Luspatercept for the Treatment of Anemia in Non-Transfusion-Dependent β-Thalassemia: Final Safety and Efficacy Data from the BEYOND Trial

Background: Suboptimal anemia management in patients (pts) with non-transfusion-dependent (NTD) β-thalassemia may lead to serious clinical complications. Improvements in hemoglobin (Hb) levels were shown to significantly decrease the risk of developing morbidities (Musallam KM, et al. Ann Hematol 2022;101:203–204; Musallam KM, et al. Ann Hematol 2021;100:1903–1905). There is a need for therapies that increase Hb levels besides red blood cell transfusions (RBCT), which lead to secondary iron overload. Primary and longer-term data from the BEYOND study showed that luspatercept led to significant increases in Hb levels and had a manageable safety profile in pts with NTD β-thalassemia (Taher AT, et al. Lancet Haematol 2022;9:e733–e744; Taher AT, et al. Blood 2022;140[suppl 1];8210–8212). Here we report the final data (up to last patient last visit [LPLV] date Nov 28, 2022), of the phase 2, randomized, double-blind, placebo-controlled BEYOND trial (NCT03342404).

Methods: Eligible pts (N = 145) had NTD (defined as receiving ≤ 5 RBC units in the 24 wk before randomization) β-thalassemia or HbE/β-thalassemia, Hb levels ≤ 10 g/dL. Pts were randomized to luspatercept (N = 96; 1.0–1.25 mg/kg) or placebo (N = 49) subcutaneously Q3W for ≥ 48 wk. Pts continued to receive best supportive care when required. Pts who crossed over from placebo to luspatercept were not included in this analysis. Mean Hb increase ≥ 1.0 g/dL (any 12-/24-wk interval) and ≥ 1.5 g/dL from baseline (any 12-wk interval), duration of mean Hb increase ≥ 1.0 g/dL and ≥ 1.5 g/dL from baseline, iron parameters, RBCT, and safety were evaluated up to the LPLV date.

Results: As of Nov 28, 2022, 93/134 (69.4%) pts on luspatercept transitioned to the long-term follow-up study (NCT04064060): 66/96 (68.8%) pts in the luspatercept arm and 27/38 (71.1%) pts who crossed over from placebo to luspatercept after study unblinding. A total of 76/96 (79.2%) pts in the luspatercept arm completed 144 wk of treatment (tx), with 55/96 (57.3%) completing 192 wk of tx; 28/96 (29.2%) pts on luspatercept discontinued tx. The median (range) luspatercept tx duration was 202.8 (15.0–242.3) wk (Table) and 61.1 (3.0–138.0) wk with placebo.

Pts had a durable increase in Hb ≥ 1.0 g/dL from baseline (by 12-wk interval) through wk 240 of luspatercept tx (Figure). In the luspatercept arm, 91/96 (94.8%) pts had a mean Hb increase ≥ 1.0 g/dL from baseline during any 12-wk interval (Table). The mean (standard deviation [SD]) total duration of mean Hb increase ≥ 1.0 g/dL from baseline (any 12-wk interval) with luspatercept was 1136.0 (491.9) days. The proportion of pts achieving a mean Hb increase ≥ 1.5 g/dL during any 12-wk interval over the entire tx period was 80/96 (83.3%) and 4/49 (8.2%) in the luspatercept and placebo arms, respectively.

At baseline, 83/96 and 42/49 pts in the luspatercept and placebo arms, respectively, were RBCT-free. The proportion of pts who remained RBCT-free was 78.1% (75/96) and 57.1% (28/49) in the luspatercept and placebo arms, respectively. For pts who received RBCT, the mean time to first RBCT was 482.0 days with luspatercept and 202.6 days with placebo.

The mean change from baseline in serum ferritin (SF) level at wk 96 was +76.45 µg/L with luspatercept (n = 90) and +211.51 µg/L with placebo (n = 35). The mean change from baseline in SF level stabilized with longer luspatercept tx (wk 192: 49.14 µg/L [n = 71]; wk 240: 52.19 µg/L [n = 51]). The mean change from baseline in liver iron content in the luspatercept arm was −1.23 mg/g dry weight (dw) at wk 96 (n = 8) and −0.79 mg/g dw at wk 144 (n = 54).

The 10 most frequently reported adverse events (AEs; ≥ 5% of pts) with luspatercept were headache, back pain, bone pain, COVID-19, arthralgia, pyrexia, pain in extremity, fatigue, diarrhea, and hypertension. Rates of serious and grade 3–4 AEs in the luspatercept arm were 24.0% [23/96] and 34.4% [33/96]), respectively. In the luspatercept arm, thromboembolic events were reported in 2/96 pts, extramedullary hemopoietic masses in 10/96 pts (all grade 1–2), and there were no malignancies. Few AEs led to tx discontinuation in the luspatercept arm (5/96 [5.2%]); no new safety signals were identified.

Summary: The final results from the BEYOND study show that pts with NTD β-thalassemia treated with luspatercept maintained clinically meaningful and durable erythroid responses. The safety profile of luspatercept was manageable, and extended luspatercept exposure did not raise new safety concerns.