Usefulness and yield of routine electroencephalography: a retrospective study

Link to article at PubMed

Intern Med J. 2021 Oct 6. doi: 10.1111/imj.15556. Online ahead of print.


BACKGROUND: The electroencephalogram (EEG) is a common diagnostic tool used to investigate patients for various indications including seizure disorders. In our study we investigate factors that predict the presence of epileptiform abnormalities on EEG and review the common indications for ordering an EEG.

METHODS: We retrospectively reviewed all routine adult EEGs performed in a hospital over a six-month period. Data collated included patient demographics, clinical indication for EEG, setting in which EEG was performed, activation procedures utilised, history of epilepsy, and whether the patient was on antiepileptic medication. Our primary objective was to evaluate the factors that were predictive of an EEG with epileptiform abnormalities.

RESULTS: 239 routine EEGs were included with indications including first seizure (25.9%), known epilepsy (25.1%), cognitive change (15.9%), syncope (15.0%), movement disorder (6.7%), psychogenic non-epileptic events (5.4%), unresponsiveness/ICU (4.6%), and psychiatric presentation (1.3%). Most EEGs were normal (48.1%). 8.9% of EEGs demonstrated epileptiform abnormalities. Using multivariate logistic regression, three variables proved significant in predicting an EEG with epileptiform abnormalities. Any seizure as an indication (first seizure or seizure in known epileptic), increasing patient age, and EEGs conducted in an inpatient setting and within 48 hours of seizure event were all statistically more likely to yield epileptiform abnormalities on EEG.

CONCLUSIONS: Our findings suggest that careful selection of patients based on appropriate indications for EEG referral would likely improve the yield of an EEG. Depending on the indication, a normal EEG result can be of similar usefulness to an abnormal EEG demonstrating epileptiform abnormalities. This article is protected by copyright. All rights reserved.

PMID:34611977 | DOI:10.1111/imj.15556

Leave a Reply

Your email address will not be published.