Introduction: Seizures are a common complication of aneurysmal subarachnoid hemorrhage (aSAH) in both acute and late stages: 10-20% acute symptomatic seizures, 12-25% epilepsy rate at five years. Compared to other acute brain injuries, aSAH provides advantages as a model of acquired epilepsy: patients are younger, have fewer comorbidities and their rate of epilepsy is significantly higher than in other structural causes (ischemic stroke 6%, intracerebral hemorrhage 12%). Furthermore, there is a notable correlation between post-SAH seizures and significantly increased mortality and severe disability. Therefore, preventing such seizures might potentially improve the prognosis for patients. The epileptogenic process initiates immediately after brain injury and persists within a range from days to years until the first clinical manifestation appears. Although the latency provides an opportunity to implement disease-modifying therapies, this strategy requires an early and precise identification of patients at higher risk of developing epilepsy. In addition, this capability would be helpful to understand the underlying mechanisms of epileptogenesis, which could be targeted for preventive therapies.
Objectives: Our aim was to determine the rate of the development of epilepsy after aSAH and identify early clinical, electroencephalogram (EEG) and computed tomography (CT) findings that could predict long-term epilepsy. In addition, we aimed to design a predictive model combining those factors to identify the patients at higher risk of developing post-SAH epilepsy.
Methods: We conducted two multicenter, retrospective, longitudinal studies involving adult patients with aneurysmal SAH. We collected and evaluated clinical characteristics (n=419), paraclinical findings (EEG and brain CT, n=278), as well as the rate of early-onset seizures (≤7 days from SAH onset), and development of epilepsy. Post-SAH epilepsy was defined as the occurrence of remote unprovoked seizures >7 days from the bleeding. Multiple Cox regression models were used to identify the variables independently associated with epilepsy development. The best-fitting regression model was used to design a predictive score, whose performance was evaluated in an external validation cohort of 308 patients using ROC curve analysis.
Results: In our sample, 50 patients (11.9%) developed post-SAH epilepsy. Higher premorbid modified Rankin Score (mRS) (HR 4.74 [1.8-12.4], p=0.001), Vasograde score (HR 2.45 [1.4-4.2], p=0.001), surgical treatment (HR 2.77 [1.6-4.9], p=0.001), and the presence of early-onset seizures (EOS) (HR 1.84 [1.0-3.4], p=0.05) were the clinical factors independently associated with the development of epilepsy. Multiple Cox regression analysis also showed a higher risk of long-term epilepsy in patients with cortical lesions on brain CT (HR 2.6 [1.3-5.2], p 0.009), focal non-epileptiform EEG abnormalities (HR 3.7 [1.6-8.2], p 0.002), and focal epileptiform discharges (HR 6.7 [2.8-15.8], p <0.001). The proposed prediction scale (RISE) scored patients based on the previously identified clinical predictors, with a score ranging from 0 to 5 points, and stratified them into three groups: low (0-1), moderate (2-3), and high (4-5) risk (2.9%, 20.8%, and 75.7% remote epilepsy rate, respectively). In the validation cohort, the proposed scale yielded a similar risk distribution and good predictive power for epilepsy development (AUC 0.82; 95%CI 0.74-0.90).
Conclusions: The independent clinical predictors for the development of post-SAH epilepsy were higher premorbid mRS, higher Vasograde score, surgical treatment, and the presence of early-onset seizures. The paraclinical predictors were the presence of cortical involvement on cranial CT and focal EEG abnormalities. Combining these variables in a prediction scale (RISE) allowed to stratify patients according to their risk of epilepsy with high reliability (AUC >80%).
| Date of Award | 25 Oct 2024 |
|---|
| Original language | Spanish |
|---|
| Supervisor | Estevo Santamarina Perez (Director) & Manuel Toledo Argany (Director) |
|---|