Inclusion Criteria:

- The proposal is to include as many patients with PT as possible including previously
diagnosed patients whether or not they have received treatment. Thus all patients are
eligible assuming they meet the diagnostic criteria and they do not have any
exclusion criterion (see below). It will be necessary to stratify patients according
to their previous treatment. This information will be collected on entry to the
study. Informed consent is of course required where there is a change of therapeutic
strategy.

The diagnostic criteria for primary thrombocythaemia are:

- Platelet count > 600x109/l.

- No evidence of overt polycythaemia(confirmed by RCM if necessary)or of polycythaemia
masked by co-existent iron deficiency.

- No Philadelphia chromosome.

- Absence of peripheral blood and/or marrow appearances of myelodysplasia, or
myelofibrosis.

- No known cause of reactive thrombocytosis. Particular care should be taken to exclude
iron deficiency in pre-menopausal women.

Notes:

- In asymptomatic patients, the platelet count should be observed for a period of at
least 2 months to confirm >600x109/l, and to allow any cause of reactive
thrombocytosis to become overt.

- If the PCV is above normal upper limit (that is, males >0.51, females >0.48) or in
high normal range in a patient with palpable splenomegaly measure RCM. Iron deficient
primary polycythaemia (polycythaemia vera) is strongly suggested if Hb/PCV is normal
in the presence of iron deficient red cell changes. In this situation, iron therapy
is potentially dangerous.

- Exceeding rarely, bcr-abl positive Philadelphia chromosome negative patients present
with high platelet counts with little or no elevation in WBC count. The features that
suggest it is necessary to examine for bcr-abl, are:- basophilia, left-shift in WBC,
granulocyte count >16x109/l, difficulty in controlling platelet count, megakaryocytes
of low ploidy (NAP is usually unhelpful).

- A normal ESR, CRP or plasma viscosity is useful in excluding a reactive
thrombocytosis.

- Written informed consent obtained in accordance with NCRI requirements.

- Patients with impaired hepatic / renal function are not excluded although the
respective biochemical tests should be monitored during therapy and reduced doses of
cytoreductive agent should be used, particularly in the case of hydroxyurea and renal
dysfunction.

Exclusion Criteria:

High risk features (any of the following):

- Age >or= 60 years

- Platelet count > or= 1500x109/l (current or previous) (a)

- History of ischaemia, thrombosis or embolic events (including erythromelalgia) (b)

- Haemorrhage considered to be related to PT (b)

- Presence of hypertension (c)or diabetes (d)

- The manufacturers of hydroxyurea state that it should be avoided in pregnancy and in
lactating women. Similarly, hydroxyurea should not be prescribed for women when
there is doubt about their use of an effective contraceptive method.

- Exclude patient from hydroxyurea therapy and, therefore, from the 'intermediate' risk
randomisation if the patient has current leg ulcers.

Notes on the definition of high risk:

- In patients with borderline counts the allocation of a patient to a high risk group
based on platelet count alone should rely on at least three samples taken on separate
occasions over at least 2 months.

- Documentation of previous thrombo-embolic, ischaemic and haemorrhagic events should
be given on the patient's entry proforma.

- Hypertension is defined as those patients requiring hypotensive therapy.

- Diabetes is defined as those patients requiring therapy with a hypoglycaemic agent.