Efficacy and safety of luspatercept in non-transfusion-dependent β-thalassemia: long-term results from the BEYOND study

Abstract

Chronic anemia due to non-transfusion-dependent β-thalassemia (NTDT) can result in clinical morbidities, particularly with inadequate management. Luspatercept was previously shown to improve hemoglobin levels in patients with NTDT in the phase 2, randomized, double-blind, placebo-controlled BEYOND trial (ClinicalTrials.gov identifier: NCT03342404). Here, we report long-term efficacy and safety results from the final analysis of BEYOND spanning an additional 26 months (∼2.2 years) of follow-up. Median treatment duration was 202.8 weeks for luspatercept and 61.1 weeks for placebo. Overall, 94.8% and 22.4% of patients in the luspatercept and placebo arms, respectively, achieved a mean hemoglobin increase from baseline ≥1.0 g/dL during any 12-week interval, with mean durations of response of 1136.0 and 203.3 days, respectively. Patient-reported tiredness and weakness showed sustained improvement with luspatercept treatment. The most common treatment-emergent adverse events in the luspatercept group were headache (45.8% vs 20.4% with placebo), bone pain (43.8% vs 6.1%), back pain (39.6% vs 12.2%), and arthralgia (38.5% vs 16.3%). Treatment-emergent extramedullary hematopoiesis events were reported in 12 (9.0%) and 2 (4.1%) patients receiving luspatercept and placebo, respectively, although differences in treatment exposures prevented informative comparisons. Of the 4 patients receiving luspatercept who reported thromboembolic events, all had >1 risk factor. These results show that luspatercept led to a sustained increase in hemoglobin levels in patients with NTDT for up to ∼4.6 years of treatment, with a consistent safety profile and no new safety findings. Luspatercept is a valuable treatment option for patients with NTDT, addressing the need for effective long-term treatment of anemia. This trial was registered at www.clinicaltrials.gov as #NCT03342404.

Graphical abstract

Figure 1.

Patient disposition. In the luspatercept (including crossover) group and placebo group, treatment discontinuations occurred in 39 of 134 (29.1%) and 31 of 49 (63.3%) patients, respectively. The main reasons for study drug discontinuation in the luspatercept and placebo groups, respectively, were lack of efficacy (2/134 [1.5%] and 17/49 [34.7%]), withdrawal by patient (23/134 [17.2%] and 10/49 [20.4%]), and AEs (8/134 [6.0%] and 4/49 [8.2%]). AEs leading to study drug discontinuation in patients receiving luspatercept (including crossover) were arthralgia, hemolytic anemia, and lupus-like syndrome (1 patient); spinal cord compression and EMH masses (1 patient); pulmonary arterial hypertension (1 patient); myocardial infarction (1 patient); transient ischemic attack (1 patient); back pain, bone pain, and palpitations (1 patient each). In the placebo group, AEs that led to study drug discontinuation were fatigue and diffuse large B-cell lymphoma (1 patient), and hypotension, asthenia, and hepatocellular carcinoma (1 patient each). ∗As of last patient last visit date (28 November 2022).

Figure 2.

Achievement and duration of 12-week and 24-week erythroid responses. (A) Achievement of 12-week erythroid response; (B) total (cumulative) duration of 12-week erythroid response by data cutoff date; (C) achievement of 24-week erythroid response; (D) total (cumulative) duration of 24-week erythroid response by data cutoff date. Median (range) treatment durations at the primary data cutoff, longer-term data cutoff, and study completion date were 99.7 (15.0-132.1) vs 61.1 (3.0-121.9), 150.1 (15.0-185.4) vs 61.1 (3.0-138.0), and 202.8 (15.0-242.3) vs 61.1 (3.0-138.0) weeks, respectively, for luspatercept vs placebo.

Figure 3.

Duration of mean hemoglobin increase from baseline. Kaplan-Meier plot of duration of mean hemoglobin increase from baseline ≥1.0 g/dL over rolling 12-week intervals. Patients receiving placebo were assessed up to crossing over to luspatercept treatment. The luspatercept group does not include crossover patients. Hemoglobin assessment was done using the consecutive rolling 12-week intervals within the double-blind treatment period (ie, from day 2 to day 85, etc). Duration assessment started from the first day of the first rolling 12-week interval during which mean hemoglobin increase from baseline ≥1.0 g/dL was achieved and ended with the last day of the last consecutive rolling 12-week interval during which mean hemoglobin increase from baseline ≥1.0 g/dL was maintained. A patient was censored at the efficacy cutoff, defined as minimum (death date, study discontinuation date, last dose date +20).

Figure 4.

Mean change from BL in hemoglobin and median hemoglobin levels during entire treatment period by BL hemoglobin. (A) Mean change from BL in hemoglobin; (B) median (95% CI) hemoglobin levels for patients with ≥8.5 g/dL hemoglobin at BL; (C) median (95% CI) hemoglobin levels for patients with <8.5 g/dL hemoglobin at BL. BL, baseline.

Figure 5.

Observed mean change from baseline in NTDT-PRO T/W domain scores over the entire treatment period. ∗P < .05 indicates statistically significant between-group difference in mean change from baseline.