This case report describes a pediatric patient with a new diagnosis of ANA-negative SLE with the initial findings of diffuse cerebral edema and acute leukoencephalopathy on imaging, characterized symptomatically only by headache, blurry vision and Grade IV papilledema on examination. Her fevers were thought to be due to active lupus as opposed to a CNS manifestation, largely because her cerebral involvement was quite diffuse as opposed to focused near the hypothalamus, where temperature regulation takes place. Hypothetically, if the fevers were secondary to CNS involvement she would have spiked fevers much more frequently given the severity of her CNS involvement on imaging. Instead, she spiked fevers intermittently. According to the literature, the finding of diffuse cerebral edema with or without leukoencephalopathy in NPSLE is extremely rare, and if present, develops later in the disease course and typically with other systemic signs of the disease. SLE patients with concern for neuropsychiatric involvement most commonly present with headaches, seizures, stroke, depression, and/or cognitive dysfunction as the sign of central neurologic involvement (4-7). There have been case reports of patients with isolated intracranial hypertension as the only sign of neuropsychiatric lupus, and a few with intracranial hypertension with associated leukoencephalopathy (see table 1), but we were unable to find an instance of a patient with isolated diffuse leukoencephalopathy as the presenting sign of lupus upon initial diagnosis. Many other patients described had other systemic signs/symptoms and already carried the diagnosis of SLE. Furthermore, most cases in the literature had a positive ANA in their workup to further assist in the diagnosis of SLE. Additional risk factors associated with development or worsening of NPSLE include generalized SLE activity or damage, history of previous or concurrent other major NPSLE, and antiphospholipid antibodies (3, 4). Our patient did not have any of these risk factors.
As shown in Table 1, most of the case reports reviewed described intracranial hypertension with or without leukoencephalopathy (8-18). Of the three cases that were children (ages, 7y, 11y, and 14y), all had a positive ANA, and one out of the three pediatric cases had a previous diagnosis of SLE. All cases reported multiple other clinical manifestations of SLE in addition to CNS involvement, unlike our patient. Imaging findings reported were consistent in showing diffuse hyperintensities on MRI suggestive of leukoencephalopathy, similar to our patient. Patient outcome across case reports were variable, with some making a full recovery and others unfortunately succumbing to their disease. Various methods were used for treatment, with high-dose steroids being a unifying treatment choice.
The pathophysiology has been explored in SLE cases of idiopathic intracranial hypertension (IIH) with diffuse leukoencephalopathy. There are multiple theories, including the possibility of immune-complex mediated damage, autoantibodies interacting (either directly or indirectly) with antigens on neuronal cell membrane, intrathecal cytokine production, and microangiopathy (1, 2). Various autoantibodies found in relation to increased incidence of NPSLE include anti-phospholipid antibodies, anti-ribosomal P antibodies, and microtubule-associated protein-2 antibodies. These autoantibodies theoretically target endothelial cells, prostacyclins, protein C-S complex, and platelets, leading to acute impact on coagulation and chronic proliferative vasculopathy (2). Cranial MRI is currently the anatomic imaging modality of choice for these patients, and displays high sensitivity but low specificity for NPSLE. Most NPSLE patients (40-80%) show small punctate focal lesions in periventricular and subcortical white matter areas on imaging, not necessarily associated with diffuse brain edema as was the case with our patient, and cerebral angiography typically is normal (2). A vast array of findings in the literature make it very difficult to determine exact pathophysiology of diffuse leukoencephalopathy with associated cerebral edema. Pathophysiology is likely multifactorial, involving autoantibody reactivity as well as an underlying propensity for cerebral damage.
Interestingly, our patient is also unique in that she was diagnosed with SLE but had a negative ANA test noted on two occasions during her hospitalization and again after discharge. Additionally, she had a positive dsDNA antibody, multiple antibodies to extractable nuclear antigens (ENA) and low complements (Table 2). When considering our patient’s infectious workup, this was possibly representative of a diffuse polyclonal B-cell response and cross-reactivity resulting in false positive mycoplasma, EBV, CMV and the abnormal fluorescent for California virus, western and eastern equine viruses, and St. Louis virus antibodies. Additional autoantibody testing was not pursued until a thorough infectious, metabolic and neurologic workup was negative or inconclusive. Upon discussion with our metabolic team, her lab workup and imaging did not support a primary metabolic disorder. Her MR spectroscopy did comment on “elevation of the glutamine/glutamate complex in short echo MRS in both the left frontal lobe and right basal ganglia” (which can be seen with urea cycle disorders but typically in the more subacute/acute phase). However, her glutamine, citrulline and arginine were normal on serum amino acids, and urine orotic acid was normal, which does not support a urea cycle disorder. L 2 hydroxyglutaric aciduria was noted to be highly unlikely in the context of normal urine organic acids and normal lysine on serum amino acids. Mitochondrial disorders were less likely due to normal lactate and normal alanine on serum amino acids. Possible hereditary leukodystrophies were considered, however without cognitive decline, motor deterioration, dysmorphism, hepatosplenomegaly, or abnormal tone this was considered less likely. APOPT1 gene (associated with cavitating leukoencephalopathy with cytochrome C oxidase deficiency) testing was sent and was negative/unremarkable. Whole exome sequencing did not identify any variants that could be interpreted to definitively explain her reported phenotype, decreasing the possibility of a hereditary leukoencephalopathy or underlying genetic defect.
Neurology followed her case closely. Due to significant cerebral swelling on imaging and tonsillar ectopia, they did not feel it was safe to do a lumbar puncture due to risk of herniation. Autoimmune encephalitis was a consideration in both neurology and rheumatology differential diagnosis. However, her neurologic exam remained normal, she was without seizure-like activity or behavioral changes, her EEG was without epileptiform discharges, and her imaging was notable for diffuse cerebral edema with leukoencephalopathy as opposed to imaging changes limited to the hippocampi (as seen in anti-α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) antibodies) or limbic system (as seen in anti-voltage-gated potassium channel (VGKC) antibodies). She had no evidence of an underlying tumor or malignancy on imaging or lab testing, which can be associated with autoimmune encephalitis. During her hospital stay, an autoimmune encephalitis panel in serum was suggested, however her workup continued to return more convincing for SLE, so this testing was not pursued.
Ultimately, it was the evidence of hypocomplementemia and a highly positive dsDNA antibody that led to the additional autoantibody testing and ultimately her diagnosis, though it took an extremely thorough workup and evaluation by other subspecialties prior to reconsidering SLE as her diagnosis. The dsDNA testing in our laboratory is completed by immunofluorescent assay with Crithidia luciliae, a flagellate parasite containing circula dsDNA without other nuclear antigens in the kinetoplast (19). Crithidia luciliae and Farr assay detect higher-avidity antibodies, with the greatest specificity for the diagnosis of SLE (19). As our testing methods are quite specific, this prompted pursuing further testing for rheumatologic disease. She had multiple positive ENA antibodies (Smith, RNP, SSA, SSB), ribosomal P antibody and neuronal antibody on further evaluation (Table 2). Our patient did not meet classification criteria for SLE based on ACR, SLICC, or ACR/EULAR. However, her clinical picture, serologies and response to treatment support the diagnosis of SLE.
Our patient was treated with both Rituximab and Cyclophosphamide for her SLE. We chose to give Cyclophosphamide given the severity of her CNS involvement and inflammation with brain edema. Rituximab was also given because of our patients’ significant auto-antibody load. B cells have an important role in the pathogenesis of SLE and lead to significant autoantibody production. Rituximab effectively helps reduce this level of auto-antibodies. As the time to effectiveness for rituximab and CNS manifestations is unknown, it was important to treat very aggressively with both Cyclophosphamide and Rituximab. She responded well to these medications and was able to discharge home shortly after obtaining immunosuppressive therapy.
Our patient is a unique representation of SLE, but her case may suggest that for some specific instances, where thorough infectious, metabolic, and neurological testing are inconclusive, that ANA testing alone may not be sufficient for screening for SLE. There are few descriptions of such cases in the literature but some do describe the importance of screening patients with suspected rheumatologic disorders with more than just the basic ANA screen (20). Testing with other markers specific to SLE disease activity, e.g. complement levels and anti-dsDNA, proved critical in the diagnosis of SLE for our patient. On treatment of her disease over time, her anti-dsDNA antibody decreased significantly in line with her disease activity (Table 2), confirming response to treatment.
Various studies propose several reasons as to why some patients with the diagnosis of SLE have negative ANA screenings. These include the prozone or hook effect, the entity of an ANA negative SLE patient, or technical issues with the ANA screen itself (21). We believed the prozone effect may have been responsible for our patient’s negative ANA result. The prozone effect occurs in cases of very high antibody concentrations and is thought to be responsible for negative immunoassays that involve the detection of antigen-antibody complexes (22). With these assays, there is dependence on agglutination to reveal the presence of the antibody and thus confirm a positive test. With the prozone effect, the antibody concentration is so high, it interferes with the clumping of antigen-antibody complexes resulting in a seemingly negative result. ANA testing was repeated in our patient on multiple occasions and was consistently reported as negative. The ANA test at our facility is conducted via immunofluorescence with HEp-2 cells as the substrate molecule. The American College of Rheumatology position statement on ANA testing states the use of IIF as the gold standard method for ANA screening, specifically IIF on HEp-2 cells (23). Her ANA test being negative initially drew the diagnosis away from the possibility of SLE, and it ultimately took a very thorough negative workup in other subspecialties before the diagnosis of SLE was re-considered. If the prozone effect is the cause of our patient’s negative ANA result, future ANA tests for our patient could turn positive with treatment of the patient’s disease and reduction in antibody burden.