The impact of voxelotor on the reduction in red blood cell transfusions in US patients with sickle cell disease

Abstract

Abstract Background: The current treatment landscape for SCD is limited, and treatment options have associated risks and can be burdensome for patients. Red blood cell (RBC) transfusions are indicated to treat or prevent complications, like severe anemia, acute multi-organ failure, vaso-occlusive crises (VOC), acute chest syndrome (ACS), and stroke. However, every RBC transfusion administered carries risks, such as iron overload, alloimmunization, or delayed hemolytic transfusion reactions (DHTR) that can be life-threatening. Voxelotor is a treatment for SCD that inhibits the polymerization of sickle hemoglobin (Hb) and has shown clinical benefit on Hb levels, which may decrease the need for RBC transfusions. Pfizer voluntarily withdrew voxelotor from the market on September 25, 2024, due to emerging preliminary data that showed the benefit of voxelotor may no longer outweigh the risk. Objective: This observational study using claims data aimed to determine if treatment with voxelotor reduces RBC transfusion rates compared to matched controls not treated with voxelotor. Methods: This non-interventional, real-world study aimed to evaluate the effect of voxelotor on the reduction of RBC transfusion burden in patients with SCD ≥ 12 years of age in the US with at least two visits with a transfusion in the prior 365 days (baseline). The study period was November 25, 2018 - January 31, 2024. The study used the Komodo Healthcare Map (KHM) database linked to Claritas prescription data (proprietary voxelotor claims) and Quest Diagnostics lab data. The study population included patients with a new prescription claim for voxelotor (no use 365 days prior) in the exposed group, while the control (non-voxelotor) group included coarsened exact matched 1:4 (voxelotor to controls) patients who did not have a prescription claim for voxelotor. Propensity score (PS) matching using variable-ratio (1:2) greedy nearest neighbor matching with a caliper distance of 1% was used to balance baseline confounders between the two groups. Potential confounding covariates were assessed for balance using absolute standardized differences (ASD); covariates that remained imbalanced (ASD > 0.1) after the matching process were adjusted for in outcome models. Patients were followed from the index date (matched voxelotor prescription date) until disenrollment, death, discontinuation of voxelotor, or the occurrence of exclusion criteria. To account for variable follow-up for each patient, outcomes were assessed as per-patient-per-year (PPPY). Results: After PS matching, the voxelotor and non-voxelotor groups consisted of 164 and 289 patients, respectively. The mean follow-up was 288 days in the voxelotor group and 333 days in the non-voxelotor group. The mean (SD) age of patients was 33.0 (16.0) years in the voxelotor group and 31.6 (15.4) years in the non-voxelotor group. The mean (SD) number of VOCs during the baseline period was 5.5 (5.4) in the voxelotor group and 5.6 (5.7) in the non-voxelotor group. The voxelotor and non-voxelotor groups were balanced with the majority of absolute standardized difference values for the covariates ≤ 0.1. The voxelotor group had 5.3 RBC transfusions PPPY at baseline and 3.4 RBC transfusions PPPY in follow-up for a reduction of 1.9 RBC transfusions PPPY, compared to the non-voxelotor group who had 5.5 RBC transfusions PPPY in baseline and 4.9 RBC transfusions PPPY in follow-up for a reduction of 0.6 RBC transfusions PPPY. The difference in the rate change was -1.3 RBC transfusions PPPY (95% CI: -2.1, -0.6; p<0.001). In the voxelotor group, 61.6% experienced >30% reduction in RBC transfusions between baseline and follow-up compared to 48.1% of the non-voxelotor group (odds ratio: 1.8; 95% CI: 1.2, 2.6). Additionally, 34.1% of the voxelotor group had no RBC transfusions during follow-up, compared to 29.1% in the non-voxelotor group (odds ratio: 1.3; 95% CI: 0.9, 1.9). Discussion: In individuals with SCD and a history of RBC transfusions, voxelotor led to a statistically significant reduction in RBC transfusions PPPY compared to matched controls in this real-world study using medical claims data. This study suggests that transfusion rate may be improved with voxelotor and further studies are needed to confirm these findings. The impact of voxelotor on other clinically relevant endpoints, such as VOCs, still needs to be evaluated.