Stroke prevention in Hispanic children with sickle cell anemia: the SACRED trial

Stroke is a common and devastating clinical complication of sickle cell anemia (SCA) that results in serious physical and neurocognitive morbidity, and often leads to early mortality.¹ Natural history studies from the United States and Jamaica have documented that 5% to 10% of untreated children with SCA will develop primary stroke during childhood,^2,3 with a cumulative incidence of 0.76 events per 100 patient-years in the first 20 years of life.⁴ 

Transcranial Doppler (TCD) ultrasonography measures intracranial arterial blood flow velocities and has strong predictive value for identifying children with SCA at highest risk for primary stroke. The maximum time-averaged mean velocity (TAMV) is recorded in the major arteries in both cerebral hemispheres, and the maximum TAMV defines the stroke risk: normal velocities (<170 cm/s) are associated with the lowest stroke risk, whereas conditional velocities (170-199 cm/s) and abnormal TCD velocities (≥200 cm/s) are associated with increased risk and highest risk, respectively.⁵ Based on the landmark National Heart, Lung, and Blood Institute–funded stroke prevention trial in sickle cell anemia (STOP) trials, children with abnormal TCD velocities should receive indefinite monthly blood transfusions to reduce their primary stroke risk,^6,7 although the National Heart, Lung, and Blood Institute–funded TCD with Transfusions Changing to Hydroxyurea trial later demonstrated that hydroxyurea at maximum tolerated dose (MTD) provides effective primary stroke prophylaxis in the setting of abnormal TCD velocities.⁸ Currently, children with conditional TCD velocities do not have a clear treatment recommendation, despite having a substantial risk of conversion to abnormal TCD velocities and even stroke over time.^9,10 

Annual TCD screening is now recommended as standard of care for children with SCA in the United States and other developed countries,^11-14 in which implementation of serial TCD screening has been effective at reducing the incidence of primary stroke in this high-risk patient population.^15,16 In low-resource settings in which the burden of SCA is highest, TCD examinations are recommended¹⁴ but there are few screening programs because of lack of equipment and trained personnel, plus potential challenges with administering monthly blood transfusions or hydroxyurea as stroke prophylaxis. In Nigeria, TCD screening with successful hydroxyurea treatment for children with abnormal TCD velocities has been reported,^17,18 and 2 recent prospective trials have documented the effectiveness of hydroxyurea in this high-risk patient population with abnormal velocities.^19,20 

In the Dominican Republic, a Hispanic Caribbean Island nation with a large sickle cell population and limited blood transfusion resources, many children with SCA experience clinically overt strokes. At the request of the Dominican Republic medical leadership, we developed a North-South global research partnership²¹ to build local capacity for TCD screening and primary stroke prevention, with the goal of establishing an effective stroke screening program. We designed a prospective research trial in close collaboration with Dominican pediatric hematologists that featured local capacity building and fostered training among young local medical graduates for TCD proficiency, hydroxyurea treatment, and overall clinical research expertise.

The Stroke Avoidance for Children in REpública Dominicana (SACRED) protocol (ClinicalTrials.gov identifier: NCT02769845) was developed with 3 main research objectives: (1) to determine the prevalence of cerebrovascular disease with elevated TCD velocities in a young Hispanic SCA population; (2) to identify clinical and laboratory correlates of TCD velocities; and (3) to obtain longitudinal data on the trajectory of elevated TCD velocities, and specifically to measure the effects of hydroxyurea treatment at MTD in Hispanic children with conditional TCD velocities.²² We previously described enrollment and early results of the SACRED trial in an American Society of Hematology global health capacity building showcase,²³ and now provide the results of extended treatment, highlighting the salutary and sustained neuroprotective effects and other clinical benefits of hydroxyurea at MTD in this Hispanic SCA cohort.

The rationale and protocol design of the SACRED trial have been published previously.²² Briefly, the SACRED trial was a prospective screening and treatment study that included systematic TCD evaluations in a large cohort of Hispanic children with SCA, with annual longitudinal evaluations and hydroxyurea treatment at MTD for children with conditional TCD velocities. The SACRED trial was conducted at the Hospital Infantil Dr. Robert Reid Cabral, a large public hospital located in Santo Domingo, the capital city of the Dominican Republic. Regulatory approvals were obtained by Centro Nacional de Investigaciones en Salud Materno Infantil, Consejo Nacional de Bioetica en Salud, and Cincinnati Children’s Hospital Medical Center institutional review board. After written informed consent by a parent or legal guardian, children aged 3 to 15 years with confirmed SCA received a thorough medical history, physical examination, and laboratory studies. TCD examinations were performed by trained and certified local medical graduates using published STOP criteria as described.²³ Funding for local personnel salaries and the hydroxyurea study treatment was supplied by the Cincinnati Children’s Research Foundation.

Children with normal or inadequate TCD velocities at initial screening were assigned to the observation arm. Children with conditional TCD velocities who were not on a regular transfusion regimen for prior stroke or abnormal TCD were eligible to begin open-label hydroxyurea on the treatment arm, beginning with fixed dose (20 mg/kg per day) for 6 months followed by escalation to MTD. Children with abnormal TCD velocities were offered blood transfusions per local standard of care, and children already receiving chronic transfusions were continued per local preference, with annual TCD examinations. Children initially assigned to the observation arm had repeat screening at month 12 and month 24, and were then eligible for hydroxyurea if their repeat TCD velocities converted into the conditional category and they were not on a transfusion regimen. After initiating hydroxyurea treatment, children had TCD exams performed at 6-month intervals. Brain magnetic resonance imaging and angiography (MRI/MRA) imaging was also performed at baseline and after 2 years of hydroxyurea treatment per protocol.

The primary objective of the SACRED trial was to screen a large convenience cohort of Dominican children with SCA for stroke risk using TCD and determine the prevalence of elevated (conditional and abnormal) TCD velocities in this cross-sectional sampling. Clinical and laboratory data were collected to identify correlates of TCD velocities including anemia, fetal hemoglobin, genetic modifiers such as glucose-6-phosphate dehydrogenase (G6PD) deficiency and α-thalassemia, as well as hydroxyurea exposure. Serial assessments provided longitudinal incidence data regarding reversion to normal TCD velocities, or conversion to abnormal TCD velocities. A common study termination date was ∼3 years from the last participant’s treatment initiation.

The primary study outcome of the SACRED trial was the treatment response to hydroxyurea in children with conditional TCD velocities, and specifically the change in maximum TAMV after 24 months of treatment. Secondary study outcomes included (1) effects of hydroxyurea on TCD velocity categories, (2) laboratory and clinical effects of hydroxyurea at MTD, (3) natural history of TCD velocity conversion without treatment, (4) laboratory correlations with TCD velocities, (5) incidence of neurological and other serious sickle-related events, (6) presence of brain MRI/MRA lesions, and (7) treatment-related complications.

Children with conditional TCD velocities were eligible for protocol-directed hydroxyurea treatment. For the first 6 months, hydroxyurea-naïve participants were administered a fixed dose of 20 mg/kg per day. During this phase, children had monthly study visits that included interval medical history, physical examination, and laboratory evaluation with automated complete blood count with leukocyte differential and manual reticulocyte count. Treatment remained at this fixed dose for the first 6 months in all participants, to allow initial assessment of laboratory and clinical adverse events (AEs), and included a repeat TCD examination at the end of that treatment phase. Hydroxyurea was then escalated every 8 weeks to MTD defined as mild marrow suppression of neutrophils and reticulocytes.¹⁹ Children already on hydroxyurea treatment by the clinic staff were eligible to have their hydroxyurea dose optimized through early escalation to MTD.

An online hydroxyurea dosing calculator was provided by the Medical Coordinating Center and Data Management Center to assist the site in achieving the target dose for each participant. Laboratory results and dosing data were reviewed by the local medical team at each scheduled clinic visit, and minor dose adjustments were made as necessary based on weight changes and blood count trends to maintain the laboratory benefits and ensure participant safety. Clinical AE and serious AE (SAE) reporting, dose-limiting toxicities using published criteria,²⁴ and all potential new neurological events were collected at interval clinic visits.

The importance of local capacity building in the SACRED trial was highlighted previously.²³ Briefly, local medical school graduates were invited to join the study team for at least 12 months; they received rigorous formal training and certification in TCD examinations from US-based STOP superexaminers with >20 years of experience, as well as laboratory sample collection and good clinical research practices. Over the clinical trial period from 2016 to 2022, a total of 10 Dominican medical school graduates were trained and certified.

Study data were collected using the Research Electronic Data Capture platform. There was no formal sample size or power calculation; instead, the trial was designed to investigate the changes in TCD velocity descriptively in this convenience sample of available children in the pediatric clinic. Descriptive analyses of TCD screening were performed to describe the study population. Comparative analysis between groups was investigated using a χ² test. For patients on hydroxyurea treatment, the highest TAMV for each time-point along with the baseline values was analyzed using repeated measures analysis of variance (multivariate analysis of variance). Baseline laboratory values were compared with on-treatment values for hemoglobin, fetal hemoglobin, absolute neutrophil count, absolute reticulocyte count, and platelet count using descriptive statistics, Wilcoxon signed-rank test, and comparative t tests. For comparisons between fixed-dose hydroxyurea and the MTD phase dosing, AE rates were analyzed using the incidence rate ratio (IRR) with 95% confidence intervals (CIs). Genetic modifiers including G6PD deficiency and α-thalassemia trait were tested for associations with TCD velocities and categories.

A total of 283 children enrolled in the SACRED trial between July 2016 and July 2017 (Figure 1). The average age at enrollment was 8.7 ± 3.4 years, and 130 (46%) were female. The study population had substantial morbidity at the time of enrollment; most of the participants (65%, 186/283) had been hospitalized at least 5 times previously for SCA-related complications, and most (89%, 251/283) had received at least 1 blood transfusion. At enrollment, 64 children (22.6%) were receiving hydroxyurea, 47 (16.6%) were receiving regular transfusions, and another 11 (3.9%) were receiving both hydroxyurea and transfusions. The remaining 161 children (56.9%) were receiving no disease-modifying therapy. A total of 20 children (7.0%) had a previous clinically overt stroke and most of those children were receiving monthly transfusion prophylaxis.

As shown in the CONSORT diagram (Figure 1), baseline TCD examinations included 200 normal velocities (70.7%), 63 conditional velocities (22.3%), 11 abnormal (3.9%), and 9 inadequate examinations (3.2%). Because the emphasis in the SACRED trial was on stroke prevention in high-risk children, those with conditional TCD velocities were initially compared with those with normal examinations. Among treatment-naïve children, younger age and lower hemoglobin concentration correlated with higher TCD velocities, whereas there were no significant associations with oxygen saturation, fetal hemoglobin, neutrophil count, or genetic modifiers including α-thalassemia trait or G6PD deficiency (Table 1). Among all children with conditional velocities, these same associations were identified (supplemental Table 1).

Among the 63 children with conditional velocities, 13 remained on transfusions, 1 withdrew, and another died before starting treatment, leaving 48 children eligible to receive hydroxyurea for stroke prevention. However, children with initial normal TCD examinations then underwent annual screening and had conditional TCD velocities identified. This process added 14 additional children at year 1 and another 12 at year 2, for a total of 74 children who received hydroxyurea per protocol (Figure 1). Among these, 57 were treatment naïve whereas 17 were already receiving hydroxyurea at an average of 17.2 ± 3.5 mg/kg per day.

The initial dose of hydroxyurea administered to the treatment-naïve cohort was 20.5 ± 1.1 mg/kg per day (Table 2). Dose escalation then began after month 6; most children (85%, 63/74) reached MTD based on mild myelosuppression at an average dose of 23.8 ± 6.0 mg/kg per day, after a treatment duration of 10 ± 4 months. Adherence to medication was assessed at each scheduled visit, and >90% of assessments documented no missed doses per caregiver report, which was further evidenced by observable myelosuppression in the laboratory values.

Baseline laboratory values in the hydroxyurea-naïve cohort revealed anemia with expected leukocytosis and reticulocytosis (Table 2). After 6 months of fixed-dose hydroxyurea, participants had robust increases in hemoglobin and fetal hemoglobin, and expected decreases in neutrophils and reticulocytes. After 24 months of treatment, the average dose had increased to 25.7 ± 5.8 mg/kg per day, with further increases documented in hemoglobin and fetal hemoglobin compared with the fixed-dose values. Conversely, the white blood cell count, absolute neutrophil count, and absolute reticulocyte count were significantly decreased, reflecting the intended mild bone marrow suppression. Extended treatment was offered to all study participants receiving hydroxyurea, and these laboratory changes were sustained even after 60 months of treatment (supplemental Table 2).

All participants with pretreatment conditional velocities underwent TCD examinations every 6 months. Among the hydroxyurea-naïve patients with a month-6 TCD exam, 34 of 56 (61%) had normalization of their velocities after completing fixed-dose hydroxyurea treatment, with the average TCD velocity decreasing by 20 cm/s from 180 to 160 cm/s (Table 3). After 24 months of hydroxyurea treatment, these lower TCD velocities were sustained, with normal values observed in two-thirds of the children. After 5 years of treatment, the average TCD velocity had decreased further to 153 cm/s, with 81% reverting to normal TCD velocities (Table 3), but 1 child converted to an abnormal TCD velocity and had new mild MRA vasculopathy. Similar results were observed for the entire cohort of children who received hydroxyurea treatment (supplemental Table 3). No significant associations were observed between the decline in TCD velocities and G6PD deficiency or α-thalassemia trait.

Baseline brain MRI studies were performed within 3 months of initiating hydroxyurea in most treatment-naïve children; 40% (19/48) had abnormalities in the brain parenchyma, typically silent infarctions in the frontal lobes or deep white matter. Paired MRI examinations in 41 children after 2 years of hydroxyurea therapy identified new vasculopathy in 4 children (Table 3). Among all children treated with hydroxyurea, 50 children had paired exams; 1 child developed a new silent infarct whereas 7 had new vasculopathy identified (supplemental Table 3).

The overall rate of clinical AE was 1.6 per patient-year, but most of these were mild grade 2 events (Table 4). Reductions in event rates over time were noted for all AEs, sickle-related AEs, vaso-occlusive crisis, and infections (Figure 2). Compared with fixed-dose hydroxyurea, dose escalation to MTD led to fewer clinical AEs with IRR of 0.56 (95% CI, 0.42-0.76; P = .00017). Similarly, sickle-related events decreased with IRR of 0.59 (95% CI, 0.36-0.98; P = .042) whereas infections also decreased with IRR of 0.53 (95% CI, 0.38-0.74; P = .00018).

A total of 37 SAEs occurred during treatment, for an overall rate of 0.1 event per patient-year, including 1 death and 1 stroke on hydroxyurea treatment. The death occurred in an 8-year-old female at month 12 of treatment due to Salmonella sepsis, whereas the stroke occurred in a 10-year-old male at month 14, for a calculated stroke rate of 0.29 per 100 patient-years. Of note, a stroke also occurred in a child on the observation arm with normal TCD velocity, whereas another occurred in a child on chronic transfusions. Dose-limiting toxicities occurred in ∼10% of blood counts, with an overall rate of 0.9 per patient-year. Laboratory toxicities were identified in all cell lines; however, anemia and reticulocytopenia were relatively overrepresented compared with neutropenia and thrombocytopenia (Table 4).

In this prospective research trial enrolling Hispanic children with SCA living in the Dominican Republic, we documented the feasibility of establishing a robust TCD screening program for identifying patients at high risk for primary stroke, and then providing hydroxyurea at MTD to decrease that stroke risk. After building local capacity for TCD examinations through rigorous training and formal certification of new medical graduates, TCD evaluations in this pediatric cohort initially documented 22% conditional velocities (Table 1), and a striking 39% cumulative incidence of conditional velocities after 3 annual examinations over a 24-month screening period (Figure 1). These results are higher than previously reported from US-based cohorts,^5,10 and support the claims by local hematologists of a high stroke rate among Dominican children with SCA, perhaps because of biological, genetic, or ethnic differences. In this high-risk population, hydroxyurea treatment was effective as primary stroke prophylaxis, with ∼60% of children with initial conditional TCD velocities reverting to normal after 6-months of fixed-dose hydroxyurea at 20 mg/kg per day, with almost 80% reverting to normal after dose escalation to MTD, with sustained treatment benefits over a 5-year treatment period (Table 3).

The presence of SCA in the Dominican Republic originated through the transatlantic slave trade that brought many thousands of Africans to the island of La Hispaniola, which was later divided into the Dominican Republic and Haiti.²⁵ Recent newborn screening efforts in adjacent Haiti and other islands have documented 7% to 12% sickle trait and a birth incidence of sickle cell disease ranging from 1:120 to 1:400 across the Caribbean.^26,27 There are no accurate statistics from the Dominican Republic itself, but even 4% to 5% sickle trait would lead to >500 affected births per year. Because of the breadth and volume of the slave trade, as well as subsequent immigration, SCA now exists across Latin America with an estimated 6000 annual births,²⁸ and there is a growing Hispanic SCA cohort in many parts of the United States and especially in California, Texas, and New York.^28-30 

Despite this increasing population of Hispanic children with SCA, little is known about the potential genetic and environmental differences that might affect their phenotype. One study evaluated the baseline fetal hemoglobin levels and the hydroxyurea treatment responses in Hispanic and non-Hispanic children with SCA, and found no significant differences,³¹ whereas another reported that an elevated white blood cell count was especially deleterious in this patient population.³² Hispanic children with SCA may have a higher proportion of acute chest syndrome and overall pulmonary disease,³³ and a recent analysis from the United States National Inpatient Sample database reported a significantly higher mortality rate.³⁴ Whether Hispanic children with SCA have similar stroke risk and treatment responses to hydroxyurea compared with non-Hispanic children with SCA has, to our knowledge, not been addressed previously.

In the SACRED trial, we documented a high cumulative incidence of conditional TCD velocities over a 2-year screening period and then a robust treatment response to hydroxyurea at MTD. Reduction in TCD velocities during hydroxyurea treatment has been demonstrated previously in prospective trials using low and moderate doses, including 3 trials in Nigeria.^18-20,35,36 Hydroxyurea at MTD decreased the maximum TCD velocity by an average of 26 and 29 cm/s in 2 US-based studies, respectively,^37,38 as well as 24 cm/s in Jamaica³⁹ and 33 cm/s in Tanzania.⁴⁰ The American Society of Hematology guidelines reported a pooled analysis with an average decline in TCD velocity of 21 cm/s (95% CI, 15-29),¹⁴ and our children in the Dominican Republic had a similar excellent response with initial decrease of 20 cm/s and a sustained average decrease of 27 cm/s after 5 years of hydroxyurea treatment (Table 3). Together, these data confirm the salutary effects of hydroxyurea to lower TCD velocities, and also for hydroxyurea at MTD to provide sustained benefits with 81% reversion to normal velocities and reduced stroke risk.

The hydroxyurea dose was safely escalated to MTD values that were similar to other treatment studies in the United States, with only transient cytopenia (Table 4). Hematological benefits were evident with significant increases in hemoglobin and fetal hemoglobin, as well as mild myelosuppression with decreased white blood cell counts, neutrophil counts, and reticulocyte counts. Clinical responses were excellent, and dose escalation to MTD provided significant reductions in the rates of all AE, and especially rates of sickle-related AEs and infections, compared with fixed-dose hydroxyurea (Figure 2). No SAEs were related to hydroxyurea treatment, including the 1 death in the treatment cohort due to Salmonella sepsis. Two children converted to an abnormal TCD velocity during treatment, and 1 child had a clinical stroke (0.29 events per patient-year). In paired analysis, there was 1 child who developed a new silent infarct on brain MRI and 7 with additional vasculopathy (supplemental Table 3), which emphasizes the need to begin preventive therapy early in life, before the onset of cerebrovascular disease. Among the children in the SACRED trial not receiving hydroxyurea, 2 children had a clinical stroke (1 with a normal TCD velocity on no treatment, and 1 with an abnormal velocity on transfusions), and another died after a hemolytic transfusion reaction.

Access to TCD in the Dominican Republic is currently limited because of the cost and availability of screening, so is primarily performed on patients in the private sector. In preparation for the SACRED trial, we used superexaminers from the United States to train and certify 3 local medical graduates who became proficient in TCD examinations and other clinical research skills; over the duration of the trial a total of 10 graduates were trained and certified. This kind of capacity building is efficient and scalable, and expanding the existing TCD resources could begin to address the high burden of clinical stroke in the SCA population across the Caribbean. The CAREST (Caribbean Network of Researchers on Sickle Cell Disease and Thalassemia) network that spans the Caribbean would provide an ideal vehicle to facilitate such training.⁴¹ 

The identification of children with elevated TCD velocities conferring risk of primary stroke has important public health implications for the Dominican Republic, as does the efficacy of hydroxyurea in reducing cerebrovascular flow velocities and reducing stroke risk. As the primary disease modifying treatment for SCA, hydroxyurea has consistently demonstrated both laboratory and clinical efficacy, including significant reduction in TCD velocities. Hydroxyurea also provides an alternative to blood transfusions for stroke prophylaxis, which is especially important in areas of the world in which the blood supply is inconsistent and/or potentially unsafe.⁴² Hydroxyurea is cost-effective in low-resource settings,⁴³ and the need for affordable access to hydroxyurea was highlighted in the recent Lancet Commission on Sickle Cell Disease.⁴⁴ With increased usage, the cost of hydroxyurea has become more affordable in the Caribbean, now $0.28 USD per 500-mg capsule in the Dominican Republic and $0.10 USD in Jamaica (M. Reid, oral communication, 7 November 2024).

In a resource-limited setting such as the Dominican Republic, in which the burden of SCA is high and morbidity is substantial, hydroxyurea offers an attractive treatment option. Research partnerships can be developed to facilitate training and technology transfer,⁴⁵ and have been successfully conducted in Latin America.^46,47 

Limitations of the SACRED trial include the convenience sample with lack of sample size and power calculations, and the inclusion of children with previous disease-modifying treatment. Strengths include building local capacity for prospective epidemiological data regarding TCD screening, stroke risk, and hydroxyurea treatment effects for Hispanic children with SCA living in the Dominican Republic. Our results document the benefits of hydroxyurea to reduce primary stroke risk in this particular population and have the potential to be generalized to other resource-limited settings across Latin America. Using innovative teaching methods and collaborative, strategic research infrastructure, the SACRED paradigm is an example of an effective method to build local capacity, optimize clinical trials management, and improve outcomes for Hispanic children with SCA. This approach can be used in other settings in the Caribbean with a large burden of SCA.

The authors acknowledge the study participants and their families for their time and effort to join the Stroke Avoidance for Children in República Dominicana (SACRED) trial. The authors also acknowledge the many individuals who accommodated the study at Hospital Infantil Robert Reid Cabral and Centro de Obstetricia y Ginecología.

The authors thank the SACRED Data Management Center (John Boesing, Justin McAdams, Rebecca Geer, and Amanda Pfeiffer), as well as the Cincinnati Children’s Research Foundation for study funding.

Contribution: R.M.N. and R.E.W. designed the trial; T.L., W.H.S., and A.C.L. participated in data curation and analysis; N.M., M.B., L.M.S., G.U., C.F., C.G., P.d.V., S.C., D.R., P.R., M.M., and E.M. participated in patient management and data collection; T.L., N.D.J., S.E.S., and R.M. helped coordinate the trial; and all authors reviewed and approved the final draft before submission.

Conflict-of-interest disclosure: R.E.W. has received research donations from Bristol Myers Squibb, Theravia, and Hemex Health; has participated on a medical advisory board for Merck and Novo Nordisk; and is a member of a data safety and monitoring board for Novartis and Vaxart. The remaining authors declare no competing financial interests.

Paola del Villar died on 14 August 2022.

Correspondence: Russell E. Ware, Pediatric Hematology, Cincinnati Children’s Hospital Medical Center, 3333 Burnet Ave, Hematology ML 7015, Cincinnati, OH 45229; email: russell.ware@cchmc.org.