Can You Die From a Seizure? SUDEP Risk, Timeline, and Real Numbers

Yes, a seizure can kill you, but it is uncommon, and the mechanism matters more than the diagnosis. Roughly 1 in 1,000 people with epilepsy die each year from sudden unexpected death in epilepsy, known as SUDEP, while status epilepticus carries a 10 to 20 percent mortality per episode. Most individual seizures stop on their own within two minutes and leave no lasting damage. Death comes through four routes: the seizure never stops, breathing fails after it ends, the body is injured during convulsions, or the heart's rhythm breaks down. Risk concentrates in convulsive, poorly controlled, and nocturnal seizures.

How a Seizure Kills: Four Mechanisms

A seizure is a storm of synchronized firing across part of the brain. Healthy brain tissue has powerful brakes: inhibitory GABA signals, rising adenosine, and falling extracellular calcium. Most seizures end within 60 to 120 seconds because those brakes engage. Death requires either the brakes to fail or a second organ system to collapse at the same moment.

Route one is the seizure that never stops. Status epilepticus means continuous seizure activity beyond 5 minutes, or repeated seizures without full recovery between them. Neurons then fire faster than the brain's energy supply can keep pace. Calcium floods the cells and glutamate accumulates between them, and neurons begin to die after roughly 30 to 60 minutes of uninterrupted activity. The body often fails first: core temperature above 104 F, muscle breakdown, acidosis, and aspiration.

Route two is respiratory failure after the seizure. Breathing commonly pauses for 30 to 60 seconds after a tonic-clonic seizure, and it can pause repeatedly for many minutes. The airway may close from laryngospasm, and fluid can leak into the lungs. This is the SUDEP pathway, and it starts after the shaking has already stopped.

Route three is physical injury. Falls, head strikes, burns, car crashes, and drowning all kill during seizures. Water is the most lethal setting, because a convulsion in a bathtub or pool removes the one thing that prevents drowning: the ability to lift the head.

Route four is cardiac. Ictal asystole, a pause in the heartbeat triggered by the seizure itself, appears in roughly 0.3 to 0.5 percent of seizures captured in epilepsy monitoring units. It usually causes fainting rather than death, but the risk is real in people with pre-existing heart disease.

The most important sentence in this article is that most seizure deaths happen after the convulsion ends. In plain terms, the dangerous window is the first minutes to two hours of recovery, not the 90 seconds of shaking.

Can You Die From a Seizure? The Quantified Risk

Death from a seizure is real, and it is also statistically unusual at the level of the single event. In community-based cohorts, SUDEP kills about 1.2 people per 1,000 with epilepsy each year. In clinics that treat drug-resistant epilepsy, the rate climbs to roughly 5 to 9 per 1,000 per year. Children with epilepsy sit far lower, at roughly 1 in 4,500 per year.

That spread is the point. Risk is not a property of the diagnosis; it is a property of seizure control, seizure type, and supervision. A person with three convulsions a month and nobody at home overnight faces a very different number from someone seizure-free for five years on one medication.

A single average also hides the comparison that matters most. For a healthy 25-year-old, the yearly chance of dying from any cause is roughly 1 in 1,000. Adding uncontrolled convulsive epilepsy roughly doubles that annual risk, which is why SUDEP stands out in young adults who have no other reason to die.

These figures use different denominators and should not be added together. The 1 in 1,000 figure is an annual risk across everyone with epilepsy, the 5 to 9 per 1,000 figure is an annual risk inside a harder-to-treat subgroup, and the 10 to 20 percent figure is the chance of dying from one status epilepticus episode.

US arithmetic makes the scale concrete. About 3.4 million Americans have epilepsy, per CDC estimates. An annual SUDEP rate of 1 to 1.2 per 1,000 predicts roughly 3,400 to 4,100 deaths a year, though death-certificate studies capture only a fraction of them.

SUDEP: The Death With No Autopsy Finding

SUDEP is defined, per the 2012 International League Against Epilepsy criteria, as a sudden and unexpected death in someone with epilepsy. The death may be witnessed or unwitnessed, and it must not be traumatic or drownings-related. Autopsy must show no structural or toxicologic cause, and documented status epilepticus must be excluded. That exclusion matters: SUDEP is not the same thing as dying in status.

Autopsy in SUDEP usually finds nothing specific, though the lungs are often heavy with fluid and the tongue may be bitten. Roughly 70 percent of SUDEP victims are found face down, and the great majority of deaths happen during sleep or when nobody is nearby. This is why the bedroom, not the emergency department, is the setting that prevention efforts target.

The leading explanation is a failure of the brain's breathing and arousal circuits in the hour after a convulsion. Postictal apnea is common and can repeat, while the person remains deeply unrousable. Prone positioning then compounds the problem: the face is down, the airway is partly blocked, and the normal reflex to turn the head is suppressed.

Two mechanistic hypotheses are still being tested. The serotonin hypothesis holds that medullary 5-HT neurons, which help drive the arousal response to rising carbon dioxide, are impaired in some people with epilepsy; mice with lesioned serotonin neurons are more likely to die after a seizure. The adenosine hypothesis points to a surge of adenosine acting on A1 receptors in the brainstem, which suppresses breathing and flattens the EEG; adenosine antagonists rescued mice in those experiments. Both theories describe the same failure: the seizure ends, but the automatic drive to breathe does not restart.

Rare variants in cardiac and neuronal channel genes, including SCN1A, KCNQ1, and SCN5A, appear in some SUDEP cases. That overlap suggests a minority of people carry a shared vulnerability in both the heart and the brain.

The Timeline: From First Convulsion to Irreversible Damage

Timing is the most actionable knowledge in seizure care. A typical generalized tonic-clonic seizure lasts 60 to 120 seconds, and most stop well before 5 minutes. The 5-minute mark is a treatment threshold, not a death sentence; it is the point at which the brain's own braking has probably failed.

The 5-minute rule replaced a 30-minute definition for a practical reason. Per the 2016 American Epilepsy Society guideline, treatment should begin at 5 minutes, because waiting 30 minutes lets status become established and far harder to break. First-line therapy is weight-based lorazepam, typically 0.1 mg per kg up to 4 mg, repeated once if needed.

Delay has a measurable cost. In a population-based study from Rochester, Minnesota, status lasting under an hour carried a mortality near 3 percent, while status lasting more than an hour carried roughly 32 percent. Age compounds it: adults over 65 die from status at roughly two to three times the rate of younger adults.

Neuronal injury in primate models begins after roughly 30 to 60 minutes of continuous seizure activity. Systemic damage arrives earlier, with fever, acidosis, and muscle breakdown appearing within 20 to 30 minutes. In plain terms, the brain has more time than the body does.

Who Is at Highest Risk: Variations and Edge Cases

Convulsion frequency is the strongest single predictor. In cohort studies, having three or more generalized tonic-clonic seizures a year is associated with roughly a 15-fold higher SUDEP risk than having none. Nocturnal and unwitnessed seizures carry additional risk, because the postictal apnea goes unobserved.

Drug-resistant epilepsy, defined as failure of two appropriately chosen medications, affects about 30 percent of people with epilepsy and concentrates most of the mortality. Abrupt medication withdrawal is a classic trigger, which is why epilepsy monitoring units run continuous observation; reported SUDEP rates there sit near 1 per 10,000 admissions.

Age shifts the pattern rather than removing it. SUDEP incidence peaks in adults between roughly 20 and 40 years old, when background mortality is low and the relative excess is largest. Children have lower rates, but sudden death in a child with epilepsy is not rare enough to ignore.

Surgery changes the trajectory. People who become seizure-free after resective epilepsy surgery see their SUDEP risk fall toward general-population levels over 5 to 10 years. Pregnancy is another edge case: seizure frequency worsens in roughly 1 in 4 to 1 in 3 pregnancies, mostly when medication levels fall.

One misconception-adjacent edge case deserves its own line. Psychogenic nonepileptic seizures do not cause SUDEP, but they are not harmless. People with that diagnosis have higher mortality than the general population, mostly from other causes, and a delayed diagnosis exposes them to unnecessary medication.

Misconceptions and Red Flags

Four beliefs are common and dangerous. The first is that a person can swallow their tongue; the tongue cannot move backward into the throat, and objects forced into the mouth break teeth and block airways. The second is that every seizure damages the brain. Brief seizures without hypoxia leave no measurable structural injury, while prolonged ones clearly do.

The third is that the danger ends when the shaking stops. Postictal apnea and SUDEP occur precisely in that window. The fourth is that SUDEP only happens in severe epilepsy. It occurs in newly diagnosed and well-controlled patients too, just less often.

Red flags that mean call 911 rather than watch and wait:

• The seizure lasts longer than 5 minutes, or a second seizure begins before the first has worn off.
• Breathing does not resume within a minute after convulsions stop, the lips stay blue, or the person keeps gasping.
• The seizure happened in water, even briefly, or the person was pulled from a bath or pool.
• There is injury, pregnancy, diabetes, or heat exposure, or the person remains unresponsive.
• This is a first-ever seizure, or a pattern that is new for this person.

What to Actually Do

Time the seizure and write down the start time, because every treatment decision downstream depends on duration. Cushion the head, clear hard objects away, and loosen anything tight at the neck. Once the convulsion subsides, roll the person onto their side into the recovery position and stay until breathing is normal and they can answer questions.

Never restrain a convulsing person and never place anything in the mouth. Do not give food, drink, or pills until the person is fully alert. If a rescue medication is prescribed, it is usually a benzodiazepine given nasally or rectally; adult nasal midazolam is dosed at 5 mg, and caregiver protocols typically allow one repeat dose after 5 to 10 minutes if the first fails.

Medication adherence is the highest-yield preventive action available. Missing doses and stopping drugs abruptly are among the strongest modifiable triggers for both status epilepticus and SUDEP. Sleep is the second lever: most adults with epilepsy need 7 to 9 hours, and sleep deprivation reliably lowers seizure threshold.

Nocturnal supervision deserves a specific plan. Room-sharing during medication changes, an audio or video monitor, and a pulse-oximetry alarm all shorten the time between a convulsion and a response. Be clear about the evidence, though: no device has been shown in a randomized trial to prevent SUDEP, and the supporting data are observational.

Bathroom and water safety are unglamorous and effective. Take showers rather than baths, keep the drain open, and avoid scalding water. Swim with a partner who knows about the epilepsy diagnosis, and never swim alone in open water.

Ask your neurologist about SUDEP directly. Surveys show most patients want the conversation and many clinicians skip it. If two appropriate medications have failed, ask for referral to a comprehensive epilepsy center, where surgery, dietary therapy, and device options are assessed.

The limits of these numbers are worth stating plainly. SUDEP incidence estimates vary two- to threefold depending on how deaths are counted, and death certificates often list epilepsy as a contributing rather than underlying cause. Risk models work at the group level and cannot tell one person what will happen to them. Much of the mechanism remains hypothesis rather than settled fact, and most incidence data come from high-income countries with strong monitoring infrastructure.