|Home | About | Journals | Submit | Contact Us | Français|
Silent infarcts have been reported most commonly in school-aged children with homozygous sickle cell disease (SCD-SS) and are associated with neurocognitive deficits. However, the prevalence of silent infarcts in younger children with SCD-SS is not well defined. In this retrospective study, brain magnetic resonance imaging and angiography (MRI/A) studies performed before six years of age in a cohort of children with SCD-SS were analyzed and the prevalence of abnormalities was calculated. Clinical and laboratory parameters were compared between the groups with and without silent infarcts. Sixty-eight of 96 children in the cohort had brain MRI/A performed prior to age 6 years. Of the 65 who were neurologically asymptomatic, 18 (27.7%, 95% confidence interval 17.3 – 40.2%) had silent infarcts (mean age 3.7 ± 1.1 years, range 1.3 – 5.9 years). Factors associated with silent infarcts included cerebral vessel stensosis by magnetic resonance angiography, lower rates of vaso-occlusive pain and acute chest syndrome, and lower hemoglobin levels. The prevalence of silent infarcts in young children with SCD-SS is similar to that of older children and anemia and severe vasculopathy may be risk factors.
Cerebrovascular complications are common in children with homozygous sickle cell disease (SCD-SS). Overt stroke, cerebral infarction accompanied by neurological symptoms, is estimated to occur in 11% of children with SCD-SS (Ohene-Frempong et al, 1998) and may result in permanent neurological and cognitive impairment. In children, overt stroke is usually related to stenosis and occlusion of large cerebral arteries of the Circle of Willis that can be detected by transcranial Doppler (TCD) ultrasonography and magnetic resonance or conventional angiography. Silent infarcts, cerebral ischemia demonstrated by magnetic resonance imaging (MRI) but without associated focal neurological symptoms, occur more commonly than overt strokes, with a prevalence of 21.8% in children with SCD-SS aged 6 to 19 years followed in the Cooperative Study of Sickle Cell Disease (CSSCD) (Pegelow et al, 2002). The aetiology of these infarcts is less clear. Despite the terminology, it is becoming increasingly apparent that silent infarcts are clinically significant given their association with subsequent overt stroke (Miller et al, 2001) and neurocognitive deficits in school-aged children (Armstrong et al, 1996; Bernaudin et al, 2000).
Most studies of silent infarcts in SCD-SS have assessed children six years and older because younger children often need sedation to complete the MRI scan. Longitudinal follow-up of children with SCD-SS ages 6 years and older performed through the CSSCD showed that the prevalence of silent infarcts is relatively stable in girls from age 6 years and in boys from age 10 years, suggesting that this brain injury begins early in life (Pegelow et al, 2002). In support of this observation, 3 of 23 (13%) infants and toddlers with SCD-SS who underwent brain MRI at a mean age of 13.7 months in the BABY HUG study had silent infarcts (Wang et al, 2008).
Similarly, a study of 36 children less than 4 years old showed a prevalence of silent infarcts of 11% (Wang et al, 1998). In that study, neuropsychological testing revealed developmental delay in 1 of 4 patients with an abnormal MRI scan, suggesting that silent infarcts may have clinical implications for young children similar to those in to older children.
A high prevalence of silent infarcts and cerebrovascular disease in very young children with SCD would suggest a need for earlier screening and implementation of interventions that might prevent neurocognitive impairment and overt stroke. We conducted a retrospective study to assess the prevalence of silent infarcts and cerebrovascular disease in a cohort of young children with SCD-SS. A secondary aim was to assess potential clinical factors associated with silent infarcts. These factors included markers of hemolysis and predictors of severe disease as defined by Miller, et al (2000).
The Institutional Review Board at The Children’s Hospital of Philadelphia approved the study. Requirement for informed consent for the retrospective study was waived.
We retrospectively identified all children actively followed at our Sickle Cell Center with a diagnosis of homozygous SCD (SCD-SS) born between January 1997 and December 2002. A subset of these children also participated in a prospective study to assess neuroradiographic findings and neuropsychological correlations (to be reported separately).
We identified all brain magnetic resonance imaging and angiography (MRI/A) performed on these patients up to December 2006. Beginning in 2000, our institutional neuroimaging screening protocol for children with SCD-SS included a baseline brain MRI/A around 2 years of age, along with TCD screening. Follow-up imaging by 5 to 6 years was also recommended. Additional interval follow-up studies were sometimes ordered based on individual hematologist preference. More frequent follow-up was routinely recommended for those with prior silent infarcts, conditional or abnormal TCD, or other risk factors for stroke.
Brain MRI/A scanning up to November, 2004 was performed on a 1.5 T Siemens Magnetom Vision (Siemens Medical Systems, Iselin, NJ) using our standard SCD imaging protocol (Zimmerman, 2005). Most subsequent MRI/A studies were performed on a 3 T Siemens Trio. The magnetic resonance angiography (MRA) parameters for time-of-flight were an echo time (TE) of 5 ms or less, a TR (repetition time) of 35 to 40 ms and a slice thickness of 1mm or less. At least two neuroradiologists reviewed each study. All images were read and a clinical report was dictated by one of several neuroradiologists. All images were reviewed by the study neuroradiologist who was blinded to the clinical report and to the clinical histories other than diagnosis of SCD-SS. The study neuroradiologist’s findings were then compared to the clinical report to assess for agreement. A third neuroradiologist read all studies for which the study neuroradiologist had read the initial clinical report. The third neuroradiologist also reviewed all studies where there was a discrepancy between the initial clinical report and the study neuroradiologist’s review, and consensus was reached.
Cerebral infarction was defined as an area of abnormally increased signal on T2-weighted and fluid-attenuated inversion recovery (FLAIR) imaging. The number, size, and location of lesions were recorded. Lesions were classified as punctate (< 5 mm), medium (5–15 mm), or large (>15 mm). Silent infarction was defined as the presence of lesions without a clinical history of stroke or motor deficits that could be attributed to the lesion on MRI. Stenosis was defined as an area of narrowing or focal signal dropout in an artery of the Circle of Willis.
We reviewed the results of all TCD studies performed on the cohort of patients up to December 2006. Screening TCD studies were obtained when the children were clinically well. A 2-MHz pulsed Doppler ultrasound system (Nicolet EME TC 2000 or Nicolet Pioneer TC 8080; Nicolet, Madison, WI) was utilized for all studies. Using right and left temporal approaches, the highest time-averaged mean velocity was recorded in the middle cerebral artery (MCA), the distal internal carotid artery (ICA) and its bifurcation, the anterior cerebral artery (ACA), and the posterior cerebral artery (PCA). The basilar artery velocity was also recorded using the suboccipital approach. Our institutional TCD screening protocol recommends annual TCD examinations beginning at age 2 years, with more frequent follow-up studies for those with conditional or abnormal results or other stroke risk factors. Thus, some children had more than one TCD examination during the study period. For each study, we recorded the highest mean velocity in the ICA, its bifurcation, or MCA on either side and using this value we classified studies as normal (less than 170 cm/s), conditional (between 170 and 199 cm/s) and abnormal (at least 200 cm/s), according to the Stroke Prevention Trial in Sickle Cell Anemia (STOP) classification schema (Adams et al, 1998). Results from the TCD examination closest to the MRI/A study and the highest velocity on any study before age six years were utilized in the analyses.
Each child’s sex and age at the time of brain MRI/A and TCD studies were recorded. To assess for potential early laboratory predictors (Miller et al, 2000) of silent infarction, hemoglobin level and white blood cell count obtained at routine clinic visits between the ages of 11 and 25 months were recorded, and the average value was calculated. Complete blood count and fetal hemoglobin levels obtained at a routine clinic visit nearest to the MRI/A study also were analyzed. Lactic dehydrogenase (LDH) values were not routinely obtained during the study period. As other markers of hemolysis, the aspartate aminotransferase (AST) and indirect bilirubin level closest to the MRI/A study were analyzed, given that these values have been shown to correlate with LDH in patients with SCD (Kato et al, 2006). Children with alanine aminotransferase (ALT) elevations were excluded from these analyses to exclude those with AST elevations that may have been related to liver disease rather than hemolysis. Laboratory values obtained while receiving chronic transfusions or hydroxycarbamide were excluded. If children received intermittent transfusions, laboratory studies at least 120 days after the transfusion were utilized. The age at first episode of dactylitis was documented. All emergency department, Hematology Acute Care Unit, and inpatient hospitalizations were reviewed and the number of visits for vaso-occlusive pain, acute chest syndrome and splenic sequestration or other acute anemic episodes were documented. History of overt stroke, seizure, or other neurological event and treatment with chronic red cell transfusions and hydroxycarbamide and the indications for treatment were recorded. Those who received at least six consecutive months of regular red cell transfusions before six years of age for a complication of SCD, such as recurrent acute splenic sequestration or abnormal TCD, were considered to have received chronic red cell transfusion therapy.
Statistical analysis was performed with STATA 9.0 software (College Station, TX, USA). Within the group with brain MRI/A studies, the prevalence with 95% confidence intervals (95% CI) was calculated separately for silent infarction and for cerebral vessel stenosis. Some children had more than one brain MRI/A performed before their sixth birthday. We first calculated the prevalence of cerebrovascular abnormalities at initial study. Then we calculated a prevalence estimate using the results of any abnormal study, if obtained before the age 6 of years.
The mean event rates were calculated for hospitalizations, emergency department or acute care unit visits for vaso-occlusive pain (including dactylitis) and acute chest syndrome in the first six years of life. Only episodes of pain or acute chest syndrome that occurred while not receiving chronic transfusions or hydroxycarbamide were included in the event-rate calculation because such therapies are expected to alter the frequency of these events.
Descriptive analyses, including means, ranges, and standard deviations were used to describe patient characteristics both overall and for the groups with brain MRI/A and without brain MRI/A studies. In exploratory analyses, continuous variables, such as age, TCD velocity, event rates, hemoglobin level, and AST were compared between groups (MRI and no MRI and silent infarct and no silent infarct) using t-tests. The independence of categorical variables such as sex, history of dactylitis before age one, history of acute anemic event (splenic sequestration or aplastic episode) was assessed using chi-square or Fisher exact tests.
The retrospective cohort consisted of 97 children. One child was excluded because his parent withdrew consent in the prospective study and did not wish his information to be included. Thus, 96 children were included in the study. Seventy-seven (80.2%) were referred to our centre as infants based on newborn (75) or prenatal (2) screening results, 15 (15.6%) were diagnosed by newborn screen but initially followed elsewhere and transferred to our centre at a mean age of 2.5 years (range 0.34 to 5.92 years), usually due to family relocation, and 4 (4.1%) were referred to our centre at a mean age of 1.9 years (range 0.98 to 2.8 years) after presenting with clinical symptoms.
Sixty-eight of the 96 (71%) children had brain MRI/A performed before age 6 years. The mean age at first brain MRI study was 3.6 ± 1.1 years (range 1.2 to 5.9 years). Two children had overt stroke at initial study at ages 1.2 and 5.9 years old and another child had viral-related global ischemic encephalopathy at age 4.9 years. These children were excluded from analyses of the prevalence of silent infarction because they were symptomatic.
Patient characteristics of the entire cohort as well as those with and without brain MRI/A studies are shown in Table I. Children with brain MRI/A were younger at the end of the study, had higher TCD velocities, and were more likely to have received chronic transfusion therapy than those who did not have MRI/A studies before age 6 years.
At initial brain MRI, 13 of 65 children had silent infarcts (20%, 95% CI 11.1 – 31.8%). Thirty-four of the 65 (52%) children had at least one follow-up brain MRI prior to the age of 6 years old. Of 28 with initial normal studies, five children initially imaged at a mean age of 3.0 years (range, 2.4 to 3.6 years), developed silent infarcts by follow-up brain MRI at a mean age of 4.8 years (range 3.5 to 5.7 years), while 23 children had repeat normal studies. Thus, the prevalence of silent infarct at the worst study obtained prior to age 6 years was 27.7% (95% CI 17.3 – 40.2%). Seven of 18 children with silent infarcts at either initial (n=6) or follow-up study (n=1), had subsequent imaging studies before age 6 years. Five children with silent infarcts had persistent stable abnormalities (two receiving transfusions), one had progression of silent infarcts (not receiving transfusions), and one developed overt stroke (prior to transfusions). The locations and sizes of silent infarcts in this cohort are shown in Table II.
At initial study, 11 of 64 (17.2%, 95% CI 8.9 – 28.7%) children aged less than 6 years had cerebral vessel stenosis detected by brain MRA (one child did not undergo MRA). Three children with abnormal MRA later had resolution of the abnormality; none of these children received transfusion therapy. One child with an initial normal MRA had stenosis detected on follow-up MRA prior to age 6 years old. Stenosis detected by MRA was significantly associated with higher TCD velocities (189 cm/s v. 158 cm/s, p = 0.004).
Lower rates of both acute chest syndrome and vaso-occlusive pain episodes were significantly associated with silent infarcts in young children with SCD-SS (Table III). Children with silent infarcts also had significantly lower hemoglobin levels than those without silent infarcts. Stenosis detected by MRA was associated with silent infarcts, but maximum TCD velocities were not statistically higher (173 cm/s v. 160 cm/s, p=0.17) in children with silent infarcts. Similarly, TCD velocities on the right, left, or the side of the silent infarct were not significantly different between the groups with and without silent infarcts (data not shown). Only three children in our cohort had a history of seizure; one had silent infarcts and two had normal brain MRI studies. Early dactylitis and hemoglobin level and white blood cell count in the second year of life were not associated with silent infarcts.
We report the brain MRI/A findings obtained before age 6 years of a large cohort of children with SCD-SS. The period prevalence of silent infarction of 27.7% was high and is similar to the prevalence of silent infarction reported in older children with SCD-SS (Pegelow et al, 2002; Steen et al, 2003). The distribution of silent infarcts in our cohort, mostly in the frontoparietal deep white matter and periventricular regions, is also similar to that reported in older children (Pegelow et al, 2002; Pegelow et al, 2001). Taken together, these data support the conclusion that silent infarcts begin to develop very early in life in SCD-SS, and support a role for early screening.
A potential selection bias exists in our study, given that 30% of the cohort did not have brain MRI studies prior to age 6 years. In part, some children were not screened due to slow implementation of our neuroimaging screening protocol, reflected by the older age of those who were not screened. More importantly, clinical severity or concern for silent infarction may have influenced the decision to obtain initial or follow-up brain MRI/A studies. In particular, children who were receiving chronic transfusions and those with higher TCD velocities were more likely to have undergone brain MRI/A. However, even if all of the children who did not have brain MRI performed before age 6 years had negative studies at that age, the prevalence of silent infarct in our cohort before age 6 years would be 19.4% (95% CI 11.9 – 28.9%), supporting the early development of cerebral ischemia.
In exploratory analyses, we identified several factors associated with silent infarcts in our cohort of young children. Young children with silent infarcts had lower painful event rates than those without silent infarcts, similar to the findings reported in older children (Kinney et al, 1999). In addition, lower rates of acute chest syndrome also were associated with silent infarcts in the present study. This suggests that the underlying pathophysiology of silent infarcts may differ from vaso-occlusive complications. The lower hemoglobin levels and the trend towards higher AST levels in children with silent infarcts in our study support a potential role of anemia and/or hemolysis in the pathophysiology of silent infarcts.
Interestingly, we also found an association of silent infarcts with cerebral vessel stenosis detected by MRA. TCD velocities also were significantly higher in children with abnormal MRA than in those with normal MRA. However, we did not find a significant association of TCD velocity with silent infarcts, similar to prior studies (Wang et al, 2000). Our results suggest that children with more severe vasculopathy, evidenced by both elevated TCD velocity and abnormal MRA, may be at higher risk of silent infarct. It is possible that those with large vessel vasculopathy also have narrowing of the microvasculature or that embolization from a larger vessel results in occlusion of smaller vessels.
In older children, silent infarcts have been shown to progress in both number and size over time (Pegelow et al, 2002) and the associated neurocognitive abnormalities also may be progressive. One of seven (14%) children in the current study with silent infarct who had follow-up studies before their sixth birthday, showed progression and a second child developed overt stroke. This supports a need for early interventions to halt progression. However, the optimal therapy for silent infarcts is currently unknown. A multicentre trial of transfusion therapy in older children with silent infarcts is currently underway and evaluation of other therapies, such as hydroxycarbamide, may be appropriate.
Thus, silent cerebral infarcts and cerebrovascular disease are common in pre-school aged children with SCD-SS, confirming that silent infarcts begin early in life. Our preliminary results suggest that anemia and vasculopathy may play a role in the pathogenesis of silent infarcts. Given the small sample size in our study, larger clinical trials to assess these potential risk factors as well as potential treatments of silent infarcts in very young children with SCD-SS are needed. Early identification of children with silent infarcts may allow for timely interventions that could prevent progression of cerebral damage and its clinical sequelae.
This work was supported by NIH Grant 2 P60HL38632.