Cerebral borderzone infarcts after phlebotomy
P.J. K o u d s t a a l * , C . A . v a n D o n s e l a a r * a n d M. V e r m e u l e n *

Introduction
Summary
In polycythaemia vera both the red cell volume
and the plasma volume are elevated 1. A high
haematocrit may also result from a normal red
cell volume in combination with a contracted
plasma volume, a disorder referred to as relative
or spurious polycythaemia 2,3.
In patients with cerebral ischaemia and a concomitant high haematocrit, phlebotomy is advised to lower the red cell volume. A decrease of
the haematocrit will lower viscosity4 and increase the cerebral blood flow s , from which patients in the acute phase of cerebral ischaemia
might benefit 6,7. Removal of whole blood in patients with a spurious polycythaemia may, however, be hazardous, as these patients have a
decreased plasma volume. We describe the
development of borderzone infarcts after
venesection in a patient with cerebral
ischaemia.

Case report
A 68-year-old left-handed farmer was referred
to our hospital because of a 'weak sensation' in
his right arm, difficulty in counting and an
unsteady gait. The symptoms had started two
days before, but details about the mode of onset
could not be given. The patient had a long history of hypertension, which had been treated
with 100 mg hydrochlorothiazide a day. The

We describe a 68-year-old patient with
cerebral ischaemia who underwent a phlebotomy to lower an elevated haematocrit (0.61
1/1. The procedure was complicated by the
development of a total hemiplegia after two
hours. CT scanning showed two borderzone
infarcts.
We conclude that the plasma volume of patients with cerebral ischaemia and a high
haematocrit may be contracted to such an extent that venesection is hazardous. The most
appropriate treatment in these patients is expansion of the plasma volume.
Key words: borderzone infarcts, phlebotomy,
spurious polycythaemia.

past history and the family history were normal.
The patient did not smoke.
On admission the pulse rate was 78, the blood
pressure 190/110 mm Hg. On general examination there were no abnormal signs, apart from
conspicuous plethora of the face. The patient
was confused, was disorientated in time, unable
to write and to tell his date of birth. He had
right/left disorientation and finger agnosia. In
simple calculations he made several mistakes. In
the limbs there were no abnormalities.
Laboratory analyses showed: haematocrit
0.61 1/1, haemoglobin 12.2 mmol/1, erythrocyte

* Department of Neurology, UniversiO, Hospital Di]kzigt, Rotterdam, the Netherlands.
Address for correspondence and reprint requests: P.J. Koudstaal, Department of Neurology, Dijkzigt Hospital, 40 Dr.
Molewaterplein, 3015 GD Rotterdam, the Netherlands.
Accepted 8.9.86
Clin Neurol Neurosurg 1986. Vol. 88-4.

279

c

d

Fig. 1.
CT scan before (a + b) and after (c + d) phlebotomy.

mal. A CT scan was normal (Fig. la + b).
Prolonged ECG monitoring and echocardiography showed no abnormalities. Doppler ultrasound examination, performed at a later stage,
showed normal carotid arteries.
After admission the diuretic was stopped. 500
ml of whole blood was removed to correct the
high haematocrit. Within two hours after completion of the phlebotomy the patient suddenly
developed a total right-sided paralysis. Systolic

count 6.6 x 109/1, MCV 83 fl, MCH 1680 fmol,
MCHC 20.2 mmol/1, leucocyte count 8.2 x 109/1,
thrombocytes 367 x 109/1, pO 2 kPa, pCO 2 5.2
kPa, bicarbonate 23.0 mmol/1, 0 2 saturation
0.92, base excess -1.
Chest X-ray and electrocardiogram were nor280

blood pressure had dropped 30 mm Hg (190 to
160 mmHg). A C T scan was repeated after three
days and showed two borderzone infarcts (Fig.
lc + d). Fluid replacement was begun intravenously and haemodilution was thus continued until the haematocrit had decreased to
0.45. In the following weeks, the patient gradually improved. Residual deficits mainly consisted of a less skilful right hand and a slight
weakness of the right leg. Extensive investigations into the cause of the polycythaemia were
all negative. The hypertension was effectively
treated with methyldopa only. The haematocrit
remained normal (0.47-0.49) during six months
follow-up.

Discussion
We attribute the development of borderzone
infarction in this patient to the removal of 500 ml
of whole blood. These infarcts are caused by
profound systemic hypotension 8, which is
usually found to be due to myocardial infarction, pulmonary embolus, internal bleeding or
septicaemia. None of these conditions could be
demonstrated in this patient. The removal of
blood was advised because it was wrongly
assumed that the high haematocrit was caused
by primary proliferative polycythaemia. This
was a very unlikely diagnosis, however, since
the platelet count was within normal limits, as
was the white cell count. Secondary
polycythaemia was an equally unlikely diagnosis
because of the absence of renal, pulmonary and
cyanotic heart disease, and the presence of normal O: saturation. Therefore, the most probable diagnosis was spurious polycythaemia.
Since the mechanism of the original stroke of
this patient is not clearly embolic, the initial
ischaemia might also have been located in the
borderzone area. The neurological examination
on admission also fits such a localization. The
sudden onset and extent of neurological
deterioration of the patient shortly after
venesection, however, imply that the aggravation cannot be simply regarded as a maturation
of the original lesion, but that it was precipitated
by the blood-letting procedure. An alternative
explanation such as carotid occlusion could be
excluded by Doppler ultrasound carotid
evaluation.

Patients who present with cerebral ischaemia
have a higher haematocrit than control patients 9. In the weeks following the ictus the
haematocrit shows a tendency to decrease spontaneously 7. The elevated haematocrit might
result from a contracted plasma volume rather
than from an absolute increase in the red cell
mass. The cause of the diminution of the plasma
volume is unclear. It has been suggested that
hypertension, stress and the use of diuretics
might play a role 2,3. The patient described here
used diuretics and after discontinuation the
haematocrit remained within normal limits
without any additional measures for at least six
months. Thus, measurement of the plasma
volume before venesecting a patient is certainly
the safest approach. It may be difficult, however, to await the results of such a time-consuming investigation before instituting a therapy in
the acute phase of a stroke.
This case report demonstrates that the plasma
volume may be decreased to such an extent that
phlebotomy results in borderzone infarcts. To
prevent the complications of phlebotomy it is
advised to reduce the haematocrit by removal of
small quantities of blood (250 ml) at a time and
to replace an equal volume of fluid and to repeat
this procedure until the desired haematocrit is
reached 5,1~ The plasma volume, however, will
probably still remain contracted. The rationale
for lowering the haematocrit in the acute phase
of cerebral ischaemia is to improve cerebral
blood flow by a reduction of blood viscosity. In
addition, cerebral blood flow can be increased
by a rise in cardiac output 11, but this will occur to
a lesser extent if the plasma volume remains
reduced. Therefore, it seems more appropriate
to treat patients with a contracted plasma
volume by expanding the plasma volume, since
this both reduces the haematocrit and supports
the increase of the cardiac output. Such a
therapeutic regimen is currently being evaluated by a group of American investigators 12. In
the near future it will be possible to compare
their results with those of two European studies,
the Scandinavian ~3and Italian multi-centre trial,
which both employ an isovolaemic haemodilution scheme.

Acknowledgement
The authors wish to thank Mrs J. Doornbosch-Konijnfor
secretarial help.
281

References
l
2
3

4

5

6

7

GARDNERPH. Treatment of polycythemia vera (P vera).
Semin Hematol 1966; 3:220-5.
WA'F'FSEJ, LEWISSM. Spurious polycythaemia- A study of
35 patients. Scand J Haematol 1983; 31:241-7.
EMERY AC, WHITCOMB WH, FROHL1CH ED. 'Stress'
polycythemia and hypertension. J A M A 1974; 229:15962.
GROTrA J, ACKERMAN, R, CORREID J e t al. Whole blood
viscosity parameters and cerebral blood flow. Stroke
1982; 13:296-301.
THOMAS DJ, MARSHALL J, ROSS RUSSELL RW et al. Effects of
haematocrit on cerebral blood flow in man. Lancet 1977;
2:941-3.
WOOD JH, FLEISCHER AS. Observations during hypervolemic hemodilution of patients with acute focal
cerebral ischemia. J A M A 1982; 248:2999-3004.
STRAND T, ASPLUND K, ERIKSSON S e t al. A randomized
controlled trial of hemodilution therapy in acute
ischemic stroke. Stroke 1984; 15:988-9.

282

8

9

10

11

12

13

BRIERLEYJR. The neuropathological sequelae of profound hypoxia. In: Brierley JB, Meldrum BS (eds.).
Brain hypoxia. London, William Heinemann Ltd. 1971;
147-51.
EISENBERGS. Blood viscosity and fibrinogen concentration following cerebral infarction. Circulation (Suppl 2)
1966; 33-34:10-4.
KIRALY JF, FELDMANN JE, WHEBY MS. Hazards of phlebotomy in polycythemic patients with cardiovascular
disease. JAMA 1976; 236:2080-1.
MESSMER K, GORNANDT L, JESCH F e t al. Oxygen transport
and tissue oxygenation during hemodilution with dextran. Adv Exp Med Biol 1973; 37:669-80.
GROTFA JC, GREED PEITIGREW L~ ALLEN S e t al. Baseline
hemodynamic state and response to hemodilution in
patients with acute cerebral ischemia. Stroke 1985;
16:790-5.
SCANDINAVIAN STROKE STUDY GROUP. Multicenter trial of
hemodilution in ischemic stroke- background and study
protocol. Stroke 1985; 16:885-90.