Acute Stroke in Children with Sickle Cell Anemia without Abnormal Transcranial Doppler in a Hydroxyurea Trial in Uganda

Introduction

Transcranial Doppler Ultrasound (TCD) screening for children with sickle cell anemia (SCA) was established in high-income regions for primary stroke prevention through targeted intervention for time-averaged mean velocity (TAMV) in the highest “abnormal” range. Additional stroke risk factors include severe anemia, low HbF, infections, acute SCA complications, e.g. acute chest syndrome (ACS), pain crisis. In low- and middle-income countries (LMICs), children may have additional risks, e.g. higher infectious burden (e.g. malaria) and malnutrition. Of 4 children who developed acute stroke in our hydroxyurea (HU) treatment trial in Uganda, none had abnormal TCD. We sought to identify stroke risk factors for children with SCA in LMICs beyond high TAMV.

Methods

Our single arm, open-label, 30-month hydroxyurea (HU) trial, “BRAIN SAFE-II,” aimed to reduce stroke and improve cognitive function. Acute stroke or death were early trial endpoints. Children aged 3-9 years attended Mulago Hospital SCA clinic and confirmed SCA by hemoglobin (Hb) electrophoresis were eligible if prior HU use <6 months and no evidence of prior stroke by pediatric stroke exam (PedNIHSS). Study procedures included enrolment TCD per the STOP trial, categorized by maximum TAMV.¹ Study staff underwent TCD training and supervision, and followed standard quality assurance measures. HU was escalated from 20±2.5mg/kg to maximum tolerated dose (MTD).

Results

In all, 264 children enrolled and started HU, mean age 5.6±1.7 years and mean Hb 7.8±1.2g/dL. A total of 252 completed the trial, as 4 had acute stroke (incidence 0.62 per 100 person years), 4 others died without evidence of stroke and 4 withdrew/ lost to follow-up. Acute stroke occurred at mean 9.3 trial months, prior to month 18 scheduled repeat TCD and HbF. These 4 were aged 3-8 years, 3 were male, mean HU dose 24.4± 4.6mg/kg. Their mean study Hb preceding the stroke was 8.5±1.0g/dL vs. 8.8 ±1.4 g/dL (P=0.23) for the entire sample at comparable study time. Mean HbF at enrolment was 4.6±1.02% vs. 12.0±8.13% (P=0.035) for the entire sample. None were malnourished per global criteria; 3 had history of headaches. Each apparent stroke precipitant was malarial infection, ACS, recent pain crisis; one was of uncertain cause. Their enrolment TCDs were: 2 normal, 1 conditional, 1 inadequate. For the entire sample, mean enrolment TAMV was 147.7cm/sec: 15.9%; conditional, 5.7% abnormal, 5.3% inadequate. At month 18, mean HU dose for 254 active participants was 25.4 ±3.04mg/kg and mean TAMV 130.7cm/sec. At month 30, mean HU dose was 25.9 ±2.7mg/kg and TMAV was 131.2cm/sec.

Discussion

In high-income countries, TCD screening is essential for primary stroke prevention in children with SCA.¹ HU reliably reduces TAMV in high-income and LMICs, raises HbF and reduces incidence of pain crises, ACS and clinical malaria. Of the 4 Ugandan participants with acute stroke despite HU therapy, TCD did not identify them as at high stroke risk. Low baseline HbF, malarial infection and episodes of ACS and pain and likely contributed to stroke risk. Children in LMICs with normal or conditional TCD and/or with acute infections or SCA complications may have elevated risk due to low HbF or other factors, potentially necessitating broader considerations for stroke prevention. A similar question was recently raised by a non-HU TCD observational study in southern Africa.²

References: ¹Adams RJ, et al., Controlled Clinical Trials, 1998; ²O’Brian NF, et al, EJHaem, 2023.