Introduction

Systemic lupus erythematosus (SLE) is a complex multisystem chronic autoinflammatory disease characterized by hyperactive autoreactive B and T cells, and inflammation induced by immune complexes that has systemic clinical manifestations and follows a relapsing and remitting course.1 Neuropsychiatric involvement in SLE (NPSLE) includes a broad spectrum of neurologic and psychiatric manifestation. In 1999, the Research Committee of the American College of Rheumatology (ACR) defined 19 neuropsychiatric syndromes: one of them is myelopathy.2 Acute transverse myelitis (ATM) is acute inflammation of gray and white matter in one or more adjacent spinal cord segments. Symptoms and sign include bilateral motor, sensory, and sphincter deficits below the level of the lesion.3 Prevalence of ATM in the general population is estimated at 1–4 new cases per million per year, while in SLE it is seen in 1–2% of patients, 1000 times greater than the general population.4 Although SLE-related myelitis is traditionally regarded as a single diagnostic entity, case reports and cohort studies show clinical and probably physiopathogenic heterogeneity.4–14 The aim is to communicate and analyze an original series of SLE-related myelitis observed consecutively in a general hospital in Buenos Aires, Argentina.

Subjects and methods

We employed a retrospective chart review of inpatient medical records at Carlos G. Durand Hospital, and identified all patients who were admitted to the hospital with SLE and myelitis during the period 2007–2014. All meet the revised criteria of ACR of SLE1 and NPSLE,2 and comply with the Transverse Myelitis Consortium Working Group (TMCWG) definition of myelitis.3 We compiled a database that included demographics, clinical variables, laboratory results, magnetic resonance imaging (MRI), autoantibodies profiles, treatment and clinical outcomes. The Systemic Lupus Erythematosus Disease Activity Index (SLEDAI)15 was measure retrospectively at myelitis diagnosis. Also, we evaluated the neurological impairment at nadir and 6 months later with the American Spinal Injury Association impairment scale (AIS) and the European Database for Multiple Sclerosis grading scale (EGS). The AIS is determined according to the motor and sensory findings on neurological examination (A=injury is complete, B=motor complete and sensory incomplete, C and D=motor incomplete with different impairment of motor function and E=normal).16 The EGS is a simplified version of the Kutzke Expanded Disability Status Scale (EDSS). This scale mainly values the ability to walk (1–3: unlimited walk distance, 4–5: walk without aid > or <500 m, 6–6.5: walk with support, 7: home restricted; 8: chair restricted; 9: bedridden and totally helpless and 10: death).14–17 All patients were assessed with analysis of cerebrospinal fluid (CSF) by lumbar puncture. Lupus anticoagulant (LA), anticardiolipin (aCL) antibody of IgG and IgM isotype, anti-β2 glycoprotein-I antibody (a-β2GP1) of IgG and IgM isotype were measured according to recommended procedures.18 In serum that were measured complement levels (C3 and C4, using nephelometry) and the presence of autoantibodies: antinuclear antibodies (ANA, indirect immunofluorescence (IFI) using Hep-2 cell culture with a cut off:1/80) and anti-double-stranded DNA (anti-dsDNA, using enzyme-linked immunosorbent assay, cut off:30) at least. Spinal and brain MRI was performed on a 1.5 tesla machine. Spinal cord lesions which extend through three or more vertebral segments on sagittal spinal MRI is regarded longuitudinally extensive transverse myelitis (LETM).

Results

Five patients with myelitis and SLE were observed. During the same period, 233 patients with SLE were recorded by the Department of Immunology. In view of these data, the prevalence of myelitis in this population can be estimated at 2.1%. In the same way, 46 acute myelitis were observed by the Department of Neurology, 10.8% of them were associated with lupus.

Table 1 shows the demographic and clinical characteristics: all were women, the mean age was 25.4 years (range:19–39). In three of five cases, myelitis was one of the initial SLE manifestations. Fever was the most frequent associated manifestation (4/5). One patient presented concomitant glomerulonephritis and another patient had acute arthritis. The SLEDAI showed high disease activity (>4) in three patients. The clinical picture was acute and severe; all with motor deficit, sensory level and urinary retention. The AIS and EGS was high.

Table 1 Demographic and clinical characteristics of our patients
Full size table

MRI findings, CSF and serological features are shown in Table 2. Spinal MRI was abnormal in all cases, three presented a longitudinally extensive myelitis (Figures 1 and 2). Brain MRI showed no significant lesions. Serum analysis revealed positive ANA at a high titer in all patients, four had low complement levels and three had anti-phospholipids positive. Only two patients were tested for the presence of aquaporin-4 antibody (AQP4-ab) by IFI with negative results. CSF analysis was abnormal in all cases with variable results.

Table 2 MRI findings, CSF and Laboratory features of our patients
Full size table
Figure 1
figure 1

Spinal MR, (a) Case 1: longitudinal extensive lesion with increase signal intensity in T2 from the cervicomedullary junction to T8. (b) Case 3: central and longitudinal extensive lesión from the T3 to conus.

Full size image
Figure 2
figure 2

Spinal MR, Case 4: central and longitudinal extensive lesion from the T1 to conus.

Full size image

All patients received high-dose intravenous (i.v.) methylprednisolone, 1 g daily for 5 days. Patients 1, 4 and 5 were also treated with cyclophosphamide. In cases that showed positive antiphospholipid antibodies (1, 2 and 5) anticoagulation was initiated with low-molecular-weight heparin. Owing to the poor response, patients 3 and 4 received plasmapheresis or i.v. immunoglobulin (Table 3). The treatment produced only partial improvement or no benefits. One patient died due to sepsis (probably secondary to mesenteric ischemia). The others showed significant disability. The neurological impairment at six months was high: AIS A=1, C=1 and D=2, and EGS=6 (walks with permanent unilateral support, walking distance <100 m without rest) to 8 (chair restricted, unable to take a step, effective use of arms).

Table 3 Treatment and Outcome of our Patients
Full size table

Discussion

NPSLE is one of the major causes of morbidity and mortality in patients with SLE. Myelitis is one of these syndromes and may be the initial SLE manifestations (60% in our series), though myelitis also occur many year after SLE diagnosis.4–14 We estimated the prevalence of myelitis in SLE at 2.1%, in line with previous reports.4 The clinical presentation may be hyperacute, acute or subacute with motor impairment, sensory level and sphincter disturbances. There are often systemic symptoms such as fever and other signs of disease activity with or without involvement of other organs.4–14 Nearly two-thirds of cases occurred in association with active lupus, and one-third occurred in low disease activity.13 In other series12–14 the prevalence of relapse is high (50–60%) but in our series no patient had a relapse probably for the short follow-up period.

In patients with suspected myelopathy a contrast-enhanced spinal cord MRI and lumbar puncture should be performed. MRI is useful to exclude cord compression and detect T2-weighted hyperintense lesions. These lesions are seen in most patients and frequently (60–75%) are LETM.9,12–14 such as was observed in our cases. CSF analysis is abnormal in most patients and show great variability:4–14 there may be only a slight increase of proteins and cells to polymorphonuclear pleocytosis with hypoglycorrhachia in hyperacute cases (for example, in case 4). Serum analysis revealed positive ANA at a high titer in all patients. Other autoantibodies, particularly anti-DNA, are also positive. Often, complement levels are low.

In the patients with LETM and SLE the AQP4-ab was positive in a variable percentage (12–57%).12,19,20 Whereas the AQP4-ab detection is a specific marker that distinguishes neuromyelitis optica (NMO) from other etiologies, in these patients, coexisting pathologies is postulated.19 In two of our patients have been tested for the presence of these antibodies by IFI with negative results. The sensitivity depends on the method of detection: assay based on cells transfected with AQP4-M23 isoform were the most sensitive but is not widely available.21

The presence of antiphospholipid antibodies is found in a high percentage (60% in our cases, 50–70% in the literature).10–14 We started anticoagulant therapy in these cases, but its usefulness is controversial and is not routinely indicated.10,11,22 The antiphospholipid antibodies may have a pathogenic role related to the interaction with certain antigens of the spinal cord and not necessarily secondary to thrombosis, or they could be an epiphenomenon.10,11,14

Birnbaum et al.12 propose the existence of different subtypes of myelitis associated with lupus: patients with ‘gray matter myelitis’ and patients with ‘white matter myelitis’. The first one has a more acute and severe clinical picture, flaccidity, hyporeflexia, with more symptoms and signs of systemic and local inflammation, extensive central lesions and worse prognosis. In this cases, recognition of fever and urinary retention as prodromes of irreversible paraplegia may allow early diagnosis and treatment. The patients with ‘white matter myelitis’ have spasticity and hyperreflexia and greater association with antiphospholipid antibodies and AQP4-ab. This group has higher recurrence rate but better prognosis. In our series, cases 1, 2 and 5 can be classified within subtype of ‘white matter myelitis’ associated to antiphospholipid antibodies and case 3 and, in particular, case 4, can fall into the subtype of ‘gray matter myelitis’.

There are no large prospective studies on the treatment of lupus myelitis.22–24 The European League Against Rheumatism in 2010 issued recommendations for NPSLE management.22 If there is high suspicion of SLE-myelitis, the combination of methylprednisolone i.v. and cyclophosphamide i.v. should be initiated promptly. A maintenance therapy with a less intensive immunosuppression may be considered to prevent recurrence. Retrospectives studies support these recommendations.14,23,24 Plasma exchange therapy, i.v. immunoglobulin and rituximab25 have been used in refractory cases and anticoagulation therapy in antiphospholipid-positive with variable results.10,11,22 In our experience the treatment produced only partial improvement.

The prognosis is variable. In our cases, despite treatment, none had complete recovery, 40% had partial improvement, 40% showed no improvement and one case died in the hospital, probably due to a complication related to SLE. In previous reports4–14 the outcomes are somewhat better, and depend on the initial clinical picture. In the most recent series published,14 ~60% have a complete or nearly complete recovery at year.

Although this study is retrospective and with limited number of patients, it is the largest case series of patients with SLE and myelitis reported in Argentina and provides the data to international casuistry. The five cases described show the heterogeneity of the clinical, radiological and serological features. Rapid diagnosis and early treatment with methylprednisolone and cyclophosphamide is recommended to improve the prognosis, but, in our experience, the cases were severe and with poor response to treatment. In accordance with expert opinions, prospective and controlled studies and perhaps the creation of registry for SLE patient with myelitis are needed to help understand this disease and establish more efficient treatment strategy.