Prioritising EEG referrals by those most likely to show epileptiform abnormalities could help improve diagnostic, time and cost benefits for neurologists and their patients, Australian clinicians say.
A retrospective study of 239 routine EEGs run at Blacktown Hospital in 2017 found up to 84.9% of patients were referred for higher-yield indications such as first seizure, (25.9%), known epilepsy (25.1%), cognitive decline or delirium (15.9%), movement disorders (6.7%), suspected psychogenic non-epileptic events (5.4%), unresponsiveness or ICU admission (4.6%) or psychiatric presentations (1.3%), yet 15.1% still received EEGs for syncope.
Overall, 10.2% of EEGs returned epileptiform abnormalities, with patients having any seizure (odds ratio [OR]: 7.70, 95% CI: 2.56 – 23.13), older age (over 60s vs under 30s OR: 7.77, 95% CI: 0.93 – 65.0) and EEGs conducted in an in-patient setting and within 48 hours of seizures (OR: 3.29, 95% CI: 1.03 – 10.52) most likely to yield positive results.
Activation procedures such as photic stimulation and hyperventilation did not significantly affect epileptiform abnormality yield on EEG, though this may be explained by the small number of susceptible patients studied, the authors wrote in the Internal Medicine Journal.
Given their higher propensity for epileptiform abnormalities, patients who are within 48 hours of their first seizure should be prioritised for EEG, where practicable, the authors suggested.
EEG should also be considered in known epilepsy patients with ongoing confusion post-seizure, where a normal result could help rule out non-convulsive status epilepticus, they added.
The same would apply for any confused, unresponsive or cognitively declining patient without epilepsy, they wrote.
While psychogenic non-epileptic seizures had low abnormality yield, EEGs can have a suggestive effect, making patients more likely to have a habitual event while recording. Capturing these events could help provide diagnostic certainty and avoid unnecessary treatments. Thus, the authors recommend referral in these patients.
Meanwhile, EEG may be a lesser priority in known epilepsy patients without confusion, since it would only confirm their diagnosis without altering treatment.
Finally, “routine EEG should be avoided if a clear clinical history of syncope is given”, the authors suggested.
“Previous publications have demonstrated that EEGs for the evaluation of suspected syncope had a low diagnostic yield and never altered the management of the patient, despite the specialty of the referring physician ordering the EEG,” they explained.
They could also result in epilepsy misdiagnosis, which could have “significant lifelong implications”.
“Our findings suggest that rational selection of patients based on clinical indication is an important factor for improving the cost and time benefit of the EEG,” the authors wrote, citing the scarcity of EEG in the public hospital system.
“Unnecessary referrals for an inpatient EEG could be avoided by understanding the usefulness and yield of certain indications.”
“Based on the results of this study, we would suggest that precedence be given to patients with indications more likely to yield an epileptiform abnormality such as first seizure presentations,” they wrote.