Background: The thrombopoietin receptor agonist eltrombopag has been shown to be safe, tolerable, and effective for adults with chronic immune thrombocytopenia. We aimed to investigate the safety and efficacy of eltrombopag for children with chronic immune thrombocytopenia.

Methods: PETIT2 was a two part, randomised, multicentre, placebo-controlled study done at 38 centres in 12 countries (Argentina, Czech Republic, Germany, Hong Kong, Israel, Italy, Russia, Spain, Taiwan, Thailand, UK, and USA). Paediatric patients aged 1–17 years who had chronic immune thrombocytopenia and platelet counts less than 30 × 10⁹ per L were randomly assigned (2:1) to receive eltrombopag or placebo. We stratified patients by age into three cohorts (patients aged 12–17 years, 6–11 years, and 1–5 years) before randomly entering them into a 13 week, double-blind period. Randomisation was done by the GlaxoSmithKline Registration and Medication Ordering System and both patients and study personnel were masked to treatment assignments. Patients who were allocated eltrombopag received tablets (except for those aged 1–5 years who received an oral suspension formulation) once per day for 13 weeks. Starting doses for patients aged 6–17 were based on bodyweight, and ethnic origin and ranged between 50 mg/day and 25 mg/day (starting dose for patients aged 1–5 years was 1·2 mg/kg/day or 0·8 mg/kg/day for east Asian patients). Patients who completed the double-blind period entered a 24 week open-label treatment period in which all patients received eltrombopag at either the starting dose (if they were formerly on placebo) or their established dose. The primary outcome was the proportion of patients achieving platelet counts of at least 50 × 10⁹ per L in the absence of rescue therapy for 6 or more weeks from weeks 5 to 12 of the double-blind period. The intention-to-treat population included in the efficacy assessment consisted of all patients who were randomly assigned to one of the treatment groups, and the safety population included all patients who received at least one dose of study drug. This trial is registered with ClinicalTrials.gov, number NCT01520909.

Findings: Beginning in March 15, 2012, 92 patients were enrolled, and the trial was completed on Jan 2, 2014. 63 patients were assigned to receive eltrombopag and 29 were assigned to receive placebo. In the double-blind period, three patients discontinued treatment because of adverse events: two patients in the eltrombopag group withdrew because of increased liver aminotransferases and one in the placebo group withdrew because of abdominal haemorrhage. 25 (40%) patients who received eltrombopag compared with one (3%) patient who received placebo achieved the primary outcome of platelet counts of at least 50 × 10⁹ per L for 6 of the last 8 weeks of the double-blind period (odds ratio 18·0, 95% CI, 2·3–140·9; p=0·0004). Responses were similar in all cohorts (eltrombopag vsplacebo: 39% vs 10% for patients aged 12–17 years, 42% vs 0% for patients aged 6–11 years, and 36%vs 0% for patients aged 1–5 years). Proportionately fewer patients who received eltrombopag (23 [37%] of 63 patients) had WHO grades 1–4 bleeding at the end of the double-blind period than did those who received placebo (16 [55%] of 29 patients); grades 2–4 bleeding were similar (three [5%] patients who received eltrombopag vs two [7%] patients who received placebo). During the 24-week open-label treatment period, 70 [80%] of 87 patients achieved platelet counts of 50 × 10⁹ per L or more at least once. Adverse events that occurred more frequently with eltrombopag than with placebo included nasopharyngitis (11 [17%] patients), rhinitis (10 [16%] patients), upper respiratory tract infection (7 [11%] patients), and cough (7 [11%] patients). Serious adverse events occurred in five (8%) patients who received eltrombopag and four (14%) who received placebo. Safety was consistent between the open-label and double-blind periods. No deaths, malignancies, or thromboses occurred during the trial.

Interpretation: Eltrombopag, which produced a sustained platelet response in 40% of patients with chronic immune thrombocytopenia, is a suitable therapeutic option for children with chronic symptomatic immune thrombocytopenia. We identified no new safety concerns and few patients discontinued treatment because of adverse events.