Luspatercept for the treatment of anaemia in non-transfusion-dependent β-thalassaemia (BEYOND): a phase 2, randomised, double-blind, multicentre, placebo-controlled trial

Background: In patients with non-transfusion-dependent β-thalassaemia, haemoglobin concentrations lower than 10 g/dL are associated with a higher risk of morbidity, mortality, and impaired quality of life. No drugs are specifically approved for anaemia management in patients with non-transfusion-dependent β-thalassaemia, other than transfusion therapy administered infrequently in accordance with patients' needs. We assessed the efficacy and safety of luspatercept versus placebo in patients with non-transfusion-dependent β-thalassaemia.

Methods: We did a phase 2, randomised, double-blind, multicentre, placebo-controlled trial in 12 centres in six countries (Thailand [n=1], Lebanon [n=1], Greece [n=2], Italy [n=5], the UK [n=1], and the USA [n=2]). Eligible patients were aged 18 years or older, had confirmed diagnosis of β-thalassaemia or haemoglobin E/β-thalassaemia (concomitant α-globin deletion, mutation, or duplication were allowed), and a baseline haemoglobin concentration of 10·0 g/dL or lower. All patients were non-transfusion-dependent. Patients were randomly assigned (2:1) to luspatercept or placebo using an interactive response technology system and stratified by baseline haemoglobin concentration (≥8·5 g/dL vs <8·5 g/dL) and baseline Non-Transfusion-Dependent β-thalassaemia-Patient-Reported Outcome Tiredness/Weakness domain score (≥3 vs <3). All patients, study site staff, and sponsor representatives (who reviewed the data), except for designated individuals, were masked to drug assignment until the time the study was unblinded. Luspatercept or placebo was given once subcutaneously every 3 weeks for 48 weeks in the double-blind treatment period. Luspatercept was started at 1·0 mg/kg with titration up to 1·25 mg/kg, or reduction in the event of toxicity or excessive haemoglobin concentration increase. The primary endpoint was achievement of an increase from baseline of 1·0 g/dL or higher in mean haemoglobin concentration over a continuous 12-week interval during weeks 13-24, in the absence of transfusions. The primary efficacy and safety analyses were done in the intention-to-treat population. This trial is registered at ClinicalTrials.gov, NCT03342404, and is ongoing.

Findings: Between Feb 5, 2018, and Oct 14, 2019, 160 patients were screened for eligiblity, of whom 145 were randomly assigned to luspatercept (n=96) or placebo (n=49). 82 (57%) patients were female and 63 (43%) were male. 44 (30%) patients were Asian, 87 (60%) were White, and 14 (10%) identified as another race. The study met its primary endpoint: 74 (77%) of 96 patients in the luspatercept group and none in the placebo group had an increase of at least 1·0 g/dL in haemoglobin concentration (common risk difference 77·1 [95% CI 68·7-85·5]; p<0·0001). The proportion of patients with serious adverse events was lower in the luspatercept group than in the placebo group (11 [12%] vs 12 [25%]). Treatment-emergent adverse events most commonly reported with luspatercept were bone pain (35 [37%]), headache (29 [30%]), and arthralgia (28 [29%]). No thromboembolic events or deaths were reported during the study.

Interpretation: Luspatercept represents a potential treatment for adult patients with non-transfusion-dependent β-thalassaemia, for whom effective approved treatment options are scarce.

Funding: Celgene and Acceleron Pharma.