Abstracts

Rationale: Antiepileptic drug (AED) withdrawal is used in epilepsy-monitoring units (EMUs) to increase the likelihood of capturing seizures to guide diagnosis and treatment decisions; however, controversy exists around the safety and best AED withdrawal approach. We completed a systematic review of the literature to determine current research and findings related to AED withdrawal in EMUs.Methods: Medline and Embase were searched using terms related to AEDs, EMUs, and medication withdrawal. Two independent reviewers screened abstracts and full-text and abstracted data from eligible studies, resolving disagreements by consensus. Included studies 1) involved EMU patients, 2) assessed AED withdrawal in relation to EMU outcomes. Excluded studies had 1) small sample sizes (N<20), 2) patients with only non-epileptic seizures, or 3) did not report AED withdrawal approaches used. Analyses were carried out in three phases to 1) assess study design (data collection, follow-up, sample size, and selection criteria), 2) summarize medication protocols, and 3) assess reported outcomes. A meta-analysis was not attempted due to the considerable heterogeneity of methods and outcomes. Results: Of 2351 abstracts identified, 55 were reviewed in full-text and 12 were selected for data abstraction (2 identified by hand searching). Studies reporting on the same patients were combined for analysis. No randomized clinical trials were identified. All were cohort studies (8 prospective, 3 retrospective). 3 prospective studies had follow-up after EMU discharge of 0.5, 6, and 12 months. 5 studies grouped patients by AED type (3/5) or temporal versus extratemporal lobe epilepsy (2/5). Median sample size was 43 patients. All studies specified selection criteria and 10/11 used ?3 criteria. The most common (9/11) were epilepsy or seizure type AED regimen. 3 studies only included patients who had seizures recorded in the EMU. Most studies (9/11) used a standard AED withdrawal protocol for all patients, 5 completed withdrawal over 24-hours and 4 over 72-hours. AED regimen was the most common patient factor associated with protocol withdrawal type (6/9). Four studies had protocols for IV lorazepam to prevent status epilepticus. Seizure frequency was reported by all studies but the reporting variation prevented statistical comparison. Mean length of stay (LOS) was reported in 7/11 studies (4.55 days). The remaining studies had standard LOS of 5 (2/11) or 9 (2/11) days regardless of AED protocol. Other outcomes included frequency of secondary generalization, seizure clustering, and time to record seizures. Most studies (10/11) completed analyses beyond descriptive statistics and 4 studies used multivariate analyses Conclusions: This study demonstrates that randomized controlled trials and other prospective studies are urgently needed to address AED withdrawal in EMUs as there are currently insufficient studies of suitable design to determine the best and safest AED withdrawal approach for those admitted to EMUs.