All patients (≥18 years) with a spontaneous SAH proven by computed tomography (CT), magnetic resonance imaging (MRI) or lumbar puncture will be considered for this trial. Upon presentation to a neurosurgical centre the patients will be treated according to the local protocol. Upon admission the patient is clinically evaluated for occurrence of clinical signs of brain herniation syndromes (anisocoria, bilateral dilated pupils, posturing). Usually first line treatment includes neurological resuscitation (placement external cerebrospinal fluid drainage in case of hydrocephalus, treatment of seizure, and general intensive care measures). Hereafter, the patient is clinically evaluated for a second time. The patients will be graded according to the usual WFNS scale and the modified "herniation WFNS" scale. The whole treatment of the patient will be according to local clinical protocols. Outcome will be measured at six and twelve months by trained investigators who are unaware of clinical data. The primary endpoint is the difference of specificities of the WFNS and hWFNS with respect to poor outcome (mRS 4-6) at 6 months after initial haemorrhage. Given that specificity and sensitivity are negatively correlated, difference in sensitivity will be the second primary outcome.

The null hypothesis to be tested is that the ratio of the true negative rates (specificity) of the hWFNS and WFNS scores is 1.35 i.e. the new score will detect 35% more patients as truly negative (good outcome) as compared to the old score. In addition and because of the negative correlation between specificity and sensitivity we will also test that the ratio of the true positive rate (sensitivity) is not below 0.82 i.e. the new score will not more than 18% less patients as truly positive (poor outcome).

Background

Aneurysmal subarachnoid hemorrhage (SAH) is a severe disease affecting approximately 400 patients per year in Switzerland. Death and severe disability after spontaneous SAH are related to the initial severity of the bleeding. The scale which is most often used and recommended to grade the severity of SAH is the World Federation of Neurosurgical Societies scale. Besides the Glasgow Coma Scale as basis for the WFNS grading, the existence of a neurologic deficit is included in this scale. In general WFNS I-III patients are termed good grade and WFNS IV and V poor grade. Morbidity and Mortality after SAH should increase with higher WFNS grade. Yet, despite low Glasgow Coma Scale (GCS) scores in poor grade SAH, 35 to 50% of aggressively treated patients show a favorable outcome.Even in the most severe SAH, i.e. WFNS grade V, a good outcome is observed in 24 to 50% of patients treated aggressively, which limits the usefulness of this scale for decisions such as whether to withhold or continue treatment. Especially the differentiation between WFNS grade IV and WFNS grade V in intubated patients is a huge problem. The reason for the discrepancy between poor grading and favorable outcome is due to the difficulty of reliable GCS assessment in the acute stage of SAH. In the acute stage the clinical status of the patient is distorted due to sedative medication, seizure and hydrocephalus and does not necessarily display the "real" clinical status. Therefore favorable outcome in poor grade SAH patients is not a matter of a very beneficial course, but rather of an initial wrong WFNS grading.

The necessity of a reliable WFNS grading becomes evident when clinical decisions are made on this basis or when scientific data of different studies has to be compared. With respect to clinical decisions - usually WFNS grade V patients are not being treated due to the limited capacity of surviving the SAH. Up to now there is no validated triage scheme to identify which SAH patients will most likely not benefit from aggressive treatment based on WFNS scores.

Because the GCS is the basis for WFNS grading, information about progressive herniation beyond the motor response and posturing - e.g. third nerve dysfunction or loss of brain stem reflexes - is unfortunately not incorporated in routine SAH grading. As a known consequence, sedated and ventilated patients who do not exhibit a motor response but have symmetrically reactive pupils and intact brain stem reflexes may be wrongly classified as having a GCS score of 3 and thus a WFNS grade V, which leads to inaccuracies in the prognostic power of the WFNS in poor grade patients.

In this study, the investigators will apply the concept following the logic of the original GCS design and the pathophysiological concept behind the scale. During progression of brain herniation, abnormal flexion (decorticate rigidity, late diencephalic syndrome) is followed by abnormal extensor posturing (decerebrate rigidity, mesencephalic syndrome). Both are positive signs, i.e. signs that can be observed by the examiner. If herniation further progresses, damage occurs to the medulla and typical signs are flaccidity, no response to pain, mydriasis and no reaction to light. Using the latter "positive" signs would help to verify a true GCS 3 patient.

The purpose of this study is to re-evaluate the prognostic power of the WFNS classification, particularly with regard to poor grade patients. The investigators compare the existing WFNS grading with a modified WFNS herniation grading model (hWFNS) with respect to predicting poor outcome and death.

Objective

The primary objective of this observational study is to test whether signs of herniation (dilated unilateral or bilateral pupils, posturing) can improve the predictive value of poor outcome at 6 months after SAH. Specifically, the study aims to show improved (superiority) specificity of the modified scale while maintaining (non-inferiority) sensitivity.

Secondary objectives are the impact of radiological factors (general/focal brain edema, intracerebral hematoma, hydrocephalus, amount of subarachnoid blood) as well as clinical factors (seizure, timing of grading, sedation) and their impact on herniation and grading of patients with SAH.

Methods

Patients with proven spontaneous SAH will be entered into the study. Upon presentation to a neurosurgical centre the patients will be treated according to the local protocol. The patient will be clinically assessed before and after neurosurgical resuscitation. Patients are screened for occurrence of clinical signs of brain herniation syndromes. The patients will be graded according to the usual WFNS scale and the modified "herniation WFNS" scale. Outcome will be measured at six and twelve months by trained investigators who are unaware of clinical data.

Inclusion Criteria:

• Written informed consent of the patient or consent of patient's next of kin
• Spontaneous SAH
• Age: ≥18
• Glasgow coma scale (GCS) ≤ 12. In intubated patients the GCS assessment will be performed after cessation of sedation or if not possible the last GCS score before intubation will be used

Exclusion Criteria

• SAH due to any other cause or structural abnormality of the brain (trauma, dissection, arterio-venous malformation, dural arterio-venous fistula)
• Foreseeable difficulties in follow-up due to geographic reasons (e.g., patients living abroad)

All patients (≥18 years) with a proven spontaneous SAH will be considered for this trial