

Weight	Descriptor
8	Seizure
8	Psychosis
8	Organic brain syndrome
8	Visual disturbance
8	Cranial nerve disorder
8	Lupus headache
8	Cerebrovascular accidents
8	Vasculitis
4	Arthritis (>2 joints)
4	Myositis
4	Urinary casts
4	Hematuria (>5 RBC/HPF)
4	Proteinuria
4	Pyuria (>5 WBC/HPF)
2	New rash
2	Alopecia
2	Mucosal ulcers
2	Pleurisy
2	Pericarditis
2	Low complement
2	Increased DNA binding
1	Fever (>38.5 C)
1	Thrombocytopenia (<100,000/ul)
1	Leukopenia (<3000/ul)

Sum of all positive scores is the SLEDAI score.

Table 3. Comparison of the three different SLEDAI.

Descriptor	SLEDAI	SELENA-SLEDAI	SLEDAI-2K
Seizure		Added the exclusion of seizure due to irreversible CNS change
Visual disturbance		Included scleritis or episcleritis
Cranial nerve disorder		Included vertigo due to lupus
Cerebrovascular accident		Added the exclusion of hypertensive causes
Arthritis	>2 joints	>2 joints	≥2 joints
Proteinuria	>0.5 g/day. New onset or recent increase of >0.5 g/day.	New onset or recent increase of >0.5 g/day	>0.5 g/day
Rash, alopecia, mucosal ulcers	Only new onset or recurrence	Included ongoing symptoms	Included ongoing symptoms
Pleurisy Pericarditis	Subjective AND objective findings	Subjective OR objective findings	Subjective AND objective findings

The biggest advantage of SLEDAI is that it is simple to score, enabling its use in clinical practice. Another advantage of SLEDAI is that it can be used for retrospective studies. The disadvantage is that only the presence of each item is scored and the severity of each item cannot be evaluated. For example, platelet counts of 1000/ul and 90,000/ul are both 1 point by SLEDAI. Second, neurological and psychiatric symptoms are slightly overemphasized, and neuropsychiatric (NP)-SLE is scored highly. On the other hand, symptoms, such as lupus headache, which is often difficult to differentiate from normal headache, are scored as 8 points. Pyuria and hematuria (4 points each) need to be evaluated carefully or the activity score will increase, leading to unstable activity judgment. Arthritis and myositis are 4 points each, whereas thrombocytopenia and serositis are only 1 and 2 points, respectively, which is not considered clinically accurate. There is also a problem that even when severe conditions, such as hemolytic anemia, lupus enteritis, transverse myelitis, alveolar hemorrhage, and pulmonary hypertension are present, they are evaluated as having no activity because there are no items for them.

Flare using SELENA-SLEDAI is finely defined [5]. Severe flare is defined as fulfilling either of the following six items. 1. Change in SELENA-SLEDAI score to greater than 12; 2. Development or deterioration of the following conditions: CNS, vasculitis, nephritis, myositis, Plt <60,000/ul or hemolytic anemia (Hb < 7 g/dL or decrease in Hb > 3 g/dL), requiring doubled prednisolone (PSL), PSL increase to >0.5 mg/kg/day, or hospitalization; 3. Increase in PSL to >0.5 mg/kg/day; 4. New administration of cyclophosphamide, azathioprine, or methotrexate for SLE activity; 5. Hospitalization for SLE activity; or 6. Increase in physician’s global assessment (PGA) score to >2.5. The definition of mild or moderate flare is defined separately [5].

BILAG (British Isles Lupus Assessment Group)

BILAG differs from SLEDAI in that it can evaluate activity for each organ, whereas SLEDAI expresses overall disease activity by the total score. The classic BILAG index [8] has been revised several times, and BILAG-2004, which was also revised in 2009, is currently used [9,10]. In the revised BILAG-2004, there are a total of 97 items in 9 domains, i.e. constitutional, mucocutaneous, neuropsychiatric, musculoskeletal, cardiorespiratory, gastrointestinal, ophthalmic, renal, and hematological domains. For each of the 97 items, 0 (not present), 1 (improving), 2 (same), 3 (worse), or 4 (new) is scored by evaluating the activity in the last 4 weeks compared with that in the previous 4 weeks (Figure 1), and using the evaluation algorithm in each of the 9 domains, A (severe), B (moderate), C (mild), D (no activity), or E (no history) is determined for each organ (domain). This determination algorithm is complicated and requires a computer, making it unsuitable for practical clinical use. An algorithm for judging mucocutaneous symptoms as an example is shown in Table 4. The 5-level evaluation of each item is set based on whether the physician wants to strengthen treatment considering the change in symptoms between the last visit and the current visit. Therefore, the BILAG system does not always evaluate the lupus activity at the current visit. The differences between the classic BILAG and BILAG 2004 are summarized in Table 5. In general, BILAG-2004 reflects disease activity change more sensitively and reports less false-positive disease activity than the classic BILAG index. Past clinical trials used the classic BILAG [16,17] for the endpoint and recent clinical trials used BILAG-2004 [25], which may have affected the outcome to some extent.

Figure 1. The BILAG-2004 index score sheet [10].

Table 4. The algorithm to score the BILAG-2004 index for the mucocutaneous domain.

Category A

Any of the following recorded as 2 (same), 3 (worse), or 4 (new):

Skin eruption – severe

Angio-oedema – severe

Mucosal ulceration – severe

Panniculitis/Bullous lupus – severe

Major cutaneous vasculitis/thrombosis

Category B

Any Category A features recorded as 1 (improving) OR

Any of the following recorded as 2 (same), 3 (worse), or 4 (new):

Skin eruption – mild

Panniculitis/Bullous lupus – mild

Digital infarcts or nodular vasculitis

Alopecia – severe

Category C

Any Category B features recorded as 1 (improving) OR

Any of the following recorded as > 0:

Angio-oedema – mild

Mucosal ulceration – mild

Alopecia – mild

Periungual erythema/chilblains

Splinter hemorrhage

Category D

Previous involvement

Category E

No previous involvement

Table 5. The main differing points between the classic BILAG index and BILAG-2004.

In general, the vasculitis domain was removed (moved to other domains), and gastrointestinal and ophthalmic domains were added. When the features that contributed to the A score were improving, they contributed to the C score in the classic BILAG index, but they contribute to the B score in the BILAG-2004.
Constitutional or General	Fatigue/malaise/weakness were removed in the BILAG-2004
Mucocutaneous	Swollen fingers, sclerodactyly, calcinosis, and telangiectasia were removed in the BILAG-2004.
Neuropsychiatric	Stroke, organic depressive illness, and episodic migrainous headaches were removed in the BILAG-2004.
Musculoskeletal	Any myositis was classified as category A in the classic BILAG index, but only severe myositis was classified as category A and mild myositis was classified as category B in the BILAG-2004.
Cardiorespiratory	Symptoms, such as pleuropericardial pain, dyspnea, or mild or remittent chest pain were removed. Cardiac failure or symptomatic effusion plus two other criteria, or four of the criteria were necessary for category A in the classic BILAG index, but only one criterion was sufficient for category A criteria in the BILAG-2004.
Gastrointestinal	Peritonitis, enteritis, protein losing enteropathy, intestinal pseudo-obstruction, hepatitis, acute cholecystitis, and acute pancreatitis were newly included in the BILAG-2004.
Ophthalmic	Orbital inflammation, keratitis, uveitis, episcleritis, and scleritis were newly included in the BILAG-2004.
Renal	No major change
Hematological	Circulating anticoagulant was removed and TTP was added in the BILAG-2004. Criteria in category B were changed as follows: White cell count <2500 = > 1000-1900, hemoglobin <11 = > 8-8.9, and platelet count <150 = > 25-49.

As the basic concept of using BILAG is to examine the activity of each organ, the lupus activity of a patient is described as, for example, having two BILAG As and 3 BILAG Bs (out of 9 domains). However, it is possible to use BILAG for overall disease activity. The BILAG-2004 index can be calculated by adding the scores of each domain as A = 12 points, B = 8 points, C = 1 point, D = 0 points, E = 0 points [29], but such usage is not recommended.

PGA (physician’s global assessment)

Although the PGA is not an activity index per se, it is one of the disease activity markers comprehensively evaluated by an attending physician, and it may be the most comprehensive disease activity index in a sense. Indeed, the PGA can be evaluated by taking fatigue into account, which cannot be evaluated by SLEDAI or BILAG. A 3-point scale (as described in the supplementary appendix of [5]) is currently used, which was introduced in the flare index of SELENA-SLEDAI [5] and the PGA is one of the items of many activity indices such as SRI(19), BICLA(14), and LLDAS [30]. As the PGA is a visual analogue scale (VAS), it should be continuous; however, there are calibrations at 0, 1, 2, and 3, and the distribution of the data is biased around them, which may be a disadvantage of this scale.

SRI (SLE responder index)

SRI is not an indicator of disease activity, but it is currently one of the most frequently used primary endpoints in clinical trials. For example, achieving SRI4 is defined as SLEDAI improvement of 4 points or more, PGA not worsening by 0.3 points or more (10% or more), and BILAG having no new As and not having two or more new Bs. This index was designed by post-hoc analysis of a failed belimumab phase II clinical trial [18], and belimumab was later approved by the FDA after phase III clinical trials [19,20] due to the development of this index. After the success of belimumab clinical trials using SRI, many lupus clinical trials adopted SRI for their primary endpoints thereafter. It was considered successful until anifrolumab (anti-type I interferon receptor antibody) failed in one of the phase III clinical trials using SRI4 (23) as a primary endpoint even though a phase II clinical trial using SRI4 [31] and another phase III clinical trial using BICLA [28] were successful. Depending on clinical trials, SRI3, SRI5, or SRI6 is used in addition to SRI4.

BICLA (BILAG-based composite lupus assessment)

BICLA was first used in the clinical trial of the anti-CD22 antibody epratuzumab [14], but it was unable to demonstrate effectiveness in the epratuzumab arm. However, BICLA was suddenly highlighted when one of the phase III clinical trials of anifrolumab [27] failed using SRI4 as a primary endpoint, but the following phase III clinical trial of anifrolumab [28] succeeded using BICLA as a primary endpoint. SRI and BICLA are similar activity indices, but SRI weighs more on SLEDAI and BICLA weighs more on BILAG. Requirements for BICLA response were: [1] BILAG-2004 improvement (all A scores at baseline improved to B/C/D, and all B scores improved to C or D); [2] no deterioration of disease activity (no new BILAG-2004 A scores and ≤1 new B score); [3] no worsening of the total SLEDAI-2K score from baseline; [4] no significant deterioration (<10% worsening) on 100-mm visual analogue PGA, and [5] no treatment failure (defined as non-protocol treatment, i.e. new or increased immunosuppressives or antimalarials; increased parenteral corticosteroids; or premature discontinuation of study treatment).

LLDAS (lupus low disease activity state)

LLDAS [30] is not an activity index as is, but it can be used as one of the endpoints in clinical trials [32,33]. LLDAS is defined as satisfying all of the following 5 items, which consist of disease activity and immunosuppressive medications: 1. SLEDAI-2K ≤4, with no activity in major organ systems (renal, CNS, cardiopulmonary, vasculitis, or fever) and no hemolytic anemia or gastrointestinal activity; 2. No new features of lupus disease activity compared with the previous assessment; 3. SELENA-SLEDAI PGA ≤1; 4. Current prednisolone (or equivalent) dose ≤7.5 mg daily; and 5. Well-tolerated standard maintenance doses of immunosuppressive drugs and approved biological agents, excluding investigational drugs. LLDAS may be a target for SLE treat-to-target (T2T) [34], and LLDAS achievement is associated with significant protection against flare and damage accrual [35].

Why rituximab failed to demonstrate efficacy?

Many experts believed that rituximab, an anti-CD20 antibody, would be effective in clinical trials because it was used for many refractory SLE patients with successful results [36]. The EXPLORER trial [15] was a double-blind placebo-controlled trial in patients with moderate to severe SLE, excluding central nervous system (CNS) disorders and renal disorders.

The EXPLORER trial included patients with a history of meeting the American College of Rheumatology (ACR) revised classification criteria, BILAG A ≥ 1 organ system or BILAG B ≥ 2 organ systems, stable use of one immunosuppressant at entry, and excluded severe CNS or organ-threatening lupus or any other active conditions requiring significant use of steroids or recent treatment by cyclophosphamide or calcineurin inhibitors. The study design was as follows: Patients were randomized at a 2:1 ratio to receive i.v. rituximab (two 1,000 mg doses given 14 days apart) or placebo on days 1, 15, 168, and 182. Immunosuppressants (azathioprine, mycophenolate mofetil or methotrexate) were continued as a background treatment. Additional daily oral prednisone (0.5, 0.75, or 1.0 mg/kg), based on the BILAG score at entry and the amount of steroids already being taken at the time of entry, was administered and tapered beginning on day 16, with the goal of reaching a dosage of ≤10 mg/day over 10 weeks and ≤5 mg/day by week 52.

The primary endpoint was the percentage of patients who achieved and maintained a major clinical response or a partial clinical response at week 52 using 8 BILAG index organ system scores. Of note, the classic BILAG-index [8] was used in this trial, not BILAG-2004. A major clinical response was defined as achieving BILAG C scores or better in all organs at week 24 without a severe flare (1 new domain with a BILAG A score or 2 new domains with a BILAG B score) from day 1 to week 24, and maintaining this response without a moderate or severe flare (≥1 new domains with a BILAG A or B score) to week 52. The definition of partial clinical response was complicated and is described in the original paper.

The patient background was not significantly different between the rituximab and placebo groups, the disease duration was approximately 8 years, and ∼50% of patients had BILAG A, ∼30% had 3 BILAG Bs, and ∼20% had only 2 BILAG Bs. The affected organs were mainly mucocutaneous, musculoskeletal, and general domains, and renal and neurological domains were negligible. The distribution of assigned prednisolone doses was ∼60%, 30%, and 10% for 0.5, 0.75, and 1 mg/kg/day, respectively.

The achievement rates of major clinical response and partial clinical response were all in the 10% range in both rituximab and placebo groups, demonstrating no significant differences between rituximab and placebo in the primary endpoint.

Assuming that rituximab is an effective treatment for SLE, possible explanations for the failure are as follows:

1. Flare criteria were overly strict (only one BILAG B during the last 6 months of the study becomes ‘no response’). There were 17 patients (14.2%) who were judged as ‘no response’ with only one BILAG B in the rituximab group and there were 4 patients (6.3%) in the placebo group.

2. The activity was not sufficiently controlled by the first steroid treatment. Indeed, more than 70% of the patients did not reach BILAG C or better at week 24. This suggested that BILAG remission was overly strict.

3. Conversely, steroids and immunosuppressants may have been overly effective. The placebo group did not relapse to the baseline level. If the observation period was longer, the placebo group may have relapsed and a difference may have been noted. This was supported by the sub-analysis focusing on treatment-resistant blacks and Hispanics yielding a difference. There was also a difference in the sub-analysis of patients who were taking MTX, but not in those who were taking MMF.

4. The classic BILAG index may be insufficient to discriminate responders vs non-responders. As shown in Table 5, BILAG-2004 can better discriminate the disease activity by increasing the chance of medium response to category A to B rather than A to C compared with the classic BILAG index.

Next are some specific examples of disease activity of BILAG B in the BILAG-2004. First, the definition of BILAG B of the mucocutaneous domain is shown in Table 3. For example, if face erythema remained, it became Category B and the major response was unable to be achieved, whereas if face erythema disappeared and re-appeared, it changed from Category C to B and was judged as relapse. In the musculoskeletal domain, if arthritis/tenosynovitis was observed in one or more joints and restricted range of motion was present and lasted for several days (more than 4 weeks), it became Category B. These examples demonstrate how easily category B can be satisfied by chance.

Why belimumab was demonstrated as effective?

After the EXPLORER trial of rituximab, a clinical trial of belimumab, a biologic targeting the same B cells, was conducted and the drug was finally approved by the FDA. What was different from the EXPLORER trial is discussed below. Belimumab failed to meet the primary endpoint in the phase II trial [18]; however, post hoc analysis revealed it to be significantly more effective in serologically active patients (71% of the total) in the belimumab group. In addition, a new treatment response index SRI was devised, and post-analysis of the phase II trial using SRI4 at 52 weeks as an endpoint yielded good results. Therefore, phase III trials (BLISS-52 (19) and BLISS-76 (22)) included patients who were positive for ANA or anti-DNA antibody at the time of screening, and adopted SRI4 at 52 weeks as a primary endpoint. The study was successful and the many of the following clinical trials used SRI as a primary endpoint. Thus, strict entry criteria and appropriate outcome measurement are essential for the success of clinical trials. SRI was considered to be the best activity index of clinical trials until the failure of anifrolumab in a phase III clinical trial [27]. However, this is understandable after examining the BLISS-76 trial in detail. In the BLISS-76 trial, there was no significant difference in efficacy (satisfying SRI4) at week 76, although it was satisfied at week 52 and the primary endpoint was met. Therefore, the trial was barely successful. Of note, in the BLISS-76 trial, the mean percent change from baseline in the SELENA-SLEDAI score was significantly different and the differences were larger according to the treatment period. Such outcomes using % change may be better endpoints than SRI.

Best SLE activity index for clinical trials?

Now, the readers can understand how difficult it is to demonstrate efficacy in a clinical trial for SLE patients. There are several important points to consider in SLE trial protocols. Of importance are the inclusion criteria, background steroid and immunosuppressant treatment, and evaluation criteria.

First, it should be discussed whether the inclusion criteria should be the entire SLE or a part of SLE such as lupus nephritis. Narrowing down to antinuclear antibody-positive cases or anti-DNA antibody-positive cases at screening has been adopted in many recent trials.

Background steroid and immunosuppressive treatments are important. It is not ethically permissible to insufficiently treat active SLE that may cause organ damage; therefore, many trials use a sufficient amount of steroids (± immunosuppressive drugs) as remission induction therapy and evaluate relapse rates. Setting an appropriate treatment protocol to reach an appropriate remission and relapse rate is difficult. Targeting relatively mild SLE or SLE in remission with stable doses of PSL (as in the MIRRA trial [37] in eosinophilic granulomatosis with polyangiitis) may be one method to demonstrate efficacy.

The most important thing for successful clinical trials is the endpoint. Next, a comprehensive disease activity index is discussed. Recently, SRI has been used most frequently due to the success of belimumab clinical trials. However, as SRI is mainly based on SLEDAI, it easily reveals the weaknesses of SLEDAI. The biggest weak point of SLEDAI is that there is only binary evaluation of each symptom or item, and it cannot evaluate improvement. As described above, the rate of SRI4 responders between the belimumab group and placebo group was not significantly different at week 76 of the BLISS76 trial, but the SELENA-SLEDAI mean % change from baseline was significant at most points. Therefore, continuous evaluation may be better than binary evaluation. SRI4 is required to improve SLEDAI by 4 points or more, making it unsuitable for such patients with cytopenia, which gives only low points, whereas patients with a high SLEDAI can easily improve SLEDAI by 4 points.

On the other hand, BILAG classifies the severity of each organ system as A (severe), B (moderate), C (mild), or D (inactive) semi-quantitatively. The BILAG system is comprehensive but complicated, and simple is sometimes better than perfect. The weakness of BILAG is that does not contain serological evaluation.

SLAM [6] and ECLAM [7] are indices created to score the severity of each organ disorder. The concept is good, and the symptoms and items are more comprehensive than SLEDAI, but the weighting may not be appropriate (e.g. 0.5 points for renal injury in ECLAM, equivalent to fatigue), and some items may not be an SLE activity such as general malaise, blood pressure, abdominal pain, and Raynaud’s phenomenon (SLAM). There are many points to be improved in SLAM and ECLAM.

Considering the overall disease activity, it is not always more severe for many organs to be involved, and a single severe organ disorder is often more severe than many minor organ disorders. Physicians usually decide treatment strength depending on the most severely affected organ (s). As such, addition does not always represent the disease activity well. Evaluating only the most severely affected organs may be a better activity assessment than summing all categories. Moreover, as the activity index differs for each patient (i.e. some patients have no increase in anti-DNA antibody), the activity index of the patient should be appropriately reflected. It may be better to evaluate the degree of improvement by continuous values, such as 50% improvement, rather than numerical values, such as SLEDAI 4-point improvement, especially for patients with low disease activity. If disease activity was evaluated using the most severely affected organ (± serological index), disease activity may be evaluable like DAS [38] for rheumatoid arthritis, and disease activity can be categorized as high, medium, or low and can be simplified.

Many studies have tried to find the best activity measurement for SLE clinical trials. Abrahamowicz et al. used the raw data of the belimumab BLISS-76 trial, and found the best combination of measurement outcomes and coefficients to discriminate the outcome of the 10-mg/kg belimumab arm from the placebo arm and validated it using the BLISS-52 trial (LuMOS formula). The LuMOS model incorporated the following response criteria: a ≥ 4-point reduction on the SELENA-SLEDAI (yes/no), increase in complement C4 level (mg/dL), a decrease in anti-double-strand DNA titers (IU/dL), and changes in BILAG scores for organ system manifestations (no deterioration of renal components and improvement in mucocutaneous components) (yes/no). The LuMOS score discriminated placebo vs belimumab groups better than SRI4.

As a continuous activity measure, SLE-DAS was recently published [39]. SLE-DAS is calculated using 21 clinical and laboratory items, most of which are evaluated by presence, but some are continuous measures (swollen joint count, proteinuria, platelet count, and leukocyte count). SLE-DAS discriminated clinically meaningful (PGA ≥ 0.3) deterioration or improvement better than SLEDAI-2K, but the complement and anti-DNA antibody titers were treated as binary measures.

Lastly, the definition of flare is important. The flare index of SELENA-SLEDAI [5] has been well demonstrated to judge mild/moderate flare or severe flare, but some revisions are desired in order to not misjudge the transient worsening of symptoms or lab tests, i.e. rash by UV exposure, proteinuria by exercise, and/or excess protein-rich food. By changing the definition of flare to ‘increasing steroid or adding immunosuppressants is necessary’, such misjudgment may be avoided. Another flare index often used is BILAG. In the EXPLORER trial of rituximab, a single BILAG B score after achieving a BILAG C score or better in all organ systems was judged as flare, which was slightly strict. Therefore, the clinical trials thereafter adopted a single new BILAG B score as acceptable for no flare (as in SRI).

Table 6 summarized the characteristics of each outcome measure. Those who use these indices should understand the advantages and disadvantages of each index and apply to clinical trials.

Table 6. Comparison of the characteristics of each outcome measure.

Outcome measure	Comprehensive organ evaluation	Organ severity assessment	Quantitative measure	Immunological variables	Incorporation of treatment measure
SLEDAI	△	×	^	^	×
BILAG		^	△	×	×
SRI	^	△	n.a.	^	×
BICLA	^	^	n.a.	△	^
PGA	^	△	^	△	△
LLDAS	^	×	n.a.	△	^
SLE-DAS	^	△	^	△	×
LuMOS	△	△	^

n.a.: not applicable.

Conclusion

Clinical trials of SLE and facing difficulties in judging a treatment as effective by one gold-standard outcome index taught us many new things. SRI is not the best, BICLA is a good candidate at present, and LLDAS may be a good practical goal but it needs to be tested in clinical trials. All measurements used SLEDAI and/or BILAG, but SLEDAI is overly simple and BILAG is overly complicated. As indices using continuous values are expected, a novel SLE activity index should be created.

Conflict of interest

KO has received research grants and/or speaker’s fee from Abbvie, Actelion, Asahikasei Pharma, Astellas, AYUMI, Bristol-Myers Squibb, Chugai, Daiichi-Sankyo, Eisai, Eli Lilly, GSK, Janssen, JB, Mitsubishi Tanabe, Nippon Kayaku, Nippon Shinyaku, Novartis, Sanofi and Takeda.