Objective We evaluated the discriminant capacity of the Lupus Low Disease Activity State (LLDAS) in post-hoc analysis of data from the BLISS-52 and BLISS-76 trials of belimumab in systemic lupus erythematosus (SLE).
Methods LLDAS attainment, discrimination between belimumab and placebo arms, and the effects in subgroups with high disease activity at recruitment were evaluated at week 52 using appropriate descriptive statistics, χ2 test and logistic regression.
Results At week 52, for belimumab 10 mg/kg, 17.0% and 19.3% of patients who achieved a Systemic Lupus Erythematosus Responder Index-4 also attained LLDAS in BLISS-52 and BLISS-76, respectively. Significantly more patients attained LLDAS on belimumab 10 mg/kg compared with placebo (12.5% vs 5.8%, OR 2.32, p=0.02 for BLISS-52; 14.4% vs 7.8%, OR 1.98, p=0.04 for BLISS-76). In a subgroup analysis, the difference in week 52 LLDAS attainment between belimumab 10 mg/kg and placebo was greater in patients who had higher disease activity at baseline, compared with the overall group.
Conclusions LLDAS was able to discriminate belimumab 10 mg/kg from placebo in the BLISS-52 and BLISS-76 trials. Our findings support the validity of LLDAS as an outcome measure in SLE clinical trials.
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
What is already known about this subject?
Lupus Low Disease Activity State (LLDAS) predicts better clinical outcomes and has been found to discriminate responders in a phase II trial of anifrolumab in systemic lupus erythematosus (SLE).
What does this study add?
This is the first study to demonstrate that LLDAS can discriminate responders in the pivotal phase III trials of belimumab (BLISS-52 and BLISS-76) in SLE, and is a more stringent outcome measure than the Systemic Lupus Erythematosus Responder Index-4 in these trials.
How might this impact on clinical practice or future developments?
This study lends weight to the potential utility of LLDAS as a novel clinical trial outcome measure for SLE randomised controlled trials.
Measurement of treatment response in systemic lupus erythematosus (SLE) clinical trials has generally been based on measurement of proportions of patients attaining a certain degree of change from baseline; in contrast, a treat-to-target analysis has seldom been applied. The Lupus Low Disease Activity State (LLDAS), a potential response indicator for lupus clinical trials, has been found to correlate with reduced damage accrual in SLE,1 suggesting that it may be a useful treatment target in the clinic. In a trial setting, LLDAS has been found to correlate with key outcome measures and has discriminated responders from non-responders in phase II trials of anifrolumab and baricitinib.2 3
In this study, we sought to evaluate LLDAS utility in discriminating drug from placebo in a post-hoc analysis of data from the pivotal phase III BLISS-524and BLISS-765trials of intravenous belimumab, an anti-BAFF (B-cell activating factor) monoclonal antibody, in patients with moderate to severe SLE.
BLISS-52 and BLISS-76 trials
The utility of LLDAS as an outcome measure was assessed in a post-hoc analysis of the data from the phase III, 52-week and 76-week BLISS-524(NCT00424476) and BLISS-765(NCT00410384) trials of intravenous belimumab in patients with SLE. In these large multicentre studies, seropositive (antinuclear or antidouble-stranded DNA [anti-dsDNA] antibody-positive) patients (≥18 years old) with moderate-severe SLE (SELENA-SLEDAI6 score ≥6), as defined by the revised American College of Rheumatology SLE classification criteria,7 were randomised in a 1:1:1 ratio to receive belimumab 1 mg/kg or 10 mg/kg, or placebo, by intravenous infusion on days 0, 14 and 28, and then every 28 days until 48 weeks4or 72 weeks,5 in addition to standard of care. The primary efficacy endpoint was improvement in the Systemic Lupus Erythematosus Responder Index-4 (SRI-4) at week 52, defined as a reduction ≥4 points in SELENA-SLEDAI6 score; no new British Isles Lupus Assessment Group (BILAG)8 A organ domain score and no more than one new B organ domain score; and no worsening (<0.3 increase) in Physician Global Assessment (PGA) score, compared with baseline. Patients with active severe or unstable neuropsychiatric SLE or lupus nephritis were excluded. Further details on BLISS-52 and BLISS-76 study design and endpoints have been published.4 5 Clinical trial data were accessed and analysed via the SAS Data Access System, through a data sharing agreement.
Lupus Low Disease Activity State
As previously published, LLDAS was attained if all of the following were present: (1) SLEDAI-2K (SLE Disease Activity Index 2000) score ≤4, with no activity in major organ systems (renal, central nervous system, cardiopulmonary, vasculitis and fever) and no haemolytic anaemia or gastrointestinal activity; (2) no new features of lupus disease activity compared with the previous assessment; (3) a physician global assessment of activity score (PGA, 0–3) ≤1; (4) current prednisolone-equivalent dosage ≤7.5 mg/day; and (5) standard maintenance dosages of immunosuppressive drugs and approved biologics allowed.1
A detailed description of how LLDAS was defined using variables in the BLISS-52 and BLISS-76 data sets is provided in online supplementary table S1.
LLDAS attainment was assessed across all study timepoints. Attainment of LLDAS, association with the primary trial endpoint (SRI-4), and discrimination between belimumab and placebo-treated patients both for the whole study population and subgroups according to the level of disease activity, glucocorticoid dose and presence of damage at recruitment were evaluated using appropriate descriptive statistics, the χ2 test and logistic regression, where appropriate, using statistical software R (V.3.4.3).
BLISS-52 and BLISS-76 patient demographics and baseline characteristics have been published4 5 and are presented in online supplementary table S2.
BLISS studies efficacy endpoints
Patients receiving belimumab treatment were more likely to achieve an SRI-4 response at week 52 than those treated with standard of care, in both BLISS-52 and BLISS-76 studies (table 1), as previously reported.4 5
LLDAS as an outcome measure in BLISS-52 and BLISS-76
LLDAS as a discriminator between placebo and belimumab in BLISS-52 and BLISS-76
Few patients (0% for BLISS-52 and 2.2% for BLISS-76) were in LLDAS at study entry (online supplementary table S3). In contrast to SRI-4, LLDAS was attained by few patients in the placebo arms at week 52 (5.8% in BLISS-52 and 7.8% BLISS-76) (table 1).
Significantly more patients attained LLDAS at week 52 on belimumab 10 mg/kg compared with placebo in both trials (12.5% vs 5.8%, OR 2.32, p=0.02 in BLISS-52; 14.4% vs 7.8%, OR 1.98, p=0.04 in BLISS-76) (table 1). A statistically significant difference was also seen at week 44 in BLISS-52 between the belimumab 1 mg/kg group and placebo (11.52% vs 6.19%, OR 1.97, p=0.05), and at week 72 in BLISS-76 between belimumab 10 mg/kg and placebo (19.4% vs 10.29%, OR 2.09, p=0.02) (online supplementary figure 1).
Of the LLDAS components at week 52, most patients (from 97.1% to 99.0% across treatment groups, in both studies) met criterion 5 (standard dose immunosuppressants allowed), with the fewest patients (22.5%–30.9%) meeting criterion 1 (SLEDAI-2K score ≤4, with no activity in major organ systems and no haemolytic anaemia or gastrointestinal activity) (table 1). Increases in LLDAS attainment across time were largely driven by attainment of criteria 1 and 3 (PGA score ≤1) (online supplementary figures 2-6).
Comparison of LLDAS with SRI-4 as an outcome measure
At week 52, in both studies, fewer patients in the treatment arms attained LLDAS compared with SRI-4 (table 1). LLDAS attainment was more stringent than SRI-4 attainment at week 52, with 13.8% of patients in BLISS-52 and 17.7% of patients in BLISS-76 who achieved an SRI-4 also attaining LLDAS across all treatment groups (table 2). However, the majority of patients who did not achieve an SRI-4 at week 52 also failed to attain LLDAS (96.7% in BLISS-52 and 95.2% in BLISS-76). Conversely only 8.7% of LLDAS responders in BLISS-52 and 20.6% of LLDAS responders in BLISS-76 failed to achieve SRI-4.
Subgroup analyses revealed that the difference in LLDAS attainment at week 52 between belimumab 10 mg/kg and placebo was greater in patients with higher disease activity at baseline, relative to all study participants (table 3). The subgroups of patients in which a greater difference was seen were those with a high anti-dsDNA antibody level ≥30 IU/mL (for both studies), low C3 (<90 mg/dL) and/or C4 (<16 mg/dL) (both studies), high anti-dsDNA antibody levels or low complement levels (both studies), SLEDAI-2K score ≥10 (BLISS-52 study only), or prednisolone dose ≥7.5 mg/day (BLISS-52 only) at study entry. Patients in the belimumab 10 mg/kg treatment arms who had baseline high anti-dsDNA levels in BLISS-76, or baseline low C3/C4 and SLEDAI-2K score ≥10 in BLISS-52, were more likely to attain LLDAS at week 52 compared with their lower disease activity counterparts (table 3). In contrast to the greater discrimination in LLDAS attainment between active and placebo seen with higher baseline disease activity, at most timepoints in both studies, patients with SLEDAI-2K score ≤9 at baseline were more likely to attain LLDAS than those with SLEDAI-2K score ≥10 (online supplementary figure 7). Additionally, patients with less organ damage at baseline (SLICC (Systemic Lupus International Collaborating Clinics) Damage Index [SDI]=0) were more likely to attain LLDAS at week 52 than those with an SDI score ≥1 (table 3).
We have demonstrated that the LLDAS is able to discriminate active treatment from placebo in the pivotal phase III BLISS studies of intravenous belimumab in moderate-severe SLE, with a greater percentage of patients receiving belimumab 10 mg/kg attaining LLDAS at week 52 compared with placebo in BLISS-52 and BLISS-76. The ability to discriminate treatment arms was evident despite the relatively modest difference measured by the SRI-4 in the original trials, and LLDAS appeared to be a more stringent outcome measure than the SRI-4, with a smaller percentage of patients overall attaining LLDAS compared with the SRI-4 and in particular a very low LLDAS attainment among placebo-treated patients. We would envisage that in a trial where the treatment is more efficacious, a larger proportion of patients in the active treatment arm would achieve LLDAS, with still few achieving this endpoint in the control arm. Of course, this will only be established when LLDAS is used as an endpoint in future trials of novel therapies, although such an outcome is strongly suggested by post-hoc analysis of the phase II anifrolumab trial.2 Although harder to achieve, a more stringent, and clinically relevant, trial endpoint has potential benefits, for example in permitting the conduct of smaller clinical trials, and distinguishing novel therapies that will effect robust clinical change. This conclusion is supported by the very low rates of LLDAS attainment in the placebo arms of both trials.
These results add to the growing body of evidence that the LLDAS is useful not only as a clinical treatment target,1 9 but that it has discriminant validity as a clinical trial endpoint in SLE randomised controlled trials.2 Recently, the LLDAS was found to discriminate responders from non-responders in the phase II clinical trials of anifrolumab2 and baricitinib in SLE.3 The frequencies of LLDAS attainment in these studies were higher than in the BLISS studies—17%, 39% and 28% of patients on placebo, anifrolumab 300 mg and anifrolumab 1000 mg at week 52, and 26%, 33% and 38% of patients on placebo, baricitinib 2 mg and baricitinib 4 mg at week 24. This could be due to differences in study patient populations and treatment protocols, and/or reflect differences in efficacy between these treatments. Our study is the first to confirm the discriminant validity of the LLDAS in two large positive phase III clinical trials of SLE treatment.
In pooled post-hoc univariable and multivariable analyses, patients with higher disease activity at baseline (SELENA-SLEDAI score ≥10, low complement, anti-dsDNA positivity and steroid dose ≥7.5 mg/day) have been found to be more likely to achieve an SRI-4 response with belimumab treatment compared with placebo, in the BLISS-52 and BLISS-76 studies.10 We similarly found that the difference in LLDAS attainment between belimumab 10 mg/kg and placebo was greater in these high disease activity subgroups than for the trial patients overall.
Limitations of this study include its post-hoc nature. However, the study is strengthened by prospectively collected data within the framework of a rigorously conducted, double-blind randomised controlled trial. The LLDAS definition was applied to available data in a stringent manner to ensure minimal or no misclassification. Gastrointestinal activity, part of LLDAS criterion 1, was not captured in the original trials and was here assumed to be captured as part of the PGA (LLDAS criterion 2); this assumption requires further validation. Our application of the LLDAS definition to the BLISS-52 and BLISS-76 data sets required for criteria 1 and 2 to have neither major organ involvement or new activity in either BILAG or SLEDAI measures of disease activity. This is a strength of our post-hoc analysis and may partly explain the relatively low frequency of LLDAS attainment in this study, in comparison with a previous post-hoc analysis of phase II data for anifrolumab in SLE, in which these criteria were assessed based on SLEDAI alone.2 Despite these stringent criteria, it is important to note that LLDAS was still attained from a relatively high baseline disease activity (average SLEDAI score at study entry was ~10). There were small differences in LLDAS attainment between the BLISS-52 and BLISS-76 studies, both in the overall group and in subgroup analyses, which could be accounted for by small differences in treatment protocol, in particular the permitted doses of background DMARDs (disease-modifying drugs) and steroids.
In conclusion, we have shown that LLDAS has discriminant validity when applied retrospectively, in two large positive phase III clinical trials of SLE treatment, and is a more stringent endpoint than that used in the original trials. These findings, and the recent report of applying the LLDAS prospectively in a phase II clinical trial,3 suggest inclusion of LLDAS as a trial endpoint in studies of novel therapies in SLE.
Clinical trial data were accessed and analysed via the SAS Data Access System, through a data sharing agreement with GSK.
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Handling editor Josef S Smolen
Contributors All authors have contributed to study design, data collection and analysis, and manuscript creation.
Funding MN is supported by a National Health and Medical Research Council (NHMRC) Career Development Fellowship (APP1126370). SO was supported by a postdoctoral fellowship from Arthritis Australia/Arthritis Victoria. Project funding was received from the St Vincent’s Hospital Research Endowment Fund.
Competing interests EFM has received consulting honoraria from GSK, unconnected to this study. No other authors have any competing interests to declare.
Patient consent Not required.
Ethics approval The studies were approved by a central or local institutional review board or ethics committee, and all patients provided written informed consent.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement Data from the original BLISS trials are available through a data sharing agreement with GSK.