Comparison of Full Outline of UnResponsiveness (FOUR) score and the conventional scores in predicting outcome in aneurysmal subarachnoid haemorrhage patients.

BACKGROUND AND AIMS: Full Outline of UnResponsiveness (FOUR) score is a more comprehensive score used to assess eye response, motor response, brainstem reflexes, and respiration that was introduced to overcome the drawbacks of Glasgow coma scale (GCS) score. Our aim was to assess which score best predicts mortality and poor outcome in aneurysmal subarachnoid haemorrhage (aSAH) patients.
METHODS: This cohort study, prospectively evaluated the use of FOUR score to assess the mortality and outcome in aSAH patients during the period from November 2015 to November 2016. For each patient of aSAH, GCS, FOUR score, Hunt and Hess (HH) score and World Federation of Neurological Surgeons (WFNS) score were determined at the time of admission to neurosurgical intensive care unit. All patients were followed till 28 days post-SAH and their outcome were assessed by Glasgow outcome scale (GOS). We calculated the sensitivity (Sn) and specificity (Sp) for each of these scores. We generated the receiver operating characteristic curve (ROC), quantified the accuracy by the area under curve (AUC), and also calculated their 95% confidence interval (95% CI).
RESULTS: A total of 75 aSAH patients were enrolled for the study. The mortality was 24/75 (32%) with 23 in-hospital deaths. FOUR score was highly specific (86.27%) and sensitive (75%) for the prediction of mortality. However, for predicting 28-day outcome, WFNS and HH grade were most specific (92.5%), whereas FOUR and HH score was moderately specific (68.57%).
CONCLUSION: FOUR score is among the most specific and moderately sensitive tool for prediction of mortality. However, WFNS and HH grade are more specific in predicting the 28-day outcome.

INTRODUCTION

Aneurysmal subarachnoid haemorrhage (aSAH) is a common cause of stroke which is associated with significant mortality and morbidity. Common scales such as Hunt and Hess (HH), World Federation of Neurological Surgeons (WFNS) are used for grading the severity of SAH and are useful indicators of outcome.[12] Most scales used to assess prognosis after SAH are based on the level of consciousness of the patient.

The most common clinical tool for determining the severity of head injury and predicting mortality and morbidity is the Glasgow coma score (GCS). GCS is used for assessing consciousness in all populations.[3] It has been a gold standard against which newer scales are compared. Over the years, its shortcomings like unreliability of verbal response in an intubated patient and absence of brainstem reflexes tests have come to focus. Full Outline of UnResponsiveness (FOUR) score was introduced to overcome the drawbacks of GCS. It is a more comprehensive score used to assess eye response, motor response, brainstem reflexes, and respiration. Each component carries equal points (0--4), making it a 17 point scale (0--16). The FOUR score has been validated in several clinical contexts, including assessment of neurocritical care unit patients, medical intensive care unit (ICU) patients, patients in the emergency department, and neurosurgical patients.[4567] Regarding interobserver reliability of the FOUR score, it has been seen that it may have a modest but definite advantage in this particular function when compared with GCS.[7] A validation study carried out in neurosurgical patients has shown FOUR score to be more robust than GCS in predicting mortality in patients with severely impaired consciousness.[7] In another recently published study, authors determined the predictive value of FOUR score for 1- and 6-month mortality post-SAH. They concluded that FOUR score at admission and day 7 post-SAH is associated with mortality, 1 and 6-month Glasgow outcome score (GOS)/modified Rankin scale.[8]

Thus, FOUR score, HH, and WFNS score all help in prognostication of patients with impaired consciousness. We hypothesised that the predictive power of FOUR score will be similar to that of conventional scores such as HH and WFNS score. We carried out this study with an aim to compare FOUR scale, HH, and WFNS scores for predicting mortality and outcome in patients having aSAH. Primary outcome was to assess which score best predicts mortality in these patients. Secondary outcome was to compare these scores for assessment of poor neurological outcome as determined by GOS at 28-days post-SAH.

METHODS

This study was approved from the institutional ethics committee and informed written consent was obtained from the patients or from next of kin if GCS <15Z. We followed the STROBE statement checklist for this observational study. This cohort study prospectively evaluated the patients suffering from aSAH admitted to our neurosurgical ICU within 3 days of ictus, over a period from November 2015 to November 2016 (1 year). In all patients, FOUR score and HH score were determined by same examiner who was well trained in assessment, within a span of 5 min. Sequence of scores to be assessed were randomised based on the computer generated randomisation chart. Scoring was done as early as possible within the first 24 h of admission to the Neuro-ICU. Patients were excluded if they were heavily sedated precluding the examiner from obtaining FOUR or GCS scores or if they had history of any psychiatric illness, psychological problem, alcoholism, and drug addiction. Patients having non-aneurysmal SAH were also excluded. Other details such as radiological findings, treatment received, and length of ICU and hospital stay were also noted.

Condition at discharge was assessed and any in-hospital mortality was recorded. At 28-days post-SAH, each patient was assessed for neurologic outcome (i.e., GOS) either by visiting in-patients or through a telephonic survey by an investigator, who was blinded to the patient's data and scores. Functional outcome was dichotomised, poor versus good, based on GOS at 28 days. Poor outcome was defined as GOS 1-3.

Statistical analysis was performed using STATA 12.0 (Stata Corp, College Station, TX). Data are presented as mean (SD) or number (%). For every possible cut-off point, sensitivity and specificity were calculated and the coordinates of receiver operating characteristic curves (ROC) were generated. Accuracy of the scales was quantified by the area under the ROC curve (AUC). An AUC of 1.0 refers to a perfect test, whereas a perfectly inaccurate test has an AUC of 0.0. An area of 0.5 represents the chance diagonal, consequently a worthless test. The relation of the FOUR score, HH score, and WFNS score to mortality (GOS-1), poor outcome (GOS <4), and favorable outcome (GOS ≥4) at 28 days were estimated. A P value of <0.05 was considered to be significant.

RESULTS

A total of 89 patients were admitted to neuro-ICU with SAH. Fourteen patients out of 89 were excluded (11 patients admitted more than 3 days post-SAH, one patient was under sedation precluding the assessment of score, and two patients had non-aSAH). Seventy-five patients were finally enrolled and followed till 28 days post-SAH [Figure 1]. None was lost to follow-up. The patient's demographics are shown in Table 1. The mean age was 52.2 ± 14.8 years. The good grade patients as per HH score (HH I--III) were 49/75 (65.3%) and according to WFNS score were 52/75 (69.3%). Total number of aneurysms were 88 in our population of which 77/88 (87.5%) were present in anterior circulation and rest 11/88 (12.5%) were involving posterior circulation. A total of 47 (62.7%) patients underwent surgical clipping to secure the aneurysm and 6 (8%) patients were managed conservatively. Rest of the patients (n = 22, 29.3%) underwent endovascular procedures to secure the aneurysm. Intraoperatively, two patients suffered aneurysm rupture during dissection. The median duration of mechanical ventilation and the length of ICU stay was 5.5 (0--53) and 8 (2--53) days, respectively. The median length of hospital stay was 11 (2--53) days. The mortality was 24/75 (32%) with 23 in-hospital deaths.

Figure 1

Consort Flow Diagram

Table 1

Baseline Characteristics and Clinical Data

Parameter	(Mean±SD)/n (%)
Age (years)	52±14.80
GCS	12.57±3.5
HH grade	3.2±1.16
WFNS	2.14±1.5
FOUR Score	13.6±3.61
Sex
Male	32 (57.3)
Female	43 (42.7)
Location of aneurysm
Anterior Communicating artery	30 (34)
Internal Carotid artery	12 (13)
Middle cerebral artery	12 (13)
Anterior Choroidal artery	9 (10)
Anterior cerebral artery	3 (3)
Superior Hypophyseal artery	1 (1)
Posterior communicating artery	10 (11)
Posterior cerebral artery	3 (3)
Superior cerebellar artery	3 (3)
Basilar top aneurysm	3 (3)
Posterior inferior cerebellar artery	3 (3)

GCS – Glasgow Coma Scale; HH Grade – Hunt and Hess Grade; WFNS – World Federation of Neurological surgeons; FOUR – Full Outline of UnResponsiveness

The specificity and sensitivity of various scores for prediction of mortality and 28-day outcome are shown in Table 2. Figures 2 and 3 depict the area under ROC (AUC--ROC) for predictive power of mortality and 28-day outcome.

Table 2

Correlation of the scales for mortality and GOS at 28 days

Outcome	TP	FP	TN	FN	Sensitivity	Specificity	AUC (95% CI)
Mortality
FOUR	18	7	44	6	75%	86.27%	0.80 (0.68-0.91)
HH grade	17	10	41	7	70.83%	80.39%	0.79 (0.67-0.90)
WFNS	17	7	44	7	70.83%	86.27%	0.76 (0.64-0.88)
GCS	18	14	37	6	75%	72.55%	0.78 (0.67-0.90)
GOS AT 28 days
FOUR	24	6	34	11	68.57%	85%	0.81 (0.72-0.90)
HH grade	24	3	37	11	68.57%	92.50%	0.82 (0.71-0.91)
WFNS	21	3	37	14	60.00%	92.50%	0.78 (0.68-0.88)
GCS	25	7	33	10	71.43%	82.50%	0.81 (0.72-0.90)

FOUR – Full Outline of UnResponsiveness; HH Grade – Hunt and Hess Grade; WFNS – World Federation of Neurological Surgeons; GCS – Glasgow Coma Scale; GOS – Glasgow Outcome Score; TP – True Positives; FP-False Positives; TN – True Negatives; FN – False Negatives; AUC – Area Under Curve; CI – Confidence Interval

Figure 2

ROC of all the scores for mortality. FOUR- Full Outline of UnResponsiveness, HH Grade- Hunt and Hess Grade, WFNS- World Federation of Neurological Surgeons, GCS- Glasgow Coma Scale, ROC- Receiver Operating Characteristic curve

Figure 3

ROC of all the scores for 28-day outcome. FOUR- Full Outline of UnResponsiveness, HH Grade- Hunt and Hess Grade, WFNS- World Federation of Neurological Surgeons, GCS- Glasgow Coma Scale, ROC- Receiver Operating Characteristic curve

In terms of the predictive power for mortality [Table 2 and Figure 2], the area under the ROC (AUC--ROC) was 0.80 (95% CI: 0.68--0.91) for the FOUR score, 0.78 (95% CI: 0.67--0.9) for the GCS score, 0.79 (95% CI: 0.67--0.9 for HH, and 0.76 (95% CI: 0.64--0.88) for WFNS. For mortality, the ROC curves of GCS and FOUR score were comparable (P = 0.61).

In terms of the predictive power for 28-day outcome [Table 2 and Figure 3], the AUC--ROC was 0.81 (95% CI: 0.72--0.9) for FOUR score, 0.81 (95% CI: 0.72--0.9) for GCS score, 0.82 (95% CI: 0.71--0.91) for HH, and 0.78 (95% CI: 0.68--0.88) for WFNS. For neurological outcome at 28 days post-SAH, the ROC curves of GCS and FOUR score were comparable (P = 0.98).

DISCUSSION

The FOUR score is a neurological assessment tool and a means of categorising aSAH patients in greater detail. This study was carried out with an intent to find out which scores best predicts the mortality and 28-day outcome in aSAH patients. The findings of our study are as follows: FOUR score was more specific for prediction of mortality (86.2%) and 28-day neurological outcome (85%) when compared with GCS. The SAH scale such as WFNS had equal specificity as FOUR score for prediction of mortality (86.27%), whereas both HH and WFNS are most specific (92.5%) scores for assessing 28-day outcome. However, both these scores fair less in sensitivity (70.83% each) for predicting mortality. For mortality and 28-day outcome, the predictive power of all these tests is comparable.

The FOUR score was more specific score for prediction of mortality as compared with GCS score (86.27% vs 72.55%). This could be attributable to its comprehensive testing configuration. However, both of them were equally sensitive for predicting mortality. As per our study results, in terms of the predictive power for mortality from AUC—ROC, we did not observe any difference between FOUR Score and GCS. Our results correlate with a recently conducted study, where authors concluded FOUR is comparable to GCS in terms of predictive ability for functional status, cognitive outcome at 90 days post-injury, and in-hospital mortality in traumatic brain injury patients.[9] The FOUR score is more specific tool in comparison to GCS in determining 28-day outcome. Sensitivity of GCS was higher compared with FOUR in determining 28-day outcome. Similar results were seen in another study where the AUC--ROC was 0.75 for FOUR score and 0.76 for GCS score as a predictor for long-term morbidity and the authors concluded that FOUR score is a good predictor of prognosis of critically ill patients.[4]

Zeiler et al. has recently concluded that FOUR score at admission and day 7 post-SAH is associated with mortality, 1-month GOS/mRS, and 6-monthGOS/mRS.[7] They carried out an analysis of components of FOUR score and concluded that the brainstem sub-score was not associated with 1- or 6-month primary outcomes. In our study, we conducted the analysis for all the conventional scores for our primary outcome of prediction of mortality and found that FOUR score is a sensitive tool for assessing mortality and 1-month outcome but when compared with other scores, these all have comparable predictive power.

However, there are certain limitations in our study. This study was carried out at a single centre and further prospective multicentric trials may hence enable us to achieve a larger sample size. We did not assess the scores for prediction of long-term outcome which can further strengthen the clinical utility of FOUR score.

CONCLUSION

The FOUR score provides a more comprehensive neurological assessment and is a comparable predictor of mortality in aSAH patients, when compared with other conventional scores. Furthermore, FOUR score is among the most specific and sensitive tool for prediction of mortality. However, WFNS and HH grade are more specific in predicting the 28-day outcome.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.