Seizure Disorders · Chapter 6
Antidepressants in Epilepsy: The Safe Lane
The fear that antidepressants “lower the seizure threshold” keeps depressed epilepsy patients untreated — and undertreatment is the real danger. The evidence says the first-line agents are safe. Here's which ones, at what dose, and the handful to actually avoid.
Bottom Line Up Front
The 30-second version
- At therapeutic doses, the first-line antidepressants don't raise seizure risk. In the pivotal analysis, patients randomized to antidepressants seized less than those on placebo — the threshold fear is largely a myth for the drugs you'd actually reach for.
- The real hazard is leaving depression untreated. Depression is the single strongest driver of poor quality of life in epilepsy, the relationship runs both ways, and withholding effective treatment does measurable harm.
- Only four antidepressants are genuinely proconvulsant at therapeutic doses: clomipramine, bupropion, amoxapine, and maprotiline. Bupropion is outright contraindicated in seizure disorder.
- Pick for the interaction profile. Sertraline, escitalopram, and citalopram are preferred — effective, and the cleanest with antiseizure medications. Fluoxetine, paroxetine, and fluvoxamine are the enzyme-inhibiting troublemakers.
- Two practical catches: enzyme-inducing ASMs lower your SSRI level (often a dose increase of 30% or more), and SSRIs stacked on oxcarbazepine or carbamazepine compound hyponatremia — check the sodium.
The Real Risk Is Undertreatment
Start here, because it reframes everything that follows: in epilepsy, depression is the strongest single predictor of poor quality of life — ahead of seizure frequency itself. The relationship is bidirectional: depression is more common in people with epilepsy, and a history of depression raises the risk of developing epilepsy. Add the class-wide suicidality signal that already sits on the ASMs these patients take, and the stakes of an untreated mood disorder are high.
Yet the reflexive worry about “lowering the seizure threshold” leads clinicians to withhold or underdose antidepressants in exactly the patients who need them. That caution is the actual danger. The evidence below exists to give you permission to treat — confidently, with the right agent.
The Myth vs the Data
The proconvulsant reputation comes from two real but narrow truths: the old tricyclics and a few specific agents can provoke seizures, and antidepressants in overdose lower the threshold. Neither describes a first-line SSRI at a therapeutic dose — but the reputation stuck to the whole class.
The cleanest evidence comes from Alper and colleagues, who pooled FDA phase II/III depression trials — roughly 75,000 patients randomized to an antidepressant or placebo. Across the second-generation antidepressants other than bupropion, seizure incidence was lower on the drug than on placebo: a standardized incidence ratio of 0.48 (95% CI 0.36–0.61). The telling detail is the placebo arm — its seizure rate ran well above the general-population rate, because depression severe enough to enter a trial itself raises seizure risk. That's the confound the class's reputation never accounted for. Two exceptions in the very same dataset prove the signal is specific, not a blanket pass: bupropion ran the other way, and several antipsychotics — clozapine, olanzapine, quetiapine — raised seizure incidence. (Real-world observational studies do report a modestly higher seizure rate on antidepressants versus non-use — but that's the placebo-controlled comparison inverted, and is best read as residual confounding by depression severity, not a true proconvulsant class effect.)
“But I saw a study saying SSRIs cause seizures.” Observational case-control studies do report associations — even adjusted analyses have found roughly a twofold higher seizure rate in antidepressant users. But these are dogged by confounding by indication: the conditions antidepressants treat (depression, anxiety, insomnia, chronic pain) are themselves seizure risk factors, and a prescription is a poor proxy for any one of them. That residual signal may be leftover confounding or a small real effect — but the randomized, placebo-controlled data are the cleaner test, and they exonerate the first-line agents at therapeutic doses. When the designs disagree, weight the one that controls for why the drug was prescribed.
The Four to Actually Avoid
The safe-lane message is not “anything goes.” A short, specific list of antidepressants can facilitate seizures at ordinary therapeutic doses — and they're the exceptions that prove the rule.
| Agent | Why it's on the list |
|---|---|
| Bupropion | Contraindicated in seizure disorder — dose-dependent risk, no safe lane here. (Full treatment in the Bupropion chapter.) |
| Clomipramine | The most proconvulsant TCA; risk is dose-related and present at therapeutic doses. |
| Amoxapine | Tetracyclic with a meaningful seizure signal, prominent in overdose but not confined to it. |
| Maprotiline | Tetracyclic with one of the highest seizure rates among antidepressants; avoid. |
Everything outside this list — the SSRIs, the SNRIs, mirtazapine, most TCAs at standard doses — carries a low, often negligible seizure risk at therapeutic dosing. The four above are the ones to keep off the prescription pad in a patient with epilepsy.
Which SSRI or SNRI
First-line is an SSRI or SNRI — the same as in any depressed patient. What narrows the choice in epilepsy is not seizure risk (they're all in the safe lane) but the interaction profile with antiseizure medications. On that axis the field splits cleanly.
| Choice | Interaction profile |
|---|---|
| Sertraline, escitalopram, citalopram | Preferred. Minimal (or, for sertraline, limited) pharmacokinetic interaction with ASMs — the cleanest choices to layer onto a seizure regimen. |
| Fluoxetine, paroxetine | Usable, but they inhibit CYP2D6 (and fluoxetine 3A4) — more interaction bookkeeping. |
| Fluvoxamine | Strong CYP1A2 inhibitor — raises clozapine and some ASM levels; usually not the first pick alongside seizure meds. |
| SNRIs (venlafaxine, duloxetine) | Also in the safe lane; reasonable when an SNRI is otherwise indicated. |
The default, then, is simple: sertraline or escitalopram for most patients, chosen because they're effective and stay out of the way of the seizure regimen. One caveat — citalopram (and to a lesser degree escitalopram) carries a dose-dependent QT signal, so respect the lower dose caps in older adults and hepatic impairment.
The Interaction Dimension
Even the “clean” SSRIs live in a two-way relationship with the seizure regimen — the same interaction engine covered earlier in the course, seen from the antidepressant side.
| Direction | What happens & what to do |
|---|---|
| Inducing ASM → your SSRI carbamazepine, phenytoin, phenobarbital, primidone | They raise clearance of sertraline, citalopram, escitalopram, and paroxetine — expect a lower level. Inducers can lift a substrate's clearance by 30–70%, so a meaningful dose increase is often needed for a full response. |
| Your SSRI → the ASM fluoxetine, fluvoxamine especially | These inhibit CYP and can raise ASM levels — a real concern with narrow-margin agents like carbamazepine and phenytoin, and added sedation with levetiracetam, lamotrigine, or phenytoin. |
| The inducer is stopped | Clearance falls, the SSRI level climbs — the de-induction window again. Reassess the dose whenever the ASM changes. |
The takeaway isn't complexity for its own sake: choose an interaction-clean SSRI and most of this disappears. It only becomes work when the antidepressant or the ASM is itself an enzyme meddler.
The Hyponatremia Stack
One combination deserves a specific flag. SSRIs cause hyponatremia through SIADH, most often in older adults — and so do oxcarbazepine and carbamazepine. Oxcarbazepine is the bigger offender of the two, with hyponatremia in a substantial minority of users and more severe drops than carbamazepine. Put an SSRI on top and the effects are additive. Worth knowing: among the SSRIs, escitalopram is not the lowest-risk for hyponatremia — so the preferred interaction profile doesn't buy a pass on checking sodium.
The practical move: in any patient on oxcarbazepine or carbamazepine — especially over 40, on polytherapy, or symptomatic (new confusion, unsteadiness, malaise, worsening seizures) — check a sodium before and after starting the SSRI, and again with dose increases. It's a cheap test that catches a common, correctable problem, and it's the single interaction most likely to actually bite in routine practice.
Starting One
The mechanics are reassuringly ordinary. Start at the usual dose, titrate to the usual targets — epilepsy doesn't call for a lower ceiling or a slower ramp on its own account. The two adjustments are agent selection (a clean SSRI) and interaction-aware dosing (anticipate the inducer effect). Counsel on the expected timeline as you would anywhere, watch for the rare early activation, and don't underdose out of residual seizure anxiety — underdosing is just undertreatment wearing a cautious face.
One genuine bonus worth mentioning to a hesitant patient: beyond being safe, SSRIs may modestly help seizure control. Small open-label and animal data suggest a possible antiseizure effect — not proven in randomized trials, so not a reason to prescribe on its own, but a reassuring counterweight to the outdated fear.
The Action Ladder
Depressed patient with epilepsy, no complicating factors
Start sertraline or escitalopram at the usual dose. Treat the depression as you would in anyone — the seizure disorder doesn't change the target.
Patient on an enzyme-inducing ASM
Same agent choice, but anticipate a lower SSRI level — inducers raise clearance ~30–70%, so be ready to push the dose up for a full response, and reassess if the ASM later changes.
Patient on oxcarbazepine or carbamazepine
Check a baseline sodium, recheck after starting and with increases. Watch for confusion, unsteadiness, or worsening seizures as hyponatremia clues.
Never reach for
Bupropion (contraindicated), clomipramine, amoxapine, or maprotiline as the antidepressant in a patient with epilepsy.
Clinical Pearls
Pearls
- Undertreatment is the real risk. Depression drives quality of life in epilepsy more than seizure frequency does — treat it.
- Placebo out-seized the drug. In the pooled trials, antidepressant arms had a lower seizure rate than placebo (SIR 0.48) — the threshold fear doesn't survive the clean data.
- Memorize the four: clomipramine, bupropion, amoxapine, maprotiline. Everything else in the safe lane at therapeutic doses.
- Sertraline or escitalopram is the default — effective and interaction-clean.
- Inducers lower your level. On carbamazepine or phenytoin, expect to dose the SSRI meaningfully higher — inducers raise clearance 30–70%.
- Watch the sodium when an SSRI meets oxcarbazepine or carbamazepine — additive SIADH, especially in older adults.
- Confounding by indication explains the scary observational headlines — the conditions ADs treat are themselves seizure risk factors.
Red Flags — Stop and Reassess
Stop and reassess
- Bupropion, clomipramine, amoxapine, or maprotiline written for a patient with epilepsy.
- Depression left untreated or underdosed out of seizure-threshold worry.
- An SSRI started on oxcarbazepine or carbamazepine with no sodium checked — especially in an older adult.
- New confusion, unsteadiness, malaise, or worsening seizures after starting an SSRI on a hyponatremia-prone ASM.
- Fluoxetine or fluvoxamine layered onto carbamazepine, phenytoin, or clozapine without accounting for the level rise.
- An SSRI dose left unchanged after an enzyme-inducing ASM is started or stopped.
Patient Counseling Script
Plain-language script
"A lot of people worry that antidepressants will trigger seizures — and for the medicine I'm recommending, the good evidence says that's not the case; if anything, treating the depression tends to help. Leaving depression untreated is the bigger risk to how you feel and function. We'll start at a normal dose. Because it works alongside your seizure medicine, I may need to adjust the amount, and if you're on certain seizure drugs I'll check your sodium with a blood test. Tell me if you feel newly confused, unsteady, or just off — and let me know before any of your other medicines change."
EMR / Documentation Template
This is one class. The course is all of them.
You've just seen how one drug class gets handled in epilepsy — the safe agents, the four to avoid, the interaction and sodium catches. The full course does the same for antipsychotics, stimulants, and benzodiazepines, plus the seizure-threshold rankings and the enzyme-interaction engine behind all of it. If this chapter earned its keep, the rest is waiting.
Prefer to keep reading first? Two member chapters that pick up directly from here:
References
- Alper K, Schwartz KA, Kolts RL, Khan A. Seizure incidence in psychopharmacological clinical trials: an analysis of FDA summary basis of approval reports. Biol Psychiatry. 2007;62(4):345–354 — antidepressant vs placebo seizure incidence, SIR 0.48 (95% CI 0.36–0.61).
- Kanner AM. Most antidepressant drugs are safe for patients with epilepsy at therapeutic doses: a review of the evidence. Epilepsy Behav. 2016;61:282–286 — the four proconvulsant agents; SSRI/SNRI first-line.
- Mula M, et al. Antidepressant drugs for seizures and epilepsy: where do we stand? — SSRIs/SNRIs first-line; citalopram/sertraline/escitalopram preferred by interaction profile; possible antiseizure signal.
- Johannessen Landmark C, Patsalos PN. Drug interactions involving antiseizure medications — enzyme-inducing ASMs raise antidepressant clearance (~30% dose adjustment).
- Berghuis B, et al. Carbamazepine- and oxcarbazepine-induced hyponatremia in people with epilepsy. Epilepsia. 2017;58(7):1227–1233 — hyponatremia frequency and risk factors.
- Lamotrigine / SSRI product labeling and StatPearls — SSRI + oxcarbazepine additive hyponatremia; citalopram QT dose caps.
Last reviewed July 2026. Part of the Psychiatry Education Forum Academy; for clinician education — it supports, and does not replace, individual clinical judgment and current local protocols. Sensitive topic note: depression and suicidality are discussed here in clinical terms; this is educational content, not individual medical advice.
Educational use only. Refer to the sources cited above and current prescribing information for clinical decisions. Psychiatry Education Forum and authors assume no liability for use of this material.
Responses