OBJECTIVES: Computed tomography is the criterion standard imaging modality to detect intracranial hemorrhage (ICH) in children and infants after closed head injury, but its use can be limited by patient instability, need for sedation, and risk of ionizing radiation exposure. Cranial ultrasound is used routinely to detect intraventricular hemorrhage in neonates. We sought to determine if point-of-care (POC) cranial ultrasound performed by emergency physicians can detect traumatic ICH in infants. METHODS: Infants with ICH diagnosed by computed tomography were identified. For every infant with an ICH, 2 controls with symptoms and diagnoses unrelated to head trauma were identified. Point-of-care cranial ultrasound was performed by an emergency physician on all patients, and video clips were recorded. Two ultrasound fellowship-trained emergency physicians, blinded to the patients' diagnosis and clinical status, independently reviewed the ultrasound clips and determined the presence or absence of ICH. RESULTS: Twelve patients were included in the study, 4 with ICH and 8 controls. Observer 1 identified ICH with 100% sensitivity (95% confidence interval [CI], 40%-100%) and 100% specificity (95% CI, 60%-100%). Observer 2 identified ICH with 50% sensitivity (95% CI, 9%-98%) and 87.5% specificity (95% CI, 47%-99%). Agreement between observers was 75%, κ = 0.4 (P = 0.079; 95% CI, 0-0.95). CONCLUSIONS: Traumatic ICH can be identified with POC cranial ultrasound by ultrasound fellowship-trained emergency physicians. Although variations between observers and wide confidence intervals preclude drawing meaningful conclusions about sensitivity and specificity from this sample, these results support the need for further investigation into the role of POC cranial ultrasound.
OBJECTIVES: Computed tomography is the criterion standard imaging modality to detect intracranial hemorrhage (ICH) in children and infants after closed head injury, but its use can be limited by patient instability, need for sedation, and risk of ionizing radiation exposure. Cranial ultrasound is used routinely to detect intraventricular hemorrhage in neonates. We sought to determine if point-of-care (POC) cranial ultrasound performed by emergency physicians can detect traumatic ICH in infants. METHODS:Infants with ICH diagnosed by computed tomography were identified. For every infant with an ICH, 2 controls with symptoms and diagnoses unrelated to head trauma were identified. Point-of-care cranial ultrasound was performed by an emergency physician on all patients, and video clips were recorded. Two ultrasound fellowship-trained emergency physicians, blinded to the patients' diagnosis and clinical status, independently reviewed the ultrasound clips and determined the presence or absence of ICH. RESULTS: Twelve patients were included in the study, 4 with ICH and 8 controls. Observer 1 identified ICH with 100% sensitivity (95% confidence interval [CI], 40%-100%) and 100% specificity (95% CI, 60%-100%). Observer 2 identified ICH with 50% sensitivity (95% CI, 9%-98%) and 87.5% specificity (95% CI, 47%-99%). Agreement between observers was 75%, κ = 0.4 (P = 0.079; 95% CI, 0-0.95). CONCLUSIONS:Traumatic ICH can be identified with POC cranial ultrasound by ultrasound fellowship-trained emergency physicians. Although variations between observers and wide confidence intervals preclude drawing meaningful conclusions about sensitivity and specificity from this sample, these results support the need for further investigation into the role of POC cranial ultrasound.