Comparing the Full Outline of Unresponsiveness (FOUR) score and the Glasgow Coma Scale (GCS) for predicting outcomes in traumatic brain injury: a comparative study Authors: Harsh Jain, Jishnu N. Nair, Davuluri Venkata Shashank, Visvanathan Krishnaswamy, Krishnamurthy Ganesh Published: 2025 ABSTRACT PurposeTraumatic brain injuries (TBIs) are a leading cause of morbidity, mortality, socioeconomic loss, and diminished quality of life among survivors of all injury types. The Glasgow Coma Scale (GCS), though commonly used in the intensive care unit (ICU), has limitations that may be addressed by the Full Outline of Unresponsiveness (FOUR) score. We compared the ability of the FOUR and GCS scores to predict outcomes in patients with TBI.MethodsThis prospective observational study included all patients presenting with head injury who survived beyond 24 hours between 2020 and 2022. Basic demographic data, prehospital care details, and neurological findings were collected. TBI severity was assessed using both the GCS and FOUR scores, and outcomes were graded using the Glasgow Outcome Score (GOS).ResultsAmong 225 patients (mean age, 42.4±19.7 years; male sex, 77.8%) with TBIs, the mortality rate was 4.4%, and 15.1% experienced a poor outcome (GOS ≤2). Receiver operating characteristic curve analysis showed good predictive ability for both the FOUR score (area under the curve [AUC], 0.85) and GCS (AUC, 0.88). FOUR score ≤12 and GCS ≤10 were identified as optimal cutoff values for predicting poor outcomes (P<0.001), with the FOUR score demonstrating sensitivity and specificity of 79.4% and 77.5%, respectively. A FOUR score ≤12 at admission was significantly associated with intubation (odds ratio [OR], 8.2; 95% confidence interval [CI], 4.6–14.4; P<0.001), tracheostomy (OR, 21.8; 95% CI. 6.9–68.8; P<0.001), and neurological complications, including mental function deficits (OR, 7.5; 95% CI, 3.2–17.7; P<0.001), cranial nerve palsy (OR, 4.1; 95% CI, 1.6–10.4; P=0.002), and motor deficits (OR, 9.4; 95% CI, 4.6–19.2; P<0.001). A strong correlation was observed between the FOUR score and GOS (r=0.7).ConclusionsThe FOUR score is a reliable tool for assessing TBI severity and predicting outcomes, with performance comparable to the GCS. An admission FOUR score ≤12 was associated with poor neurological outcomes, increased need for advanced airway management, and long-term disability. By capturing additional neurological parameters beyond the GCS, the FOUR score may offer clinical advantages in ICU settings for early prognostication and management of patients with TBI. INTRODUCTION Background Traumatic brain injury (TBI) is an acquired brain injury that occurs when the head is subjected to a sudden or forceful impact, or when an object penetrates the skull and enters brain tissue [1]. The severity of TBIs ranges from mild concussions to severe and irreversible brain damage. Mild TBI, often referred to as a brain concussion, was once considered a benign event but has since gained considerable attention due to its potential for long-term neuropsychological sequelae [1]. Moderate to severe TBIs remain leading causes of injury-related mortality and long-term disability. Although TBIs are reported in large numbers each year, approximately 80% are classified as mild head injuries [1]. The incidence of TBIs is approximately 500 cases per 100,000 population, with adolescents and older adults being the most affected age groups [1]. The frontal and temporal lobes are the regions most frequently injured [1]. In India, it is estimated that nearly 1 million people sustain a TBI annually, with approximately 200,000 resulting in death [2]. TBIs continue to be a major public health concern, representing a leading cause of morbidity, mortality, disability, socioeconomic loss, and reduced quality of life among survivors across all types of injury. Scoring systems are commonly utilized in intensive care settings to assess consciousness levels and facilitate standardized communication among healthcare providers regarding neurological status. The Glasgow Coma Scale (GCS) is the most widely used and thoroughly studied coma scale to date [3]. Originally designed to provide a shared language for evaluating the depth and duration of impaired consciousness and coma, the GCS has become the gold standard against which newer scales are compared. Despite its extensive use, the GCS has notable limitations, including variability in interrater reliability and predictive validity, inability to assess the verbal component in intubated patients, and failure to evaluate breathing patterns, brainstem reflexes, or subtle changes in neurological examination [4,5]. To address these shortcomings, the Full Outline of Unresponsiveness (FOUR) score was developed in 2005 at the Mayo Clinic as a modification of the GCS. The FOUR score provides a more comprehensive assessment of neurological function by evaluating four components: eye response, motor response, brainstem reflexes, and respiratory patterns. Objectives In this study, we aimed to compare the predictive performance of the GCS and FOUR score in TBI outcomes. We also sought to determine the FOUR score cutoff above which the risk for adverse outcomes, including intubation, cranial nerve deficits, and other neurological complications, increases significantly. METHODS Ethics statement This study was approved by the Institutional Ethics Committee of Sri Ramachandra Institute of Higher Education and Research, (No. Csp/20/DEC/88/236). Written informed consent was obtained from the patients or their family member. Study design A prospective, observational study was conducted in the neurosurgery department at a tertiary care hospital in South India over an 18-month period between 2020 and 2022. All patients presenting to the emergency room and subsequently admitted to the neurosurgery department with clinical or radiological evidence of head injury (either in isolation or in combination with other injuries) were included in the study after obtaining informed consent. Study population Patients admitted to the neurosurgical intensive care unit (ICU) following direct head injury who survived beyond 24 hours after admission were included in the study. Exclusion criteria comprised any history of neuromuscular agent intake or neuromuscular disease, as determined through interviews with patients, their attendants, or review of hospital records. The study was conducted by a team of five individuals: a medical student, two neurosurgery residents, and two attending neurosurgeons. Exposure variables The primary exposure variables were the GCS and FOUR scores at presentation, both recorded by two neurosurgical residents after being trained by the attendings. Detailed information regarding patient demographics, prehospital care, type of injury, radiological findings, clinical examination, neurological findings, and management strategies was collected. Based on the GCS score, TBI cases were classified as mild (13–15), moderate (9–12), or severe (≤8). Patients were also graded using the FOUR score (maximum of 16) according to eye response, motor response, brainstem reflexes, and respiration pattern (Table 1). Importantly, in patients who were intubated at admission and therefore unable to have their verbal response assessed, the GCS score was approximated by assigning a verbal score of 1 added to the sum of their eye and motor responses. This approach highlights a major drawback of the GCS, as no accurate method exists to assess verbal response in intubated patients. Outcome variables The primary outcome was the Glasgow Outcome Score (GOS) at the time of discharge, recorded by the neurosurgical residents. Patients were monitored daily by neurosurgical residents under the supervision of the attendings for any neurological improvement or deterioration, the occurrence of complications, and the need for ventilatory support or tracheostomy. Outcomes were further dichotomized as favorable (GOS >2) or poor (GOS ≤2, indicating death or persistent vegetative state). Although this threshold is more stringent, it was chosen to focus on patients with the most severe neurological impairment, and is supported by previous studies that have adopted similar dichotomization in TBI populations [6,7]. Statistical analysis Data were presented as proportions for categorical variables, and as means with standard deviations for continuous variables. Differences between categorical variables were assessed using the chi-square test, while continuous variables were compared using the two-tailed t-test. The significance threshold was set at α<0.05. For continuous variables not normally distributed, the nonparametric Mann-Whitney test was employed. The Fisher exact test was used for categorical variables when the sample size was very small. Receiver operating characteristic (ROC) curve analysis was performed to calculate the area under the curve (AUC). The Youden index was calculated and used to determine cutoff values that maximize sensitivity and specificity. All analyses were conducted using IBM SPSS ver. 29 (IBM Corp). RESULTS Patient demographics Of the 225 patients admitted to the neurosurgical ICU with TBIs during the study period, the mean age was 42.4±19.7 years, and 175 (77.8%) were male. The most common mode of injury was motor vehicle collision, with 159 patients (70.7%), and 176 patients (78.2%) experienced loss of consciousness following the injury. Subdural hemorrhage was observed in 72 patients (32.0%), while 59 (26.2%) had a subarachnoid hemorrhage. The mean FOUR and GCS scores at presentation were 12.8±3.6 and 11.8±3.5, respectively (Table 2). Across the entire cohort, 10 patients (4.4%) died during the hospital stay, and 34 (15.1%) experienced a poor outcome (GOS ≤2). When comparing patients with GOS ≤2 and those with GOS >2 at presentation, patients with a poor outcome were older (51.8±19.5 years vs. 40.7±19.3 years, P=0.002), more likely to have hypertension (26.5% vs. 11.5%, P=0.021), and more often presented with altered sensorium (32.4% vs. 15.7%, P=0.017) and motor deficits (85.3% vs. 33.5%, P<0.001). At admission, those with GOS ≤2 had a lower FOUR score (8.3±4.3 vs. 13.6±3.4, P<0.001) and GCS score (7.2±3.4 vs. 12.6±2.8, P<0.001). Radiological imaging revealed a higher incidence of intracerebral hemorrhage (32.4% vs. 7.3%, P<0.001) among patients with poor outcomes. During hospitalization, patients with GOS ≤2 were significantly more likely to require ventilatory support (97.1% vs. 15.2%, P<0.001) and tracheostomy (64.7% vs. 6.8%, P<0.001), and to develop a higher mental function deficit (64.7% vs. 3.1%, P<0.001) (Table 2). FOUR and GCS scores as predictors of outcomes ROC curves were plotted for both the FOUR score and GCS (Fig. 1), and their reliability in predicting outcomes for survivors and nonsurvivors was assessed. The AUC for the FOUR score was 0.83, and 0.84 for the GCS score. Both the GCS and FOUR scores were significantly lower in nonsurvivors compared to survivors (both P<0.001). Patients with a FOUR score of <7 at admission were significantly more likely to die during their hospital stay (P<0.001). Due to the low mortality rate (4.4%) observed in our study, we further analyzed the predictive reliability of both scores for poor outcomes (GOS ≤2). The corresponding ROC curve is depicted in Fig. 2. Based on AUC values, both the FOUR score (AUC, 0.85; P<0.001) and GCS (AUC, 0.88; P<0.001) demonstrated good discriminative power (Table 3). The optimal cutoff points, determined by the highest Youden index (sensitivity + specificity – 1), were 12.5 for the FOUR score and 10.5 for the GCS score. For further analysis, as both GCS and FOUR scores are ordinal, patients were categorized as having FOUR score ≤12 or >12, and GCS score ≤10 or >10. Patients with a GCS score ≤10, and those with a FOUR score ≤12 at admission, were more likely to have a poor prognosis (GOS ≤2) and to end up in a vegetative state or die during their hospital stay (P<0.001). The sensitivity and specificity of the FOUR score in predicting poor outcomes were 79.4% and 77.5%, respectively. Furthermore, patients with a FOUR score of ≤12 at admission were significantly more likely to require intubation (odds ratio [OR], 8.2; 95% confidence interval [CI], 4.6–14.4; P<0.001) and ventilation (OR, 8.5; 95% CI, 4.8–15.0; P<0.001), undergo tracheostomy (OR, 21.8; 95% CI. 6.9–68.8; P<0.001), develop a higher mental function deficit (OR, 7.5; 95% CI, 3.2–17.7; P<0.001), experience cranial nerve palsy (OR, 4.1; 95% CI, 1.6–10.4; P=0.002), and have motor deficits (OR, 9.4; 95% CI, 4.6–19.2; P<0.001). The likelihood of developing a complication during the hospital stay is summarized in Table 4. A Pearson correlation analysis demonstrated a strong correlation between the FOUR score and GOS (r=0.7). DISCUSSION The current study included patients with TBIs and compared the performance of the GCS and FOUR scores in predicting patient outcomes. Both the GCS and FOUR scores were significantly lower in patients who died during hospitalization. A lower FOUR score at admission was strongly associated with an increased likelihood of mortality. Because the mortality rate in our study was relatively low, we further analyzed the ability of the GCS and FOUR scores to predict poor outcomes (GOS ≤2), including severe disability or a vegetative state. Both scores demonstrated good predictive accuracy. Patients with a FOUR score ≤12 at admission were more likely to require intubation, ventilation, and tracheostomy during hospitalization. These patients also faced a higher risk of neurological complications, including cognitive deficits, cranial nerve palsies, and motor impairments. Although the FOUR score has been extensively studied over the last decade, research evaluating its validity in differentiating vegetative state outcomes and establishing precise cutoff values below which patient outcomes worsen remains limited. Despite evidence of the FOUR score’s efficacy, its clinical adoption has been restricted, particularly in the Indian subcontinent. A strong correlation was observed between the FOUR score and patient outcomes, underscoring its prognostic value. Although the AUC for the GCS score was slightly higher than for the FOUR score, it is important to recognize that, for intubated patients, the GCS verbal response was defaulted to 1, thus approximating the lowest possible score. Some of these patients may have been able to vocalize if not intubated, meaning their actual GCS score could have been higher. To date, no literature describes a method to precisely approximate the GCS for intubated patients. This inability to accurately assess the GCS in a relatively sicker, intubated cohort is a key reason for advocating the use of the FOUR score, which addresses this limitation. Our findings are consistent with previous studies reporting that higher FOUR and GCS scores were significantly associated with lower mortality rates and more favorable prognoses [8–10]. In a meta-analysis by Foo et al. [11], the AUC for the FOUR and GCS scores for predicting 1-month mortality ranged between 0.62 and 0.99. In another study, Ghelichkhani et al. [12] found a higher AUC for FOUR (0.99) than for GCS (0.97), similarly, the AUCs were 0.82 and 0.79, respectively, in a study by Khanal et al [8]. A prospective study by Zeiler et al. [13] evaluated the ability of FOUR score to predict outcomes in patients with aneurysmal subarachnoid hemorrhage and found a statistically significant association between the FOUR score and mortality [14]. In our study, the optimal cutoff scores determined by the Youden index for FOUR and GCS scores were 12.5 and 10.5, respectively, higher than those reported by Khanal et al. [8] (cutoff 6.5 for both scores) and Yousefzadeh-Chabok et al. [9] (cutoff of 8 for GCS score). The higher cutoff values observed in our study may be attributable not only to the relatively low mortality rate in our cohort but also to our broader definition of poor outcome. Unlike prior studies that focused solely on mortality, we defined poor outcome as GOS ≤2, including both death and persistent vegetative state. This inclusive definition encompasses a wider range of severely impaired patients and may naturally yield higher cutoff scores using ROC-based analysis. Our study also adds to the limited literature on the FOUR score as a predictor of in-hospital clinical course and deficits at discharge. Patients with a FOUR score ≤12 at admission were significantly more likely to experience a complicated hospital course, with higher risks of neurological decline and a greater need for ventilatory support. We also observed a strong correlation between the FOUR score and GOS (r=0.7), similar to findings by Khanal et al. [8] (coefficient, 0.91) and Iyer et al. [10] (0.98 for the first rater, 0.92 for the second rater). Said et al. [14] compared the FOUR and GCS scores as predictors of successful extubation at 14 days, 28-day mortality, and neurological outcomes at 3 months. They reported that a GCS score of ≤7 and a FOUR score of <10 predicted extubation failure at 14 days, with a higher AUC for the FOUR score. The 28-day mortality and 3-month neurological outcomes were predicted with similar accuracy by both scores [12]. Notably, prior studies have reported mortality rates ranging from 10% to 34.8%—substantially higher than the 4.4% mortality rate observed in our study [13,15–17]. Although our mortality rate was low, a significant proportion of patients experienced unfavorable outcomes, which are equally important and devastating for patients and their families. It is therefore essential that families are adequately counseled regarding the likely course of the patient’s condition. Accordingly, we further analyzed the FOUR score’s ability to predict adverse events, the need for intubation, and overall unfavorable outcomes, concluding that a FOUR score of ≤12 at admission significantly increased the risk of complications. Limitations This study has several limitations. First, the sample size is known to influence calibration, and although our sample size was larger than most available studies, further large-scale studies are needed in South Asia to better assess the reliability of the FOUR score in this population. Second, differences in clinical settings, case mix, quality of care, and institutional policies may introduce bias and affect results. Developing setting-specific models may enhance the predictive performance of these scoring systems. Third, the absence of follow-up data after discharge limited our ability to assess long-term patient outcomes. Fourth, we used a more stringent definition of poor outcome (GOS ≤2), focusing on death and vegetative state. While several studies define poor outcome as GOS ≤3, similar thresholds have been used in prior TBI research and align with our objective to identify patients with the most profound impairment. Lastly, we approximated the GCS score for intubated patients by assigning the lowest verbal response (verbal score of 1), which may introduce bias. However, this limitation further justifies the adoption of the FOUR score, which does not depend on the verbal component. Conclusions In this prospective study of patients with TBI, the FOUR score demonstrated strong predictive ability for poor outcomes, comparable to the GCS score. Notably, a FOUR score of ≤12 at admission was significantly associated not only with mortality and poor functional outcomes, but also with the need for critical care interventions such as intubation and tracheostomy, as well as with specific neurological complications including cranial nerve palsy and motor deficits. These findings suggest that, beyond its value in prognostication, the FOUR score may serve as a valuable tool for anticipating clinical deterioration and guiding early management decisions in patients with TBI. Incorporating the FOUR score into the initial neurological assessment may enhance triage and intervention strategies in acute care settings.