Cerebral blood flow trajectories in paediatric sickle cell anaemia by age, region, and treatment associations

Abstract

Opposite to the age trend in healthy children, cerebral blood flow increases with age in children with sickle cell anaemia. Early treatment with hydroxyurea (hydroxycarbamide) may help slow this abnormal trend and protect brain functions.

To the Editor,

Sickle cell anaemia (SCA) is caused by mutations in the β‐globin gene that lead to haemoglobin polymerization under hypoxic conditions, producing sickled red blood cells prone to haemolysis and vascular occlusion, leading to ischaemic complications, including in the brain. ¹ , ² In SCA, chronic anaemia reduces arterial oxygen carrying capacity, prompting compensatory increases in cerebral blood flow (CBF) to maintain adequate oxygen delivery to brain tissue. ³ However, elevated CBF is linked to a risk of stroke, silent cerebral infarcts (SCI) and neurocognitive impairment. ² , ⁴ , ⁵

In healthy children, CBF rises rapidly after birth, peaking around age 6 to meet the high metabolic demands of synaptogenesis, myelination and neural network formation. CBF then declines through adolescence with synaptic pruning and plateaus at adult levels. These developmental changes in CBF are regionally distinct, reflecting differences in maturation rates and functional demands in different brain regions. ⁶ , ⁷

Although elevated CBF in SCA is well established, little is known about age‐related and regional variations in children. Moreover, the impact of hydroxyurea (hydroxycarbamide), a key disease‐modifying therapy for children with SCA, on developmental trajectories of CBF remains unknown. Hydroxyurea increases fetal haemoglobin as well as total haemoglobin, ⁸ reduces vaso‐occlusive episodes and SCI risks ⁸ , ⁹ and has been linked to lower cerebral metabolic stress ¹⁰ and improves cognition ¹¹ , ¹² in patients with SCA. It may lower CBF by decreasing compensatory hyper‐perfusion through better oxygen delivery with reduced anaemia. ¹²

We hypothesised that children with SCA have age‐related CBF patterns distinct from healthy controls. We also aimed to characterise temporal and spatial CBF patterns in children with SCA and assess hydroxyurea treatment effects within this context.

This prospective, non‐therapeutic, non‐randomized longitudinal imaging study was approved by the Institutional Review Board of St. Jude Children's Research Hospital. Written informed consent was obtained from the guardians, and children provided assent to participate. Patients aged 6–19 years with HbSS or HbS/β⁰‐thalassaemia were enrolled. Exclusion criteria included contraindication to undergo an magnetic resonance imaging (MRI), sedation needed for MRI, prior hydroxyurea initiation or transfusion therapy before baseline and history of stem cell transplant, myelosuppressive therapies or an overt stroke.

Participants underwent baseline (TP1) and 1‐year follow‐up (TP2) assessment, including MRI, blood tests, neurocognitive testing and transcranial Doppler (TCD). Hydroxyurea treatment, when indicated, was initiated after TP1 at 20 mg/kg/day and escalated to the maximum tolerated dose.

Thirty‐two children with SCA were included in the analysis (median age: 12.3 years, range: 7.2–17.8; 16 females). At TP1, five participants (16%) had conditionally elevated TCD velocities (170–199 cm/s), which decreased to two participants (7%) by TP2; all others had normal TCD velocities. SCI was present in 19 children at TP1, with three new cases identified at TP2 (Table S1). Of the 32 children, 21 adhered to prescribed hydroxyurea and were assigned to the treatment group; 11 were classified as untreated due to non‐initiation or poor adherence. The hydroxyurea treatment group partially overlaps with a previously published cohort. ¹²

Thirty‐two age‐ and gender‐matched healthy controls (median age: 12.5 years; range: 6.6–18.3; 16 females) were selected via propensity score matching from the cohort of our previously reported study on CBF developmental pattern in healthy children. ⁷

CBF was computed ⁷ , ¹² from the arterial spin labelling (ASL) images acquired with a Q2TIPS sequence ¹³ on a 3T Siemens MRI scanner and co‐registered to high‐resolution 3D T1 images for segmentation. Grey and white matter CBF values were extracted for bilateral frontal, parietal, temporal and occipital lobes. Image analysis was performed using custom‐built Matlab programs and SPM12 software (www.fil.ion.ucl.ac.uk/spm/software/spm12).

Baseline TP1 data were used to model age‐related trends and regional CBF differences. Pearson correlations assessed left–right and grey–white matter relationships; repeated‐measures ANOVA with post hoc t‐tests evaluated regional variations; permutation‐based Welch's two‐sample t‐tests compared CBF values between SCA and healthy groups; and multivariable linear models compared trends in CBF with age for SCA and healthy groups.

A linear mixed‐effects model assessed age‐related CBF trajectories in treated, untreated and healthy groups comparing data from TP1 and TP2. Permutation‐based Welch's t‐tests compared CBF among the groups.

All p‐values were Bonferroni‐corrected for multiple lobar comparisons using a predetermined significance level of α = 0.05. Statistical analyses were performed in R (v4.1.2, www.r‐project.org).

We found that, in children with SCA, grey and white matter CBF were highly correlated, with consistently higher values in grey matter (Figure S1). Left–right hemisphere CBF values were also strongly correlated (Figure S2), indicating no asymmetry across regions. Similar to healthy children, occipital lobes had the highest CBF values (p < 0.0001) among all brain lobes in children with SCA. However, children with SCA showed significantly higher CBF in frontal (p = 0.004) and parietal (p = 0.0008) lobes than healthy children (Table S2).

Children with SCA had significantly different (p ≤ 0.001) CBF‐age slope, that is, the rate of CBF change with age, than healthy children across brain regions (Figure 1A,B). While CBF decreases with age in healthy children, it increased significantly with age (p < 0.001), with a rate of 2.22 mL/100 g/min per year in the whole brain grey matter in children with SCA. Regression lines for the two groups crossed between ages of 9 and 11 years, after which group differences widened through adolescence. White matter CBF in SCA followed a similar trend (Figure S3).

FIGURE 1.

Age‐related changes, regional variations and hydroxyurea treatment effects on CBF. (A) CBF maps from four representative children with SCA. Each was the median‐aged patient from each of the age quartile at baseline. CBF shows a progressive increase with age. (B) CBF‐age trajectories. Grey matter CBF for all participants at baseline are shown. In children with SCA, CBF increased with age across brain regions: Frontal lobe +2.44 (p = 0.002), temporal lobe +2.01 (p = 0.01), parietal lobe +2.79 (p < 0.001) and occipital lobe +2.44 (p = 0.04) mL/100 g/min per year. In healthy controls, CBF declined with age: Frontal lobe −1.29 (p = 0.10), temporal lobe −2.06 (p = 0.01), parietal lobe −1.34 (p = 0.08) and occipital −3.08 (p = 0.01) mL/100 g/min per year. Shaded areas in the graph represent 95% confidence intervals. CBF‐age slopes, that is, rates of CBF change per year, differed significantly between SCA and healthy in all lobes (p ≤ 0.001). (C) Hydroxyurea treatment effects on CBF–age trajectories. At TP1, the CBF‐age slopes did not differ significantly between treated and untreated groups in both grey matter and white matter. By TP2, the CBF‐age slopes in the treated children with SCA declined in both grey and white matters and even became negative in white matter, approaching healthy children's trajectories, while the CBF‐age slopes in the untreated patients remained unchanged. Differences between treated and untreated groups were significant (grey matter p = 0.009; white matter p = 0.001). Grey dashed lines connect data from the same subject across the two time points. CBF, cerebral blood flow; SCA, sickle cell anaemia.

Hydroxyurea treatment effects are shown in Table 1 and Figure 1C. Mean CBF values did not differ between treated and untreated SCA groups at TP1, but both exceeded healthy controls in specific lobes: hydroxyurea‐treated children in the frontal (p = 0.008) and parietal (p = 0.004) lobes and untreated children in the parietal lobe (p = 0.008). By TP2, both SCA groups exhibited significantly higher mean CBF in all four lobes compared to healthy controls (p < 0.001).

TABLE 1.

Effects of hydroxyurea treatment on CBF in different brain regions.

	Brain lobes	TP 1 (baseline)						TP2 (12 months)
		Healthy children	TX a	unTX a	Adjusted p‐values d			Healthy	TX	unTX	Adjusted p‐values d
					TX vs. unTX	TX vs. healthy	unTX vs. healthy				TX vs. unTX	TX vs. healthy	unTX vs. healthy
Mean CBF b (mL/100 g/min)	Frontal	47.43	59.38	61.24	ns	0.008	ns	37.40	61.35	63.62	ns	<0.001	<0.001
	Occipital	77.55	83.38	96.20	ns	ns	ns	59.67	78.63	103.11	0.001	<0.001	<0.001
	Parietal	48.82	62.25	67.43	ns	0.004	0.008	37.96	61.12	69.83	ns	<0.001	<0.001
	Temporal	56.57	62.91	66.57	ns	ns	ns	45.01	61.83	66.77	ns	<0.001	<0.001
CBF‐age slopes c ([mL/100 g/min]/year)	Frontal	−1.41	2.16	3.41	ns	0.003	0.002	−1.42	0.15	4.67	0.03	ns	0.001
	Occipital	−3.31	1.93	4.14	ns	0.003	0.001	−1.97	−0.88	2.44	ns	ns	ns
	Parietal	−1.46	2.68	3.18	ns	0.0003	0.003	−1.62	0.34	5.06	0.02	ns	0.0002
	Temporal	−2.26	1.80	2.78	ns	0.0008	0.002	−2.35	0.46	3.20	ns	ns	0.004

Note: At TP1, mean CBF did not differ between treated and untreated SCA groups but was higher than in healthy children in the parietal lobes; frontal CBF was elevated only in the treated group. CBF–age slopes were similar between SCA groups but differed significantly from healthy controls. At TP2, both SCA groups had higher CBF than healthy controls across all regions. In the treated children with SCA, CBF–age slopes declined and no longer differed from healthy children, whereas in untreated children with SCA, their slopes increased further, reaching significance in all lobes except the occipital lobe.

Abbreviations: CBF, cerebral blood flow; SCA, Sickle cell anaemia.

^a TX: SCA group treated with hydroxyurea; unTX: SCA group untreated with hydroxyurea.

^b Permutation Welch test was used to compare the CBF values between the groups.

^c Linear mixed effect model with random intercept was used for the analysis.

^d p‐values were adjusted with Bonferroni correction for multiple comparisons; ns = not significant.

At TP1, CBF‐age slopes did not differ between hydroxyurea‐treated and ‐untreated SCA groups, but differed significantly from healthy controls across all lobes (p ≤ 0.003). By TP2, CBF‐age slopes no longer differed from healthy controls in the treated group, but remained significantly higher (p ≤ 0.001) in all lobes, except the occipital lobe, in the untreated group. Region‐specific effects were noted: untreated children showed a higher parietal CBF‐age slope (p = 0.02), while treated children developed a negative occipital slope. Treatment effects were more prominent in white matter, where CBF‐age slopes became negative across all lobes in treated children by TP2—approaching the healthy trajectory—while remaining positive in the untreated group (additional details in Figures S4 and S5 and Table S3).

The contrasting CBF–age trend in children with SCA likely reflects progressive compensation for chronic anaemia and impaired oxygen delivery. Given the steep decrease of CBF before puberty in healthy children ⁶ , ⁷ and the heightened stroke risk in children with SCA under age 10 years, ² , ⁵ the crossing point at around this age may represent a critical period for cerebrovascular pathology. This period may also signal heightened vulnerability to neurocognitive decline, as slowed cognitive growth and deficits in school‐aged children with SCA were reported. ¹¹ These findings underscore the importance of initiating effective therapy early in life to mitigate disease progression and prevent cerebrovascular injury.

Although prior studies have suggested neuroprotective effects of hydroxyurea, ¹⁰ , ¹² reported whole‐brain CBF reductions were often not significant. ¹⁰ , ¹² We show that, without considering the age trend, CBF values across the brain appeared significantly higher in both treated and untreated SCA groups at TP2 (Table 1), potentially masking treatment effects, particularly in paediatric patients. However, hydroxyurea significantly attenuated age‐related CBF increases across all regions (Table 1), with stronger effects in white matter, which is particularly vulnerable to ischaemia ¹⁰ and linked to disrupted connectivity underlying cognitive deficits. ¹⁴ Our findings support the neuroprotective role of hydroxyurea by attenuating abnormal cerebral perfusion and suggest that emerging gene‐based therapies that elevate fetal haemoglobin and total haemoglobin levels ¹⁵ may yield similar cerebral benefits. We also suggest the age‐ and region‐specific CBF profiles should be considered in assessing treatment efficacy.

Our results also highlight concerning trends in children with SCA who had no treatment or subtherapeutic exposure. Their CBF increases with age and progressively diverges from healthy trajectories across brain regions, even if they have normal TCD velocities, as TCD may not capture tissue‐level perfusion or reliably predict SCI risk. Our findings suggest silent progression of haemodynamic abnormalities during critical brain development. Such insults may underlie cumulative neurological injury and cognitive vulnerability, reinforcing the need for early interventions, even before conventional thresholds, like abnormal TCD, are met.

A limitation of this study is the small sample size (n = 32) and restricted age range (7–18 years), which limits generalizability and precludes assessment of early childhood CBF trajectories. The age restriction reflects the challenges of performing unsedated MRI in young children. Future studies with larger cohorts, including infants and preschoolers, are needed, as this critical period of neurodevelopment may yield important insights into early disease pathophysiology and long‐term outcomes.

In conclusion, we characterised age‐related and regional CBF changes in children with SCA and demonstrated that hydroxyurea treatment modulates these trajectories. Detailed CBF mapping is valuable for evaluating SCA therapies and may aid in risk stratification and early treatment decisions.

AUTHOR CONTRIBUTIONS

Ping Zou Stinnett: Conception, study design and implementation, image acquisition and analysis, data analysis and visualization, result interpretation, manuscript writing; Robert J Ogg: Conception, study design and implementation, image acquisition; Kathleen Helton: Conception and study design; Winfred Wang: Study design, patient treatment and reviewing, result interpretation, manuscript editing; Andrew M. Heitzer: Introduction and manuscript editing; Zachary Abramson: Result interpretation and discussion, manuscript editing; Yimei Li: Statistical modelling and data analysis; Tushar Patni: Statistical modelling and data analysis; Paul VanGilder: Manuscript editing; Akshay Sharma: Clinical correlation of the data and introduction, critical review and manuscript editing; Ranganatha Sitaram: Project supervision, review of findings, discussion and manuscript editing.

FUNDING INFORMATION

This study was supported by the Cancer Center Support (CORE) grant CA21765 from the National Cancer Institute, grant RR029005 from the National Center for Research Resources, and by ALSAC, the fund‐raising arm of St. Jude Children's Research Hospital. Andrew M. Heitzer was supported by K23HL166697 from the National Institutes of Health, National Heart, Lung, and Blood Institute (NHLBI). Akshay Sharma is also supported by a 1U01HL163983 from the National Institutes of Health, NHLBI. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

CONFLICT OF INTEREST STATEMENT

Andrew M. Heitzer received consultancy fees from Global Blood Therapeutics. Akshay Sharma has received a consultant fee from Spotlight Therapeutics, Medexus Inc., Vertex Pharmaceuticals, Sangamo Therapeutics and Editas Medicine. He is a medical monitor for the RCI BMT CSIDE clinical trial for which he receives financial compensation. He has also received research funding from CRISPR Therapeutics and honoraria from Vindico Medical Education. AS is the St. Jude Children's Research Hospital site principal investigator of clinical trials for genome editing of sickle cell disease sponsored by Vertex Pharmaceuticals/CRISPR Therapeutics (NCT03745287), Novartis Pharmaceuticals (NCT04443907) and Beam Therapeutics (NCT05456880). The industry sponsors provide funding for the clinical trial, which includes salary support paid to Akshay Sharma's institution. Akshay Sharma has no direct financial interest in these therapies. The other authors have no conflicts of interest to disclose.

ETHICS APPROVAL STATEMENT

The study was approved by the institutional review board of St. Jude Children's Research Hospital.

PATIENT CONSENT STATEMENT

Written informed consent was obtained from the guardians, and children provided assent to participate.

Supporting information

Figure S1.

Figure S2.

Figure S3.

Figure S4.

Figure S5.

Table S1.

Table S2.

Table S3.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding authors upon reasonable request and with appropriate institutional approvals.