Prevalence of transcranial Doppler abnormalities in Nigerian children with sickle cell disease

Transcranial Doppler (TCD) ultrasonography helps to identify children with sickle cell disease (SCD) who are at an increased risk of stroke, making primary stroke prevention a reality. A cross-sectional study of 145 Nigerian children aged ≥3 years with SCD was carried out to describe the pattern of cerebral blood flow (CBF) abnormalities. The mean time-averaged mean velocity (TAMV) was 152 ±27.0 cm/sec and 122 ±22.0 cm/sec in Hb SS and Hb S+C group, respectively. Abnormal velocities were recorded in six (4.7%) of the Hb SS patients and none of the Hb S+C while conditional risk (CR) velocities were recorded in 19.7% of Hb SS (low conditional 11.0%, high conditional 8.7%) and low conditional in 5.6% of Hb S+C cases. Cerebral flow velocities showed a negative correlation with age and hematocrit. Compared with African-American children, Nigerian children with Hb SS disease have a considerably higher prevalence of CR velocities. A major breakthrough in the field of SCD-related brain injury has been the success of TCD ultrasonography to identify asymptomatic children with SCD who are at an increased risk of stroke and the administration of chronic transfusion to prevent its occurrence [1, 2]. SCD is the most common inherited genetic disorder affecting Nigerians and its incidence in the country is amongst the highest in the world [3]. It is estimated that 20–25% of the Nigerian population carry the sickle cell gene and about 2–3% of children born annually have SCD, thus about 150,000 children are born annually with the disease [4, 5]. Reports suggest that the incidence of stroke in patients with sickle cell anemia in Africa is as high as 1.3/100 patient years [6, 7] compared with 0.6/100 patient years in the USA [8]. The estimated prevalence of stroke among Nigerian children with SCD is between 5 and 7% and it is estimated that 40–75% of children with SCD who suffer a first stroke will have a recurrence without any intervention [9-11]. Studies have shown the efficacy of TCD in identifying stenotic intracranial vessels in the circle of Willis and thus asymptomatic children who are at an increased risk of stroke [1, 12]. The narrowing of intracranial vessels in SCD results from fibrous proliferation of the intima and in some cases, thrombi formation. The lesions apparently progress for months or years before cerebral symptoms develop and so there is an opportunity to detect vascular narrowing before cerebral lesions develop. Flow velocities of ≥200 cm/sec in the middle cerebral or the internal carotid arteries has been reported to be associated with 40% risk of developing a stroke while flow velocities of <170 cm/sec and 170–199 cm/sec are associated with a stroke-risk of 2% and 7%, respectively [12]. Low arterial oxygen saturation <95% and a history of ≥3 febrile episodes in the preceding year were reported to predict a high CBF in a cohort of Kenyan children with SCD [13]. Reports have also shown a significant correlation between CBF velocities and stroke and age, hemoglobin concentration, and white count [8, 14]. There have been no previous reports on TCD studies in Nigerian children with SCD. This article presents our findings on the pattern of flow velocities in the major intracranial arteries as measured by TCD ultrasonography in a cohort of Nigerian children with SCD and the clinical and hematological parameters that are associated with increased TAMV in the major intracranial arteries. A total of 145 children with SCD were studied, 89 males and 56 females, with a male to female ratio of 1.5:1. All the children and caregivers who were introduced to TCD consented to have the test done giving 100% uptake of the test. The ages of the children ranged from 3 to 16 years, with a mean of 9 years, 2 months ± 4.0 years (110.0 ± 47.6 months). One hundred and twenty-seven (87.6%) children had homozygous Hb SS while the remaining 18 (12.4%) had hemoglobin S+C disease. There were no cases of HbS-β0/β+ thalassaemia and none of the children was on hydroxyurea or transfusion therapy. The hematocrit ranged from 16 to 35%, mean 23.6% ± 4.0% for the Hb SS cohort and from 23 to 35%, mean 28.8% ± 3.3% in the Hb S+C group. In the children with homozygous Hb SS disease, the total white blood cell (WBC) count ranged from 2,200 to 24,600 cells/mm3, with a mean of 10,448 ± 4,734 cells/mm3 and the absolute neutrophil count ranged from 1,215 to 15,060 cells/mm3, with a mean of 4,838 ± 2,751 cells/mm3. The platelet count varied from 160,000 to 861,000/mm3, with a mean of 344,975 ± 170,988/mm3. The mean arterial oxygen saturation (SaO2) was 98% ± 3.0% and this ranged from 91 to 100%. The TAMV ranged from 77 to 235 cm/sec in the 127 children with Hb SS, with a mean of 152 ± 27.0 cm/sec and from 90 to 180 cm/sec in the Hb S+C group with a mean of 122 ± 22 cm/sec. Normal velocities were recorded in 96 (75.6%) of the 127 children with Hb SS and 17 (94.4%) of 18 with Hb S+C. The highest velocities in each patient were recorded in the middle cerebral arteries and the internal carotid arteries in 95 (65.5%) and 50 (34.5%) of the cases, respectively. The range of flow velocities recorded in the intracranial arteries is shown in Table I. Of the 32 children with TAMV >170 cm/sec, 31 (96.9%) had Hb SS disease while only one (3.1%) had Hb S+C disease. Table II shows the pattern of TCD abnormalities in the children with HbSS and Hb S+C disease. All the six children with abnormal velocities had Hb SS. The study showed a negative correlation between hematocrit (P = 0.002) and age (P = 0.001) and TAMV among Hb SS cases. Higher TAMV were associated with a lower hematocrit and a younger age. There was no correlation with total WBC count, absolute neutrophil count, total platelet count, or arterial oxygen saturation (Table III). None of the variables was significantly correlated with TAMV among Hb S+C cases. Variables significant on bivariate analysis for children with Hb SS were entered into a multiple linear regression model. Age and hematocrit remained significant correlates for TAMV. A month increase in age was associated with a reduction in TAMV of about 0.22 cm/sec while the TAMV reduces by about 2.5 cm/sec for every unit increase in hematocrit. Identification of risk factors for stroke is clinically important because it offers the chance to prevent the first stroke in SCD [15, 16]; however, the present study did not allow to determine the predictive factors of abnormal velocities as it would be possible in a prospective cohort study [17, 18] where the presence of G6PD deficiency, the absence of α-thalassemia and the degree of hemolysis were retained as independent significant predictors. This is the first report on TCD studies in Nigerian children with SCD. Our study showed an excellent uptake of TCD by children with SCD and their caregivers. This can be attributed to the fact that TCD is painless, noninvasive and the scanning time is relatively short. Our study cohort showed a male to female ratio of 1.5:1 which may be a reflection of the superstitious and cultural high value placed on the male child in the African setting. The mean TAMV of 152 ± 27.0 cm/sec recorded in the Nigerian HbSS cohort is higher than the mean of 141 ± 35 cm/sec in the African American cohort [12] and the Creteil newborn cohort [17] in which velocities were maximum between age 4 and 5 years (140–145 cm/sec) and 130 cm/sec at age 9 years, which is the mean age in this study. The presence of TAMV ≥170 cm/sec in 24.4% of the Hb SS cohort is higher than the prevalence of 17.5% reported by Adams et al. [12] in a cohort of 315 African-American children, with 19.7% of the cases in this study being in the CR category, as opposed to 9% reported in the African-American cohort. The prevalence of CR in this study is however lower than the 27% reported in the prospective Creteil study [17]. It is however interesting that despite the higher stroke prevalence in African children with SCD, only 4.7% of the Hb SS cases had abnormal velocities, compared to 8% in the African-American cohort. The reasons for this disparity are not quite clear. Could it be that Nigerian children with SCD develop stroke at relatively lower velocities than has been previously reported or those with abnormal velocities had suffered a stroke prior to the availability of routine TCD screening? A long-term follow-up of the children with CR velocities may provide some answers to this question as conversion to abnormal velocities has been reported in a significant number of children with CR velocities and thus an increased risk of stroke [19]. Our findings are at variance with the report from Kenya [13] in which only 3 (2.8%) of the 105 children with sickle cell anaemia (SCA) studied had CR velocities and none had abnormal velocities. The reasons for this disparity are uncertain particularly as the Bantu haplotype which is associated with low levels of hemoglobin F and severe manifestations of SCA is known to be predominant in Kenya while the relatively less severe Benin haplotype predominates in Nigeria [20, 21]. Abnormal velocities were found in 4.7% of the children with Hb SS and despite appropriate education and counseling, none of the caregivers consented to the option of periodic blood transfusion which has been shown to confer 90% reduction in the risk of stroke in children with high risk velocities [15]. This observation may be related to the fact that this treatment plan is cost-intensive, rigorous, and time consuming and would impose tremendous burden on the meager resources of the families as payment for healthcare services in Nigeria are largely out-of-pocket. There is a need to explore other treatment options that would be cost-effective and efficacious in primary stroke prevention in children with abnormal velocities in countries with limited resources. It has been suggested that hydroxyurea may be a potential alternative for primary and secondary stroke prevention in SCD in resource-poor settings [22-26]. Zimmerman et al. [24] have shown that significant decreases occurred in CBF velocities in children with conditional velocities and its use was associated with minimal toxicities. Beneficial effects of hydroxyurea have also been reported in a cohort of Belgian children with abnormal velocities [25]. On the other hand, the SWiTCH (Stroke with transfusion changing to hydroxyurea) study was prematurely halted because seven strokes were recorded in hydroxyurea arm and none in the transfusion arm [27]. More controlled multi-center trials on the use of hydroxyurea would be extremely valuable in determining the efficacy of this drug in primary and secondary stroke prevention, particularly in Africa where blood is not readily available, quite expensive, and transfusions are associated with a potential risk of transmission of infections [28]. The stroke risk among patients with CR velocities is estimated to be 2–5%/year compared with 9%/year for children with abnormal velocities [1, 12]. However in spite of this lower relative risk of stroke in children with CR, more children have CR velocities such that the absolute number of children with CR who will develop stroke without therapy is comparable to those with abnormal velocities. In addition, children with CR velocities have a potential risk of converting to abnormal velocities by a progressive increase in their flow velocities over time to the high risk range [15, 19, 29]. This, therefore, makes the detection of CR velocities in one of five Nigerian children with Hb SS a major concern and may partly account for the relatively higher prevalence of stroke in Nigerian children with SCD. The study showed a significant negative correlation between age and hematocrit and TAMV. This is consistent with the findings on the Creteil cohort [17] which showed a progressive decline in flow velocities with increasing age after the age of 4–5 years and the majority of the children in our study were older than age 5years. Glucose 6 phosphate deficiency, anemia, and hemolysis have been reported as significant and independent risk factors for abnormally high velocities [18, 30]. CR velocities occur more frequently in Nigerian children with HbSS disease than their African-American counterparts. Low hematocrit and young age are associated with higher TAMV in children with SCD. All consecutive cases of SCD diagnosed by hemoglobin electrophoresis, aged 3 years and above seen at the Pediatric Hematology clinic of the University College Hospital (UCH), Ibadan, Nigeria over a period of 18 months, July 2009 to December 2010 were the study subjects. As routine TCD screening was just being introduced into routine care of children with SCD at the centre, information leaflets were given out to caregivers, together with health talks at every clinic visit by the public health nurses on the technique and benefits of routine TCD. Informed consent was obtained from all the caregivers who mainly included parents and in a few cases, close family relations who brought the children to the clinic. Children with SCD who were aged less than 3years and those who had a stroke were excluded from the study. Children who had recent blood transfusions, acute illnesses, such as painful crisis, fever, infections, and those who were hospitalized were not screened until they recovered completely. Children were only considered eligible for TCD at least 4 weeks after they were certified as having recovered from an acute illness and at least 4 weeks after the last transfusion. At enrolment, the following information was obtained from all the cases: name, age, gender, age at diagnosis of SCD, frequency of vaso-occlusive crisis, previous hospitalizations, and blood transfusions. Their past medical histories and hospital records were reviewed to identify the complications of SCD that had occurred in each child. Anthropometric measurements were taken, arterial oxygen saturation (SaO2) was measured using a pulse oximeter, and blood samples were drawn into EDTA bottles for full blood count estimation prior to the TCD examination at the same clinic visit. One of the investigators (IL) had received 6 weeks training on the technique of TCD ultrasonography at the Royal London Hospital, London, United Kingdom, under the tutelage of one of the co-investigators, PT. All the examinations were done at the Pediatric Hematology clinic of the UCH, Ibadan, Nigeria using a Doppler-Box™ with DWL Doppler system and QL software. The technique was not an imaging technique but similar to the Nicholette used by Adams in the STOP trial [15]. The measurements were taken following the protocols for the Stroke Prevention in sickle cell disease study (STOP) [15]. With the patient in the supine position, a 2 MHz hand-held probe was placed in the temporal region of the scalp. By manipulating probe angulation, probe position and the depth setting of the instrument, the ultrasound sample volume was placed on major arterial segments of the circle of Willis. The TAMV were measured in the middle cerebral arteries, the internal carotid arteries, the anterior cerebral arteries, and the posterior cerebral arteries in centimeters per second. Multiple measurements were taken at varying depths on either side and for each vessel, between 40 and 60 mm. The highest recorded mean velocity for each artery was assumed to be the most representative and this was recorded as the time-averaged mean of the maximum velocity (Vmax). All the measurements were recorded. In cases with elevated velocities of ≥170 cm/sec, a repeat TCD scan was done one week later and the child was taken as having conditional or abnormal velocities only if the mean flow velocity (MFV) remained so at the second examination. According to the STOP trial [15], MFV of <170 cm/sec was taken as normal, 170–184 cm/sec as low conditional, 185–199 cm/sec as high conditional, and ≥200 cm/sec as abnormal. All children with abnormal velocities were counseled on the preventive role of long-term periodic blood transfusion in stroke prevention. Acceptance or decline of this preventive measure was noted. The TAMV was correlated with age, hematocrit, total WBC count, absolute neutrophil count, platelet count, and oxygen saturation using Pearson correlation analysis. Comparison of means between groups was done using two-sided t-tests and F tests. Relationship between TAMV and clinical and hematological variables were explored with multiple linear regression analysis. Level of significance was set at P < 0.05. Ethical approval for the study was given by the Health Research Ethics Committee of the University of Ibadan. The authors thank Dr. B. Adedokun for the statistical analysis of the data and Mr. Afolabi for the laboratory support given for the project. The support of Dr. B.J. Brown and the resident doctors in the Department of Paediatrics, UCH, Ibadan, Nigeria is appreciated. IkeOluwa Lagunju Tel.: +234-8035857326. ilagunju@yahoo.co.uk* , Olugbemiro Sodeinde* , Paul Telfer §, * Department of Paediatrics, College of Medicine, University of Ibadan, Ibadan, Nigeria, Consultant Paediatrician, University College Hospital, Ibadan, Nigeria, Queen Mary University of London, London, United Kingdom, § Honorary Consultant Haematologist, Barts and The London NHS Trust, London, United Kingdom.