Effects of sleep deprivation on the pediatric electroencephalogram.

Background: The routine electroencephalogram aids in epilepsy syndrome diagnosis. Unfortunately, routine outpatient electroencephalogram results are normal in roughly half of children with epilepsy. To increase the yield, practice guidelines recommend electroencephalograms with sleep and sleep deprivation. The purpose of this study was to rigorously evaluate this recommendation in children.

Methods: We conducted a randomized, blinded comparison of routine electroencephalograms versus sleep-deprived electroencephalograms in 206 children aged 0 to 18 years. Electroencephalograms were ordered for standard indications after a neurologist's clinical assessment indicated > or =1 seizure (83%) or unclear spell (17%). The primary outcome was the proportion of normal routine electroencephalogram results versus sleep-deprived electroencephalogram results. Logistic regression modeling was used to assess the influence of sleep, as well as other clinical factors.

Results: Although children with sleep-deprived electroencephalograms had less sleep the night before (4.9 vs 7.9 hours) and more sleep during electroencephalograms (73% vs 55%), the increase in electroencephalogram yield was borderline significant (56% normal sleep-deprived electroencephalogram versus 68% normal routine electroencephalogram). Moreover, sleep during the electroencephalogram did not increase its diagnostic yield. Sleep-deprived electroencephalogram yield tended to be higher in children with preelectroencephalogram clinical diagnosis of seizure(s) and at older ages (>3 years).

Conclusions: Sleep deprivation, but not sleep during the electroencephalogram, modestly increases the yield of the electroencephalogram in children diagnosed with seizures by neurologists. Compared with a routine electroencephalogram, the number needed to test with sleep-deprived electroencephalogram to identify 1 additional child with epileptiform discharges is approximately 11.