Answers

Short answers

1 Infarction in the territory of the right middle cerebral artery (confirmed by magnetic resonance imaging; see figure 1).

2 Erythrocytosis.

3 Seek evidence of chronic respiratory, cardiac, or renal disease. Investigations should include bone marrow/trephine, serum erythropoietin, JAK2 and BCR-ABL gene mutation analysis, and, in certain instances, red cell mass. In this patient the cause of the erythrocytosis was chronic alcohol misuse.

View larger version (51K): [in this window] [in a new window] [PowerPoint Slide for Teaching]   Fig 1 Infarction in the territory of the right middle cerebral artery consistent with a stroke of ischaemic origin

 
Long answers
1 Underlying neurological diagnosis
The patient presented with signs consistent with stroke, defined as a focal neurological deficit of sudden onset that lasts for more than 24 hours with no apparent cause other than that of vascular origin. Several other conditions, such as hemiplegic migraine, were excluded on the basis of the history, examination, and cerebral imaging. Magnetic resonance imaging (the preferred imaging modality in young people with a putative diagnosis of stroke) (fig 1) showed infarction in the territory of the right middle cerebral artery consistent with a stroke of ischaemic origin, which was investigated accordingly (box 1).

Box 1: Classification and pathogenesis of strokes    Ischaemic stroke Large artery atherosclerosis Cardioembolic Small vessel occlusion    Intracerebral haemorrhage Hypertensive small vessel disease Intracranial vascular malformations Amyloid angiopathy Bleeding diathesis    Subarachnoid haemorrhage Saccular aneurysm: Congenital Polycystic kidney disease Ehlers-Danlos syndrome A-V malformations

2 Underlying cause of the neurological condition
Cardiovascular disease—The patient remained normotensive throughout his illness. His electrocardiogram showed sinus tachycardia with no evidence of myocardial arrhythmia or ischaemia, his echocardiogram showed mildly sclerotic aortic valve leaflets. The extracranial carotid artery tree was normal on duplex examination.

Haematological disorders—These are known to underlie a small proportion of ischaemic strokes. Screening was negative for factor V Leiden, prothrombin gene mutation, protein C and protein S, antithrombin deficiency, and serum antiphospholipid antibodies. However, his raised haemoglobin concentration was accompanied by a concomitant increase in his packed cell volume consistent with a diagnosis of erythrocytosis.^{1 2 3} Haematological disorders which induce a thrombotic tendency contribute to overall risk of ischaemic stroke and may cause cerebral infarction in the absence of other risk factors,⁴ as in this case.

Stroke in younger people—The differential diagnosis was widened, in view of his young age, to include hyperlipidaemia, carotid dissection, homocystinuria, anticardiolipid syndrome, vasculitides, and mitochondrial cytopathy, but none of these disorders was present. Excessive alcohol intake is known to predispose to stroke, and is a major cause of ischaemic stroke in young people^{5 6 7 8 9 10}; he was questioned repeatedly about his alcohol intake and vehemently denied drinking to excess.

3 Further investigations to assess the underlying condition
Erythrocytosis is an abnormal increase in the numbers of red cells and can be relative or absolute. Relative erythrocytosis occurs as a result of haemoconcentration after fluid loss secondary to, for example, burns or dehydration, or after an alcohol related diuresis, which is mediated via inhibition of secretion of antidiuretic hormone.¹¹ In relative erythrocytosis, the red cell mass is normal. In this patient the packed cell volume exceeded 60%; an absolute erythrocytosis is invariably present at this level,¹² obviating the need to measure his red cell mass and effectively excluding a diagnosis of relative erythrocytosis caused by plasma volume contraction.

In absolute erythrocytosis the volume of red cells is disproportionately high in relation to other haematological constituents. It may be primary or secondary (box 2).

The primary form, polycythaemia vera, is a myeloproliferative disorder caused by overactive haematopoiesis. About 95% of patients with the condition harbour a mutation in the Janus Kinase 2 gene (JAK2) that allows erythroid progenitor cells to proliferate in the absence of erythropoietin.¹³ The incidence of polycythaemia vera is 5-17 per 1 000 000; the average age of onset is 60 years, with a male predominance.¹⁴ Polycythaemia vera is associated with a hypercoagulable state; almost half of affected individuals develop arterial or venous thromboses, or both; often they are recurrent.¹⁵ Overall, this condition probably underlies only a small proportion of strokes.⁴

Secondary erythrocytosis is associated with several conditions (box 2). The incidence of stroke in secondary erythrocytosis is unknown but is thought to be comparable to that of polycythaemia vera.⁴

Box 2: Classification of absolute erythrocytosis1 2 3    Primary erythrocytosis (polycythaemia vera) JAK2 positive (95%)—High packed cell volume (men >0.52, women >0.48) or red cell mass (>25% above predicted) and JAK2 mutation JAK2 negative—High packed cell volume or red cell mass, absence of JAK2 mutation, no cause of secondary erythrocytosis and palpable splenomegaly or acquired gene abnormality in haematopoietic cells, or two of: thrombocytosis, neutrophil leucocytosis, radiological evidence of splenomegaly, endogenous erythroic colonies or low serum erythropoietin    Secondary erythrocytosis Congenital Erythropoietin receptor mediated: High oxygen affinity haemoglobin 2,3-biphosphoglycerate mutase deficiency Erythropoietin receptor mutation Acquired Erythropoietin mediated: Hypoxia driven (central)—Lung parenchymal disease; vascular shunts; hypoventilation syndromes Local renal hypoxia—Renal artery stenosis; end stage renal failure; polycystic kidneys Pathological erythropoietin production: Cerebellar haemangioblastoma Hypernephroma Meningioma Exogenous erythropoietin Medicinal or illegal use of erythropoietin Medicinal or illegal use of androgen Idiopathic erythrocytosis

His blood film showed normal numbers and morphology of cellular constituents, and the bone marrow aspirate and trephine biopsy showed normocellular marrow with normal haematopoiesis of the erythrocytes, leucocytes, and platelets; he was negative for both the common JAK2 mutation (Val617Phe) and exon 12 mutation. His circulating erythropoietin concentration was 3.8 U/l (reference range 5-25 U/l). He did not, therefore, fulfil the criteria for classic polycythaemia vera or JAK2-negative polycythaemia vera, nor did he have any of the recognised causes of secondary erythrocytosis. Thus, we made an initial diagnosis of idiopathic erythrocytosis.

During his admission his haemoglobin fell to a low of 137 g/l after venesection and his platelet count returned to normal. No further increase in his haemoglobin was seen over the next year.

A unique case
Despite his vehement denials, this patient had a history of alcohol misuse, and subsequent changes in his haemoglobin concentration closely reflected changes in his alcohol consumption (fig 2), supporting a diagnosis of alcohol related erythrocytosis.

View larger version (34K): [in this window] [in a new window] [PowerPoint Slide for Teaching]   Fig 2 Changes in haemoglobin concentration and platelet count in relation to venesection and alcohol consumption

 
We first saw him in May 2006. In May 2007, his serum aspartate aminotransferase (AST) increased to 56 U/l (reference range <37 U/l) and GGT to 83 U/l (8-61 U/l), and a collateral history of alcohol misuse was eventually obtained; his liver biopsy showed features of alcohol related liver injury.

Since then his alcohol intake has fluctuated. Periods of abstinence have been associated with normal haemoglobin concentrations; periods of alcohol misuse have been accompanied by a return of erythrocytosis (fig 2). His daily alcohol intake at the time of presentation in May 2006 with stroke (arrowed) was approximately 16 units (128 g) of alcohol; his haemoglobin was 228 g/l and his platelet count was 92x10⁹/l. He remained abstinent for about eight months but over the next five months returned to drinking approximately 8 units (64 g) of alcohol daily. This was accompanied by an increase in his haemoglobin to 176 g/l, but his platelet count remained normal at 279x10⁹/l. A further two months of abstinence was followed by a drinking period, during which he consumed at least 16 units (128 g) of alcohol, culminating in a rise in his haemoglobin to 176 g/l and a fall in his platelets to 178x10⁹/l. At this point, April 2008, he developed a pulmonary embolus in the absence of peripheral venous thrombosis. He was successfully treated with anticoagulants and remains under review.

The pink elephant in the consulting room
Excessive consumption of alcohol predisposes to both haemorrhagic and ischaemic stroke.^{5 6 7 8 9 10} The relative risk of all subtypes of stroke is approximately doubled in heavy drinkers.⁶ Postulated mechanisms include raised systolic blood pressure, supraventricular arrhythmias, cerebral vasospasm, and rebound thrombocytosis.⁷ Alcohol misuse is also associated with relative erythrocytosis¹¹; cardiac and pulmonary infarction have both been reported in association with this condition.¹⁶ In this patient, alcohol misuse was associated with absolute erythrocytosis, the mechanism of which is unknown but may involve the erythropoietin signalling pathway.³

The patient did not develop features of alcohol withdrawal after admission to hospital; there were no superficial or cutaneous features suggesting alcohol misuse, nor was his liver palpable. His serum GGT was mildly raised at the time of presentation but was not accompanied by a rise in either his mean cell volume or serum AST. Thus there was little or no evidence on which to base a diagnosis of alcohol related erythrocytosis in the face of his persistent denial of alcohol misuse. The only clue, in retrospect, was his initial thrombocytopenia, which resolved spontaneously within 7-10 days of admission but which was attributed to his intracerebral thrombosis.¹⁷

In this patient the underlying cause of the stroke was alcohol related erythrocytosis. Doctors should be alert to this association and should endeavour to obtain detailed and accurate alcohol histories from individuals presenting with stroke or with erythrocytosis.

Cite this as: BMJ 2009;338:b90