Long-term follow-up for up to 5 years on the risk of leukaemic progression in thrombocytopenic patients with lower-risk myelodysplastic syndromes treated with romiplostim or placebo in a randomised double-blind trial

Background: Treatment options for thrombocytopenia in myelodysplastic syndromes are scarce. As described previously in a randomised phase 2 study (n=250), 58 weeks of romiplostim treatment in patients with International Prognostic Scoring System (IPSS)-defined lower-risk (low-risk or intermediate-1 risk) myelodysplastic syndromes led to reduced platelet transfusions (p<0·0001) and increased International Working Group-defined haematological improvement-platelet rates (p<0·0001) versus placebo. However, the study drug was discontinued because of the potential risk for progression to or incorrect diagnosis or treatment for acute myeloid leukaemia, based on an acute myeloid leukaemia interim hazard ratio (HR) of 2·5; the subsequent 58-week acute myeloid leukaemia HR was 1·2 (95% CI 0·4-3·8).

Methods: This study is a 5-year follow-up of a phase 2, multicentre, double-blind trial of romiplostim treatment in patients with lower-risk myelodysplastic syndromes. Eligible patients were recruited at 109 centres in North America, Europe, Russia, and Australia, were aged 18-90 years, and had platelets of 20 × 10⁹ per L or less with or without a history of bleeding or 50 × 10⁹ platelets per L or less with a history of bleeding. Patients were randomly assigned by interactive voice response system with stratification by baseline platelet count (≥20 × 10⁹ per L or <20 × 10⁹ per L) and IPSS risk (low or intermediate-1) to receive either placebo or 750 μg romiplostim subcutaneously once per week for 58 weeks. The primary outcomes for this long-term follow-up were survival and progression to acute myeloid leukaemia. Progression to acute myeloid leukaemia was defined as either 20% blasts or more after 4 weeks from romiplostim discontinuation; as per pathology; or by initiation of antileukaemia treatment. The primary outcome was assessed per protocol in all patients with available data. This study is registered with ClinicalTrials.gov, NCT00614523.

Findings: Patients were recruited from July 21, 2008, to Dec 16, 2010. 167 patients were assigned to receive romiplostim treatment and 83 were assigned to receive placebo. 210 (84%) of 250 patients entered the 5-year long-term follow-up (139 patients in the romiplostim group and 83 in the placebo group). At the end of follow-up, proportions of patients with acute myeloid leukaemia (20 [12%] of 167 in the romiplostim group vs nine [11%] of 83 in the placebo group; HR 1·06 [95% CI 0·48-2·33]; p=0·88) and proportions who died (93 [56%] vs 54 [54%]; HR 1·03 [0·72-1·47]; p=0·89) were not significantly different between the two groups.

Interpretation: Following the decision to stop the study drug, 5-year long-term follow-up HRs for transformation to acute myeloid leukaemia and HRs for death did not differ between patients treated with romiplostim and those treated with placebo, indicating that use of romiplostim is probably not associated with any increased risk of acute myeloid leukaemia or death, despite initial concerns.

Funding: Amgen Inc.