Trends and Outcomes of Voxelotor Treatment in Pediatric Patients With Sickle Cell Disease: A Retrospective Study

Abstract

OBJECTIVE

Sickle cell disease (SCD) is characterized by hemoglobin S polymerization, leading to anemia, pain, and organ damage. Voxelotor is a therapy for SCD that was withdrawn from worldwide markets in September 2024. We evaluated trends in voxelotor use and its impact on pediatric SCD patients before withdrawal.

METHODS

A retrospective study was conducted using patient records from 4 Children’s Healthcare of Atlanta (CHOA) hospitals. Data on patient demographics, length of time on voxelotor, clinical markers, hospital usage, side effects, and discontinuation reasons were analyzed.

RESULTS

Two hundred forty-five patients were taking voxelotor between 2019 and 2024 outside of clinical trials. Of these, 188 (77%) patients were on voxelotor for at least 6 consecutive months, with an average of 25.0 months on treatment. The main reason for discontinuation was due to a side effect (47%) for patients on treatment for less than 6 months and market withdrawal (69%) for those taking 6 months or more. On average, patients had a significant increase in mean hemoglobin (0.5 g/dL, p < 0.001), a decrease in hemolysis markers, and a mean decrease in annualized emergency department visits. At the time of market withdrawal, 123 patients were actively taking voxelotor.

CONCLUSIONS

Overall, patients had a positive hematologic and clinical response to voxelotor, with the majority remaining on treatment for at least 6 consecutive months. Despite market withdrawal due to safety and efficacy concerns, a subset of patients derived clinical benefits from voxelotor, highlighting the need for continued development of effective therapeutics for SCD.

Introduction

Sickle cell disease (SCD) is an inherited blood disorder that affects approximately 100,000 people in the United States.¹ SCD can lead to acute and chronic complications that can contribute to stroke, organ dysfunction, and mortality. SCD is characterized by the polymerization of deoxygenated sickle hemoglobin, which causes erythrocytes to form a sickle shape that damages the cell membrane. These damaged erythrocytes have a shortened lifespan and contribute to vaso-occlusive crises (VOC), which are episodes of severe pain that may require opioid use, as well as other complications of chronic hemolytic anemia such as acute and chronic end-organ damage.¹

Voxelotor is a medication that inhibits sickle hemoglobin polymerization and works by binding reversibly to hemoglobin. It received accelerated approval by the Food and Drug Administration for the treatment of SCD in November 2019 for patients 12 years and older, based on data from the Hemoglobin Oxygen Affinity Modulation to Inhibit Sickle Hemoglobin Polymerization (HOPE) trial, which demonstrated the drug’s safety and efficacy.² The approval was notably dependent on the novel endpoint of an increase in hemoglobin, which was a result of reduced red blood cell hemolysis.² The HOPE trial was a randomized, double-blind, placebo-controlled phase 3 trial of 274 patients between 2016 and 2018.³ After 72 weeks, patients receiving voxelotor showed a significant increase in mean hemoglobin concentration and improvement in markers of hemolysis. The study concluded that voxelotor 1500 mg once daily resulted in improvements in hemoglobin concentrations maintained over 72 weeks and could potentially change the morbidity associated with hemolytic anemia in SCD.⁴ In 2021, voxelotor Food and Drug Administration approval was expanded to include children ages 4 years and older based on phase 2 trial data.⁵

In both clinical trials and real-world data, patients taking voxelotor have reported adverse events. The most common adverse events have been headache, diarrhea, and rash, which have generally been mild in severity.^3,6,7 In real-world studies, between 39% and 46% of patients taking voxelotor experienced at least 1 side effect. At the same time, the clinical trials found that 94% to 98% of patients in the 1500-mg group experienced at least 1 side effect, but only 26% to 28% of these were severe adverse events.^3,4,7,8 Between 15% and 29% of patients in the real-world studies required dose modification due to side effects, and 5% to 23% discontinued voxelotor treatment altogether.^8,9

On September 25, 2024, voxelotor was abruptly withdrawn from all markets by the manufacturer after a “higher than anticipated” number of deaths were reported and clinical evidence showed that the drug’s benefits no longer outweighed the risks. A recall notice cited emerging data that suggested an unfavorable imbalance in VOCs and fatal events in patients treated with voxelotor.¹⁰ Beginning September 27, 2024, following market withdrawal, Children’s Healthcare of Atlanta (CHOA) patients were advised to reduce their voxelotor dose by a third each week over 2 to 3 weeks, according to recommendations by the Sickle Cell Disease Association of America. Patients were asked to have blood tests performed 1 to 2 weeks after stopping voxelotor to monitor effects and any complications.

At the time of withdrawal, voxelotor was 1 of 4 available disease-modifying therapies (DMTs) for SCD, along with hydroxyurea, crizanlizumab, and L-glutamine. Since the original clinical trial, there have been few large, real-world descriptions of voxelotor use in pediatric patients with SCD. CHOA has the largest pediatric SCD clinic population in the US, seeing more than 2000 patients a year. We report the volume and experience of patients who received voxelotor at our institution.

Materials and Methods

All patients with SCD ever seen at CHOA were eligible for review. Potentially eligible patients were identified via 2 methods. Our institution maintains a Sickle Cell Clinical Database, which is a prospective cohort of all patients with a laboratory and hematologist-confirmed diagnosis of SCD seen at our institution. The charts of these patients are manually reviewed, and their use of SCD DMTs is recorded in the Sickle Cell Clinical Database. All patients who were noted to have ever received voxelotor in the Sickle Cell Clinical Database were pulled for review. As a secondary method, all prescriptions ever written at our institution for “voxelotor” or “Oxbryta” were pulled from the electronic health record (Epic) via the SlicerDicer tool.

The medical records were reviewed for all patients identified through either method to determine if they had ever started voxelotor, along with the start and stop dates if applicable. These determinations were based on a review of prescriptions and clinical notes. In cases where notes were unclear, the most likely date was estimated (eg, “the patient started in June” without further details was assigned a start date of June 15). Patients who did not take their medication for more than 3 months were considered to have stopped and were assigned a secondary start date if they resumed consistent use. Any side effects, concurrent DMTs, and reasons for discontinuation were manually abstracted from the chart. Side effects were defined as any new or worsening symptom that was related or thought to be related to voxelotor per the notes in the chart. Patients who died, transitioned to adult care, or moved outside the CHOA network but were still actively taking voxelotor at the time of the event were considered censored, and no further data were collected for them beyond that point.

All patients who started taking voxelotor commercially or as part of a noninterventional trial (open-label extension or expanded access protocol) at our institution were included in the study. Patients who were on voxelotor as part of an interventional trial were only included if they subsequently switched to the commercial medication, and only their time on the commercial or noninterventional medication was analyzed. Before market withdrawal, voxelotor initiation at CHOA was based on patient and provider preference but typically recommended for those patients who still had pain despite optimization of hydroxyurea use or significant symptomatic anemia as defined by significant fatigue, chronic headache, or low hemoglobin as determined by the treating physician, or for families who declined to take hydroxyurea. Voxelotor was available in a 500- or 300-mg tablet for oral suspension. Standard dosing was 1500 mg orally once daily for patients 12 years and older. For pediatric patients aged 4 years to younger than 12 years, the recommended once-daily oral dose was as follows: 10 kg to under 20 kg: 600 mg; 20 kg to under 40 kg: 900 mg; 40 kg or greater: 1500 mg. Patients using the tablets for oral suspension were advised to dissolve the tablet in a clear liquid, such as water or apple juice. Dose modification could be attempted for patients who experienced side effects upon voxelotor initiation, and the medication would subsequently be either continued at the lower dose or increased back to the original dose once symptoms had resolved.

For all patients, demographics, experience with voxelotor, and other concurrent DMTs were abstracted and summarized. Comparisons were made between patients who received voxelotor for less than 6 months and those who received the drug for 6 months or more using Χ² or Fisher’s exact tests. Patients who stopped and restarted voxelotor were counted in the over 6 months category only if they took it consistently for at least 6 months; however, their cumulative time ever taking voxelotor was used in calculating the total time on treatment. The volume of voxelotor use at our institution was measured by the number of patients actively taking voxelotor at any point during the calendar year.

Further analysis was done on a subset of patients who took voxelotor consistently for at least 6 months and had available baseline data. For this subgroup, clinical markers and healthcare usage were assessed while taking voxelotor compared to baseline. Average hemoglobin, absolute reticulocyte count (ARC), lactate dehydrogenase (LDH), and total bilirubin values obtained during clinic visits in the year before starting voxelotor were used for baseline, and the average of labs obtained during clinic visits while on voxelotor was used for on-treatment measures. For patients who were previously taking voxelotor as part of a clinical trial, baseline was calculated for the year before being on any voxelotor product. The annualized number of transfusions and blood units at baseline compared with while on voxelotor was analyzed among patients who had at least 1 intermittent transfusion in the baseline year and were not on chronic transfusions.

Baseline hospital usage was defined as the number of any-cause ED encounters and hospitalizations, and hospital days for SCD pain or acute chest syndrome (ACS) in the year before starting voxelotor for patients who had at least one ED encounter or admission during baseline. Annualized utilization while on voxelotor was calculated as the number of these measures per time on treatment. Additionally, a change in admissions was calculated among patients who were high hospital users at baseline, defined as having had 5 or more admissions for pain or ACS in the year. Sub-analyses were conducted on patients taking voxelotor at the time of market withdrawal. All comparisons were made using paired t-tests with statistical significance set at p < 0.05, and analyses were performed using R (R version 4.4.0; R Core Team 2021).

Results

Population overview.

A total of 289 patients with SCD with at least 1 prescription written for voxelotor between 2019 and 2024 were reviewed. Forty-four patients (15%) were excluded; the majority (34; 77%) were prescribed voxelotor but never started taking it, primarily due to insurance approval or cost, patient or caregiver deciding against it, or being prescribed right before the recall. The remaining 10 patients were excluded for being on voxelotor as part of an interventional trial but never transitioning to commercial medication at our institution (n = 5), not starting voxelotor until after transitioning to adult care (n = 2), taking voxelotor at a previous institution but never starting it at our institution (n = 1), starting voxelotor after an unsuccessful bone marrow transplant (n = 1), or having a voxelotor prescription entered in error (n = 1).

The characteristics of the remaining 245 patients are described in Table 1. The majority of patients were male (53%), had an Hb SS genotype (94%), and an average age of starting voxelotor of 12.7 years (SD 4.2 years, range 4.2–21.1 years). Fifty-seven (23%) patients were on voxelotor for less than 6 months, and 188 (77%) were on voxelotor for 6 months or more. The average time taking voxelotor was 9.9 months (SD = 15.6). Three patients started taking voxelotor less than a month before market withdrawal, and the longest length of time on voxelotor was 61 months. Seventeen patients had irregular usage, stopping and restarting the medication. Most of these patients either stopped due to adherence issues and restarted with improved adherence (n = 4) or stopped due to side effects and restarted to see if the side effects would improve (n = 8). The majority of patients were taking at least 1 additional DMT (hydroxyurea, crizanlizumab, L-glutamine, or chronic transfusion therapy) at the same time as voxelotor (n = 218, 89%; Table 1).

Table 1.

Characteristics of pediatric patients with sickle cell disease on voxelotor at a single institution

	All Patients (N = 245)	Taking < 6 mo (n = 57)	Taking ≥ 6 mo (n = 188)
Sex, n (%)
Female	116 (47)	26 (46)	90 (48)
Male	129 (53)	31 (54)	98 (52)
Age first started taking voxelotor, yr, mean (SD)	12.7 (4.2)	12.9 (4.5)	12.6 (4.2)
Sickle cell genotype, n (%)
Hb SS	230 (94)	52 (91)	178 (95)
Hb S Beta Zero Thalassemia	8 (3)	4 (7)	4 (2)
Hb SS or Hb S Beta Zero Thalassemia	4 (2)	0	4 (2)
Hb S Beta Plus Thalassemia	1 (0.4)	1 (2)	0
Hb SC	2 (0.8)	0	2 (1)
Concurrent DMTs*, n (%)
Hydroxyurea	212 (87)	51 (89)	161 (86)
L-glutamine	8 (3)	1 (2)	7 (4)
Crizanlizumab	11 (5)	0	11 (6)
Chronic transfusion therapy	12 (5)	1 (2)	11 (6)
None	27 (11)	6 (11)	21 (11)
Total time on voxelotor treatment, mo, mean (SD)	9.9 (15.6)	2.9 (2.1)	25.0 (14.1)
Reason for discontinuing†, n (%)	n = 206	n = 51	n = 155
Side effect‡	43 (21)	24 (47)	19 (12)
Did not feel medication was working	5 (2)	3 (6)	2 (1)
Unable to tolerate medication‡	9 (4)	5 (10)	4 (3)
High medication/pill burden	2 (1)	0	2 (1)
Cost/insurance issues	3 (1)	1 (2)	2 (1)
Poor compliance	6 (3)	2 (4)	4 (3)
Transaminitis	2 (1)	2 (4)	0
Transition to other therapy/BMT	11 (5)	1 (2)	10 (6)
Market withdrawal‡	118 (57)	11 (22)	107 (69)
Unknown	12 (6)	3 (6)	9 (6)
Censored†, n (%)	n = 43	n = 7	n = 36
Transitioned to adult care	36 (84)	6 (86)	30 (83)
Moved outside CHOA network	5 (12)	1 (14)	4 (11)
Deceased	2 (5)	0	2 (6)

BMT, Bone Marrow Transplant; CHOA, Children’s Healthcare of Atlanta; DMT, disease-modifying therapy.

^*Patients may have been on >1 concurrent DMT.

^†Patients may have had more than one discontinuation reason. Patients who re-started may be in the discontinued and censored categories.

^‡p value < 0.05

Among all patients who started voxelotor, 206 discontinued at some point, while 43 patients were censored from our analysis but were taking voxelotor at the time of censoring (Table 1). The majority of patients who were censored transitioned to adult care (n = 36) or moved outside of our care network (n = 5). Among patients taking voxelotor for less than 6 months, the primary reason for discontinuation was a side effect (47%), while for those who took voxelotor for over 6 months, the main reason was market withdrawal (69%). Two patients died while taking voxelotor: 1 in a car accident and 1 from hyperhemolysis and multiorgan failure with COVID-19. Neither death was believed to be directly related to the medication.

The timeline of voxelotor prescriptions at our institution is shown in Figure 1. The first year that patients were taking voxelotor in a noninterventional research study was in 2019 (n = 18). Each subsequent year saw an increase in patients taking voxelotor, peaking in 2023 at 176 patients. In 2024, the number of patients on voxelotor during the year fell slightly to 152. At its peak in 2023, approximately 10% of all patients with SCD aged four or older at our institution were taking voxelotor.

Figure 1.

Pediatric patients with sickle cell disease on voxelotor at a single institution by year, 2019 to 2024.

Side effects.

The side effects that patients experienced while taking voxelotor are shown in Table 2. There were 167 (68%) patients who did not experience any side effects, while 78 (32%) did. Most patients experienced only 1 side effect (42; 54%), while 26 (33%) had 2 side effects, and 10 (13%) endured 3 or more. Among those who experienced at least 1 side effect, 43 (55%) discontinued voxelotor due to side ­effects. The most commonly reported side effects were gastrointestinal issues (81%), pain (28%), headaches (21%), and allergic reactions (15%) (Table 2). The top-reported gastrointestinal symptoms included nausea, diarrhea, and vomiting; the majority of patients who experienced pain reported abdominal pain (79%), and allergic reactions primarily presented as a rash (67%). The highest proportion of patients who discontinued voxelotor due to side effects was those with allergic reactions (83%), headaches (75%), chest distress (67%), and dizziness (67%). There was no difference in the occurrence of side effects between patients who were taking an additional DMT versus those who were only taking voxelotor.

Table 2.

Patient-reported side effects on voxelotor among pediatric patients with sickle cell disease

	Total With Side Effect(s) (n = 78), n (%)	Number of Patients Who Discontinued Due to Listed Side Effect(s), n (%)
Allergic reaction	12 (15)	10 (83)
Itching 2
Rash 8
Swelling 2
Chest distress	3 (4)	2 (67)
Chest tightness 1
Shortness of breath 2
Dark nails	1 (1)	0
Dehydration	1 (1)	0
Dizziness	3 (4)	2 (67)
Double vision/vertigo	2 (3)	0
Fatigue	1 (1)	1 (100)
Fever	1 (1)	1 (100)
Headache	16 (21)	12 (75)
GI distress	63 (81)	31 (49)
Acid reflux 1
Anorexia 1
Constipation 3
Diarrhea 12
Heartburn 1
Indigestion 1
Nausea 25
Stomach cramps 1
Unspecified 9
Vomiting 9
Pain	22 (28)	11 (50)
Abdominal 14
Back 2
Chest 2
Unspecified 4
Priapism	1 (1)	0
Weight change	2 (3)	1 (50)
Gain 1
Loss 1

Patients may have experienced >1 one side effect.

Clinical outcomes.

Markers of hemolysis and clinical outcomes were assessed for 183 patients who were consistently on voxelotor for at least 6 months and had available baseline data (Table 3). While on voxelotor, patients experienced a mean increase in hemoglobin of 0.5 g/dL (p < 0.001), and a mean decrease in ARC of 34.8 10⁹/L (p < 0.001) and in total bilirubin of 0.6 mg/dL (p < 0.001). A small but non-significant decrease in LDH of 20.1 U/L was also seen (p = 0.22). Of the 182 patients with available hemoglobin measurements, 83 (46%) had an average increase of at least 0.5 g/dL and 41 (22.5%) had an average increase of 1 g/dL or greater during their time on voxelotor (range –2.4 to 3.7, Figure 2). No difference was seen in mean hemoglobin change between patients who were on an additional DMT versus those only taking voxelotor. Among the 58 patients not on chronic transfusions with at least 1 intermittent transfusion at baseline, there was an average decrease of 0.8 transfusions per year (p = 0.002) and 0.9 transfused units per year (p = 0.03) in the time on voxelotor compared to baseline. The majority of the 58 patients had a reduction in the number of transfusions (45; 77.6%).

Table 3.

Hospital utilization, markers of hemolysis, and transfusion rates before and during voxelotor treatment among patients on at least six months of therapy

		N	Baseline, mean (SD)	On Treatment, mean (SD)	Mean Difference, mean (SD)	p
Utilization
	Annualized ED Encounters	91	2.2 (2.5)	1.4 (1.5)	−0.9 (2.0)	<0.001
	Annualized hospital admissions	91	2.8 (2.9)	2.3 (2.7)	−0.5 (2.5)	0.06
	Hospital days per admission	72	6.0 (2.7)	6.1 (3.7)	0.09 (3.5)	0.83
Labs
	Hemoglobin, (g/dL)	182	8.2 (1.1)	8.7 (1.3)	0.5 (0.9)	<0.001
	Absolute reticulocyte count, (109/L)	180	297.0 (98.8)	262.2 (99.6)	−34.8 (81.1)	<0.001
	Lactate dehydrogenase (U/L)	119	544.4 (197.2)	524.2 (162.9)	−20.1 (176.4)	0.22
	Total bilirubin (mg/dL)	160	3.5 (2.0)	2.9 (2.2)	−0.6 (1.7)	<0.001
Intermittent transfusions
	Annualized transfusions	58	2.3 (1.9)	1.5 (2.5)	−0.8 (1.9)	0.002
	Annualized transfused units	58	3.1 (2.5)	2.2 (3.9)	−0.9 (3.1)	0.03

Figure 2.

Waterfall plot of the mean change in hemoglobin while on voxelotor from baseline in pediatric patients with sickle cell disease.

There was a significant decrease in annualized ED encounters (−0.9 encounters/year, p < 0.001) and a non-significant decrease in annualized hospital admissions for SCD pain or ACS (−0.5 admissions/year, p = 0.06) while on voxelotor as compared with baseline among patients with at least 1 encounter in the baseline year (n = 91). There was no change in the number of hospital days per admission among patients who had admissions in both periods (n = 72). Patients who were high hospital utilizers at baseline (n = 18) had a mean decrease of 2.4 admissions per year (SD = 4.3, p = 0.03).

Post-market withdrawal.

At the time of market withdrawal, 123 patients were actively taking voxelotor. Only a subset of patients who stopped voxelotor due to market withdrawal had labs obtained after the medication was withdrawn through the end of the year (available data ranged from n = 38 for LDH to n = 67 for hemoglobin and ARC). Among those with available data, there was an average decrease in hemoglobin of 0.67 g/dL (p < 0.001) and an average increase in ARC (22.3 10⁹/L, p = 0.05), LDH (52.6 U/L, p = 0.01), and bilirubin (0.67 mg/dL, p = 0.005) in the time off voxelotor compared with while taking it.

Overall, no change was seen in monthly ED visits or admissions in these patients after stopping voxelotor through the end of 2024, compared with while on voxelotor. Seventeen (13.8%) patients were admitted for pain or ACS between October and December 2024 after the market withdrawal. During hospitalization, these patients experienced an average hemoglobin drop of 2.5 g/dL (range 0.5–5.4) compared with their pre-admission hemoglobin levels on voxelotor, and 12 patients (70.6%) required transfusions during their stay.

Discussion

This retrospective study assessed hematologic responses and clinical outcomes of voxelotor use in pediatric patients with SCD at a single institution. Between 2019 and 2024, 245 patients were treated with voxelotor at our institution, with a peak of 176 patients on voxelotor in 2023, or approximately 10% of all patients with SCD aged 4 and older. We found that, on average, patients experienced increased hemoglobin concentrations, fewer blood transfusions, reduced ED visits, and that the most frequently admitted patients had fewer admissions while on voxelotor compared with baseline.

Over three-quarters of patients who ever started voxelotor continued it for at least 6 months, suggesting it was well tolerated in our population. Most patients (68%) had no documented side effects from the medication, and more than half of those who discontinued voxelotor only did so due to market withdrawal, further indicating a positive experience for many of our patients. However, voxelotor was not tolerated equally well by all patients. Over 20% of those who stopped taking voxelotor did so due to side effects from the medication. We found that the most common reported side effects included gastrointestinal symptoms (primarily nausea and diarrhea), pain, headache, and allergic reactions (such as itching and rash). A high proportion of patients who experienced headaches and allergic reactions discontinued voxelotor because of these side effects. These adverse effects are similar to those reported in both trial data and other real-world studies.^3,7,8,11 In addition to side effects, patients also stopped taking voxelotor because of difficulty tolerating the taste if they were prescribed the tablets for oral suspension or because the patient or their caregiver did not feel the medication was effective. Insurance approval issues (n = 14) were also seen as both reasons for discontinuing voxelotor and explanations for why some patients never started it.

Overall, patients taking voxelotor for at least 6 months had a positive clinical response to the medication. In our pediatric population, there was a significant average increase in mean hemoglobin from a baseline of 0.5 g/dL (p < 0.001), with 46% of patients experiencing a mean increase of 0.5 g/dL or greater and 22.5% experiencing an increase of 1 g/dL or greater. This is lower than what was reported in the HOPE trial, where patients had a mean change of 1.1 g/dL in the 1500 mg group and 0.6 g/dL in the 900 mg group, and in the HOPE-KIDS-1 trial, which found that 47% of children aged 4 to 11 experienced at least a 1 g/dL increase from baseline.^3,11 Other real-world literature on voxelotor in adolescent and adult populations found higher hemoglobin responses, from an average increase of 2.0 g/dL to 50% to 60% of voxelotor patients having an increase of at least 1 g/dL.^7,12–14 The lower hemoglobin response in our population may be due to inconsistent medication adherence, which we were not able to measure. Consistent with existing literature, we found that patients had an improvement in ARC and total bilirubin and a reduction in intermittent transfusions while taking voxelotor.^3,7,12,13 We also found a small, non-significant mean change in LDH from baseline.

Among patients with baseline acute care usage, we found a significant decrease in annualized ED encounters (p < 0.001) and a non-significant decrease in annualized admissions for SCD pain or ACS (p = 0.06). However, the 18 patients who had 5 or more admissions in 1 year at baseline had a mean decrease of 2.4 hospital admissions per year. Existing literature has shown mixed results of the impact of voxelotor on SCD pain. Although the HOPE trial found that the 2 treatment groups had a lower number of annualized VOCs than the placebo group, there was no difference in the percentage of patients who had at least 1 VOC between groups.³ A review of patient experience on voxelotor using a large claims database found a significant annualized mean decrease in VOC frequency (10.9–8.4), VOC-related hospitalizations (7.2–4.8), and length of stay (21.2–14.7 days) from baseline (p < 0.001).¹² Importantly, we were only able to assess all-cause ED encounters and hospital admissions for VOC or ACS. We were not able to capture times when the patient experienced a VOC at home and did not seek medical attention.

We observed worsening hematologic markers after stopping voxelotor among a subset of patients who were instructed to discontinue the medication in the Fall of 2024 due to market withdrawal. On average, these patients had a significant decrease in hemoglobin (p < 0.001) and a significant increase in LDH (p = 0.01), ARC (p = 0.05), and total bilirubin (p = 0.005). Although we did not see an increase in acute care use following market withdrawal in the entire group, a subset of patients was admitted with clinical symptoms and more severe anemia. Longer-term analysis is needed to better determine the clinical implications of discontinuing voxelotor in patients who were experiencing clinical benefit while on the medication.

Although the impact of SCD is widespread, affecting daily functioning, school, work, and family life, the development of effective therapies has been slow, and recent setbacks have highlighted the gap in available treatments. Crizanlizumab (Adakveo) was removed from the European Union market in 2023 after the European Medicines Agency concluded that its benefits no longer outweighed its risks.¹⁵ Currently, crizanlizumab remains approved and available for patients with SCD in the US, making it 1 of only 3 approved SCD therapies in the US alongside hydroxyurea and L-glutamine.

Our findings highlight real-world clinical outcomes and side effects of voxelotor in a larger group of children, compared with previously published literature and clinical trial data, and how this may impact the risk-benefit profile of new medications. Despite withdrawal because of safety and efficacy concerns, there may be a subset of patients who derive clinical benefit from the medication, such as those in our study with high baseline hospital use who had a significant decrease in VOC admissions on voxelotor.

Limitations of study.

Several limitations should be acknowledged. As a retrospective chart-review study, causality cannot be definitively established, and our data relied on documentation completeness and voxelotor-specific notes in medical records. Adherence to voxelotor could not be reliably assessed because documentation was often incomplete or missing from medical records, and we had no objective method to measure medication consumption. Precise start and stop dates for voxelotor were inconsistently recorded, limiting the ability to accurately determine treatment duration and clinical outcomes that may have resulted from voxelotor. Identifying side effects was limited to those explicitly documented in medical records. This may underestimate the true occurrence of side effects from voxelotor.

Furthermore, the severity of reported side effects could not be fully assessed, except in cases where patients stopped voxelotor because of these side effects. Similarly, we were limited to using changes in acute care use to evaluate clinical outcomes on voxelotor. Many patients likely had VOCs at home but did not present for care, and thus we were not able to fully assess a change in pain events. This study design did not allow for control of confounding variables, including concurrent DMTs, which may have influenced clinical outcomes independently of voxelotor. Finally, we were limited to small samples of patients with available data to assess changes in outcomes after stopping voxelotor.

Despite these limitations, our study’s strength lies in the large number of patients we reviewed, including data on patients under 12 years old. By manually reviewing patient charts, we were able to gather more detailed and sensitive information that would have been missed through database extraction.

Conclusions.

This retrospective review provides a real-world overview of the pediatric and adolescent experience of voxelotor treatment at a single institution. Overall, patients responded positively to treatment and had good hematologic responses, although to a lesser degree than seen in clinical trials. The majority of patients who discontinued voxelotor before 6 months of use did so due to side effects. Currently, similar agents that target Hb S polymerization are in development, and it will be important in future studies to consider the interplay of medication side effects, patient adherence and access, and concurrent DMTs on drug effectiveness. The withdrawal of voxelotor from worldwide markets highlights the need for the continued development of safe and effective novel therapeutics for SCD.

ABBREVIATIONS

ACS

acute chest syndrome

ARC

absolute reticulocyte count

CHOA

Children’s Healthcare of Atlanta

DMT

disease-modifying therapy

ED

emergency department

HOPE

Hemoglobin Oxygen Affinity Modulation to Inhibit Sickle Hemoglobin Polymerization

LDH

lactate dehydrogenase

SCD

sickle cell disease

VOC

vasco-occlusive crisis