Design: Double-blind randomised placebo-controlled phase 2 trial
Population: 110 adults aged ≥60 with newly diagnosed AML receiving intensive induction
Primary outcome: Overall survival
Effect: HR 1.59 (95% CI 0.65–3.89); median platelet transfusions 8 vs 10

EPAG-2015 randomised 110 patients aged 60 or over (median 70) with newly diagnosed AML at 18 French hospitals to eltrombopag or placebo, started on day 11 after intensive induction and continued until response assessment or platelet recovery.

Complete response rates were similar. Overall survival, the primary end point, did not differ (HR 1.59, with a wide interval). Eltrombopag did not increase severe adverse events and reduced the median number of platelet transfusions from 10 to 8.

Thrombopoietin receptor agonists in AML have raised concern about stimulating leukaemic cells; this trial found no safety signal when the drug started after chemotherapy washout. But a two-transfusion saving does not change outcomes.

Do not add eltrombopag to AML induction outside a trial on this evidence.
Continue standard platelet transfusion thresholds during induction.
Note that starting eltrombopag after chemotherapy washout showed no safety signal in this trial.

Why it matters

It settles a safety worry but not the case for routine use.

The statistics, in plain English

The survival hazard ratio of 1.59 has a confidence interval of 0.65 to 3.89, so the trial cannot rule out either benefit or harm. The transfusion reduction is a secondary outcome.