Another reason that patients with epilepsy are at risk for sleep disorders is that treatment with antiepileptic drugs (AEDs) may directly contribute to daytime sleepiness or insomnia, or may exacerbate underlying sleep disorders like obstructive sleep apnea. On the other hand, these drugs may also improve sleep organization.

Studies of people taking AEDs have found that they affect sleep organization in highly variable ways: 11,12

  • Phenobarbital shortens sleep latency and decreases the number of arousals in patients with epilepsy.
  • Benzodiazepines decrease time to sleep onset, increase the amount of NREM stage 2 sleep, decrease the quantity and the amplitude of delta (stages 3 and 4) NREM sleep, decrease the amount of REM sleep, prolong REM sleep latency, and decrease the number and duration of awakenings and arousals.
  • Phenytoin increases the amount of delta NREM sleep subacutely (after 4–6 weeks of treatment) in patients with epilepsy, but the only chronic effect is shortened sleep latency. 13
  • Ethosuximide decreases delta sleep and increases stage 1 NREM sleep in epilepsy patients. 14
  • Valproic acid, when given in high doses (1,000 mg) to healthy subjects, decreased REM activity and increased delta activity with period analysis, but visual-analyzed data did not show a statistically significant difference. 15
  • Carbamazepine decreased sleep fragmentation, increased delta NREM sleep, and increased total sleep time when used to treat subjects with bipolar disease. 16 Newly diagnosed patients with epilepsy had an improvement in the fragmentation of their sleep owing to awakenings, after treatment for 1 month. 6 Paradoxically, in one report, initiation of controlled-release carbamazepine in epilepsy patients provoked a reduction and fragmentation of REM sleep and an increase in the number of sleep stage shifts.17 These effects were almost completely reversed after 1 month of treatment, however, and no significant difference was noted between the baseline condition and long-term follow-up.
  • Felbamate had stimulantlike effects in patients with epilepsy who took it as part of a 2-week randomized double-blind placebo-controlled trial, followed by open felbamate monotherapy.18 Another study of the side effects of felbamate reported insomnia in 25% of 60 epilepsy patients.19 Psychiatric rating scales demonstrated stimulantlike effects (e.g., insomnia, anorexia, and anxiety) in both acute and chronic phases of treatment.
  • Lamotrigine can also cause insomnia.20 Among 109 patients taking lamotrigine, 7 (6.4%) had insomnia requiring a change in therapy. The insomnia appeared to be dose-dependent. In contrast, a polysomnographic study of lamotrigine in 7 subjects with epilepsy showed no effect on total sleep time, sleep efficiency, sleep latency or REM latency, or the percentage of time spent in NREM sleep stages. Stage shifts and arousal indices were reduced and the percentage of REM was increased.21

The effects of the more recently introduced AEDs on sleep in epilepsy patients await further investigations. With the large number of AED choices currently available and with additional AEDs in clinical testing, clinicians may base their AED choice not only on the epilepsy syndrome, but also on its effects on sleep. For example, sedating AEDs may benefit epilepsy patients with insomnia, and stimulating AEDs may benefit epilepsy patients with daytime sleepiness.

In addition to their direct pharmacologic effects on sleep, AEDs may indirectly improve sleep in the epilepsy patient through the reduction of seizures and, in some cases, interictal epileptic discharges.

Conversely, AEDs may adversely affect sleep by contributing to sleep disorders. For example, in a patient predisposed to obstructive sleep apnea, barbiturates and benzodiazepines may worsen the frequency of apneas and hypopneas by decreasing upper airway resistance or arousal mechanisms. AEDs that are associated with weight gain, such as valproate, may also worsen this condition. Avoiding these agents for patients with untreated obstructive sleep apnea may be advisable, especially if alternative AEDs are available.