Abstract

A 67-year-old woman with a history of rheumatoid arthritis (RA) presented to a local hospital for a 3-month history of episodes of confusion, disorientation, loss of touch with reality, and failure to recognize familiar people, with complete symptom resolution in between episodes. The day prior to transfer to our hospital, she was found in a confused state by neighbors. Magnetic resonance imaging (MRI) at the local hospital demonstrated white matter lesions in the brain parenchyma. She was transferred to the Oregon Health & Science University hospital for further management. Upon arrival she was completely asymptomatic. There were no focal neurologic symptoms, and a review of systems was unremarkable. She had had seropositive nodular, erosive RA for approximately 20 years, with severe hand arthritis needing surgical fusion of the right wrist, bilateral thumbs, and index finger metacarpophalangeal joints. She took methotrexate and hydroxychloroquine for many years and had taken infliximab for the past 3 years, with a good response. Her last infliximab infusion was 2 weeks prior to the current admission. Her RA was in remission except for a recent flare, which was treated with a short course of prednisone (total of 130 mg of oral prednisone over 10 days) and an intraarticular steroid injection (10 mg of methylprednisolone). She also carried a diagnosis of celiac disease due to several years of intermittent diarrhea and cramping, which had dramatically improved on a gluten-free diet. She had a previous diagnosis of Hashimoto thyroiditis, which was being treated with thyroid hormone replacement. She had a history of smoking one-half pack per year previously, but had quit 1 year prior. Her current medications included methotrexate 25 mg injection every week, infliximab 300 mg (6.2 mg/kg) infusion every 8 weeks, hydroxychloroquine 400 mg by mouth daily, folic acid 1 mg by mouth daily, levothyroxine 88 μg by mouth daily, oxycodone 5 mg by mouth as needed, estradiol 0.5 mg daily, and medroxyprogesterone 2.5 mg by mouth daily. On the initial examination, the patient was awake, alert, and oriented to person, place, and time. She weighed 48.5 kg (body mass index 19 kg/m2). Her oral temperature was 36.6°C, blood pressure was 106/45 mm Hg, respiratory rate was 12 breaths/minute, and heart rate was 70 beats/minute. The general examination was unremarkable. The musculoskeletal examination was notable for small rheumatoid nodules over the right foot, with changes consistent with previous fusion surgeries as mentioned above and without active synovitis in any joint. The neurologic examination was normal. The initial blood evaluation was unremarkable (Table 1). Inflammatory markers as well as an immunology panel were negative. The initial infectious evaluation, including cerebrospinal fluid (CSF) evaluation, was negative. An MRI scan of the brain on admission showed cortical and subcortical nonenhancing T2 and fluid-attenuated inversion recovery hyperintense foci (Figure 1A). On hospital day 4, the patient developed generalized tonic–clonic seizures with prolonged postictal confusion and rapid mental decline. A computed tomography (CT) scan and CT angiogram of the head were essentially unremarkable. A CT scan of the chest, abdomen, and pelvis revealed right upper lobe pulmonary nodules, including a cavitary lesion (Figure 1B). A biopsy of these nodules revealed necrotizing granulomas with surrounding fibrosis and mixed cellular infiltrate, without evidence of vasculitis. All of the cultures from this lung specimen were negative. A transthoracic echocardiogram failed to show vegetations. Meanwhile, the patient's mental status continued to worsen; she became comatose, now demonstrating repetitive orofacial myoclonus. Several electroencephalograms failed to demonstrate epileptic focus and suggested a nonspecific global encephalopathy. A followup brain MRI scan showed interval worsening of brain lesions showing diffusion restriction (Figures 1C and D). A brain biopsy of a right frontal lesion obtained under CT guidance revealed reactive microglia, astrocytes, diffuse edema, and patchy loss of axons and myelin (Figure 2). Small vessels displayed focal perivascular macrophages and lymphocytes, rare areas of incomplete fibrinoid necrosis, without associated mural inflammation or vessel wall destruction suggestive of frank vasculitis. The biopsy also failed to show any fungal, bacterial, mycobacterial, or Whipple's disease organisms, and features of prion disease, cerebral amyloidosis, and cerebral autosomal-dominant arteriopathy with subcortical infarcts and leukoencephalopathy (CADASIL). A, Coronal fluid-attenuated inversion recovery (FLAIR) image (hospital day 1) showing cortical and subcortical foci of hyperintense signal in the frontal lobes without mass effect (arrows). B, Axial chest computed tomography scan (hospital day 1) demonstrates a cavitary right pulmonary nodule (arrow). C and D, Coronal diffusion-weighted FLAIR images (hospital day 10) show multifocal areas of increased T2 signal affecting the cortex and subcortical white matter (arrows). E, Axial T2-weighted and F, coronal FLAIR images on HD45 show marked interval progression of multifocal T2 hyperintensity with relative sparing of the cortex (straight arrows) in both cerebral hemispheres, with mass effect, compression of the right lateral ventricle, and a mild midline shift to the left (curved arrow). Scale of the figure is 25 μm. A, Subcortical white matter with patchy parenchymal loss. B, Vascular congestion with reactive plump endothelial cells (black arrow) and vascular wall edema and degenerative changes (blue arrow) show widespread tissue vacuolization indicative of edema. C, Scant perivascular infiltrates (arrows) without vascular infiltration or necrosis to suggest vasculitis. D, Scant perivascular CD3-positive T lymphocytes. E, Occasional perivascular CD68-positive macrophages without overt vascular damage. F, CD68 immunohistochemistry showing markedly increased activated microglial cells in the cerebral cortex and white matter. In the absence of a unifying diagnosis and rapid deterioration, the patient was empirically treated with pulse methylprednisolone 1 gm daily for 3 days. She was also started on antituberculosis (anti-TB) and antifungal therapy, both of which were stopped once infections were ruled out. After approximately 1 week, the patient started to show improvement. Despite a followup MRI on HD45 showing worsening edema with a midline shift (Figures 1E and F), the patient showed continued clinical improvement in mental and motor functioning, and was subsequently discharged with no residual neurologic abnormalities. This is a case of a 67-year-old woman with longstanding seropositive erosive RA treated with immunosuppressive medications, including methotrexate and a tumor necrosis factor α (TNFα) inhibitor, with an additional history of possible Hashimoto thyroiditis and celiac sprue. Her presentation with confusion and MRI abnormalities suggested a broad differential diagnosis, including infections, malignancy, an adverse effect of a medication or toxin, or an immune-mediated disease. Metabolic, endocrine, and vascular disorders, including posterior reversible encephalopathy syndrome (PRES), Hashimoto encephalitis, amyloid angiopathy, and other miscellaneous causes, were also considered and are listed in Table 2. With the patient receiving immunosuppression and clinical improvement after stopping infliximab, infection was the foremost consideration. Among various central nervous system (CNS) infections, CNS TB can present in a slowly progressive manner, presenting as meningitis, tuberculoma, and tubercular vasculitis (1). A lack of meningeal enhancement on MRI scans, negative tissue cultures, and polymerase chain reaction (PCR) testing and rapid improvement provided evidence against CNS TB. Viral, fungal, protozoan, and spirochete infections of the brain as well as Whipple's disease and prion disease were excluded with serology, cultures, PCR, and histopathology, as appropriate (Table 1). The patient's continued improvement in the absence of a specific antimicrobial treatment also suggested against an infectious etiology. Immunosuppression is an important risk factor for primary CNS lymphoma (PCL) (2). Characteristic MRI findings of PCL include iso- to hypointense deep brain or periventricular white matter lesions (3) with sharp borders in most (4) and mass effect in more than half of the cases, as opposed to the confluent T2 hyperintense lesions seen in our patient. The rapid progression of MRI lesions and rapid clinical improvement in our patient are quite unusual for PCL. Metastasis to the brain was also possible, but a full-body CT scan failed to reveal any other suspicious areas of malignancy. The CSF evaluation was negative for malignant cells and the brain biopsy conclusively ruled out a malignancy. Primary angiitis of the CNS (PACNS) is a vasculitis of small and medium-sized arteries of the brain parenchyma (5), presenting with headache, seizures, and encephalopathy (6). CSF evaluation shows a lymphocytic pleocytosis with an elevated protein level. Brain MRI mostly shows multiple focal parenchymal infarctions (7). Cerebral angiogram shows vasoconstriction or “beading” in multiple vascular beds. Brain biopsy shows classic granulomatous vasculitis with Langerhans' or foreign body giant cells (8). In our patient, CSF did show mild lymphocytic pleocytosis, but the MRI showed confluent white matter lesions suggestive of diffuse vasogenic edema without infarction, and a brain biopsy failed to show vasculitis, making a diagnosis of PACNS unlikely. The brain biopsy was performed under CT guidance, which decreased the likelihood of missing the pathology. Antineutrophil cytoplasmic antibody–negative granulomatosis with polyangiitis (Wegener's) was considered due to the finding of necrotizing granulomas in the lung. However, the lack of sinus symptoms and renal abnormalities, and both the lung and brain tissues failing to show vasculitis, made it less likely. The absence of recurrent oral or genital ulcers, uveitis, arterial or venous thromboses, glandular enlargement, and negative SSA/SSB antibodies, as well as neuroimaging and brain tissue suggestive of vasogenic and cytotoxic edema as opposed to infarction, demyelination, or vasculitis typical of Sjögren's syndrome or Behçet's syndrome, made these and other vasculitides less likely. Rheumatoid vasculitis was also thought to be less likely because of its usual presentation in patients with severe uncontrolled nodular RA, with high titers of rheumatoid factor and vasculitic ulcers present in almost 90% of cases. Our patient had well-controlled RA on anti-TNF therapy without skin ulcers, and the brain biopsy failed to show vasculitis. Antinuclear antibody–negative systemic lupus erythematosus (SLE) with CNS involvement was considered less likely because of the absence of typical mucocutaneous, serosal, hematologic, and renal manifestations of lupus, a negative extractable nuclear antigen antibody panel, the presence of normal complements, and an erosive, as opposed to a nonerosive, arthritis typical of SLE. Nonvasculitic autoimmune inflammatory meningoencephalitis is an ill-defined syndrome of unclear pathogenesis (9, 10) that presents as rapidly progressive dementia, elevated inflammatory markers, and evidence of autoimmunity (e.g., high titers of antithyroperoxidase antibodies in Hashimoto encephalitis). Brain biopsy shows an intense perivascular lymphocytic infiltrate and reactive gliosis without frank vasculitis. These biopsy features were present in our patient, but a normal erythrocyte sedimentation rate and the absence of antithyroperoxidase and thyroglobulin antibodies provided evidence against a diagnosis of Hashimoto encephalitis. The patient's age, CSF findings, neuroimaging, and brain biopsy did not support a diagnosis of multiple sclerosis, CADASIL, or cerebral amyloid angiopathy (Table 1). Several reports of demyelination associated with TNFα inhibitor therapy have caused concern in many cases such as ours. However, it is an extremely rare complication (11), with approximately 500 cases reported to date and the majority occurring during the first year of therapy. Our patient had been receiving a TNF inhibitor for more than 3 years, and MRI scans and a brain biopsy did not show specific demyelinating lesions. At the 1-month followup visit, the patient had no focal neurologic findings. A followup MRI 5 months after the initial presentation (Figure 3) showed complete resolution of cerebral lesions. The presence of seizures in association with cerebral dysfunction and this neuroimaging pattern is characteristic of PRES, and a diagnosis of PRES was made retrospectively. The patient was restarted on methotrexate and hydroxychloroquine, and prednisone was successfully tapered off. At the 6- and 12-month followup visits, she remained without neurologic deficits except for very mild cognitive impairment. Coronal fluid-attenuated inversion recovery and T2-weighted axial images A, at the 1-month followup visit and B, 5 months after the initial presentation show resolution of vasogenic edema corresponding with complete neurologic recovery. This was a challenging case of an elderly woman with longstanding seropositive erosive RA receiving immunosuppression with anti-TNF therapy who developed rapidly progressing neurologic dysfunction followed by complete recovery. With the exhaustive negative infectious, malignancy, and autoimmune evaluation, the imaging pattern and brain biopsy are most consistent with the diagnosis of PRES. The first cases of reversible brain lesions in patients with lupus were described by Aisen et al (12) and Sibbitt et al (13). The condition was named reversible posterior encephalopathy syndrome in a 1996 case series (14). The pathophysiology of PRES remains unclear. Fifty to 70% (15) of these patients have high blood pressure. Therefore, dysfunction in cerebral autoregulation secondary to hypertension is thought to be one possible mechanism. However, up to 20–30% of patients are normotensive (15), and PRES has been seen in association with eclampsia, posttransplant immunosuppression, and sepsis, suggesting that endothelial injury with increased levels of adhesion and inflammatory molecules may underlie its pathogenesis (15). Hypoperfusion has been suggested as another possible mechanism, as evidenced by low relative cerebral blood volume and increased lactate concentrations on proton MR spectroscopy (15) in the respective brain areas. It is unclear why posterior circulation is favored, but it could be due to a relative lack of sympathetic innervation. PRES is typically seen in association with a variety of clinical settings (Table 3). In a large case series (16), the mean age of PRES patients was 48 years, with women affected more frequently than men. The most common clinical presentations included seizures (74%), encephalopathy (28%), headache (26%), and visual disturbances (20%). Common conditions associated with PRES include hypertension, cytotoxic medications, sepsis, preeclampsia or eclampsia, and multiple organ system dysfunction. Autoimmune diseases can be found in as many as 45% of patients with PRES (16), including thrombotic thrombocytopenic purpura, SLE, inflammatory bowel disease, and RA, among others (Table 3). Lately, PRES has been noted with increasing frequency in patients with SLE (17). PRES has also been associated with immunosuppressive medications, with common ones being cyclophosphamide, tacrolimus, cyclosporine, and mycophenolate. Treatment with infliximab and etanercept complicated by PRES has been reported previously (18, 19). Tanno et al reported 2 cases of CNS demyelination following infliximab administration, both showing partial resolution in brain edema after discontinuation of treatment (20). Interestingly, both patients also received methotrexate and prednisone. Neuroimaging remains the diagnostic modality of choice in PRES, typically showing symmetric hemispheric edema in occipital lobes (15), but other areas are also frequently affected. According to one report, cerebellar, cortical, and parietooccipital involvement was more commonly found in patients with autoimmune diseases, sepsis, and preeclampsia/eclampsia, respectively (16). Diffusion-weighted MRIs show diffusion restriction in approximately one-quarter of cases (21). Catheter angiography shows diffuse or focal vasoconstriction and can even show a “string of beads” appearance (22) reminiscent of reversible cerebral vasoconstriction syndrome, which can include concurrent PRES in approximately 10% of cases. Histopathologic examination of the brain parenchyma of patients with PRES shows vasogenic edema and reactive astrocytes and microglia, scattered macrophages, and lymphocytes, often without substantial inflammation, ischemia, or neuronal damage (23). There have been no randomized controlled trials looking at the treatment of this condition. Most patients show near complete resolution of brain lesions within weeks (16). Early detection and management are critical because although reversible by definition, complications such as status epilepticus, hemorrhage, or infarction can cause substantial morbidity and mortality in approximately 15% of cases (15, 21). Management is largely supportive, with cautious lowering of blood pressure, control of seizures with anticonvulsants, withdrawal of the offending agent, and management of the underlying disorder (24). Lowering immunosuppression may be important in this setting; however, in a Mexican cohort, high-dose corticosteroids were used for treatment of PRES (25). Understanding the pathophysiology and treatment remains an area of great interest and discussion. Posterior reversible encephalopathy syndrome. 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. Garg 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. Garg, Woltjer, Hamilton, Neuwelt, Rosenbaum. Acquisition of data. Garg, Woltjer, Hamilton, Neuwelt, Rosenbaum. Analysis and interpretation of data. Garg, Woltjer, Hamilton, Rosenbaum.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call