Reconsidering Bupropion's Contraindication After Sustained Remission from Bulimia
Bupropion carries a flat contraindication in bulimia nervosa, current or past, because of a seizure signal found decades ago in actively purging patients. This man has been in confirmed remission for eight months. Whether the original mechanism still applies to him — or whether the label was never really about him in the first place — is the whole disagreement.
D.H., a 34-year-old man, has coached his son's Little League team for three seasons now, a commitment he says has mattered more to his recovery than almost anything on his actual treatment plan, since it gave him somewhere to be every Tuesday and Thursday evening that wasn't alone with his own thoughts. He manages a hardware store his father used to own, work he describes as steady but not something he particularly enjoys, and has struggled with depressive episodes on and off since his early twenties — untreated for most of that time, in his own words, because "guys in my family don't really do therapy." Bulimia nervosa entered the picture roughly three years ago, developing quietly alongside a depressive episode he never separately sought help for, until his wife found the pattern and pushed him into treatment. He has now been in confirmed remission — documented abstinence from binge-purge episodes, corroborated by his therapist and his own consistent self-report — for eight months.
His depression, though, has not moved the way his eating disorder has. He remains moderately symptomatic on sertraline despite an adequate trial, and has asked directly about bupropion, partly because a friend described real benefit from it and partly because he's still smoking half a pack a day and knows bupropion is approved for cessation too — a second problem he'd genuinely like solved alongside the first. The obstacle is that bupropion carries a flat contraindication for any current or past diagnosis of bulimia or anorexia nervosa, a warning that traces to a specific finding: Horne and colleagues, 1988, found an unacceptably high seizure rate — roughly four seizures among 55 bulimic patients treated with bupropion, far above the drug's baseline rate in the general population — in a trial population that was, at the time, actively purging. The mechanism proposed was never that bulimia itself lowers seizure threshold as a fixed trait; it was that active purging produces electrolyte disturbances — hypokalemia, metabolic alkalosis — that independently lower seizure threshold, and bupropion's own seizure risk is known to be dose- and titration-dependent on top of that. D.H.'s electrolytes today are unremarkable, and he has not purged in eight months. The label, though, is written to a diagnosis rather than to a physiology — "current or prior," not "currently purging" — precisely because the behavior it guards against is the kind patients under-report, sometimes to themselves. His labs describe the mechanism as absent today; the wording was chosen by people who did not trust today's labs to settle the question.
Medication review, eight months into remission
I don't think we should prescribe this, even given eight good months. The label says "current or prior diagnosis of bulimia or anorexia nervosa" — not "active purging" — and that wording is deliberate. Eating-disorder relapse is well documented to be under-reported, sometimes even to a patient's own awareness of what's happening. We have other antidepressants and other smoking- cessation options that don't carry a seizure signal at all. The asymmetry matters: if we're wrong about how solid this remission is, the downside is a seizure, not a missed opportunity.
I take the detection-uncertainty point seriously, but I want to be precise about where the original signal actually came from. Horne and colleagues found roughly four seizures among 55 bulimic patients on bupropion — far above expected background rate — but that trial population was actively purging at the time. The mechanism proposed was never "bulimia as a diagnosis lowers seizure threshold." It was that active purging causes real, measurable electrolyte disturbance — hypokalemia, metabolic alkalosis — which independently lowers seizure threshold, on top of bupropion's own known dose- and titration-dependent seizure risk. D.H.'s potassium and bicarbonate are normal today, corroborated by his therapist, not just self-reported. The mechanism that produced the original warning isn't present in his current physiology.
I'd also note he's not asking us to guess at his remission — his therapist is independently confirming it, which is a real check the original 1988 trial population didn't have.
I'm not going to argue the mechanism reasoning is wrong — the electrolyte story is real, and his labs back it up. What I'd flag instead is timing. His PHQ-9 is 15 despite an adequate sertraline trial — he's not euthymic right now, he's in an active depressive episode, and depression itself is a documented relapse risk factor for bulimia independent of anything to do with seizures. Adding a new variable to his regimen during a vulnerable mood period isn't the same decision as adding it to a fully recovered, currently stable patient, even if the seizure-specific reasoning checks out completely.
I don't have a clean answer here, and I want to say that directly rather than manufacture false confidence. I'd lean toward first optimizing his depression treatment through another route — augmentation or a different antidepressant class — and revisiting bupropion specifically once his mood is more stable, rather than treating tonight as the decision point either way.
Agreed: hold bupropion for now, without ruling it out permanently. Augment his current sertraline with buspirone as an interim step for his active depressive episode, offer nicotine replacement therapy for his smoking goal in the meantime, and revisit the bupropion question explicitly once his mood has stabilized.
Not agreed, and left genuinely open rather than resolved: whether the mechanism-specific reasoning (normal electrolytes, corroborated remission) will be considered sufficient to prescribe bupropion once his depression improves, or whether the diagnosis-based label language should govern regardless of how stable he becomes. The Attending Psychiatrist believes the physiological case will only get stronger with more time in remission; the Clinical Pharmacologist maintains the label's conservative wording exists for a reason that doesn't fully go away with any amount of documented stability. D.H. was told directly that this remains unresolved and that the team wants to revisit it with him once his mood is in a better place, not that the door has been closed.