Using disease-modifying therapies in sickle cell disease

A 28-year-old man with hemoglobin SS (HbSS) sickle cell disease (SCD) complicated by retinopathy, acute chest syndrome (ACS), nephropathy, and frequent pain episodes requiring healthcare utilization was seen in clinic for follow up. He was on hydroxyurea (∼20  mg/kg/d, his maximum tolerated dose) and losartan (25  mg/d) and was fairly adherent. Laboratory studies showed a white blood cell count (WBC) of 6.2  ×  10⁹/L; an Hb level of 9.7  g/dL; a mean corpuscular volume of 110 fL; a platelet count of 292  ×  10⁹/L; an absolute neutrophil count of 3.5  ×  10⁹/L; an absolute reticulocyte count of 114.8  ×  10⁹/L; a creatinine level of 0.8   mg/dL; and a cystatin C level of 0.8   mg/L. Hb electrophoresis showed a sickle Hb (HbS) level of 81.2%; fetal Hb (HbF), 15.6%; and HbA2, 3.2%. Given his frequent pain episodes, options for optimizing his care were discussed.

SCD affects millions of individuals worldwide.¹  Although a rare disease in the United States, an estimated 230 000 children (representing the vast majority of worldwide births) were born with sickle cell anemia (referring to HbSS and HbSβ⁰) in sub-Saharan Africa in 2010.²  In addition to the presence of HbS, SCD is characterized by hemolytic anemia, vaso-occlusive complications, and progressive end-organ dysfunction. The mortality rate for children with sickle cell anemia remains high in sub-Saharan Africa, with estimated rates of 36.4% for those younger than 5 years and 43.3% for those younger than 10 years,³  but most children in resource-rich countries live to adulthood.⁴  Despite increased survival to adulthood, individuals with SCD in resource-rich nations have a reduced life expectancy compared to the general population.^5-8  SCD can be cured following allogeneic stem cell transplantation and possibly following gene therapy and gene editing. However, as most patients do not have access to these potentially curative treatments, the availability of safe, effective, and affordable drugs remains highly desirable.

This article reviews important historic and recent drug trials for SCD, highlighting regulatory agency–approved drug therapies and our approach to the use of these agents.

The development of effective drug therapies for SCD requires an adequate understanding of its pathophysiology. The primary event in the pathophysiology of SCD is the polymerization of HbS following deoxygenation.⁹  The polymerization of HbS depends on several factors, including the degree of HbS deoxygenation, the intracellular HbS concentration, and the amount of HbF.⁹  HbS polymerization as well as its multiple consequences, including endothelial cell injury, endothelial dysfunction, increased oxidant stress, inflammation, coagulation and platelet activation, and complement activation, is a therapeutic target in SCD. The clinical manifestations of SCD appear to be driven by 2 major pathophysiological processes: vaso-occlusion with ischemia-reperfusion injury and hemolytic anemia.¹  SCD may also be divided into 2 overlapping subphenotypes: viscosity-vaso-occlusion (characterized by higher Hb levels, possibly increased blood viscosity, and clinical complications such as acute pain episodes, ACS, and avascular necrosis of bone) and hemolysis-endothelial dysfunction (characterized by lower Hb and higher levels of markers of hemolysis, including reticulocyte count, indirect bilirubin and lactate dehydrogenase, and clinical complications such as leg ulcers, priapism, stroke, and pulmonary hypertension).¹⁰  This classification, while controversial, may facilitate an increased understanding of the pathobiology of SCD-related complications and the effects of therapeutic agents. The pathophysiology of SCD is beyond the scope of this article and is reviewed elsewhere.¹¹ 

Although SCD affects multiple body organs, most trials of potentially disease-modifying drugs have focused on acute pain episodes (commonly referred to as vaso-occlusive crises, or VOCs) as their primary end point. The general approaches to the management of acute pain episodes are support, intervention, and prevention (Figure 1). No drugs have been approved for shortening the duration of acute vaso-occlusive complications (Table 1). As such, acute pain episodes are usually managed supportively. The majority of drug trials have focused on disease-modifying therapies to prevent acute pain episodes. For many years, hydroxyurea was the only drug approved by the US Food and Drug Administration (FDA) for sickle cell anemia. More recently, however, 3 other drugs, L-glutamine, crizanlizumab, and voxelotor, have been approved for SCD. The following sections focus on phase 3 and select multicenter phase 2 studies of disease-modifying agents.

Multiple mechanisms can prevent HbS polymerization: 1) inhibition of sickle fiber intermolecular contacts; 2) increase in HbF; 3) decrease of intracellular HbS concentration; 4) increase of oxygen affinity; and 5) decrease in 2,3-diphosphoglycerate concentration.¹² 

Hydroxyurea, an inhibitor of ribonucleotide reductase, is thought to exert its therapeutic effects largely by inducing HbF, although the mechanisms of HbF induction remain unclear. Hydroxyurea was approved for adults based on the results of the double-blind, placebo-controlled, multicenter trial of hydroxyurea in 299 patients with sickle cell anemia, which showed significant reductions of VOCs (median, 2.5 vs 4.5 crises per year; P < .0001), hospitalizations due to crises (median annual rates, 1.0 vs 2.4; P < .001), ACS (25 vs 51 patients; P < .0001), and blood transfusions (48 vs 73 patients; P  =  .001; and 336 vs 586 units of blood; P  =  .004) following treatment with hydroxyurea vs placebo.¹³  Similar findings were observed in the multicenter BABY HUG trial in which treatment with hydroxyurea significantly reduced disease-related acute complications in young children with sickle cell anemia.¹⁴  Based on this, a National Heart, Lung, and Blood Institute (NHLBI) expert panel strongly recommended offering hydroxyurea to children as young as 9 months with sickle cell anemia.¹⁵  Hydroxyurea was approved in the United States for children aged 2 years or older in 2017 based on findings from an open- label, single-arm trial that showed significant decreases in acute vaso-occlusive events and transfusion requirements.¹⁶  Although no significant differences were observed in the intention-to-treat analysis, hydroxyurea reduced the risk of conversion from conditional to abnormal transcranial Doppler (TCD) velocity compared with observation (0 vs 50%; P  =  .02) in post–host analysis of a multicenter trial.¹⁷  In individuals with abnormal TCD on chronic blood transfusion and no severe vasculopathy, hydroxyurea was also noninferior to chronic red blood cell (RBC) transfusion in preventing stroke.¹⁸  In this trial the final model-based TCD velocities for standard transfusions vs hydroxyurea were 143  cm/s (95% CI, 140-146) vs 138 cm/s (95% CI, 135-142), with a difference of 4.54 (95% CI, 0.10-8.98; P  =  8.82  ×  10⁻¹⁶ for noninferiority, and P  =  .023 for superiority post hoc). In the REACH study, the treatment of 606 children from 4 sub-Saharan countries with hydroxyurea significantly increased total Hb (an increase of 1.0  g/dL; 95% CI, 0.8-1.0) and HbF levels (an increase of 12.5%; 95% CI, 11.8-13.1) and decreased the rates of vaso-occlusive pain (98.3 vs 44.6 events per 100 patient-years; incidence rate ratio [IRR], 0.45; 95% CI, 0.37-0.56), RBC transfusions (43.3 vs 14.2 events per 100 patient-years; IRR, 0.33; 95% CI, 0.23-0.47), and mortality (3.6 vs 1.1 deaths per 100 patient-years; IRR, 0.30; 95% CI, 0.10-0.88) when compared with the pretreatment period.¹⁹  In a double-blind, parallel-group, phase 3 trial of 220 children with sickle cell anemia and abnormal TCD velocities conducted in Nigeria, no significant difference was seen in the stroke incidence rate with low-dose hydroxyurea (10  mg/kg) compared with moderate-dose hydroxyurea (20  mg/kg), although the incidence rate for all-cause hospitalization was lower with moderate-dose hydroxyurea.²⁰  Furthermore, no significant difference in the incidence rates of the primary outcome measures (stroke, transient ischemic attack, and death) was seen with low-dose vs moderate-dose hydroxyurea for secondary prevention of stroke.²¹ 

Voxelotor is an HbS polymerization inhibitor that reversibly binds to Hb and stabilizes it in the oxygenated (relaxed) state.²²  In the multicenter phase 3 HOPE study, 274 patients (12 years and older) with SCD were randomized to receive a daily dose of 1500  mg of voxelotor, 900  mg of voxelotor, or placebo for 72 weeks. In the intention-to-treat analysis, 51% (95% CI, 41-61) of participants who received 1500  mg/d achieved a Hb increase of greater than 1  g/dL from baseline after 24 weeks of therapy compared to 7% (95% CI, 1-12) in the placebo group. Treatment with voxelotor at 1500  mg/d also resulted in a significant increase in Hb (mean change, 1.14  g/dL vs −0.1  g/dL; P < .001) and significant decreases in indirect bilirubin (mean change, −29.1% vs −3.2%; P < .001) and percent reticulocyte count (mean change, −19.9% vs 4.5%; P < .001) compared to placebo.²³  Similarly, a Hb increase of more than 1  g/dL from baseline after 24 weeks of voxelotor was observed in 47% of children with HbSS or HbSβ⁰ in the HOPE KIDS-1 trial.²⁴ 

Senicapoc, a potent blocker of the Gardos channel, a calcium- activated potassium channel of intermediate conductance in RBCs, improves RBC hydration by reducing the loss of solute and water. However, an increase in Hb (mean change, 0.59  g/dL vs −0.1  g/dL; P < .001) and decreases in percent reticulocyte count (mean change, −2.46% vs −0.79%; P < .001) and indirect bilirubin (mean change −16.6  µmol/L vs −0.3  µmol/L; P < .001), both markers of hemolysis, following senicapoc administration were not accompanied by a significant reduction in acute pain episodes.²⁵  Magnesium inhibits K⁺ efflux through the potassium chloride cotransport channel in RBCs and consequently prevents RBC dehydration. Treatment with intravenous (IV) magnesium when compared with placebo did not shorten the length of hospitalization (median, 56 hours; interquartile range [IQR], 27.0-109.0 vs 47 hours [IQR, 24.0-99.0]; P  =  .24), reduce opioid use (median, 1.46  mg/kg vs 1.28  mg/kg morphine equivalents; P  =  .12), or improve quality of life in children and young adults who were hospitalized for acute pain episodes.²⁶ Table 2 lists actively recruiting studies of antisickling agents.

Oxidative stress is a major contributor to the pathophysiology of SCD. L-glutamine, a conditionally essential amino acid, is a precursor for nicotinamide adenine dinucleotide (NAD), increases the NAD redox ratio within sickle RBCs, and may improve RBC health by reducing oxidative stress. In a multicenter, placebo-controlled trial of 230 patients with HbSS or HbSβ⁰, L-glutamine, administrated twice daily, significantly reduced acute pain episodes (3.0 vs 4.0; P  =  .005) and hospitalizations (2.0 vs 3.0; P  =  .005) compared to placebo.²⁷  Treatment with L-glutamine also resulted in a significantly lower cumulative number of hospital days and fewer occurrences of ACS compared with placebo. Treatment with another antioxidant, N-acetylcysteine (NAC), at a dose of 600  mg twice a day for 6 months did not significantly decrease the rate of SCD-related pain days per patient-year, days of hospital admission, number of admissions, or days with home analgesic use compared with placebo.²⁸ 

Antiadhesion agents may improve flow in the microvasculature by reducing abnormal cell-cell (RBC, leukocyte, platelet, endothelial cell) interactions. Crizanlizumab, a humanized monoclonal anti–P-selectin antibody, blocks the interaction between P-selectin (expressed on activated endothelial cells and platelets) and P-selectin glycoprotein ligand 1. In the SUSTAIN trial, a multicenter, randomized, placebo-controlled phase 2 trial, crizanlizumab administered at a dose of 2.5  mg/kg or 5  mg/kg via IV every 4 weeks (following an initial loading dose) in a 52-week period was compared with placebo in individuals with SCD.²⁹  Treatment with high-dose crizanlizumab (5  mg/kg) resulted in a significantly lower median rate of painful crisis (1.63 vs 2.98; P = .01), a lower median rate of uncomplicated crisis (1.08 vs 2.91; P = .02), and longer median times to occurrence of the first (4.07 months vs 1.38 months; P = .001) and second pain crises (10.32 vs 5.09 months; P  =  .02). The benefit in reducing pain crisis was observed regardless of the ongoing use of hydroxyurea, pain crisis frequency in the previous 12 months, or SCD genotype. However, recently reported results from the phase 3 STAND trial showed no significant difference between either crizanlizumab at 5  mg/kg or 7.5  mg/kg and placebo on the annualized rates of VOC leading to a healthcare visit over the first year postrandomization, although there were no new safety concerns.³⁰  Based on these results, the European Medicines Agency's Committee for Medicinal Products for Human Use concluded that the benefits of crizanlizumab do not outweigh its risks and recommended the revocation of its marketing authorization in the European Union.

Rivipansel (previously called GMI-1070) is a small-molecule pan- selectin inhibitor with highest affinity to E-selectin. In a multicenter phase 2 trial, rivipansel reduced the cumulative IV opioid dose during acute pain episodes by 83% compared to placebo.³¹  However, in a multicenter, placebo-controlled phase 3 trial (RESET), no benefit was seen with rivipansel in shortening the times to readiness for discharge, hospital discharge, or discontinuation of IV opioids.³²  In a post hoc analysis, the early initiation of rivipansel within 26.4 hours of crisis onset resulted in clinically meaningful reductions in median time to readiness for discharge by 56.3 hours (from 122.0 to 65.7 hours; hazard ratio [HR], 0.58; P  =  .033), median time to discharge by 41.5 hours (from 112.8 to 71.3 hours; HR, 0.54; P  =  .010), and time to discontinuation of IV opioids by 50.5 hours (from 104.0 to 53.5 hours; HR, 0.58; P  =  .026), compared with placebo.³²  Sevuparin, a low-molecular- weight heparin-derived polysaccharide with antiadhesive properties but minimal anticoagulant activity, did not shorten the times to VOC resolution or discontinuation of IV opioids vs placebo when administered during acute pain episodes.³³ 

Poloxamer-188, a nonionic block copolymer surfactant that improves microvascular blood flow and reduces hydrophobic cell-cell interactions, has been evaluated in patients hospitalized for acute pain episodes. In an earlier study, poloxamer-188 significantly shortened the duration of acute pain episodes when compared with placebo (mean, 133 hours [standard deviation, 41] vs 141 hours [standard deviation, 42]; P  = .04).³⁴  However, with an absence of documentation of the study's crisis resolution criteria in approximately 24% of participants (30% in the placebo arm vs 18% in the poloxamer-188 arm), another trial was conducted. This subsequent study showed no significant difference in the time to discontinuation of IV opioids when poloxamer-188 was compared to placebo (81.8 hours vs 77.8 hours; P  = .09).³⁵ 

Multiple agents have been evaluated to mitigate the effects of inflammation in SCD.³⁶  Regadenoson, a partially selective adenosine A_2A receptor agonist that decreases the activation of invariant natural killer T cells, did not decrease the duration of hospitalization (3.96 days vs 3.99 days; P  =  .80), total opioid use (median morphine equivalent dose, 0.03  mg/kg/h vs 0.04   mg/ kg/h; P  = .34), or pain scores (−2.68 vs −2.80; P = .91) when compared with placebo.³⁷  Treatment with SC411, a docosahexaenoic acid ethyl ester formulation, for 8 weeks significantly reduced levels of D-dimer (P  = .025) and soluble E-selectin (P  = .0219) and increased Hb (mean change, 0.97  g/dL vs 0.33  g/dL; P  = .04) when compared with placebo.³⁸  Canakinumab (ACZ885), a monoclonal anti–interleukin 1 beta antibody, was well tolerated in a 6-month study. Although the trial did not achieve its prespecified primary end point for diary-reported daily pain scores, treatment with canakinumab resulted in reductions in markers of inflammation (high-sensitivity C-reactive protein, absolute counts of leukocytes, monocytes) and number/duration of hospitalizations as well as trends for improvement in pain intensity, fatigue, and absences from school or work when compared with patients in the placebo arm.³⁹  Montelukast, a leukotriene inhibitor, did not significantly decrease levels of soluble vascular cell adhesion molecule 1 and reported pain when compared to placebo following 8 weeks of treatment.⁴⁰  Actively recruiting studies of antioxidants, antiadhesive and anti-inflammatory agents are shown in Table 3.

Abnormalities of the nitric oxide (NO)-cyclic guanosine monophosphate (cGMP)-dependent signaling pathway may play a role in the inflammation and vascular dysfunction seen in SCD. As a result of ongoing hemolysis, scavenging of NO, and subsequent endothelial dysfunction, NO and related agents may provide benefit to patients with SCD. In a phase 2 multicenter study of SCD patients experiencing acute vaso-occlusive episodes, inhaled NO did not significantly shorten the median time to resolution of vaso-occlusive episodes (73 hours [95% CI, 46.0-91.0] vs 65.5 hours [95% CI, 48.1-84.0]; P = .87) or the median length of hospitalization (4.1 days [IQR, 2-6] vs 3.1 days [IQR, 1.7-6.4]; P  = .30) or reduce the median opioid usage (2.8  mg/kg [1.4-6.1] vs 2.9  mg/kg [1.1-9.9]; P  = .73) or the rate of ACS when compared to placebo.⁴¹  In a separate study, inhaled NO did not reduce the rate of treatment failure in adult patients with mild to moderate ACS.⁴²  L-arginine is an obligate substrate for NO production. In a multicenter, double-blind, placebo-controlled study of children in Nigeria, oral L-arginine therapy administered within 6 hours of presentation for a pain crisis significantly reduced total analgesic usage, quantified using the mean analgesic medication quantification scale (73.4 [95% CI, 62.4-84.3] vs 120.0 [96.7-143.3]; P < .001), pain scores (1.50 [1.23-1.77] vs 1.09 [0.94-1.24]; P  = .009), time to crisis resolution (75.8 ± 36 hours [95% CI, 63.4-88.2] vs 93.3 ± 32.7 hours [95% CI, 81.7-104.9]; P  = .02), and the total length of hospital stay (105 hours [IQR, 72-144] vs 141 hours [IQR, 117-205]; P  = .002).⁴³  However, patients treated with sildenafil, a phosphodiesterase-5 inhibitor that increases NO-mediated effects by inhibiting cGMP degradation, experienced more serious adverse events, predominantly hospitalization for pain, but no clinical benefits when compared to placebo.⁴⁴ 

Platelet activation occurs in SCD at steady state, with further activation during acute pain episodes.^45-49  Although ticlopidine was previously shown to decrease the number, the mean duration, and the severity of acute pain episodes,⁵⁰  more recent phase 3 trials of the newer-generation P2Y₁₂ receptor blockers, prasugrel and ticagrelor, did not show a benefit in reducing the frequency of vaso-occlusive episodes compared to placebo.^51,52  Tinzaparin, a low-molecular-weight heparin, significantly reduced the durations of acute pain episodes (mean difference in duration of painful crises, −1.78 days; 95% CI, −1.94 to −1.62; P < .0001) and hospitalization (mean difference in duration of hospitalization, −4.98 days; 95% CI, 5.48 to −4.48; P < .0001) when compared to placebo.⁵³  However, it is uncertain whether the beneficial effects were a result of its anticoagulant or antiadhesive effects. Table 4 shows actively recruiting trials of NO and related agents, antiplatelet agents, and anticoagulants.

As most patients have limited access to curative therapies, pharmacotherapy may offer the best hope for improved patient outcomes at this time. In the absence of clinical trials comparing available drugs, the choice of initial therapy may be guided by a patient's clinical presentation as well as the availability and cost of drugs (Table 5). Patients with frequent vaso-occlusive complications (such as acute pain episodes or ACS) may obtain benefit from the use of hydroxyurea, L-glutamine, and crizanlizumab, while hemolytic anemia may be improved with the use of hydroxyurea and voxelotor. Despite the negative results of the STAND trial, we continue to use crizanlizumab on a case-by-case basis as several studies other than the SUSTAIN trial suggest a benefit to decreasing the frequency of painful episodes leading to health center visits.^54-56  As approved drug therapies have limited clinical efficacy, most complications related to SCD are unlikely to be ameliorated by a single drug. Consequently, patients are most likely to obtain maximum benefit using a combination of drugs with different mechanisms of action and nonoverlapping side effects. While more data are necessary to evaluate the effects of drug combinations, previous studies of L-glutamine, crizanlizumab, and voxelotor showed that these agents in combination with hydroxyurea were beneficial, without increased toxicity.

In the absence of identifiable precipitating factors and following confirmation of adherence with hydroxyurea, other treatment options for acute pain episodes were extensively discussed. While continuing hydroxyurea, the patient began on L-glutamine because he wished to avoid monthly infusion clinic visits. He was, however, switched to crizanlizumab due to poor tolerance of L-glutamine. He experienced a substantial reduction in the frequency of acute pain episodes over the next year.

Although SCD is an orphan disease in the United States, it is common worldwide. With advances in the understanding of disease pathophysiology, multiple drugs have been approved by regulatory agencies, with more in various stages of clinical testing. The development of new drugs for SCD offers opportunities to test drug combinations in the hope of improved clinical outcomes. Although the majority of drug trials in SCD have evaluated acute pain episodes as the primary clinical end point, other SCD- related complications and surrogate end points are increasingly being assessed. Demonstrating the benefit of drug therapies on end-organ dysfunction in SCD will provide further evidence for their role in improving patient outcomes.

Kenneth I. Ataga is supported by awards from the FDA (R01FD006030) and NHLBI (HL159376).

Parul Rai: consultancy: Global Blood Therapeutics.

Kenneth I. Ataga: research funding: Novartis, Novo Nordisk, Takeda Pharmaceuticals; advisory board member: Novartis, Novo Nordisk, Fulcrum Therapeutics, Agios Pharmaceuticals, Pfizer; consultancy: Roche, Biomarin; data monitoring committee: Vertex.

Parul Rai: nothing to disclose.

Kenneth I. Ataga: nothing to disclose.