The Diagnostic Accuracy of Prolonged Ambulatory Versus Routine EEG

Abstract

Prolonged ambulatory electroencephalography (paEEG) is increasingly used in clinical practice but its diagnostic accuracy relative to that of routine EEG (rEEG) remains uncertain. We examined a consecutive sample of 72 individuals who had undergone 32-channel paEEG immediately after an rEEG, creating perfectly matched EEG samples. Each recording was prospectively assessed for epileptiform discharges (ED) and nonepileptiform abnormalities. The median paEEG duration was 22.5 hours (interquartile range: 22.0-23.0). The sensitivity of paEEG was 2.23 times greater than that of rEEG [sensitivity ratio: 2.23 (95% CI = 1.49-3.34)] if a positive test was limited to the presence of epileptiform discharges. This benefit of paEEG versus rEEG was no longer evident if the definition of a positive test included nonepileptiform abnormalities (sensitivity ratio 1.26; 95% CI = 1.02-1.55). The specificity of the 2 tests was not evidently different (specificity ratio 0.67; 95% CI = 0.17-2.67). Twenty-six percent of paEEG recorded epileptic seizures while none of the rEEG did (absolute difference 26.0% (95% CI = 11.8-40.2). Our findings quantify the benefit of 32-channel paEEG, relative to rEEG, and support its role in the diagnosis and characterization of epilepsy.