Summary Androgens are considered the treatment of choice for the anaemia of myelofibrosis with myeloid metaplasia (MMM). Good results have been reported in a few patients treated with danazol, a synthetic attenuated androgen. The long‐term efficacy and tolerability of danazol as treatment for the anaemia of MMM was evaluated in 30 patients, who received 600 mg/d, with progressive tapering to the minimum effective dose in the responders after 6 months. Complete response (CR) was defined as transfusion cessation with normal Hb and partial response (PR) as an Hb increase ≥1·5 g/dl with transfusion‐independent Hb values >10 g/dl maintained for at least 8 weeks. Median follow‐up was 20·5 months (range: 3·5–58 months). Response was achieved in 11 patients (37%), including eight CRs and three PRs. Median time to response was 5 months (range: 1–9 months). Four patients stopped responding at 6–24 months, two responders discontinued treatment because of toxicity, and five maintained response at 3·5–42 months. Pretreatment variables associated with response were lack of transfusion requirement (P= 0·001) and higher Hb at treatment start (P= 0·02). Toxicity was usually moderate, leading to treatment withdrawal in only two cases. Danazol is effective and well tolerated in a substantial proportion of MMM patients with anaemia.

Thrombosis is a frequent complication of polycythemia vera and essential thrombocythemia, but its incidence and predisposing factors in idiopathic myelofibrosis (IM) are unknown. In 18 (11.6%) of 155 patients diagnosed with IM in a single institution, 31 thrombotic events (19 arterial, 12 venous) were registered after a mean follow-up of 4.2 (s.d.: 4.5) years. In six patients, the thrombosis was simultaneous to or appeared a few months before IM diagnosis and 14 had one or more thrombotic episodes. When compared with the general population, a significant increase was observed in the incidence of venous thrombosis (odds ratio 17.5, 95% confidence interval: 10.3-31.4). At multivariate analysis, the initial variables associated with an increased risk of thrombosis were thrombocytosis (platelets 4450 Â 10 9 /l, P ¼ 0.001), presence of one cardiovascular risk factor (arterial hypertension, smoking, hypercholesterolemia, or diabetes, P ¼ 0.003), cellular phase of myelofibrosis (P ¼ 0.005), and Hb 411 g/dl (P ¼ 0.02). Considering post-diagnosis events, the 5-year thrombosis-free survival probability was 90.4% in the series, 80.6% for patients with platelets 4450 Â 10 9 /l, 82.6% for patients with one cardiovascular risk factor, and 85.1% for those in cellular phase. These results indicate an increased thrombotic risk for IM patients with hyperproliferative features and/or coexistent cardiovascular risk factors.