Dr Christopher Allen's discussion

When I started thinking about this case, the words of Geoffrey Donnan (from his clinicopathological conference (CPC)discussion here) came to mind: ‘One must consider two important principles: first, the CPC is the domain of the rare—either rare presentations of common conditions or common presentations of rare conditions—and second, that treatable conditions are rarely presented’. CPCs tend to have several other common characteristics: there must be pathology to show at the end; they are generally ‘interesting’ (also known as rare) conditions; the history is usually challenging and does not give the game away; and there are usually some clinical tests to show (such as CSF and neuroimaging). I also did a little background research, which revealed that in 60 previously published CPCs the most common diagnoses were: lymphoma (4), acute disseminated encephalomyelitis (3), subacute sclerosing panencephalitis (SSPE) (3) and central nervous system (CNS) angiitis (3).

So with this in mind let us review this case. The history is of 45-year-old female nurse lecturer, living alone, with no previous medical or psychiatric history and taking no medication. She was stopped for driving erratically and noted to be confused (the breathalyser test was negative). We are told she was disorientated in time and place, unable to give a history, but denied any symptoms. Her sister reported 12 months of behavioural change during which time she started using cannabis, but other than this she was normal until 4 weeks ago. Her deterioration can therefore be considered a subacute or acute encephalopathy (or ‘delirium’).

The situation is, however, complicated by the apparent preceding behavioural change, and several questions immediately come to mind. Most importantly, was this change neurological or psychological in nature? If neurological, it may be part of the same condition as the encephalopathy and would indicate more longstanding orbitofrontal or limbic dysfunction. On the other hand, it may have had a psychological basis—we are not told about other life events and stressors that may be relevant. If the latter is the case, some of the behaviours may have been contributory to the eventual presentation. For example, we know about the cannabis use, but this gives rise to the possibility of other illicit drug use (such as cocaine). And if other aspects of her behaviour, such as her sexual behaviour, also changed, sexually transmitted diseases such as HIV and syphilis come into the frame.

Coming to the acute presentation, she has a precipitate decline over 4 weeks with worsening memory and dyspraxia. We are told about an episode of self-resolving confusion with vomiting, incontinence and leg weakness. This may have been a seizure, but she did not seek medical attention and we do not know much more about this. There are no clues as to precipitants, either for her overall deterioration or for this episode. We are told she was a modest drinker and a smoker; cannabis is the only known illicit drug she was using (although she may have been taking others); and we are not told of any relevant family history or foreign travel. Unfortunately, there are few additional clues from the examination; we are told that ‘there was no fever, no meningism and general examination was normal’. This makes acute infective meningitis or encephalitis unlikely, and (although it is not explicitly stated) we can probably infer that there was no rash, no obvious internal abnormality such as organomegaly or heart murmurs, and there was nothing to suggest a ‘smelly’ inherited metabolic disease such as maple syrup urine disease or carnitine deficiency. Neurological examination revealed disorientation with a global cognitive failure, with the presentation more in keeping with encephalopathy than dementia. Although there was no weakness, there were some long tract signs in the form of brisk reflexes and extensor plantars, suggesting either brain pyramidal tract or spinal cord involvement. Otherwise there were no abnormalities in eye movements, and no cerebellar, extrapyramidal or sensory features.

So what is the ‘bedside syndrome’? This is a case of subacute encephalopathy, with potentially relevant distinguishing features being: a possible 1-year history of preceding fronto-limbic dysfunction; new onset of cannabis use; at least one recent self-limiting episode of altered awareness (possibly representing a seizure, drug toxicity or metabolic dysfunction); and pyramidal signs but no other focal neurological features.

The ‘bedside differential diagnosis’ here is wide and includes:

• ■ genetic causes (eg, mitochondrial disease, adult onset leukoencephalopathies, inborn error of intermediate metabolism);

• ■ vascular causes (eg, vasculitis, venous sinus thrombosis, antiphospholipid syndrome, cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy (CADASIL), amyloid angiopathy, multiple emboli);

• ■ drugs (eg, cannabis, cocaine, inhaled heroin, other self-poisoning);

• ■ infections (eg, HIV, syphilis, SSPE, Lyme disease, Whipple's);

• ■ autoimmune or inflammatory causes (eg, multiple sclerosis, acute disseminated encephalomyelitis, limbic encephalitis, Hashimoto's encephalopathy)

• ■ neoplastic/paraneoplastic causes (eg, multiple metastases, primary tumour, limbic encephalitis);

• ■ metabolic causes (eg, postanoxic, B₁₂ deficiency, thiamine deficiency, porphyria);

• ■ degenerative causes (eg, variant Creutzfeldt–Jakob disease, fronto-temporal dementia).

So what do we learn from the investigations that may help us? The MRI is interesting. We are given the report of ‘diffusely abnormal supratentorial white matter high signal without associated mass effect, with patchy high signal in the corpus callosum. Normal medial temporal structures’ and two illustrative slices (figure 2). However, I have the benefit of having reviewed the complete MRI images, which clearly also show bilateral globus pallidus cavitation (figure 3). So this enables us to generate a rather more precise ‘clinicoradiological syndrome’ of subacute leukoencephalopathy with globus pallidus cavitation. Before refining our differential diagnosis, we should therefore review the causes of leukoencepahlopathy in adults, and then causes of globus pallidus cavitation.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Additional MRI images showing bilateral globus pallidus cavitation (arrowed).

There are many causes of leukoencephalopathy in adults. Genetic causes can be divided broadly into two groups, complex molecular defects and defects in intermediary metabolism. Of the complex molecule defects, metachromatic leukodystrophy and adrenoleukodystrophy are the most well recognised, but the list of described conditions is long (box 2). And the most frequently encountered defects in intermediary metabolism are the mitochondrial respiratory chain disorders (mitochondrial encephalomyopathy, lactic acidosis and stroke-like episodes (MELAS), etc) and the homocysteine remethylation defects (methylenetetrahydrofolate reductase (MTHFR) deficiency and cobalamin C disease), but again the complete list of conditions is more extensive (box 2). These are reviewed in detail in an informative review article by Sedel et al.1

Toxic causes are the next big group of conditions to consider, with therapeutic drugs, environmental agents and drugs of abuse all potentially implicated. In this scenario, postanoxic and carbon monoxide (CO) poisoning are possible environmental agents that warrant consideration, and with the history of recent cannabis use, drugs of abuse are also possible causes for her presentation. Cocaine, ecstasy and inhaled heroin have all been associated with leukoencephalopathy, although there is no definite link with cannabis.

In terms of autoimmune and inflammatory conditions, limbic encephalitis (either autoimmune or paraneoplastic) would be a potential cause for a presentation of this type, as would a vasculitis affecting the CNS. Other inflammatory conditions including multiple sclerosis, acute disseminated encephalomyelitis and acute haemorrhagic encephalomyelitis, while possible, seem much less likely, and the pattern does not really point towards sarcoid or systemic lupus erythematosus although these conditions can present in a variety of forms and so should not be entirely dismissed.

A variety of chronic infections including HIV, progressive multifocal leukoencephalopathy, SSPE, syphilis and Lyme can present with a leukoencephalopathy. Primary CNS lymphoma and paraneoplastic syndromes such as limbic encephalitis are worthy of consideration, as is gliomatosis cerebri, although the imaging is not typical for this. Metabolic causes, particularly hypoxia, are in the differential; eclampsia, hypertensive encephalopathy and cobalamin or folate deficiency could cause some of these features but do not fit so well with the overall pattern. And vascular causes seem less likely, although CADASIL should be considered, and both Binswanger's disease and amyloid angiopathy can cause a leukoencephalopathy.

So what about the globus pallidus lesions? There are a number of recognised causes, which are summarised below (box 3). Several of these are clear possibilities in this case. Of the common causes, hypoxic ischaemic injury and CO poisoning are plausible; of the less common causes, drug abuse and Leigh syndrome should both be kept in the differential. And of the rare causes, cobalamin C disease, Wilson's disease and neurosyphillis should all be considered. The other causes listed would be very unlikely in the clinical context.

So, having identified the clinicoradiological syndrome and considered the causes of leukoencephalopathy and globus pallidus cavitation can we make a differential diagnosis? At this point, my list would be as follows:

• ■ Genetic

  • ■ Leucodystrophy (metachromatic leukodystrophy, adrenoleukodystrophy)

  • ■ Mitochondrial disease

  • ■ Inborn errors of metabolism (homocysteine remethylation defects, eg, cobalamin C disease)

  • ■ Wilson's disease

• ■ Toxic

  • ■ Anoxia

  • ■ CO

  • ■ Cocaine, 3-4 methylenedioxymethamphetamine (ecstasy)

  • ■ Heroin (inhaled)

  • ■ Cannabis

• ■ Infection

  • ■ HIV

  • ■ Progressive multifocal leukoencephalopathy

  • ■ SSPE

  • ■ Neurosyphilis

• ■ Neoplastic

  • ■ Primary CNS lymphoma

  • ■ Paraneoplastic

• ■ Autoimmune

  • ■ Limbic encephalitis

  • ■ Vasculitis

• ■ Vascular

  • ■ CADASIL

But now we need to consider the other information we are given, in particular the blood tests. First, what do we learn from the normal blood tests? The normal full blood count, erythrocyte sedimentation rate, urea and electrolytes, calcium, liver function tests and C reactive protein suggest there is no systemic metabolic or inflammatory condition, and the normal B₁₂ and folate (and absence of macrocytosis) exclude some, although not all, remethylation defects. A normal ACE would go against sarcoidosis, and normal ammonia would effectively exclude a urea cycle defect (such as ornithine transcarbamylase deficiency). The normal lactate makes MELAS less likely, and normal lead excludes lead poisoning (eg, from adulterated cannabis). We know she is not hypoadrenal, due to the normal short Synacthen test, which goes against adrenoleukodystrophy (manifesting carrier of this X linked disorder), and the normal autoantibodies make lupus, systemic vasculitis and paraneoplastic causes much less likely. The negative tests for HIV, hepatitis B and C and toxoplasma effectively exclude these conditions. Normal white cell β-galactosidase excludes GM1-gangliosidosis and a normal MTTL1 gene excludes 80% of MELAS cases.

So what about the abnormal blood tests? There are a few. First, the measles immunoglobulin G (IgG) titre is elevated; however, this merely reflects past exposure or vaccination and so can be dismissed as irrelevant. Second, anticardiolipin IgM is elevated on two separate tests, although Russell viper venom time ratio and kaolin clotting time ratio are normal. I think this is also a red herring: false positive anticardiolipin tests are not uncommon (biological false positives without prior treponemal infection). A diagnosis of antiphospholipid syndrome requires at least one of the clinical and one of the laboratory criteria listed below to be met (box 4)2; our patient does not meet any of these. So I think this finding is also irrelevant.

However, the white cell arylsuphatase A (ArSA) is clearly low at 0.1 nmol/min/mgP (reference range 0.4–1.6); this is 25% of the lower limit of normal. The question here is: is this low enough to make a diagnosis of metachromatic leukodystrophy? Or is this pseudodeficiency?

Data from the Willink lab in Manchester (where our samples from Cambridge are sent) suggest that most patients with metachromatic leukodystrophy have ArSA levels of 10% of the lower limit of normal (or lower). However, confirmed metachromatic leukodystrophy cases have been documented with ArSA levels at up to 20% of the lower limit of normal. Prior to performing genetic tests for metachromatic leukodystrophy, the lab recommends confirming low ArSA levels with a repeat sample; we just have the one result here. Set against this background, ArSA pseudodeficiency, in which enzyme activity is significantly reduced but not sufficiently to cause the metachromatic leukodystrophy state, is said to occur in 0.5–2% of the population. This seems more likely to be the situation in this case. Interestingly, ArSA pseudodeficiency (10–30% of normal enzyme activity) has been reported in association with posthypoxic demyelination.3

In terms of other tests there is little extra to be learnt. The EEG shows just generalised non-specific slowing, and the ECG trace (from the EEG) looks normal. The CSF is essentially normal, with no evidence of CNS inflammation; the only abnormality is a positive 14-3-3 protein, but this is a very non-specific feature.

It is also worth considering tests which were not done but which might have been helpful. There is quite a list, including:

• ■ a drug screen (for drugs other than cannabis);

• ■ peripheral neurophysiology (for metachromatic leukodystrophy, adrenoleukodystrophy, MTHFR deficiency, cobalamin C disease);

• ■ antivoltage-gated potassium channel and antiN-methyl-D-aspartate antibodies (for limbic encephalitis);

• ■ urinary suphatide/ArSA gene (for metachromatic leukodystrophy);

• ■ very long chain fatty acids (for adrenoleukodystrophy);

• ■ leucocyte galactocerebrosidase (for Krabbe's disease);

• ■ leucocyte α-galactosidase (for Fabry's disease);

• ■ serum homocysteine/methylmalonic acid (for cystathionine B-sulphase, MTHFR deficiency, cobalamin C disease);

• ■ plasma cholestanol (for cerebrotendinous xanthomatosis);

• ■ urinary organic acids (for methylmalonic aciduria and other inborn errors of metabolism);

• ■ CSF lactate and other mitochondrial genes (for mitochondrial disorders);

• ■ caeruloplasmin and urinary copper (you should never forget Wilson's!);

• ■ urinary porphobilinogen (for acute intermittent porphyria);

• ■ specific serology for syphilis and borrelia, CSF cytology (for lymphoma);

• ■ MRI cervical spine (for cobalamin C disease);

• ■ 12-lead ECG (for long QT syndrome).

So, we can refine the list of potential diagnoses in the light of these additional investigations. I think a toxic cause, particularly postanoxia, is a distinct possibility (maybe in the context of ArSA pseudodeficiency). Unrecognised cocaine use could cause this presentation, as could several genetic conditions including mitochondrial disease and cobalamin C disease. Neurosyphillis is possible, as is limbic encephalitis, but these are less likely in the overall clinical context, and the remainder of the differential diagnosis discussed above seems increasingly unlikely.

So before making a final diagnosis, are there any final clues from the clinical progression?

We are told that on admission she started a treatment, but initially deteriorated further. Was this treatment an anti-infective agent such as antibiotics or acyclovir, or was she given steroids at this point? We do not know. We do know that after 10 days she had a further MRI and MR venogram; however, it is not clear whether this was because the diagnosis was still unclear at that stage, or whether the diagnosis which had been reached raised the suspicion of a thrombotic tendency. We are also told that 3 weeks after the admission she began improving and that another treatment was initiated; however, it is not clear whether this treatment was started prior to any improvement or if it was initiated after a spontaneous improvement had already begun. The nature of this treatment is also unknown; it could be IVIg or plasma exchange for an autoimmune condition, anticoagulants for antiphospholipid syndrome, a treatment for drug abuse (including psychotherapy), treatment for an infection (eg, neurosyphilis), a specific treatment for MTHFR deficiency, cobalamin C disease or another inborn error of metabolism or even a treatment for Wilson's disease (you should never forget Wilson's!).

We are then told that an MRI head, repeated after 3 weeks, shows improvement but with some residual white matter changes and increasing cerebral atrophy. Was the clinical and radiological improvement because of, or despite, treatment? This sort of remitting pattern could fit with delayed postanoxia, but also with an inflammatory condition or a mitochondrial disorder or inborn error of metabolism. She was discharged home from rehabilitation after 3 months, by which time her cognition was good enough for her to return to work 2 months after discharge, and by 8 months she seems to have recovered completely. This would be very much against metachromatic leukodystrophy or other leukodystrophy, and would put us in the category of treatable or spontaneously relapsing and remitting conditions. Possibilities would include: mitochondrial disorders, MTHFR deficiency or cobalamin C disease, delayed postanoxic leukoencephalopathy, an inflammatory or infective disorder, a prothrombotic condition and, just possibly, Wilson's disease.

But then, despite her improvement, we are told that 9 months later she was found dead at home. This appears to be a sudden unexpected death, and so a variety of direct causes of death should be considered, including cardiovascular (pulmonary embolus, myocardial infarction), cerebrovascular (possibly haemorrhage from anticoagulants), a cardiac arrhythmia, suicide or drug overdose, seizure, or recurrence of a metabolic encephalopathy. So finally let us take a step back and reconsider the overall course (figure 4); the presentation is of a reversible leukoencephalopathy followed by sudden death.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4

Graphical representation of the disease time course.

Taking everything into consideration, my first bet would be cobalamin C disease with ArSA pseudodeficiency as a cause for the subacute encephalopathy, with death resulting from a massive pulmonary embolus (due to the associated hyperhomocysteinaemia). A series of patients with cobalamin C disease and with similarities to this case has been published.4

However, I would like to make a couple of ‘side bets’! I think this could also be a mitochondrial disease (more typical of Leigh disease than MELAS), exacerbated by cannabis smoking, and with death caused by a recurrence of encephalopathy. Alternatively, this could be delayed posthypoxic encephalopathy (after drug overdose or cardiac arrhythmia) in a patient with ArSA pseudodeficiency, with death caused by a further overdose or arrhythmia.

Discussion

Dr Jonathon Schott: Just a comment, particularly for any trainees here. In a presentation of this sort, voltage-gated potassium channel antibodies should always be checked. While this is not the diagnosis here, an autoimmune encephalopathy is a potentially treatable cause of subacute cognitive decline that should always be considered and excluded.

Audience member: Was carbon monoxide poisoning considered on the basis of the MRI findings?

Dr Richard Davenport: Toxic causes were raised as a possibility, but as Dr Allen has described, the differential diagnosis for this clinicoradiological pattern is extensive.

Dr Chris Allen: I had dismissed the possibility of carbon monoxide poisoning as she was found driving her car, rather than with a hosepipe into the car.