Diagnostic value of extending outpatient EEG to record sleep - A retrospective study

Evidence for the utility of sleep electroencephalography (EEG) in the diagnosis of epilepsy is inconsistent, often necessitating repeat visits and sleep deprivation (SD), which increases patient and resource burden. In our setting in India, patients are often referred for EEG immediately following consultation (single visit) and EEG sessions are extended to allow for spontaneous sleep acquisition without prior SD. We aimed to evaluate the success rate of achieving sleep during these extended outpatient recordings and its incremental diagnostic yield for interictal epileptiform discharges (IEDs).

We retrospectively analyzed 1762 outpatient EEGs from patients (45% female) with suspected or known epilepsy. All patients underwent an extended routine EEG (~40-50 min), incorporating standard activation procedures and aiming to capture spontaneous sleep where possible. Successful sleep acquisition was defined as the attainment of at least NREM stage N2. IED detection rates were compared between wakefulness and sleep across different epilepsy types.

Sleep was achieved in 70.5% of patients (median age, 28 years; IQR: 16-55; range, 0.3-90). The proportion attaining sleep increased with recording duration, from 36.7% at 20 min to 55.6% at 30 min and 66.7% at 50 min, with a median time to sleep onset of 20 min. Among 1199 patients with both awake and sleep states captured, wakefulness identified IEDs in 29% (n = 350), while sleep captured unique IEDs in an additional 12% (n = 141), increasing the total yield significantly. In a sub-analysis of confirmed epilepsy cases (n = 751), wakefulness showed IEDs in 42% (n = 315), with sleep adding another 16% (n = 123). The diagnostic yield of sleep recordings was significantly higher than that of wakefulness alone (OR: 1.31; 95% CI: 1.02-1.68; p = 0.043), particularly among adults with idiopathic generalized epilepsy (IGE) (OR: 2.20; 95% CI: 1.30-3.73; p = 0.004) and children with self-limited epilepsy (OR: ∞; p = 0.031).

Despite a busy outpatient setting, 70.5% of patients attained N2 sleep during an extended EEG recording without SD. Sleep capture increased progressively with recording duration, with the greatest incremental gain between 20 and 30 min, suggesting that a modest extension beyond 20 min substantially increases the likelihood of capturing sleep. Sleep recording increased IED detection over wakefulness alone, particularly in children, in adults with IGE, and in self-limited epilepsies. Whether this approach can substitute for a sleep-deprived EEG requires prospective, head-to-head evaluation.