The very-long-term follow-up of patients initially included in the PVSG protocols provides useful information. The excess risk of cancer after chlorambucil appears to persist for 5 years after stopping this treatment. The risk of leukaemia induced by marrow suppression (32P or chemotherapy) was marked before the 10th year, but low thereafter. Phlebotomy is unacceptable as permanent treatment because of the poor clinical tolerance and the frequency of vascular complications. This treatment is also associated with a risk of early progression towards myelofibrosis with myeloid splenomegaly. In the very long term, 15 years or more after the diagnosis, this complication is the major clinical risk, affecting almost 50% of our patients surviving at this time. The prevention of this type of complication could constitute one of the objectives of future protocols dealing with this disease.

Fifty-one cases of pure, primary erythrocytosis were identified and followed at Hôpital Saint-Louis, Paris, and compared with 350 cases of polycythemia vera (PV) observed during the same period. At the initial evaluation, these cases did not differ from PV cases with respect to age, sex ratio, degree of red cell volume increase, and clinical symptoms. They did differ by the absence of splenomegaly, granulocytosis and thrombocytosis. At a late stage of evolution only a few cases developed classical criteria of PV. From this group of apparently homogeneous cases, two subgroups evolved. Sixty percent of the cases were highly responsive to myelosuppression with 32P. The median duration of the first remission was greater than five years, the mean yearly dose of 32P was very low, and there was a low incidence of complications. The other group (40% of cases) was relatively resistant to myelosuppressive agents. The development of better methods of investigate this disorder might help in discriminating these two groups from both an etiological and pathophysiological viewpoint. The thromboembolic risk of these diseases suggests that myelosuppressive therapy should be utilized in older patients with higher risk of vascular accidents, reserving phlebotomy for younger patients and those who are shown to be resistant to 32P therapy.

Over 20 years, 58 cases of PV in young people (46 meeting the full PVSG criteria, 12 with elevated red cell volume and leucocytosis or thrombocytosis, without splenomegaly) were studied and have been followed for periods of 3-24 years. These cases represent approximately 5% of the cases of PV referred to the Department of Nuclear Medicine of St Louis Hospital during this period. They differ from older patients in the initial clinical severity, the short interval between the first symptoms and the diagnosis, frequent presentation with a life-threatening complication (two cases of hepatic vein thrombosis, six thrombotic or haemorrhagic events, six splenectomies, two abortions) and a very enlarged spleen in half the cases. However, after the initial complications, the overall survival is very long (exceeding 70%, even when including the initial complications, at 15 years). The vascular accidents occur exclusively in the phlebotomized patients, the main risk factor being the poor stability of the haematocrit. Only one acute leukaemia was observed among the 14 cases treated by radioactive phosphorus and/or alkylating chemotherapy. The most frequent late complication was evolution towards myelofibrosis. This spent phase seemed to occur earlier in patients treated by phlebotomy. On the basis of this data, we would advise the following therapeutic strategy: phlebotomies, as soon as the diagnosis is established, and a systematic long-term treatment by hydroxyurea with the hope of reducing the number of vascular complications and of delaying the evolution towards the spent phase and the myelofibrosis.

SUMMARY The site of sequestration of 51Cr‐labelled platelets has been studied in 465 subjects, of whom 317 suffered from idiopathic thrombocytopenic purpura. The validity of the method was demonstrated in several ways: a given subject usually showed the same site of platelet sequestration when investigated more than once even after a long interval; there were characteristic and very different platelet sequestration curves in the thrombocytopenias due to bone marrow hypoplasia, hypersplenism or ITP; and there was a correlation between the preoperative in vivo results and the radioactivity found in the spleen after splenectomy. In ITP the destruction of labelled platelets was more often splenic in children and in patients whose thrombocytopenia responded to steroid therapy. There was a perfect correlation between the site of platelet destruction and the platelet rise immediately after splenectomy. There was a good correlation between the site of platelet destruction and the long‐term effectiveness of splenectomy.

The analysis of 288 cases of polycythemia vera (PV) with a minimal follow-up of 10 years enabled us to study the characteristics of acute leukemia as observed in 33 patients (11.4%). In 50% of the patients (16 of 33), the malignant transformation is of the refractory anemia with excess of blasts (RAEB) type. Half of these further transform to acute nonlymphocytic leukemia (ANLL). Their life expectancy is not better than patients who abruptly develop ANLL. Leukemic transformation shows a frequency peak in the eighth year after initial evaluation in PV treated with chemotherapy and in the 11th year in patients treated with radiotherapy. In 30% of the patients myelofibrosis, or the spent phase of PV, is present before the transformation to acute leukemia (AL). This complication is, however, part of the natural history of PV and is observed in 20% of PV patients at 10 years when leukemic transformation is absent. Marrow fibrosis can therefore not be considered as a preleukemic phase. It was also noted that the occurrence of myeloid metaplasia/myelofibrosis is more frequent and begins earlier in patients treated by phlebotomy alone, and who do not transform to leukemia. The clinical characteristics of these AL, including high frequency of partial marrow invasion, difficulties in cytologic classification, a peak incidence similar to that in patients treated by chemotherapy or radiotherapy for a prior malignancy, multiple chromosome abnormalities, and poor response to therapy are all highly suggestive of secondary leukemias.