Achieving clinically meaningful changes in haemoglobin levels in patients with non‐transfusion‐dependent β‐thalassaemia treated with luspatercept: A post hoc analysis of the phase 2 BEYOND trial

Abstract

In this post hoc analysis of the phase 2 BEYOND trial, the majority of patients with non‐transfusion‐dependent β‐thalassaemia achieved clinically meaningful haemoglobin levels ≥10.0 g/dL and increases from baseline ≥1.0 g/dL, thresholds associated with reduced risk of morbidity and mortality and recommended as an indication and target for treatment by international management guidelines, respectively.

To the Editor,

Ineffective erythropoiesis in β‐thalassaemia leads to chronic anaemia of varying severity. ¹ , ² Although patients with non‐transfusion‐dependent β‐thalassaemia (NTDT) are perceived to have milder forms of disease not requiring intervention, chronic anaemia in these patients is associated with significant morbidity and an increased risk of mortality, with complications increasing with age. ² , ³ , ⁴ , ⁵ Long‐term morbidity and mortality have been shown to be lower among patients with haemoglobin levels >10 g/dL. ² , ³ , ⁴ Patients with NTDT and haemoglobin levels ≤10 g/dL also have worse health‐related quality of life than the general population. ⁶ Furthermore, haemoglobin increases of 1 g/dL are strongly correlated with reduced odds of developing morbidity. ² , ³ , ⁷ Thus, treatment strategies that increase haemoglobin levels may improve outcomes for patients with NTDT; however, until recently, the therapeutic options to manage anaemia beyond sporadic red blood cell (RBC) transfusions remained sparse. ⁸ , ⁹ , ¹⁰

Luspatercept, a first‐in‐class erythroid maturation agent, demonstrated efficacy and safety in patients with NTDT in the phase 2, randomized, double‐blind, placebo‐controlled BEYOND trial (NCT03342404), where 77.1% of patients receiving luspatercept achieved ≥1.0 g/dL increase from baseline in mean haemoglobin during weeks 13–24 versus 0% receiving placebo. ¹¹ Based on these results, the European Medicines Agency approved luspatercept for the treatment of anaemia in adults with NTDT. ¹²

This post hoc analysis of the BEYOND trial evaluated the achievement of haemoglobin thresholds associated with a lower risk of morbidity and/or mortality (absolute haemoglobin ≥10.0 g/dL and haemoglobin increase ≥1.0 g/dL) across different baseline haemoglobin subgroups over the full trial period.

Detailed methods from the BEYOND trial have been reported previously. ¹¹ Briefly, patients were ≥18 years old with NTDT or non‐transfusion‐dependent haemoglobin E/β‐thalassaemia—defined as receiving ≤5 RBC units per 24 weeks and being RBC transfusion‐free >8 weeks before randomization—and baseline haemoglobin ≤10.0 g/dL. All patients participating in the BEYOND trial provided written informed consent. Patients were randomized 2:1 to luspatercept or placebo, given subcutaneously every 3 weeks. Placebo treatment was administered during the double‐blind treatment period (≥48 weeks); patients randomized to placebo discontinued placebo treatment after study unblinding. Luspatercept was started at 1.0 mg/kg with titration up to 1.25 mg/kg; dosing was reduced in the event of toxicity or if the increase in haemoglobin was ≥2.0 g/dL compared to the previous dose. Patients received best supportive care, including sporadic RBC transfusions and iron chelation therapy. ¹¹

This post hoc analysis used final BEYOND trial data (trial end date, 28 November 2022). The total proportion of patients who achieved a haemoglobin level ≥10.0 g/dL in the absence of RBC transfusions was assessed at 12‐week intervals through 240 weeks in the intention‐to‐treat (ITT) population and by baseline haemoglobin level (<7.0 g/dL, 7.0–8.5 g/dL, and >8.5 g/dL). Predictors of achieving haemoglobin ≥10.0 g/dL with luspatercept at week 144 were evaluated using a multivariable logistic regression model. Week 144 was used as it was the latest assessment time point with haemoglobin results for the majority of patients receiving long‐term luspatercept (76 patients [79.2%] completed 144 weeks of treatment). Covariates included in the initial multivariable regression model were age, sex, splenectomy status, baseline mean haemoglobin, genotype category (category 1 [β0/β0, β+/β+, β+/β0 without α‐thalassaemia] and category 3 [β0/β, β+/β with α gene duplication] genotypes versus category 2 genotypes [β0/β0, β+/β+, β+/β0 with α‐thalassaemia]) and baseline liver iron concentration (LIC) category (≤5 vs. >5 mg/g dry weight [dw]). A backward elimination variable selection procedure with a significance stay level of 0.15 was utilized to select the final model, which included baseline haemoglobin and splenectomy status. Additionally, the proportion and odds ratio (OR) of patients achieving erythroid response, defined as a mean increase in haemoglobin ≥1.0 g/dL from baseline, was evaluated during any 12‐ or 24‐week interval over the entire treatment period of the BEYOND trial by baseline haemoglobin level. Kaplan–Meier methodology was used to evaluate the time to first achievement of haemoglobin ≥10.0 g/dL and the time to first erythroid response during any 12‐week period in the ITT population and by baseline haemoglobin level. Hazard ratios (HRs) and statistical significance were based on an unstratified Cox proportional hazards model and unstratified log‐rank test, respectively. Haemoglobin measurements in these analyses were >21 days after RBC transfusions and therefore considered not to be influenced by RBC transfusions received. ¹¹

The ITT population comprised 145 patients (96 randomized to luspatercept and 49 to placebo). The median age was 40.0 years (range 18.0–71.0) and 57% were female. Most patients (73%) had a diagnosis of β‐thalassaemia (with or without co‐inheritance of α‐thalassaemia), and 27% had a diagnosis of haemoglobin E/β‐thalassaemia. The median (range) baseline haemoglobin level was 8.2 g/dL (5.3–10.1) in the luspatercept group and 8.1 g/dL (5.7–10.1) in the placebo group; 41% of patients had a mean baseline haemoglobin level >8.5 g/dL; 37%, 7.0–8.5 g/dL; and 22%, <7.0 g/dL (Table S1). Median (range) treatment duration at the end of the BEYOND trial was 202.8 weeks (15.0–242.3) for luspatercept and 61.1 weeks (3.0–138.0) for placebo.

In total, 75% (72/96) of patients receiving luspatercept achieved a haemoglobin level ≥10.0 g/dL by week 240 compared with 31% (15/49) of patients receiving placebo in the ITT population (Figure 1A). This trend favouring luspatercept was observed across baseline haemoglobin levels (Figure S1). Patients receiving luspatercept also achieved ≥10.0 g/dL haemoglobin significantly faster than patients receiving placebo: the median time to the first achievement was 7.43 weeks (95% confidence interval [CI] 5.86–16.71) in the luspatercept group and not reached (95% CI 88.86–not evaluable [NE]) in the placebo group (HR 3.41 [95% CI 1.95–5.96]; p < 0.0001; Figure 1B). As expected, patients with higher baseline haemoglobin levels achieved ≥10.0 g/dL more frequently and faster than patients with lower baseline haemoglobin (Figure S2). For patients with baseline haemoglobin <7.0 g/dL, three patients in the luspatercept group achieved haemoglobin ≥10.0 g/dL during the trial versus no patients receiving placebo. Together, these results indicate that luspatercept treatment can elicit fast, clinically meaningful increases in haemoglobin to a level associated with reduced risk of morbidity and mortality in patients with NTDT regardless of baseline haemoglobin level. ² , ³ , ⁴

FIGURE 1.

Total number of patients achieving haemoglobin ≥10.0 g/dL (A) and time to first achievement of haemoglobin level ≥10.0 g/dL (B) for patients in the ITT population. Time to response was defined as (first day of response) − (date of first dose of study drug) + 1. Responses were defined as achieving haemoglobin ≥10.0 g/dL from the date of first dose. CI, confidence interval; HR, hazard ratio; ITT, intention‐to‐treat; NE, not evaluable; NR, not reached.

The multivariable regression analysis adjusting for age, sex, splenectomy status, baseline haemoglobin, genotype category and baseline LIC showed that a history of splenectomy was associated with 82% lower odds of achieving ≥10.0 g/dL haemoglobin at week 144 (OR 0.18 [95% CI 0.037–0.908]; p = 0.0377) and that every 1.0 g/dL unit increase in baseline haemoglobin level was associated with a nearly 13‐fold increase in the odds of achieving ≥10.0 g/dL haemoglobin at week 144 (OR 12.79 [95% CI 4.04–40.47]; p < 0.0001). It is possible that patients who had undergone splenectomy may have had more severe ineffective erythropoiesis, thus less frequent achievement of a haemoglobin level ≥10.0 g/dL. Results of the univariable analysis are shown in Figure S3.

To further characterize the effect of luspatercept on improving haemoglobin levels, erythroid response, defined as a mean increase in haemoglobin ≥1.0 g/dL from baseline, was also evaluated. A greater proportion of patients treated with luspatercept than placebo achieved erythroid response over any 12‐ and 24‐week period of the BEYOND trial regardless of baseline haemoglobin level (Figure 2A,B). In patients treated with luspatercept, those with baseline haemoglobin >8.5 and 7.0–8.5 g/dL had numerically higher odds of achieving erythroid response in any 24 weeks (OR [95% CI] 1.92 [0.45–8.15] and 2.93 [0.58–14.77], respectively) than patients with baseline haemoglobin <7.0 g/dL. These data suggest that erythroid response with luspatercept is more common among patients with higher baseline haemoglobin, potentially due to luspatercept's capacity to better enhance effective erythropoiesis in patients with residual activity. Consequently, for patients with very low haemoglobin who have a lower likelihood of responding, transfusions should still be considered to achieve target haemoglobin levels.

FIGURE 2.

Total number of patients achieving haemoglobin ≥1.0 g/dL increase from baseline (by baseline haemoglobin level) during any 12‐week period (A) and 24‐week period (B), and time to first achievement of mean haemoglobin increase from baseline ≥1.0 g/dL over any 12 weeks for patients in the ITT population (C). Time to first achievement of mean haemoglobin increase from baseline ≥1.0 g/dL (any 12 weeks) was defined as (first day of response) − (date of first dose of study drug) + 1. CI, confidence interval; HR, hazard ratio; ITT, intention‐to‐treat; NE, not evaluable; NR, not reached.

When evaluated during any 12‐week period, patients receiving luspatercept achieved erythroid response significantly faster than patients receiving placebo (median [95% CI] time to first response: 0.29 weeks [NE–NE] with luspatercept and not reached [93.29–NE] with placebo; HR 12.22 [95% CI 6.42–23.26]; p < 0.0001; Figure 2C), with more than half of the patients receiving luspatercept achieving a first erythroid response after one dose. Faster responses with luspatercept compared with placebo were also observed across baseline haemoglobin subgroups (Figure S4).

In summary, the majority of patients with NTDT receiving luspatercept in the BEYOND trial achieved clinically meaningful haemoglobin levels ≥10.0 g/dL and increases from baseline ≥1.0 g/dL, thresholds associated with reduced risk of morbidity and mortality in these patients and recommended as an indication and target for treatment, respectively, by the Thalassaemia International Federation guidelines for the management of NTDT. ¹⁰ Although some patients receiving placebo experienced ≥1.0 g/dL increases from baseline, this is likely attributable to natural variation in haemoglobin levels. In contrast, a substantially higher proportion of patients receiving luspatercept than placebo achieved haemoglobin ≥10.0 g/dL, particularly among patients with baseline haemoglobin ≤8.5 g/dL who might be expected to have more difficulty achieving this level. Together, these findings support the efficacy of luspatercept to achieve clinically relevant improvements in haemoglobin irrespective of baseline anaemia severity and further characterize the benefit of luspatercept treatment for patients with NTDT, potentially improving patient outcomes.

AUTHOR CONTRIBUTIONS

A.T.T., J.B.P., A.K. and M.D.C. contributed to data acquisition. M.B.G., L.M.B., P.M.‐R., M.R., L.F.M., M.D. and K.W. accessed and verified the data. K.W. performed the statistical analysis. All authors contributed to data interpretation. A.T.T. and M.D.C. designed the BEYOND trial and A.T.T. is the chief investigator. All authors carefully reviewed the manuscript, approved the final version and had final responsibility for the decision to submit for publication.

FUNDING INFORMATION

The BEYOND trial was supported by Celgene, a Bristol‐Myers Squibb Company, in collaboration with Acceleron Pharma Inc., a wholly owned subsidiary of Merck & Co. Inc., Rahway, NJ, USA.

CONFLICT OF INTEREST STATEMENT

K.M.M. reports grants or contracts from Agios Pharmaceuticals and Pharmacosmos and consultancy fees from Agios Pharmaceuticals, Bristol Myers Squibb, CRISPR Therapeutics, Novartis, Novo Nordisk, Pharmacosmos and Vifor Pharma. A.T.T. reports consultancy fees and research support from Agios, Bristol Myers Squibb, Novo Nordisk, Pharmacosmos and Roche. J.B.P. reports consultancy fees and research funding from Silence Therapeutics; honoraria for advisory boards from Agios, Bristol Myers Squibb and Vifor. A.K. reports grants or contracts from Bristol Myers Squibb; honoraria/payment for lectures, presentations, speakers bureau, manuscript writing or educational events from Agios Pharmaceuticals, Bristol Myers Squibb, Chiesi and Vertex; receiving travel support from Bristol Myers Squibb; participation on a data safety monitoring board or an advisory board for Agios Pharmaceuticals, Bristol Myers Squibb, Vertex and Vifor Pharma. M.B.G. reports being employed by and owning stock in Bristol Myers Squibb. L.M.B. reports being employed by and owning stock in Bristol Myers Squibb. P.M.‐R. reports being employed by Bristol Myers Squibb. M.R. reports being employed by and owning stock in Bristol Myers Squibb. L.F.M. reports being employed by and owning stock in Bristol Myers Squibb. M.D. reports being employed by Bristol Myers Squibb. K.W. reports being employed by Bristol Myers Squibb. M.D.C. reports honoraria/payment for lectures, presentations, speakers bureau, manuscript writing or educational events from Agios, Bristol Myers Squibb and Sanofi; receiving travel support from Sanofi; participation on a data safety monitoring board or an advisory board for Sanofi, Silence and Vertex.

ETHICS STATEMENT

The BEYOND protocol was approved by the institutional review board/independent ethics committee at each study site.

PATIENT CONSENT STATEMENT

All patients participating in the BEYOND trial provided written informed consent.

CLINICAL TRIAL REGISTRATION

ClinicalTrials.gov Identifier: NCT03342404.

Supporting information

Table S1. Patient demographic and clinical characteristics.

Figure S1. Total number of patients achieving haemoglobin ≥10.0 g/dL at 12‐week intervals through week 240 by baseline haemoglobin level.

Figure S2. Time to first achievement of haemoglobin level ≥10 g/dL by baseline haemoglobin level.

Figure S3. Univariable regression analysis of predictors of achieving ≥10.0 g/dL haemoglobin at week 144.

Figure S4. Time to first achievement of mean haemoglobin increase from baseline ≥1.0 g/dL over any 12 weeks by baseline haemoglobin level.

DATA AVAILABILITY STATEMENT

The Bristol Myers Squibb policy on data sharing may be found at https://www.bms.com/researchers‐and‐partners/independent‐research/data‐sharing‐request‐process.html.