Rationale:
Infantile spasms (IS) are a severe epileptic encephalopathy of infancy requiring prompt diagnosis and treatment to improve neurodevelopmental outcomes. Electroencephalography (EEG) plays a central role in diagnosis, particularly the identification of hypsarrhythmia or related patterns. While sleep EEG is believed to enhance diagnostic yield by increasing the visibility of interracial abnormalities, its added value following a routine awake EEG remains unclear. This study aims to evaluate whether performing a sleep EEG after a normal routine EEG significantly improves diagnostic accuracy in suspected cases of infantile spasms, thereby guiding more efficient and timely investigations.
Methods:
We retrospectively reviewed EEG data from 100 children aged 1 to 24 months referred for suspected infantile spasms or stereotyped motor events between November 2020 and August 2024. Of these, 97 children underwent an awake EEG as their initial test, 2 were asleep during the initial EEG, and 1 had a sleep EEG from the outset. Twenty-four children were referred for follow-up sleep EEGs due to normal or inconclusive awake EEG results. Children referred for unrelated paroxysmal events (e.g., febrile seizures, absence seizures) were excluded. We assessed whether sleep EEG identified new or clearer epileptic abnormalities that could aid diagnosis.
Results:
Of the 97 initial awake EEGs, 2 confirmed epileptic spasms (1 repeated post-treatment and returned to normal), 5 showed epileptiform activity not diagnostic of IS, and 89 showed no abnormalities (1 report was missing). Among 24 follow-up sleep EEGs:
- 1 case with a normal awake EEG showed suspicious findings on sleep EEG (not definitive for IS)
- 21 had normal sleep EEGs confirming the normal awake EEG
- 1 with abnormal awake EEG had abnormal sleep EEG (not IS)
- 1 with abnormal awake EEG had normal sleep EEG
The diagnostic yield of sleep EEG following a normal awake EEG was therefore 4.2% (1/24). This is lower than yields reported in prior studies: Lux et al. (2010) found 10–15%, a 2018 UK audit reported 10% (2/20), and Severn et al. (2021) found 7% in similar contexts. The lower yield in our audit may reflect stricter referral criteria, a high proportion of non-epileptic events, or improved awake EEG interpretation.
Conclusions:
Sleep EEG had limited incremental diagnostic value when awake EEG was normal in infants with suspected spasms. While sleep EEG may still be appropriate in select cases where clinical suspicion remains high, routine use after a normal awake EEG may not be justified. Careful clinical assessment should guide the decision for further EEG evaluation, supporting more targeted and efficient use of resources.
Funding:
This work was unfunded and carried out as part of a routine clinical audit project.