Sickle cell disease (SCD) is a cause of strokes in children [1] [2]. Strokes occur in 7 – 11.5% of children with SCD, at an average age of 4 – 9 years, and reoccur in up to 50% cases [1] [2] [3] [4]. Deaths and severe neurologic deficits of hemipareisis, aphasia, swallowing incoordination, learning difficulties, loss of cognitive function and seizure disorders have been reported after the first stroke [1] [2] [3] [4].

The few reported predictors of stroke in SCD include Hb SS genotype (SS anaemia), low steady state haematocrit, leucocytosis, elevated systolic blood pressure, arterial hypoxaemia and previous episodes of acute chest syndrome [1] [2]. Transcranial Doppler (TCD) ultrasound studies of the brain, a simple non-invasive tool for assessing large cerebral artery flow dynamics, has been used to identify children with SCD at risk of stroke [2] [5].

Stroke in SCD affects the large vessels of the arterial circulation of the brain [1] [2] [4] [6]. There is progressive occlusion of these arteries by fibrin proliferation and thrombi leading to endoluminal narrowing. This causes fast and turbulent blood flow within the vessels, measurable by TCD ultrasound waves [2] [4] [6]. Studies show that increased blood flow velocities in the internal carotid and middle cerebral arteries correlate with an increased risk of stroke [5] [6] [7] [2]. Generally, the time averaged mean velocities (TAMV) of blood flow in these vessels measured by TCD of ≥ 200 cm/sec indicate a significant risk of stroke of 40%, while velocities of 170 - 199 cm/sec and < 170 cm/sec show a risk of 7% and 2%, respectively [2] [7].

Although TCDs cannot predict all strokes [2] [4] [6] [7], they offer an opportunity to identify children at risk for strokes and reduce the number of first strokes by up to 90% by enabling appropriate therapy such as chronic blood transfusions, as determined by the Stroke Prevention Trial in SCD (STOP) [8]. TCD studies is therefore the only recommended screening tool for primary stroke prevention in SCD by the National Heart Lung and Blood Institute of the US National Institute of Health, and the American Heart Association/American Stroke Association [2] [9].

Few reports of the use of TCDs for primary stroke prevention in SCD have emanated from Africa. These include those by Lagunju [10] in Nigeria and Makani [11] in Kenya. In most of the developing world, the use of TCDs as a screening modality for primary stroke prevention in SCD is not yet a routine standard of care. This study aimed to identify children with SCD at risk of strokes using TCD ultrasound studies of the cerebral vessels of the brain, to determine the prevalence of and pattern of TCD ultrasound abnormalities, evaluate the relationship of findings with age, and degree of anaemia, as well as offer treatment as required. It is hoped that this information will increase awareness to the use of TCD ultrasound studies of the cerebral vessels as a primary prevention screening tool for strokes in SCD.

This is a descriptive prospective study of children with SCD at the age of 0 - 16 years seen at the sickle cell clinic of the paediatric unit of the National Hospital Abuja, a referral tertiary center located in the capital city of Nigeria over a 4-year period from January 2009 – January 2013. Sickle cell disease was diagnosed by cellulose acetate haemoglobin electrophoresis methods at the Haematology unit of the hospital.

Parents and caregivers of children with SCD were informed about TCDs at the sickle cell clinic by the health care personnel. Those children 2 years and older were routinely requested to undergo TCDs as part of the standard of care of the clinic. The cost of the TCDs was borne by the parents and care givers of the children. Therefore, only those children whose parents could afford the TCDs had the tests done. The TCDs were done when the children were in steady state; and those who had strokes previously were exempt from it. We define steady state in SCD as that period when the children are in an optimal state of health, free from all forms of crises or other acute complications of SCD in the preceding month; and stroke as an acute neurologic syndrome secondary to occlusion of a cerebral artery or haemorrhage with resultant ischaemia and neurologic signs and symptoms lasting greater than 24 hoursm as defined by the Cooperative Study in Sickle Cell Disease (CSSCD) [1]. Children already on hydroxyurea or receiving chronic blood transfusions (CTT) were also excluded from the study, as both therapies have been shown to influence TCD results.

The following information was obtained from the medical notes of children: demographic data and steady state haematocrits as measured by the Packed Cell Volume (PCV). The PCVs were determined by Haematology auto analyzer (Sysmex KX-21N) carried out at the Haematology Department of the hospital following manufacturer’s instructions.

The transcranial Doppler ultrasound studies (TCDs) of the cerebral vessels were carried out by a Consultant Radiologist. The examinations were done using an imaging Doppler ultrasound machine (Sonoace X6, Medison Co. Ltd, Seoul, Korea. 2011). Blood flow velocities were assessed over the internal carotid (ICA), middle cerebral (MCA), anterior cerebral (ACA), posterior cerebral (PCA) and basilar arteries. The procedure for the TCD studies was as follows: The child laid supine, neck extended backwards (over a pad put under the neck to aid the extension). The chin was turned opposite to the side of the head that was to be examined for the trans-temporal approach to assess the ICA, MCA, ACA and PCA. Both sides of the head were examined. For the trans-foraminal approach to examine the posterior cerebral circulation (mainly for basilar artery) the child laid on his/her side, the neck was flexed with the chin pointing to or touching the chest. The children were not sedated.

Using a 2MHz probe, (2.5MHz to 5MHz multi-frequency Imaging & Doppler transducer) placed at the temporal area of the head, the internal carotid artery (ICA) was identified, and scanned moving in 2mm increments, the probe’s depth and angle being adjusted to get the most representative velocity spectrum. The sample volume size selection was between 4 to 6mm. The same steps were taken to identify and scan the MCA, ACA, PCA and basilar arteries. The higher mean flow velocity (MFV) for each artery was recorded. The results were printed out graphically and documented.

The STOP trial [8] had utilized time averaged mean velocities (TAMV) over the proximal portion of the MCA for stratification of stroke risk. This study utilized the mean flow velocity (MFV) measured over the same portion of the MCA for that purpose. Therefore, TCD findings (based on the higher recording on at least one side of the examination) in the children were classified according to the STOP [8] trial thus: Normal - when the TCDs were < 170 cm/sec, Conditional – when the TCDs were 170 -199.9 cm/sec, Abnormal – when the TCDs were ≥ 200 cm/sec, and Inadequate – when the scan could not be interpreted due to inadequate examination for any reason (related to the patient or machine).

Therapy offered the children based on the TCD results over the MCAs following recommended guidelines [2] [9] [12] were: Normal TCD results of < 170cm/sec: followed up in the clinic and requested to have yearly TCDs [2] [9] [12] ; Abnormal TCD results of ≥ 200 cm/sec (high risk): offered preventive therapy of chronic transfusion therapy (CTT), and to have a repeat TCD in 6 months [2] [9] [12] and Conditional TCD results of 170 – < 200 cm/sec: followed up in the clinic and repeat TCDs in 1 - 3 months [12].

Children with conditional or abnormal TCD results were offered Chronic Transfusion Therapy (CTT), according to the guidelines stated above. If they declined CTT, they were commenced on hydroxyurea therapy, after counseling the parents and caregivers. Hydroxyurea was commenced at 15 mg/kg and gradually increased to 30mg/kg as the child tolerated the drug. Patients were closely followed up in the clinic with initial 2 weekly, then monthly and extending to 2 monthly visits. At each visit, they had clinical evaluations, and laboratory investigations of Full Blood Count, as well as Liver and Kidney functions tests of serum alanine amino transferase and creatinine levels respectively.

For those with inadequate results, the TCDs were repeated at a later date at no cost to the patient.

The results were presented as simple frequencies and percentages and shown in tables. Descriptive statistics were expressed as means +/- 2 standard deviations. Statistical evaluations were done by Microsoft Excel 2007 and Statistical Package for Social Sciences SPSS version 20. Associations between variables were done by the chi square test as appropriate. A p value of less than 0.05 was considered statistically significant.

There were 129 children, 67 males and 62 females (M:F=1.08:1) aged 1 year 7 months to 16 years, (mean 7.0 years, SD 3.5) who had TCDs done in the period under review. They comprised only 24.0% of the 537 children registered in the clinic. The haemoglobin genotypes were Hb SS 127(98.4%) and Hb SC 2 (1.6%). The Packed Cell Volumes (PCVs) ranged between 13 – 33% (mean 24.1%, SD 3.6).

The Mean Flow Velocities (MFVs) of the TCDs, based on the higher recording of both MCAs were classified as normal (< 170cm/sec) 105 (81.4%), conditional (170 - < 200 cm/sec) 15 (11.6%) and abnormal (≥ 200cm/sec) 9 (6.9%) (Table 1). There were 12 (9.3%) children who had markedly discrepant (50% and greater) TCDs between the left and right MCAs; and 21(16.3%) children with low TCDs < 70 cm/sec. There were 13/15 (86.6%) and 7/9 (77.7%) children with conditional and abnormal TCD velocities respectively in the age group 3 – 8 years, although the finding regarding this age group was not statistically significant. There were no significant differences in the TCDs to age, sex or PCVs (Table 2).

Of the treatment modalities offered the children, there was zero (0%) uptake of chronic blood transfusions for abnormal and conditional TCDs despite counseling offered parents and caregivers of affected children by doctors and nurses in the clinic. The drug hydroxyurea was accepted by 23 out of 24 (95.8%) parents and caregivers of these children as an alternate treatment modality.

This study has looked at the pattern of Trans cranial Doppler ultrasound studies (TCDs) of the cerebral vessels as screening for primary prevention of stroke in children with sickle cell disease (SCD) in steady clinical state attending a tertiary hospital in Nigeria. While this intervention is not yet a routine standard of care for children with SCD in Nigeria, the Federal Government of Nigeria is currently proposing to commence TCD screening at the recently established Centers of Excellence for SCD in the country. It is however hoped that the TCDs will be provided free at these COEs as the cost of SCD management is at present borne exclusively by parents/caregivers of those affected.

The sex distribution of M:F 1.08:1 of the subjects of this study is similar to previous reports of sickle cell clinic populations in Nigeria; as is the mean steady state hematocrit – Packed Cell Volume (PCV) levels of 24.1% (range 13 – 33%, SD 3.6) [10]. Almost all (98%) of the children had SS anaemia, and therefore the study results can only represent this form of SCD.

Operator, patient or machine variables may affect the measurement of TCDs [2]. While operator and patient variables were optimized in our study, machine variables were different from the STOP trial [8], which had utilized conventional TCD ultrasound machine in the measurement of cerebral vessel velocities. The present study has utilized duplex power Doppler ultrasound studies, an imaging TCD (TCDI) modality for the measurement of cerebral vessel velocities similarly following the protocol of the STOP trial. The use of TCDIs for this purpose has also been previously reported [13]. TCDI, when compared to conventional TCD ultrasound studies, provide direct visualization of blood vessels, is easier to use with more technical variability, and can provide equivalent predictive results depending on the insonation angle [13] [14]. However, vessel velocities from TCDI have ranged between similar values obtained from conventional TCDs to up to 20% lower [13] [15] [16]. Furthermore, validation of the results from TCDIs in comparison to conventional TCDs has been carried out [13] [15]. However, the differences in imaging modality may limit comparisons been made between TCD velocities measured by the various methods.

The prevalence of abnormal TCDs from this study was 6.9%. This is higher than the previous report by Lagunju from Nigeria [10] of a prevalence of 4.5% abnormal TCDs among children with SCD; or from Sao Paolo, Brazil [17] which found a prevalence of only 1.6%. In the study by Makani from Kenya, none of 105 children screened had abnormal results [11]. It is however lower than 9.3% prevalence of abnormal TCDs found in the STOP [8] trial and 12.3 - 34% reported from the German [18], France [19] or African American [5] SCD cohorts studies. However, we note that the criterion for abnormal TCDs in the African American study [5] was 170cm/sec, and the prevalence of about 20% abnormal TCDs from that study was inclusive of children with velocities > 170cm/sec. Furthermore, the sample size of 47children from the Germany [18] study was rather small, and may not be truly representative of children with SCD. The 11.6% conditional TCDs found in our study is lower than the prevalence of 24.4% earlier reported from Nigeria [10] as well as the 17.5% and 26.7% found in the African-American [5] and French [19] cohort studies respectively. On the other hand, it is much higher than the prevalence of 2.8% and 8.1% conditional TCDs reported from Kenya [11] and Brazil [17] respectively. The relevance of these conditional TCDs is quite significant, as long term follow up of children with such results have shown a conversion from conditional to abnormal TCDs in 9.4 - 34.5% of patients over a period of 9 months – 2 years [5] [6] [7] [18] [20]. The interpretation of these results would therefore be that an additional 11.6% of the children from this present study with conditional TCDs are at increased risk of stroke.

Overall, there appear to be differences in the pattern of TCD abnormalities found in our study to the findings from previous studies, even in similar geographic areas of the world. It has been suggested that genetic or environmental factors may contribute to these differences [4] [6] [9] [16]. This may influence the disparity of results from Nigeria as the earlier study [10] was from a location where majority of people were of a single ethnic group, while our study in the capital of the country had people of different ethnic groups. It may also be related to the different sickle cell haplotypes found in these areas [4] [6] [16]. However, these differences suggest that more standardized studies, preferably long term, are required to actually identify the proportion of children with SCD at risk of stroke in different locales.

There were 12 children with markedly discrepant TCDs (50% and above) of both MCAs, although these values were within normal limits (<170cm/sec). This implies there is relatively greater stenosis of the blood vessel on one side of the brain as compared to the other. As the pathophysiology of arterial occlusion in SCD suggests ongoing progressive lesions over time [4] [6] [16], it is likely that this relative stenosis may in fact become significant. There is paucity of information as to treatment guidelines for this group of children. Close follow up will be required as well as research to determine its best mode of treatment. Furthermore, our study identified 21 (16.3%) children with low TCD velocities < 70 cm/sec. Low TCD velocities may be indicated of severe arterial stenoses [16]. Thus, these affected children are at additional risk of stroke. They should be closely followed up with regular monthly TCDs as recommended by the Brazilian guidelines on primary prevention of stroke in SCD [12], with further investigations and/or interventions as required.

Knowledge of risk factors for abnormal TCDs in SCD would enhance identification of the child at risk for stroke. Negative correlations has been reported with haematocrit and age to rising velocities on TCDs [10] [16], meaning that younger children and those with lower haematocrits tend to have higher TCD velocities. Our study did not find this negative correlation. However we found that the age groups 3 – 8 years had the highest concentration of children with abnormal and conditional TCDs; which is quite similar to what was found in the CSSCD study [1] and STOP trial [8]. This distribution suggests emphasis be placed on screening children at these ages with TCDs if available resources are unable to screen all ages for risk of stroke in SCD.

Treatment was offered the children based on their TCD results in accordance with the recommended guidelines [2] [9]. None of the parents or caregivers of the children with abnormal TCDs accepted the treatment strategy of chronic blood transfusions, which is associated with up to 90% reduction in stroke risk in SCD and is the optimal standard of care [2] [8] [9] [12]. This observation was similarly reported by Lagunju [10] in the earlier paper from Nigeria on the subject. Our experience has been that long term blood transfusions are difficult to maintain even when required for secondary prevention of strokes in SCD, as blood is not readily available, expensive, time consuming, and creates a burden on the family [10] [21]. This emphasizes the need for other forms of therapy for primary stroke prevention in SCD.

The drug hydroxyurea has been found beneficial in SCD, reducing the rate of painful crises, acute chest syndrome, number and duration of hospital attendances and admissions, blood transfusions and mortality [4] [6] [9] [22]. In addition, trials have shown it is well tolerated, does not impair growth or development, and can be safely used in young children [23]. However, there are varying reports on the efficacy of hydroxyurea for primary prevention of stroke in SCD. In children with conditional and abnormal TCDs, it was found to significantly decrease TCD velocities into the normal and conditional range; thereby decreasing the risk for stroke [24] [25] [26]. Furthermore, the incidence of first stroke in patients on hydroxyurea has been reported as 0.36 – 0.52 per 100 patient years [24] [25] [26], lower than in non transfused patients. However, data from long term follow up studies in which children that had already normalized their TCD velocities after chronic transfusion therapies (CTT) were switched onto hydroxyurea in the Stroke with transfusion changing to hydroxyurea (SWiTCH) trial reported increased numbers of strokes occurring in children on hydroxyurea as compared to those receiving CTT [27]. Similarly, Bernaudin et al reported a return to high TCD velocities in up to 40% of children on hydroxyurea on long term follow up [28]. These data indicate the importance of the ongoing randomized controlled trials to determine the efficacy of or otherwise of hydroxyurea for primary stroke prevention in children with SCD.

As our standard of care in the unit, we prescribed hydroxyurea for children with abnormal, conditional and discrepant TCDs for its multiple positive effects on SCD, as well as the potential for decreasing TCD values. We found its uptake was more acceptable than chronic blood transfusions. However, the drug was relatively expensive and difficult to access particularly for the younger children in whom the medication needed to be compounded, as there were no liquid formulations or small dose (<500mg) capsules available then. There is the need to have hydroxyurea made readily available to ensure adherence to therapy. The routine laboratory investigations required for monitoring children on hydroxyurea further contributed to an increase in their health care costs.

In conclusion, this study has shown that among Nigerian children with sickle cell anaemia, there is a high risk for stroke as reflected by the finding of abnormal and conditional TCD velocities in up to 18% of patients studied. Children with discrepant and low TCDs which may indicate a further risk of stroke were also identified. It was not possible from this study to deduce risk factors associated with elevated TCDs that may identify children at risk of stroke, although we noted that the prevalence of abnormal and conditional TCDs was maximal in the age group 3 – 8 years. The use of hydroxyurea was preferable by our patients as opposed to chronic blood transfusions for primary stroke prevention.

It is recommended that primary prevention of strokes by TCDs of the cerebral arteries as a screening tool is done on regular basis in all children with SCD so as to initiate preventive measures in those identified at risk of stroke. However, access to health facilities where TCD evaluations can be done, and the cost of the test may well be a barrier to the implementation of this highly effective stroke prevention strategy, even among children with health insurance in Nigeria, which at present does not pay for such intervention. There should be easy availability of TCD machines at specific health facilities across the country as well as enhancement of training of personnel to ensure the success of this intervention. Acceptance of CTT as a primary prevention strategy for those identified at risk of stroke is also a major challenge. Widespread health education on its advantages would have to be done. It is imperative that the National Blood Transfusion Services of Nigeria be supported by the government to ensure easy access for blood transfusions as required. The available treatment option of hydroxyurea for primary prevention of strokes in SCD would need to be further evaluated in randomized controlled trials.

We note that a point of bias of this study was that parents had to pay for the TCD screening of their children. It constituted a limitation in as much as only those children whose parents could afford the test had it done, as opposed to all children with SCD been screened.

We wish to acknowledge the support of the members of the Paediatric Haematology unit of the National Hospital Abuja Nigeria, for their support in data collection towards the preparation of this manuscript.

There are no conflicts of interest to declare.