Validation of the Scandinavian Guidelines for Management of TBI in Adults

Overview

Traumatic brain injuries (TBI) are one of the most common reasons for patients to attend the emergency department (ED). 90% of patients with TBI are defined as mild TBI (mTBI). A small minority of patients with mTBI would show pathological computed tomography (CT) results and even fewer need neurosurgical intervention. Nevertheless, complications would be so severe, if neurosurgical intervention is delayed, that it has become common practice to subject all patients with mTBI to CT. The high number of CT scans has an impact on health care resources but may also involve risk by subjecting patients through potentially harmful ionizing radiation. Several independent research groups have attempted to optimize CT use in mTBI patients by forming guidelines that aim to identify patients at high risk for intracranial complications. Most guidelines have been published in the past 15 years and have been validated both prospectively internally and externally; all guidelines have been shown to be safe when implemented in clinical use with few missed complications. However the number of CT scans has not been reduced dramatically, in some cases it has even increased. In 2013, the new Scandinavian guidelines (SNC13) were published. They are the first guidelines that use a biomarker, S100B, as a tool for managing patients with mTBI. Although S100B has a low specificity for intracranial complications, a high sensitivity makes it suitable to be implemented into clinical practice as a tool for CT reduction. Previous SNC guidelines have been compared to other prominent guidelines with impressive results. The SNC13 have been externally validated in a retrospective study from the USA that was underpowered for important outcomes. Nevertheless, SNC13 have already been partially implemented in clinical practice in Scandinavia. However, a strict multicenter validation has not been performed yet nor a systematic comparison to other available guidelines. Our primary aim is to validate the performance of the SNC13 in predicting intracranial complications in adult patients presenting with traumatic head injury in Swedish hospitals. A secondary aim is to compare the performance of SNC 13 with 6 other clinical guidelines, with respect to important outcomes. Moreover, to explore the performances of different biomarkers in predicting intracranial complications in predefined subgroups of TBI. Finally, to evaluate the possibility of further improvement of the SNC13.

Design Investigators will perform a prospective, multicenter, pragmatic, observational study of adults presenting with traumatic head injury at the ED. All data necessary for analysis including predictor variables and outcome data for all the seven guidelines included in the study will be registered (table 1). Patients will be managed clinically accordingly to the judgment of the responsible physician and/or local guidelines. Study setting and population The study will be set in Halmstad, Malmö, Lund, Örebro and Linköping, Sweden. Hallands Hospital Halmstad (HS) is a level II trauma centre, Skåne University Hospital in Malmö and Lund (SUS), Örebro University Hospital, Linköping University Hospital are level I trauma centers. The coordinating site for the study will be HS where the statistical and the comparative biomarker analysis will be performed. Inclusion criteria From September 2017 investigators will prospectively enroll all adult patients with a GCS 9-15 that seek the ED within 24h after TBI. Exclusion criteria * patients younger than 18 years of age; * patients without a Swedish personal identification number due to difficulties in performing the follow up phase; * all patients that refuse to participate. Data registration and follow-up Details of how patients are managed, including patient characteristics, injury type, patient history, clinical examination results, current medications and CT findings will be documented in a pre-determined case-report form by the triage nurse and/or physician on call. All patients will be asked to answer a questionnaire sent by mail 3 months after the injury. The questionnaire will be re-sent if no answer is received. If no answer is received from these attempts, patients will be contacted via telephone. The questionnaire includes questions that would identify a significant intracranial lesion, data concerning sick-days, new contacts with medical professionals and information concerning quality of life. In cases where patients can not be reached by mail or telephone, medical records and national mortality databases will be consulted for evidence of complications and/or death. The Swedish health care system allows full visibility of data for persons with a Swedish personal identification number for medical records and mortality database over the whole country. Patients who suffer significant (enough to seek medical care) intracranial complications after discharge would therefore be identified. Details on study period are specified on figure 1 with an algorithm for patient eligibility and data analysis. Data will be registered in an Excel® file. Descriptive statistics will be analysed using IBM SPSS® Statistics Version 20 software. S100B analysis A 5ml blood sample is drawn from patient's cubital vein in the ED. Samples are analysed with the fully automated Elecsys® S100 (Roche AB) at the Clinical Chemistry Department of HS, SUS, Örebro University Hospital and Linköping University Hospital, Sweden. Cut-off level for normal levels of S100B according to the SNC guidelines is less than 0.10μg/L and a window of sampling of within 6 hours from the time of the injury. From all patients seeking care within 24h form injury with medium and low risk TBI, according to SNC13 (including multitrauma patients), a 5ml blood sample will be drawn, centrifuged and frozen at -70 degrees Celsius. Samplings will be coded and registered for analysis of GFAP, SBP-50 and TAU. CT examinations CT scans are always analysed by a board certified radiologist. Sample size The Scandinavian guidelines will recommend discharge (i.e. neither CT nor admission) in approximately 50% of patients and a prevalence of our primary outcome of 5% (from our own observations and from data derived from a pre-selected cohort) . Allowing for one missed case, a sensitivity of \>99% with a lower 95% confidence interval, a sample size of 2490 patients is required to detect traumatic intracranial complications according to the SNC13. Allowing for a 10% lost to follow-up, the desired sample size is 2767 patients. Interim analysis After 1000 patients, the prevalence for the primary outcome will be measured in order to be able to reevaluate sample size. Ethics Ethical approval was granted from the Regional Ethical Review Board of Lund (approval number 2012/574). Informed verbal consent will be obtained and registered by nurses responsible for triage at ED. Patients' data and social security number will be stored and handled accordingly to Swedish Personal Data Act, (PUL 1998: 204). Written consent will be obtained from all patients from whom the extra blood sampling for biomarker analysis will be requested. Sampling will be coded and patients will be able at any time to refuse to be part of the study.