Abstracts

Rationale: Hypoxemia, as identified by oxygen desaturation, occurs in over 30% of epileptic seizures. It is widely assumed not to occur in psychogenic non-epileptic seizures (PNES), but to our knowledge has not been studied in this patient group.Methods: We reviewed all inpatient video-EEG monitoring studies (including video, 22 channel EEG and fixed pulse oximetry) from May 2005 to date June 2008, identifying a total of 281 events from 74 patients. Of these, 203 events from 45 patients were epileptic attacks and 78 events from 29 patients were PNES. 2 patients had both PNES and epileptic seizures. We reviewed oxygen saturations in all events, using the widely accepted threshold of below 92% to define desaturation. We excluded from analysis all events in which desaturation appeared to be artifactual. We undertook statistical comparison of desaturation in epileptic seizures versus PNES.Results: Oxygen saturation data was available in 225(80%) of the 281 events (166 epileptic and 59 PNES). In the PNES group, after exclusion of 3 events due to suspicion of artefact, desaturation was seen in 14(25%) events. In the epilepsy group, oxygen desaturation was observed in 64(39%) events. Overall,9( 30%) of patients with PNES demonstrated oxygen desaturation in at least one attack compared with 26(57%) of subjects with epilepsy. Two patients with PNES and oxygen desaturation had events from sleep, with clinical episodes apparently arising from sleep-related desaturations from undiagnosed obstructive sleep apnea (OSA). Overall oxygen desaturation in PNES was observed in 25% of attacks, compared with 38.5% of true epileptic seizures. This difference was not statistically significant(( 2 = 3.376, p=0.66).Conclusions: Hypoxemia is widely recognized in epileptic seizures, and oxygen desaturation is sometimes used as an objective clinical marker to distinguish epileptic seizures from PNES. Our findings indicate that oxygen desaturation can occur in a surprisingly high proportion of patients with PNES, suggesting that the diagnostic value of this parameter is uncertain. The mechanism by which desaturation occurs in PNES is unclear, although breath holding (possibly after a period of hyperventilation) is a plausible explanation for some attacks. Some sleep-related PNES may be directly related to apnoeic episodes in individuals with OSA. Artefact, particularly from movement, may account for some cases, although we excluded events in which this was clearly the case. Caution should be used when interpreting oxygen desaturation as evidence of an epileptic basis to seizures.