Prospective hydroxyurea impact on cerebral magnetic resonance imaging from a 30-month trial of children with sickle cell anemia in Uganda

Abstract Introduction: Children with sickle cell anemia (SCA) often develop cerebral vasculopathy early in life. Vascular injury is manifested by stroke, silent cerebral infarcts (SCI), arterial stenosis and elevated arterial flow velocity measured by transcranial doppler ultrasound (TCD). We had previously reported infarct frequency of 68% in a cross-sectional Ugandan SCA sample who were largely untreated with hydroxyurea (HU) or other disease-modifying therapy. Here we present prospective MRI-MRA imaging results from our recent HU trial, BRAIN SAFE-II, with a new sample of participants ages 3-9 years. To our knowledge, no prospective studies of MRI imaging has been reported for children with SCA on HU therapy in sub-Saharan Africa. Methods: Our HU open-label, single-arm, dose-escalated 30-month HU trial of Ugandan children were enrolled at ages 3-9 years. Ethical considerations precluded trial randomization. Enrollment criteria included no more than a total of 6 months of prior HU treatment and a lack of previous stroke by history and stroke-focused neurological examination. TCD determined time-averaged mean velocity (TAMV) using the STOP criteria (excluding basilar artery). A randomly selected trial subset of 89 children, age 5 years or older (to limit sedation use), underwent prospective MRI-MRA imaging by a 1.5T scanner at trial entry and completion. Cerebral infarcts and vascular stenoses were gauged using the SWiTCH trial criteria, modified by the Silent Infarct Transfusion (SIT) Trial. MR imaging was independently reviewed by each of 2 neuro-radiologists, with adjudication of differences performed by joint review. Data were analyzed by ANOVA for continuous and Chi squared for categorical data. Results: At trial entry, mean age of the imaged subset (N=89) was 6.6±1.4 years, 46.1% male. Initial mean values were Hb 7.8±1.1g/dL, HbF 10.3±7.7%, O2 saturation 95.4±3.2% and systolic BP 105.0±11.3mm Hg. Mean TAMV was 150.1±31.4cm/sec, with 7 (7.9%) abnormal and 11 (12.4%) conditional TAMV. Six (6.7%) children of the 89 had 1 or more arterial stenosis; each of those had 1 infarct. Overall, 62 (69.7%) participants had 1-3 SCI, mean age 6.6±1.5 and 29 (46.8%) male. This subset included all of those with an abnormal TCD and all but 1 with conditional. Children with SCI had marginally higher TAMV: 155.1±33.4 vs. 137.0±21.9cm/sec (p=0.06) compared to those without SCI. Initial trial HU dose of 20.2±0.9mg/kg was escalated to 25.9±2.7, per standard CBC criteria. At trial completion, 6 participants could not be re-imaged: 4 had reached a premature trial endpoint of stroke (3) or death (1) and 2 declined study continuation. Of the 83 re-imaged, 1 (1.2%) had abnormal TCD and 8 (9.6%) had conditional. Of the 27 without initial SCI, none acquired any MRI-MRA abnormalities. In contrast, 8 (12.9%) of the 62 with prior infarcts developed additional SCI over 30 months; 6 of these 8 also had vascular stenosis at trial entry. SCI progression was associated with lower TCD at trial completion: 113.6±11.0 vs. 136.0±27.1 cm/sec (p=0.027). Of the 62 pre-existing infarctsr and e-imaged, SCI progression with HU therapy was estimated at 3.6 per 100 patient trial years, demonstrating relatively modest evolution. Conclusions: At trial entry, the SCI prevalence in this sample of Ugandan children confirmed our prior findings of very high burden of cerebral vasculopathy. At trial completion, no participant with normal imaging at entry developed new infarcts or stenosis. This finding was associated with lower TCD but no other key clinical factors examined. Their lack of progression without pre-existing vascular injury is consistent with U.S. reports of HU prevention of SCI if treatment is initiated in young children without pre-existing SCI. For those with infarct(s) at trial entry, re-imaging demonstrated modest SCI progression despite HU therapy. SCI progression with lower TCD TAMV suggests that cerebrovascular architecture may eventually shift towards impaired blood flow or even neovascularization. On ethical grounds, we were unable to conduct a randomized, controlled trial to directly compare MRI effects of SCA with HU therapy over time. Our findings add to the existing literature demonstrating that early HU initiation protects children from SCA-associated brain injury and contributes unique insights into the impact of HU therapy in sub-Saharan Africa children.