Relevant papers were identified in MEDLINE, EMBASE and CINAHL bibliographic databases following the initial SHARE methodology. Studies were eligible for inclusion if they fulfilled the following criteria: (a) published in English language; (b) from 2012–September 2021; (c) original research studies regarding management of JSLE, LN and/or childhood-onset APS; and either (d) pediatric studies or e) adult SLE studies meeting the following criteria were included: (i) 3 or more patients <18 years of age, or (ii) when lower limit of age range ≤16 years: include if study has more than 15 patients AND a mean/median age ≤30 years.

Publications were excluded for the following reasons: (a) publications on other diseases (e.g., vasculitis, adult SLE alone); (b) with a focus on aspects other than management; (c) case report with <3 patients; (d) conference abstracts only or full text unavailable; e) reviews; (f) adult studies not fulfilling age criteria; (g) non-human data; and h) not published in English. Further literature searches were performed to assess if there was any specific new evidence within the pediatric or adult SLE literature related to each of the existing SHARE treatment recommendations for JSLE in general, LN and/or APS.

Three reviewers (PP, KG, ES) independently screened all publications (titles, abstracts) according to the studies predefined inclusion/exclusion criteria. All relevant publications were retrieved and reviewed by at least two of the three reviewers. For publications where the age criteria were unclear, corresponding authors were contacted. Of the 10 authors contacted, three provided further information.

Data extraction was performed using the original SHARE data extraction sheet for treatment (see Supplemental File 1). The extraction sheet included: general study information, study population characteristics, study methods, results, conclusions/discussions, validity assessment and category of evidence. Data was collected by one author and reviewed by two more authors independently. For publications where there was any uncertainty, a face-to-face discussion was held.

Figure 1 summaries the results of the literature search in JSLE and LN. Both searches contained the term “lupus” (lupus nephritis AND juvenile systemic lupus erythematosus), therefore the same publications were captured by each search (n = 1,100). For JSLE in general, after screening of the titles and abstracts, 225 publications were identified as relevant to the management of JSLE. Full text publications were then assessed against the inclusion/exclusion criteria, with 6 publications identified as meeting the full criteria (two of which related to the same trial). For LN, after screening of the titles and abstracts, 118 publications were identified as relevant to the management of LN. Of those, 16 publications met the full inclusion/exclusion criteria. Figure 2 summaries the results of the literature search in APS. The literature search produced 395 publications. After screening of the titles and abstracts, 55 publications were identified as relevant to the management of APS but no publications met the full inclusion/exclusion criteria.

The SHARE recommendations state “all children with lupus should be on hydroxychloroquine routinely” (1). No new pediatric studies relating to this recommendation could be identified. In adult-SLE, a Canadian cohort study reported that “more consistent” use of antimalarials in the first 5-years following SLE diagnosis (defined as patient reported antimalarial agent use >60% of the time) was associated with reduced risk of Systemic Lupus International Collaborating Clinics/American College of Rheumatology Damage Index (SLICC-SDI) score defined damage, increased achievement of low disease activity (defined as a clinical-SLEDAI-2K score of ≤2, not including serology) and reduced cumulative glucocorticoid dose after 5-years of follow-up.

A study examining long-term outcomes in Dutch adults with JSLE has also shown current hydroxychloroquine monotherapy to be associated with absence of SLICC-SDI defined damage (23). A large population based study using hospital episode statistics and national death certificates (from 1987 to 2012) has demonstrated that hydroxychloroquine use is associated with a 45% reduction in the hazards of mortality in adult-SLE (24). Collectively these more recent adult-SLE studies support this SHARE recommendation.

The increasing evidence regarding long term side effects of hydroxychloroquine should however be considered. Due to advances in eye screening, hydroxychloroquine retinopathy has been found to be more common than previously thought, with a 2014 study showing a prevalence of 7.5% in adult-SLE patients taking hydroxychloroquine for a minimum of 5-years. A total daily dose of >5 mg/kg (using actual body weight) was found to be associated with increased risk. Renal impairment and tamoxifen therapy also increased the risk of retinopathy (25). The SHARE recommendations currently advise that yearly eye screening should be “considered” in children taking hydroxychloroquine (1). The UK Royal College of Ophthalmologists guideline (for adult use) states that when long-term hydroxychloroquine treatment is planned, patients should receive a baseline examination (within 12-months), followed by annual screening from year 5 of treatment onwards. In patients with additional risk factors for retinopathy (e.g., Tamoxifen use, impaired renal function (estimated glomerular filtration rate of <60 ml/min/1.73 m²), hydroxychloroquine dose >5 mg/kg/day) annual monitoring from baseline is recommended. Despite a lack of evidence in patients <18-years, these guidelines advise that these patients on long term hydroxychloroquine should also be referred for monitoring as per the adult criteria (26).

SHARE advises that “in all decisions of treatment change or modification, compliance should be actively checked”. Patient compliance is known to be a challenge in JSLE, with a previous study showing only 32% of adolescents and young adults with SLE to be compliant with hydroxychloroquine (27), and a further study demonstrating 43–75% of adult-SLE patients to be non-compliant with hydroxychloroquine (28). Despite this, drug levels are not routinely monitored in clinical practice. The 2012 PLUS study found that tailoring hydroxychloroquine dose to a target therapeutic blood concentration did not reduce the frequency of SLE flares. However, hydroxychloroquine blood levels increased spontaneously between study inclusion and randomization, suggesting improved adherence to hydroxychloroquine treatment in all patients, likely in response to the information that was sent to patients about the study. This study therefore concluded that despite the trial not meeting its primary endpoint, routine measurement of hydroxychloroquine levels may improve patient adherence to treatment (29). A recent meta-analysis found a good association between whole-blood hydroxychloroquine levels and reported non-adherence (30).

In 2018, a web-based education and a social media intervention was shown to significantly improve adherence to medications in adolescents and young adults with SLE (31). Within this study, self-reported medication adherence was significantly higher than objectively measured indices of adherence (e.g., the medication possession ratio), highlighting the need for objective measures of adherence, such as blood levels. Novel approaches such as social media interventions may help empower patients to manage their own medications effectively (31).

In relation to corticosteroid therapy, the SHARE recommendations advise that “when it is not possible to taper the prednisone dose, a DMARD should be added to the therapy”. No new original pediatric research was found relating to this recommendation. However, the 2019 update to the EULAR recommendations for SLE also advise addition of methotrexate, azathioprine, or mycophenolate mofetil (MMF) in patients whose symptoms are not controlled with corticosteroids and hydroxychloroquine. Cyclophosphamide is suggested for severe organ threatening or life-threatening SLE or for patients who do not respond to other immunosuppressive agents. These recommendations advise that belimumab is used for patients with frequent relapses or those not able to taper steroid dose despite the above standard of care. Consideration of rituximab is suggested for organ-threatening disease refractory to standard immunosuppressive agents or where these are contraindicated or not tolerated (32).

SHARE advises that “in mild/moderate hematological involvement: when haemolysis is present and Hb is lower than normal, a DMARD should be added to the therapy”. In 2015, a retrospective cohort study assessed 24 JSLE patients treated with rituximab for refractory cytopenias, 19 of whom had haemolytic anemia. Overall, 96% of patients showed complete response after the first course of rituximab (defined as Hb >120 g/L for haemolytic anemia and platelet count >100 x 10⁹/L for patients with thrombocytopenia). The median time to complete response for patients with haemolytic anemia was 85 days (33). A Turkish study examining hematological involvement in JSLE also found benefit from rituximab in cases of haemolytic anemia resistant to steroid and intravenous immunoglobulin (IVIG) treatment (34).

The SHARE recommendations state that “if rituximab is required, the recommended dose is either 750 mg/m²/dose (up to a maximum of 1 g) at day 1 and day 15, or 375 mg/m²/dose once a week for four doses”. No new pediatric research was identified examining dose regimes for rituximab. In 2014, a UK study of rituximab use in patients with JSLE over a 10-year period (2003–2013) of 63 patients, all received a dose of 750 mg/m²/dose ~2-weeks apart (35). The new evidence relating to general JSLE SHARE treatment recommendations is summarized in Supplemental Table 1.

JSLE SHARE treatment recommendations also included those directed toward neuropsychiatric manifestations (1). They recommend that “When neuropsychiatric manifestations are caused by an immune or inflammatory process and non-SLE-related causes are excluded, corticosteroids and immunosuppressive therapy are indicated”. Limited new evidence could be found relating to this recommendation in JSLE. A retrospective study of 144 children with autoimmune and inflammatory disorders of the central nervous system (CNS, 18/144 with NP-SLE) treated with rituximab demonstrated “definite” clinician-defined improvement with rituximab in 5/18 patients, “probable” in 7/18, “possible” in 5/18 and “no improvement” in 1/18 patients (36). A 2013 study of Saudi children included two cases of lupus cerebritis which improved with combined rituximab and cyclophosphamide treatment (37). A Chinese study of 20 children with SLE reported that in 10/20 (50%) cases, delirium and cognitive disorders improved after one-month of rituximab treatment (38). An Indian study of 88 adult patients with NP-SLE treated with MMF and deflazacort showed complete response (defined as complete resolution of initial neuropsychiatric signs and symptoms) in 83.9% of patients at 1-year follow up, and in 92.3% of patients at last follow up (median 33-months) (39).

SHARE also recommended that “antiepileptic drugs are usually not necessary after a single seizure in the absence of MRI lesions and definite epileptic abnormalities on EEG following recovery from the seizure. Long-term antiepileptic therapy should be considered for recurrent seizures” (2). No new evidence could be found relating to this recommendation. Overall, the evidence base for NP-SLE treatment in children is minimal and further studies are needed. The EULAR adult-SLE guidelines for NP-SLE are in-keeping with the SHARE recommendations, namely that “Treatment of SLE-related neuropsychiatric disease includes glucocorticoids/immunosuppressive agents for manifestations considered to reflect an inflammatory process” (32). The final SHARE recommendation is that “there is a need for pediatric NP-cSLE research regarding treatment” (2), and this clearly continues to be the case. New evidence relating to neuropsychiatric JSLE SHARE treatment recommendations is summarized in Supplemental Table 1.

Sixteen studies relating to management of LN in JSLE were identified: two pediatric studies, one adolescent study and 13 adult studies including JSLE patients. Tables 2–4 summarizes these studies, and the key findings are discussed below.

Pediatric evidence–The largest exclusively pediatric study is from the UK JSLE Cohort Study, comparing MMF and cyclophosphamide as induction treatments for proliferative LN. 34/51 (67%) of patients received MMF and 17/51 (33%) received cyclophosphamide (56). No significant differences were identified between the treatment groups in terms of their numerical BILAG disease activity scores, urine albumin/creatinine ratio, serum creatinine, ESR, anti-dsDNA antibody, C3 levels and patient/physician global scores at 4–8 months, 10–14 months after renal biopsy, and at last follow up. There were also no differences in SLICC-SDI scores at 13-months, and last follow up. The time to achievement of inactive LN, and time to subsequent renal flare was also comparable between the groups (Table 2) (56). Despite being the largest JSLE study comparing MMF/cyclophosphamide LN induction treatment to date, this study was limited by its numbers, particularly in comparison to adult SLE studies.

Mixed pediatric and adult SLE evidence–In a study including 35 Nepalese patients (mean age 25.43 ± 10.17–years), Sedhain et al. demonstrated low dose MMF (maximum daily dose of 1.5 g) to be comparable to monthly cyclophosphamide pulses (dose of 0.5–1g/m²) as induction treatment for proliferative LN (41). Both treatments led to similar reductions in proteinuria, improvements in kidney function (serum creatinine, eGFR) and achievement of complete remission, with less adverse events in the MMF group. Rathi et al. randomized 100 SLE patients (mean age 28.3-years) to low dose cyclophosphamide (six fortnightly infusions of 500 mg each) or MMF (daily doses 1.5–3g), accompanied by three intravenous methylprednisolone infusions, followed by oral corticosteroids. Maintenance therapy of azathioprine and low-dose corticosteroid was started after 6-months of induction therapy. They demonstrated similar rates of treatment response in each group (p = 1.0), with complete renal remission achieved by 50% of the cyclophosphamide group, and 54% of the MMF group (p = 0.84) after 24-weeks treatment. They concluded that low dose cyclophosphamide is comparable in safety and efficacy to oral MMF as an induction treatment for less severe LN (class III, IV, V LN, but excluding those with crescentic LN or a serum creatinine over 265 μmol/l) (42). Mendonca et al. conducted a 24-week prospective, randomized, open-label trial comparing oral MMF with monthly IV cyclophosphamide as induction therapy for active biopsy proven Class III and IV LN, in 40 Indian patients. MMF and cyclophosphamide were demonstrated to be comparable in terms of the rates of complete remission, partial remission and cumulative probability of response at 6 months (Table 2) (43).

The Aspreva Lupus Management Study (ALMS trial) published it's result in late 2011 (57). ALMS was a large, multinational, prospective, two-stage, parallel-group, phase III RCT including patients with LN>12-years old. In the first phase of the study (24-week induction), patients were randomized to oral MMF (target dose 3 g/day) or intravenous cyclophosphamide (0.5–1 g/m²/month), plus prednisone. Responders then went into phase 2 of the study (36-month maintenance), where patients were randomized 1:1 to MMF (1.0 g, bd) or oral azathioprine (AZA) (2 mg/kg/day), plus prednisone. Of the 370 patients enrolled, 24 were aged <18-years (mean age of 14.8, standard deviation 1.48-years), and the results of sub-analyses for this age group were published in 2012, showing induction treatment with MMF and intravenous cyclophosphamide to be equally efficacious. During the maintenance phase, MMF was demonstrated to be at least as effective as azathioprine. The results of the JSLE patients were largely comparable to those of the adult SLE patients, but of note, adolescent patients more commonly developed serious infections, regardless of the treatment arm (Table 2) (44). Whilst the results of this study are encouraging, larger trials involving purely pediatric and adolescent study populations are needed.

A single center RCT has compared high dose cyclophosphamide (six four-weekly cycles of 750 mg/m², maximum of 1.5 g/pulse) with low dose cyclophosphamide (six fortnightly cycles of 500mg). This study included 75 proliferative LN patients (mean age 30.7±10.04-years, standard deviation years in the low dose cyclophosphamide group, and 27.24±10.60-years in the high dose group). At 52-weeks, high dose cyclophosphamide was shown to be more effective than low dose in achieving a partial and complete response (73 vs. 50%, p = 0.04), and in preventing LN relapse (3 vs. 24%, p = 0.01). There was a significantly lower number of non-responders in the high dose cyclophosphamide group (27 vs. 50%, p = 0.04, Table 2) (45). Further studies involving multiple centers and younger patients are required, as this was single center study involving both JSLE and adult SLE patients.

A JSLE and adult-SLE Nepalese prospective observational study has assessed the performance of an unconventional cyclophosphamide regimen which differs to the more commonly used Euro Lupus (500 mg every 2-weeks for 3-months) (58) or the National Institute of Health (NIH) regimens (0.5–1g/m² monthly for 6-months) (59). In the Nepalese study 500 mg of cyclophosphamide was given per month, for 6-months. The study included 41 patients with a mean age of 26.9 ± 10.6-years, with biopsy proven class III, IV, V, or mixed III/IV+V LN. 43.9% of patients achieved complete remission and 39% achieved partial remission (overall response rate of 83%) using this cyclophosphamide regimen (Table 2) (46). The overall response rate of 82.9% is comparable to those of the Euro Lupus trial where 71% of the low-dose cyclophosphamide group achieved renal remission (58) and the NIH trial where 85% achieved renal remission (60).

Intravenous cyclophosphamide is usually combined with high dose intravenous methylprednisolone or oral corticosteroids for the management of LN. A prospective observational study evaluated the use of IV cyclophosphamide without additional methylprednisolone/high dose oral prednisolone in patients presenting with their first episode of LN. In this study, the use and dose of prednisone was based solely on the presence of mild to moderate extrarenal SLE manifestations, with dose tapering decided upon based upon extrarenal activity alone. Fourty patients with a mean age of 29.7 ± 10.1-years received 12 IV cyclophosphamide pulses over 24-months (6-monthly pulses, and six quarterly pulses). The initial cyclophosphamide dose was 0.5 g/m²; subsequent doses were increased by 250 mg, with a maximum of 1,500 mg per pulse. After 24-months, 62.5% of patients met the criteria for complete renal response and 20% met the criteria for partial renal response. Mean starting dose of prednisone was 23.9+/-23.8 mg/day. Post-hoc analysis compared outcomes for patients treated with prednisone doses ≥20 mg/day (Group A, n = 19) and <20 mg/day (Group B, n = 21). Complete renal response was achieved in 52.6% of Group A patients vs. 71.4% of Group B patients (p = 0.37); and partial renal response was seen in 26.3 vs. 14.3% of group A and B patients respectively (p = 0.58). Overall, renal outcomes were the same irrespective of initial prednisone doses (p = 0.46, Table 2) (47). These findings warrant further exploration in JSLE, ideally within a randomized trial comparing different corticosteroid dosing regimens in children and young people with LN.

A pediatric study including 44 JSLE patients with active LN (ISN RPS class III/IV/V) aged 3.5–13.8-years (median 8.4) compared outcomes in patients treated with induction treatment consisting of methylprednisolone followed by either rituximab (n = 17), MMF (n = 12) or IV cyclophosphamide (n = 15), with a tapering dose of oral prednisolone. MMF was added as maintenance immunosuppression (800 mg/m² daily) in all children from 3-months. At 36-months, flare-free survival was highest in the rituximab group than other treatment groups (100% RTX vs. 83% MMF vs. 53% for CYC, p = 0·006). The mean daily dose of prednisolone was also significantly lower in the rituximab group after 3-months (rituximab vs. MMF, p = 0.005; rituximab vs. cyclophosphamide, p = 0.0001). There was a numerical difference in the proportion of patients achieving complete remission (76.5% achieved complete remission with rituximab, 41.7% with MMF and 46.7% with cyclophosphamide, however this did not reach statistical significance (p = 0.28, Table 3) (48).

In a study including 144 JSLE and adult SLE patients (mean age 30.6-years), Rovin et al. evaluated the efficacy and safety of rituximab in a randomized, double-blind, placebo-controlled phase III trial in patients with class III and IV LN treated concomitantly with MMF and corticosteroids. Patients received rituximab (1,000 mg) or placebo on days 1, 15, 168, and 182. The primary end point of the study was assessment of renal response status at 12-months. The overall (complete and partial) renal response was 56.9% in the patients treated with rituximab (in addition to a background of MMF and corticosteroids) and 45.8% in the patients receiving placebo (p = 0.18). Of note, more placebo treated patients required cyclophosphamide rescue therapy during the 12-months of follow-up, and there were significantly greater reductions in anti-dsDNA and C3/C4 levels in patients receiving rituximab. The study also showed that combination of rituximab with MMF and corticosteroids did not result in any new or unexpected safety alerts (Table 3) (49).

In 2016, Tanaka et al. investigated the efficacy and safety of rituximab in an open-label study of 34 Japanese patients with active SLE (17/34 with LN) who had been refractory to conventional therapy. The study included JSLE patients who were >16-years, but did not specify how many were recruited or the mean age of the study population. 76.5% of these previously refractory patients responded to rituximab therapy at week 53; with 47.1% achieving remission (defined as a change from British Isles Lupus Assessment Grade (BILAG) A or B score to a BILAG C or D score in every organ system) and 29.4% achieving partial remission (change from a BILAG A or B score to a C or D score in at least one organ system, but with presence of one BILAG A or B score in another organ system). In the patients with LN, 52.9% of patients demonstrated an overall renal response (29.4% complete renal response, 23.5% partial renal response) at 52-weeks. The response rate was higher in patients with biopsy proven class III/IV LN than other LN patients. A significant reduction in prednisolone was observed following rituximab treatment (45 mg/day, inter-quartile range, IQR: 35–55) at baseline to 6 mg/day (IQR: 5–9) at week-53. Most adverse events were graded mild to moderate, however there were a few serious adverse events (cerebral infarction, cholecystitis, endometritis, and hypoferric anemia) which were likely associated with the underlying diseases/concomitant illnesses rather than rituximab (Table 3) (50).

Ocrelizumab is a recombinant humanized monoclonal antibody that selectively targets and depletes CD20+ B-cells in the peripheral circulation. A randomized, double blind phase III study has compared patients treated with placebo or IV ocrelizumab (either 400 or 1,000 mg) in addition to standard care, which comprised of corticosteroids plus either MMF or Euro-Lupus regimen treatment (cyclophosphamide induction and azathioprine maintenance treatment). The overall renal response rates were not significantly different between treatment groups, and ocrelizumab was associated with a higher rate of serious infections leading the study to be terminated early (Table 3) (51).

Calcineurin inhibitors such as cyclosporin A and tacrolimus have been investigated in several studies, both in isolation and as part of a multitarget regimen in adult SLE (61–64). In 2013 Tanaka et al. published a small open-label, prospective, long-term tacrolimus-based treatment study involving 19 young patients (mean age 18-years) with biopsy proven LN. 15/19 (79%) had a history of LN and experienced a “lupus flare”, defined as a sustained increase in urinary protein excretion by more than 25% of the baseline value, associated with a significant decrease in serum C3 levels and/or increase in the serum anti-dsDNA antibody titer, and/or other signs of active SLE. Their usual cytotoxic was discontinued and replaced by tacrolimus (3 mg/day) with concomitant prednisolone (maximum 30 mg). 4/19 patients had new onset LN and were treated with a multitarget regimen consisting of Tacrolimus plus mizoribine (selective inhibitor of inosine monophosphate dehydrogenase in the purine synthesis pathway, acting in a similar manner to MMF) in combination with prednisolone. 12/19 (63%) achieved a complete renal response and 5/19 (26%) demonstrated a partial response, with two patients showing no response. There were no serious adverse effects (Table 4) (52). This study is of interest as Tacrolimus use is not reported frequently in JSLE, however, it is clearly limited by the sample size, lacks a wash out period and blinding.

A small Chinese study including 70 JSLE and adult SLE patients (all >16-years old, mean age of the study population not specified) has compared maintenance tacrolimus and azathioprine treatment, showing similar, low LN relapse rates in both treatment arms, with tacrolimus demonstrating a more favorable safety profile than azathioprine (53). In another small study from China, 26 patients with LN and persistent proteinuria of >1.5g/24-h despite treatment with cyclophosphamide (>8 g in <6-months), were commenced on 2–3mg of tacrolimus daily. 23/26 patients demonstrated an overall renal response (10 complete and 16 partial renal response). Most patients had biopsy confirmed LN (class III=5, class IV=2, class V=5, class III+V=7, class IV+V=4 and unknown n = 3), with patients with class V LN demonstrating higher rates of remission (Table 4) (54). Further research is required in JSLE to evaluate the role of tacrolimus in studies that are sufficiently powered.

In a small randomized controlled trial including 18 JSLE and SLE patients (mean age 29-years) with WHO class III/IV LN, no additional effect was seen in those treated with human umbilical cord-derived mesenchymal stem cells over and above standard immunosuppression (intravenous methylprednisolone and cyclophosphamide followed by maintenance oral prednisolone and MMF, Table 4) (55). The trial was stopped early when it became clear that it would not demonstrate a positive treatment effect.

This review highlights that treatment paradigms in JSLE are often needing to be extrapolated from adult SLE, whilst RCTs in JSLE are particularly scarce, especially any that are sufficiently powered to demonstrated statistical significance. Most available treatment options are not targeted (conventional DMARDs), and known to cause significant associated adverse events and toxicity, particularly in vulnerable children and young people (40, 77). Although biologic therapies are used extensively for many autoimmune conditions, there have been several notable setbacks in developing a robust evidence base for SLE, with only belimumab so far licensed for use in SLE in the past 50-years (40). Difficulties with definitions and use of outcome measures in SLE clinical trials have contributed to these setbacks. The Belimumab in JSLE (PLUTO trial) summarized above (12) raises important questions about the applicability of adult SLE outcome measures in JSLE. In this trial, the adult SLE primary outcome measure (SRI4) was not met in the pediatric age group, but the pediatric-derived major secondary outcome measure (PRINTO/ACR 30, 50) was achieved. Given the known differences in disease activity, severity and damage demonstrated between pediatric, adolescent, and adult SLE (78, 79), it is important that lessons are learnt from such studies.

Most of the more recent published evidence relate to treatment of LN, with a marked dearth of studies on NP-JSLE and APS. T2T approaches are hoped to offer an opportunity to improve further the clinical management of JSLE patients by using existing treatments in a structured way with the aim of more aggressively controlling disease activity at an early stage, preventing organ damage and improving HRQOL (7, 8). Such approaches are already part of routine clinical care in many areas of adult medicine (e.g., rheumatoid arthritis, hypertension, diabetes) (80), with growing international evidence for the potential role of T2T in JIA in recent years (10, 81, 82). Development and testing of such an approach as part of the TARGET LUPUS research programme is eagerly awaited.