Differentiation between essential thrombocythemia and polycythemia vera with marked thrombocytosis.

Accurate distinction between essential thrombocythemia and thrombocytotic polycythemia vera requires determination of the red cell mass in the presence of adequate iron stores, but this is not always possible. We therefore compared the clinical and laboratory features at the time of presentation of 50 patients with unequivocal essential thrombocythemia and 27 patients with thrombocytotic polycythemia vera. Univariate analysis failed to identify any single parameter capable of reliably separating the groups. A logistic regression algorithm incorporating hematocrit, white cell count, and spleen size markedly increased the diagnostic accuracy (92%) compared with predictions based on the hematocrit alone (52%). The algorithm's usefulness for patients with intermediate hematocrits was confirmed by analysis of independent samples of essential thrombocythemia and thrombocytotic polycythemia vera patients, and also by analysis of patients with probable essential thrombocythemia in whom the diagnosis could not be confirmed because of inadequate exclusion of polycythemia vera. Furthermore, comparison of survival data suggests that differentiating these disorders is prognostically important. The algorithm is recommended as an alternate method for differentiating essential thrombocythemia from thrombocytotic polycythemia vera whenever the red cell mass is unavailable or iron deficiency cannot be excluded.