Stroke Prevention with Hydroxyurea Enabled through Research and Education: A Phase 2 Primary Stroke Prevention Trial in Sub-Saharan Africa

Abstract

Introduction

Stroke is a severe complication of sickle cell anemia (SCA), with devastating sequelae. Transcranial Doppler (TCD) ultrasonography predicts stroke risk, but implementing TCD screening with suitable treatment for primary stroke prevention in low-resource environments remains challenging. SPHERE (NCT03948867) is a prospective phase 2 open-label hydroxyurea trial for SCA in Tanzania.

Methods

After formal training and certification, local personnel screened children 2–16 years old; those with conditional (170–199 cm/s) or abnormal (≥200 cm/s) time-averaged mean velocities (TAMVs) received hydroxyurea at 20 mg/kg/day with dose escalation to maximum tolerated dose (MTD). The primary study endpoint is change in TAMV after 12 months of hydroxyurea; secondary endpoints include SCA-related clinical events, splenic volume and function, renal function, infections, hydroxyurea pharmacokinetics, and genetic modifiers.

Results

Between April 2019 and April 2020, 202 children (average 6.8 ± 3.5 years, 53% female) enrolled and underwent TCD screening; 196 were deemed eligible by DNA testing. Most had numerous previous hospitalizations and transfusions, with low baseline hemoglobin (7.7 ± 1.1 g/dL) and %HbF (9.3 ± 5.4%). Palpable splenomegaly was present at enrollment in 49 (25%); average sonographic splenic volume was 103 mL (range 8–1,045 mL). TCD screening identified 22% conditional and 2% abnormal velocities, with hydroxyurea treatment initiated in 96% (45/47) eligible children.

Conclusion

SPHERE has built local capacity with high-quality research infrastructure and TCD screening for SCA in Tanzania. Fully enrolled participants have a high prevalence of elevated baseline TCD velocities and splenomegaly. SPHERE will prospectively determine the benefits of hydroxyurea at MTD for primary stroke prevention, anticipating expanded access to hydroxyurea treatment across Tanzania.

Introduction

Sickle cell anemia (SCA) is an inherited hematological disorder characterized by chronic hemolytic anemia, recurrent acute vaso-occlusive events, cumulative multisystem organ damage, and premature death [1]. Stroke is a common and catastrophic event for children with SCA. In Jamaica, 7.8% of children with SCA developed clinical stroke by 14 years old [2]. In the US Cooperative Study of Sickle Cell Disease, 11% of children with HbSS had a stroke by 20 years of age with the greatest risk between ages 2–5 years when overall incidence reaches 0.75 per 100 patient-years [3]. Before the introduction of disease-modifying therapy with blood transfusions or hydroxyurea, stroke recurrence was common [4] and often caused disability and death [5].

Stroke risk is one of the few predictable aspects of SCA disease management, as the velocity of cerebral arterial blood flow can be measured noninvasively with transcranial Doppler (TCD) ultrasonography to assess the risk of developing primary stroke [6]. Children with elevated TCD velocities have a higher risk of stroke and can be prescribed more aggressive treatment to decrease the risk of a future neurological event. Particularly for children with abnormal TCD velocities (≥200 cm/s), regular transfusions decrease stroke risk by improving anemia and replacing endogenous sickle erythrocytes with healthy erythrocytes [7]. The use of systematic TCD screening and transfusion prophylaxis has drastically lowered the incidence of first stroke in high-income countries [8].

SCA is most common in sub-Saharan Africa [9] where mortality is extremely high [10], and the consequences of stroke are more severe [11, 12]. Blood supply and transfusion services in Africa are currently inadequate to support a safe and effective chronic transfusion program for SCA [13, 14]. Hydroxyurea is an attractive alternative for both primary and secondary stroke preventions. The NOHARM and REACH trials have documented the benefits of hydroxyurea at maximum tolerated dose (MTD) for children with SCA living in Africa [15, 16], and recently the NOHARM MTD trial was halted prematurely when a higher dose of ∼30 mg/kg/day was superior to a lower fixed dose of ∼20 mg/kg/day without increased toxicity [17]. Challenges remain, however, about the most practical approach for implementing an integrated TCD screening and treatment program in Africa, specifically for primary stroke prevention.

Stroke Prevention with Hydroxyurea Enabled through Research and Education (SPHERE, ClinicalTrials.gov NCT03948867) is a prospective, phase 2 open-label dose escalation trial of hydroxyurea to evaluate the feasibility and benefits of hydroxyurea escalated to MTD for primary stroke prevention in children with SCA living in Tanzania, while also enhancing local research capacity. We now describe the study design and baseline enrollment characteristics of SPHERE, which will determine the efficacy of hydroxyurea in this setting.

Materials and Methods

Clinical Performance Site

SPHERE is conducted at Bugando Medical Centre (BMC), a 900-bed referral and teaching hospital in northwest Tanzania, on the shores of Lake Victoria. Mwanza is the second largest urban center in Tanzania with approximately 1 million people in Mwanza city and 3 million people in the Mwanza region. As one of five national referral hospitals, BMC provides tertiary care for 16 million people. Northwest Tanzania has the highest prevalence of SCA in the country, recently measured at 1.2% of all births [18].

BMC houses the only sickle cell clinic for the entire catchment area, which operates 2 days per week and cares for >700 children with SCA. Children are seen for routine visits approximately every 1–3 months depending on severity of disease. Families are provided an insecticide-treated bed net and patients are prescribed daily penicillin prophylaxis and folic acid supplementation. The national vaccine program includes Haemophilus influenza type b and 13-valent pneumococcal conjugate vaccines. Per recently published national guidelines [19], hydroxyurea is now recommended for all children with sickle cell disease, but few families can afford the medication. Blood transfusions are available for severe anemia, but chronic scheduled transfusion programs are not feasible due to a limited blood supply.

Capacity Building

A group of potential TCD examiners was invited to attend a 2-day in-person training educational seminar reviewing TCD theory and standards, along with a practical hands-on workshop where participants learned to obtain high-quality TCD measurements. This was followed by ongoing remote training. Practice exams performed locally were reviewed by an outside expert TCD examiner prior to certification. Each prospective TCD examiner was required to submit at least 20 high-quality exams before achieving formal certification. Principals of TCD techniques, common errors, and best practices were reviewed during a regular teleconference, and an annual in-person refresher workshop was conducted.

Patient Population

Children with SCA 2.0–16.0 years of age, attending the BMC clinic with a parent or guardian, were invited to join the SPHERE trial. Patients were not enrolled until 2 weeks had elapsed since any recent febrile illness, hospitalization, or blood transfusion. Enrolled participants with elevated TCD velocities, either conditional (170–199 cm/s) or abnormal (≥200 cm/s), were prescribed hydroxyurea therapy unless they had a medical condition making hydroxyurea ill-advised (e.g., malignancy, uncontrolled infection, pregnancy). Children with normal TCD velocities (<170 cm/s) will be rescreened after 12 and 24 months, to document further TCD conversions and allow more children to receive hydroxyurea treatment.

Sample Size

The sample size was calculated to detect a mean decrease in TCD velocity of 15 cm/s with a standard deviation of 30 cm/s. It was determined that 36 evaluable subjects would provide 90% power to detect this decrease from baseline to month 12 from a one-sided t test with 0.05 level of significance. To allow for dropout, the study aimed to enroll approximately 200 children in the screening phase so that approximately 40 would be eligible for hydroxyurea therapy.

Data Collection

Baseline data included demographic characteristics, comprehensive clinical history and physical examination, immunization history, concomitant medications, and laboratory results. The clinical history was self-reported by the parent. For each participant, Z-scores were calculated for height-for-age and weight-for-height using published WHO growth metrics [20]. Data were entered into an Internet-based Research Electronic Data Capture (REDCap^TM) database [21, 22]. Primary source documents were uploaded into REDCap^TM to allow remote study monitoring.

Study Treatment

Hydroxyurea is initiated at 20 mg/kg/day, averaged over the course of a week in the form of available 500 mg capsules, using an online dosing calculator. Laboratory values are checked monthly, and hydroxyurea is escalated every 8 weeks to achieve mild myelosuppression, when the following escalation laboratory criteria are met: absolute neutrophil count (ANC) >4.0 × 10⁹/L, Hb >6.5 g/dL, absolute reticulocyte count (ARC) >150 × 10⁹/L, and platelets >150 × 10⁹/L. The MTD is determined when consistent mild myelosuppression is achieved on 2 consecutive visits, with a maximum dose of 35 mg/kg/day. Children return every 3 months after reaching a stable MTD dose for a medication refill. Hydroxyurea treatment is temporarily suspended if any of the following dose-limiting toxicities occur: ANC <1.0 × 10⁹/L; Hb <4.0 g/dL; Hb <6.0 g/dL with an ARC <100 × 10⁹/L; ARC <80 × 10⁹/L with Hb <7.0 g/dL; or platelets <80 × 10⁹/L. If the toxicity resolves within 2 weeks, the same dose is restarted. If it persists beyond 2 weeks, hydroxyurea is restarted at a lower dose after the toxicity resolves, as previously described [15, 16].

Endpoints

The primary study endpoint is the change in TCD velocities from enrollment until 12 months of hydroxyurea treatment. Secondary endpoints include SCA-related clinical events; changes in splenic volume and function; change in renal function; incidence of infection, especially malaria; hydroxyurea pharmacokinetics; and genetic modifiers of disease including pharmacogenomics.

Safety

All serious adverse events and clinical adverse events, Grade 2 or above, are recorded for children receiving hydroxyurea study treatment. Those with normal enrollment TCD examinations are rescreened at 12 months and 24 months, and all neurological events and serious adverse events are recorded. A Data Safety and Monitoring Board (DSMB) that includes clinicians and researchers in Tanzania with expertise in SCA, hydroxyurea treatment, biostatistics, and global health research was formed and reviews the data every 6 months.

Laboratory

Automated complete (full) blood counts are performed using a commercial hematology analyzer (DH76; Dymind Biotech, Shenzhen, China), while ARCs are calculated by microscopy with vital staining. Quantitative measurement of fetal hemoglobin (%HbF) is performed locally using HPLC (SmartLife PolyLC, Columbia, MD, USA) and calculated as a percent of total HbF + HbS present. At baseline, whole blood was collected on FTA cards (GE Healthcare Bio-Sciences, Pittsburgh, PA, USA) and stored at −80°C until further analysis.

TCD and Splenic Ultrasound

The maximum time-averaged mean velocity (TAMV) is measured by TCD ultrasonography in each of the major intracerebral arterial segments as performed in the STOP protocol [23] using a 2-MHz nonimaging Sonara/tek instrument. Each exam is centrally reviewed and scored by an expert TCD specialist blinded to all clinical and laboratory data, as previously described [24]. The TCD exam is considered adequate if velocities can be measured in all three major vessels in at least one hemisphere. Spleen size is measured using a GE Logiq V5 Expert ultrasound with 3.5 MHz curvilinear probe. With the participant in the supine position, the maximum distance in the longitudinal (cranio-caudal), anterior-posterior, and transverse dimensions is recorded. Spleen volume is calculated using the formula for an ellipsoid shape: longitudinal length × anterior-posterior depth × transverse width × 0.523 [25].

Genetic Modifiers

FTA cards were kept frozen until transport to CCHMC, where DNA-based studies were performed to (1) confirm the diagnosis of SCA; (2) identify concomitant glucose-6-phosphate dehydrogenase (G6PD) deficiency; and (3) document the presence of alpha thalassemia trait. Genomic DNA was prepared from FTA cards using the InstaGene Matrix protocol (Bio-Rad, Hercules, CA, USA) [26]. As previously described, the diagnosis of SCA was determined using a customized TaqMan probe designed for the HbS rs334 polymorphism; the G6PD A⁻ variant was identified using three commercially available TaqMan probe sets; and detection of 3.7 kb α-globin gene deletional (−α^3.7) thalassemia trait was performed using a copy-number variant TaqMan assay with custom TaqMan probes [27].

Results

Enrollment

Enrollment occurred briskly, and 202 children underwent TCD screening, slightly exceeding the projected enrollment pace and goal (shown in Fig. 1). A total of 196 had confirmed SCA (HbSS), continued in the trial, and are included in this analysis. Figure 2 illustrates a CONSORT diagram of the study enrollment and current status of the participants.

Fig. 1.

SPHERE performed 202 screening transcranial Doppler ultrasound exams over the course of 12 months, exceeding the proposed enrollment timeline.

Fig. 2.

SPHERE enrollment CONSORT diagram. SPHERE aimed to enroll about 200 children. TCD, transcranial Doppler ultrasound.

Enrollment Characteristics

The average age (mean ± SD) at enrollment was 6.8 ± 3.5 years, and 53% were female (Table 1). Overall, the children were small and undernourished: the mean weight-for-height was −1.03 ± 1.19 for all participants and slightly lower in children with elevated TCD velocities (−1.31 ± 0.94) than normal velocities (−0.94 ± 1.25), p = 0.066. Most had previous dactylitis (77%), painful vaso-occlusive episode (93%), and malaria (88%). Previous splenic sequestration was common (35%), but only 1 (0.5%) had undergone surgical splenectomy at the time of enrollment. The majority (68%) had received at least one previous transfusion, and 31% of the children had over 5 previous hospitalizations. None of the enrolled participants had previous history of stroke, and only 7 (4%) of the children had previously used hydroxyurea.

Table 1.

Baseline characteristics of children with SCA enrolled in SPHERE

	All participants1	Normal TCD	Elevated TCD	p value
N	196	149	47
Demographics
Age at enrollment, mean±SD, years	6.8±3.5	6.9±3.6	6.3±3.2	0.288
Male, N (%)	93 (47)	71 (48)	22 (47)
Female, N (%)	103 (53)	78 (52)	25 (53)	1.000
Intake physical examination
Height, cm	111.5±18.5	112.0±19.3	110.1±15.8	0.553
Weight, kg	18.6±6.6	19.0±6.7	17.6±6.2	0.213
Splenectomy, N (%)	1 (1)	1 (1)	0 (0)	1.000
Spleen palpable, N (%)	49 (25)	36 (24)	13 (28)	0.772
Spleen palpable ≥5 cm, N (%)	32 (16)	23 (15)	9 (19)	0.146
Oxygen saturation (%), mean±SD	93.0±4.3	94.0±3.9	91.7±5.0	0.001
Enlarged tonsillar tissue, N (%)	149 (74)	108 (73)	41 (87)	0.130
Growth parameters, Z-score
Weight-for-height	–1.08±1.36	–1.01±1.46	–1.31±0.94	0.066
Height-for-age	–1.35±1.10	–1.43±1.10	–1.11±1.09	0.082
Past medical history, N (%)
Vaso-occlusive painful event	182 (93)	137 (92)	45 (96)	0.465
Dactylitis	150 (77)	118 (79)	32 (68)	0.108
Malaria	173 (88)	132 (89)	41 (87)	0.415
Acute chest syndrome/pneumonia	70 (36)	53 (36)	17 (36)	1.000
Splenic sequestration	68 (35)	54 (36)	14 (30)	0.708
Previous stroke	0 (0)	0 (0)	0 (0)	–
Bacteremia/sepsis	73 (37)	60 (40)	13 (28)	0.262
Transfusion	135 (69)	100 (67)	35 (74)	0.372
Past hydroxyurea use	7 (4)	6 (4)	1 (2)	1.000
Lifetime hospitalizations, N (%)
≤5 hospitalizations	135 (69)	105 (70)	30 (64)	0.602
6–10 hospitalizations	30 (15)	21 (14)	9 (19)
>10 hospitalizations	31 (16)	23 (15)	8 (17)
Baseline parameters, mean±SD
Hemoglobin, g/dL	7.7±1.1	7.9±1.1	7.1±1.0	<0.001
MCV, fL	78±10	79±10	82±9	0.037
Fetal hemoglobin, %	9.3±5.4	9.6±5.7	8.5±4.0	0.251
ARC, ×103/L	566±243	558±241	594±249	0.364
WBC count, ×109/L	14.5±5.4	14.2±5.8	15.3±3.7	0.251
ANC, ×109/L	5.5±2.3	5.4±2.3	5.9±2.3	0.172
Platelets, ×109/L	379±164	378±162	382±173	0.899
ALT, U/L	21±16	20±15	23±18	0.355
Creatinine, mg/dL	0.22±0.12	0.23±0.11	0.19±0.12	0.013
TCD velocities, mean±SD
TAMV, cm/s	149±25	138±18	182±12	<0.001

¹A total of 202 children were enrolled, but 6 were withdrawn following confirmatory diagnostic testing. MCV, mean corpuscular volume; WBC, white blood cell; ANC, absolute neutrophil count; ALT, alanine transferase.

Baseline laboratory results were in the expected ranges for children with SCA who are not receiving disease-modifying therapy. Hemolytic anemia was documented with a low hemoglobin concentration and compensatory reticulocytosis, as well as the expected leukocytosis (Table 1). The average baseline HbF was 9.3 ± 5.4%. Comparing children with elevated TCD velocities to those with normal TCD velocities, the average hemoglobin concentration was lower (7.1 vs. 7.9 g/dL, p < 0.001) as was the oxygen saturation measured by pulse oximetry (91.7% vs. 94.0%, p < 0.01), but other laboratory parameters were similar between the two groups (Table 1).

TCD Velocities

TCD examinations were performed in all children at enrollment, and only 2 participants (1%) had inadequate exams. Among the 194 adequate exams, the average maximum TAMV was 149 ± 25 cm/s with 76% normal, 22% conditional, and 2% abnormal velocities. Participants with an elevated TCD velocity were slightly younger than those with normal velocities (6.3 vs. 6.9 years), but this difference was not statistically significant. For those with elevated TCD velocities, the average TAMV was 182 cm/s, compared to 138 cm/s for the normal cohort (Table 1).

Splenic Measurements

Baseline assessment by physical examination revealed a palpable spleen in 49 children (25%), and over half of these (32/49) were ≥5 cm below the costal margin (Table 2). Larger spleen size by palpation correlated with a history of splenic sequestration, and a lower baseline hemoglobin, but not with age. Baseline abdominal ultrasonography documented splenic tissue in 177 children (90%), with an average splenic volume of 103 ± 124 mL (range 8–1,045 mL). The splenic volume correlated with the amount of palpable splenic tissue (p < 0.001). Splenic volume also correlated with participant height, history of splenic sequestration, lower hemoglobin concentration, lower platelet count, and fewer copies of the alpha globin gene. Almost half (39%) of participants had splenic volumes that were above the 95th percentile for height (shown in Fig. 3).

Table 2.

Associations between palpable splenomegaly and baseline clinical and laboratory parameters

Palpable splenomegaly	None	1.0–4.9 cm	5.0–9.9 cm	≥10 cm	p value
N	148	16	28	4
Demographics
Age at enrollment, mean±SD, years	6.9±3.5	5.7±3.2	6.6±3.5	7.3±3.7	0.559
Male, N (%)	71 (48)	7 (44)	12 (43)	3 (75)
Female, N (%)	77 (52)	9 (56)	16 (57)	1 (25)	0.701
Baseline examination, mean±SD
Height, cm	112.5±18.5	105.3±19.2	109.5±18.0	116.0±10.8	0.438
Weight, kg	18.9±6.7	16.8±6.7	17.9±6.2	19.6±7.7	0.589
O2 saturation, %	93.0±4.5	93.3±3.3	95.4±3.3	96.0±2.9	0.029
Medical history, N (%)
Painful event	135 (93)	16 (100)	27 (100)	4 (100)	0.423
Dactylitis	114 (78)	11 (69)	23 (82)	2 (50)	0.357
Malaria	132 (91)	14 (88)	25 (93)	2 (50)	0.094
Acute chest syndrome	51 (35)	4 (25)	14 (50)	1 (25)	0.350
Splenic sequestration	39 (31)	4 (40)	21 (84)	4 (100)	<0.001
Previous stroke
Bacteremia/sepsis	55 (43)	4 (33)	13 (54)	1 (33)	0.640
Transfusion	98 (66)	9 (56)	24 (86)	4 (100)	0.060
Hydroxyurea use	4 (3)	1 (6)	2 (7)		0.341
Lifetime hospitalizations, N (%)
≤5 hospitalizations	104 (70)	13 (81)	17 (61)	1 (25)	0.019
6–10 hospitalizations	24 (16)	1 (6)	4 (14)	1 (25)
>10 hospitalizations	20 (14)	2 (13)	7 (25)	2 (50)
Laboratory parameters, mean±SD
Hemoglobin, g/dL	7.8±1.1	7.8±0.7	7.3±1.2	5.4±0.8	<0.001
MCV, fL	79.5±10.2	78.1±9.8	81.4±10.4	84.1±10.4	0.590
Fetal hemoglobin, %	9.2±5.4	11.7±5.8	8.6±5.0	10.4±5.9	0.280
ARC, ×103/L	537±223	734±250	599±282	760±349	0.005
WBC, ×109/L	14.1±4.4	15.1±3.9	16.8±9.1	10.7±5.7	0.046
ANC, ×109/L	5.4±2.3	6.2±2.3	6.1±2.2	5.0±3.6	0.266
Platelets, ×109/L	404±164	318±107	317±152	119±45	<0.001
ALT, U/L	21±15	16±6	22±23	28±9	0.412
Creatinine, mg/dL	0.2±0.1	0.3±0.1	0.2±0.1	0.2±0.1	0.318
Spleen volume, mL
Mean±SD	71.5±55.9	118.5±78.9	173.7±148.2	589.2±382.4	<0.001
Median	57.7	97.8	110.5	540.6	<0.001
IQR (25%, 75%)	(35.7, 91.0)	(66.0, 137.1)	(83.7, 221.7)	(297.1, 832.6)
TCD velocity, mean±SD
TAMV, cm/s	149±25	147±26	146±28	167±18	0.472

Fig. 3.

Correlation of splenic volume and body length for children enrolled in SPHERE. Black line = correlation in SPHERE study. Red lines = 95% confidence interval for a normal population of children without sickle cell disease living in a high-income country [24].

Genetic Modifiers

Six children were found to be ineligible after DNA testing confirmed sickle cell trait. All of the other children were confirmed to have HbSS, and specifically none had heterozygous beta thalassemia trait that is present in East Africa [28]. Alpha thalassemia trait was common with half of the study participants having either 1 gene deletion (38%) or 2 gene deletions (12%). G6PD deficiency (A⁻ variant) was present in 15% of boys, while 27% of girls were heterozygous and 5% of girls were homozygous for this mutation (Table 3).

Table 3.

Prevalence of α-thalassemia and G6PD deficiency observed in SPHERE participants

Genotype	Clinical effect	Observation	Hydroxyurea	p value
α thalassemia, N (%)		n = 149	n = 47
5 copies, αα/ααα	1-gene duplication, unaffected	1 (1)	0 (0)	0.133
4 copies, αα/αα	0-gene deletion, unaffected	64 (43)	29 (62)
3 copies, αα/–α3.7	1-gene deletion, silent carrier	60 (40)	14 (30)
2 copies, -α3.7/-α3.7	2-gene deletion, trait	20 (13)	3 (6)
Indeterminate		1 (1)	1 (2)
No sample tested		3 (2)	0 (0)
G6PD deficiency, N (%)
Males (n = 93)		n = 71	n = 22	0.737
B or A+	Unaffected male	59 (83)	18 (82)
A–	Affected male	10 (14)	4 (18)
Indeterminate
No result		2 (3)	0 (0)	0.233
Females (n = 103)		n = 78	n = 25
BB, BA+, A+A+	Homozygous wild type, unaffected	54 (69)	13 (52)
BA–, A+A–	Heterozygous A– variant, carrier	19 (24)	9 (36)
A–A–	Homozygous A– variant, affected	3 (4)	2 (8)
Indeterminate		1 (1)	1 (4)
No sample tested		1 (1)	0 (0)

G6PD, glucose-6-phosphate dehydrogenase.

Treatment Initiation

Of 47 children eligible for hydroxyurea for elevated TCD velocities, 45 successfully initiated treatment. Two children were not treated; 1 withdrew due to distance from the clinic, and 1 developed acute splenic sequestration and died before initiating study treatment (shown in Fig. 2). The starting dose was 20.2 ± 1.4 mg/kg/day (range 17.1–22.4 mg/kg/day).

Discussion/Conclusion

We describe the study design and baseline data for the prospective SPHERE trial, which has fully enrolled a large cohort of Tanzanian children with SCA. With investment in local infrastructure and families who are eager to participate, the trial enrolled ahead of schedule (Fig. 1). Using well-trained and formally certified local examiners, the children underwent baseline TCD screening that identified a high proportion (24%) with either conditional (22%) or abnormal (2%) TCD velocities. These data document a higher prevalence of elevated TCD velocities, compared to earlier US-based studies that identified 10–17% of untreated children with high risk for primary stroke [6, 29, 30]. In addition, we found lower mean hemoglobin concentration, oxygen saturation, serum creatinine, and mean corpuscular volume in children with elevated TCD velocities, confirming that previously identified markers of disease severity correlate with elevated TCD velocities in the sub-Saharan African context. This higher overall prevalence of elevated TCD velocities in our Tanzanian cohort is similar to the prevalence of elevated TCD velocities documented among children with SCA in East Africa [24], West Africa [31], and the Caribbean [32, 33]. Whether this higher prevalence reflects biological, genetic, or environmental differences is unclear, but highlights the need for reliable TCD screening to prevent pediatric stroke. Particularly in low-resource countries where medical rehabilitative services are limited and the socioeconomic impact of stroke is high, an effective TCD screening program with direct linkage to disease-modifying therapy is an essential part of care for children with SCA.

For children in the USA with SCA and elevated TCD velocities, an impressive decrease in clinical neurovascular complications using chronic transfusions was first proven over 20 years ago [7], but this approach has been extremely challenging to implement in low- and middle-income countries within sub-Saharan Africa. First, there are difficulties in establishing a robust TCD screening program, beginning with the costs of instrumentation and the need for a formal training and certification process. Issues related to a safe and adequate blood supply are also formidable, since demand outstrips the supply necessary to support population needs, even before estimating the blood needed to adequately provide chronic transfusion therapy for children with SCA [34]. Even with an adequate supply, however, blood banking practices with full cross-matching and extended red cell typing are typically unavailable even at major referral centers [35]. Erythrocyte alloimmunization is not uncommon from sporadic transfusions in sub-Saharan Africa [36], and regularly scheduled transfusions for stroke prevention will likely increase the incidence of alloimmunization and complicate future care [37]. Blood is only screened for four pathogens: HIV, hepatitis A, hepatitis B, and syphilis. Finally, repeated simple transfusions will inevitably lead to iron overload, which carries additional morbidity.

A suitable alternative to transfusions for stroke prevention is oral hydroxyurea, which modifies red cell and vascular physiology through multiple mechanisms [38]. The stimulation of fetal hemoglobin production and consequent inhibition of polymerization and hemolysis enable greater red cell survival, higher hemoglobin, and greater oxygen delivery to the tissues including the brain. The reduction in circulating leukocytes, free heme, and reactive oxygen species, along with greater bioavailability of nitric oxide, decrease the chronic inflammatory state that contributes to development and progression of vasculopathy. US-based studies have shown that hydroxyurea when escalated to achieve mild myelosuppression reduces TCD velocities [29] and maintains the reduction achieved by chronic transfusion in children receiving transfusions for primary stroke prevention [39]. A recent publication from Jamaica also documented the reduction in elevated TCD velocities from hydroxyurea at MTD [40].

The NOHARM MTD trial in Uganda demonstrated the feasibility and superiority of escalating hydroxyurea to MTD at ∼30 mg/kg/day, compared to fixed dose at ∼20 mg/kg/day, for decreasing complications of SCA [17]. Despite a report suggesting that moderate fixed dose hydroxyurea can reduce primary stroke rates [41], children with elevated TCD velocities are an especially high-risk group who could derive particular benefit from dose escalation to MTD if it can reasonably be accomplished, especially at referral centers in Africa where TCD screening becomes established. Access to and interpretation of laboratory monitoring, and medication titration using fixed pill sizes (usually 500 mg capsules) may be challenging. In the SPHERE trial, we are facilitating dosing adjustments using an online calculator that incorporates current laboratory values and weight, plus the previously prescribed dose, to avoid dosing errors and ensure that doses are held or reduced when hematological toxicities are observed. The dosing calculator also averages the weight-based dose over the course of a week so that a weekly number of capsules can be prescribed for children who take less than one pill per day. In SPHERE, clinical and laboratory monitoring becomes less frequent after a child has achieved a stable, optimized dose.

We observed a high 25% prevalence of baseline palpable splenomegaly in the SPHERE cohort, with half of these spleens ≥5 cm below the left costal margin. Palpation of the spleen is a relatively imprecise technique for detection of splenomegaly with low sensitivity [42], so in SPHERE we have included sonographic measurement of the spleen in 3 dimensions to calculate splenic volumes more accurately. The calculated volumes did correlate with palpated size. Larger size by palpation and greater volume by ultrasound were associated with lower hemoglobin concentration and platelet count, as well as an association with history of splenic sequestration. Volume by ultrasound revealed a correlation with height, age, and alpha thalassemia status that was not apparent when compared to palpable spleen size. Few previous studies adequately describe an appropriate control population for splenic size in SCA, since they were conducted in a different setting [25], focused on another specific disease [43], included restricted ages [44], or reported only one splenic dimension [45]. These SPHERE data therefore provide a useful reference range for spleen size and volumes in a large pediatric cohort with SCA living in sub-Saharan Africa.

For children with SCA living in the USA, the usual pattern of recurrent vaso-occlusion and tissue infarction leads to functional asplenia and eventual fibrotic involution [4, 46]. In sub-Saharan Africa, however, splenomegaly is a common problem due to parasites such as malaria and schistosomiasis that influence the size and function of the spleen. Palpable splenomegaly is relatively common in young children with SCA [47] and can persist into the third and fourth decades of life [48]. A report that spleen sizes decreased after treatment with proguanil suggested that chronic asymptomatic malaria parasitemia may be due to the spleen harboring malaria [49], but a later study showed no relationship between splenic size, malaria, or proguanil use [50]. For children enrolled in SPHERE, the causes of baseline splenomegaly, functional status of the enlarged spleens, and relationships to parasitic infections are currently under investigation.

As the treatment phase of the SPHERE trial continues, the relationship between splenomegaly and hydroxyurea therapy in sub-Saharan Africa will be followed closely, as the REACH cohort has reported an increased prevalence of palpable splenomegaly over time in association with hydroxyurea escalated to MTD [51]. This relationship and its effects need further elucidation, as the findings have implications for widespread prescription of hydroxyurea for SCA across Africa. SPHERE participants with initial normal TCD exams will also provide contemporary comparison data about the longitudinal effects of hydroxyurea on spleen size and function, and evaluation of splenic filtrative function will improve our understanding of its physiological effects.

In summary, children with SCA in Tanzania have an elevated risk for primary stroke and many have splenomegaly. Fully enrolled, SPHERE has built local clinical capacity with high-quality research infrastructure and TCD screening, and will prospectively determine the benefits of hydroxyurea at MTD for primary stroke prevention and its effects on splenomegaly, as access to hydroxyurea continues to expand access across Tanzania.

Statement of Ethics

The SPHERE trial is approved locally by the Catholic University of Health and Allied Sciences (CUHAS) and Bugando Medical Centre (BMC) Combined Research Ethics Committee, approval number CREC/302/2018, nationally by the Tanzanian National Institute for Medical Research's (NIMR) National Health Research Ethics Committee (NatHREC), approval number NIMR/HQ/R.8a/Vol. IX/2952, and the Tanzanian Medicines and Medical Devices Authority (TMDA), approval number TFDA0018/CTR/0016/06, and also by the Cincinnati Children's Hospital Medical Center (CCHMC) Institutional Review Board, approval number 2018-5483. Written informed consent was obtained from the parent/guardian of each child. Assent was obtained from children ≥10 years old.

Conflict of Interest Statement

Russell E. Ware is a consultant for Nova Laboratories; receives research donations from Bristol Myers Squibb, Addmedica, and Hemex Health; and chairs Data and Safety Monitoring Boards for Novartis and Editas. Susan E. Stuber is a consultant for the American Society of Hematology Research Collaboration. None of these disclosures is relevant to the results and conclusions of the SPHERE trial. The remaining authors have nothing to disclose.

Funding Sources

Luke R. Smart receives support from a training grant from the National Heart, Lung, and Blood Institute (K23 HL153763) and an American Society of Hematology Research Training Award for Fellows. Emmanuela E. Ambrose is supported through an American Society of Hematology Global Research Award.

Author Contributions

Janet Adams, Thad A. Howard, Maria Nakafeero, Sara M. O'Hara, and Primrose Songoro assisted with data collection and review, and critically read, edited, and approved the final version of the manuscript. Emmanuela E. Ambrose, Luke R. Smart, and Russell E. Ware conceptualized and designed the trial, analyzed the data, wrote the manuscript, and critically read, edited, and approved the final version of the manuscript. Georgina Balyorugulu, Mwesige Charles, Protas Komba, Kathryn E. McElhinney, Jodie Odame, and Idd Shabani assisted with data collection, and critically read, edited, and approved the final version of the manuscript.. Adam Lane and Teresa S. Latham analyzed the data, wrote the manuscript, and critically read, edited, and approved the final version of the manuscript. Abel N. Makubi critically read, edited, and approved the final version of the manuscript. Susan E. Stuber designed the trial and critically read, edited, and approved the final version of the manuscript.

Data Availability Statement

Requests for de-identified data may be directed to the corresponding author and must include an IRB-approved proposal. The data that support the findings of this publication are not publicly available due to the ongoing nature of the clinical trial and the need for approval from the Tanzanian national regulatory board. De-identified data will be made available following approval by an internal review committee and execution of a data-sharing agreement between all parties.

Clinical Trial Registration: NCT03948867.

Luke R. Smart and Emmanuela E. Ambrose contributed equally.

Funding Statement

Luke R. Smart receives support from a training grant from the National Heart, Lung, and Blood Institute (K23 HL153763) and an American Society of Hematology Research Training Award for Fellows. Emmanuela E. Ambrose is supported through an American Society of Hematology Global Research Award.