Cerebral Infarcts and Cerebrovascular Disease in Neurologically Intact Tanzanian Children with Sickle Cell Anaemia

Introduction:

Children and adolescents with sickle cell anemia (SCA) without history of neurological manifestations remain at risk of stroke and require cost effective screening as evidence-based stroke prevention interventions are developed. Although there are data on the prevalence of Transcranial Doppler (TCD) abnormality, there are relatively few studies on the prevalence of silent cerebral infarction (SCI) on magnetic resonance imaging (MRI) and vasculopathy on MR angiography (MRA), paricularly in Africa, where the burden of disease is greatest. In this study in Africa, we determined the prevalence of SCI on MRI and cerebral vasculopathy on MRA, and explored associations with age, sex, internal carotid/middle cerebral artery (ICA/MCA) and basilar cerebral blood flow velocities (CBFV) on TCD, and hematological variables.

Patients and methods:

We prospectively studied children with homozygous SCA (HbSS) without prior clinically overt stroke or TIA or seizures. Clinical information and blood for full blood count was collected for all patients. All were offered TCD using non-imaging equipment (Compumedics) and MRI and MRA of intracranial arteries on a Philips 1.5 Tesla scanner. MRA was graded as 1 Turbulence, 2 Stenosis, 3 Occlusion, 4 Occlusion with collaterals (moyamoya). Imaging was reviewed by 2 neuroradiologists (MJ and DS) and consensus was reached.

Results:

Within a period of 12 months, 395 children with SCA were recruited. Mean age was 12.7+/-4 (range 5-19 years); 199 (50.4%) were male. Mean hemoglobin was 7.6±1.1 (range 4.1-13.5) g/dl. 381 had TCD, of whom 227 (57%) have had brain MRI and MRA so far.

Only 14/395 (3.5%) of patients had abnormal CBFV, 8 (2%) with maximum CBFV <50 cm/sec, 3 with CBFV 150-169 cm/sec, 1 with conditional (170-199 cm/sec) and 2 (0.5%) with abnormal CBFV (>200cm/sec).

Prevalence of silent cerebral infarction was 29% (65/227) and was similar in males (30%; 34/114) and females (28%; 31/110) but was lower in children aged <9 years (20%; 10/51) than in those aged 10-14 years (33%; 27/83) or those aged 15-19 years (32%; 28/89); this was not significant (p=0.2).

The prevalence of cerebral vasculopathy on MRA, defined as stenosis (Grade 2), occlusion (Grade 3) or occlusion with collaterals (moyamoya; Grade 4), was 9% (20/227). The internal carotid and/or middle cerebral artery was affected in all 20 and the anterior cerebral artery was affected in 7 patients. Vasculopathy was not observed on MRA in the posterior cerebral artery and no patients had moyamoya. An additional 9 (4%) patients had turbulence (Grade 1) in at least one artery on MRA.

SCI were more common in those with MRA vasculopathy Grades 2-3 (12/64; 19%) and in those with Grade 1 Turbulence (2/9; 22%) than in those with normal MRA (8/165; 5%) (p<0.05). Patients with Grade 1 Turbulence had higher ICA/MCA CBFv (mean 122+/-36 cm/sec) than those with normal MRA (mean 101+/-25 cm/sec) or Grades 2-3 (mean 109+/-21 cm/sec) vasculopathy. The difference in those with Grade 1 turbulence compared with those with normal MRA was significant (mean difference 20 cm/sec, 95% confidence intervals 1, 39 cm/sec; p=0.037). Basilar velocity was also significantly higher in those with Grade 1 turbulence (87+/-21 cm/sec) than in those with normal MRA (76+/-14 cm/sec) and those with Grade 2-3 vasculopathy (80+/-11 cm/sec)(mean difference 11.5 cm/sec, 95% confidence intervals 0.3, 22.6 cm/sec; p=0.043). Basilar velocity weakly correlated with hemoglobin (r=-0.18) but the relationships between hemoglobin and ICA/MCA CBFV or either presence of SCI or MRA abnormality were not significant.

Conclusions:

The prevalence of SCI on MRI is high in children with SCA without neurological history living in Africa even when TCD CBFV is normal. Children and adolescents with all grades of vasculopathy on MRA are at higher risk of brain parenchymal injury. The lack of association between SCI and hemoglobin may be related to the relatively severe anemia in our African study. As MRA contrast depends on velocity of blood flowing in vessels, it is not surprising that Grade 1 turbulence on MRA appears to reflect high CBFV on TCD. Importantly, high TCD velocity and Grade 1 turbulence may reflect the potentially reversible early stages in the development of SCA vasculopathy when cerebral blood flow is high before stenosis and occlusion supervene. TCD is operator dependent and there may be a case for MRI and MRA screening if follow-up shows that these measures predict stroke.