Disease-modifying therapies in systemic lupus erythematosus for extrarenal manifestations

Abstract

Our 2022 published working definition of disease modification in systemic lupus erythematosus (SLE) was ‘minimising disease activity with the fewest treatment-associated toxicities and slowing or preventing organ damage progression’. The objective of this review was to classify current SLE treatments according to the proposed non-renal disease modification criteria excluding toxicities. Based on a review of select clinical trial (n=32) and observational study (n=54) publications for 14 SLE medications across different therapeutic classes, and the authors’ clinical experience, we evaluated disease modification potential as per the proposed framework at three time points. Specific criteria used to determine disease modification potential included a drug’s capacity to reduce: (1) non-renal disease activity, (2) severe flares, (3) use of steroids/immunosuppressants and (4) organ damage accrual. Criteria 1–3 were assessed at 1 year and 2–5 years and, when positive, were considered evidence for disease modification potential; criterion 4 was used to confirm disease modification at >5 years. Each treatment received one of four mutually exclusive designations at each time point: (a) criterion met, (b) indications of criterion met despite insufficient evidence in the literature, (c) inconclusive and (d) no available supportive data. This review excludes an assessment of potential toxicities. Eight of the 14 SLE treatments met ≥1 disease modification criteria up to year 5. Hydroxychloroquine improved overall survival at >5 years, suggesting long-term disease modification, but no data on specific organ systems were reported. Belimumab was the only treatment to meet all criteria. Belimumab and hydroxychloroquine met disease modification definitions across three time points. Evidence for other SLE therapies was incomplete, particularly at >5 years. Future studies are warranted for other treatments to meet the disease modification criteria. We discuss challenges to classification and possible updates to our published criteria.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• We previously proposed a working definition of disease modification in systemic lupus erythematosus (SLE) as: ‘the minimisation of disease activity with the fewest treatment-associated toxicities and slowing or preventing the progression of organ damage’.

WHAT THIS STUDY ADDS

• We evaluated available literature for 14 currently administered SLE treatments to determine for the first time which therapies fulfil the proposed disease modification criteria. Outcomes at year 1 and years 2–5 indicated disease modification potential, with disease modification confirmed beyond year 5 by evidence of delayed organ damage accrual.

• Based on the available literature, hydroxychloroquine and belimumab are the only treatments for SLE that currently satisfy disease modification criteria.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• Several treatments other than hydroxychloroquine and belimumab have data indicating disease modification potential, and these will need formal evaluation in longer term studies, with available data often not available or incomplete, currently precluding their classification as disease-modifying.

• We believe a consensus style approach with a larger group of experts is required to refine the SLE disease modification definition and fully address disease modification in current SLE treatments.

Introduction

SLE is a chronic and heterogeneous autoimmune disease in which 30–50% of patients develop organ damage within 5 years of diagnosis.¹ Despite the available treatments for SLE, a systematic literature review of studies published after 2015 reported remission prevalence ranging from 21% to 70% at 5 years, and only 10% of patients achieved remission at 10 years.²

Disease modification is a well-established concept in other therapeutic areas, including rheumatoid arthritis, multiple sclerosis, Parkinson’s disease and Alzheimer’s disease.^3–9 Characterisation of the available treatments used in SLE as disease modifying can aid in the comparison of treatments and decision-making for physicians and patients. Optimal use of disease-modifying treatments could ultimately reduce the disease burden of SLE and consequently have beneficial impacts on health-related quality of life (HRQoL), patient-reported outcomes and the economic burden of disease.

The proposed working definition of disease modification in SLE was previously published as: ‘the minimisation of disease activity with the fewest treatment-associated toxicities and slowing or preventing the progression of organ damage’.¹⁰ To substantiate a disease modification claim beyond 5 years, evidence is required of delayed or prevented progression of organ damage, usually indicated as the absence of an increase in the Systemic Lupus International Collaborating Clinics/American College of Rheumatology (ACR) Damage Index (SDI); however, long-term clinical trial data are often limited. In this review, we further explore the concept of disease modification in SLE by applying the proposed criteria to treatments currently used in SLE, where the assessment of treatment tolerability is excluded and the focus is on disease activity and irreversible organ damage criteria. We also discuss further refinements of the definition.

Methods

The classes of SLE treatments considered in this review comprised glucocorticoids, antimalarials (hydroxychloroquine), immunosuppressants, calcineurin inhibitors and biological therapies. Medicines used predominantly for comorbidities (eg, analgesics and antihypertensives) were not included. 14 SLE therapies from the aforementioned classes were selected for analysis.

We searched PubMed for clinical trial and clinical practice studies that investigated the 14 selected treatments in patients with SLE. Publications were initially screened for relevance to this analysis using their title and abstract, and publications containing data deemed relevant to the disease modification criteria¹⁰ were selected for analysis. Additional publications known to the authors that were not identified using PubMed were also included if relevant.

We identified clinical trial (n=32) and clinical practice/observational study (n=54) publications and summarised the study outcomes. Using these study outcomes, along with the authors’ clinical experiences, we determined whether the therapies met criteria of the published disease modification framework at three time points.¹⁰ Specific criteria used to determine disease modification potential included a treatment’s capacity to reduce: (1) non-renal disease activity, (2) severe flares, (3) use of steroids/immunosuppressants and (4) organ damage accrual. Criteria 1–3 were assessed at 1 year and at 2–5 years and, when favourable, were considered evidence for disease modification potential. Criterion 4, reducing organ damage accrual, was used to confirm disease modification at >5 years. Each treatment received one of four mutually exclusive designations at each time point: (a) criterion met, (b) indications of criterion met despite insufficient evidence in the literature, (c) inconclusive and (d) no available data. Notably, this review focused on extrarenal disease activity and irreversible organ damage criteria and does not include an assessment of potential reversible or short-term damage associated with treatment tolerability.

Results

Treatments with disease modification potential: year 1 outcomes

To demonstrate disease modification potential, outcomes at year 1 required fulfilment of one or more of the following criteria: (1) a significant reduction in disease activity (measured using a validated tool—for example, SLE Disease Activity Index (SLEDAI), or British Isles Lupus Assessment Group (BILAG), or a composite index, such as the SLE Responder Index-4); (2) significant reduction in severe flare using a validated tool (ie, Safety of Estrogens in Lupus Erythematosus National Assessment-SLEDAI Flare Index (SFI) or BILAG); and (3) a reduction in use of glucocorticoids and/or immunosuppressants (table 1). The reduction in glucocorticoid use required a taper to ≤7.5 mg/day prednisone-equivalent, as per the 2019 European Alliance of Associations for Rheumatology (EULAR) SLE treatment recommendations (NB: updated 2023 recommendations, which suggest a taper to ≤5 mg/day, became available after these criteria were proposed¹¹) and the Lupus Low Disease Activity State definition.^{12 13}

Table 1.

Application of the proposed matrix for non-renal disease activity and organ damage disease modification criteria

Product	Disease modification potential		Disease modification confirmed (beyond 5 years)
	Outcomes: year 1	Outcomes: years 2–5	Outcomes: year 5 or later
	❶ Significant reduction in disease activity measured using a validated tool (ie, SELENA-SLEDAI, BILAG, SRI-4) ❷ Significant reduction in severe flare measured using a validated tool (ie, SFI or BILAG) ❸ Reduction in use of steroids* and/or immunosuppressants	❶ Sustained improvement in multiple organ domains/no worsening in multiple organ domains ❷ Prevention of severe flares ❸ Continued reduction in use of steroids* and/or immunosuppressants	No change in SDI or delayed progression
Glucocorticoids
Hydroxychloroquine
Immunosuppressants
Azathioprine
Cyclophosphamide
Leflunomide
Methotrexate
Mizoribine
Mycophenolate mofetil
Calcineurin inhibitors
Ciclosporin
Tacrolimus
Biologics
Abatacept
Anifrolumab
Belimumab
Rituximab

, criterion met;, insufficient evidence in the literature to meet the specific criterion, but strong general indications of met criterion;, inconclusive;, data not available in the literature to support criterion met;, negative impact on criterion.

*≤7.5 mg/day prednisone-equivalent as per 2019 EULAR SLE treatment guidelines and LLDAS; ≤5 mg/day as per DORIS remission definition.

BILAG, British Isles Lupus Assessment Group; DORIS, Definition of Remission in SLE; EULAR, European Alliance of Associations for Rheumatology; LLDAS, Lupus Low Disease Activity State; SDI, Systemic Lupus International Collaborating Clinics/American College of Rheumatology Disease Index; SELENA-SLEDAI, Safety of Estrogens in Lupus Erythematosus National Assessment-SLE Disease Activity Index; SFI, SELENA-SLEDAI Flare Index; SRI-4, SLE Responder Index-4.

Glucocorticoids,¹⁴ hydroxychloroquine,^{15 16} cyclophosphamide,¹⁷ leflunomide,¹⁸ mycophenolate mofetil (MMF),^{17 19–21} tacrolimus,^{22 23} abatacept,²⁴ belimumab,^25–36 anifrolumab^37–41 and rituximab^42–54 met year 1 criterion 1 of a reduction in disease activity (table 1). Of these, belimumab^{25–36 55} and anifrolumab^{37–41 56} met all three criteria for disease modification potential at year 1, and rituximab also showed a reduction in the number of severe SFI criteria at 6 months versus baseline after the first course, thereby meeting two of the criteria at year 1.⁴⁵ Glucocorticoids,^57–60 hydroxychloroquine, azathioprine, methotrexate, MMF, ciclosporin, tacrolimus, abatacept and rituximab had inconclusive evidence for other year 1 criteria (for a full list of publications examined, please see online supplemental tables 1–5). Glucocorticoid and hydroxychloroquine studies involved comparisons of reduced dose versus higher dose, or withdrawal versus maintenance, which are challenging to match directly with the criteria for disease modification potential. Glucocorticoids and hydroxychloroquine were therefore classified as inconclusive for criterion 2 at year 1.

It is worth noting that data were lacking on whether glucocorticoid use impacts immunosuppressant use at year 1. Therefore, as criterion 3 at year 1 requires a reduction in glucocorticoid and/or immunosuppressant use, we deemed glucocorticoids to have not met this criterion. Taken together, although relevant published data are lacking, based on authors’ clinical experiences and published treatment recommendations,⁶¹ glucocorticoids were considered to have met the reduction in disease activity criterion at year 1.

No data were found to support any disease modification potential at year 1 for mizoribine.

Treatments with disease modification potential: years 2–5 outcomes

The criteria required to demonstrate disease modification potential at years 2–5 comprised at least one of the following: (1) sustained improvement in multiple organ domains/no worsening in multiple organ domains; (2) prevention of severe flares; and (3) a continued reduction in the use of glucocorticoids and/or immunosuppressants (table 1).

For belimumab26 27 31 32 62–65 and anifrolumab,^{41 66 67} there were sufficient data to meet all three criteria for disease modification potential at years 2–5 (table 1). Hydroxychloroquine met the criteria for improvement/no worsening in multiple organ domains, where its use was associated with reduced risk and delayed onset of organ damage^{16 68–70}; however, criteria 2 and 3 (the prevention of severe flares, and continued reduction in use of glucocorticoids and/or immunosuppressants, respectively) were classified as inconclusive based on the available literature^{16 71–74} and author experience. Withdrawal of hydroxychloroquine resulted in a larger proportion of patients with flares (34.1%), defined as new SLEDAI-2K scores ≥4, compared with patients who continued treatment (17.9%).⁷³ However, any abrupt discontinuation of an SLE treatment may be expected to exacerbate flare rates versus tapering, and so the interpretation of these results in the context of the published framework should not be considered. In a small retrospective extension study, the relative risk for major flares was not significantly reduced with hydroxychloroquine compared with placebo,⁷⁴ whereas in another study, the severe flare rate in the low-dose hydroxychloroquine group was not significantly different from the rate observed in the high-dose hydroxychloroquine group.⁷² Overall, these conflicting observations resulted in an inconclusive classification for hydroxychloroquine.

Azathioprine and MMF met criterion 3, as one clinical trial⁷⁵ showed that among patients taking ≥7.5 mg/day prednisone at enrolment, 83.5% of patients receiving azathioprine and 94.9% of patients receiving MMF achieved reductions in prednisone dose to <7.5 mg/day by 24 months (online supplemental table 3).

Glucocorticoids had available data; however, criteria 1 and 2 were classified as not met given the known negative impact on organ damage from long-term glucocorticoid use, although as mentioned for year 1, glucocorticoids have short-term disease activity-reducing properties.^{58 60} As studies found conflicting evidence of the proportion of patients receiving immunosuppressants between prednisone regimens,^{14 58} we also did not classify glucocorticoids as meeting criterion 3 (a reduction in use of glucocorticoids or immunosuppressants).

Some year 2–5 criteria were classified as inconclusive for azathioprine,⁷⁵ MMF^{20 75 76} and rituximab43 44 77–79 based on a lack of clear evidence related to the definitions. Azathioprine and MMF were classified as inconclusive for criterion 1 as both treatments resulted in similar resolution of disease activity for most individual body systems (from BILAG A/B to BILAG D) in one clinical trial.⁷⁵ Data from observational studies of MMF were also difficult to classify according to the criteria because of a lack of statistical significance in comparisons of groups, as well as these data having been generated in small retrospective studies.^{20 76} Data for rituximab were inconclusive for criterion 1, largely due to a lack of data related to multiple organ domains. Most available data were from small observational studies, with one study noting no significant difference versus placebo in SDI after a median follow-up time of 4.5 months,⁷⁸ whereas another study did report significant reductions in BILAG (but the study did not report BILAG by individual organ domains).⁴⁴ Improvements in individual organ domains from baseline were reported with rituximab in small observational studies in Korea (p<0.05) and Colombia (no statistical testing performed).^{43 79} Together, we concluded that the available data were not sufficiently strong for rituximab to fully meet criterion 1 at years 2–5.

For many SLE treatments, criteria at years 2–5 were not met owing to insufficient data—either due to few studies covering the years 2–5 time frame, or because outcomes specific to the definition were not reported (cyclophosphamide, leflunomide, methotrexate, mizoribine, ciclosporin, tacrolimus and abatacept (for a full list of publications examined, please see online supplemental tables 1–5)). For example, data for reduction in glucocorticoid use were often not reported according to the defined ≤7.5 mg/day prednisone-equivalent threshold, or flare data did not contain information about severe flares specifically, which is the focus of the disease modification criterion 3.

Confirmed disease-modifying treatments for SLE: outcomes beyond year 5

Confirmed disease modification requires no change in SDI or delayed progression at 5 years or later (table 1); for example, this may include a lower SDI score versus placebo at a given time point, a reduced rate of change in SDI score over a certain period or a delayed progression of organ damage. Disease modification 5 years or later was confirmed for belimumab based on the results of a propensity score-matched comparative analysis between patients enrolled in the phase 3 belimumab extension study and patients in the Toronto Lupus Cohort,⁶⁴ and a similar study of US and non-US patients in long-term extension studies propensity score-matched to patients in the Toronto Lupus Cohort.⁸⁰ As anifrolumab lacks data >5 years, the long-term impact on disease modification is not conclusive. Although some data for hydroxychloroquine at earlier time points were inconclusive or insufficient, available long-term data for hydroxychloroquine were, in fact, suggestive of long-term disease modification. Hydroxychloroquine has been shown to improve cumulative 15-year survival rates (0.95 vs 0.68 (p<0.001) with or without hydroxychloroquine, respectively).⁷⁰ However, while survival was not a criterion of the published definition of disease modification for SLE, the improvements in survival led us to conclude that disease modification at year 5 or later was met for hydroxychloroquine.

Glucocorticoids did not meet the criteria for disease modification at 5 years or later, given their widely reported negative impacts on long-term organ damage.^{58 81–83} For the remaining SLE treatments, published data for 5 years or later relating to SDI or organ damage progression were not available (table 1).

Discussion

This is the first attempt to determine which currently administered SLE treatments fulfil the proposed SLE disease modification criteria.¹⁰ Due to the use of multiple agents within studies, differences in study design, patient populations and definitions of treatment response, the categorisation of treatments as per the published working definition¹⁰ was challenging.

In this review, we summarised the available data according to the proposed criteria and confirmed disease modification at >5 years for hydroxychloroquine and belimumab, with several other therapies also meeting individual criteria for disease modification potential at the year 1 and years 2–5 time points. The lack of available data for hydroxychloroquine versus placebo or other standard treatments at earlier time points could be due to its recognised benefits and wide use for several decades as a cornerstone SLE therapy for many patients.^{11 12} In addition, it should be noted that we designated hydroxychloroquine as having ‘insufficient evidence in the literature to meet the specific criterion, but strong general indications of criterion met’ for disease modification at 5 years or later. In this case alone, we used survival data as a surrogate for SDI data; however, we do not suggest survival data be used in other cases as part of the disease modification definition, due in part to the duration required to gather these data. Belimumab was the only treatment to meet the confirmed disease modification criteria at >5 years as stated in the original criteria. Anifrolumab had available data for all criteria at year 1 and years 2–5 but lacked data for the year 5 or later confirmation time point, as the TULIP-LTE long-term extension only provided an additional 3 years of data following the first phase of the two phase 3 studies.⁶⁷ However, for other treatments, data as per the working definition were lacking or inconclusive in many instances. Criteria were designated as inconclusive for the following reasons: when comparisons were between high-dose and low-dose regimens (eg, glucocorticoids); inconsistent outcome measures compared with the proposed framework outcome definitions (eg, ‘major’ flares vs severe flares); conflicting results across outcome measures or across studies; or when significant reductions in steroids were reported, but not in relation to the ≤7.5 mg/day threshold stated in the criteria. Therefore, further research is needed for most treatments to fully evaluate their disease modifying ability.

A central question regarding the proposed framework for the definition of disease modification is whether a reduction in disease activity at year 1 should be considered disease modifying at this early time point. In this review and the published disease modification framework,¹⁰ disease modification is proposed to only be confirmed after year 5, with criteria before year 5 instead indicating potential for long-term disease modification. However, there may be value in categorising treatments as disease modifying at earlier time points to help determine as soon as possible when alternative therapy needs to be initiated. The early use of disease-modifying treatments could be important in advancing a treatment paradigm shift. Given a general lack of data >5 years for existing SLE treatments, the available data at earlier time points may serve as a surrogate for a drug’s potential to attain long-term disease modification potential. As such, longer-term data from cohorts, registries or pragmatic trials will be needed to confirm if this potential is realised over a longer time horizon. On the contrary, it should be noted that in our analysis, disease modification potential demonstrated at early time points did not always translate into long-term confirmed disease modification. For instance, the criteria fulfilled by glucocorticoids on disease activity at year 1 did not persist into years 2–5 given the change in criteria between these time points, and that glucocorticoids instead worsened criteria at later time points and do not have long-term disease modifying effects.

An additional question raised by our approach was whether any or all of the proposed criteria at year 1 and years 2–5 time points should be mandatory for disease modification potential to be fulfilled, or whether there should be a hierarchy in terms of the importance of individual criteria. For example, should a reduction in SLE disease activity be required throughout all time periods for a treatment to be classified as disease modifying? This aspect should be addressed in future studies using a consensus approach. Although some treatments carry substantial concerns around long-term toxicity and organ damage, they could still theoretically have disease modifying potential at earlier time points, as evidenced by the continued recommendations around glucocorticoid use as a ‘bridging therapy’ by EULAR.^{11 61} While on a strict reading of the ‘reduction in disease activity’ criteria, glucocorticoids may show potential for disease modification at year 1; in our collective expert view, the known and well-documented detrimental effects of glucocorticoids support the EULAR 2023 recommendation that they should be used as a bridging therapy,¹¹ as there will be a crossover point at which potential early disease modification benefits become outweighed by associated risks of irreversible organ damage. Since, in the case of glucocorticoids, the safety risks are closely related to the administered dose and duration of treatment, physicians should consider this when prescribing glucocorticoids early in the disease. The cumulative glucocorticoid dose throughout each time point would be an ideal and perhaps more accurate approach than requiring a reduction below a certain threshold at a set time point. However, unfortunately, we have found that studies do not often report cumulative glucocorticoid dose data, and typically provide dose values and averages at specific time points, currently making this approach unfeasible.

Hydroxychloroquine and belimumab had evidence of SDI reductions before 5 years, which raises the question of whether such evidence should be sufficient for treatments to be confirmed as disease modifiers without the need for data beyond 5 years, as is required in the current framework.¹⁰ It often takes several years for damage to begin to accumulate; for instance, a study of the Toronto Lupus Cohort found that mean SDI increased from 0.33 in year 1 to 1.90 at 15 years.⁸⁴ For this reason, our author group decided upon a threshold of >5 years, after which SDI can be evaluated to confirm disease modification. This is intended to ensure that any changes to the underlying pathophysiology are robust, sustained and meaningful. However, with the existing definition, a lack of data for all treatments other than hydroxychloroquine and belimumab at >5 years currently precludes the classification of these treatments as a confirmed disease modifier, although several available treatments have a high potential for long-term disease modification based on <5-year data. We recommend a more formal review of this complex issue by implementing a Delphi-based consensus approach involving a large and diverse panel of rheumatology and SLE experts, which could also consider the grades of evidence available for each treatment (eg, data from randomised controlled trials vs real-world observational studies).

Other considerations that we identified during our analysis in classifying existing treatments as disease modifying include the typical duration of treatment, their use as monotherapies versus combination therapies and application according to disease severity. For example, hydroxychloroquine is recommended by EULAR for all patients with SLE, which may be as monotherapy, or in combination with glucocorticoids, immunosuppressants or biologics.^{11 61} However, this complexity makes the classification of individual treatments as disease modifying difficult to resolve. Since combining more than one SLE therapy could increase the probability of achieving disease modification, classification of commonly used combination regimens, as opposed to individual treatments, as disease modifying could be considered in future studies. Additional concepts that could be developed further include whether it will be possible to directly compare disease modification potential between drugs. For example, a recent electronic health record-based SLE cohort study of preliminary data assessed the risk of organ damage progression after 5 years between belimumab and standard therapies, showing that belimumab was associated with a lower risk of damage progression over 5 years than MMF, but not methotrexate or azathioprine.⁸⁵ Whether dose or duration of treatment should be considered is yet another topic to be addressed.

Through this endeavour to examine the existing data for SLE treatments according to the published criteria for disease modification, it has raised the question of whether alterations should be made to the disease modification definition itself, which was proposed to include ‘slowing or preventing organ damage progression’. It must be decided whether ‘slowing’ (ie, lower rates of damage accrual at the end of the study period vs the comparator’s rate of damage accrual) is sufficient, or whether ‘delaying’ (ie, no damage accrual at the end of the study period) should be required. Additionally, in the case of lupus nephritis, ‘delaying or preventing progression to end-stage kidney disease’ may be too late of a measure in the disease course, and ‘progression to chronic kidney disease’ may be more appropriate, although the renal disease modification criteria was not the focus of this review. Finally, the reduction in immunosuppressants at year 1 criterion could be removed, as reducing immunosuppressants within 1 year is not recommended in treatment recommendations¹¹ or reflective of immunosuppressant use in clinical practice, where patients are typically maintained on immunosuppressants for at least 1 year.

We also considered whether the disease modification criterion should continue to require improvements across multiple organ domains, or whether improvement in a single domain is sufficient. Specifically, improvement in a single organ domain at years 2–5 could be appropriately viewed as indicative of disease modifying potential. It may be the case that different targeted therapies are disease modifying for different manifestations, depending on the pathogenesis of a given manifestation. Furthermore, the glucocorticoid criteria of the 2019 EULAR SLE treatment recommendations, used to develop the initial disease modification criteria, specified a threshold of ≤7.5 mg/day prednisone-equivalent,^{12 13} or ≤5 mg/day as per the Definition of Remission in SLE remission definition⁸⁶ to meet criteria; however, a change in the threshold to reduce glucocorticoids further than ≤7.5 mg/day or a reduction in glucocorticoid dose by ≥50% could also be considered. The updated EULAR 2023 recommendations also now suggest a threshold of ≤5 mg/day,¹¹ promoting support for an alteration of the glucocorticoid disease modification criteria to this dose level. However, as studies have more often used a ≤7.5 mg/day threshold as an endpoint than ≤5 mg/day, this alteration may limit the amount of relevant literature for evaluating the glucocorticoid disease modification criteria. A greater number of future studies may use ≤5 mg/day prednisone-equivalent as an endpoint following the recent updates to the EULAR recommendations.¹¹ Additionally, although severe flares are specifically mentioned in the current criteria at years <5, there may also be merit in evaluating data for all flares regardless of severity, given mild or moderate flares are more common. However, many experts would argue that mild or moderate flares may be too low a threshold to be considered for disease modification, but this topic is worthy of further deliberation. Severe flares can also be defined differently depending on the instrument used for measurement, adding further challenges when evaluating with the current proposed framework.

Limitations of this review include its narrative approach rather than a more formal systematic literature review, and the general lack of consistent study outcomes aligned with the proposed disease modification framework criteria. In the case of hydroxychloroquine, this necessitated a special case consideration for use of survival data as a surrogate for SDI data. We did not consider the timing of treatments (for instance, induction/initial or maintenance/subsequent) when evaluating treatments as disease modifying. We also did not distinguish between background treatments, or comparators, which were very variable between studies, and across the decades of included research. We cannot rule out the possibility that the results were impacted by changes over time in standard of care and dosing regimens (eg, glucocorticoid dosing), as the evaluated criteria had been devised according to current standard practices, which could have limited the relevance of older studies. However, we note that hydroxychloroquine, which has been used in SLE for many decades, had a positive disease modification classification despite the age of many of the cited studies, suggesting that the study date was not a critical factor in the classification of treatments. As previously noted, this review did not consider data pertaining to treatment tolerability and focused instead on available evidence for the published framework criteria. Although HRQoL, patient-reported outcome measures and biomarkers were suggested as supportive elements to the main disease modification definition,¹⁰ they were not included in this analysis. Further research should consider these aspects for currently available SLE treatments, as ACR, EULAR and the WHO recommend that HRQoL be more frequently considered in clinical trials and practice, given that impact to quality of life is often overlooked in favour of disease activity and organ damage measures.⁸⁷ Finally, this evaluation was funded by GSK, the manufacturer of belimumab. However, any risks of unconscious bias were mitigated through the involvement of several prominent SLE experts as authors. Each author contributed significantly to the approach, the study selection, data evaluation and application of the criteria.

Conclusions

This was the first attempt to classify disease-modifying treatments for SLE by applying criteria from the recently published working definition of disease modification in SLE. Based on the available literature, hydroxychloroquine and belimumab are currently the only two treatments for SLE that satisfy disease modifying criteria. However, several other treatments do have data indicating disease modification potential, and these will need formal evaluation in longer-term studies. We have also identified some aspects of the existing working framework that could be revised, but we believe a consensus-style approach with a larger group of experts is required to refine the definition and fully address disease modification in current SLE treatments.

Ethics statements

Patient consent for publication

Not applicable.

Ethics approval

Not applicable.