Risk adapted diagnostics and hospitalization following mild traumatic brain injury.

Traumatic brain injury (TBI) remains a leading cause of hospital admission and mortality, intracranial hemorrhage (ICH) presents a severe complication. Low complication tolerance in developed countries and risk uncertainty, often cause excessive observation, diagnostics and hospitalization, considered unnecessary and expensive. Risk factors predicting ICH, progression and death in patients hospitalized with mild TBI have not been identified yet. Mild TBI cases indicated for cranial computer tomography (CT) and hospitalization, according to international guidelines, at our Level I Trauma Center between 2008 and 2018 were retrospectively included. Multivariate logistic regression was performed for ICH, progression and mortality predictors. 1788 mild TBI adults (female: 44.3%; age at trauma: 58.0 ± 22.7), were included. Skull fracture was diagnosed in 13.8%, ICH in 46.9%, ICH progression in 10.6%. In patients < 35 years with mild TBI, chronic alcohol consumption (p = 0.004) and skull fracture (p < 0.001) were significant ICH risk factors, whilst in patients between 35 and 65 years, chronic alcohol consumption (p < 0.001) and skull fracture (p < 0.001) revealed as significant ICH predictors. In patients with mild TBI > 65 years, age (p = 0.009), anticoagulation (p = 0.007) and neurocranial fracture (p < 0.001) were significant, independent risk factors for ICH, whilst increased age (p = 0.01) was a risk factor for mortality following ICH in mild TBI. Late-onset ICH only occurred in mild TBI cases with at least two of these risk factors: age > 65, anticoagulation, neurocranial fracture. Overall hospitalization could have been reduced by 15.8% via newly identified low-risk cases. Age, skull fracture and chronic alcohol abuse require vigilant observation. Repeated CT in initially ICH negative cases should only be considered in newly identified high-risk patients. Non-ICH cases aged < 65 years do not gain safety from observation or hospitalization. Recommendations from our data might, without impact on patient safety, reduce costs by unnecessary hospitalization and diagnostics.