Clinical and Neuropathological Variability in Clinically Isolated Central Nervous System Whipple's Disease

Abstract

Central nervous system Whipple's disease (CNS‐WD) with clinically isolated neurological involvement is a rare condition fatal without an early diagnosis. We aimed to present clinical and neuropathological features of three cases of pre‐ or post‐mortem polymerase chain reaction confirmed CNS‐WD with distinct clinical presentation, outcome and pathological findings. One patient had an acute onset with spinal and brainstem involvement and died without CNS‐WD diagnosis after 14 weeks. Neuropathology showed extensive inflammatory and necrotizing lesions with abundant foamy periodic‐acid‐Schiff (PAS)+ macrophages. The second patient had a subacute evolution with late CNS‐WD diagnosis and death occurring 18 months after onset despite antibiotic treatment. Brain examination showed inflammatory lesions in the brainstem, thalamus and cerebellum, and abundant foamy PAS+ macrophages. The third case was diagnosed within 4 weeks of onset and treated with an excellent response. He died after a disease‐free period of 24 months of unrelated causes. Neuropathology showed cystic residual lesions devoid of microorganisms without inflammatory reaction. CNS‐WD may have an acute or subacute course with variable response to treatment. Accordingly, subjacent lesions may be those of a severe acute necrotizing encephalitic process or subacute inflammatory lesions involving diencephalic, brainstem, cerebellar and spinal regions. Chronic, cavitary brain lesions may be sequelae of a successful treatment. Early diagnosis should allow appropriate treatment and improve prognosis.

Introduction

Whipple's disease (WD) is a rare, multisystemic chronic infectious disease. It is caused by the Gram‐positive bacterium Tropheryma whipplei, described for the first time by George Hoyt Whipple in 1907 24. The first successful treatment with antibiotics was reported in the 1950s 17. It usually affects middle‐aged men who present with arthralgia/arthritis, diarrhea, weight loss, fever and lymphadenitis. Without appropriate treatment, the disease is fatal in all cases. Neurologic involvement has been reported in about 15%–25% of cases, usually coexisting with systemic signs and symptoms and frequently related with antibiotic discontinuation 11, 12. Clinically isolated neurologic involvement of WD is very rare and is misdiagnosed in many occasions because of its protean clinical manifestations that can mimic almost every neurological condition 6, 16. The diagnosis of isolated central nervous system (CNS)‐WD requires a positive polymerase chain reaction (PCR), periodic‐acid‐Schiff (PAS) positivity or inmunoreactivity for T. whipplei antigens of cerebrospinal fluid (CSF) macrophages or brain biopsy samples 20. In some cases, without a prompt correct diagnosis, the evidence of CNS infection by T. whipplei is made only at autopsy 9. Classical neuropathological features include those of a lymphocytic encephalitis with the presence of abundant perivascular foamy macrophages filled with PAS, Gram‐ and Silver‐positive rod‐shaped structures, also called sickle particle containing cells or Sieracki cells 7, 19, 22, 23, 25.

We present the clinical and neuropathological features of three cases of pre‐ or post‐mortem PCR‐confirmed WD with clinically isolated CNS involvement. All three patients had a distinct clinical, neuroimaging and outcome pattern associated with acute active, subacute and chronic brain lesions.

Material and Methods

Clinical data

Clinical charts were retrospectively reviewed by the treating physicians. All the available clinical data as well as complementary investigations were used to compose the clinical pictures of the patients.

PCR and sequencing

CSF

DNA extraction used 200 μL of CSF and was carried out with the DNA extraction kit EZ1 virus minikit 2.0 and the EZ1 extractor, following manufacturer's protocol (QIAGEN, Hilde, Germany). Five microliters of the extracted DNA were used for PCR. PCR amplifying the hsp65 gene of T. whipplei was performed using the conditions previously described 14. The PCR product was loaded in a 1.5% agarose gel, purified using Wizard SV gel and PCR clean‐up system (Promega, Madison, WI, USA) and it was further sequenced using a Big Dye Terminator sequence kit (v. 2.0 Applied Biosystems, Foster City, CA, USA) according to the manufacturer's instructions.

Brain tissue

Approximately 10–40 mg of brain tissue was incubated with 250 μL of buffer G2 from EZ1 Tissue kit. Ten microliters of proteinase K were added and the mixture was incubated overnight at 56°C. The DNA was extracted with the EZ1 extractor following manufacturer's protocol (QIAGEN, Hilde, Germany). PCR was carried out following the conditions mentioned earlier.

Neuropathological work‐up

Written informed consent for brain donation for diagnostic and research purposes was obtained from patients or next of kin. The left brain hemisphere and hemicerebellum were sliced in the fresh state in 0.5–1.0‐cm thick coronal sections and frozen at −80°C, whereas the right hemisphere, hemicerebellum and alternate sections of brainstem were fixed in 10% buffered formaldehyde solution for 4 weeks. For histopathological evaluation, 5‐μm thick sections were cut from formalin‐fixed and paraffin‐embedded tissue from multiple brain areas: frontal, temporal, parietal and occipital cortices, motor cortex, anterior cingulum, anterior and posterior basal ganglia, anterior, medial and posterior thalamic nuclei, hippocampus and parahippocampal gyrus, amygdala, n. basalis Meynert, midbrain, pons, medulla oblongata, olfactory bulb, cerebellar vermis and dentate nucleus, as well as cervical, and where available, thoracic and lumbar segments of spinal cord.

Sections were stained with hematoxylin and eosin, luxol fast blue, PAS, PAS followed by diastase digestion, and silver methenamine. Immunohistochemistry was performed using different monoclonal (mc) and polyclonal (pc) primary antibodies: anti‐bA4‐amyloid (DAKO, Glostrup, Denmark, mc, clone 6F/3D, dilution 1:400), anti‐phosphorylated tau (Thermo Scientific, Rockford, IL, USA; mc, clone AT8, dilution 1:200), anti‐ubiquitin (DAKO, pc, dilution 1:400), anti‐alpha‐synuclein (Novocastra, Newcastle, UK; mc, clone KM51, dilution 1:500), anti‐TDP43 (Abnova, Taipei, Taiwan; mc, clone 2E2‐D3, dilution 1:500). For characterization of inflammatory reaction, we used anti‐CD4 (DAKO), anti‐CD8 (DAKO), anti‐CD20 (DAKO), anti‐granzyme B, and anti‐CD68 (DAKO) antibodies.

Immunohistochemistry for the detection of T. whipplei on formalin‐fixed and paraffin‐embedded brain tissue samples was performed at the University of Marseille, France, according to established protocols 2.

Results

Case 1

A 49‐year‐old man was admitted to the hospital for acute onset paraparesis. Several hours after admission he developed progressive tetraparesis with loss of deep tendon reflexes and impaired consciousness that required mechanical ventilation. An MRI evaluation showed T2 signal abnormalities in cervical and thoracic spinal cord with contrast enhancement (Figure 1A) with a discrete increase in CSF protein levels with normal cell count. Treatment with acyclovir, methylprednisolone and intravenous immunoglobulin was initiated with no improvement. A screening of infectious agents [HIV, Epstein–Barr virus (EBV), cytomegalovirus (CMV), human herpesvirus 6, enterovirus, Borrelia burgdorferi, Mycobacterium tuberculosis, Mycoplasma pneumoniae and Treponema pallidum] was negative in CSF and blood. Immunological analyses, cellular immunophenotyping, immunofixation (serum and urine) and hormonal studies were normal. During the following 2 weeks, the subject's condition worsened and he suffered two episodes of recovered asystole. A follow‐up MRI showed extensive white‐matter damage in cervicothoracic spinal cord, brainstem and both thalami (Figure 1B–D). Six weeks after the clinical onset, the patient stabilized in a minimal consciousness state. He died 14 weeks after disease onset due to cardiac arrest.

Figure 1.

Case 1 magnetic resonance imaging (MRI) evolution. A. Initial MRI showing a T2 hyperintense cervicothoracic lesion. B, C and D: MRI performed 4 weeks later showing extensive lesions of the spinal cord (B), brainstem (D), both thalami and internal capsule (C).

Neuropathological examination

Unfixed brain weight was 1290 g. Macroscopically, thickening and yellowish discoloration of spinal leptomeninges was observed. There was diffuse softening of brainstem and cervical spinal cord (Figure 2A). Histologically, multiple lesions were detected throughout the brain parenchyma showing different grades of severity. The most severely affected region was the spinal cord at all levels, which showed almost complete necrosis. Adjacent leptomeninges were thickened and showed perivascular macrophages and a moderate amount of inflammatory infiltrates. There was relative preservation of the anterior and posterior nerve roots. In addition, severe involvement of brainstem including midbrain, pons and medulla oblongata was also observed, with more diffuse, non‐necrotic lesions characterized by abundant parenchymal and perivascular macrophages, lymphocytes, diffuse microglial activation and astrogliosis (Figure 3A1–A4). Inferior olives showed diffuse gliosis and moderate inflammatory infiltrates. In cerebellum, segmental loss of Purkinje cells with frequent axonal swellings in granule cell layers (torpedoes) (Figure 4A2) was detected, associated with white matter rarefaction and diffuse gliosis of dentate nucleus (Figure 4A1). In cerebrum, the most involved areas were basal ganglia, that showed perivascular macrophages and diffuse microglial activation, that were accompanied by extensive macrophage tissue infiltration with perivascular inflammatory cuffs in pallidum, putamen and thalamus, and adjacent white matter. Inflammatory infiltrates were mainly composed of CD3‐ and CD8‐positive T‐cells (Figure 5E), a few CD4 and less CD20 immunoreactive B‐cells. Some of the CD8‐ and granzyme B‐positive T‐lymphocytes were in close contact with morphologically intact neurons (Figure 5F,G). In occipital cortex, a segmental laminar sclerosis involving primary visual cortex was observed, along with secondary atrophy of the lateral geniculate body. Focal areas of perivascular white matter pallor with scattered lymphocytes, macrophages, loss of myelin and better preserved axons were seen in frontal white matter adjacent to the corpus callosum, internal and external capsules, insular white matter, and brainstem white matter. Perivascular macrophages contained PAS‐positive rod‐shaped structures in their cytoplasm. These were also stained by silver stains but not by Ziehl‐Neelsen. An initial PCR from frozen tissue was positive, while the immunohistochemistry for T. whipplei was negative. A second PCR assessment from deep frozen basal ganglia was performed 4 years later in a different laboratory, and gave again a positive result. No concomitant neurodegeneration‐related abnormal protein aggregates such as tau, alpha‐synuclein, beta‐amyloid or TDP43 were detected.

Figure 2.

Comparative neuropathological features of the three cases. A–C: Macroscopic appearance of representative coronal sections of the right hemisphere of the three cases. In A (Case 1) a severe softening of spinal cord was observed (lower inset). In B (Case 2), small necrotic foci are detected in basal ganglia (arrows). In C (Case 3), cystic lesions in frontal white matter are indicated by arrows and in the upper inset.

Figure 3.

A1–C3: In lesional areas, abundant and diffusely distributed macrophages are most prominently seen in Case 2 (B 1, B 2: anti CD68). There are also dense perivascular lymphocytic infiltrates, that are more prominent in Case 2 (B 3), than in Case 1 (A 3) or Case 3 (C 3).

Figure 4.

A1–C1: The cerebellar dentate nucleus showed severe neuronal loss, gliosis and mineralizations in Case 2 (B 1), with less involvement in Case 1 (A 1), and no involvement in Case 3 (C 1). A2–C2: In the cerebellum, there was a marked segmental depletion of Purkinje cells in Cases 1 and 2 (A 2, B 2), with the formation of some torpedoes (B 2, arrows). No cerebellar involvement was observed in Case 3 (C 2).

Figure 5.

Histological appearance of large foamy macrophages on hematoxylin and eosin‐stained sections. (A), that show dense rod‐shaped periodic‐acid‐Schiff‐positive material in the cytoplasm (B), that stains partly by silver impregnation techniques (C). Macrophages are strongly immunoreactive for CD68 (D). In affected areas, accompanying inflammatory infiltrates are predominated by CD8+ T‐cells (E), some of them in close contact with morphologically intact appearing neurons (F). In G, small granzyme B‐positive granules bearing lymphocyte is seen close to a neuron.

Case 2

This 48‐year‐old man consulted for gait impairment and dysarthria that had begun 3 months before. Neurological examination showed bilateral facial weakness with moderated right‐sided paresis and sensory loss, brisk reflexes, right Babinski sign, ataxia of right lower limb and gait ataxia. Neuropsychological tests revealed frontal and subcortical cognitive impairment. Infectious screening (HIV, hepatitis B virus, hepatitis C virus, EBV, CMV, M. pneumoniae, B. burgdorferi, T. pallidum and M. tuberculosis) was negative, and immunological analyses, cellular immunophenotyping, immunofixation (serum and urine) and hormonal studies were normal (except for low testosterone). CSF analysis revealed increased protein levels with a normal cell count. No microorganism or malignant cells were identified. The patient had no oligoclonal bands, and negative serologies and PCRs for infectious agents, John Cunningham virus and Cryptococcus neoformans antigen in the CSF. Spinal magnetic resonance imaging (MRI) was normal. Cranial MRI revealed T2 abnormalities in subcortical white matter affecting left centrum semiovale and left corona radiata extending to internal capsule, cerebral peduncles, pons, cerebellum and corpus callosum with slight heterogeneous gadolinium enhancement of cerebellar and internal capsule lesions (Figure 7). High doses of steroids were administered with no improvement. The patient's condition worsened over the following months and he developed progressive cognitive impairment, stuttering, echolalia and tetraparesis, and cerebellar and extrapyramidal dysfunction. One year after the first evaluation, palatal myoclonus and oculofacial myorhythmia appeared. A further brain and spinal MRI were performed showing an increase in size of the supra and infratentorial brain lesions and appearance of lesions affecting the lateral columns of the cervical and thoracic cord bilaterally. The diagnosis of WD was suspected and a CSF PCR for T. whipplei was positive. Although treatment with ceftriaxone and trimethoprim‐sulfamethoxazole was started, there was a progressive clinical and brain MRI deterioration. Despite antibiotic therapy for 3 months, the patient died 18 months after clinical onset due to aspiration pneumonia.

Figure 7.

Case 2 magnetic resonance imaging (MRI) findings: Brain MRI showing extensive hyperintense lesions on fluid‐attenuated inversion recovery sequences in brainstem, cerebellar peduncles, internal capsule, left corona radiata and corpus callosum.

Neuropathological examination

Unfixed brain weight was 1410 g. On gross examination, diffuse softening of brain parenchyma was detected, predominantly involving brainstem and cerebellum. On coronal sections, there were multiple small ill‐defined necrotic lesions with irregular borders involving internal capsule, putamen, pallidum, thalamus and occipital white matter (Figure 2B, arrows). No brainstem lesions were detected macroscopically. Histologically, multiple, partly necrotic inflammatory lesions were detected mainly in the thalamus and brainstem, but also in the upper cervical spinal cord, with severe involvement of cerebellar white matter and dentate nucleus, which showed abundant mineralizations (Figure 4B1). Segmental Purkinje cell loss and torpedoes were also detected (Figure 4B2, arrows). Parieto‐occipital white matter showed also extensive lesions. Lesions were characterized by dense macrophage infiltration of parenchyma and perivascular spaces, where they were accompanied by dense inflammatory infiltrates in some areas (Figure 3B1–B3). These were composed of both B‐ and T‐cells, whereas in the parenchyma CD8‐positive T‐cells and macrophages predominated. The surrounding parenchyma showed abundant large reactive astrocytes and activated microglia. Macrophages contained PAS‐positive material, which was also Gram positive and argyrophilic (Figure 5A–C). In apparently unaffected brain areas, prominent perivascular macrophage accumulation was also seen throughout the brain (Figure 6B1,B2). PCR from frozen brain tissue was positive for T. whipplei, while immunohistochemistry for this agent was negative. Three years later, a second PCR from frozen tissue performed in a different laboratory gave a negative result. No toxoplasma or herpes virus antigens were detected by immunohistochemistry. No neurodegeneration‐related abnormal protein aggregates such as beta‐amyloid, tau, alpha‐synuclein or TDP43 were detected.

Figure 6.

A1–C1: In perilesional areas, perivascular macrophages are observed in Case 1 (A1) and 2 (B1), but are nearly absent in Case 3 (C1). A2–C2: In distant, non‐lesional areas, perivascular macrophages are still evident in Case 2 (B2), but not in Case 1 (A2) or Case 2 (C2).

Case 3

An 80‐year‐old retired engineer consulted a physician for gait ataxia and developed rapidly progressive cognitive impairment within days. On hospital admission, he presented fluctuating cognitive impairment (minimental state examination—oscillated within hours from normal to a score of 20 out of 30) with normal consciousness and fluctuating right hemispheric signs like anosognosia and left‐sided neglect. Generalized myoclonus in upper and lower limbs causing a marked gait ataxia was present. The rest of the neurological examination was normal. Blood profile, vitamin levels, hormonal study (including antiperoxidase‐antimicrosoma antibodies), infectious screening (blood and CSF) and the immunological profile (antinuclear antibodies, rheumatoid factor) were unremarkable. A test for onconeuronal antibodies was negative. CSF revealed no cells and had normal glucose and protein levels including 14.3.3 protein. Brain MRI showed no focal lesions and no grey matter hyperintensities. The electroencephalogram ruled out epileptic seizures, but displayed a diffuse low‐frequency low‐voltage activity suggestive of global encephalopathy with no periodic sharp wave complexes.

Treatment with high‐dose prednisone was initiated. One week after admission the patient developed a supranuclear palsy of eye movements in addition to a tendency to hypothermia, with a core temperature ranging from 35.2 to 35.5°C. After corticosteroids, myoclonus slightly improved, but the clinical fluctuations worsened and a left‐sided paralysis appeared. A new brain MRI showed non‐enhancing high T2/fluid‐attenuated inversion recovery white matter signal hyperintensities in the right hemisphere (Figure 8A,B), that were not present in the initial MRI performed 2 weeks earlier. PCR for T. whipplei in CSF gave a positive result. Treatment with high‐dose ceftriaxone and cotrimoxazole was initiated with a gradual improvement. Four months later, the patient's cognition recovered completely and he could walk independently. A follow‐up PCR for T. whipplei in CSF performed 6 months after the hospital discharge was negative. MRI showed residual cystic lesions (Figure 8C,D) in the right periventricular white matter and corona radiata. Treatment with cotrimoxazole was maintained during the following 18 months and the subject lived an autonomous life. He died 24 months after the clinical onset due to a heart attack.

Figure 8.

Case 3 magnetic resonance imaging (MRI) evolution. A and B: Hyperintense lesions on fluid‐attenuated inversion recovery MRI sequences performed 2 weeks after clinical onset and clinical worsening. C and D: MRI performed 6 months later showing residual cystic lesions in right periventricular white matter/corona radiata.

Neuropathological examination

Unfixed brain weight was 1305 g. Macroscopically, a 1‐cm large cystic lesion was detected in the right frontal pole and small confluent lesions at the level of the right amygdala (Figure 2C). Histologically, there were several well‐delineated necrotic areas in frontal white matter extending to the overlying cerebral cortex and corpus callosum with surrounding reactive gliosis and few foamy macrophages (Figure 3C1–C3) containing moderate amounts of finely dotted, but not rod‐like PAS‐positive material. Another partly cystic lesion with macrophages was observed in internal pallidum and in parietal white matter without inflammatory reaction. No rod‐like PAS‐positive material and no perivascular macrophages or inflammatory infiltrates were observed. Only isolated perivascular lymphocytes were seen in lesional areas. No cerebellar involvement was detected (Figure 4C1,C2). Immunohistochemistry and PCR (in two occasions and two different laboratories) from frozen brain tissue were both negative for T. whipplei. There were additional moderate neurodegenerative changes of the type of argyrophilic grain disease associated with moderate neurofibrillary pathology (Braak stage III) and a high density of neuritic plaques [the Consortium to Establish a Registry for Alzheimer's Disease (CERAD) age‐related plaque score C]. No alpha‐synuclein or TDP43 protein aggregates were observed.

Discussion

We report the clinical and neuropathological features of three patients with clinically isolated CNS‐WD. All three had a distinct clinical presentation, outcome and neuropathological findings. The clinical picture of WD of the CNS is highly heterogeneous, making the clinical diagnosis very challenging in the absence of highly characteristic clinical signatures like oculomasticatory myorhythmia or supranuclear gaze palsy, especially in patients who do not have systemic symptoms 6.

One of the patients died with the clinical diagnosis of acute disseminated encephalomyelitis, the second was diagnosed very late as CNS‐WD, and the third was diagnosed early as CNS‐WD, had a good response to treatment and died more than 18 months later because of an unrelated cause. To the best of our knowledge, neuropathological features of a clinically cured CNS‐WD have not been reported so far. The hints for clinical suspicion of a CNS‐WD in Case 2 was the onset of oculomasticatory myorhythmia, in Case 3 the presence of fluctuating cognitive impairment, supranuclear gaze palsy and central hypothermia, whereas in Case 1 no diagnosis was made during life, especially because isolated spinal involvement is rare in CNS‐WD. Only few cases of CNS‐WD presenting as spinal lesions have been described so far 4, 5, 6, 13, 21. All of them had in common an isolated large cervical hyperintense lesion, although with a more chronic course, lack of long‐lasting response to steroids but marked improvement with antibiotic treatment.

The classical neuroradiological findings of CNS‐WD are diverse and range from normal to high midline T2 signal intensity involving the hypothalamus, brainstem, or mesial temporal lobes with minimal enhancement and no restricted diffusion 3, 6. In addition, high T2 signal intensity alterations in a peripheral disposition in white matter preserving midline diencephalic structures have also been reported 6, 15. Recently, CNS‐WD with high signal intensity in corticospinal tracts has been described 3. As in Case 1 and late clinical stage of Case 2 of the present article, a recent article describing in detail neurological, neuroimaging and laboratory features of 18 subjects with CNS‐WD found clinical spinal cord involvement and pachymeningitis in two of them 6; thus, CNS‐WD, although infrequent, must be taken into account when evaluating spinal cord inflammatory lesions.

Current diagnostic recommendations for WD 20 require two out of three positive tests (PAS staining, PCR, immunohistochemistry) from intestinal samples or clinically affected systems. None of the three patients have had clinically apparent extra‐CNS involvement, although a general autopsy was not performed.

Patients 1 and 2 of the present description fulfil these recommendations: Case 1 had a positive PAS staining and two different PCR determinations from post‐mortem brain tissue, and Case 2 a positive PCR in CSF during life, PAS staining and a positive PCR from post‐mortem brain tissue. In contrast, Case 3 who had a positive PCR in CSF in the context of highly suggestive clinical symptoms, and without clinically apparent neurological disease at the time of death had two negative PCR and immunohistochemistry on post‐mortem brain tissue, suggesting a good response to an adequate treatment. Although we cannot rule out a false positive result in the latter patient, as it has been previously reported in subjects with other CNS diseases or in healthy individuals (1% to −30% depending on the sample and PCR technique used) 1, 8, 10, 18, the clinical improvement associated with long‐term appropriate antibiotic therapy makes this possibility unlikely. The most probable explanation for the posttreatment in vivo and the two post‐mortem negative PCR results is to consider that the patient was cured in life by antibiotics without disease reactivation prior to death.

The discrepancy observed in Case 2 between the initial positive PCR in post‐mortem brain tissue and a posterior negative result could be related to differences in test methodology between the two laboratories.

Classically, the distribution of pathology has been reported to be widespread and to involve cortical areas and subpial regions, basal ganglia, hypothalamic nuclei, periaquaeductal grey matter, brainstem nuclei and dentate nucleus of cerebellum 25. The size of the lesions may vary from multiple microscopic foci to large confluent lesions. Romanul et al described three types of lesions in the same patient: one consisting of massive aggregates macrophages, that stained finely with PAS at the periphery and coarse and very intense in the center, probably representing membrane remnants of degraded bacteria, along with perivascular lymphocytes in few lesions and occasional in apparently unaffected brain areas 19. The second type of lesion involved thalamus and the tegmentum of the brainstem and was seen only microscopically as small microglial nodules containing finely stippled PAS‐positive structures and Gram‐positive bacteria. The third type was that of a glial scar and was considered to represent a spontaneously or treatment‐induced healed lesion. We had the opportunity to compare the lesions of three different cases: one non‐treated (Case 1) and two treated subjects with CNS‐WD, one unsuccessfully (Case 2) and the other with a very good outcome (Case 3), and we have also observed a gradient of severity of lesions.

The first, untreated case with a disease duration of 4 months showed a complete necrosis of spinal cord and multiple, non‐necrotic, moderately inflammatory foci in basal ganglia, diencephalon and brainstem, with abundant PAS‐positive macrophages. In patient 2, with a disease duration of 18 months and treated for 3 months, the most severe and active inflammatory reaction was dominated by parenchymal CD8‐positive T‐cells and perivascular B‐cells, and abundant parenchymal and perivascular macrophages in thalamus and midbrain, in medulla oblongata and cerebellum with severe degeneration of inferior olives and dentate nucleus with mineralizations. In contrast, in the successfully treated case (patient 3), only cavitary residual lesions devoid of microorganism, with no inflammation and no perivascular macrophages were detected. In conclusion, WD can be considered in the differential diagnosis of diencephalic, brainstem, cerebellar and spinal necrotizing, macrophage and T‐cell dominated encephalitic processes with marked involvement of inferior olives and dentate nucleus. Even if the clinical picture may be rather non‐specific, Case 3 emphasizes the importance of an early diagnosis in CNS‐WD in order to ensure a good outcome. Successfully treated WD may leave no sequelae or residual cystic lesions devoid of microorganisms.