The Epworth Sleepiness Scale in epilepsy: Internal consistency and disease-related associations

The Epworth Sleepiness Scale (ESS) is the most common instrument for measuring subjective sleep propensity in people with epilepsy but has not yet been validated in this population.

We aimed to systematically assess the validity, performance, and internal consistency of the ESS, as well as correlations between the ESS and disease-specific variables and patient-reported outcome measures in a cohort of adults with epilepsy (AWE).

Ninety-five AWE completed sleep and seizure diaries, in-laboratory polysomnography (PSG) and patient-reported outcome measures, including the ESS, Insomnia Severity Index (ISI), and the Beck Depression Inventory (BDI). Demographic information and data from 95 matched controls referred for PSG for suspected obstructive sleep apnea (OSA) was taken from the electronic medical record. Frequencies of high ESS item ratings (item score ≥2) were calculated for each group. Cronbach's α and factor analysis were performed to assess the internal consistency and validity of the ESS within cases and controls. Multivariable linear models were used to assess the association between ESS and predictors of interest, adjusting for demographic and disease-specific variables, including seizure type, frequency, and anti-seizure medication (ASM) therapy.

While suspected OSA controls had significantly greater mean ESS total scores (9.9 vs 7.9, p = 0.004) and proportion with ESS >10 (42% vs 25%, p = 0.014), there were no significant differences in the severity of item responses, with the exception of "lying down to rest in the afternoon when circumstances permit," for which more controls rated as likely/very likely (79% vs 64%), p = 0.024). AWE with ESS >10 had higher mean standardized ASM dose (2.5 vs 1.7, p = 0.026). All ESS items were significantly correlated with the total score within each group. Cronbach's α was 0.75 for cases and 0.85 for controls, indicating good internal consistency of the ESS for both groups. After adjusting for demographic and sleep characteristics, higher ESS scores were associated with greater insomnia scores on the ISI (p = 0.024) and depressive symptoms on the BDI (p = 0.018).

This study provides validity for the use of the ESS in adult populations with epilepsy.