Prevalence of cerebral sinovenous thrombosis in abusive head trauma.

BACKGROUND: Cerebral sinovenous thrombosis (CSVT) has been proposed in legal settings to be an atraumatic mimic of abusive head trauma (AHT). OBJECTIVE: The objective of this study was to determine the prevalence of CSVT and subdural hemorrhage (SDH) in a large AHT population. MATERIALS AND METHODS: This retrospective cohort study measured the prevalence of CSVT and SDH on magnetic resonance venograms in 243 patients diagnosed with AHT at a single center. We also reported additional intra- and extracranial injuries, head injury severity and length of hospital stay. RESULTS: Among 243 patients diagnosed with AHT, 7% (16/243) had CSVT. SDH was present in 94% (15/16) of the CSVT cases. Cytotoxic edema and subarachnoid hemorrhage were in 88% (14/16) and 69% (11/16) of the CSVT cases, respectively. Extracranial signs of abuse were also in 100% (16/16) of the patients with CSVT. Critical to maximal head injury severity (abbreviated injury scale >=5) was in 75% (12/16) of the CSVT population vs. 33% (82/243) in the total AHT population. Length of hospital and pediatric intensive care unit stay was greater in those with CSVT (10 vs. 21.9 and 3.5 vs. 7.3 days). CONCLUSION: These findings suggest that CSVT is uncommon in AHT and is associated with additional traumatic injuries and greater injury severity.

ICD-10-CM Codes for the Identification of Abusive Head Trauma in Administrative Datasets.

OBJECTIVE: We sought to develop and validate a list of ICD-10-CM codes identifying abusive head trauma (AHT). METHODS: Subjects included all children under 2 years with head trauma seen in the emergency department or admitted to one of 5 medical centers. Cases were classified as AHT, accidental head injury, or indeterminate based on chart review of the medical record. ICD-10-CM code list to identify cases of AHT was developed based on prior head injury code lists. Sensitivity and specificity of the final code list were calculated. RESULTS: There were 2883 patients in the study population of whom 524 had AHT, 2123 had accidental injury, and 236 were indeterminate cases. The final list of AHT codes had a sensitivity of 76.1% (95% CI 72.5-79.8) and a specificity of 98.5% (95% CI 98.0-99.0) when limiting analyses to the groups with identified cause of injury (accidental vs abusive). Misclassification of cases based on codes resulting in false positives and false negatives was due to coding errors. CONCLUSIONS: The list of ICD-10-CM codes can be utilized to identify and track cases of AHT at a national level in large administrative datasets though likely underestimates true injury burden.