Study design and cohorts. The analytic sample size for the primary study end point was 63 patients (non–transfusion dependent [NTD], n = 31; transfusion dependent [TD], n = 32); this included all patients who received 0.6 to 1.25 mg/kg of luspatercept in either the initial or extension stage of the study. Doses <0.6 mg/kg of luspatercept were not considered efficacious. *Patients assigned to the dose-escalation cohorts received luspatercept at dose levels of 0.2 to 1.25 mg/kg; patients were assigned to the following dose levels: 0.2 mg/kg (n = 6 patients; all NTD), 0.4 mg/kg (n = 6 patients; all NTD), 0.6 mg/kg (n = 5 NTD patients; n = 1 TD patient), 0.8 mg/kg (n = 3 NTD patients; n = 3 TD patients), 1.0 mg/kg (n = 2 NTD patients; n = 4 TD patients), and 1.25 mg/kg (n = 1 NTD patient; n = 4 TD patients). †Patients assigned to the expansion cohort and the extension stage received luspatercept at dose levels of 0.8 to 1.25 mg/kg; patients were assigned to the expansion cohort after review of adverse event (AE) and efficacy data for patients in the dose-escalation cohorts. ‡Patients entering the extension with treatment interruption were those who had finished the initial stage of treatment and had completed the end-of-study visit at least 28 days before receiving the first dose in the extension. §Patients enter the 2-month follow-up after discontinuing from or completing the extension stage; because of the treatment period of 60 months, no patients have yet entered the follow-up period.