Introduction

Prompt disease control of flares in patients with systemic lupus erythematosus (SLE) is a priority in treatment strategy planning. Glucocorticoids (GCs) have always been used by physicians to treat lupus patients to obtain swift disease control; however, due to the high burden of short- and long-term dosage-related collateral effects, the feasibility of alternative, steroid-sparing treatment approaches has been intensely investigated.

In the last decade, we have witnessed a growing number of studies exploring the safety and efficacy of novel biological agents for both inducing and maintaining low disease activity and remission, especially in the context of lupus nephritis (LN), which is one of the most common organ-related complications of SLE. While biologics are opening a new era for SLE/LN management, clear evidence supporting their net benefit as GCs-sparing agents is still lacking.

Due to these facts, it would be useful to explore the recent evidence on the potential steroid-sparing effect of biologic therapies currently used to treat SLE patients, especially those affected by LN, in phase II and phase III randomized, placebo-controlled trials (RCTs). Similarly, it might be worth exploring unmet needs in available studies, especially from a methodological perspective. To answer these questions, we designed a scoping review to synthesize the available evidence, while highlighting points of improvement that could be addressed by future researches [1].

Methods

Protocol and registration

To conduct this scoping review, we adhered to the guidelines outlined in the Preferred Reporting Items for Systematic Reviews and Meta-Analysis extension for scoping reviews (PRISMA-ScR) [2]. Although a protocol was devised, it was not formally registered.

We formulated two key research questions:

  1. 1)

    What insights does the literature provide regarding the steroid-sparing effect of biological agents in individuals with SLE?

  2. 2)

    Can we substantiate a clear steroid-sparing effect associated with any specific biologic treatment utilized in the available SLE trials?

Eligibility criteria and search strategy

Trials involving patients with SLE, regardless of age, were identified from the MEDLINE (via PubMed) and EMBASE databases on April 18, 2022, using the following key search strategy (as detailed in the supplementary material S1). Search results were confined by using PubMed filters to human clinical trials in English, with full-text articles available for review (excluding conference abstracts alone). Inclusion criteria, following JMB recommendations for scoping reviews, encompassed the population (phase II or III trials involving SLE patients, with or without LN), concept (studies providing detectable steroid data indicating the potential for biological treatments to exhibit a steroid-sparing effect) and context (studies applying biological treatments). The steroid-sparing effect was defined as the ability of the investigated immunosuppressant regimen to reduce both the daily dose and the cumulative prednisone dosage compared to standard treatment at the study’s conclusion. Studies that did not meet these criteria were explicitly excluded from the research. The results of the search and study inclusion process are depicted in a PRISMA-ScR flowchart (Supplementary 2). The risk of bias was assessed using the Cochrane Collaboration’s tool for RCTs [3].

Data extraction

Studies that met the inclusion criteria were independently reviewed for eligibility confirmation based on title and abstract by two researchers (SS and SGF). In case of disagreement, a third researcher (MR) was consulted to reach a consensus in the evaluation stage. Data items were defined and extracted independently by the two researchers (SGF and SS), while data registration in an Excel Table (Microsoft, Redmond, 110 WA, USA) resulted after disagreement discussion and identification of consensus. As the data on eligibility were dichotomous (eligible: yes/no), agreement at both the title and abstract review and the full article review stages was determined by calculation of Cohen’s kappa coefficient (k > 0.8). If consensus could not be achieved, a third part (MR) cleared up the disagreement.

Extracted data included pertinent details about the participant, the concept, the context and the key findings relevant to the review questions.

After reaching a consensus on papers selection and data extraction, three analyses were conducted:

  1. 1)

    Line I data analysis: included both phase II and phase III studies to evaluate the whole impact of steroid-sparing therapies on GC utilization (considered both as cumulative dosage and as for dosage at the end of the study follow-up);

  2. 2)

    Line II data analysis: compared GC tapering potential of the counterposed immunosuppressant therapies in phase II and phase III trials in LN cohorts;

  3. 3)

    Line III: evaluated the quality of evidence of phase III RCTs (clean data analysis) considering both patients with and without LN.

Cochrane Risk of Bias Tool for RCTs was applied to papers selected for the clean data analysis and for the sub-analysis. Bias assessment was independently evaluated and then critically discussed by the two researchers (SGF and SS). When consensus on one of the defined items was not reached, a third part (MR) resolved the disagreement.

Results

As a result, a total of 30 articles were identified through the literature search. During the data analysis phase, eight RCTs met the inclusion criteria [4,5,6,7,8,9,10,11]. In summary, seven studies evaluated B-cell-targeted therapies [4, 5, 7,8,9,10,11], while one study investigated the use of anifrolumab [6]. Table 1 provides an overview of study populations, biological treatments and the presence or absence of a net steroid-sparing effect. Detailed treatment regimens and trial results are outlined in Table 2. Supplementary Table 2. Supplementary 2 presents the screening process in a PRISMA flowchart.

Table 1 Main findings in terms of GC-sparing effect of the trials included in the analysis
Full size table
Table 2 Main characteristics of the trials included in the analysis
Full size table

In the TULIP-LN trial conducted by Jayne et al., the biologic utilized was anifrolumab [6]. The trial explicitly outlined a precise GCs regimen from the study’s outset, applying to both the placebo and treatment arms. This regimen involved methylprednisolone infusion (500 mg IV, within 10 days of randomization or at randomization) and/or oral administration, with a corresponding meticulous tapering schedule (mandatory dosage: 10 mg/day by week 12th; < 7.5 mg/day by week 24th). While a sustained prednisone tapering was similar in the anifrolumab basic regimen and placebo, a higher proportion of intensified anifrolumab patients demonstrated a distinct steroid-sparing effect. This effect, defined as the treatment’s ability to reduce daily steroid intake (55.6% vs. 33.3%), was more pronounced compared to the placebo [6].

In the EMBRACE (Ginzler et al.) and PLUTO (Brunner et al.) trials [4, 5], both investigations centred on the utilization of belimumab. Steroid treatment was permitted across all arms throughout the analysis weeks, and no mandatory steroid tapering was prescribed. However, the assessment of prednisone dosage reduction was considered in the final analysis, revealing no discernible differences in steroid consumption between patients treated with biologics and those receiving standard care. Specifically, in the Ginzler trial [5], during weeks 40 to 52, 14.7% of patients in the treated arm compared to 12.6% in the placebo arm achieved a prednisone reduction exceeding 25% of the initial dosage and/or a daily steroid intake < 7 mg/day (OR 1.3, P = 0.4996). In the Brunner study [4], between the 44th and 52nd weeks, 20% of belimumab-treated and 21.1% of placebo patients experienced a ≥ 25% reduction in prednisone baseline dosage (which was lower at baseline in the belimumab-treated arm, OR 0.92, 95% CI 0.29–2.88), while the median GCs dosage did not decrease by week 52 in either arm. In BEL113750 (Zhang et al.) trial [10], though no forced prednisone tapering was requested, cumulative steroid dosage and tapering maintenance were considered as secondary endpoints. Belimumab population showed a significant reduction in the cumulative steroid dosage undertaken by patients in the 52 weeks of observation (4758.1 mg vs 4190.0 mg, P = 0.0005), a more stable and tight prednisone decrease when compared to placebo (40th–52nd weeks prednisone tapering ≥ 25% in relation to baseline and/or ≤ 7.5 mg/day (15.6% vs 10.9%; OR 1.68, P = 0.0721); daily dose reduction for patients with baseline steroid intake > 7.5 mg/day favours belimumab (P = 0.0288)).

In BEL112341 (Stohl et al.) trial [8], the studied biologic was belimumab. In detail, a steroid tapering regimen was not forced, but steroid dosage was considered as a secondary endpoint. The belimumab arm showed a greater reduction in prednisone dosage among weeks and/or less dosage adjustments when compared to placebo (with significance between weeks 20th and 52nd except for week 32nd (8.1% vs 13.2%, OR 0.55 [95% CI 0.34–0.87]; P = 0.0117). Indeed, the whole cumulative dose of prednisone in the belimumab arm resulted to be lower, but without reaching significance when compared to placebo (3933 mg vs 4567 mg; P = 0.4299).

In the BLISS-76 trial conducted by Furie et al. [11], investigating the use of belimumab, any prednisone dosage was permissible through week 24. Subsequently, only temporary and well-defined adjustments to steroid intake were allowed; however, prednisone reduction was at the investigator’s discretion. The placebo group experienced more treatment failures (14.9% vs. 7.5%, P = 0.005 or 8.1%, P = 0.01) and prednisone dosage adjustments between weeks 16 and 76, although statistical significance was not reached. Additionally, a greater proportion of belimumab-treated patients reduced prednisone by ≥ 25% and to ≤ 7.5 mg/day between weeks 40 and 52 and between weeks 64 and 76 compared to the placebo, but these differences were not statistically significant (1 mg belimumab: 19%, 10 mg belimumab 18%; placebo 13%, P > 0.05).

Data from the BLISS-52 trial (Navarra et al.) [7], which investigated belimumab and specified the same restriction in steroid usage from week 24, indicated a significant superiority in both overall prednisone dosage reduction (10 mg/kg belimumab at 52 weeks, OR 1.78 (CI 1.13–2.79), P = 0.0122) and sustained steroid tapering (tapering for more than 12 weeks, 1 mg/kg belimumab, P = 0.0465; 10 mg/kg belimumab, P = 0.0032) in the belimumab-treated arms compared to the placebo. In the trial NTC00071487 by Wallace et al. [9], exploring the utilization of belimumab, unlimited changes in prednisone and immunosuppressive medications were allowed, but prednisone dosage reduction was considered a secondary endpoint. No difference emerged between groups in prednisone reduction; however, placebo patients with no steroid or low-dosage steroid at the beginning of the study manifested a higher increase in prednisone dosage compared to the belimumab-treated arm (P = 0.0459).

Lastly, as previously mentioned, a sub-analysis was carried out considering only those papers assessing the GCs sparing effect in patients with LN (both considering phase II and phase III trials). Only one paper [6] was eligible for the analysis. In TULIP-LN trial [6], even though all included patients received one intravenous methylprednisolone pulse before randomization and a mandatory regimen at weeks 12th and 24th (< 10 mg/day and < 7.5 mg/day respectively), intensified anifrolumab arm showed greater sustained steroid reduction in comparison to other arms. It is worth noting, however, that the placebo group had better renal response and GCs tapering maintenance in comparison to basic anifrolumab regimen.

Quality of the evidence

When specifically assessing phase III trials through a meticulous data analysis [5, 7,8,9,10,11], encompassing both LN and SLE patients without LN, the overall quality of the studies was deemed fair or good, as outlined in Table 3, considering the risk of bias. In particular, four studies [5, 7, 10, 11] exhibited a low risk of bias in various aspects, including random sequence generation, allocation concealment, blinding procedures, data outcomes, selective reporting and other potential biases. However, in the trials led by Stohl and Wallace [8, 9], there is no explicit statement regarding the methodology applied in random sequence generation, allocation concealment and blinding of personnel and participants. Nevertheless, a thorough discussion of all remaining biases is presented, and in conclusion, there is no solid data supporting the notion that the validity of results is less than fair.

Table 3 Cochrane Risk Bias Tool for randomized controlled trials
Full size table

Discussion

The utilization of biological treatments demonstrated a clear steroid-sparing effect in five investigations, comprising four RCTs focusing on belimumab [8, 10, 11] and one on anifrolumab [6]. What insights can be gleaned from this analysis? Firstly, a steroid-sparing effect is achievable with various treatments. Secondly, to accurately identify and assess a drug’s potential to reduce the cumulative dose of prednisone per patient, certain steps are crucial:

  • establish a consensus on the definition of a steroid-sparing effect;

  • advocate for the inclusion of the steroid-sparing effect as an outcome in all future drug efficacy studies involving SLE patients.

Strengths and limitations

While the novelty of our scoping review lies in attempting to address an unmet need for patients with SLE and LN, we acknowledge several limitations that affect the robustness of our conclusions. Firstly, the identified trial protocols exhibit heterogeneity, and the inclusion of a strong standard of care regimen with a robust steroid supplement in the early phase of the trial may have obscured the effective and beneficial role of the new drug in reducing overall GCs consumption, leading to a significant knowledge gap.

Furthermore, steroid dosage adjustments were permitted in all studies at various time points and dosages, contributing to the potential impact of prednisone on the efficacy of immunosuppressive treatment. The choice of a permissive complementary treatment was influenced by the characteristics of the population under examination and the ethical imperative of preserving the patient’s health. Active SLE patients, especially those with LN or refractory LN, require prompt disease control, explaining the fluctuation in steroid regimens, which are still considered rescue treatments in SLE management.

Considering all these points, it is challenging to provide clear indications on when and if it is possible to consider these described molecules as “steroid-sparing” agents. It may be more accurate to currently view these drugs as valuable add-on therapies to pre-existing standard of care treatments. In phase III trials on LN, the complete steroid-sparing effect of belimumab [15] is not systematically specified or analysed (due in part to the study design).

Finally, consistent with the scoping review design, it was not the purpose of this analysis to provide aggregate data on the potential steroid-sparing effect of each regimen. We performed a searching with Embase/PubMed filter for RCTs, so we recognize that it is possible that some studies of relevance may have not been identified by using this approach. Similarly, some potentially relevant studies investigating the use of biologics in SLE might have been excluded because the steroid-sparing effect (e.g., steroid tapering protocol and/or cumulative dose) was not quantifiable. Future systematic reviews are required to broaden the spectrum of evaluable works. In conclusion, our study lays the foundation for future research needed to define which patients might benefit most from specific treatment protocols and to establish a clear definition of the steroid-sparing effect guaranteed by new immunosuppressant therapies, aiming to enhance tailored management.