Predictors of Acute Intracranial Pathology Identified by Computerized Tomography in Children with Sickle Cell Disease.

Headache and other symptoms and signs of possible CNS disease are common in children with sickle cell disease (SCD). Computerized tomography (CT) is a rapid and widely available method to image the skull and brain. However, the use of head CT to evaluate possible acute CNS complications in SCD has not been adequately addressed. We conducted a retrospective cohort study to characterize predictors of acute intracranial pathology detected by head CT. We included patients <22 years of age with SCD who had a head CT at Johns Hopkins Hospital from January 1994 to March 2006. We identified cases using the hospital’s discharge database (ICD-9 codes for SCD and CPT codes for head CT). We defined acute intracranial pathology as any intracranial hemorrhage, new hydrocephalus, cerebral edema, tumor, abscess, or ischemic stroke dated within the previous 7 days by clinical findings and confirmed by head CT. We recorded baseline characteristics, symptoms and signs before head CT, results of neuroimaging, treatments given, and duration of hospitalization. We compared dichotomous variables and calculated odds ratios by exact methods and by univariate and multivariate logistic regression with adjustment for clustering and robust estimates of errors. We calculated the sensitivity and specificity of predictors of acute intracranial pathology. We identified 89 patients (mean age 13.3 years, range 2 weeks to 21.9 years) who had 133 head CTs. Indications for neuroimaging included headache (63), altered mental status (20), suspected stroke (19), seizure (9), trauma (7), suspected infection (5), syncope (3), ataxia (3), and other (4). Fifteen CTs showed evidence of acute intracranial pathology, including 8 hemorrhages (5 intraparenchymal, 2 subdural, and 1 subarachnoid), 5 acute ischemic strokes, and 2 brain contusions. Associated clinical characteristics included antecedent trauma (OR 12, 95% CI 2.6–55, P<0.001), loss of consciousness (OR 2.3, 95% CI 1.3–24, P=0.02, amnesia (OR 7.3, 95% CI 1.1–49, P<0.05), and/or neurological deficits on examination (OR 10, 95% CI 3.0–34, p<0.001). An abnormal Glasgow Coma Scale (GCS) was weakly associated with intracranial pathology (OR 1.2 per 1 point decrease, 95% CI 1.0–1.4, p=0.06). A history of trauma or a focal neurological deficit on examination identified all patients with intracranial pathology by CT with reasonable specificity (Table 1). Forty-two head CTs were followed by MRI +/− MRA of the brain within 14 days. Of these, 13 revealed new intracranial pathology (5 acute ischemic strokes, 3 silent cerebral infarcts, and 5 stenoses of cerebral arteries) and 3 mild sinusitis not detected by CT. Two MRIs failed to confirm possible ischemic lesions noted on head CT. Use of head CT could be reduced by 34% by limiting their use to those with trauma, a focal deficit, or GCS <15 and even further by rapid access to MRI. Our findings require prospective validation, but provide information to guide neuroimaging in children and young adults with SCD. Eliminating unnecessary head CTs could reduce both cost and exposure to radiation.