Epidemiology of Seizures and Association With Mortality in Adult Patients Undergoing ECMO: A Systematic Review and Meta-analysis
Jaeho Hwang,Armaan F Akbar,Lavienraj Premraj,Eva K Ritzl,Sung-Min Cho
DOI: https://doi.org/10.1212/WNL.0000000000209721
IF: 9.9
2024-08-27
Neurology
Abstract:Background and objectives: Extracorporeal membrane oxygenation (ECMO) provides lifesaving support to patients with cardiopulmonary failure. Although seizures increase mortality risks among critically ill patients broadly, studies specific to adult ECMO patients have largely been limited to single-center studies. Thus, we aimed to perform a systematic review and meta-analyses of seizure prevalence, mortality, and their associations in adult ECMO patients. Methods: PubMed, EMBASE, Cochrane trial registry, Web of Science, and SCOPUS were searched on August 5, 2023. Following the Preferred Reporting Items for Systematic Reviews and Meta-analyses guidelines, we included studies of adults undergoing venovenous ECMO (VV-ECMO), venoarterial ECMO (VA-ECMO), or extracorporeal cardiopulmonary resuscitation (ECPR) that reported seizures during ECMO. The extracted data included study characteristics, patient demographics, ECMO support, EEG monitoring, and seizures, organized by ECMO types. Forest plot and meta-regression analyses were performed. Bias assessment was performed with the Egger test and Newcastle-Ottawa Scale. Results: Twenty-three studies (n = 40,420, mean age = 51.8 years, male = 62%) were included. Data were extracted by ECMO type as follows: VV-ECMO (n = 16,633), non-ECPR VA-ECMO (n = 11,082), ECPR (n = 3,369), combination of VA-ECMO and ECPR (n = 240), and combination of all types (n = 9,096). The pooled seizure prevalence for all ECMO types was 3.0%, not significantly different across ECMO types (VV-ECMO = 2.0% [95% CI 0.8-4.5]; VA-ECMO = 3.5% [95% CI 1.7-7.0]; ECPR = 4.9% [95% CI 1.3-17.2]). The pooled mortality was lower for VV-ECMO (46.2% [95% CI 39.3-53.2]) than VA-ECMO (63.4% [95% CI 56.6-69.6]) and ECPR (61.5% [95% CI 57.3-65.6]). Specifically, for VV-ECMO, the pooled mortality of patients with and without seizures was 55.1% and 36.7%, respectively (relative risk = 1.5 [95% CI 1.3-1.7]). Similarly, for VA-ECMO, the pooled mortality of patients with and without seizures was 74.4% and 56.1%, respectively (relative risk = 1.3 [95% CI 1.2-1.5]). Meta-regression analyses demonstrated that seizure prevalence was not associated with prior neurologic comorbidities, adjusted for ECMO type and study year. Discussion: Seizures are infrequent during ECMO support. However, they were associated with increased mortality when present. Multi-institutional, larger-scale studies using standardized EEG monitoring are necessary to further understand the risk factors of specific classes of seizures for individual ECMO types, and their effects on mortality. Limitations of our study include missing data for details on seizure types, sedating/antiseizure medications used during ECMO, other ECMO-related complications, and EEG recording protocols.