Hydroxyurea Therapy for Neurological and Cognitive Protection in Pediatric Sickle Cell Anemia in Uganda (BRAIN SAFE II): Protocol for a single-arm open label trial

Ethics approval and consent

This protocol and all other participant-facing study documents were approved by the Makerere University School of Medicine Research and Ethics Committee (SOMREC protocol #2019-147) and Columbia University (Columbia protocol AAAS8955). Administrative clearance was received from Mulago National Referral Hospital (MHREC# 1802), Uganda National Council of Science and Technology (HS864ES) and the National Drug Authority (CTA0147).

Study location

Mulago Hospital sickle cell clinic (MHSCC), a well-established specialized clinic at the Mulago National Referral Hospital, is staffed by pediatricians and pediatric hematologists in collaboration with the Makerere University College of Medicine.(10, 15) MHSCC is the largest SCA clinic in Uganda, providing cost-free specialized care to affected children. Over 5000 pediatric patients are actively followed, with 350-450 weekly visits.(16) The clinic offers standard of care for SCA in accordance with the Uganda Ministry of Health national SCA guidelines, including monthly supplies of folic acid, malaria prophylaxis and, for children to the age of age 5 years, penicillin V.(17)

Eligibility criteria and informed consent

Based on MHSCC records, parents/guardians of children who met the screening eligibility criteria (Table 1) were contacted and invited to the study clinic to provide written consent. Among the criteria, enrolment required the absence of evidence of prior stroke detected by a standardized stroke examination, PedNIHSS.(10) Consent (and assent for children ages ≥8 years) for trial procedures were provided in English or a local language (Luganda), per parent/guardian and participant preference.

Intervention

This is an open-label single arm trial. Hydroxyurea is a well-studied medication for children with SCA for reducing disease-associated complications.(18-20). Hydroxyurea therapy increases hemoglobin levels and induces a dose-dependent increase in fetal hemoglobin (HbF) levels. HbF, a minor component of total hemoglobin, reduces hemolysis and anemia by inhibiting the polymerization of sickle hemoglobin, and thereby the sickling of red blood cells. Other hematological effects include reduced numbers of white blood cells and platelets, thought to be useful for decreasing vascular activation and inflammation.

An earlier Ugandan pediatric trial demonstrated hydroxyurea safety in children with SCA, despite endemic malaria, as well as its efficacy in reducing acute pain crises and need for blood transfusions.(21) This study led to the Uganda approval of hydroxyurea for children. More recently, a Ugandan randomized trial of standard hydroxyurea dose versus dose escalation demonstrated clinical advantage to the latter.(22) Due to these established benefits, all the participants in the BRAIN SAFE II trial were started on hydroxyurea with planned dose escalation.

Study drug and dosing

Study drug is a hydroxyurea product, Siklos® (AddMedica/Theravia, France). It is formulated for accurate pediatric dosing and swallowing in readily dispersible scored tablets of 100 and 1000mg. A gradual escalation to maximum tolerated dose (MTD) up to 30mg/kg/day is planned, based on current standard international guidelines for escalation and haematological toxicity, especially myelosuppression (Table 1).(22) Dose escalation is planned for all participants, regardless of laboratory or clinical improvement on lower doses.

• Hydroxyurea dose escalation will occur in 2.5 mg/kg/day increments, adjusted every 8 weeks unless a haematological toxicity occurs or reaches MTD;

• Target absolute neutrophil count (ANC) on hydroxyurea therapy is 2.0 - 4.0 x 10⁹/L, similar to goals of a previous hydroxyurea clinical trial.(22) Dose-limiting MTD may also be based on low absolute reticulocyte count (ARC).

Strategies for drug adherence

Residual pill counts are performed by study staff at each study visit. Adherence is also assessed by parental/guardian report of adherence and by haematological laboratory evidence of reduced dose exposure, e.g. rising neutrophil counts and reduced mean corpuscular volume (MCV) compared to earlier changes.

Participants with excess residual study tablets and the above evidence are identified as less adherent than study requirements. Staff will provide counselling for adherence to parents/guardians and work to identify addressable adherence barriers. Caregivers will be reminded not to share study drug with other family members with SCA.

Criteria for discontinuing trial participation

Participants may be removed from the trial by the Principal Investigators (PIs) for any of these reasons: severe allergic reaction or life-threatening drug-related toxicity; withdrawal of approval by the parent/guardian; participants who miss two scheduled study visit in any 12-month period; participant withdrawal for poor adherence despite staff interventions.

Medical care during the trial

In addition to the study drug, parents/caregivers are educated on the care of children with SCA, including hydration, care during cold weather, pain management and when to seek medical help. Hospitalizations for SCA complications and other serious medical events will use MHSCC admission criteria.

Sample size

Sample size of 270 participants will provide sufficient power (alpha=0.05; beta=84%) to detect decreased arterial TCD velocity by 15 cm/second over 3 years in children with abnormally elevated velocities at baseline.(23) This difference is clinically significant, and sufficient to move a child into or towards the normal range, thereby lowering stroke risk.(24) Sample size determinations included an additional 8% to accommodate expected losses due to stroke, death or other loss to follow-up. The proposed sample size is also expected to provide sufficient power to detect even modest changes in cognition over time.

Recruitment

We randomly selected, consented, and enrolled a cohort of 270 children ages 3-9 years from among eligible patients attending MHSCC. Children ≥8 years were asked for assent. Final screening included a neurological exam using the PedNIHSS (Figure 2).(25, 26) Eligibility required lack of with evidence of prior stroke by a standardized pediatric stroke scale (PedNIHSS). Participants are provided with and monitored on hydroxyurea for the entire study period. Most scheduled study assessments will be those we have previously described.(10)

A sample of 110 siblings or close family relatives without SCA were recruited as controls and to create age-specific z-scores for neurocognitive testing, malnutrition and other study assessments. Controls were recruited at each expected participant age group throughout the trial: ages 3-12 years. Eligibility required lack of neurologic evidence of prior stroke PedNIHSS or other neurologic condition; and confirmed lack of SCA by hemoglobin electrophoresis.

Implementation

Baseline assessments consisted of a detailed history, including age at SCA diagnosis, number and reasons for previous hospitalizations, blood transfusions and severe SCA complications. In addition, standard anthropometric, neurologic, TCD and neurcognitive testing, and neuroimaging testing, complete blood count and renal and hepatic function tests were obtained (Figure 2, Tables 2, 3). Subsequently, scheduled visits include a general health assessment, weight and blood count monitoring for study drug safety and ongoing weight-based dosing.

After completing baseline assessments, participants were initiated on daily hydroxyurea therapy of 20+/-2.5mg/kg/day. Families of enrolled children are provided with dose instructions and sufficient study drug supply until the next scheduled visit. Dose escalation is performed at scheduled visits and according to standardized safety consideration for blood counts (described below).(22). In addition, periodic dose adjustments are performed for weight changes to maintain stable daily per-kilogram intake. Potential hematological, hepatic and renal toxicities are monitored prior to dosing and at annual intervals (Table 3).

Primary outcomes

The primary efficacy outcome measures are the frequency, age-specific prevalence and severity of SCVI manifested by >1 of the 3 end points after 3 years compared to baseline assessments (Figure 1, Table 4) and the age-specific prevalence of these outcomes determined by our prior cross-sectional study, BRAIN SAFE (10):

1. Reduced maximal TCD velocity by >15cm/second and by improved TCD using standard categorical assessment - normal, conditional or abnormal - for children with SCA; (27)

2. Stabilized or improved cognitive impairment using age-specific z-scores and categorical (impaired/not impaired) using standard criteria of z-scores ≤-2 using compared to baseline;

3. New stroke, as assessed by PedNIHSS, or death where stroke appears to be the inciting event. Rate of incident stroke will be compared to reports of children with SCA in the region, including the 5-month period of our prior cross-sectional study at MHSCC. (10)

Secondary outcomes

Secondary outcome measures are:

1. Proportion of the sub-sample with magnetic resonance imaging who have abnormalities detected (yes/no); and stabilized or worsened structural changes on brain magnetic resonance and angiography imaging (MRI and MRA) at the follow-up scan at 36 months compared to baseline.(10, 28)

2. Higher haemoglobin and HbF, lower CRP levels

3. And, improved nutrition status by WHO standards (the latter by age-adjusted international z-score) and by bioimpedance analysis (BIA) to assess changes in body composition at months 18 and 36 compared to baseline values.(10, 29)

Measurement of outcomes

Assessment for all 3 primary endpoints are scheduled for enrolment, months 18 and 36 (Figure 2) and (supplementary Table 1):

Secondary outcomes

Data collection and management

Participants were assigned a non-identifying numeric code for secure de-identification to maintain confidentiality for all study records. Data are collected and retained using both paper-based and electronic record forms. The password-protected study database uses REDCap (https://www.project-redcap.org), hosted by Global Health Uganda with cloud back-up. Paper records are kept in a locked file cabinet in the Uganda PI’s office. Data will be kept electronically in compliance with prevailing laws on data storage.

Participant retention

Participants are reimbursed for round-trip transportation costs to clinic for enhanced visit attendance. Contact information/telephone and location and directions to their homes are recorded. Caregivers are called if children miss scheduled follow-up visit.

Laboratory evaluations and storage of biological specimens

Table 3 summarizes the schedule and specifics of laboratory testing at enrolment and scheduled assessments.

1. At enrolment, laboratory testing included a full blood count, HbF level and safety labs. At subsequent study visits, safety labs are performed in accordance with the MHSCC clinical protocol.

2. Additional plasma samples are obtained at time 0, 18 months and 36 months for CRP testing and storage for biomarker testing.

3. Cell pellets from the plasma sample will be stored for future DNA testing for SCA variants and other polymorphisms.

Patient characteristics and baseline comparisons

At enrolment, descriptive statistics will characterize baseline demographic and clinical features. These characteristics shall include the child’s age, sex, anthropometric measurements, mother’s (or primary caregiver’s) age and educational attainment and family socioeconomic status. Caregiver questionnaire reports SCA complications, serious infections and hospital admissions in the pretrial 12 months, haemoglobin level. For each primary outcome, baseline characteristics of children who did/did not experience an event an experience an event will be compared using a Student’s t- and Chi-squared tests. Non-parametric tests shall be used as needed. Variables that show an association with the occurrence (or non-occurrence) of the events (P<0.2) will be selected for the multivariable analysis in the multivariable models for the respective outcomes.

Primary analysis

The primary outcome analysis will include neurological examination for incident stroke, TCD and neurocognitive outcomes as distinct endpoints to gauge the effect of hydroxyurea at subsequent time points (18 and 36 months) compared to baseline measures performed by intention to treat. In addition, analysis of participants who reach an early endpoint of stroke or death, thereby terminated the trial early, will be compared to the remaining sample. Crude and adjusted mixed effects models will be adjusted for baseline characteristics with P<0.2 at bivariate analyses. Generalized and linear mixed effects models (GLMM) account for the correlation of repeated measures on the same child. Unadjusted analyses shall be conducted, followed by adjusted analyses shall be conducted by adding other baseline characteristics with P<0.2 to the models in addition to the time variable. We will treat age as a time-varying covariate to account for age differences. No data imputation will be conducted for missing data.

Assessment of efficacy

TCD shall be measured both as a quantitative (by arterial flow velocity) and as categorical (normal vs. conditional/abnormal) variable. We will use linear mixed effects models (LMM) to assess within and between participant changes over time from month 0 to 36 for the continuous TCD velocity measure, where participant-level variation will be captured by a participant-specific random effect.

Neurocognitive tests shall also be both quantitative (z-scores) and categorical (less than -2 z-scores vs. -2 z-scores and above, defining neurocognitive deficits present or absent, respectively. Changes in binary neurocognitive measures over time will be modelled using GLMM.

Incident stroke will be determined through NIHPedSS, to be dichotomized into yes/no at each measurement time point (baseline 0, 18, and 36 months), considering time to the stroke. The proportional hazards assumption shall be evaluated both using log-log plots and Schoenfeld residuals and by testing the significance of interaction terms of model variables with time. The proportional cause-specific hazard assumptions shall also be assessed using Schoenfeld residuals.

Analysis of secondary outcomes

Levels of haemoglobin, HbF, CRP and malnutrition status (the latter by z-score) over time will be compared to baseline measurements.

Brain magnetic resonance and angiography imaging (MRI-MRA) will be performed on a subset of 90 SCA participants (33%) at baseline and study month 36. We will assess imaging markers for SCVI at enrolment and subsequently for evidence of stabilization or worsening treatment over time(28). Changes in cerebral infarcts and vascular stenoses will be assessed as the proportion of: a) additional participants with infarct or stenoses; b) additional or larger/more severe infarcts or stenoses at trial completion for those with these abnormalities are baseline. Outcomes will be tested as a dichotomous outcome of MRI-MRA abnormality yes/no and stable or worse, using standardized SCA cerebral vasculopathy evaluations to identify patients whose scans worsen. Logistic regression will be used to identify risk factors such demographic, clinical and laboratory results for worsening versus stable imaging results.

Changes in anaemia, CRP, and malnutrition status

Changes will be assessed for haemoglobin, CRP, and malnutrition z-scores at months 0, 18 and 36. To estimate hydroxyurea effects for each endpoint, we will use linear mixed effects models to assess within and between participant changes over time for the continuous measures of sickle cerebrovascular injury biomarkers (e.g. haemoglobin, CRP, and malnutrition z-scores), where participant-level variation is captured by random effects.

Oversight and monitoring

All adverse events (AEs) and serious adverse events (SAEs) will be recorded, tabulated by type, intensity, seriousness, duration, and relationship to study drug and study procedure. Oral hydroxyurea is currently standard of care in Uganda for abnormal TCD or secondary stroke prevention in children with SCA. Moreover, its safety for this Ugandan population was demonstrated.(45) Hydroxyurea is used widely as standard of care in the U.S., France and elsewhere for children with SCA.(18, 46) For these reasons, interim safety analysis and futility testing were not required. Nonetheless, we constituted a Data Safety and Monitoring Committee (DSMC) to review any safety issues that may emerge compared to local standard of care. Its trial functions are to assure correct study procedures and participant safety. The independent members and the study biostatistician are each connected to SCA and clinical trials. The DSMC meets regularly at least twice annually to evaluate study progress and safety. All adverse events (IAEs) are documented following GCP principles. All severe AEs (SAEs) are promptly reported to the sponsor, DSMB and ethics committees within seven days of the investigators becoming aware of the event.

All scientific decisions are made jointly by the two PIs. Global Health Uganda oversees compliance with “Good Clinical Practice” (GCP), trial scientific integrity and oversight for all procedures, including data collection. Regularly scheduled communication, as well as ad hoc communication among the study team leadership supports high-quality trial conduct and quality assurance. Internal institutional compliance monitoring is conducted by Makerere University School of Public Health Clinical Trials Unit. External monitoring is conducted by the East African Consortium for Clinical Research Monitors.

Trial status

Trial enrollment occurred from April to December 2021, following a 1-year delay due to COVID-19 pandemic starting in March 2020. Staff followed study-specific standard operating procedures (SOPs) to minimize risk of viral infection for staff, participants and their families. Baseline clinical assessments and initiation of hydroxyurea treatment for all participants commenced at or shortly after enrollment. An unselected subset of 90 ages ≥5 underwent MRI-MRA imaging.

Participants continue to undergo per protocol monitoring and study procedures. As of July 2023, all trial participants had completed study month 18, when a optional midpoint analysis was performed. Based on the month-18 interim analyses, Makerere Institutional Review Board (IRB) and the DSMC approved a shortening of the trial from 36 to 30 months. The 12-month following the midpoint assessment would permit assessment of continued or even further benefit from study drug. Further, a 12-month gap between neurocognitive testing for children is considered the minimum time to avoid practice effects from repeated testing.