Hydroxyurea to decrease stroke risk in children with sickle cell anemia: a systematic review and meta-analysis Open Access

Sickle cell anemia (SCA) is among the most prevalent inherited blood disorders globally, present in 75% of ∼400 000 annual births occurring in Africa.¹ This heavy burden places SCA high on the list of noncommunicable diseases causing early mortality, especially in African countries.^1,2 The disease’s complex pathophysiology involves hypoxia-induced polymerization of hemoglobin S (HbS) that deforms red blood cells (RBCs), damages RBC membranes, and increases inflammation, leading to severe complications such as chronic anemia and organ damage, which are particularly detrimental to young children.³ 

Among the complications of SCA, stroke is one of the most severe and can occur as early as 2 years of age, peaking between age 5 and 9 years.⁴ Without disease-modifying therapy, children with SCA have a high cumulative incidence of stroke, reaching 7.8% by age 14 years in a Jamaican cohort⁵ and 11% by age 20 years in a US Cooperative Study (CSSCD).⁴ In Sub-Saharan Africa, pooled stroke prevalence estimated mostly in cross-sectional hospital-based convenience samples is 4.2% (range, 0.7-16.9) and slightly higher in studies that included imaging.⁶ Early deaths from other causes might contribute to a lower than expected estimate of cumulative incidence.⁷ In the United States, both hemorrhagic and ischemic stroke are observed in children, but ischemic stroke is more common before 20 years of age.⁴ There are fewer reports of hemorrhagic stroke in Sub-Saharan African studies, in which imaging confirmation is not always available to distinguish between hemorrhagic and ischemic insult.⁸ Regardless of etiology, stroke has devastating consequences, and up to one-third of affected patients die.^4,9,10 Those who survive often face neurocognitive disabilities, and recurrent strokes are common, especially in the years immediately after the initial event.¹⁰ 

The risk of stroke in SCA increases as the cerebral blood flow velocity rises, a phenomenon measured using transcranial Doppler (TCD) ultrasound. Adams et al¹¹ validated TCD as an accurate predictive tool for primary stroke risk assessment in children with SCA. Conditional TCD velocities of 170 to 199 cm/s moderately increase stroke risk, and abnormal TCD velocities of ≥200 cm/s drastically increase stroke risk.^12,13 For every 10 cm/s increase in velocity above 170 cm/s, the stroke risk rises by 30%.^14,15 Disease-modifying therapies, such as hydroxyurea and chronic monthly RBC transfusions, have shown great clinical benefit.¹⁶ The landmark STOP (Stroke Prevention in Sickle Cell Disease) trial demonstrated that chronic transfusion therapy lowers TCD velocities and reduces primary stroke risk.¹⁷ STOP2 confirmed that continuation of transfusion therapy is better than cessation of all disease modification.¹⁸ Hydroxyurea monotherapy has never been studied in comparison with placebo for primary stroke prevention, but the TWiTCH (TCD With Transfusions Changing to Hydroxyurea) trial showed that transitioning to hydroxyurea at maximum tolerated dose is noninferior to continuation of transfusion therapy if children with elevated TCD velocities are treated with chronic transfusion therapy for 1 year, TCD velocities normalize, and vasculopathy is absent on magnetic resonance imaging/magnetic resonance angiography (MRI/MRA).¹⁹ 

In higher resource settings, TCD screening and transfusions have reduced stroke incidence by >90%,^11,13,15 although complications such as RBC alloimmunization, iron overload, and treatment noncompliance remain concerning.¹⁹ However, in lower resource settings, key diagnostic tests such as MRI/MRA are often unavailable, the blood supply is limited, and chronic transfusion programs are not sustainable.²⁰ In the absence of viable alternatives, hydroxyurea treatment is much better than providing no disease-modifying therapy at all.²¹ Hydroxyurea works differently than transfusion therapy, but it is more accessible and affordable in many countries and is increasingly being used as an alternative treatment strategy.²² Despite its overall clinical effectiveness,²¹ there is still no consensus on the use of hydroxyurea in children with high stroke risk, leading to variability in treatment practices. Therefore, this review aims to consolidate evidence on hydroxyurea's efficacy in reducing stroke risk by lowering elevated TCD velocities, particularly in lower resource settings, to guide future treatment strategies and improve patient outcomes.

The research followed the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines.²³ 

We included all prospective controlled clinical trials that enrolled children with SCA, performed TCD screening before hydroxyurea treatment initiation, prescribed protocol-directed hydroxyurea, and collected serial measurements of TCD velocities and stroke incidence during hydroxyurea treatment, regardless of the date when the trial was conducted.

Male or female children with SCA (HbSS and HbSβ⁰).

Hydroxyurea treatment, any dosing regimen.

Change in TCD velocity between pretreatment and posttreatment. Stroke incidence during treatment.

Laboratory benefit and toxicities while receiving hydroxyurea.

We conducted a comprehensive search in 5 major medical databases: CINAHL, EMBASE, Trip Medical Database, Scopus, and PubMed. The database query occurred on 23 July 2024. An information specialist (D.P.J.) developed a detailed search strategy with input from 2 board-certified pediatricians (L.R.S. and E.E.A.).

The search strategy combined key concepts using controlled vocabulary and keywords to identify prospective, controlled clinical trials that included children with SCA who underwent TCD screening before hydroxyurea initiation, and documented TCD results during hydroxyurea treatment. No date or language limitations were applied. A detailed description of the search is included in supplemental Table 1. The initial search was conducted in PubMed and subsequently adapted for EMBASE, CINAHL, Trip Medical Database, and Scopus.

Search strategies were validated using known authoritative articles, provided by L.R.S. and E.E.A. Once each search strategy retrieved these key articles, the searches were considered complete. Institutional review board approval was not required for this systematic review.

After deduplication, the title and abstract of each citation was screened independently by 2 investigators (E.E.A. and L.R.S.). If any doubt remained, the full-text article was obtained for review before making the final decision to exclude.

A trial was eligible for inclusion if it was designed as a prospective, controlled clinical trial, enrolled children aged <18 years, included children with SCA, conducted TCD screening to assess primary stroke risk before administration of hydroxyurea (or any other disease-modifying therapy), administered hydroxyurea alone for disease-modifying therapy, and obtained follow-up TCD measurements during hydroxyurea therapy. Any publications related to an included trial were identified and included for full-text review.

Studies were excluded if they were observational cohorts, used chronic transfusion therapy before or in conjunction with hydroxyurea therapy, or did not report TCD velocities before and after hydroxyurea initiation.

Data were independently extracted by 2 authors (L.R.S. and E.E.A.) for comparison between trials. The primary results publication was the main source of data for description of change in TCD velocity after hydroxyurea treatment. Conference abstracts that preceded the publication, or secondary publications from the clinical trial were reviewed for supplemental data. Data extraction focused on children that received hydroxyurea.

Two authors (L.R.S. and E.E.A.) independently assessed the methodological quality of each trial using the Cochrane risk of bias tools nonrandomized studies of intervention,²⁴ as recommended.²⁵ 

Heterogeneity was assessed with the I² statistic.

A funnel plot was generated to assess reporting bias.

The search was conducted on 23 July 2024 and identified 836 citations from the 5 major databases: 162 in PubMed, 42 in CINAHL, 399 in EMBASE, 28 in Trip Medical database, and 205 in Scopus. After deduplication, 449 citations were screened for eligibility. After 381 citations were excluded, 68 reports were retrieved for full-text review. Based on full-text review and reference lists, an additional 57 citations were identified for full-text review. Of these citations, 20 were excluded after full-text review. The remaining 105 citations describing 13 clinical trials were included in the systematic review. Figure 1 shows the PRISMA diagram depicting the search, screening, and inclusion results. A list of all citations reviewed for each clinical trial included is located in the supplemental Appendix.

Two ongoing trials were identified that might generate useful data for inclusion in future systematic reviews.^26,27 

Trials had a moderate risk of bias based on the ROBBINS-I (Risk Of Bias In Nonrandomized Studies - of Interventions) criteria (supplemental Figures 2 and 3). The possibility of publication bias cannot be ruled out based on the funnel plot (supplemental Figure 4).

Characteristics of the 13 clinical trials are found in Table 1.^28-40 

Trials were conducted in 7 countries on 4 continents: Europe (United Kingdom), North America (United States, Jamaica, and the Dominican Republic), South America (Brazil), and Sub-Saharan Africa (Nigeria and Tanzania). Five were multisite,^{31,32,36,38,39} and 8 were single-center trials.^{28-30,33-35,37,40} 

The design of most trials was open-label,^{28-30,33-35,37,38,40} and typically involved a single arm of treated participants. One trial compared children receiving open-label hydroxyurea treatment with age-matched controls who did not receive hydroxyurea,²⁸ and 1 trial compared children receiving hydroxyurea for abnormal TCD velocities with children undergoing observation for normal or conditional TCD velocities.³³ Four trials were randomized controlled trials: 1 trial randomized participants to hydroxyurea vs observation,³² 2 randomized participants to hydroxyurea vs placebo,^31,39 and 1 randomized participants to different dosing regimens (10 mg/kg per day vs 20 mg/kg per day) and also included a third observation group with normal or conditional TCD velocities.³⁶ One trial enrolled 3 groups of patients, only 1 of which qualified for inclusion in this review.³⁴ 

All trials included HbSS and HbSβ⁰, and several also permitted people with HbSO_Arab, and/or HbSD to enroll.^29,32,34 Populations differed significantly with respect to several important variables: age, TCD velocity, and previous stroke. Two trials enrolled a very young population: BABY HUG enrolled infants aged 9 to 18 months,³¹ and HU Prevent enrolled infants aged 12 to 52 months.³⁹ Because of the relative protection of residual fetal Hb (HbF) in young children,⁴¹ these trials were excluded from the forest plot depicting TCD response to hydroxyurea; however, they are summarized in the tables that describe each trial and a supplemental figure (supplemental Figure 1). The Jamaica liquid hydroxyurea trial enrolled participants aged 0.5 to 18 years, and only 6 participants were aged <24 months.³⁸ Thornburg et al³⁰ enrolled participants aged 1.5 to 5 years, and only 3 participants were aged <24 months. All others enrolled populations that were aged >24 months.^{28,29,32-37,40} 

With regards to TCD velocity, 4 trials enrolled individuals with any TCD velocity (normal, conditional, or abnormal),^28,30,35,38 1 enrolled only normal velocities (<170 cm/s),³⁹ 1 enrolled normal or conditional velocities (<200 cm/s),³¹ 2 enrolled only conditional velocities (170-199 cm/s),^32,40 3 enrolled only elevated velocities (≥140 cm/s or ≥170 cm/s),^29,34,37 and 2 enrolled only abnormal velocities (≥200 cm/s).^33,36 One trial included a few participants for hydroxyurea treatment who had previous stroke or abnormal TCD if transfusions were not possible or had failed.²⁸ One trial included participants with stroke, but only data from those without stroke were included in this review.³⁴ Most trials explicitly excluded children with evidence of previous clinical stroke.^{31-33,35-37,39,40} Three trials did not mention an exclusion of those with stroke.^29,30,38 

Hydroxyurea dosing regimens varied between clinical trials. A fixed weight–based dose regimen was used by 3 trials^31,33,36: 1 used open-label hydroxyurea 20 mg/kg,³³ 1 randomized participants to fixed low dose (10 mg/kg) or fixed moderate dose (20 mg/kg),³⁶ and 1 randomized participants to placebo or hydroxyurea 20 mg/kg.³¹ Ten trials used an escalated dosing regimen.^{28-30,32,34,35,37-40} Hydroxyurea was initiated at 15 to 20 mg/kg per day and increased to 2.5 to 5.0 mg/kg per day every 8 to 12 weeks to a maximum of 30 to 35 mg/kg per day if certain laboratory criteria were met. Hematological toxicity criteria that required pause of hydroxyurea differed slightly between trials as displayed in Table 2. Hydroxyurea dose was decreased for repeated (>1 consecutive complete blood count) or prolonged (>1-2 weeks) toxicities.

All but 1 trial used nonimaging TCD equipment,²⁸ and all trials reported time-averaged mean maximum velocity and STOP criteria with 3 categories for stroke risk (normal <170 cm/s, conditional >170-199 cm/s, and abnormal ≥200 cm/s). The number of vessels examined varied from the full STOP protocol to an abbreviated version that focused on the middle cerebral arteries.³³ The TCD examination was performed by a variety of personnel, from “certified examiners” with undocumented professional background, to neurodiagnostic technicians, to radiologists. The person reporting/interpretating the results varied from a “certified examiner” to a radiologist. Trials reported TCD results every 3 to 12 months. Final changes in TCD velocity (trial primary outcomes) were reported as early as 6 months after initiation²⁸ and as late as 2.75 years,³⁹ but results were reported up to 5 years.⁴⁰ 

Two trials used an explicit definition for stroke (World Health Organization,³³ Pediatric National Institutes of Health Stroke Scale³⁶). In the SPRING trial, a video recording of the examination was validated by another neurologist.³⁶ Other trials did not describe an explicit clinical definition of stroke, so it was assumed that stroke was diagnosed using examination (supplemental Table 2).

Two trials obtained baseline MRI/MRA²⁹ or MRI.^31,42 Five trials obtained MRI³⁹ or MRI/MRA^30,34,35,40 at baseline and interval follow-up. No scheduled imaging was obtained in 6 trials.^{28,32,33,36-38} 

Table 3 contains a summary of the populations enrolled and the results achieved in the individual trials.

Most trials enrolled <50 participants. BABY HUG and SPRING both enrolled nearly 100 participants per arm.^31,36 A total of 575 children received hydroxyurea after baseline TCD ultrasound assessments.

The average hydroxyurea dose was a low, fixed, weight-based dose in 1 trial at 10.8 mg/kg per day³⁶; a moderate, fixed, weight-based dose in 3 trials ranging from 19.2 to 20.6 mg/kg per day^31,33,36 and was escalated in other trials, with the final mean dose ranging from 23.3 to 27.9 mg/kg per day. The mean duration of treatment before the primary outcome in the trials was 1.6 ± 0.7 years, with a range from 0.5 to 2.8 years.

The change in TCD velocity varied based on the timing of initiation. Two trials that enrolled young infants with TCD velocities of <170 cm/s or <200 cm/s were excluded from Figure 2.^31,39 One documented a rise in the average TCD velocity while receiving hydroxyurea, which was slower than the rise observed in the placebo arm.³¹ The other enrolled very few children (n = 12), and reported stable TCD categories but not the changes in mean velocity.

All other trials enrolled older children and revealed a mean decline in TCD velocities from −11 to −55 cm/s in individual trials (Figure 2). The mean overall decrease was a significant decline of −30 cm/s with a 95% confidence interval (CI) of −41 to −19 cm/s. Many of the individual trials included in the summary reported a CI that included the possibility of a rise in TCD velocity in some individuals.

Trials that enrolled children with multiple categories of TCD values (normal, conditional, and/or abnormal), sometimes reported that children in the elevated group (conditional and/or abnormal) had a more significant decline than those in the normal group.^28-30 For example, in a trial of 24 children treated over 6 months, Kratovil et al²⁸ observed a decline of −34.8 cm/s in those with elevated TCDs (>170 cm/s), but only −11.6 cm/s in those with normal TCDs. Similarly Zimmerman et al²⁹ found that a higher baseline TCD measurement was associated with a larger reduction after hydroxyurea therapy (r² = 0.12; P = .04).

The largest TCD velocity reductions were observed in 2 trials that enrolled only children with abnormal TCD velocities that exceeded 200 cm/s at baseline.^33,36 Aside from BABY HUG with its unique infant population, a significant decrease was observed in trials regardless of the dosing strategy or the baseline TCD measurement. Almost all trials reported the proportion of children with higher risk TCD velocities (>170 cm/s) who returned to the normal velocity (<170 cm/s) during the course of the trial (Table 3). On average, 64.9% ± 19.5% of children treated with hydroxyurea had return of their TCD velocity to the normal category (<170 cm/s). Significant TCD reduction was observed as early as 3 months of therapy initiation,³⁶ and the durability of response was documented through 5 years of treatment.⁴⁰ In trials that collected multiple measurements over time, there was continued decline in the mean velocity over the first year of treatment.^29,33,36,40 

Silent cerebral infarct and vasculopathy were identified at baseline in 6 of trials that used scheduled imaging, and remained relatively stable when follow-up imaging was obtained (supplemental Table 2).

Most of the trials reported that no strokes occurred among children treated with hydroxyurea.^{28,30-35,37-39} This included trials that enrolled normal TCD velocities,^38,39 conditional TCD velocities,^31,32 and a combination of conditional and abnormal TCD velocities.^28,29,37 Three trials reported strokes in children treated with hydroxyurea.^29,36,40 All strokes occurred in children whose baseline TCD was abnormal at >200 cm/s. Zimmerman et al²⁹ reported 1 new event (0.52 per 100 patient-years). SACRED reported 1 new event (0.29 per 100 patient-years).⁴⁰ In SPRING in Nigeria, strokes occurred in both the low-dose and moderate-dose treatment arms (1.19 per 100 patient-years vs 1.92 per 100 patient-years; incident rate ratio of 0.62; 95% CI, 0.10-3.20; P = .77).³⁶ SPIN and SPRING reported a similar, or higher incidence of stroke in children with untreated SCA enrolled in a comparator arm that included children with conditional or normal TCD velocities.^33,36 In the BABY HUG trial, 1 stroke was reported in the placebo group, but none occurred in children randomized to hydroxyurea.³¹ 

After hydroxyurea treatment, most trials observed a 1.0 to 2.0 g/dL rise in Hb. The trial that enrolled the youngest participants recorded the lowest incremental improvement in Hb (+0.2, BABY HUG) because the naturally protective HbF helps preserve baseline Hb in infancy and therefore an incremental gain in Hb is not expected. Similarly, HbF declined slightly in BABY HUG (−3.2%) because the baseline HbF was high in the infants at the time of enrollment, but it rose in all other trials that enrolled older children whose HbF was low (1.9%-15.7%) at the time of hydroxyurea initiation (Table 3).

Hematological toxicities as defined in Table 2 for each trial occurred most frequently in BABY HUG. SPHERE, EXTEND, and SACRED observed a moderate number of hematological toxicities. All were reversible, temporary, and did not cause any clinical effects. SPIN and SPRING describe an absence of hematological toxicities. Many of the trials do not discuss hematological toxicities explicitly, as summarized in Table 3.

This systematic review aimed to evaluate the benefits of hydroxyurea therapy in reducing elevated TCD velocities, a known proxy for stroke risk in children with SCA. Hydroxyurea was consistently associated with a decline in TCD velocities in all but 2 trials.^31,39 These 2 trials were unique in that they enrolled very young children with few sickle cell complications who were still benefiting from high levels of HbF and Hb before HbS begins to predominate and severe hemolytic anemia develops, along with disease-related TCD increases. TCD velocities naturally rise in early childhood, especially in children with SCA. BABY HUG showed that hydroxyurea slowed this rise compared with placebo.³¹ The HU Prevent trial also enrolled very young children, but <10 in each arm, and the TCD velocities were not explicitly reported, but all children in the treatment arm remained in the normal risk category.³⁹ 

In trials with children who are aged >2 years, hydroxyurea treatment showed a consistent reduction in TCD velocities, with a mean decrease of −30 cm/s (95% CI, −41 to −19) observed over a treatment period of 0.5 to 2.6 years (Figure 2). The greatest reductions (−45 to −55 cm/s) were observed in trials that enrolled children with exclusively abnormal TCD velocities (>200 cm/s),^33,36 and the least reduction was observed in trials that enrolled children with normal TCD velocities.^28,35,38 The mean decline in many of the individual trials included a wide 95% CI that included the possibility of a rise in TCD velocity in some participants.^{28,30,32,34,35,38,40} 

Observational studies were not included in this review, but cohorts followed-up in Belgium, France, Nigeria, and the United States corroborate these findings,^43-46 underscoring hydroxyurea's dual role in both prevention and treatment. Particularly important for lower resource settings is the long-term efficacy observed in the SPPIBA (Stroke Prevention Programme in Ibadan) cohort in Ibadan, Nigeria, which began treating children with SCA and TCD velocity of ≥170 cm/s with hydroxyurea in 2009 because of lack of access to, and acceptability of, chronic transfusion therapy.⁴⁴ They observed only 2 strokes over 2254 patient-years of follow-up (0.08 per 100 patient-years), and the mean decline in TCD velocity was −44 cm/s (95% CI, −39 to −48).

In the included trials, a reduction in TCD velocities was observed as early as 3 months after initiation,³⁶ with further improvements or maintenance over 2 to 5 years of follow-up.^34,37,40 Most trials reported no incident stroke among participants treated with hydroxyurea.^{28,30-35,37-39} In the 3 trials that reported incident strokes, the incidence was lower than expected for untreated children.^29,36,40 Low stroke incidence while using hydroxyurea in Africa has also been reported in the SPPIBA observational cohort (0.08 per 100 patient-years)⁴⁴ and the REACH open-label dose escalation trial (0.1 per 100 patient-years).²¹ SPIN and SPRING trials reported a low incidence of stroke compared with an observation arm of lower risk children followed up concurrently without treatment,^33,36 as well as compared with prior historical studies.^4,11 

The reduction in TCD velocities is because of, in large part, the improvement in Hb that is observed during hydroxyurea therapy, as increased oxygen delivery and extraction presumably drives the arterial flow rate.⁴⁷ It does so by inducing the production of HbF, which in turn inhibits the polymerization of sickle Hb, improves cell rheology, and preserves the RBC life span, allowing more oxygen carrying capacity and delivery.^48,49 Hydroxyurea’s multifaceted effects on SCA pathophysiology also play a role, especially the salutary effects on inflammation. Reduced hemolysis decreases free Hb, preserves nitric oxide bioavailability, improves vascular tone, and reduces blood flow velocities. Overall, hydroxyurea’s comprehensive impact on blood viscosity, vascular health, and inflammation supports the significant improvements seen in TCD velocities across multiple studies.^29,32,35,37 

Some features of the included clinical trials highlight elements that increase the feasibility of stroke prevention programs in lower resource settings.⁵⁰ Other health care personnel were trained to operate the TCD machine in some trials. Several trials dispensed only 500 mg capsules, even for younger children to improve ease of dosing and storage.^33,37,40 One of the most debated aspects of hydroxyurea use is the optimal dosing strategy, especially in lower resource settings. In the setting of these clinical trials dose escalation was accomplished with serial laboratory monitoring, but the frequency of monitoring required in clinical practice has been debated. Fixed weight-based regimens require fewer dose adjustments and laboratory monitoring but do not adapt to accommodate individual variability in response and tolerance. This approach may be useful in some lower resource settings.

In the included trials, there was not a clear trend toward greater TCD velocity reduction with higher dosing strategies, but the review question did not embrace all aspects of SCA complications. Only the SPRING trial compared 2 dosing strategies,³⁶ and it showed that the TCD reduction and incidence of stroke was similar in people who received either the low-dose (10 mg/kg per day) or the moderate-dose (20 mg/kg per day) regimen, but other clinical complications were significantly better with a higher dose of hydroxyurea. This is congruent with the findings in the NOHARM-MTD trial that found numerous benefits with higher doses that are escalated until mild myelosuppression is observed, and cytopenias were no higher in the dose-escalated arm than the fixed weight–based dose arm.⁵¹ Interestingly, the SPPIBA observational cohort showed that 69% of participants required a hydroxyurea dose of >20 mg/kg per day to achieve significant reductions in TCD velocity.⁴⁴ Personalized pharmacokinetic guided dosing may prove a useful tool for some locations to accelerate the start of a higher dose from the beginning of therapy.²⁶ Notably, several clinical trials successfully achieved a dose escalation even in a lower resource setting.^32,34,37,40 

Although cytopenias occasionally required dose adjustments, severe adverse events were rare. Trials generally monitored hematological values regularly to assess safety. Unfortunately, not all trials clearly described the incidence of hematological toxicities, and toxicity thresholds varied between trials. Cytopenias were commonly observed, but most trials reported manageable levels of toxicity with mild and reversible cytopenia, even with escalated dosing strategies, supporting the feasibility of hydroxyurea as a long-term treatment. Although myelosuppression sometimes represented drug-related marrow suppression that warranted treatment pauses, single cell-line cytopenias also result from other factors such as infection, splenomegaly, or autoimmune conditions. Overall, cytopenias were more common with escalated dosing, but dose reductions were rarely necessary, indicating these cytopenias were generally transient and not solely due to hydroxyurea, and trials did not report any clinical adverse effects of the cytopenias observed.

We made every effort to minimize potential bias in our review. A comprehensive search strategy was implemented, encompassing multiple databases and clinical trial registries to capture all relevant studies. No language restrictions were applied, allowing the inclusion of studies regardless of their original publication language. Each article’s relevance was rigorously assessed, with all screening and data extraction conducted in duplicate to enhance accuracy. Outcomes were prespecified before analysis to avoid post hoc adjustments. However, the limited number of trials available for meta-analysis may still affect the robustness of our conclusions. We were not able to clearly document the effect on stroke incidence because of the short duration, small size of each cohort, and few strokes reported. Perhaps this is because of the effectiveness of hydroxyurea, but it is impossible to be certain. Trials without scheduled imaging could not identify silent cerebral infarcts (SCI) or vasculopathy, and most trials identified stroke by examination.

However, the findings from this review suggest that hydroxyurea is an effective option for decreasing stroke risk in children with SCA, based on its ability to reduce TCD velocities and improve hematological parameters. In practice, both fixed and escalated dosing regimens can be valuable tools, with the choice of regimen ideally tailored to the resources available in the health care setting. Fixed dosing may be particularly beneficial in lower resource settings in which frequent laboratory monitoring is challenging, whereas escalated dosing is better suited for environments with robust monitoring capabilities. The cost of hydroxyurea is still a significant barrier that needs to be overcome in many settings. A 500-mg capsule is the most commonly available dose form, which can be administered even to young children with alternate-day dosing.³⁷ The Lancet Commission on Sickle Cell Disease recommended a target cost of $0.10 per 500 mg capsule, an achievable goal with the assistance of global stakeholders.^52,53 

This review highlights the need for further research to understand the real-world feasibility and practical implications of hydroxyurea dosing strategies and to explore its long-term safety and efficacy in primary stroke prevention. Additionally, there is a need for research into alternative formulations with longer shelf lives and better stability, which would support hydroxyurea’s use in areas with limited pharmaceutical infrastructure. Longitudinal studies on hydroxyurea's impact on TCD velocities, stroke incidence, brain imaging abnormalities, and other long-term health outcomes will also be crucial for understanding the full potential of this treatment in pediatric sickle cell populations globally.

The authors thank Catholic University of Health and Allied Sciences, Bugando Medical Center, and Cincinnati Children’s Hospital Medical Center for their financial and time support, which made this review possible. The authors thank the staff of the Donald C. Harrison Health Sciences Library at the University of Cincinnati for their invaluable assistance with literature searches and data management. The authors also thank their colleagues and peer reviewers whose insights have enriched this manuscript. The authors thank the health care providers and clinical staff who provided context on the practical use of hydroxyurea, and the patients and families affected by sickle cell anemia whose participation in clinical trials has advanced our understanding of stroke prevention.

L.R.S. is supported by a grant from the National Heart Lung and Blood Institute (grant number K23 HL153763). L.R.S. and E.E.A. both benefited from the support of the American Society of Hematology Clinical Research Training Institute.

Contribution: E.E.A. and L.R.S. designed the research protocol; E.E.A., L.R.S., and D.P.J. performed the research (data screening, collection, and extraction); E.E.A., L.R.S., and P.A.S. analyzed the data; E.E.A., L.R.S., D.P.J., P.A.S., A.N.M., and R.E.W. interpreted the data and wrote the manuscript; and all authors made substantial contributions to the analysis and interpretation of data, drafting the manuscript, reviewing it critically for important intellectual content, and giving final approval of the version to be published.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Emmanuela E. Ambrose, Bugando Medical Centre and Catholic University of Health and Allied Sciences Pediatrics, 1370 Wurzburg Rd, Mwanza 33109, Tanzania; email: [email protected].