| Abstract |
Background
Head injury in children is a common presentation in acute care settings, with a significant incidence ranging from 47-280 / 100,000 and mortality rates that can reach 7%. The effects of craniocerebral trauma on children's developing brains are the subject of international studies, with emergency brain CT beign the test of choice for diagnosing brain pathology. CT scans of the brain are mandatory in patients with moderate to severe TBI, according to TBI management protocols that are applied worldwide. However, the performance of CT in children with a mild injury is debatable.
Clinical prediction models were designed to identify children with a seemingly mild injury, who require a CT scan to identify a clinically occult but significant intracranial damage and simultaneously prevent the performance of unnecessary CT scans and therefore unnecessary ionizing radiation in children with mild TBI.
At the same time, imaging severity scores have been studied as independent predictive indicators of TBI in adults. There is limited literature on the usefulness of imaging severity scores in the pediatric population with few recent publications that evaluate severe TBI in children.
Objective
The evaluation of the clinical prediction models PECARN, CHALICE and CATCH for the necessity of performing emergency CT of the brain in children and adolescents who were treated at the University Hospital of Heraklion for TBI during the years 2014-2018.
The comparative evaluation of the imaging severity scores according to Marshall, Rot-terdam, Helsinki and their correlation with clinical and laboratory indicators during the admission of patients with TBI as well as with the outcome.
Methods
The retrospective study was conducted on 232 children and adolescences with TBI who had admitted during the years 2014-2018, in the University Hospital of Heraklion, Crete. Numerous parameters were collected and included demographic and clinical characteristics of patients such as age, sex, mechanism of injury, GCS, the need for surgery, and brain injury outcome. The diagnostic accuracy of prediction models PECARN, CHALICE and CATCH was assessed. Moreover, each patient’s CT scan was evaluated and scored using the Marshall, Rotterdam and Helsinki scoring systems. Statistical correlation with significant clinical variables was performed.
Results
211 patients were studied, with an average age of 6.3 years, 66.3% were boys, and the main cause of injury was falls (74%) while 14.2% of children were treated in the ICU. Mortality was 1%, and 1 in 3 ICU patients underwent surgery. 187 (91%) patients had mild, 9 (4.3%) moderate and 15 (7, 1%) severe TBI.
141 patients were evaluated by PECARN, 107 by CATCH and 208 by CHALICE. PECARN evaluated 141 patients with CGS> 13, recommended CT for 28 which eventually was performed in 36 children. CATCH evaluated 107 children with CGS> 12, suggested CT for 41 which eventually was performed in 59. CHALICE, on the other hand, rated 208 children at all Glasgow Coma Scales, proposed 126 CTs, and 93 CTs were performed. 15% more CTs were performed than those proposed by PECARN, 17% more CTs than those proposed by CATCH and 15% less than those proposed by CHALICE. PECARN was more sensitive in detecting patients with clinically imortant TBI (ciTBI) (100%) and patients who needed surgery, followed by CATCH and CHALICE.
The severity of TBI was related to all 3 imaging scoring systems, with the Helsinki pre-dominating in terms of predicting the outcome of patients with ciTBI (p-value 0.001). GCS scoring was related to the scores by imaging (p <0.001).
Conclusion
All patients with severe TBI and / or patients hospitalized in the ICU underwent a brain CT scan according to current guidelines for severe head injuries.
A higher number of CTs were performed in children with mild TBI compared to the rec-ommended CTs by the clinical prediction models while CT scans were performed in all patients with ciTBI. The PECARN model was more sensitive for ciTBI compared to CALICE and CATCH models.
Imaging severity scores Marshall, Rotterdam and Helsinki correlate with the severity of TBI and together with the GCS constitute predictive indicators of the outcome in children with TBI.
|
Προγνωστικά μοντέλα έκβασης σε σοβαρή κρανιοεγκεφαλική κάκωση σε παιδιά και εφήβους
