In this work, we were able to create an EBI score comprising seven key variables, which allowed prediction of the neurological outcome of the patients after aSAH with very high accuracy. We named this EBI prediction score the Subarachnoid Hemorrhage Associated Early Brain Injury Outcome Prediction score (SHELTER-score), which consists of the following independent predictors: age, WFNS, anisocoria, early deterioration, CPR, midline shift, and early ischemia. The SHELTER-score has numerous implications for clinical decision-making and patient care, as it can help clinicians predict the course of the disease and determine appropriate treatment plans. Additionally, the score highlights the importance of EBI events in the pathophysiology of aSAH, emphasizing the need for further research in this area.

In recent years, there has been an increasing understanding of the pathophysiological processes after aSAH. Despite decades of research on cerebral vasospasm, the lack of development of successful treatment options has shifted the focus of research toward detecting early clinical events after subarachnoid hemorrhage to develop novel diagnostic, outcome, and therapeutic tools. The early pathophysiological processes in the phase of EBI have progressively come into focus, whereby a cascade of sterile inflammation following tissue injury after aSAH leads to cellular stress, apoptosis, blood–brain barrier disruption, and microvascular dysfunction, which in turn maintains and even drives an inflammatory response [5, 9, 22]. The SHELTER-score developed in this study therefore focuses on risk factors from the EBI phase and aims to predict the influence of EBI on the neurological outcome of patients with aSAH, independent of the phase of DCI. Through the very high predictive value and accuracy of our SHELTER-score, we see the immense importance of the tissue damage within the EBI for the overall outcome in patients with aSAH.

The parameters used in the SHELTER-score were selected based on the results of forward selection. Our study population showed that age strongly influences patient outcomes, which is consistent with previous studies [24,25,26,27]. Elderly patients are more likely to have preexisting conditions that can worsen the severity and duration of complications, leading to a significantly worse outcome [24,25,26,27]. To reflect this, our score divides age into five categories with ascending score values. Patients over the age of 80 are assigned the maximum score (4) in the entire score, emphasizing the critical role of age in determining outcomes.

Another well-established predictor of patient outcome after aSAH is the initial neurological condition of the patient, commonly assessed by using the WFNS grading system [28,29,30]. In our study population, the WFNS grade was found to be the parameter with the highest predictive power, and it was assigned a point value of 0.5 for each consecutive grade in the SHELTER-score.

We also included the EBI parameter of prehospital CPR in our score, as previous studies have shown that aneurysmal hemorrhage can be associated with cardiac arrest requiring CPR in the prehospital setting [31,32,33]. Our data demonstrate that patients who required CPR generally had worse outcomes than those who did not (Fig. 1d). This may be due, in part, to an initial massive brain damage and maximum intracranial pressure increase that can trigger cardiac arrest [31, 32, 34]. Furthermore, cardiac arrest in the context of aSAH can also mask the underlying bleeding problem and lead to a worse outcome by delaying adequate therapy. In our study population, the prognostic value of CPR was also evident, although only 2 points were attributed to this parameter after weighting in our SHELTER-score.

Another significant neurological symptom in clinical settings is the presence of dilated unilateral or bilateral pupils, which often indicates the need for immediate emergency surgery [35, 36]. Therefore, it is not surprising that unilateral or bilateral mydriasis is also an important predictor of patients’ outcomes in our patient cohort. In clinical practice, bilateral dilated pupils typically indicate the most severe brain or brainstem damage [37, 38]. Our SHELTER-score reflects this condition with an increased score of one point for anisocoria with mydriasis and two points for bilateral mydriasis. It is essential to note that the score takes into account not only the possible initial anisocoria at the time of ictus but also any anisocoria that may occur during the EBI phase.

Anisocoria is typically caused by a unilateral space-occupying process, primarily intraparenchymal hemorrhage or, more rarely, an acute subdural hematoma in the acute phase of aSAH [39, 40]. In the later course (secondarily), brain swelling due to hypoxia leading to severe edema is another reason for a space-occupying process [39]. These ultimately result in a displacement of the remaining intact brain tissue, visible in CT imaging as a midline shift. Therefore, there is a strong collinearity between these space-occupying effects and the occurrence of a midline shift. This collinearity between intraparenchymal blood volume in milliliters and midline shift in millimeters is also evident in our patient cohort (Supplementary Fig. 4), so we included only the stronger prognostic factor, the midline shift, in the SHELTER-score. The midline shift is not only the better prognostic factor but also easier and faster to measure than the amount of bleeding on CT. The increasing predictive value for poor patient outcome with increasing midline shift is reflected in the distinction between a midline shift greater than 10 mm (score: 0.5) and a midline shift greater than 20 mm (score: 1) in our SHELTER-score.

Another parameter that is important to consider when assessing the severity of EBI is early deterioration, which refers to a patient’s clinical decline within the first 72 h without any other identifiable cause [5, 9]. This deterioration is caused by multiple pathophysiological processes that lead to early brain damage, which in turn promotes further damage in the course of the disease [5, 9]. In line with the literature, our study found that early deterioration has a high prognostic value for patient outcome, which is why it was included in our SHELTER-score with a score value of 1.

Early deterioration reflects early brain damage, whereas early ischemia, as the most severe form, has already resulted in demarcated infarcted areas in the brain visible on CT imaging, as a manifestation of tissue hypoxia. The occurrence of early ischemia is heavily weighted in our SHELTER-score, with a score value of 2 given its high predictive accuracy in our study population.

In conclusion, by using the EBI variables, we developed a score with a very high predictive accuracy for patient outcome in the studied population of patients with aSAH. The strong correlation of the SHELTER-score with patient outcome highlights the critical nature of early-phase damage following subarachnoid hemorrhage for the overall disease course. Our SHELTER-score may prove valuable to clinicians in predicting the course of aSAH and making treatment decisions for affected patients with this life-threatening condition.

When interpreting these findings, it’s important to acknowledge that this is a single-center, retrospective study, and the resulting score needs to be replicated and validated in other populations to estimate the generalizability of the SHELTER-score. Furthermore, only the 6-month mRS was used as an outcome variable, and other important outcome parameters such as cognitive status or quality of life of the affected patients were not taken into account in the compilation of the SHELTER-score. Additionally, future studies with a longer follow-up duration would be valuable for providing an even more comprehensive prediction of long-term outcomes. Additionally, preexisting health conditions of the patients were not taken into account and our study examined a specific set of clinical and radiological variables, and there may be additional prognostic factors that were not considered during the development of the scoring system (e.g., lab values). Incorporating these factors could potentially enhance the performance of the score. However, it is crucial to strike a balance between practicality and comprehensiveness. In our study, we prioritized variables that could be readily recorded in a timely manner. This decision was made to ensure the score’s practicality and ease of use in the clinical setting.