Uncontrolled Seizures and a Drug Reaction That Closed Off the Best Option
A DRESS reaction to carbamazepine has closed off the drug family that controls his seizures best — and every alternative in that family carries some version of the same risk.
Nate B., a 34-year-old man, was three weeks into a new job managing inventory for a regional distributor when a diffuse rash, high fever, and facial swelling sent him to the emergency department eight months ago — what turned out to be DRESS syndrome, triggered by carbamazepine started six weeks earlier for focal epilepsy that had been poorly controlled since a head injury five years ago. He was hospitalized for twelve days, treated with systemic corticosteroids, and required two additional weeks off work once he recovered enough to return; his liver enzymes, markedly elevated during the acute illness, have since normalized completely. Since stopping carbamazepine he has been maintained on levetiracetam alone, and his seizure frequency has climbed to roughly one breakthrough seizure a week — a level his neurologist considers both dangerous and inadequately controlled by any reasonable single-agent standard. He lost his driving privileges the week after his DRESS diagnosis, per state law on recent seizure activity, and has had two falls during breakthrough events since, one requiring stitches over his eyebrow — the kind of accumulating, ordinary-life cost that doesn't show up on a single lab value but that he brings up at every visit regardless.
Carbamazepine belongs to the aromatic anticonvulsant family, alongside phenytoin, oxcarbazepine, and lamotrigine, and cross-reactivity for severe cutaneous adverse reactions among drugs in this family is well documented — Hirsch and colleagues' retrospective series is the most-cited estimate, at roughly 40 to 60% among aromatic agents — high enough that a prior DRESS or SJS/TEN reaction to one aromatic agent is generally treated as a caution against the entire class, not just the specific drug involved. His neurologist has raised lamotrigine specifically, since it controls his seizure type well and its chemical structure differs from carbamazepine's more than oxcarbazepine's does — but lamotrigine carries its own well-established, largely titration-dependent SJS/TEN risk independent of any cross-reactivity question, a risk that rises sharply with rapid dose escalation regardless of a patient's carbamazepine history. No formal genetic testing has yet been done to inform which, if any, of these agents might be safer for him specifically — a gap his neurologist and allergist agree should close before any aromatic agent, however structurally distinct, goes back into consideration.
Choosing the next anticonvulsant after DRESS closed off the first one
Levetiracetam alone isn't controlling him, and weekly breakthrough seizures carry their own real risk — injury, status epilepticus, the accumulating cognitive and social cost of uncontrolled epilepsy. Lamotrigine controls his seizure type well, and its chemical structure is meaningfully different from carbamazepine's, more so than oxcarbazepine's is. With an appropriately slow titration, I think it's reasonable to try it rather than accept his current seizure burden indefinitely.
I'm not proposing this casually — I'm proposing it because the alternative on the table right now is a seizure frequency none of us would accept in a patient who hadn't already had a drug reaction.
I understand the seizure burden is real and dangerous, but the aromatic anticonvulsant cross-reactivity data don't distinguish as cleanly by structure as the argument for lamotrigine implies. Cross-reactivity for severe cutaneous reactions among aromatic anticonvulsants — carbamazepine, phenytoin, oxcarbazepine, and lamotrigine — runs around 40 to 60% in Hirsch's series, high enough that a prior DRESS reaction to one is treated as a caution against the whole family, not a specific exception carved out for whichever member differs most on paper. Before any aromatic agent goes back on the table, I want HLA-A*31:01 and HLA-B*15:02 testing done — HLA-B*15:02 from Chung's 2004 Taiwanese SJS/TEN association and HLA-A*31:01 from McCormack's 2011 work on carbamazepine hypersensitivity including DRESS, and while lamotrigine's own risk isn't identical to carbamazepine's, this is exactly the situation genetic testing exists to inform rather than guess through.
'A different structure' is true and worth naming, but it's a matter of degree within the same family the reaction already targeted once — it isn't the same as choosing a drug outside that family altogether.
I want to add the piece of this that's actually modifiable regardless of which side of the class question wins. Lamotrigine's SJS/TEN risk is separately, and heavily, tied to titration speed — rapid dose escalation is one of the best-established risk factors for lamotrigine-associated severe cutaneous reactions, independent of any carbamazepine cross-reactivity question. If lamotrigine is used here at all, an unusually slow titration, well below the standard schedule, is not optional caution, it's the single biggest lever we actually control. And I'd want the HLA testing back before starting anything in this family, not obtained in parallel with a first dose already given.
Agreed: HLA-A*31:01 and HLA-B*15:02 testing obtained before restarting anything in the aromatic anticonvulsant class; results pending. In the meantime, a non-aromatic second agent, valproate, added to his levetiracetam to reduce his current seizure burden without touching the disputed class at all.
Not agreed: whether lamotrigine should ultimately be added later if HLA testing comes back reassuring, which the neurologist still wants to pursue given how well it controls his seizure type, or whether a documented DRESS reaction to any aromatic agent should functionally close the whole class regardless of genetic testing, which the allergist leans toward. The pharmacologist's titration-speed point was adopted regardless of how that larger disagreement resolves: any future lamotrigine trial, if it happens, uses an extended titration schedule well beyond the standard one.