In its deliberations, the Hematology Disease Site Group (DSG) places particular emphasis on (a) results from published randomized controlled trials (RCTs) (where available) and (b) the recognition of a hierarchy of outcomes that should influence treatment decisions, with priority being given to therapies found to improve clinically important outcomes.
Because there was not much strong evidence to inform the question of possible risk factors for thrombosis or hemorrhage, the DSG sought consistency across study findings, regardless of design type, as an indicator of a predictive relationship. While, among prospective trials, the presence of initial thrombotic symptoms was not predictive of subsequent events, most retrospective studies found initial symptoms to be predictive. The magnitude of the platelet count at diagnosis, or during treatment, did not predict for thrombosis. Patients with a very high platelet count may be at a higher risk of bleeding overall, but the incidence of major bleeds reported in this series is low, and there is very little evidence of mortality or permanent morbidity. Age and other vascular risk factors were inconsistently predictive. Other groups, notably the Italian Society of Hematology, have recommended platelet-lowering treatment for patients over 60 years of age, for those with platelet counts over 1500x109/L, or for patients aged 40-60 with counts over 1000x109/L and with cardiovascular risk factors. However, in light of the available evidence, the DSG feels a definite group at high risk of bleeding or thrombosis cannot be identified with strong certainty, though the evidence seems to suggest initial symptoms are a predictor of subsequent thrombosis.
Several quality RCTs addressed the possible benefit of cytoreductive therapy for controlling thrombocytosis with respect to outcomes such as major thrombosis and hemorrhage, myelofibrosis, or survival. There is good evidence to show that hydroxyurea results in a reduction in the incidence of total arterial thrombosis in essential thrombocythemia (ET) when compared with anagrelide or with no treatment. However, no effect of hydroxyurea has been shown for stroke, myocardial infarction, or overall survival. In one of the RCTs of patients with ET, the biggest reduction was in the incidence of transient ischemic attacks, and, in the other, both transient ischemic attacks and digital microvascular ischemia. Anagrelide is inferior to hydroxyurea in controlling arterial thrombosis, and its efficacy in comparison to no cytoreductive therapy has not been established. It does not prolong overall survival in ET. Although venous thrombosis was reduced in the anagrelide arm of the study comparing anagrelide with acetylsalicylic acid (Aspirin) [ASA] to hydroxyurea with ASA, it is unclear whether the rate was increased by hydroxyurea or decreased by anagrelide.
There was no published evidence to show that controlling thrombocytosis with any of the agents reviewed reduces the incidence of major or minor bleeding. Serious bleeding was increased with anagrelide in the study comparing anagrelide with ASA to hydroxyurea with ASA. This is likely to have been caused by the functional inhibition of platelets by anagrelide. Similarly, there is very little evidence available on the use of agents in non-elderly patients. Two studies observed thrombosis rates greater than 20% in treated patients with long-term follow-up, showing elevated risk in this population as well.
Unlike ET, there is no randomized placebo-controlled trial of hydroxyurea in polycythemia vera (PV). The two RCTs by the French PV study group evaluated hydroxyurea in comparison to 32P and pipobroman, and observed no differences between agents in terms of thrombohemorrhagic outcomes. The DSG regards hydroxyurea as an efficacious agent in the PV population because of the biologic similarity between it and ET and because of the benefit established for hydroxyurea in the latter population.
The incidence of myelofibrosis in PV patients treated with phlebotomy alone is no different than for those who are treated with cytoreductive therapy. In addition, in randomized studies of patients with PV, hydroxyurea is not different from 32P and inferior to pipobroman, with respect to the subsequent rate of myelofibrosis. The natural history of myelofibrosis in ET is unknown. In the RCT that compared anagrelide and hydroxyurea, there was less myelofibrosis in the hydroxyurea arm. Whether hydroxyurea or anagrelide is responsible for this is not known.
With regard to the potential for cytoreductive therapies to induce transformation to acute leukemia (AL), strong data from randomized studies indicate that hydroxyurea is leukemogenic in patients with myeloproliferative disorders (MPD) when used after busulphan or in conjunction with 32P. There is some indication that hydroxyurea may be leukemogenic when used alone in myeloproliferative disorders (MPD); the Cortelazzo et al RCT found an elevated risk in the treatment group (in comparison to no-treatment controls). The leukemogenic potential of hydroxyurea and pipobroman in previously untreated younger patients with polycythemia vera, as reported in the RCT from Najean et al included in this review, are approximately equal and higher than would be anticipated in a phlebotomy-only group.
The Medical Research Council (MRC) PT1 RCT did not show any acute myeloid leukemia/myelodysplastic syndrome (AML/MDS) in patients treated with either agent alone. However, the median follow-up in this study was only just over three years, which is probably too short to exclude a leukemogenic effect. Whether or not hydroxyurea is leukemogenic in individuals without myeloproliferative disorders is unknown. Data from the Polycythemia Vera Study Group (PVSG)-01 RCT showed that both chlorambucil and 32P are leukemogenic, and anagrelide and interferon are believed to be non-leukemogenic from their mechanism of action.
The evidence shows that cytoreductive therapy carries with it significant leukemogenic risk and should not, therefore, be used unnecessarily. There are no studies confirming benefit in terms of superior rates of major thrombohemorrhagic events, myelofibrosis, or overall survival for asymptomatic patients, although there is an observed benefit of hydroxyurea in terms of the reduction of arterial thrombosis. The Italian Society of Hematology recommended hydroxyurea as first-line therapy in all patients over 60 years of age, and in patients aged 40-60 without childbearing potential and with a previous thrombotic event. In the absence of conclusive evidence of benefit, and with clear evidence in support of harms, the Hematology Disease Site Group adopts a somewhat more conservative stance and recommends that treatment without cytoreductive therapy in the asymptomatic population is reasonable.
In patients with PV, high-dose aspirin (900 mg/day) was not found to be beneficial, and data suggested the possibility for harm. Short-term follow-up data from the European Collaboration on Low-dose Aspirin in Polycythemia Vera (ECLAP) RCT showed a benefit for low-dose aspirin (100 mg/day) in reducing thrombotic events. There is little evidence to inform this issue for patients with ET (the one retrospective study reported a low event rate of 2.4 events/100 patient years in aspirin-treated patients; this rate is comparable to rates observed in studies of cytoreductive therapy-treated patients, notably the randomized Medical Research Council PT1 trial whose patients received aspirin therapy with cytoreductive therapy). Because PV and ET are similar diseases, clinical observations showing that ASA relieves the symptoms of microvascular occlusion and that low-dose aspirin therapy has a low risk of harm make it reasonable to anticipate that they would also be effective in this population.