The British Society for Rheumatology Guideline for the Management of Systemic Lupus Erythematosus in Adults

Caroline Gordon; Maame-Boatemaa Amissah-Arthur; Mary Gayed; Sue Brown; Ian N. Bruce; David D'Cruz; Benjamin Empson; Bridget Griffiths; David Jayne; Munther Khamashta; Liz Lightstone; Peter Norton; Yvonne Norton; Karen Schreiber; David Isenberg


Rheumatology. 2018;57(1):e1-e45. 

In This Article


Overview of the Management of Severe Lupus

Patients who have serious manifestations with organ- or life-threatening disease require treatment with intensive immunosuppression followed by a prolonged period of less aggressive maintenance therapy to prevent relapse (summarized with suggested dosing regimens in Table 7). In some cases there may be a thrombotic component to the clinical features that requires anticoagulation, for example in patients with APS as well as lupus. There is most evidence for the management of LN, less for neuropsychiatric disease and very little for other organ-specific manifestations.

The authors of this guideline have not reviewed the evidence for the management of LN as they suggest that the EULAR/ERA-EDTA recommendations for the management of adult and paediatric LN[24] are followed. The main recommendations and SOAs with them are shown in Table 3. Further details about these recommendations and the evidence for them have been published.[24]

For the management of severe non-renal SLE, the evidence for treatment with high-dose CSs, AZA, CYC, MMF, rituximab, IVIG and plasma exchange (plasmapheresis) is discussed below. The evidence for use of belimumab and of the calcineurin inhibitors ciclosporin and tacrolimus, particularly for cytopenias due to lupus, has already been reviewed above. Suggested initial target dosing regimens and lower maintenance regimens to prevent flares once patients are stable are shown in Table 7. The actual regimen used for individual patients will depend on the clinical picture and the treatment history. Patients with refractory disease, especially those being considered for belimumab and rituximab, should be discussed with and/or seen by a specialist lupus centre (see Figure 1 flowchart for eligibility and response criteria). It is important to review the response regularly and to increase the dose and/or change the treatment if patients fail to respond.

CSs for Severe SLE

Summary. The emphasis in the last 10 years has been on finding steroid-sparing regimens to treat severe lupus, using other immunosuppressants in conjunction with CSs (either orally, intravenously or both), to induce and maintain response with the least risk of adverse events, particularly infection. In general, there is an increasing tendency to use oral prednisolone at a dose of 0.5 mg/kg/day with i.v. MP pulses (3 × 500–750 mg) rather than higher doses of i.v. MP pulses and/or higher dose of oral prednisolone (e.g. 0.75–1 mg/kg/day) as done in the past for all severe manifestations of lupus.

Evidence. I.v. MP pulses as an alternative to, or in addition to, high-dose oral prednisolone was first reported as a treatment for LN.[24,325,326] I.v. MP pulses were introduced for the management of non-renal lupus in the early 1980s.[147] An open-label cohort study[146] and an open-label trial[145] using i.v. MP pulses followed by alternate day oral CSs found that pulse therapy led to rapid improvement in clinical symptoms and anti-dsDNA and C3 levels, but that an alternate day oral regimen was associated with relapses. A small double-blind, placebo-controlled RCT with mostly non-renal SLE patients[144] found that 3 i.v. MP pulses resulted in faster and more complete improvement in the first 2 weeks in 12 patients with SLE, but there was no significant difference in efficacy or safety parameters at 4 weeks or 6 months compared with the placebo group; however, all patients received 40–60 mg of oral prednisolone daily.[144]

A double-blind RCT[143] comparing three daily i.v. MP pulses of either 1000 or 100 mg in 21 patients with SLE causing fever, cardiorespiratory, renal or NP manifestations (with individualized outcomes based on entry manifestations) suggested no difference in efficacy between the regimens. A retrospective study compared low-dose i.v. MP pulses (⩽1500 mg over 3 days) with high-dose pulses (3–5 g over 3–5 days) for the treatment of severe flares.[148] This study suggested that the lower dose was sufficient and safer for controlling SLE flares than the high-dose regimen, which was associated with an increased number of infections.[148]

Conclusions. There is limited evidence for any particular CS regimen for specific manifestations of severe non-renal lupus. Overall the LOE for i.v. MP pulses and oral prednisolone in non-renal severe lupus is 2+ and the GOR is C.

AZA in Severe SLE

Summary. AZA (2–3 mg/kg/day) is sometimes used as first-line therapy with CSs in severe non-renal lupus (see Table 7), based on the evidence discussed in the section on the use of AZA for the management of moderate lupus. It is most often used in women planning pregnancy or pregnant, as it is much safer in pregnancy than CYC or MMF, which are contra-indicated in such situations.[239]

Evidence. There was only one open-label controlled trial, with 24 patients with severe (life-threatening) multisystem manifestations of lupus,[151] which showed no definite benefit from the addition of AZA compared with 40–60 mg prednisone alone for 6 months, before tapering over the next 18 months, although there was some steroid-sparing benefits seen at 12 months. It has been used as primary treatment at a dose of 2 mg/kg/day as an alternative to MMF or CYC in low-risk renal patients without adverse prognostic factors and when these drugs are contra-indicated, not tolerated or unavailable.[24]

AZA has been used more often as maintenance therapy after a course of CYC for severe lupus, based on the evidence from studies undertaken in patients with LN.[24,25] The rate of major extra-renal flares in the maintenance phase of the Aspreva Lupus Management Study (ALMS) study was low in the AZA group at 6.3% (7/111) and similar to the frequency of 6.9% (8/116) in the MMF group.[160] There is some evidence that AZA may be less effective at preventing renal flare in patients in this LN study than MMF, as discussed in the section on MMF.[160] However in a predominantly Caucasian LN population, in the MAINTAIN study, no difference in number or time of severe systemic flares in the AZA group (4/43) compared with the MMF group (3/53) was observed.[161] There are no trials or controlled studies addressing AZA as a primary treatment for neuropsychiatric lupus or any other specific serious non-renal manifestations of lupus, but it has been used after CYC for the treatment and prevention of recurrence of lupus psychosis in 13 patients.[328]

The systematic review of non-biologic immunosuppressants in non-renal SLE by Pego-Reigosa et al.[133] only considered the unblinded RCT (showing no benefit) from 1975[151] and a cohort study (showing a reduced rate of flare[155] in patients on AZA) and concluded that there was little evidence to support the use of AZA in non-renal lupus.

Conclusions. Overall, the LOE for AZA in non-renal severe lupus is 2+ and the GOR is C.

CYC in Severe SLE Including LN and Neuropsychiatric Lupus

Summary. CYC, although not licensed for lupus, has been used for the treatment of severe lupus, particularly LN and organ- or life-threatening non-renal disease, since the late 1960s, with the first open-label trial in LN reported in 1971.[364] Oral CYC is associated with an increased risk of bladder cancer and has been replaced by i.v. CYC pulses in the management of severe lupus. There is most experience with i.v. CYC pulses in LN and NPSLE (Table 3 and Table 7). CYC is teratogenic and is contra-indicated in women trying to conceive, or who are pregnant or breast-feeding. It is gonadotoxic and can cause infertility, and men should not father children while on CYC.[239]

Evidence. The first controlled trial comparing prednisone with CYC in LN, non-renal lupus and PM was reported in 1973,[365] and a similar design was used to compare oral CYC and AZA in lupus not responsive to 15 mg prednisolone,[366] but numbers were small and the aim of matching individual patients and comparing their outcomes was unsuccessful. Since then, studies have used different trial designs and evidence supporting the use of various doses of oral and later i.v. pulse CYC regimens to reduce disease activity and prednisolone dosage and to improve outcomes in patients with LN and non-renal lupus have been reported. The best-known regimens are based on the National Institutes for Health i.v. CYC protocol (monthly i.v. CYC at 500–1000 mg/m2 body surface area for 6 months, followed by 3 monthly i.v. CYC for 2 years)[367] and the Euro-Lupus protocol, which uses lower doses (500 mg fixed dose i.v. CYC 2-weekly for a total of 6 doses, followed by oral AZA)[368] and appears to be as effective and safer for LN in Europe than high-dose regimens.[369] In recent years, the 3-monthly i.v. CYC maintenance pulses for 2 years in the National Institutes for Health protocol have been replaced by oral MMF or AZA.[25,370]

I.v. CYC pulses were the most widely used regimes for all but the mildest cases of acute proliferative glomerulonephritis until MMF was found to be comparable in efficacy and safer.[24,25] It should be noted that neither of these drugs is licensed for the treatment of LN, but both are supported as appropriate treatment for the management of LN in the EULAR/ERA-EDTA recommendations for the management of adult and paediatric LN[24] (Table 3) and the ACR guidelines for screening, treatment and management of LN.[25]

Treatment regimens tested in LN have often been applied to severe non-renal lupus disease as there are fewer non-renal studies and they include heterogeneous patient populations. A systematic review[133] evaluated 29 studies, including 4 unblinded RCTs in which 3742 patients with non-renal lupus were treated with a variety of CYC regimens. There are more data on the efficacy and safety of using CYC to treat non-renal lupus than of any other drug treatment; however, there are fewer high-quality studies than for LN, and diverse end points have been used, making it hard to compare the studies.

Data from the ALMS RCT comparing i.v. CYC (0.5–1.0 g/m2 monthly × 6) and MMF (target 3.0 g/day) as induction therapy for LN[159] showed that i.v. CYC therapy was associated with almost 95% response in all of the non-renal systems, apart from the haematology, which was confounded by drug-induced cytopenias and anaemia of uncertain cause. There was no difference in response between i.v. CYC or MMF in any of the systems studied, including renal.

Some of the best evidence supports the use of pulse i.v. CYC in NP lupus, with one small RCT favouring an i.v. CYC regimen over i.v. MP alone.[186] That trial used more CSs than we would recommend now and was based on a previous retrospective cohort study that suggested that i.v. CYC was useful in the management of NPSLE.[371] The RCT[186] recruited 32 SLE patients with active severe NP manifestations without thrombosis (such as seizures, optic neuritis, peripheral or cranial neuropathy, coma, brainstem disease or transverse myelitis) that had developed within the previous 15 days. All of the patients received oral prednisolone 1mg/kg/day for up to 3 months and then tapered depending on response and 1 g of i.v. MP daily for 3 days. One group received further 1 g of i.v. MP daily for 3 days repeated monthly for 4 months then bimonthly for 6 months and finally 3 monthly for one year. The other group received i.v. CYC 0.75g/m2 body surface monthly for 12 months then this dose was repeated every 3 months for another year. The primary end point was at least 20% improvement from baseline using clinical, laboratory or specific neurological criteria and was met in 18/19 (95%) receiving CYC and 6/13 (46%) receiving MP.[186] A Cochrane systematic review of the treatment of NPSLE[372] calculated a relative risk of 2.05 (95% CI: 1.13, 3.73) for 20% response at 24 months with CYC therapy, but most patients responded by 5 months. CYC treatment was also associated with greater improvement in other lupus manifestations, a significant reduction in SLEDAI score at 6 and 12 months, greater reduction in prednisolone dosage and more patients completing the protocol compared with the MP group. There was no difference in adverse events, including infections and deaths. Recruitment to the study was stopped early due to the higher failure rate of the MP arm. Although the RCT is not of high quality[372] due to the small number of patients studied, the heterogeneity of the NP events, the variable outcome measures used for their assessment, and potential confounding by variable oral CS dosing, it is clear that the i.v. MP regimen was not sufficient and that CYC was better at controlling active NPSLE and preventing relapse.

Further evidence for the use of CYC in NPSLE comes from a previous open-label, controlled pilot study on the use of low-dose i.v. CYC, with a mean dose of 21 mg/day oral prednisone in 37 NPSLE patients, compared with oral prednisone alone in 23 patients (mean dose 21 mg/day),[187] and a cohort study[373] in which a low-dose regimen of i.v. CYC was used in 25 patients with NPSLE with benefit and a low risk of adverse events. A case series[328] found that treating 13 patients with lupus psychosis with oral prednisolone starting at 1 mg/kg/day for 8 weeks and oral CYC (1–2 mg/kg/day) for 6 months followed by oral AZA (1–2 mg/kg/day) led to improvement within a mean of 44 days and only one relapse with psychosis after 2 years; however, 23% developed other NP features and 38% had non-NP flares over the mean follow-up of 7 years. Anti-psychotic agents were used in nine patients for a mean of 6 months. Evidence for CYC and other treatments in neuro-ophthalmic manifestations of lupus have been reviewed in a systematic review,[374] but the data on treatment is mostly based on case reports and small case series, for example cases with neuromyelitis optica treated with or without CYC.[374]

In contrast to the studies assessing low-dose regimens, high-dose CYC has been studied as well in the hope of achieving better responses in severe lupus. An open-label, uncontrolled study[375] reported the initial safety and efficacy of high-dose CYC (50 mg/kg × 4 days) without stem cell transplantation in 14 patients with refractory moderate to severe SLE despite CSs and at least one immunosuppressant. A prospective RCT[188] was designed to compare the efficacy and safety of a widely used standard i.v. CYC regimen (monthly i.v. CYC at 750 mg/m2 body surface area for 6 months, followed by 3 monthly i.v. CYC for 2 years) with this high-dose i.v. CYC regimen. Entry criteria included moderate-to-severe lupus with renal (22 patients), neurologic (14 patients) or other organ system involvement (11 patients). There was no evidence that response differed between the regimens, but non-responders to monthly i.v. CYC could be rescued with high-dose i.v. CYC. There was no difference in serious adverse events, infections, premature ovarian failure or deaths between the two groups. Leuprolide (a gonadotropin-releasing hormone analogue) was not used to protect against ovarian failure.[376] This should be considered with i.v. CYC moderate- and high-dose regimens,[188] as amenorrhoea and ovarian failure are dose- and age-related adverse events of CYC,[370,377] but are rare with the European low-dose i.v. CYC regimen (500 mg 2-weekly for 3 months only) recommended for LN.[24]

The remaining data[133] supporting the use of CYC for other serious non-renal manifestations of lupus are obtained predominantly from a variety of cohort studies, small case series and case reports, including 5 patients with systemic lupus vasculitis,[378] 11 patients with myocarditis[379] and 5 patients with heart failure due to myocarditis.[380] There is one open-label RCT comparing i.v. CYC with enalapril for 6 months in the treatment of pulmonary hypertension, which showed greater benefit from CYC but an increased risk of infection and gastrointestinal side effects.[189]

Conclusions. There is considerable evidence supporting the use of i.v. CYC to reduce disease activity and CS usage in severe lupus, for both renal and non-renal disease, including NPSLE. There is no evidence that CYC prevents chronic damage, and all regimens are teratogenic, but there is less risk with the Euro-Lupus regimen of adverse events (such gastrointestinal side effects, alopecia, infection, amenorrhoea and infertility due to ovarian failure) than with higher dose regimens.[12,16,24,25,133,372] Overall, the LOE for the use of CYC in non-renal severe lupus, including NPSLE, from 1 systematic review including 29 studies and 1 systematic Cochrane review of NPSLE is 2 ++, and the GOR is B.

MMF in Severe SLE

Summary. There is considerable evidence supporting the use of MMF in the management of LN, and this has been discussed in the Joint EULAR/ERA-EDTA recommendations for the management of adult and paediatric LN[24] (Table 3) and the ACR guidelines for screening, treatment and management of LN.[25] The mean SOA of all of the authors of this guideline with each of the main EULAR/ERA-EDTA recommendations for the management of LN is shown in Table 3. There is very little evidence for the use of MMF in NPSLE, but it is being used to reduce other types of moderate and severe non-renal lupus disease activity (Table 7), to prevent flare and for its steroid-sparing properties, as an alternative to CYC or AZA, especially in cases where inefficacy, drug intolerance and concerns about toxicity arose. It is not compatible with conception, pregnancy or breast-feeding.[239]

Evidence. As mentioned in the section on moderate lupus, there is a systematic review of non-biologic immunosuppressants in non-renal SLE[133] that summarizes the data from 8 papers (covering 768 patients with moderate/severe lupus), which assessed the efficacy and safety of MMF in the treatment of non-renal SLE, including the ALMS RCT comparing the use of MMF with that of CYC as induction therapy for LN,[159] and 7 cohort studies including 6 discussed above[162–166,351] and an abstract that does not meet the criteria for this guideline.

Conclusions. Overall, the LOE for MMF in non-renal lupus from 2 systematic reviews, 2 open-label RCTs in LN and 7 cohort studies is 2 ++, and the GOR is B.

Rituximab in Severe SLE

Summary. According to the NHS England Interim Commissioning Policy Statement for rituximab in SLE,[267] rituximab may be considered in patients with severe or moderate SLE (BILAG system category A or ⩾2B system scores, or SLEDAI >6) who fail treatment with MMF or CYC, either because of lack of effect or due to adverse events, providing they have already failed another immunosuppressant or it would be contra-indicated, or who require unacceptably high long-term CS dosing to control their lupus activity (see Figure 1 flowchart for eligibility and response criteria).

Evidence. Clinical examples of severe lupus are shown in Table 7, and the evidence for rituximab is summarized in Table 2. The systematic reviews by Duxbury et al.[201] and Cobo-Ibáñez et al.[200] provide evidence supporting the use of rituximab for non-renal severe manifestations of lupus, such as NP involvement (5 cohort studies[381–385]), haematological manifestations (6 cohort studies[383,385–389]) and at least 10 other cohort studies[382,383,385,387,390–395]). The data for improvement in NPSLE are still limited and uncontrolled, but showed 73–100% response in small numbers of patients. There is some evidence for improvement (50–100%) in mostly refractory lupus patients and idiopathic autoimmune thrombocytopenia and haemolytic anaemia. There are some specific reports on the use of rituximab in neuro-ophthalmological cases in a systematic review of these conditions,[374] and pooled data from European cohorts[396] on the effects of rituximab in LN, as mentioned in the EULAR/ERA-EDTA recommendations for the management of adult and paediatric LN.[24] There are insufficient data to comment on other specific severe lupus manifestations at present, but rituximab is accepted as having steroid-sparing properties (three open-label studies[192,193,199]).

Conclusions. Overall, the LOE for rituximab from 3 systematic reviews and 30 studies, including 1RCT and 3 open-label trials for reducing lupus disease activity and for steroid-sparing properties, is 2+, and the GOR is C.

IVIG in Severe SLE

Summary. IVIG has been used most in patients with refractory cytopaenias, thrombotic TTP and the catastrophic variant of APS. It can be used in pregnancy (but does not prevent heart block or fetal loss) and in patients with infection. It is rarely indicated as there is not much evidence for its use (Table 2).

Evidence. Much of the initial data are from case reports or small case series reporting treatment of acute events in small numbers of patients.[223–226] A systematic review and meta-analysis covering 3 controlled and 10 observational studies in SLE concluded that IVIG led to a reduction in SLE disease activity scores and a rise in complement levels in 31% of patients (P = 0.001, 95% CI: 22.1, 41.3) . There were insufficient data to assess response using other outcome measures, although serious adverse events were rare and mild.[227] The observational studies often did not report concomitant medication and used a variety of outcome measures and treatment regimens, as discussed below.

IVIG at a dose of 400 mg/kg/day for 5 consecutive days was used monthly for 6–24 months with some benefit in an open-label, uncontrolled trial with 12 refractory SLE patients.[210] Another open-label study[213] assessed 13 female SLE patients with a flare who received 0.4 g/kg body weight IVIG daily for 5 days. Short-term benefit was seen irrespective of concomitant therapy. IVIG–related adverse effects were mild and rare, and there was no worsening of renal function.[213]

Low-dose IVIG was used to treat histologically confirmed cutaneous lupus in 12 patients starting with doses of 1 g/kg × 2, followed by 400 mg/kg monthly until disease remission or for 6 months.[214] Five patients showed complete or almost complete (>75%) clearing of their skin lesions, two had partial improvement (>50%) and three had poor responses (<50%). There were few side effects in this study, but renal patients were avoided because nephrotoxicity has been reported in other studies.[397]

A retrospective chart review of 62 patients treated with low-dose IVIG (~0.5 g/kg) on average every 5 weeks for a mean of 6 courses showed a steady reduction in SLEDAI score over 8 months.[215] Patients with fever, rash, mucosal ulcers, pleurisy, pericarditis, urinary casts and urinary red cells responded in over 50% of cases, but only 30% of arthritis cases responded. Patients with thrombocytopenia, vasculitis and alopecia did not respond. Another group also found a disappointing response to IVIG in thrombocytopenia[216] in a retrospective analysis of 59 patients with immune-mediated severe thrombocytopenia, 44 of whom had definite lupus. A transient response to IVIG was reported in three patients with haemolytic anaemia in another study.[217]

The effect of high-dose IVIG (30 g of sulfonated IVIG on days 1–4 and 21–24) in 12 mild to moderate active lupus patients[218] was only temporary in most patients. High-dose IVIG treatment in 17/20 (85%) SLE patients given 1–8 treatment courses consisting of 2 g/kg monthly given over 5 days[219] led to some improvement in arthritis, fever, thrombocytopenia and NP lupus.[219] A retrospective chart review of 17 patients (including 11 with SLE), with a mean follow-up of 30 months and long-term high-dose IVIG treatment monthly for 6 months then every 2–3 months,[220] found that there was a significant reduction in the SLEDAI score with significant steroid-sparing effects, and remission was achieved in 12 patients.[220]

A case–control study[221] compared 12 pregnant SLE patients with a history of recurrent spontaneous abortions who were on high-dose IVIG (0.5 g/kg every 3–33 weeks) with 12 similar patients treated with prednisolone and NSAIDs. Patients in the IVIG group stopped prednisolone (n = 4) and NSAIDs (n = 9). Disease activity decreased by the end of pregnancy (P < 0.0001) and there was a reduction in autoantibodies and normalization of complement levels in the IVIG group. Such improvements were not seen in the control group, and there were three fetal losses due to spontaneous abortion in this group compared with none in the IVIG group. However, other studies have not confirmed that IVIG can prevent fetal loss,[239] and it is possible that NSAIDs contributed to fetal loss in the control group.[240]

A multicentre, prospective, open-label study of pregnant women with anti-SSA/Ro antibodies in the mother and birth of a previous child with CHB/neonatal lupus rash was undertaken to determine whether IVIG (400 mg/kg) given every 3 weeks from weeks 12 to 24 of gestation could prevent the development of CHB.[211] CHB was detected at 19, 20 and 25 weeks in 3 babies at a stage when 20 mothers had completed the IVIG protocol before the trial was stopped. An additional child without CHB developed a transient rash consistent with neonatal lupus.[211] Another European prospective study showed similar results.[212]

A large retrospective, single-centre cohort study was published by Camara in 2014,[222] which included 52 SLE patients with predominantly cutaneous, haematological, NP and cardiac manifestations who received at least one cycle of IVIG (400 mg/kg/day for 5 days). IVIG was given to 27 patients with infection and active lupus disease, and 17 (63%) patients showed some response. In 18 (69%) of 26 patients with refractory active disease without infection, some response was seen also. This study was too recent to be included in the comprehensive review on the use of IVIG in rheumatic diseases[228] that covered the case–control study in pregnancy by Perricone et al.,[221] 4 prospective open-label studies,[210,213,215,218,219] a retrospective cohort study[220] in lupus and a small RCT in LN not discussed here.[228]

Conclusions. IVIG, particularly the high-dose regimen, can have some beneficial effects in the short term on disease activity, but has to be continued with intermittent courses for sustained benefit to be seen and only then has steroid-sparing properties. It has a low rate of adverse events in non-renal patients, but can cause nephrotoxicity, especially with pre-existing renal disease. The evidence supporting its use is weak compared with that of other treatments that are cheaper and easier to administer, so it should be reserved for patients in whom other treatments are contra-indicated or have failed. Overall, the LOE for IVIG in non-renal severe lupus from 2 systematic reviews (including a meta-analysis, 3 open-label trials, 10 cohort studies and 4 case series) is 2–, and the GOR is D.

Plasma Exchange (Plasmapharesis) for Severe SLE

Summary. Plasma exchange in SLE has been used in small numbers of patients with conflicting results since the late 1970s. A systematic review was published while this paper was in preparation.[238] It is rarely indicated, because there is inadequate data to support its use except in thrombotic TTP (Table 2).

Evidence. The evidence supporting treatment with plasma exchange, which is expensive and often difficult to organize, remains poor except for thrombotic TTP,[229,398] the catastrophic variant of APS[238] and refractory neuropsychiatric, haematological and renal lupus.[238] Even for rapidly progressive glomerulonephritis, the evidence is limited.[399]

Studies have shown that plasmapheresis can reduce immune complexes and anti-dsDNA antibodies, but there is a rapid rebound of complexes and antibodies to pre-treatment levels, as shown originally in 5/8 patients.[230] Marked improvement after plasma exchange was seen in 7/11 (64%) SLE patients in another study[231] lasting up to 3 years, but one (9%) patient with a severe relapse died, and plasma exchange was ineffective in 3 (27%) patients. In another small study of nine patients, 5 (56%) improved, 2 (22%) progressed to end-stage renal failure, and 2 (22%) died due to complications of severe SLE.[232]

There was less support for the use of plasma exchange in SLE after a trial comparing plasma exchange in combination with CYC and CSs with standard therapy revealed no benefit from the plasma exchange for 40 patients with severe LN.[400] However, to avoid the rebound increase in autoantibodies after plasma exchange, a synchronized protocol was developed by the Lupus Plasmapharesis Study Group, consisting of plasmapheresis (3 × 60 ml/kg) followed by high-dose pulse CYC (36 mg/kg) then 6 months of oral immunosuppression. This treatment led to rapid improvement in disease activity in the initial 14 patients with various severe SLE manifestations, sufficient for immunosuppressants including CSs to be withdrawn in 12 (86%) patients at 6 months. Treatment-free clinical remission was sustained in 8 (57%) patients for a mean of 5.6 years.[233] However, there has been concern that improvements seen in this and 2 other uncontrolled studies[234,235] with 23 patients may have been due to the concomitant immunosuppressants. It is notable that the Lupus Plasmapharesis Study Group never reported on the final disappointing results of a randomized international multicentre trial comparing their synchronized protocol[233] with the administration of pulse CYC alone.

The evidence for treating patients who have diffuse alveolar haemorrhage, thrombotic TTP or catastrophic APS with lupus is predominantly from case reports and small case series.[229,236,237] Given the high mortality in TTP in general, but especially with lupus,[229,398] it is essential that patients with TTP are referred early for plasma exchange and specialist care.[398,401] Further details about the experience with and potential use of plasma exchange and immunoadsorption in lupus and APS, including LN, are covered by the systematic review.[238]

Conclusions. There remains a need for further research to better define the patients who are most likely to benefit from plasma exchange, but in general they are considered to be those who have TTP, severe refractory disease or contra-indications to conventional treatment (such as pregnancy). Overall, the LOE for plasma exchange for the treatment of non-renal severe lupus from one systematic review and nine studies is weak,[3] and the GOR is D, but for TTP it is strongly recommended (grade B), as for non-lupus patients with TTP.