Efficacy and Safety of Nonbiologic Immunosuppressants in the Treatment of Nonrenal Systemic Lupus Erythematosus: A Systematic Review

Authors


Abstract

Objective

To analyze the efficacy and safety of nonbiologic immunosuppressants in the treatment of nonrenal systemic lupus erythematosus (SLE).

Methods

We conducted a sensitive literature search in Medline, Embase, and the Cochrane Central Register of Controlled Trials up to October 2011. The selection criteria were studies including adult patients with SLE, a treatment intervention with nonbiologic immunosuppressants, a placebo or active comparator group, and outcome measures assessing efficacy and/or safety. Meta-analyses, systematic reviews, clinical trials, and cohort studies were included. The quality of each study was evaluated using Jadad's scale and the Oxford Levels of Evidence.

Results

In total, 158 of the 2,827 initially found articles were selected for detailed review; 65 studies fulfilled the predetermined criteria. Overall, the studies were low quality, with only 11 randomized controlled trials (RCTs). Cyclophosphamide demonstrated efficacy for neuropsychiatric SLE, preventing relapses with an additional steroid-sparing effect, although its use was associated with cumulative damage, development of cervical intraepithelial neoplasia, and ovarian failure. Other immunosuppressants (azathioprine, methotrexate, leflunomide, mycophenolate mofetil, and cyclosporin A) demonstrated efficacy in reducing nonrenal activity and flares with a steroid-sparing effect, although only on occasion in non–placebo-controlled RCTs of small numbers of patients.

Conclusion

Several immunosuppressants demonstrated their safety and efficacy in nonrenal SLE. A specific drug for each particular manifestation cannot be recommended, although cyclophosphamide may be used in more severe cases, and methotrexate may be the first option in most cases of moderately active SLE. High-quality RCTs of larger numbers of patients are needed.

INTRODUCTION

Systemic lupus erythematosus (SLE) is a heterogeneous disease that may affect all organ systems. The disease may be highly active, requiring aggressive therapy in one system or a few systems but inactive in all the others. To date, there have been few randomized controlled trials (RCTs) addressing the efficacy and safety of the available treatments for specific manifestations of SLE. The first controlled clinical trials were conducted in the 1980s and 1990s by researchers from the Mayo Clinic and National Institutes of Health and focused primarily on lupus nephritis ([1-3]). Over the past decade, most RCTs that investigated the effect of different therapies on lupus, mainly cyclophosphamide (CYC), azathioprine (AZA), and mycophenolate mofetil (MMF), also focused primarily on lupus nephritis ([4-10]). Recently, several biologic therapies for the treatment of renal and nonrenal manifestations of SLE have been studied ([11-18]). The heterogeneity of the disease means that a trial showing that a treatment is effective against one manifestation of lupus cannot necessarily be extrapolated to other types of organ involvement. To date, few RCTs have analyzed the effect of the available therapies on nonrenal manifestations of SLE. Specifically, there is limited evidence on the efficacy of nonbiologic immunosuppressants for the treatment of nonrenal manifestations of the disease. Despite this, several nonrenal manifestations of SLE are frequently treated with off-label medications such as methotrexate (MTX), MMF, and so forth. Thus, the objective of this study was to systematically review the available literature regarding the efficacy and safety of nonbiologic immunosuppressive therapies in the treatment of nonrenal manifestations of SLE.

Box 1. Significance & Innovations

  • There are only 11 randomized controlled trials (RCTs) assessing the effect of different nonbiologic immunosuppressants in the treatment of nonrenal systemic lupus erythematosus (SLE).
  • In those circumstances, several immunosuppressants have demonstrated their safety and efficacy in reducing nonrenal activity with a steroid-sparing effect.
  • Cyclophosphamide in more severe cases and methotrexate in most cases of moderately active SLE may be the first therapeutic option.
  • High-quality RCTs of larger numbers of patients are needed.

MATERIALS AND METHODS

This study was performed by experts of the Evidence-Based Medicine Study Group and the Systemic Autoimmune Diseases Study Group of the Spanish Society of Rheumatology.

Search strategy

The studies were identified by sensitive search strategies in the following main bibliographic databases: Medline from 1961 to October 2011, Embase from 1980 to October 2011, and the Cochrane Central Register of Controlled Trials up to October 2011. An expert librarian checked the search strategies. Finally, a hand search was performed by reviewing the references of the studies included. Details about the strategies are shown in Supplementary Appendix A (available in the online version of this article at http://onlinelibrary.wiley.com/doi/10.1002/acr.22035/abstract).

Selection of studies

Initially, we performed a systematic literature review on the efficacy and safety of biologic and nonbiologic immunosuppressive drugs in the treatment of nonrenal manifestations of SLE. The studies retrieved by the mentioned strategies were included if they met the following preestablished criteria: 1) the population was adult patients diagnosed with SLE; 2) the intervention was treatment with a nonbiologic immunosuppressive agent; 3) the comparator was a placebo or active comparator; 4) the outcome measures assessing efficacy were nonrenal manifestations, scores by activity indices, SLE flares, a steroid-sparing effect, and so forth; and 5) the outcome measures assessing safety were infections, cardiovascular events, malignancies, and so forth. Only meta-analyses, systematic reviews, clinical trials, and cohort studies were included. We excluded studies specifically about efficacy in lupus nephritis or discoid lupus, studies assessing antimalarials or biologic therapies, and studies with insufficient data for analysis.

Screening of studies, data collection, and analysis

The titles and abstracts of all articles retrieved by the search strategy were independently reviewed for selection criteria by 2 reviewers (TC-I and EL-S). The reviewers collected the data from the included studies by using ad hoc standard forms. One of the reviewers (TC-I) entered the data from the forms into spreadsheets. In case of any discrepancy in the information between both reviewers, a consensus was reached by reading the original article or by asking the mentor.

The level of evidence and grades of recommendation were established by a reviewer (TC-I) based on a scale from the Oxford Centre for Evidence-Based Medicine ([19]); Jadad's scale was additionally used to grade quality in case the study was an RCT ([20]) (for definitions, see Supplementary Appendix B, available in the online version of this article at http://onlinelibrary.wiley.com/doi/10.1002/acr.22035/abstract). Evidence tables were produced with all the included studies and a qualitative analysis was performed (see Supplementary Appendix C, available in the online version of this article at http://onlinelibrary.wiley.com/doi/10.1002/acr.22035/abstract).

RESULTS

The complete literature search produced 2,827 items. After removing 359 items that were duplicated, 2,468 studies were reviewed and 65 were ultimately included. The result of the search strategies is shown in Figure 1. The excluded studies and reasons for exclusion are shown in Supplementary Appendix D (available in the online version of this article at http://onlinelibrary.wiley.com/doi/10.1002/acr.22035/abstract). Most of the items included were cohort studies and only 11 were RCTs. The detailed information from these RCTs is shown in Table 1.

Figure 1.

Flow chart of the results from the search strategies.

Table 1. RCTs assessing nonbiologic immunosuppressants in the treatment of nonrenal SLE*
Author, year (ref.), design, followupPopulationInterventionOutcome measureEfficacySafetyComments
  1. RCT = randomized controlled trial; SLE = systemic lupus erythematosus; IC = inclusion criteria; NPSLE = neuropsychiatric SLE; EC = exclusion criteria; LN = lupus nephritis; IV = intravenous; CYC = cyclophosphamide; PRD = prednisone; EEG = electroencephalogram; EP = evoked potentials; AEs = adverse events; PAH = pulmonary arterial hypertension; sPAP = systolic pulmonary artery pressure; COPD = chronic obstructive pulmonary disease; ↓ = decrease; NYHA = New York Heart Association; RR = relative risk; 95% CI = 95% confidence interval; GI = gastrointestinal; NP-APS = NP antiphospholipid syndrome; CNS = central nervous system; MP = methylprednisolone; IS = immunosuppressants; RIFLE = responder index for lupus erythematosus; AZA = azathioprine; PCB = placebo; MTX = methotrexate; ↑ = increase; SLEDAI = Systemic Lupus Erythematosus Disease Activity Index; VAS = visual analog scale; SLAM-R = Systemic Lupus Activity Measure-Revised; SDI = Systemic Lupus International Collaborating Clinics/American College of Rheumatology Damage Index; NSAIDs = nonsteroidal antiinflammatory drugs; PRDL = prednisolone; LEF = leflunomide; HCQ = hydroxychloroquine; anti-dsDNA = anti–double-stranded DNA; MMF = mycophenolate mofetil; BILAG = British Isles Lupus Assessment Group; SELENA–SLEDAI = Safety of Estrogens in Lupus Erythematosus National Assessment–SLEDAI; CSA = cyclosporin A.
Stojanovich et al, 2003 ([21]), open RCT, ≥6 moN = 60 IC: NPSLE EC: LN, othersIV CYC (200–400 mg/mo × 6 mo) + PRD 20.5 mg/d (n = 37) vs. PRD 20.5 mg/d (n = 23)Clinical improvement, relapses, EEG improvement, EP improvement, AEsClinical improvement at 6 mo: 62.2% vs. 21.7% (P = 0.005); relapses at 3 mo: 37.8% vs. 78.3% (P = 0.005); EEG recovered: 75% vs. 18% (P = 0.003); EP recovered: 80% vs. 0% (P = 0.003)2 herpes zosterOxford 3a Jadad 1
Gonzalez-Lopez et al, 2004 ([22]), open RCT, 6 moN = 34 IC: PAH due to SLE (sPAP >30 mm Hg) EC: embolism, pulmonary fibrosis, asthma, COPD, and othersIV CYC: 0.5 g/m2/mo × 6 mo (n = 16) vs. enalapril 10 mg/d × 6 mo (n = 18)↓ sPAP, NYHA functional class improvement, AEs↓ sPAP: from 41–28 mm Hg (P < 0.001) vs. 39– 27 mm Hg (P = 0.02); significant difference (P = 0.04); only CYC improves NYHA class (P = 0.02)CYC: more infections (RR 1.6 [95% CI 1.001, 2.47]) and more GI AEs (RR 14.6 [95% CI 2.15, 99.7])Jadad 3
Barile-Fabris et al, 2005 ([23]), open RCT, 24 moN = 32 IC: new-onset NPSLE EC: NP-APS, CNS infection, metabolic encephalopathy, and othersIV MP (1 g/d × 3 d) + IV CYC (0.75 g/m2/mo × 12 mo, then every 3 mo × 1 year) (n = 19) vs. IV MP (1 g/d × 3 d + MP 1 g/d × 3 d/mo × 4 mo, then 3 d/2 mo × 6 mo, then 3 d/3 mo × 12 mo) (n = 13) Comedication: PRD 1 mg/kg/d and ↓Response to treatment: ≥20% improvement in clinical, serologic, and neurologic measures, AEsResponse rate at 2 y: 95% vs. 46.2% (P < 0.03)AEs: no differenceJadad 3 36.8% (CYC) and 76.9% (MP) were lost to FU
Petri et al, 2010 ([24]), open RCT, 30 moN = 47 IC: moderate/severe SLE and lack of response to moderate- to high-dose steroids or ISIV CYC 0.75 g/m2/mo × 6 mo, then every 3 mo × 2 y (n = 26) vs. high-dose IV CYC 50 mg/kg × 4 d (n = 21)RIFLE (complete or partial response, no change, or worsening), AEsComplete response at 30 mo: 65% and 48%; partial response at 30 mo: 10% and 19% (P = not significant for both, overall and by major organ system)No difference in serious AEs, hospitalizations, infections, deaths, and ovarian failureJadad 3 Oxford 2b
Hahn et al, 1975 ([50]), open RCT, 24 moN = 24 IC: active, life-threatening SLE, and no IS or >20 mg/d the preceding 6 weeks EC: drug-induced SLEAZA 3–4 mg/kg + PRD 60 mg/d (n = 11) vs. PRD 60 mg/d (n = 13) Comedication: PRD 40–60 mg/d × 4–6 mo; if failure to respond: double PRD dose 4–6 w; if no response: removal from study; if response: ↓ PRDClinical improvement, mean PRD dose, AEsClinical improvement: no difference at 3, 6, 12, 18, and 24 w in arthritis, serositis, dermatitis, polyneuritis, CNS involvement, fever, hemolytic anemia, and thrombocytopenia Mean dose PRD: no differenceNo difference in AE due to steroids; in AZA group: hepatotoxicity in dosages ≥200 mg/d Deaths: 2 vs. 4; 5/6 due to SLE activityJadad 3 Majority in AZA group were white and majority in PRD group were black
Carneiro and Sato, 1999 ([52]), double-blind PCB-controlled RCT, 6 moN = 41 IC: SLE, PRD <0.5 mg/kg/d, at least 1: arthralgia >3 joints, arthritis, discoid lesion or malar rash, pleuritis, pericarditis, vasculitis, proteinuria, or urinary castsOral MTX 20 mg/w if >50 kg and 15 mg/w if <50 kg (n = 20) vs. PCB (n = 21) Comedication: stable PRD dose the first mo, then ↓ or ↑ depending on activitySLEDAI, VAS for joint pain, ↓ PRD dose, joint and skin improvement, AEsSLEDAI: MTX ↓ (P < 0.05) vs.usb PCB ↑ at 6 mo (P < 0.05) VAS: MTX ↓ vs. PCB ↑ at 6 mo (P < 0.05) ↓ PRD: in 65% MTX patients vs. 5% PCB patients; P < 0.00170% MTX vs. 14% PCB patients: side effects, mainly dyspepsia and hepatotoxicityJadad 5 Oxford 1b Similar withdrawal rates No intent-to-treat analysis
 EC: creatinine ≥2 mg/dl, recent loss of renal function, IS use ≥4 mo, and others  Arthralgia/arthritis and discoid SLE/malar rash: significant improvement with MTX vs. PCB (P < 0.001 for both)  
Fortin et al, 2008 ([53]), double-blind PCB-controlled RCT, 12 moN = 86 IC: moderate SLE (SLAM-R ≥8), SDI ≤15 EC: CYC or AZA in the last 4 w, renal failure, LN, and othersMTX 7.5 mg/w and ↑ to 20 mg/w (n = 41) vs. PCB (n = 45) Comedication: folic acid 2.5 mg/d, 6 d/w, PRD Stable PRD, antimalarials, and NSAIDs dose the previous 4 wSLAM-R, ↓ PRD dose, AEMTX ↓ SLAM-R at 12 mo: −0.86 (96% CI −1.71, −0.02), P = 0.039 MTX ↓ SLAM-R in patients with SDI 0 at 12 mo: −1.41 (96% CI −2.42, −0.39), P = 0.008 MTX ↓ mean PRD daily dose at 12 mo: −22.3 (96% CI −36.2, −5.4), P = 0.010No difference in AEs overall; differences in GI (56.1% vs. 33.3%) and psychological AEs (9.8% vs. 0%), P = 0.05 for bothJadad 5 Intent-to-treat analysis No difference in % lost to followup
Tam et al, 2004 ([60]), double-blind PCB-controlled RCT, 24 wN = 12 IC: active SLE (SLEDAI ≥6), PRDL <0.5 mg/kg/d EC: need for CYC or AZALEF 100 mg/d × 3 d, then 20 mg/d (n = 6) vs. PCB (n = 6) Comedication: HCQ, PRDL 15 mg/d and ↓, NSAIDsSLEDAI, proteinuria, C3, anti-dsDNA, PRDL dose, AEs↓ SLEDAI at 24 w: mean ± SD 11.0 ± 6.1 vs. 4.5 ± 2.4 (P = 0.02) Similar changes in proteinuria, C3, anti-dsDNA, PRDL doseAEs: no differenceJadad 5 No difference in % lost to followup
Ginzler et al, 2010 ([62]), open RCT, 24 wN = 370 IC: SLE and LN EC: ≥2 w dialysis before randomization, anticipated dialysis for ≥8 wMMF 0.5 g/12 h and ↑ to 1.5 g/12 h (n = 185) vs. IV CYC 0.5–1 g/m2/mo × 6 mo (n = 185) Comedication: PRD 60 mg/d and similar ↓ in both groupsNonrenal outcomes, BILAG, SELENA–SLEDAI, flaresNo difference in nonrenal outcomes in % patients with unchanged BILAG, % patients with improved BILAG, mean change in SELENA–SLEDAI and remission, and BILAG and SELENA–SLEDAI flaresNot reportedJadad 3 Similar withdrawal rates
Dammacco et al, 2000 ([70]), open RCT, 12 moN = 18 IC: moderate SLE EC: severe SLE, othersIV MP (1 g/d × 3 d) in both groups, then CSA <5 mg/kg/d and ↓ + PRD 0.5–1 mg/kg and ↓ to 5 mg/d (n = 10) vs. the same doses of PRD only (n = 8)SLEDAI, mean cumulative PRD dose, AEs↓ SLEDAI at 12 mo: 16.3 vs. 11.6 (P < 0.05) Mean ± SD cumulative PRD dose at 12 mo: 179.4 ± 40.1 vs. 231.8 ± 97.1 (P < 0.005)AEs: 60% vs. 62.5%; no differences per each AEJadad 1 1 vs. 5 patients were lost to followup by worsening
Griffiths et al, 2010 ([71]), open RCT, 12 moN = 89 IC: severe SLE (requiring a new IS and PRDL ≥15 mg/d) EC: hypertension, abnormal serum creatinine, othersCSA 1 mg/kg/d and ↑ to 2.5–3.5 mg/kg/d (n = 47) vs. AZA 0.5 mg/kg/d and ↑ to 2–2.5 mg/kg/d (n = 42) Comedication: stable PRDL, antimalarials, and NSAIDsMean change in PRDL dose, BILAG, BILAG flares, AEs↓ mean PRDL dose at 12 mo by >50% in both groups (P < 0.001); no difference in the change between groups (P = 0.2) No differences in BILAG activity and BILAG flaresNo patient had severe hypertension or persistent rise in creatinine One-third of patients in both groups discontinued the drugs due to AE or lack of efficacyJadad 3 CSA group was younger, had more nonwhite patients, and more damage on the SDI

The inclusion criteria and nonrenal manifestations analyzed in the diverse studies were varied. Likewise, the outcome variables used to measure the effects of the drugs were diverse (disease activity and clinical response measured by different indices, flares, serologic response, other specific testing variables, corticoids requirement, adverse events, etc.). The main recommendations, levels of evidence, and grades of recommendation from our review are shown in Table 2.

Table 2. Main recommendations and conclusions, levels of evidence, and grade of recommendation from the review of the evidence of the use of nonbiologic immunosuppressants in nonrenal SLE*
Recommendation/conclusionLEGR
  1. SLE = systemic lupus erythematosus; LE = level of evidence; GR = grade of recommendation; CYC = cyclophosphamide; IV = intravenous; NPSLE = neuropsychiatric SLE; MP = methylprednisolone; NYHA = New York Heart Association; sPAP = systolic pulmonary artery pressure; PAH = pulmonary arterial hypertension; CIN = cervical intraepithelial neoplasia; AZA = azathioprine; MTX = methotrexate; NSAIDs = nonsteroidal antiinflammatory drugs; LEF = leflunomide; MMF = mycophenolate mofetil; anti-dsDNA = anti–double-stranded DNA; CSA = cyclosporin A; TAC = tacrolimus.
CYC  
IV CYC may be useful for the treatment of NPSLE and reduction of relapses3aC
IV CYC plus prednisone is better than prednisone for the short-term treatment of NPSLE and reduction of relapses3aB
IV CYC is better than MP for the long-term treatment of NPSLE and reduction of relapses2bB
High-dose IV CYC has the same efficacy in the treatment of nonrenal SLE and the same adverse event rate than a traditional IV CYC regimen2bB
IV CYC is better than enalapril to improve the NYHA functional class and reduce the sPAP for the treatment of PAH in SLE, although it has a higher nonsevere infection rate2cB
IV CYC use is associated with cumulative damage ([28, 38])2aB
IV CYC use is associated with development of CIN ([41])2cB
IV CYC decreases leukocyte, neutrophil, and lymphocyte count, but the effect size is very small; therefore, severe myelotoxicity is infrequent ([32])2bB
In women with SLE, oral or IV CYC is independently associated in the short term with ovarian failure ([27])2aB
In women with SLE, the risk of ovarian failure increases with the cumulative dose of oral or IV CYC and is higher with longer IV CYC regimens ([25, 37, 47])2bB
In women with SLE, the risk of ovarian failure is associated with an older age at commencement of both oral and intravenous CYC; age itself is a risk factor for ovarian failure ([29, 30, 39])2cB
AZA  
The association of AZA with prednisolone treatment might reduce the flare rate2cB
MTX  
In patients with active nonrenal SLE manifestations despite prednisone, the association of MTX (15–20 mg/day) reduces in the short term (6 months) the global, articular, and cutaneous activity of the disease, with an additional short-term steroid-sparing effect1bA
In patients with moderate activity and nonrenal SLE manifestations despite prednisone, NSAIDs, and antimalarials, treatment with MTX (20 mg/day) reduces in the medium term (12 months) the activity of the disease, particularly in patients without damage, with an additional medium-term steroid-sparing effect1bA
LEF  
In patients with mild to moderate active SLE in spite of prednisolone, the addition of LEF is more effective than placebo in improving the activity of the disease with similar short-term (6 months) side effects1bC
MMF  
MMF can be used for the treatment of nonrenal SLE manifestations in patients with lupus nephritis because it is not inferior to IV CYC to decrease activity, induce remission, and reduce flares of nonrenal manifestations and to improve serologic parameters2bB
MMF is safer on the hematologic system than AZA and MTX, and may increase platelet and leukocyte counts and hematocrit ([64])2bB
MMF can be used to improve nonrenal activity in patients with nonrenal and/or renal refractory SLE and to reduce the need for corticosteroids3aB
MMF may prevent short-term (6 months) SLE flares when added to the treatment of patients with increasing anti-dsDNA titer2bC
In patients with renal and/or nonrenal SLE, MMF may cause non–dose-dependent adverse events (particularly in the gastrointestinal system) and drug survival is acceptable with a low withdrawal rate due to adverse events in the medium term (12 months) ([69])3aC
CSA  
In patients with renal and/or nonrenal SLE refractory to steroids, the addition of CSA may improve disease activity and induce remission in the short term but causes frequent adverse events2cB
In patients with renal and/or nonrenal SLE refractory to steroids, the addition of CSA may improve disease activity and have a steroid-sparing effect in the long term but causes frequent adverse events2bB
In patients with active SLE refractory to steroids, CSA is not less effective than AZA in reducing renal and/or nonrenal activity and both drugs have a similar steroid-sparing effect in the medium term with no significant difference in adverse events2b, 2cB
TAC  
In patients with active nonrenal SLE despite conventional treatment, the addition of TAC may be useful to improve disease activity in the medium term but causes frequent adverse events2cC

CYC

Twenty-nine studies evaluated the efficacy and/or safety of CYC in the treatment of nonrenal manifestations of SLE; 4 were unblinded RCTs ([21-24]) (Table 1), 1 was an open prospective study ([25]), and 24 were cohort studies ([26-49]) that included 3,742 patients overall. Different nonrenal manifestations were treated, although neuropsychiatric SLE (NPSLE) was studied in a more rigorous way ([21, 23, 24]). The CYC regimens and duration of CYC treatment and the comedications allowed in those studies were varied. The outcome variables used were varied, with the most frequent being clinical response to treatment measured by different activity and response indices, serologic response, rate of disease flares, decrease in the dose of prednisone, and adverse events. Some of the studies specifically addressed safety issues such as ovarian failure, neoplasias, or association with damage. The main conclusions and quotes of these studies are shown in Table 2. In summary, the evidence for using CYC in the treatment of nonrenal SLE is based on studies of a larger number of patients than those that assessed any other nonbiologic agent and RCT information is available, particularly for NPSLE. However, only a small percentage of patients were included in high-quality studies.

AZA

There were only 2 articles that assessed the efficacy and/or safety of AZA in the treatment of nonrenal SLE; 1 was an unblinded RCT ([50]) (Table 1) and 1 was a cohort study ([51]) that included 85 patients overall. The retrospective cohort study analyzed the influence of AZA (≥2 mg/kg/day) and prednisolone (7–12 mg/day) on the frequency of SLE flares and evaluated the predictors of these flares in 61 patients (38 without renal disease) over a mean followup period of 7.5 years ([51]). In comparison with a preceding period without AZA, this combined regimen resulted in a significant reduction in flares and an increase in flare-free patient-years. In summary, there is little evidence for using AZA in the treatment of nonrenal SLE because there is only 1 unblinded, non–placebo-controlled RCT of a small number of patients.

MTX

Seven articles evaluated the efficacy and/or safety of MTX in the treatment of nonrenal manifestations of SLE; 2 were double-blind, placebo-controlled RCTs ([52, 53]), 1 was a crossover open study ([54]), and 5 were cohort studies ([55-59]) that included 230 patients overall. Detailed information on both RCTs is shown in Table 1.

The crossover open study ([54]) assessed the efficacy of oral MTX (7.5 mg/week) in SLE patients without major organ involvement and active disease despite >10 mg/day of prednisone. The patients received treatment for 2 periods: 1) 3 months (followed by a 3-month control period without treatment), and then 2) 6 months (followed by a 6-month control period without treatment). In the 13 patients who finished the study, there was a significant reduction of lupus flares during the MTX treatment periods compared with the control phases (P = 0.02) without significant differences in the requirements of prednisone. In summary, the evidence for using MTX in nonrenal SLE treatment is based on high-quality studies, 2 double-blind, placebo-controlled RCTs. However, a small number of patients were included in these studies (only 61 patients were treated with MTX).

Leflunomide (LEF).

Two articles assessed the efficacy and/or safety of LEF in the treatment of nonrenal SLE; 1 was a double-blind, placebo-controlled RCT ([60]) (Table 1) and 1 was a cohort study ([61]) that included 30 patients overall. The cohort study retrospectively assessed the efficacy and safety of LEF (100 mg/day for 3 days, followed by 20 mg/day) in 18 SLE outpatients ([61]). After 2–3 months of therapy, most patients had subjective improvement and significantly lower Systemic Lupus Erythematosus Disease Activity Index (SLEDAI) scores. In summary, there is very little evidence for use of LEF in nonrenal SLE because the only reported double-blind, placebo-controlled RCT included only 6 patients treated with LEF.

MMF

Eight articles evaluated the efficacy and/or safety of MMF in the treatment of nonrenal manifestations of SLE; 1 was an RCT ([62]) (Table 1) and 7 were cohort studies ([63-69]) that included 769 patients overall. The RCT by Ginzler et al ([62]) explored as secondary end points the nonrenal findings of the Aspreva Lupus Management Study (ALMS) ([8]), a prospective, open-label, parallel-group RCT that assessed the effect of MMF compared with CYC as induction treatment for lupus nephritis. Some of the cohort studies specifically addressed safety issues, and their main results and quotes are shown in Table 2. In summary, the evidence for using MMF in nonrenal SLE treatment is based on studies with a large number of patients and RCT information is available. However, most patients were included in low-quality studies, and the RCT assessed the nonrenal response in patients with lupus nephritis who received induction treatment including high-dose corticosteroids.

Cyclosporin A (CSA).

Eight articles evaluated the efficacy and/or safety of CSA in the treatment of nonrenal SLE; 2 were unblinded RCTs ([70, 71]) (Table 1), 1 was a prospective open study ([72]), and 5 were cohort studies ([72-77]) that included 319 patients overall. The prospective open study investigated the effect of CSA (2.5–5 mg/kg/day) in 16 patients with active SLE over an average treatment period of 30.3 months ([72]). The European Consensus Lupus Activity Measurement score decreased significantly (P < 0.005) after 6 months, but not at the end of the observation period. The most frequent side effects were hypertension and deterioration of renal function (3 of 16 patients) and hypertrichosis (5 of 16 patients). In summary, there is little evidence for using CSA in nonrenal SLE treatment because one of the 2 unblinded, non–placebo-controlled RCTs that assessed this drug included only 10 patients treated with CSA, and in the other RCT, almost one-third of all patients discontinued the drug because of adverse events or a lack of efficacy.

Tacrolimus (TAC).

Two studies assessed the efficacy and/or safety of TAC in the treatment of nonrenal manifestations of SLE; both were cohort studies ([78, 79]) that included 31 patients. In the open-label prospective 24-week study by Suzuki et al ([78]), 21 patients with mild active SLE treated with oral TAC (1–6 mg/day) were studied. The mean SLEDAI score decreased significantly at 24 weeks (P < 0.01). In 8 cases, treatment was discontinued within 24 weeks because of inefficacy (6 cases) and adverse effects (2 cases). Nonserious side effects were observed in only 5 cases (23.8%). The retrospective cohort study investigated whether oral TAC (1–3 mg/day) was effective for treating SLE patients without active nephritis (n = 10) ([79]). The mean SLEDAI score and the mean dose of prednisolone decreased significantly after 1 year (P < 0.05 for both). Four of the 10 patients had adverse events and 2 patients discontinued treatment. In summary, there is very little evidence for using TAC because only 2 small studies have been reported, neither of which were RCTs, and almost one-third of all patients studied discontinued the drug because of a lack of efficacy or adverse effects.

Combinations of nonbiologic immunosuppressants

Six articles assessed the effect of several regimens combining different nonbiologic immunosuppressants for the treatment of nonrenal SLE; all were cohort studies ([80-85]) that included 2,262 patients. Although there were no RCTs, some interesting results were found. One study that analyzed risk factors for thrombosis in a large (n = 1,930), multiethnic SLE cohort found that an AZA treatment history (odds ratio 1.36, P = 0.023) and CYC treatment history (odds ratio 1.42, P = 0.025) were significant risk factors for thrombosis ([82]). Noel et al ([83]) carried out a retrospective study to analyze infectious complications and their risk factors in a cohort of 87 SLE patients and found that intravenous corticosteroids and immunosuppressants (oral or pulse CYC, AZA, or MTX) were independent risk factors for infection. In summary, there is very little evidence for using combinations of nonbiologic immunosuppressants in nonrenal SLE because there are no RCTs reported and the vast majority of patients were included in low-quality studies that were not designed to assess efficacy.

DISCUSSION

We conducted a systematic review of the literature to analyze the efficacy and safety of nonbiologic immunosuppressive drugs in the treatment of nonrenal manifestations of SLE. The following 3 facts justify the interest of our review: 1) to date, the vast majority of the studies have focused on the effect of the different available therapies for renal SLE; 2) a treatment may not be effective for all manifestations of SLE, but may be effective for specific manifestations; and 3) nonbiologic immunosuppressants are frequently used as off-label treatment for nonrenal SLE, even when there are no clear recommendations for their use in this situation. There are actually very little RCT data for most of the treatments currently being used.

In our systematic review, we found a great number of studies, although in general their quality was low and the number of patients included in high-quality studies was small. Although there are several RCTs addressing the treatment of nonrenal SLE, it is almost impossible to combine the data in a single meta-analysis because of an important variability in the selected patients, treatment doses, and outcome measures. The main objective of any treatment for SLE in clinical practice is a decrease in disease activity. Most of the studies we analyzed used some validated activity indices, although there was no uniformity among them. An additional aim of using an immunosuppressant is the reduction of the doses of steroids used to control disease activity, preventing their side effects. Several of the immunosuppressants analyzed demonstrated their efficacy and safety in the treatment of nonrenal SLE with a steroid-sparing effect.

Our systematic review demonstrated that MTX has the strongest level of evidence for the treatment of nonsevere extrarenal SLE, with 2 double-blind, placebo-controlled RCTs showing the same results ([52, 53]). The design of both trials and the characteristics of the recruited patients were similar, with >90% having musculoskeletal and/or cutaneous manifestations, particularly arthralgias/arthritis, malar rash, and discoid lesions. These studies demonstrated the efficacy of the drug in reducing global, cutaneous, and articular activity, with an additional steroid-sparing effect in the short and medium term (6–12 months) with a good safety profile. Thus, our systematic review supports the use of MTX as the first immunosuppressive therapy recommended in the treatment of moderately active nonrenal SLE.

CYC was successfully tested in difficult clinical situations such as NPSLE ([21, 23, 24]) and pulmonary hypertension due to SLE ([22]). Although it is not possible to establish a general CYC schedule, maintenance therapy with CYC is associated with a significant reduction in NP relapses. However, our systematic review demonstrated that CYC is an important risk factor for cumulative damage, including ovarian failure and different neoplasias. Thus, although CYC may be recommended as the first immunosuppressive agent in the treatment of more severe cases of nonrenal SLE, decisions about CYC use must be evaluated as a balance between the benefits of treating life-threatening complications of SLE and risks of severe adverse events that are generally associated with a longer duration and higher cumulative dose of both intravenous and oral CYC.

Although AZA is one of the immunosuppressants most frequently used in combination with steroids in nonrenal SLE, there is very little evidence supporting its use. AZA was tested in 1 unblinded, non–placebo-controlled RCT of a very small number of patients treated with the drug (n = 11) that was unable to show any benefit of AZA plus prednisone over prednisone alone in any of the nonrenal clinical manifestations in the short term (3 months) and long term (24 months) or in the reduction of steroid doses ([50]). However, a recommendation about this result cannot be made because there is an important design bias; both groups continued to receive high-dose prednisone (40–60 mg/day) for 4–6 months, which may explain the absence of differences. LEF also was more effective than placebo in treating mild to moderate active SLE patients with a favorable safety profile, although more RCTs assessing this drug in a greater number of patients are required.

The evidence of the efficacy of MMF for the treatment of nonrenal SLE is limited because it is based on low-quality studies and the evaluation of the secondary end point (nonrenal features of SLE) of the ALMS, an RCT specifically designed to evaluate MMF in comparison with CYC for the induction treatment of lupus nephritis ([8, 62]). However, although induction treatment for lupus nephritis with high-dose corticosteroids and intravenous CYC or MMF is recommended in clinical practice, these regimens are not the standard of care for SLE patients with nonrenal SLE. Therefore, the results of this study cannot be extrapolated to patients with nonrenal SLE who are usually treated less aggressively. Further well-designed RCTs assessing the efficacy and safety of MMF as a primary end point for the treatment of nonrenal SLE are eagerly awaited.

There is very little evidence for the use of calcineurin inhibitors because CSA was assessed in 2 small unblinded, non–placebo-controlled RCTs ([70, 71]) and TAC was assessed in 2 small low-quality non-RCTs ([78, 79]). The withdrawal rate was high for both drugs; therefore, concern about their side effects is a barrier to generalizing their use.

Based on its clinical and serologic heterogeneity, some authors have recently considered SLE as a syndrome rather than a single disease ([86]). This approach to SLE and the accumulated knowledge of its different pathogenic factors might allow a better classification of this syndrome and the use of targeted therapies for specific manifestations of SLE in the future. Until then, our approach to the treatment of SLE should be more general and based on the data of the studies we reviewed.

In conclusion, several immunosuppressants have demonstrated their efficacy, safety, and steroid-sparing effects in the treatment of nonrenal SLE. However, the number and quality of the studies are limited. A specific drug for each particular manifestation cannot be recommended, although CYC may be used in more severe cases and MTX may be the first option in moderately active SLE. The results of our review may help the clinician make better therapeutic decisions and serve as a reference for further development of clinical practice guidelines or clinical trials addressing system-specific nonrenal manifestations in larger populations of SLE patients.

AUTHOR CONTRIBUTIONS

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Pego-Reigosa had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Pego-Reigosa, Cobo-Ibáñez, Calvo-Alén, Loza-Santamaría, Muñoz-Fernández, Rúa-Figueroa.

Acquisition of data. Pego-Reigosa, Cobo-Ibáñez, Muñoz-Fernández.

Analysis and interpretation of data. Pego-Reigosa, Cobo-Ibáñez, Calvo-Alén, Loza-Santamaría, Rahman.

ROLE OF THE STUDY SPONSOR

Roche Spain had no role in the study design or in the collection, analysis, or interpretation of the data, the writing of the manuscript, or the decision to submit the manuscript for publication. Publication of this article was not contingent upon approval by Roche Spain.

Acknowledgments

We would like to thank our librarian, Ms María Piedad Rosario-Lozano (Research Unit of the Spanish Society of Rheumatology), who checked the search strategies and performed the hand search.

Ancillary