Analysis of diagnostic informative value of the Full Outline of UnResponsiveness Scale in patients with spontaneous supratentorial intracerebral hemorrhage

Material and methods. Prospective cohort study of 138 patients in acute period of SSICH was conducted, which included clinical assessment (using the Full Outline of UnResponsiveness (FOUR) Scale and the Glasgow Coma Scale (GCS) scores) and neuroimaging estimation of cerebral injury severity. A comparative analysis of Spearman’s rank correlation coefficients (R) and different areas under the receiver operating characteristic curves (derived from the same cases) was conducted with the help of Z statistic.

Spontaneous supratentorial intracerebral hemorrhage (SSICH) is the leading cause of death and disability of young and middle aged people in most countries of the world [3].
The significant role of intracerebral hematoma volume effect along with the perifocal edema in progressive midline shift development and precondition for unfavorable course and outcome of SSICH acute period has been convincingly proved [12].
The Glasgow Coma Scale (GCS) and the National Institutes of Health Stroke Scale are the most common to quantify the severity of cerebral and focal syndromes [2]. In spite of the fact that the Glasgow Coma Scale is the gold standard for assessing the level of consciousness in clinical angioneurology, its significance for the determination of midline shift severity in patients with cerebral hemorrhagic supratentorial stroke has recently been challenged. The presence of aphasic disorders in patients with dominant hemisphere damage reduces the informative and prognostic value of the Glasgow Coma Scale in the acute period of the disease. In addition, the Glasgow Coma Scale does not include some important signs of the brainstem functional activity evaluation, namely: respiratory pattern, pupillary reactions, as well as corneal, conjunctival and pharyngeal reflexes [15]. All of the above justifies the expediency of more informative clinical scales research for the objectification of midline shift severity and the further determination of vital and functional outcome prognosis in the acute period of the disease.
Taking all of this into account, we pay our particular attention to the Full Outline of UnResponsiveness Scale, which includes the assessment of eye movements, respiratory pattern, motor reactions, brainstem reflexes and motor reactions. The evaluation of each function consists of 5 gradations (0-4) [1,16]. This scale does not take into consideration verbal reactions, which excludes the lateralization of lesion influence on the total score [2]. The abovementioned suggests the prospective viability of the Full Outline of Un-Responsiveness Scale use in patients with SSICH, however, the information as for the diagnostic informative value of this scale in the acute period of the disease are absent, which justifies the viability of the further studies in this area.

The aim
The aim of the study was the analysis of the diagnostic informative value of the Full Outline of UnResponsiveness Scale in patients with SSICH based on the comparison with neuroimaging criteria of midline shift severity.

Materials and methods
In order to achieve this goal, a comparative study was conducted on 138 patients (79 men and 59 women, mean age 63.6 ± 1.0 years), who underwent a therapy in the Brain Circulation Disorders Department of the Municipal Institution "Zaporizhzhia City Clinical Hospital # 6".
Inclusion criteria: 1) men and women with SSICH confirmed by the results of clinical and neuroimaging study; 2) admission to the hospital within the first 24 hours of the disease onset; 3) informed consent signed by the patient for study participation.
Exclusion criteria: 1) ≥2 lesions; 2) combined stroke; 3) acute brain circulation disorders in the past medical history; 4) decompensated somatic pathology; 5) oncopathology. The diagnosis of SSICH was confirmed based on the results of neuroimaging study, which was conducted on admission to hospital using a computed tomography scanner "Siemens Somatom Spirit" (Germany). Lesion size and midline shift severity were considered.
Clinical and neurological study included the assessment of neurological deficit severity on the National Institute of Health Stroke Scale (NIHSS). The Glasgow Coma Scale (GCS) and the Full Outline of UnResponsiveness Scale (FOUR) were used in order to assess the severity of dislocation syndrome clinical signs.
Statistical analysis of the obtained data was carried out with the help of Statisticfa for Windows 13 (StatSoft Inc., № JPZ804I382130ARCN10-J) and MedCalc (version 16.4). The distribution normality of the studied traits was assessed by Shapiro-Wilk criterion. Since the majority of indicators distribution did not comply with the laws of normality, descriptive statistics were presented in the form of the median (Me) and interquartile range (Q 25 -Q 75 ). Spearman's rank correlation coefficient (R) was used for the interrelation between quantitative characteristics assessment. The diagnostic criteria were determined by the ROC analysis. The optimal cut-off values were estimated with the help of Youden index method. A comparative analysis of correlation coefficients and different areas under the receiver operating curves (derived from the same cases) was conducted with the help of Z statistic. Chi-squared test was used for the intergroup differences of diagnostic criteria accuracy assessment. P < 0.05 was considered to indicate a statistically significant difference.
Midline shift was verified in 78 (56.5 %) cases using neuroimaging studies. The results of the septum pellucidum and pineal gland displacement severity assessment in patients are presented in Table 1.
The analysis results of the informative value of the FOUR Scale within the clinical assessment of septum pellucidum displacement severity in comparison with the GCS are presented in Table 2.
As the data shows, the AUC value when using the FOUR Scale as the tool for the verification of clinical signs of severe septum pellucidum displacement (>10 mm) constituted 0.97 ± 0.02 (Р ˂ 0.0001), which exceeded the values of FOUR Scale for the clinical detection of mild septum pellucidum displacement (1-5 mm) (AUC = 0.84 ± 0.05, P = 0.0158). AUC values were also high when FOUR Scale was used for the verification of the clinical signs of severe pineal gland displacement as well as those of mild pineal gland displacement. They constituted 0.99 ± 0.01 (Р ˂ 0.0001) and 0.92 ± 0.03 (Р ˂ 0.0001) respectively ( Table 3).
It was determined that the FOUR Scale was statistically more significant than the GCS as for the diagnostic informative value when used as a tool for the clinical detection of midline shift (AUC = 0.81 ± 0.03 versus AUC = 0.67 ± 0.04, P = 0.0002), as well as for the severity assessment of the septum pellucidum displacement <4 mm (AUC = 0.80 ± 0.04 versus AUC = 0.73 ± 0.05, P = 0.0286) and pineal gland displacement <4 mm (AUC = 0.80 ± 0.05 versus AUC = 0.74 ± 0.05, P = 0.0306). However, the aforementioned scales were compared on the basis of the discriminate potential of the midline shift >5 mm and >10 mm clinical signs detection.

Original research
Optimal cut-off values of the FOUR score and GCS score were determined based on a ROC analysis. They serve as clinical criteria of the midline shift presence, as well as criteria for the clinical assessment of septum pellucidum and pineal gland displacement severity (Tables 4  and 5).
The analysis results of the diagnostic accuracy of the FOUR Scale score when used as the criteria for the clinical assessment of midline shift severity are presented in Tables 6 and 7.
As the presented data show, the FOUR Scale has a high accuracy when used as the assessment tool for the severity of septum pellucidum (86.7-95.7 %) and pineal gland displacement (94.2-96.4 %) clinical signs. The FOUR Scale was as accurate as the GCS when used for the clinical signs of moderate and severe midline shift assessment. However, the FOUR Scale was more accurate than the Glasgow Coma Scale as for the clinical signs of midline shift detection   Based on a correlation analysis along with the calculation of Spearman's rank correlation coefficients (R) it was determined that the FOUR score was statistically more associated with septum pellucidum displacement (R 95 % CI = -0.65 (-0.54; -0.74), P ˂ 0.0001 versus -0.42 (-0.27; -0.54), P ˂ 0.0001 for GCS score, Z statistic = -2.69, P = 0.0071) and pineal gland displacement (R 95 % CI = -0.65 (-0.54; -0.71), P ˂ 0.0001 versus -0.45 (-0.31; -0.57), P ˂ 0.0001 for the GCS score, Z statistic = -2.39, P = 0.017).

Discussion
It was determined that AUC values for the FOUR Scale corresponded to "very high" and "excellent" gradations in accordance with the International Expert Scale of binary classification quality assessment M. H. Zweig, G. Campbell (1993) [17]. The obtained results showed a high diagnostic informative value of the FOUR Scale for the clinical signs of midline shift detection in patients with SSICH in the acute period, furthermore the AUC value was the highest when the FOUR Scale was used as a tool for the verification of clinical signs of severe septum pellucidum displacement (0.97 ± 0.02 versus 0.84 ± 0.05, P = 0.0158) and severe pineal gland displacement (0.99 ± 0.01 versus 0.92 ± 0.03, Р = 0.0269). Based on the ROC analysis optimal cut-off values of the FOUR Scale score were determined for the clinical assessment of midline shift severity.
The obtained data are consistent with the results of other studies, which showed, that the FOUR Scale was superior to the GCS in detecting risk of lethal outcome in critically ill patients [14] with acute ischemic stroke [9] and traumatic brain injury [8,11].
In our opinion, a higher diagnostic informative value of the FOUR Scale in comparison with the GCS was due, on the one hand, to the fact that the verbal reaction assessment in the structure of the FOUR Scale was absent, and the latter was influenced by the lateralization of the lesion in patients with supratentorial intracerebral hemorrhage. On the other hand, the spectrum of the assessed indexes had some additional criteria for the clinical assessment of the supratentorial structures functional state (respiratory pattern, brainstem reflexes). The study of Y. Hu et al. (2017) showed, that motor response and brainstem responses were verified as independent predictors of conscious awareness recovery in patients with acute ischemic stroke [7]. By the results of M. A. Gorji et al. (2015), the FOUR score was more accurate and practical in intubated patients regarding a lack of verbal response factor in early prediction of mortality in the GCS [6].
All of the above together with numerous studies convincingly proving a high inter-rater agreement of the FOUR Scale using [4,5,10,13,16] makes this scale more useful and preferable than the GCS for the assessment of midline shift clinical signs severity in patients in the acute period of SSICH.

Conclusions
1. The FOUR Scale is a highly informative tool for the clinical detection of midline shift presence and severity (AUC > 0.80, P < 0.0001) in patients with SSICH, whereas the diagnostic informative value of this scale within the assessment of severe midline shift clinical signs is higher than for the detection of mild midline shift (AUC = 0.97 ± 0.02 versus AUC = 0.84 ± 0.05 for septum pellucidum displacement, P = 0.0158; AUC = 0.99 ± 0.01 versus AUC = 0.92 ± 0.03 for pineal gland displacement, P = 0.0269).
The perspective for the further scientific research is the assessment of the FOUR score prognostic value in patients with SSICH.

Funding
The study is funded as a part of scientific research work at Zaporizhzhia State Medical University "Optimization of the diagnostic, treatment and rehabilitation approaches in patients with acute and chronic cerebral bloodflow violations", number of state registration 0113U000798 (2013-2017).