Early Neurosurgical Procedures Enhance Survival in Blunt Head Injury: Propensity Score Analysis

Machine learning-based prediction of emergency neurosurgery within 24 h after moderate to severe traumatic brain injury

Background

Rapid referral of traumatic brain injury (TBI) patients requiring emergency neurosurgery to a specialized trauma center can significantly reduce morbidity and mortality. Currently, no model has been reported to predict the need for acute neurosurgery in severe to moderate TBI patients. This study aims to evaluate the performance of Machine Learning-based models to establish to predict the need for neurosurgery procedure within 24 h after moderate to severe TBI.

Methods

Retrospective multicenter cohort study using data from a national trauma registry (Traumabase®) from November 2011 to December 2020. Inclusion criteria correspond to patients over 18 years old with moderate or severe TBI (Glasgow coma score ≤ 12) during prehospital assessment. Patients who died within the first 24 h after hospital admission and secondary transfers were excluded. The population was divided into a train set (80% of patients) and a test set (20% of patients). Several approaches were used to define the best prognostic model (linear nearest neighbor or ensemble model). The Shapley Value was used to identify the most relevant pre-hospital variables for prediction.

Results

2159 patients were included in the study. 914 patients (42%) required neurosurgical intervention within 24 h. The population was predominantly male (77%), young (median age 35 years [IQR 24–52]) with severe head injury (median GCS 6 [3–9]). Based on the evaluation of the predictive model on the test set, the logistic regression model had an AUC of 0.76. The best predictive model was obtained with the CatBoost technique (AUC 0.81). According to the Shapley values method, the most predictive variables in the CatBoost were a low initial Glasgow coma score, the regression of pupillary abnormality after osmotherapy, a high blood pressure and a low heart rate.

Conclusion

Machine learning-based models could predict the need for emergency neurosurgery within 24 h after moderate and severe head injury. Potential clinical benefits of such models as a decision-making tool deserve further assessment. The performance in real-life setting and the impact on clinical decision-making of the model requires workflow integration and prospective assessment.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13017-022-00449-5.

Comparative Effectiveness of Surgery for Traumatic Acute Subdural Hematoma in an Aging Population

There is uncertainty as to the optimal initial management of patients with traumatic acute subdural hematoma, leading to regional variation in surgical policy. This can be exploited to compare the effect of various management strategies and determine best practices. This article reports such a comparative effectiveness analysis of a retrospective observational cohort of traumatic acute subdural hematoma patients in two geographically distinct neurosurgical departments chosen for their - a-priori defined - diverging treatment preferences. Region A favored a strategy focused on surgical hematoma evacuation, whereas region B employed a more conservative approach, performing primary surgery less often. Region was used as a proxy for preferred treatment strategy to compare outcomes between groups, adjusted for potential confounders using multivariable logistic regression with imputation of missing data. In total, 190 patients were included: 108 from region A and 82 from region B. There were 104 males (54.7%). Matching current epidemiological developments, the median age was relatively high at 68 years (interquartile range [IQR], 54-76). Baseline characteristics were comparable between regions. Primary evacuation was performed in 84% of patients in region A and in 65% of patients in region B (p < 0.01). Mortality was lower in region A (37% vs. 45%, p = 0.29), as was unfavorable outcome (53% vs. 62%, p = 0.23). The strategy favoring surgical evacuation was associated with significantly lower odds of mortality (odds ratio [OR]: 0.43; 95% confidence interval [CI]: 0.21-0.88) and unfavorable outcome (OR: 0.53; 95% CI: 0.27-1.02) 3-9 months post-injury. Therefore, in the aging population of patients with acute subdural hematoma, a treatment strategy favoring emergency hematoma evacuation might be associated with lower odds of mortality and unfavorable outcome.

Prehospital Imaging-Based Triage of Head Trauma with a Mobile Stroke Unit: First Evidence and Literature Review

BACKGROUND

An ambulance equipped with a computed tomography (CT) scanner, point-of-care laboratory, and telemedicine capabilities (Mobile Stroke Unit [MSU]) has been shown to enable delivery of thrombolysis to stroke patients at the emergency site, thereby significantly decreasing time to treatment. However, the MSU frequently assesses patients with cerebral disorders other than stroke. For some of these disorders, prehospital CT scanning may also be beneficial.

METHODS

Our institution manages a program investigating prehospital stroke treatment of patients with neurological emergencies. We assessed a patient with head trauma for whom prehospital CT scanning and laboratory tests allowed cause-based triage to the most appropriate hospital. We examined implications of this case for clinical practice in light of a literature review.

RESULTS

The MSU was dispatched to assess a 74-year-old woman with suspected head trauma or stroke, found lying on the floor with a left frontal laceration. Her Glasgow Coma Scale score was 13, apart from drowsiness she exhibited no neurologic deficit. A CT scan ruled out intracranial hemorrhage and skull fracture. On the basis of these prehospital diagnostic findings, the patient was taken to the nearest primary care hospital rather than to a trauma center with neurosurgery facilities.

CONCLUSION

Patients with neurologic disorders other than stroke, such as traumatic brain injury, may also benefit from prehospital CT studies. This case report and the results of our analysis of the literature support the potential benefit of prehospital imaging in correctly triaging patients with suspected traumatic brain injury to the appropriate target hospital.

The accuracy of alternative triage rules for identification of significant traumatic brain injury: a diagnostic cohort study

INTRODUCTION

Traumatic brain injury (TBI) is a leading cause of death and disability in young adults. Reorganisation of trauma services with direct triage of suspected head injury patients to trauma centres may improve outcomes following TBI. This study aimed to determine the sensitivity of principal English triage tools for identifying significant TBI.

METHODS

We performed a diagnostic cohort study using data prospectively collated from the Trauma Audit and Research Network database between 2005 and 2011. Adult head injury patients were retrospectively classified according to London Ambulance Service (LAS) and Head Injury Transportation Straight to Neurosurgery study (HITS-NS) triage criteria. Sensitivity and specificity were then calculated against a reference standard of significant TBI, comprising head region abbreviated injury score (AIS) ≥3 or neurosurgical operation.

RESULTS

6559 patients were included in complete case analyses. The LAS and HITS-NS triage tools demonstrated sensitivities of 44.5% (95% CI 43.2 to 45.9) and 32.6% (95% CI 31.4 to 33.9), respectively, for identifying significant TBI patients. False negative significant TBI cases were relatively older, more likely to be female, more frequently secondary to low-level falls, and were less likely to have very severe AIS five or six head injuries, p<0.01.

CONCLUSIONS

A considerable proportion of significant head injury patients may not be triaged directly to trauma centres. Investment is therefore necessary to improve the accuracy of existing triage rules and maintain expertise in TBI diagnosis and management in non-specialist emergency departments.

Redefining the golden hour for severe head injury in an urban setting: the effect of prehospital arrival times on patient outcomes

BACKGROUND

In patients with severe head injuries, transportation to a trauma centre within the "golden hour" are important markers of trauma system effectiveness but evidence regarding impacts on patient outcomes is limited.

OBJECTIVE

To determine the effect of patient arrival within the golden hour on patient outcomes.

METHODS

A retrospective cohort of adult patients with severe head injuries (head AIS ≥ 3) arriving within 24h of injury was identified using the trauma registry from 2000 to 2011. Survival analysis was used to determine the effect of patient arrival time on overall mortality. Study outcomes were in hospital mortality and survival to hospital discharge without requiring transfer for ongoing rehabilitation or nursing home care.

RESULTS

There was a significant association with mortality with each incremental minute of patient arrival (HR 1.002, 95%CI 1.001-1.004, p=0.001). There was however no survival benefit observed for patients arriving within 60 min of injury time (HR 0.77, 95%CI 0.50-1.18, p=0.22) but an apparent benefit for those presenting within 2h of injury time (HR 0.31, 95%CI 0.15-0.66, p=0.002). Patient arrival within 60 min of injury time was associated with increased odds of survival to hospital discharge without requiring ongoing rehabilitation (OR 1.78, 95%CI 1.14-2.79, p=0.01).

CONCLUSION

A survival benefit exists in patients arriving earlier to hospital after severe head injury but the benefit may extend beyond the golden hour. There was evidence of improved functional outcomes in patients arriving within 60 min of injury time.

Should Uncooperative Trauma Patients with Suspected Head Injury be Intubated?

In trauma patients with a suspicion for traumatic brain injury (TBI), a head computed tomography (CT) scan is imperative. However, uncooperative patients often cannot undergo imaging without sedation and may need to be intubated. Our hypothesis was that among mildly injured trauma patients, in whom there is a suspicion of a head injury, uncooperative patients have higher rates of TBI and intubation should be considered to obtain a CT scan. We found that uncooperative patients intubated for diagnostic purposes were more likely to have moderate to severe TBI than nonintubated patients (21.4 vs 8.4%, P < 0.0001) and uncooperative behavior leading to intubation was an independent predictor of TBI (odds ratio, 2.5; 95% confidence interval, 1.5 to 4.5). Of patients with brain injury, intubated patients more often had a head Abbreviated Injury Scale score of 4 (20.8 vs 7.9%, P = 0.04). Uncooperative intubated patients had longer hospital stays (3.6 vs 2.6 days, P = 0.003) and higher mortality (0.9 vs 0.2%, P = 0.02) than nonintubated patients. Uncooperative behavior may be an early warning sign of TBI and the trauma surgeon should consider intubating uncooperative trauma patients if there is suspicion for brain injury based on the mechanism of their trauma.

Impact of operative intervention delay on pediatric trauma outcomes

BACKGROUND

Establishing quality indicators is an essential step in improving mortality and disability among pediatric patients with trauma. We hypothesized that timing of craniotomy, intracranial pressure (ICP) monitoring for traumatic brain injury, and abdominal operation for solid organ injury correlates with a reduced risk of death, shorter stay, and reduced risk of requiring assistance at discharge.

METHODS

This was a retrospective cohort study of 99,513 pediatric patients with trauma, using the National Trauma Data Bank.

RESULTS

For patients who had an ICP monitor placed within 4 hours compared with those whose ICP monitor was delayed, there was no difference in mortality; however, there was a shorter stay in the hospital (relative risk [RR], 0.84; 95% confidence interval (CI), 0.72-0.97) and in the intensive care unit (ICU) (RR, 0.76; 95% CI, 0.66-0.86) in those that survived to discharge. Patients who had craniotomy within 4 hours had higher mortality (RR, 1.98; 95% CI, 1.11-3.51) compared with those that were delayed. After excluding those that died, there was a shorter overall stay (RR, 0.69; 95% CI, 0.59-0.81) and ICU stay (RR, 0.69; 95% CI, 0.57-0.83). Similar length of stay results were seen in pediatric patients with solid organ injuries. Excluding those that died, length of stay (RR, 0.58; 95% CI, 0.47-0.73) and ICU stay (RR, 0.52; 95% CI, 0.37-0.74) were shorter.

CONCLUSION

Early intervention in those who survive their initial operation is associated with shorter ICU and hospital stay for traumatic brain and solid organ injuries.

LEVEL OF EVIDENCE

Therapeutic study, level III.