A 32-year-old right-handed man presented with a two day history of gradual weakness and clumsiness of the right hand, followed by sudden weakness of the right leg. His partner described a transient episode of slurred speech accompanied by right-sided facial droop on the day before admission, of which the patient had been unaware. There was no history of any preceding infectious or other illness, and no fever or headache at presentation. Six months previously he had had a number of brief episodes of “twitching” of the right leg during which he had needed to lean on walls or furniture for support. One month before admission he had developed similar episodes involving the left hand. He had a background history of type 1 diabetes mellitus. His glycaemic control was said to have been poor in the past, and he had had an episode of diabetic ketoacidosis in 1993. Blood pressure (BP) recordings gleaned from his general practitioner records suggested gradually increasing hypertension over at least five years: 11.09.90–120/80; 16.08.94–148/70; 08.11.94–158/80; 27.09.99–160/100 mm Hg. His only regular medications were subcutaneous insulins. He was a non-smoker. There was no significant family history.

On examination at the time of admission his BP was 189/100, heart rate 80 beats per minute regular. He was apyrexial, alert and orientated. Cranial nerve examination was normal. There was a flaccid right-sided MRC grade 3–4 hemiparesis, pyramidal in distribution, and he was unable to walk. Tendon reflexes were bilaterally brisk, and plantar responses “equivocal”. Coordination and sensation were normal. The general examination was unremarkable.

THE FIRST TESTS

Non-contrast brain CT showed multifocal bilateral subcortical low attenuation. There was also low attenuation involving the anterior corpus callosum and the left caudate (fig 1).

Non-contrast MRI of the brain also showed multifocal, periventricular and subcortical white matter signal abnormality, most prominent on T2 and FLAIR sequences. This abnormality was particularly evident in the left frontal lobe and anterior corpus callosum (fig 2A and B). These areas were associated with diffusion restriction on DWI (fig 2C and D), with low signal on the ADC map (fig 2E and F). In addition, an area of low signal abnormality in the left frontal lobe, evident on gradient echo sequences, was felt to reflect microhaemorrhage (fig 2G and H).

The full blood count and coagulation screen were normal, ESR was 16. Urea and electrolytes and liver function were also normal. Blood glucose was 17.5 mmol/l, HbA1C 11.9% and cholesterol 6.1 mmol/l. The C-reactive protein (CRP) was <5. Rheumatoid factor, ANA, ENA, ANCA and anti-cardiolipin antibodies were all normal or negative.

The cerebrospinal fluid (CSF) opening pressure was 13 cm H₂O. Microscopy showed 7 white cells and 6 red cells per cumm, and no organisms. There was no growth on standard cultures. Protein was raised at 0.91 g/l. Glucose was 4.4 mmol/l versus serum glucose 7.9 mmol/l. Oligoclonal bands were not detected. Cytopathology showed small numbers of lymphocytes and occasional macrophages. There were no abnormal cells.

MANAGEMENT AND PROGRESS

He was initially treated with intravenous methylprednisolone, 1 g/day for 3 days. By day 2 there was significant functional improvement in his right hand, and by day 5 he was able to walk with a zimmer frame. But on day 7 he appeared “frustrated”, and fell while trying to transfer from bed to chair. Persistent hypertension was noted in the range 160–200/90–110 mm Hg and he was started on amlodipine. On day 9 his right arm was weaker. On the following day his right leg was weaker too and he required assistance to transfer. On day 11 he fell while attempting to transfer, and thereafter required hoisting. His speech was noted to be slurred, and he developed slight right-sided facial weakness. His speech continued to fluctuate during the following day, and his hypertension worsened with readings of up to 210/110 mm Hg. Enalapril was added.

On day 13 he developed episodes of extensor spasm involving the right arm and both legs, with trismus and tonic right sided facial grimacing. He remained fully responsive during the 1–2 min duration of these episodes, and was able to move his left arm and face voluntarily. These episodes were accompanied by severe hypertension up to 260/120 mm Hg, sinus tachycardia and profuse sweating. He was incontinent of urine during the following night, and was felt to be “confused”.

On day 15 he remained severely hypertensive, sweaty and confused. He was not consistently able to obey simple commands. There was severe spasticity of his right arm and both legs with generalised hyper-reflexia and bilateral extensor plantar responses. Later on the same day his conscious level fell, with eye opening to speech, confused sentences, and localising motor response only. He had further episodes of tonic extension, now involving all four limbs and accompanied by twitching of the right face. These became increasingly frequent, with incomplete recovery between episodes. He was transferred to the high dependency unit and the “spasms” were managed with intravenous benzodiazepines, and a further course of high dose intravenous methylprednisolone was started.

FURTHER INVESTIGATIONS

At this stage, repeat brain MRI with contrast was attempted. Although the images were grossly degraded by motion artefact, there was a suggestion of increased signal in the frontotemporal regions bilaterally. Doppler examination of the extracranial cervical arteries was normal. He was investigated for his persistent severe hypertension but there was no evidence of hyperaldosteronism. Urinary albumin was raised at 3024 mg/l, with a microalbumin/creatinine ratio of 358 mg/mmol. Urinary catecholamines showed no evidence for phaeochromocytoma. The full blood count showed a rising neutrophil leucocytosis (white cells 25.4×10⁹/l, neutrophils 23.73×10⁹/l), but was otherwise normal. Coagulation studies and liver function remained normal. There was a modest rise in urea and creatinine (12.1 mmol/l and 105 umol/l respectively) and the serum albumin had fallen to 23 g/l. CRP was 5. LDH was 465 U/l. Cryoglobulin and cryofibrinogen were not detected. HIV testing was negative. Repeat CSF microscopy showed no white cells, 15 red cells and no organisms. There was no subsequent growth on culture. CSF protein was raised at 1.11 g/l, and glucose 5.8 mmol/l. PCR for herpes simplex virus 1+2 and varicella zoster was negative. There was insufficient sample for JC virus PCR on the CSF. A routine EEG showed a low amplitude background with some scattered irregular slow waves, but no interictal epileptiform abnormalities or any evidence of ongoing seizures.

FINAL COURSE OF THE ILLNESS

In the early hours of the 16th day he deteriorated further with eye opening to pain, incomprehensible sounds and symmetrical extensor response to pain only. He remained severely hypertensive with tachycardia, tachypnoea and profuse sweating. On the following day there was no eye opening and no verbal response, and he continued to extend to pain. Oculocephalic reflexes were impaired. Phenytoin and aciclovir were started. He was transferred to the intensive care unit. A brain biopsy was planned, but repeat CT showed diffusely swollen brain with small ventricles, widespread loss of grey/white matter differentiation and a swollen brainstem (fig 3). Following discussion with his family, active treatment was withdrawn and he died later on the evening of the same day.

DISCUSSION: DR DAVID HILTON-JONES

The background seems straightforward. We have a man in his early 30s who had diabetes, and it appears that for a number of years this had not been well controlled. We also know that during the 1990s he had high blood pressure which was not being treated. My immediate thought was, what is this likely to be? What are the common things that hypertensive diabetics get? And what I thought of were: vascular disease; hypertensive encephalopathy; metabolic (both hyperglycaemia and hypoglycaemia can cause seizures); and infection (diabetes can of course be associated with a number of unusual infections). But the lingering thought in the back of my mind was, what if all this background were a red herring? What if the present problems are all entirely unrelated, and this unleashed a range of further potential diagnoses.

Going through the history, the first event appears to have been 6 months before admission when he developed intermittent “twitching” of the right leg. He then developed similar “twitching” of his left hand. These could have been focal seizures. They did not sound really typical of the “paroxysmal” attacks sometimes seen in multiple sclerosis because they were too brief, and in my experience limb movements in multiple sclerosis tend to be more gross—for example, dystonic posturing of the whole limb. Could they have been positive phenomena associated with transient ischaemic attacks (TIA), so-called limb jerking attacks? One would usually only think of this in people with severe carotid disease in the neck and precipitated by things like standing up or a hot bath, but I wondered if microvascular ischaemic disease could cause similar events. But in the end I thought these episodes were due to focal epilepsy.

The next history is from two days before admission, when the patient developed problems with his right hand of gradual onset, and so not really suggestive of a vascular cause. Then, shortly after that he developed sudden weakness of the right leg, sounding much more like an ischaemic event. One does sometimes see patients presenting with a stuttering onset of hemiparesis due to vascular disease, but it struck me as odd that these two events involved different vascular territories. We are also told that after the acute onset, the weakness continued to progress. This can be seen following stroke, and is often attributed to cerebral oedema, but it does still seem slightly odd for a vascular event. On the day before admission, there was a further episode [of right facial weakness and slurred speech] which does sound rather more like a TIA.

On admission, he had no headache and there were no signs of systemic illness and no fever, but he was significantly hypertensive. He was cognitively normal but had a right hemiparesis, and also signs suggesting bi-hemispheric disease with bilaterally brisk reflexes and “equivocal” plantars.

The unenhanced CT and MR imaging narrowed down the differential diagnosis fairly quickly. I felt that these changes could all be explained by the consequences of small vessel disease caused by hypertension and diabetes. But the changes were also consistent with any vasculitic process causing small infarcts. It also struck me that these lesions were really all in the white matter, with few if any convincing lesions in the cortex, and so we also have to think about an inflammatory process. Clearly this is not the typical picture of multiple sclerosis but perhaps it could be acute disseminated encephalomyelopathy (ADEM). The normal blood count and coagulation screen would not be in keeping with thrombotic thrombocytopenic purpura. The CRP and a series of autoantibodies were normal, giving no support to the possibility of a systemic vasculitic process. The CSF examination was generally unhelpful

The patient was then given methylprednisolone, suggesting that the team looking after him were thinking about vasculitis. The impression is that during the first five days after admission things were getting better, but then started to go downhill. He remained sufficiently hypertensive to cause concern, primary hyperaldosteronism and phaeochromocytoma were considered but then ruled out. He had a moderately raised urinary albumin—on the border of nephrotic syndrome, and we later learn that his serum albumin was low. This would suggest kidney damage, which could still be in keeping with microvascular disease as a result of diabetes and hypertension.

He was started on treatment for his hypertension, which remained refractory, and from this point on there was rapid deterioration. He started to suffer “spasms” of the sort which are reminiscent of the terminal phase of any disease in which there is acute brain swelling. They do not sound epileptic—we are told that he remained responsive throughout—and although they were said to stop with diazepam, this may have been coincidental. He deteriorated further, becoming sweaty and confused. Hypertension persisted with signs of autonomic arousal. He was started on a further course of intravenous methylprednisolone, which had no effect.

He had repeat MR imaging which, although grossly degraded by motion artefact, showed very extensive damage throughout frontal and temporal regions bilaterally. The EEG was unhelpful. He had a CT scan the day before he died, showing considerable brain swelling.

THE LIKELY DIAGNOSTIC CANDIDATES

In my mind there are three possibilities: intracranial small vessel disease compounded by hypertensive encephalopathy, ADEM, and vasculitis (primary angiitis of the central nervous system (CNS)).

Intracranial small vessel disease compounded by hypertensive encephalopathy

This could be a “common or garden” problem presenting in an unusual way—small vessel disease associated with his pre-existing diabetes and hypertension, the small vessel disease being further exacerbated by the hypertension, leading to hypertensive encephalopathy. The initial “fits” might have been due to cortical infarcts, and the acute presentation might have represented the “stuttering” onset of infarction. One feature of the history against this is the fact that when he presented he had no headache. On the other hand, the CT and MRI findings on admission would be consistent with this possibility, the ADC map is consistent with cytotoxic damage as would be expected with stroke, and the CSF changes could be consistent with infarction. But why did he become so hypertensive during his admission? I thought initially that this was hypertensive encephalopathy, but one alternative might be that the blood pressure was rising in response to raised intracranial pressure. We have, however, very little evidence of this.

The cognitive changes would, I think, fit well with the encephalopathic features of hypertensive encephalopathy. The “spasms”, according to this hypothesis, would have to have been due to very extensive cerebral damage with failure of autoregulation. In hypertensive encephalopathy the posterior part of the brain tends to be more involved than the anterior (we are all familiar with the reversible posterior encephalopathy associated with hypertension), so it is perhaps a little odd that the dramatic MRI change on the second scan was all anterior.

Cotton et al¹ described a very similar case of acute hypertensive encephalopathy with widespread small vessel disease on MRI in a diabetic patient. Their patient was also hypertensive, and poorly compliant with treatment, but presented with headache. He subsequently developed focal features with fits, cognitive impairment and eventually died, all within 14 days. The abnormalities on the images presented, although more extensive than in our case, showed a similar distribution. The patient did not have a postmortem but the authors suggested that there was pre-existing small vessel disease due to diabetes and hypertension, with fragile vessels which were further damaged by hypertensive encephalopathy.

Acute disseminated encephalomyelitis

The possibility of ADEM is suggested by the imaging appearances although it is difficult to explain the symptoms six months before admission. Many patients have a history of a prodromal illness or vaccination, or some trigger for the event, but this is only the case in 50–75% of cases, probably more frequently in children than adults. Similarly, although patients often have headache, some do not, and headache is less common in adults. ADEM would however be rather more in keeping with the paretic symptoms and the improvement in response to steroids. The imaging findings in ADEM in the literature are broadly similar to this case, with high T2 and FLAIR and variable DWI, except in most reports the ADC is normal. There is however some suggestion that the ADC, among other imaging findings, might predict the outcome of ADEM. Axer et al² described two patients—in one the ADC was normal, and in the other the ADC was reduced, implying a vasogenic versus a cytotoxic picture. The patient with the reduced ADC had a much poorer outcome. With regard to the other investigations, the CSF can be normal in ADEM, but usually there is a lymphocyte pleocytosis. Oligoclonal bands are only present in a minority of cases.

The progressive hypertension would have to be explained on the basis of brainstem involvement during the later course of his illness. The “spasms” described seem rather similar to the tonic spasms of multiple sclerosis, and I wondered if such paroxysmal symptoms could also be seen in ADEM.

Primary angiitis of the central nervous system

This usually presents with headache, followed by a combination of encephalopathic features and focal deficits. The course in some patients is subacute, or even acute, whereas in others it can be chronic and lingering. One problem is that it is not at all clear that all the patients reported to have it in the literature, actually had it—as proven histologically. Again in this case it would be difficult to explain the initial symptoms six months before admission. Lukas et al³ reported a similar patient thought to have primary angiitis of the CNS except for the presence of headache, and he went on to develop paresis, confusion and focal deficits, followed by rapid progression and death within 10 days. The MRI findings were similar, but with grey as well as white matter involvement; however, these are essentially the same as the changes seen in small vessel disease, and the CSF findings do not help us differentiate between them.

Conclusion

Overall, the most likely diagnostic candidates are ADEM and primary angiitis of the CNS, but there are no clinical features that point conclusively to either one of these. On balance the temporal course—in particular the gradual onset and progression of the right limb weakness, the absence of a prodromal illness, the absence of headache, the lack of early encephalopathic features—and the predominance of white matter lesions, with very few if any cortical lesions on MRI—all favour ADEM over primary angiitis of the CNS.

DR DAVID HILTON-JONES’ DIAGNOSIS

Acute disseminated encephalomyelitis.

PATHOLOGY: PROFESSOR JAMES IRONSIDE

Macroscopic appearances

The heart was enlarged with quite substantial concentric left ventricular hypertrophy, in keeping with the patient’s known hypertension. The lungs were markedly oedematous, with areas of consolidation in keeping with brochopneumonia. The fixed brain weighed 1450 g and showed symmetrical generalised swelling with transtentorial herniation. Dissection of the brain showed numerous central pontine haemorrhages secondary to intracranial herniation. There were multiple areas of ill-defined softening (corresponding to the areas of abnormality on the MRI scan) in the frontal and parietal cortex, central white matter, basal ganglia, thalamus and corpus callosum (fig 4) The cerebellum and spinal cord were macroscopically unremarkable.

The pathological differential diagnoses based on these macroscopic appearances are summarised in the box. The most likely is multiple infarcts—embolic, vasculitic or septic. They could also represent a multifocal tumour and in particular we considered the possibility of intravascular lymphoma. A glioma is less likely, but gliomatosis cerebri would be a remote possibility. A variety of infectious processes—fungal, bacterial, viral and protozoal—can cause multiple necrotising lesions in the brain. Finally we considered the possibility of demyelination, but felt that this was unlikely, given the grey matter involvement.

Microscopic appearances

Sections through the necrotic area in the basal ganglia confirmed infarction, and showed blood vessels with abnormal walls. Some of these vessels had small collections of cells around them (fig 5A and B). At higher power we saw necrotic brain with numerous macrophages. The vessel walls were thickened, and within the walls the nuclei of mononuclear cells were seen. A myelin stain showed myelin breakdown within the area of infarction, but the surrounding white matter did not show any evidence of demyelination.

Posterior to the area of infarction in basal ganglia we saw additional inflammatory changes involving the walls of the blood vessels which were expanded and infiltrated by small numbers of mononuclear cells. Staining for CD68 showed significant numbers of infiltrating macrophages in and around the vessels. Staining for CD3 showed a population of T cells within the vessel wall, with smaller numbers of B cells (figs 5C and D). These features are indicative of vasculitis, with no giant cells or granulomata present. If this were lymphoma we would expect to have seen a monomorphous population of neoplastic B cells within the vessel lumens. There was no evidence of a T cell lymphoma of the brain (which would be very unusual) and the lymphocytes did not show any morphological features suggestive of lymphoma.

Elsewhere in the brain the appearances were very similar. The changes were not confined to the white matter, and involved areas of cortex (fig 6A–C). Similar changes were seen in small vessels in the subpial and subarachnoid spaces. The larger blood vessels in the brain were normal, so this appears to be an inflammatory process confined to the small vessels within the brain parenchyma.

Examination of the hippocampus showed substantial mineralisation of the small capillaries, in keeping with longstanding diabetes. The thalamus showed oedema and other acute change suggestive of compression and the pons showed fresh haemorrhages consistent with transtentorial herniation (fig 6D). The microscopic appearances of the cerebellum, spinal cord, dorsal root ganglia, nerve roots and cauda equina were normal.

Additional information

• Virology and bacteriology cultures from autopsy brain were negative.

• Immunohistochemistry for bacteria, herpes simplex, varicella-zoster and toxoplasma were negative.

• Special stains for bacteria, tuberculosis and fungi were negative.

• Immunohistochemistry for monoclonal lymphoid cell population was negative and there was no evidence of increased lymphoid cell proliferation.

• No evidence of vasculitis outside the central nervous system.

Pathological diagnosis

• Primary angiitis of the central nervous system.

• Cerebral infarction.

• Intracranial herniation.

• Bronchpneumonia.

• Left ventricular hypertrophy due to hypertension.

COMMENT

Primary angiitis of the CNS is a rare vasculitic disorder of unknown pathogenesis in which the CNS is the only, or the dominantly involved organ. Its diagnosis is made difficult by the wide variety and low specificity of the presenting symptoms and signs, and the absence of a reliable non-invasive test. There have been no randomised controlled trials of treatment.

The condition was described in 1959 as a “non-infectious granulomatous angiitis with a predeliction for the nervous system”.⁴ The term “granulomatous angiitis” has fallen out of use as subsequent biopsy and postmortem evidence has revealed non-granulomatous pathology in a significant proportion of cases.⁵ Equally, the term “isolated angiitis of the central nervous system” has been called into question by the finding of clinically silent foci of vasculitis affecting organs outside the CNS in cases where there had been no clinical or laboratory evidence of organ dysfunction.⁶

The literature is relatively small, and due to a historical reliance on angiography as a “gold standard” investigation, the diagnosis in many cases was not histologically confirmed by biopsy or at postmortem. When analysed separately, clear differences are apparent in the demographic characteristics, clinical presentation and outcome between the pathologically and angiographically diagnosed subgroups.⁷ Some authors have used the term “angiopathy” or even “benign angiopathy” to suggest a distinct condition of unknown pathology and relatively benign course for patients diagnosed only by angiography.

Primary angiitis of the CNS predominantly affects men in the fourth to sixth decades. The mode of onset and clinical course are highly variable. The onset is often chronic, but can be subacute, or even acute. The classic course is of progressive, cumulative multifocal neurological dysfunction, but relapsing and remitting “MS-like” cases are well recognised.⁸ Headache is common, but by no means invariable, affecting as few as 50% of cases in some series.^7,9 The great majority of patients develop focal neurological deficits, but almost always in the setting of more diffuse problems such as impaired cognition or reduced conscious level. A wide variety of focal deficits are reported. Hemiparesis occurs most commonly, but isolated cranial neuropathies, visual disturbance, aphasia, dysarthria, apraxia, ataxia, tremor, chorea, myoclonus, tetraparesis and paraparesis, have all been described.^7,9 The deficits may look like TIAs, stroke-like or quasi-demyelinating events. Focal seizures are common. Less commonly the presentation may mimic a mass lesion, chronic meningitis or dementia. There are seldom any systemic symptoms.

Routine laboratory blood tests are usually unhelpful. The ESR may be raised, but not to the same extent as is in giant cell arteritis. The CSF is almost always abnormal, but the findings are non-specific, reflecting aseptic meningitis with a mild to moderate increase in protein and a modest lymphocyte pleocytosis.

Magnetic resonance imaging is sensitive, but non-specific. Typical abnormalities include leptomeningeal enhancement, either linear of punctuate; high T2 signal periventricular and subcortical white matter abnormalities; and multifocal infarcts.⁹ Less commonly, enhancing focal lesions may resemble a neoplasm.^10,11 The appearances most frequently seen in primary angiitis of the CNS overlap considerably with the typical findings in small vessel disease, demyelination, and amyloid angiopathy.

Angiography, despite its frequent use as the sole or primary diagnostic investigation in suspected primary angiitis of the CNS, has proved to be relatively insensitive and non-specific when assessed in biopsy-proven cases. Chu and co-workers studied 30 consecutive patients undergoing brain biopsy for suspected vasculitis. Among the 10 patients with biopsy confirmed vasculitis, two had normal angiograms, five had only subtle luminal narrowing, and only one had the “typical” appearance of severe multifocal “sausage-like” stenoses and occlusions. These severe angiographic abnormalities were in fact seen more frequently among patients whose biopsies showed hypertensive vasculopathy or atherosclerosis.⁹ Similar findings are reported elsewhere.⁵

Brain biopsy also has limited sensitivity when measured against postmortem diagnosis and may give false negative results in as many as 25% of cases, presumably because the vasculitic lesions are scattered.¹² Sensitivity may be maximised by sampling an imaging lesion, and the leptomeninges should be included in the biopsy.

The pathological findings are of inflammation affecting primarily leptomeningeal and cortical vessels. Larger vessels may be involved, but never to a greater extent than smaller vessels. This inflammation is often patchy, with skip lesions. There may be a variety of histological appearances including angioinvasive granulomatous pathology, lymphocytic infiltration, and polyarteritis nodosa-like fibrinoid necrosis (without the predeliction for vessel bifurcations). Giant cells are often present. All these appearances may coexist in the same patient.

In this case, the patient was of typical age and sex for primary angiitis of the CNS. The appearance and subsequent remission of the first symptoms six months before admission is not unexpected in the clinical course. The absence of headache is not unusual, and the evolution of a progressive multifocal neurological deficit in parallel with global cognitive decline, impairment of consciousness and finally death is characteristic of the condition. The relative normality of routine laboratory tests, and the CSF is expected, and the MR imaging abnormality affecting predominantly subcortical and periventricular white matter, with additional involvement of deep nuclei is typical, although non-specific. The pathological features of this case are also typical, although no giant cells or granulomata were present.

EDITOR’S NOTE

We are extremely grateful to the patient’s family for agreeing to a postmortem examination so that those looking after him were able to discover the cause of the neurological syndrome, for agreeing for the findings to be discussed at the Advanced Clinical Neurology Course in Edinburgh so that the participants could be better educated, and finally for agreeing that the case could be written up in Practical Neurology for the education of a much wider group of neurologists and their trainees. Primary angiitis of the CNS is a rare disorder, but maybe if more postmortems were done we might find it is not as rare as we think. It is very difficult to diagnose during life, the treatment is by no means always successful, and we need to know much more about it. Unless we ask for postmortems in an open and informative way, and notwithstanding the increased regulations surrounding this process, then we are making it much more difficult for ourselves when trying to help future patients with similar problems.

This patient was discussed at the 2006 Edinburgh Advanced Clinical Neurology Course.