Among people with epilepsy, there is a 20-fold higher risk of dying suddenly and unexpectedly compared with the general population. This phenomenon is called sudden unexpected death in epilepsy (SUDEP) and the term is used when sudden death occurs in an otherwise reasonably healthy person with epilepsy and the autopsy is unrevealing. In most cases, SUDEP occurs during sleep and is unwitnessed. Risk factors for SUDEP include the presence or number of generalized tonic-clonic seizures (GTCS), nocturnal seizures, young age at epilepsy onset, longer duration of epilepsy, dementia, absence of cerebrovascular disease, asthma, male gender, symptomatic aetiology of epilepsy and alcohol abuse. Suggested factors predisposing to SUDEP have included long-QT-related mutations, impaired serotonergic brain stem control of respiration, altered autonomic control and seizures with a pronounced postictal suppression and respiratory compromise. Final events that may lead up to SUDEP are a postictal CNS shutdown with pronounced EEG suppression, ictal or postictal apnoea, and ictal cardiac arrhythmia. It is unknown whether antiepileptic drugs (AEDs) modify the risk for SUDEP. Studies have consistently found that the presence or number of GTCS is associated with an increased risk for SUDEP. Since continued presence of GTCS clearly necessitates the use of AEDs, both factors must be taken into account to determine whether one or both increases the risk for SUDEP. Some studies suggest that AEDs, such as lamotrigine and carbamazepine, may increase the risk of SUDEP, but rarely adjust for GTCS. Other studies, which have found that AEDs are associated with a decreased SUDEP risk, either adjust for the number of GTCS or are meta-analyses of randomized clinical trials. Studies assessing the impact of AEDs on the risk for SUDEP are limited because SUDEP is a rare event, making randomized clinical trials impossible to conduct. Observational studies focus on whether or not an AED was prescribed. When postmortem AED concentrations are assessed they are usually low or absent, perhaps due to sampling in deceased individuals, making it difficult to fully resolve whether AEDs increase or decrease SUDEP risk. Despite these caveats, the evidence suggests that AEDs are not associated with an increased risk for SUDEP on a population level, although some individuals may be susceptible to effects of AEDs. Recent evidence from a meta-analysis of randomized clinical trials of adjunctive AEDs at efficacious doses provides strong support for AED treatment as mono- or polytherapy to increase seizure control and protect against SUDEP in patients with refractory epilepsy. For patients for whom seizure control is unattainable, supervision or monitoring may prevent SUDEP, though this has never been formally tested.