International Mission For Prognosis And Analysis Of Clinical Trials Score For Predicting Outcome In Moderate To Severe Traumatic Brain Injury

Introduction: Traumatic brain injury disease of major importance globally. Prognostic models are useful for making decisions in the clinical practice. The aim of this study was to assess the accuracy of International Mission for Prognosis and Analysis of Clinical Trials in TBI (IMPACT) score in predicting outcome in moderate to severe TBI at 3 months. Materials and Methods: All patients admitted to National Trauma Center, National Academy of Medical Sciences with moderate to severe traumatic brain injury from February 2020 to February 2021 were included in the study. IMPACT scores (core/extended core/ lab) were recorded separately at admission. Outcome was measured with Glasgow Outcome Scale (GOS) at the time of discharge and at six months. Correlation between observed and predicted outcomes was evaluated by Pearson’s correlation coefficient (r). Sensitivity and specificity were plotted in the receiver-operating characteristic (ROC) curve, and the area under the curve (AUC) was calculated to determine the discrimination ability of this prognostic model. Results: A total of 112 patients were enrolled in the study. Eighty (71.4 %) patients had moderate and 32 (28.57 %) had severe TBI. The median age was 33 years with male preponderance (M: F=4:1). Thirty three (29.5 %) patients died within 6 months of TBI, and 38 (33.9 %) patients had an unfavorable outcome. Pearson correlation coefficient showed good correlation between observed and predicted outcomes. Hosmer-Lemeshow test showed good model fit for IMPACT core, IMPACT extended and IMPACT lab in diagnosing mortality and unfavorable outcome in six months (p>0.05). The ROC curve indicated that all 3 models could accurately discriminate between favorable and unfavorable outcomes, as well as between survival and mortality (unfavorable outcome AUC= 0.905, 0.940, 0.955; mortality AUC= 0.875, 0.914, 0.917 respectively) in our patient population. Conclusion: The IMPACT score is a good prognostic model to predict 6-month outcomes in moderate to severe TBI at admission in Nepalese patient population. Among the three IMPACT models, IMPACT lab has the greatest discriminating ability.

BACKGROUND AND OBJECTIVES:Traumatic brain injury (TBI) is a major public health challenge in India but there is a lack of high-quality data on its clinical characteristics and outcomes. We aimed to describe the TBI population of a tertiary care center in India, identify predictors of inpatient mortality, and assess the performance of existing prognostic tools.METHODS:We conducted a prospective observational cohort study of patients admitted to a high-volume tertiary care center in Vellore, India, after a TBI between 2013 and 2019.RESULTS:We identified 3172 patients (2667 males, 84%) admitted after a TBI (median age = 34 years [IQR 23-48]). Two-wheeler road traffic accidents caused 2259 (71%) injuries, in which 13 (0.6%) patients were wearing a helmet. There were 174 (5%) inpatient deaths (median length of stay = 6 days [IQR 4-10]) and overall mortality (median follow-up = 6 months [IQR 3-9]) was 17% (n = 540). Age, Glasgow Coma Scale motor score, systolic blood pressure ≤90 mm Hg, and key computed tomography imaging features were independently associated with inpatient mortality. Existing prognostic models predicted inpatient mortality with good performance (International Mission for Prognosis and Analysis of Clinical Trials in TBI: Brier = 0.0876, area under the curve (AUC) = 83% [95% CI 79%-87%]; Rotterdam CT: Brier = 0.0890, AUC 79% [95% CI 75%-83%]), but showed poorer performance for post-discharge mortality (International Mission for Prognosis and Analysis of Clinical Trials in TBI: Brier = 0.134, AUC = 75% [95% CI 72%-78%]; Rotterdam CT: Brier = 0.145, AUC 66% [95% CI 63%-69%]).CONCLUSION:In a tertiary care center in India, we described a predominantly young male TBI population with a high contribution of 2-wheeler road traffic accidents and significant post-discharge mortality. Existing prognostic models showed poor performance when predicting which patients died after discharge. These findings should inform public health interventions to reduce the significant burden of TBI in India.

We sought to assess the prognostic value of incorporating daily inpatient physiological biomarker trajectories to the International Mission for Prognosis and Analysis of Clinical Trials in TBI (IMPACT) model for mortality and morbidity six months after severe traumatic brain injury (TBI). Patients with severe TBI (presenting Glasgow Coma Scale ≤ 8) were prospectively collected in a single-center database (n = 598). Morbidity (yes/no) was defined as Glasgow Outcome Scale Extended of 1-4. Daily blood labs (e.g., glucose, sodium, platelets, hemoglobin, neutrophils, lymphocytes, creatinine, blood urea nitrogen) were extracted for the first 14days after injury. IMPACT was compared with IMPACT-extended (IMPACT + daily lab trajectories) for area under the curve (AUC). Net reclassification index (NRI) assessed the number of correctly reclassified cases for IMPACT-extended compared with IMPACT. IMPACT-extended had a better AUC than IMPACT for mortality (AUC 0.93 vs. 0.84, P < 0.001) and morbidity (AUC 0.84 vs. 0.80, P < 0.001). NRI analyses revealed IMPACT-extended improved correct classifications of patients who survived to 6months by 41%. NRI analyses revealed that IMPACT-extended modestly improved correct classification of controls by 4% compared with IMPACT. Incorporating trajectories for daily blood biomarkers of physiological function significantly improves the discrimination and clinically relevant performance of existing prognostic models for both morbidity and mortality six months following severe TBI.

Is It Reliable to Predict the Outcome of Elderly Patients with Severe Traumatic Brain Injury Using the IMPACT Prognostic Calculator?

Computed tomography (CT) is an important imaging modality for guiding prognostication in patients with traumatic brain injury (TBI). However, because of the specialized expertise necessary, timely and dependable TBI prognostication based on CT imaging remains challenging. This study aimed to enhance the efficiency and reliability of TBI prognostication by employing machine learning (ML) techniques on CT images. A retrospective analysis was conducted on the Collaborative European NeuroTrauma Effectiveness Research in TBI (CENTER-TBI) data set (n = 1016). An ML-driven binary classifier was developed to predict favorable or unfavorable outcomes at 6 months post-injury. The prognostic performance was assessed using the area under the curve (AUC) over fivefold cross-validation and compared with conventional models that depend on clinical variables and CT scoring systems. An external validation was performed using the Comparative Indian Neurotrauma Effectiveness Research in Traumatic Brain Injury (CINTER-TBI) data set (n = 348). The developed model achieved superior performance without the necessity for manual CT assessments (AUC = 0.846 [95% CI: 0.843-0.849]) compared with the model based on the clinical and laboratory variables (AUC = 0.817 [95% CI: 0.814-0.820]) and established CT scoring systems requiring manual interpretations (AUC = 0.829 [95% CI: 0.826-0.832] for Marshall and 0.838 [95% CI: 0.835-0.841] for International Mission for Prognosis and Analysis of Clinical Trials in TBI [IMPACT]). The external validation demonstrated the prognostic capacity of the developed model to be significantly better (AUC = 0.859 [95% CI: 0.857-0.862]) than the model using clinical variables (AUC = 0.809 [95% CI: 0.798-0.820]). This study established an ML-based model that provides efficient and reliable TBI prognosis based on CT scans, with potential implications for earlier intervention and improved patient outcomes.

This study extends our previous investigation regarding the effect of nondifferential dichotomous Glasgow Outcome Scale (GOS) misclassification in traumatic brain injury (TBI) clinical trials to the effect of GOS misclassification on ordinal analysis in TBI clinical trials. The impact of GOS misclassification and ordinal outcome analysis was explored via probabilistic sensitivity analyses using TBI patient datasets from the IMPACT database (n = 9205). Three patterns of misclassification were explored given the pre-specified misclassification distributions. For the random pattern, we specified a trapezoidal distribution (minimum: 80%, mode: 85%, and 95%, maximum: 100%) for both sensitivity and specificity; for the upward pattern, the same trapezoidal distribution for sensitivity but with a perfect specificity; and for the downward pattern, the same trapezoidal distribution for specificity but with a perfect sensitivity. The conventional 95% confidence intervals and simulation intervals, which accounts for the misclassification and random errors together, were reported. The results showed that given the specified misclassification distributions, the misclassification with a random or upward pattern would have caused a slightly underestimated outcome in the observed data. However, the misclassification with a downward pattern would have resulted in an inflated estimation. Thus the sensitivity analysis suggests that the nondifferential misclassification can cause uncertainties on the primary outcome estimation in TBI trials. However, such an effect is likely to be small when ordinal analysis is applied, compared with the impact of dichotomous GOS misclassifications. The result underlines that the ordinal GOS analysis may gain from both statistical efficiency, as suggested by several recent studies, and a relatively smaller impact from misclassification as compared with conventional binary GOS analysis.

Cognitive impairment is a common sequela following traumatic brain injury (TBI). This study aimed to identify risk factors for cognitive impairment after 3 and 12 months of TBI and to create nomograms to predict them. A total of 305 mild-to-moderate TBI patients admitted to the First Affiliated Hospital with Nanjing Medical University from January 2018 to January 2022 were retrospectively recruited. Risk factors for cognitive impairment after 3 and 12 months of TBI were identified by univariable and multivariable logistic regression analyses. Based on these factors, we created two nomograms to predict cognitive impairment after 3 and 12 months of TBI, the discrimination and calibration of which were validated by plotting the receiver operating characteristic (ROC) curve and calibration curve, respectively. Cognitive impairment was detected in 125/305 and 52/305 mild-to-moderate TBI patients after 3 and 12 months of injury, respectively. Age, the Glasgow Coma Scale (GCS) score, >12 years of education, hyperlipidemia, temporal lobe contusion, traumatic subarachnoid hemorrhage (tSAH), very early rehabilitation (VER), and intensive care unit (ICU) admission were independent risk factors for cognitive impairment after 3 months of mild-to-moderate TBI. Meanwhile, age, GCS score, diabetes mellitus, tSAH, and surgical treatment were independent risk factors for cognitive impairment after 12 months of mild-to-moderate TBI. Two nomograms were created based on the risk factors identified using logistic regression analyses. The areas under the curve (AUCs) of the two nomograms to predict cognitive impairment after 3 and 12 months of mild-to-moderate TBI were 0.852 (95% CI [0.810, 0.895]) and 0.817 (95% CI [0.762, 0.873]), respectively. Two nomograms are created to predict cognitive impairment after 3 and 12 months of TBI. Age, GCS score, >12 years of education, hyperlipidemia, temporal lobe contusion, tSAH, VER, and ICU admission are independent risk factors for cognitive impairment after 3 months of TBI; meanwhile, age, the GCS scores, diabetes mellitus, tSAH, and surgical treatment are independent risk factors of cognitive impairment after 12 months of TBI. Two nomograms, based on both groups of factors, respectively, show strong discriminative abilities.

BackgroundThe inflammasome plays an important role in the inflammatory innate immune response after central nervous system (CNS) injury. Inhibition of the inflammasome after traumatic brain injury (TBI) results in improved outcomes by lowering the levels of caspase-1 and interleukin (IL)-1b. We have previously shown that inflammasome proteins are elevated in the cerebrospinal fluid (CSF) of patients with TBI and that higher levels of these proteins were consistent with poorer outcomes after TBI when compared to patients that presented these inflammasome proteins at lower levels.Methods and findingsHere we extend our work by analyzing serum from 21 TBI patients and CSF from 18 TBI patients compared to 120 serum samples and 30 CSF samples from no-TBI donor controls for the expression of caspase-1, apoptosis-associated speck-like protein containing a caspase recruitment domain (ASC), interleukin(IL)-1b and IL-18. Analysis was carried out using the Ella Simple Plex system (Protein Simple) to determine the sensitivity and specificity of inflammasome proteins as biomarkers of TBI. Receiver operator characteristic (ROC) curves, confidence intervals and likelihood ratios for each biomarker was determined. ROC curves, confidence intervals, sensitivity and specificity for each biomarker examined revealed that caspase-1 (0.93 area under the curve (AUC)) and ASC (0.90 AUC) in serum and ASC (1.0 AUC) and IL-18 (0.84 AUC) in CSF are promising biomarkers of TBI pathology. Importantly, higher protein levels (above 547.6 pg/ml) of ASC (0.91 AUC) were consistent with poorer outcomes after TBI as determined by the Glasgow Outcome Scale-Extended (GOSE).ConclusionThese findings indicate that inflammasome proteins are excellent diagnostic and predictive biomarkers of TBI.

Objective To investigate the clinical practicability and prognostic value of Helsinki CT score in patients with traumatic brain injury (TBI). Methods A retrospective case series study was conducted to analyze the clinical data of 124 TBI patients admitted to First Affiliated Hospital of Xinjiang Medical University from September 2016 to October 2018. There were 91 males and 33 females, aged 14-84 years, with an average age of 49 years. Glasgow coma score (GCS) at admission ranged from 3-8 points in 45 patients, 9-12 points in 42 patients, and 13-15 points in 37 patients. According to Glasgow outcome scale (GOS) at 6 months after injury, 26 patients were classified into the poor prognosis group with GOS of 1-3 points and 98 patients were in the good prognosis group with GOS of 4-5 points. The prognosis-related risk factors were analyzed, and the role of Helsinki CT score to predict the adverse prognosis and mortality of TBI patients in the two groups was investigated. The sensitivity and specificity of Helsinki CT Score for 6-month poor prognosis were evaluated by receiver operation characteristic (ROC) curve and area under the curve (AUC). Results Univariate analysis suggested that there were significant differences in terms of subdural hematoma, intracranial hematoma, extradural hematoma, hematoma volume >25 cm3, intraventricular hemorrhage and suprasellar cistern pressure between the poor prognosis group and good prognosis group (P 0.05). The Helsinki CT score could independently predict the adverse prognosis and mortality of TBI patients at 6 months (multivariate logistic regression: ORdeath=1.21, ORadverse prognosis=1.14). Helsinki CT score had a better predictive ability of 6-month mortality (AUC=0.85) than that of 6-month adverse prognosis (AUC=0.76), and had a predictive value for 6-month mortality and adverse prognosis. Conclusions Subdural hematoma, extradural hematoma, intraventricular hemorrhage and suprasellar cistern state (compression or disappearance) are the risk factors for the poor prognosis of TBI patients. Intraventricular hemorrhage and suprasellar cistern state are the main risk factors for predicting the mortality of 6 months. Helsinki CT score can independently predict the adverse prognosis and mortality of TBI patients at 6 months, and has relatively better value in predicting the mortality. Key words: Brain injuries; Tomography, X-ray; Helsinki CT score

Laboratory studies have identified numerous potential therapeutic interventions that might have clinical application for the treatment of human traumatic brain injury. Many of these therapies have progressed into human clinical trials in severe traumatic brain injury. Numerous trials have been completed, and many others have been prematurely terminated or are currently in various phases of testing. The results of the completed Phase III trials have been generally disappointing, compared with the expectations produced by the successes of these interventions in animal laboratory studies. In this review, we summarize the current status of human traumatic brain injury clinical trials, as well as the animal laboratory studies that led to some of these trials. We summarize criteria for conducting clinical trials in severe traumatic brain injury, with suggestions for future improvements. We also attempt to identify factors that might contribute to the discrepancies between animal and human trials, and we propose recommendations that could help investigators avoid certain pitfalls in future clinical trials in traumatic brain injury.

Clinical trials in traumatic brain injury (TBI) have been fraught with failure due in part to heterogeneity in pathology and insensitive outcome measurements. The International Mission for Prognosis and Analysis of Clinical Trials in TBI (IMPACT) prognostic model has been purposed as a means of risk adjustment and outcome prediction for use in trial design and analysis. The purpose of this study was to evaluate the performance of the IMPACT model in predicting 6-month functional outcome and mortality using prospectively collected data at a large, Level 1 neurotrauma center. This population-based cohort study included all TBI patients ≥14 years of age admitted with a Glasgow Coma Scale (GCS) score of ≤8 (severe TBI) to the University of Pittsburgh Medical Center between July 1994 and May 2009. Clinical data were prospectively collected and linked to 6-month functional outcome (Glasgow Outcome Scale [GOS]) and mortality. The discriminatory power and calibration of the three iterations of the IMPACT model (core, extended, and lab) were assessed using multiple regression analyses and indicated by the area under the receiver operating characteristic curve (AUC). A sample of 587 patients was available for analysis; the mean age was 37.8±17 years. The median 6-month GOS was 3 (IQR 3); 6-month mortality was 41%. The prognostic models were composed of age, motor score, and pupillary reactivity (core model), Marshall grade on head CT and secondary insults (extended), and laboratory values (lab); all of these displayed good prediction ability for unfavorable outcome and mortality (unfavorable outcome AUC=0.76, 0.79, 0.76; mortality AUC=0.78, 0.83, 0.83, respectively). All model iterations displayed adequate calibration for predicting unfavorable outcome and mortality. Prospective, independent validation supports the IMPACT prognostic model's prediction of patient 6-month functional status and mortality after severe TBI. The IMPACT prognostic model is an effective instrument to assist TBI study design and analysis.

INTRODUCTION: Nearly three quarters of deaths in clinical trials of severe traumatic brain injury (TBI) occur due to withdrawal of life sustaining therapies, often due to perceived poor neurological prognosis. Most of these deaths occur within the first week when physicians historically lack enough accuracy to make these decisions. Prognostic models, such as the International Mission for Prognosis and Analysis of Clinical Trials in TBI, are designed to prognose outcomes upon admission and fail to utilize clinical information accumulated after admission. Quantitative prognostic models capturing this informaiton are needed to help guide care decisions. METHODS: We built our model using 266 patients from a prospective database of severe TBI patients from 2002 through 2018 with an MRI obtained 1 to 9 days post-trauma. We developed a customized, multi-channel, deep convolutional neural network model with an AlexNet backbone and pre-trained with ImageNet. The cohort was split into 80% training, 10% validation, and 10% independent testing set. We built separate models using T2 FLAIR and DWI sequences. To reduce the heterogeneity of imaging after a decompressive hemicraniectomy (DHC), we only used DHC patients for model training. RESULTS: No adverse clinical events occurred while obtaining early MRIs. The overall mortality was 25% (68) and 20% (34) in the 170 patients who did not have a DHC. In the independent test cohort, the T2 FLAIR model performed best, with an area under the curve (AUC) of 0.80 (95% Confidence Interval: 0.57-0.94) and accuracy of 83%. The DWI model had an AUC of 0.64 (0.50 – 0.79) with an accuracy of 78%. With a specificity of 100% (i.e., never recommending withdrawal of care in a patient who would otherwise survive), the sensitivity for mortality was 57%. For identifying patients with favorable recovery potential, the sensitivity was 83% with a specificity of 80%. CONCLUSION: Early MRIs are safe to obtain in the severe TBI patient population. Deep learning analysis of these MRIs can predict 6-month mortality to help guide neurological prognostication.

Prognostic models for outcome prediction in patients with traumatic brain injury (TBI) are important instruments in both clinical practice and research. To remain current a continuous process of model validation is necessary. We aimed to investigate the performance of the International Mission on Prognosis and Analysis of Clinical Trials in TBI (IMPACT) prognostic models in predicting mortality in a contemporary New York State TBI registry developed and maintained by the Brain Trauma Foundation. The Brain Trauma Foundation (BTF) TBI-trac® database contains data on 3125 patients who sustained severe TBI (Glasgow Coma Scale [GCS] score ≤ 8) in New York State between 2000 and 2009. The outcome measure was 14-day mortality. To predict 14-day mortality with admission data, we adapted the IMPACT Core and Extended models. Performance of the models was assessed by determining calibration (agreement between observed and predicted outcomes), and discrimination (separation of those patients who die from those who survive). Calibration was explored graphically with calibration plots. Discrimination was expressed by the area under the receiver operating characteristic (ROC) curve (AUC). A total of 2513 out of 3125 patients in the BTF database met the inclusion criteria. The 14-day mortality rate was 23%. The models showed excellent calibration. Mean predicted probabilities were 20% for the Core model and 24% for the Extended model. Both models showed good discrimination with AUCs of 0.79 (Core) and 0.83 (Extended). We conclude that the IMPACT models validly predict 14-day mortality in the BTF database, confirming generalizability of these models for outcome prediction in TBI patients.

Objective: The Neurological Outcome Scale for Traumatic Brain Injury (NOS-TBI) is a measure designed to assess neurological functioning in traumatic brain injury (TBI) across the spectrum of recovery. We hypothesized the NOS-TBI would demonstrate predictive validity and be more sensitive to change than other well-established measures. Background Clinical trials in TBI frequently rely on a single global outcome measure. Available tools do not provide an assessment of neurological functioning, per se. We recently developed a measure of neurological functioning in TBI, the NOS-TBI. While the NOS-TBI has been validated in patients with TBI undergoing rehabilitation up to 18 months post-injury, its performance in TBI-related clinical trials and its sensitivity to change over time are unknown. Design/Methods: We analyzed data from the National Acute Brain Injury Study: Hypothermia-II (NABIS:H-II) clinical trial. Patients were 16-45 years with severe TBI, assessed at 3, 6, and 12 months post-injury. A subset was assessed at 1 month. For analysis of criterion-related validity, Spearman correlations were performed comparing the NOS-TBI to Glasgow Outcome Scale (GOS), GOS-Extended (GOS-E), Disability Rating Scale (DRS), and Marshall CT classification. Sensitivity to change was analyzed using the Wilcoxon signed-rank sum test. Results: Significant correlation was observed between the NOS-TBI and the GOS, GOS-E, DRS, and NRS-R total scores, demonstrating the concurrent validity of these tools. The GCS, Marshall CT classification, and lesion volume demonstrated overall limited ability to predict outcome at 3, 6, and 12 months post-injury. In contrast, the NOS-TBI total score at both 1 and 3-months post-injury demonstrated a significant ability to predict outcome at 3, 6, and 12 months post-injury. Trial subjects whose recovery appeared to have stabilized based on GOS and GOS-E scores, demonstrated significant changes in their level of neurologic recovery using the NOS-TBI. Conclusions: The NOS-TBI demonstrated significantly better outcome prediction and sensitivity to change compared with well-established outcome measures. Supported by: In part by grant NS 43353 from the NIH/NINDS, Guy L. Clifton, Principal Investigator. Disclosure: Dr. Moretti has nothing to disclose. Dr. McCauley has nothing to disclose. Dr. Wilde has nothing to disclose. Dr. Levin has nothing to disclose. Dr. Clifton has nothing to disclose.

Predicting functional outcome in critically ill patients with traumatic brain injury (TBI) strongly influences end-of-life decisions and information for surrogate decision makers. Despite well-validated prognostic models, clinicians most often rely on their subjective perception of prognosis. In this study, we aimed to compare physicians' predictions with the International Mission on Prognosis and Analysis of Clinical Trials in TBI (IMPACT) prognostic model for predicting an unfavorable functional outcome at 6months after moderate or severe TBI. PREDICT-TBI is a prospective study of patients with moderate to severe TBI. Patients were admitted to a neurocritical care unit and were excluded if they died or had withdrawal of life-sustaining treatments within the first 24h. In a paired study design, we compared the accuracy of physician prediction on day 1 with the prediction of the IMPACT model as two diagnostic tests in predicting unfavorable outcome 6months after TBI. Unfavorable outcome was assessed by the Glasgow Outcome Scale from 1 to 3 by using a structured telephone interview. The primary end point was the difference between the discrimination ability of the physician and theIMPACT model assessed by the area under the curve. Of the 93 patients with inclusion and exclusion criteria, 80 patients reached the primary end point. At 6months, 29 patients (36%) had unfavorable outcome. A total of 31 clinicians participated in the study. Physicians' predictions showed an area under the curve of 0.79 (95% confidence interval 0.68-0.89), against 0.80 (95% confidence interval 0.69-0.91) for the laboratory IMPACT model, with no statistical difference (p = 0.88). Both approaches were well calibrated. Agreement between physicians was moderate (κ = 0.56). Lack of experience was not associated with prediction accuracy (p = 0.58). Predictions made by physicians for functional outcome were overall moderately accurate, and no statistical difference was found with the IMPACT models, possibly due to a lack of power. The significant variability between physician assessments suggests prediction could be improved through peer reviewing, with the support of the IMPACT models, to provide a realistic expectation of outcome to families and guide discussions about end-of-life decisions.