ABSTRACT
Objectives: Status epilepticus (SE) is a serious neurological condition with a high risk of complications and mortality. We described adult patients who required admission to the intensive care unit (ICU) for SE and assessed their prognosis.
Methods: We retrospectively analyzed consecutive adult ICU admissions for SE between 2016 and 2021. We performed multivariable logistic regression to determine the predictors of an unfavorable outcome (discharge Glasgow Coma Scale <14 or death).
Results: We studied 115 patients (median age 38 years; males 62.6%; prior antiepileptic therapy 59.1%; median presenting Glasgow Coma Scale 7; median Status Epilepticus Severity Score 4). Brain computed tomography revealed acute/sub-acute stroke in 5.2% of cases, intracranial hemorrhage in 2.6%, and brain tumors in 5.2%. Cerebrospinal fluid analysis in 41 patients detected bacteria or herpes simplex virus in 4. Intravenous lorazepam was used to abort seizures in 80.0% of patients (median dose 3 mg). Mechanical ventilation was needed for 87 patients for a median of 4 days. Thirty-five patients (30.4%) experienced an unfavorable outcome, and hospital mortality was 5.2%. Infections present on admission were associated with an unfavorable outcome (odds ratio 3.61, 95% confidence interval 1.35-9.63).
Conclusion: Most ICU admissions for SE involved young patients with few having abnormalities on brain imaging and cerebrospinal fluid and most needing intubation. Hospital mortality was 5.2% with the presence of an infection on admission linked to an unfavorable outcome.
Status Epilepticus (SE) is a life-threatening neurological crisis that requires urgent medical attention.1,2 Its definition has evolved over time as our understanding of SE grew.1-3 Currently, SE is defined as a continuous seizure lasting more than 5 minutes or multiple episodes of seizures within 5 minutes without full recovery of consciousness between episodes (10 minutes for nonconvulsive or focal SE).4 Although convulsive SE is more common, nonconvulsive SE may follow if convulsive SE remains uncontrolled.3 The crude incidence rate of SE in population-based studies was estimated at approximately 13 per 100,000 people annually (95% confidence interval [CI] 10-15) with variation among countries, likely due to differences in awareness, diagnostic capacity and medical resources.5
Current clinical practice guidelines consistently recommend benzodiazepines (intramuscular midazolam, intravenous lorazepam, or intravenous diazepam) for as first-line therapy of SE.1,2 Evidence indicates better seizure control when using benzodiazepines compared with other antiepileptic drugs and when treatment is instituted earlier.1,2,6 When seizures persist despite initial therapy, multiple second-line drugs can be used.1,2 A meta-analysis suggested that valproate, levetiracetam and phenobarbital are reasonable options and may be more effective than phenytoin.7 However, a subsequent randomized controlled trial found similar efficacy among valproate, levetiracetam, and fosphenytoin in controlling SE and improving consciousness within 60 min.8 Despite the availability of evidence-based guidelines, adherence remains suboptimal. A systematic review of 22 studies that appraised adherence to SE guidelines showed that the rates of deviation from SE guidelines ranged from 10.7% to 66.1%.9 The most common forms of deviations were excessive usage of benzodiazepines and delay in therapy initiation, both were associated with poor outcomes.9 The excessive usage of benzodiazepine was associated with serious adverse events including respiratory depression, intubation, and intensive care unit (ICU) admission.9 The delay in the treatment initiation was associated with a longer seizure duration and requirement for ICU admission.9 Among patients with SE requiring ICU admission, complications include altered consciousness, delirium, mechanical ventilation, hemodynamic instability, and, in some cases, death.10 The overall mortality for convulsive SE in adults remains high at approximately 16%, with little change observed between 1990 and 2017.11
There is a paucity of studies on SE in adult populations within Saudi Arabia. In this study, we aimed to describe the clinical characteristics of adult patients who were admitted to a tertiary-care ICU for SE, evaluate the used management approaches and assess their outcomes.
Methods
Study design and setting
This retrospective study was conducted in the Intensive Care Department at King Abdulaziz Medical City, a tertiary-care referral medical center in Riyadh, Saudi Arabia. The department had 104 beds in 8 different ICUs and employed a multidisciplinary approach for patient care. Patients in these ICUs received continuous care, 24/7, from board-certified critical care physicians in collaboration with specialists from various fields, including neurologists. Patients admitted due to neurologic/neurosurgical conditions were prioritized for admission to an 8-bed neuro ICU. Given the nature of the study, the Institutional Review Board of the Ministry of National Guard - Health Affairs granted approval and waived the requirement for informed patient consent (IRB number: IRB/0846/22).
Study patients
This study included all consecutive patients who were 14 years and older and admitted to the ICU for SE between January 1, 2016, and December 31, 2021. In accordance with the hospital’s administrative rules, any patient aged 14 years or older was admitted to the adult ICU.12 The operational definition of SE for our study was a seizure lasting 5 minutes or longer, or shorter consecutive seizures without recovery to baseline consciousness in between seizures.4 For patients with multiple ICU admissions during a single hospitalization, only the first SE-related ICU admission was analyzed. We excluded patients with post-cardiac arrest SE (due to uniformly poor prognosis), those with SE during pregnancy (part of preeclampsia/eclampsia syndrome) and patients transferred to the ICU from other hospitals with a diagnosis of SE (as initial management details were unavailable).
Data collection
Data were retrieved from the hospital’s electronic medical records (BestCare) by identifying patients assigned epilepsy-related codes based on the International Statistical Classification of Diseases and Related Health Problems: Tenth Revision Australian Modification (ICD-10 AM). The medical records of all identified patients were reviewed for inclusion and exclusion criteria. For eligible patients, we obtained their baseline demographic and clinical data that included chronic comorbidities, presence of metabolic/genetic diseases, SE semiology, level of consciousness at the time of presentation to the hospital as assessed by the Glasgow Coma Scale (GCS), and Status Epilepticus Severity Score (STESS). The STESS is used to predict outcomes and treatment response through four clinical predictors: age, seizure type, level of consciousness, and prior history of seizure.13 Additionally, we noted the findings from brain computerized tomography (CT) scans and electroencephalograms, and relevant laboratory results including those of cerebrospinal fluid (CSF). We also noted if an infection was clinically suspected on admission and whether it was subsequently confirmed. We also collected management-related data including the administration of antiepileptic drugs, antibiotics, antivirals, and vasopressors, intubation, and renal replacement therapy. We evaluated the following outcome variables: death in the ICU and hospital, GCS at discharge from ICU and hospital, occurrence of delirium in the ICU, performance of tracheostomy, duration of mechanical ventilation, and length of stay in the ICU and hospital. In this study, the primary endpoint (unfavorable outcome) was a composite bivariate variable, defined as a discharge GCS <14 or death in the hospital.
Data analysis
We presented categorical data as frequencies with corresponding percentages and continuous data as medians with interquartile ranges (IQRs). We employed the chi-square test or Fisher’s exact test, as appropriate, to compare the primary outcome across age groups (14-17 years, 18-49 years, 50-64 years, and ≥ 65 years) and periods (pre-COVID-19 [before March 2020], during the curfew for COVID-19 [March to June 2020], and after the curfew [July 2020 and onwards]). To identify the predictors of an unfavorable outcome, we performed multivariable logistic regression analysis. We entered in the model clinically relevant variables, including baseline characteristics (age, hypertension, diabetes, chronic kidney disease, old stroke, initial GCS, and STESS), brain CT findings, presence of confirmed infection on admission, vasopressor therapy and mechanical ventilation, as independent variables. As STESS incorporates age for its score calculation, we checked for collinearity assumption and did not find a significant correlation. We presented the results of the regression model using odds ratios (ORs) with the corresponding 95% CIs. We used Statistical Package for Social Sciences (SPSS) version 23 software (IBM Corp., Armonk, New York) for data analysis. The level of significance α was two-tailed and was set at 0.05 for all tests.
Results
Characteristics of SE patients
Between January 2016 and December 2021 (72 months), 115 patients required admission to the ICU for SE. They had a median age of 38 years (IQR 26-59), with 54.8% of patients being 19 to 49 years old and 80.0% being younger than 65 years. Most patients were males (62.6%). Hypertension, diabetes, and prior stroke were present in 33.0%, 29.6%, and 18.3% of patients, respectively (Table 1). Seventy patients (62.5%) had a prior history of seizures with 68 (59.1%) on anti-epileptic therapy. Three patients had Lennox-Gastaut syndrome. Most cases (n=93; 80.9%) presented to the emergency department before ICU admission (Table 1). With March 2020 being the cutoff date, 80 patients (69.6%) were admitted before the start of COVID-19 pandemic and 35 patients (30.4%) after. Five patients (4.3%) were admitted during the 3-month city-wide curfew for COVID-19 with an SE admission rate of 1.7 SE cases per month compared with 1.3 cases per month in the prior 51 month-period (Table 1).
- Baseline characteristic of the study patients.
On presentation, most patients (63.3%) were stuporous or comatose, while 40 patients (36.7%) were either alert or somnolent/confused. The median presenting GCS score was 7 (IQR 3-10), and the lowest recorded GCS score within 6 hours of presentation was 3 (IQR 3-7). The predominant seizure type was generalized-convulsive seizures (n=90; 85.7%). Nonconvulsive SE was rare (n=2; 1.9%). The median STESS score was 4.
Investigations
Metabolic derangements on admission included hypoglycemia (below 3.9 mmol/L) in 3 patients (2.6%), hyponatremia (below 120 mmol/L) in only 1 patient (0.9%), hypocalcemia (below 1.9 mmol/L) in 20 patients (17.4%), hypomagnesemia (below 0.65 mmol/L) in 7 patients (6.1%), and hypophosphatemia (below 0.6 mmol/L) in 4 patients (3.5%). Infection was clinically suspected upon presentation in 53 patients (46.1%) and was confirmed in 39 (33.9%). The source of infection was the central nervous system in 23 patients (20%), and the respiratory tract in 23 (20%). Lumbar puncture for CSF evaluation was performed in 41 patients (35.7%). A high white blood cell count in CSF (> 5 cells/mm3) was present in 8 of these 41 patients (23.5%), and a high protein level (>0.40 g/L) was present in 22 patients (59.5%). Two patients (8.7%) had positive serology for herpes simplex virus. Three patients (2.6%) tested positive for COVID-19 by a respiratory reverse transcription polymerase chain reaction.
Electroencephalography was performed for 70 out of 115 patients (60.9%). The most common finding was the presence of generalized slowing suggesting encephalopathy (n=50). Epileptiform discharges were noted in only 8 patients (6 focal and 2 generalized). The electroencephalogram was normal in 9 patients and was unsatisfactory due to artifacts in 2 patients.
Management in the ICU
Lorazepam was the most commonly used first-line agent to abort SE (n=92; 80%) with a median dose of 3 mg (IQR 2-6). Levetiracetam (n=62; 53.9%) and phenytoin (n=35; 30.4%) were the most commonly used second line agents (Figure 1). Intubation and invasive mechanical ventilation were needed for 87 patients. For seizure control, midazolam infusion was used in 25 patients (21.7%) and propofol infusion in 22 patients (19.1%). Barbiturate infusion was used in only 3 patients (2.6%). Levetiracetam was continued as a maintenance treatment to approximately two-thirds of patients (n=67; 58.3%), and phenytoin for one third (n=38; 33%). In the first 3 days of admission, vasopressors were administered to 32 patients (27.8%). Renal replacement therapy was provided in 22 patients (19.1%) during the ICU stay.
- Antiepiletic drugs used during status epilepticus (abortive) and after seizure control.
Outcomes
Table 2 summarizes the outcomes of the study patients. Death in the hospital occurred in 6 patients (5.2%). Among hospital survivors, the median GCS score at discharge was 15 (IQR 9-15). An unfavorable outcome (death in hospital or live discharge with a GCS score < 14) occurred in more than one-third of patients (n=35; 30.4%). An unfavorable outcome was observed in only 3 patients aged 14-18 years (27.3%), 16 patients aged 19-49 years (25.4%), 6 patients aged 50-64 years (31.6%), and 10 patients aged ≥ 65 years (45.5%) (between-group p=0.37). An unfavorable outcome was also observed in only 2/8 patients with high white blood cell count in CSF (25%) and 8/22 patients with high protein in CSF (36.4%). The rates of an unfavorable outcome were similar before, during and after the COVID-19 curfew (p=0.71). The ICU mortality was 3.5%. The median GCS score at ICU discharge was 14 (IQR 9-15). During ICU stay, most patients had delirium (101/115, 87.8%). A tracheostomy was required for 21 patients (18.3%).
- Outcomes of the study patients.
On multivariable logistic regression analysis (Table 3), the presence of a confirmed infection (OR 3.61, 95% CI 1.35, 9.63) and chronic kidney disease (OR 7.22, 95% CI 1.00-57.28) were associated with an unfavorable outcome. STESS (OR 1.45, 95% CI 0.96-2.21; p=0.08) and mechanical ventilation (OR 2.50, 95% CI 0.89-6.99; p=0.08) were not statistically significant.
- Multivariable logistic regression analysis for the predictors of an unfavorable outcome (Glasgow Coma Scale score < 14 or death in hospital) in the study patients (N=115).
Discussion
In this study of 115 adult patients who required ICU admission for SE, most patients were young aged 19 to 49 years. Most had a prior history of seizures suggesting that uncontrolled epilepsy was a common etiology, a finding consistent with other studies.14 Hypertension, diabetes mellitus, and old stroke were the most reported comorbidities. Stroke, in particular, is known to increase the risk of seizures and is probably the leading of SE in the elderly.15 Early seizures affect approximately 3% of patients after ischemic strokes, and late seizures occur at an annual rate of 1.8%.16 In our cohort, a minority of patients had abnormalities in brain imaging and/or CSF. We also did not observe significant changes in hospital admissions related to SE during the COVID-19 pandemic in Riyadh, even though emergency room visits declined in many countries during that period.17 A similar pattern of no change in hospital admission for SE was observed in a population-based study in Salzburg.18
Timely seizure control, ideally within the first 1–2 hours after SE onset, is important to avoid systemic complications.19 Treatment is more likely to work the earlier it is given.20 In the current study, intravenous lorazepam was used as the first-line agent for aborting SE in most patients (80.0%). Benzodiazepines are recommended as first-line drugs for SE management.1,2,21 The efficacy of intravenous lorazepam is probably comparable to that of other benzodiazepines. However, intravenous lorazepam may offer potential advantages, such as its longer duration of action, faster termination of SE, and requirement of fewer doses and fewer other antiepileptic drugs to reach an abrupt abortion of SE.20 Medications other than benzodiazepines have been evaluated as first-line drugs. For instance, an open pilot study that evaluated intravenous lorazepam (0.1 mg/kg) versus intravenous levetiracetam (20 mg/kg) found comparable seizure control (75.6% and 76.3% of patients, respectively) but a significantly higher need for intubation in the lorazepam group (10/21 patients versus 4/23 patients, p=0.03).22 Although valproate and levetiracetam may be safe and effective as first-line drugs for SE, evidence is inadequate to recommend their routine use.20 After intravenous lorazepam monotherapy, almost one in three patients fail to respond.1 These patients are treated with second-line antiepileptic drugs. Intravenous phenytoin, and alternatively valproic acid, levetiracetam, and phenobarbital are second-line anti-epileptic drugs. In a randomized controlled trial involving 384 patients with benzodiazepine-refractory SE, levetiracetam, valproate, and fosphenytoin demonstrated comparable efficacy.8 Nearly 50% of patients experienced seizure control and enhanced alertness within 60 minutes with comparable adverse events across the three treatment groups.8 However, a recent systematic review of 11 trials (1933 patients with SE) reported more clinical seizure control and lower risk of acute respiratory depression and life-threatening hypotension with levetiracetam compared with phenytoin.23 In our study, levetiracetam was the most frequent second-line drug and the most common epilepsy maintenance treatment, which reflects the changes in evidence and practice.24
Complications are common in patients with SE.10,14 In the current study, almost three in four patients required intubation, and one in four received vasopressor therapy. Indications for endotracheal intubation in SE include the requirement for general anesthesia (management stage 3 [midazolam and/or propofol infusion] and stage 4 [pentobarbital infusion]),1 and acute respiratory failure. Contrary to assumptions, respiratory depression was less common in patients with convulsive SE treated with benzodiazepines versus placebo in one review, indicating that undertreatment of convulsive SE poses a greater risk.21 A secondary analysis of a clinical trial evaluating intravenous lorazepam versus intramuscular midazolam for prehospital SE found that 218 out of 1023 patients (21.3%) required intubation.25 In a Swiss study exploring SE complications, 42% of patients received mechanical ventilation and 20% vasopressors.10
In the current study, most patients had good recovery, with a median GCS of 14 at ICU discharge and 15 at hospital discharge. This is consistent with earlier observations, including one study where 75.2% of the study patients improved by the time of discharge.14 The mortality rate in our study was 5.2%, which is notably lower than the 12.1% to 38% range reported in previous studies.10,26 The lower mortality in our study may be attributable to the relatively younger age of our cohort and the exclusion of post-cardiac arrest SE, which usually carries a poor prognosis.27 It may also reflect the improvements in intensive care in general and in SE management in particular and the quality of delivered care in the study hospital.
Although prompt seizure control improves outcomes,20 SE prognosis is largely influenced by nonmodifiable factors, including older age, comorbidities, and SE etiology.20 We found that the presence of an infection at presentation was a predictor of an unfavorable outcome. Sepsis is a risk factor for SE,2 and sepsis with SE carries a higher mortality risk than sepsis alone.28 Timely and appropriate treatment of sepsis associated with SE is important and may improve outcomes. The current study also illustrated that chronic kidney disease was a predictor of an unfavorable outcome. Another study similarly observed that chronic comorbidities —namely hypertension, diabetes, and chronic kidney disease—were linked to poor outcomes.14 Chronic kidney disease may complicate the care in the ICU due to challenges in volume management, metabolic homeostasis and drug dosing.29 This may explain the worse outcomes in this population. We found no significant association between outcome and age group, STESS score, or mechanical ventilation. Although the STESS score is a validated prognostic tool,13 its predictive value in our study was limited (p=0.08). This suggests that further validation is needed before STESS can be reliably used to guide SE management.1 Similarly, although another study reported prolonged ventilation and vasopressor therapy as independent predictors of poor outcomes,10 we did not find significant associations with these variables. The COVID-19 pandemic also did not appear to impact SE-related hospitalization rate in our study.
This study has both strengths and limitations. The key strengths include comprehensive data collection and assessment of patient-centered outcomes. However, as a retrospective, single-center study, it is subject to potential selection bias and limited generalizability. Patients were identified using hospital coding, which may have led to under-ascertainment. Despite these limitations, the findings provide valuable insight into SE management and outcomes.
Conclusions
In summary, the majority of ICU admissions for SE involved relatively young patients, and only a minority exhibited abnormalities on brain imaging or CSF analysis. Electrolyte abnormalities at baseline were common. The most commonly used abortive medication was lorazepam followed by levetiracetam. Although most patients showed clinical improvement, approximately one-third experienced an unfavorable outcome, one-fifth required tracheostomy, and 5% died during hospital stay. The presence of infection and chronic kidney disease were predictors of poor outcome. These findings align with earlier studies and suggest that the mortality of patients with SE has improved with time, potentially reflecting advances in critical care and SE management. Future multicenter studies are warranted to enhance generalizability and validate our findings.
Acknowledgments
We thank the staff of the Research Unit of the College of Medicine, King Saud bin Abdulaziz University for Health Sciences, Riyadh for their support and insightful comments on the research proposal. We also thank Enago (www.enago.ae), the editing brand of Crimson Interactive Inc, for English language editing. This service was provided through King Abdullah International Research Center, Riyadh.
Footnotes
Disclosure. Authors have no conflict of interests, and the work was not supported or funded by any drug company.
- Received December 15, 2024.
- Accepted July 31, 2025.
- Copyright: © Neurosciences
Neurosciences is an Open Access journal and articles published are distributed under the terms of the Creative Commons Attribution-NonCommercial License (CC BY-NC). Readers may copy, distribute, and display the work for non-commercial purposes with the proper citation of the original work.
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.