Effects of Luspatercept on Iron Overload and Impact on Responders to Luspatercept: Results from the BELIEVE Trial

Introduction: BELIEVE is an ongoing phase 3, randomized, double-blind, placebo-controlled study evaluating the efficacy and safety of luspatercept, a first-in-class erythroid maturation agent, in adults with β-thalassemia requiring regular RBC transfusions (ClinicalTrials.gov identifier: NCT02604433). Treatment with luspatercept significantly reduced RBC transfusion burden and was associated with a favorable safety profile in this patient (pt) population (Piga A, et al. Blood. 2019;133:1279-89).

Here we report iron-related efficacy endpoints of the BELIEVE study, including change from baseline in serum ferritin, liver iron concentration (LIC), and myocardial iron, as determined by T2*-weighted magnetic resonance imaging (MRI), and the effect of baseline iron levels on response to luspatercept.

Methods: Adults with β-thalassemia or hemoglobin (Hb) E/β-thalassemia who required regular transfusions of 6-20 RBC units in the 24 weeks (wks) prior to randomization (with no transfusion-free period ≥ 35 days during that period) were randomized 2:1 to luspatercept (starting dose level 1.0 mg/kg; titration up to 1.25 mg/kg) or placebo, administered subcutaneously every 3 wks for ≥ 48 wks. Pts in both arms could receive RBC transfusions and iron chelation therapy (ICT) throughout the study to maintain their baseline Hb levels.

Results: 332/336 randomized pts received treatment with luspatercept or placebo; 209/223 (93.7%) and 104/109 (95.4%) pts in the luspatercept and placebo arms had history of iron overload, defined as serum ferritin > 1,000 µg/L or LIC > 7 mg/g dry weight. 97.3% of luspatercept-treated and 97.2% of placebo-treated pts received ICT at baseline; 62.3%, 41.3%, and 37.2% received deferasirox, deferiprone, and deferoxamine, respectively, in the luspatercept arm, vs 57.8%, 36.7%, and 35.8% in the placebo arm. Fewer pts received ICT as combination therapy vs monotherapy (luspatercept arm: 28.6% vs 68.8%; placebo arm: 24.1% vs 72.3%, respectively). Mean baseline serum ferritin, LIC, and myocardial iron (by T2* MRI) for luspatercept vs placebo arms were 2,096.91 vs 1,845.05 μg/L, 12.04 vs 10.09 mg/g, and 33.52 vs 34.76 ms, respectively. As of Jan 7, 2019, 157 (70.4%) continue to receive luspatercept and 92 (84.4%) crossed over from the placebo arm.

During the last 12 wks of the 48-wk randomized treatment period, mean change from baseline in serum ferritin was −248.02 μg/L vs +106.62 μg/L for pts in the luspatercept and placebo arms, respectively (least squares mean [LSM] difference: −347.80; P < 0.0001). 158/224 (70.5%) and 33/112 (29.5%) pts in the luspatercept and placebo treatment arms achieved a ≥ 33% reduction in RBC transfusion burden of ≥ 2 RBC units during any 12-wk interval. Among the responders, mean change from baseline in serum ferritin was −343.22 μg/L and +47.30 μg/L for pts receiving luspatercept and placebo, respectively (LSM difference: −389.50; P = 0.0047). Of 140 luspatercept-treated pts with baseline ferritin ≥ 1,000 μg/L, 29 (20.7%) achieved post baseline ferritin of < 1,000 μg/L when assessed over the entire treatment period; results were similar in responders and non-responders.

During the treatment phase, mean change in myocardial MRI iron signal (T2*) from baseline to Wk 48 was −1.83 ms for luspatercept and +0.02 ms for placebo (LSM difference: −2.39; P = 0.0391). Over 96 wks, 6 (20%) luspatercept-treated pts with baseline myocardial iron ≤ 20 ms had post-baseline results > 20 ms; results were similar between responders and non-responders. Mean LIC changes from baseline to Wk 48 (mg/g dry weight) were +0.10 and +0.08 for luspatercept and placebo arms, respectively (LSM difference: +0.11; P = 0.8685). Over 96 wks, mean change in LIC among luspatercept-treated patients overall was −0.5 mg/g and −1.1 mg/g among responders.

When assessing impact of baseline iron loading there was a general consistency in achievement of ≥ 33% reduction in RBC transfusion with luspatercept over any 12 wks, regardless of iron loading subgroup (baseline serum ferritin, LIC, and myocardial iron) (Table).

Conclusions: Luspatercept treatment resulted in clinically meaningful reductions in serum ferritin levels. Baseline iron overload did not seem to affect response to luspatercept; treatment resulted in clinically meaningful reductions in RBC transfusion burden regardless of baseline serum ferritin level, LIC, or myocardial iron loading.

View largeDownload slide

View largeDownload slide

Close modal