Felbamate for Refractory Lennox-Gastaut Syndrome, Despite the Aplastic Anemia Risk
A single pediatric patient with Lennox-Gastaut syndrome who has failed five prior antiepileptic drugs. The remaining option that might actually help carries a real risk of aplastic anemia.
N.J., an 8-year-old boy, was diagnosed with Lennox-Gastaut syndrome at age three and has failed valproate, clobazam, lamotrigine, rufinamide, and cannabidiol, each tried at adequate doses over adequate trial periods, without meaningful reduction in his drop attacks — sudden loss of muscle tone that sends him to the floor ten to fifteen times most days, resulting in two emergency-department visits this year for facial lacerations and, six months ago, a fractured wrist. He wears a protective helmet during school hours, and his parents have reorganized their home, removing sharp furniture corners and carpeting hard floors, around managing the physical risk of his seizures rather than around anything resembling a typical eight-year-old's home.
Felbamate is a real option for exactly this situation — genuine efficacy in refractory LGS specifically, part of why it remains on the market at all despite a dramatic prescribing collapse in the mid-1990s after post-marketing surveillance identified aplastic anemia at rates far above the general population background, alongside a real, separate hepatotoxicity signal. Both are potentially fatal, both can develop with limited warning despite monitoring, and both are the specific reason felbamate now carries a boxed warning and is generally reserved for refractory cases exactly like N.J.'s, prescribed under informed consent after other reasonable options have genuinely failed.
Baseline and serial CBC and liver function monitoring are standard practice for any patient started on felbamate — real, recommended, and genuinely useful for catching some complications early. What monitoring cannot do — and what his parents need to understand plainly before consenting — is lower the underlying risk. The felbamate label states outright that routine blood testing cannot be relied upon to reduce the incidence of aplastic anemia; it may, in some cases, detect hematologic change before the syndrome declares itself clinically. Post-marketing estimates put that incidence somewhere near 1 in 4,000 to 1 in 10,000 exposures, roughly a hundredfold above background, with mortality approaching 30%. Aplastic anemia can also develop rapidly enough between scheduled checks to be caught too late. Monitoring shortens the interval to detection. It does not change the odds.
The informed-consent conversation, before the first dose
I think felbamate is the right next step, and I want to say that plainly rather than hedge it. He's failed five adequate trials of reasonable alternatives, his drop attacks are causing real, repeated physical injury, and felbamate has genuine, specific efficacy in refractory LGS — not a marginal option, one of the more effective agents left for exactly this syndrome once the more conventional choices have failed.
I'm not arguing against using it — in a case this refractory, with injuries this real, I think it's a defensible choice. But the aplastic anemia risk needs to be named without softening: post-marketing figures put it somewhere around 1 in 4,000 to 1 in 10,000 exposures, roughly a hundredfold above background, with mortality approaching 30%. And I want to correct something in how monitoring usually gets described — the felbamate label states outright that routine blood testing cannot be relied on to reduce the incidence of aplastic anemia. It sometimes catches hematologic change before the syndrome declares itself. That is a different and much weaker claim than risk reduction.
Describing felbamate as 'reserved for refractory cases under informed consent' can read as though monitoring resolves the risk — it doesn't resolve it, and by the label's own language it doesn't reliably reduce it either; at best it shortens the interval to detection. His parents deserve to hear that distinction stated directly, not implied.
Both of those are right, and the honest version of this conversation holds them together rather than picking one tone. Start felbamate with baseline labs already in hand and frequent CBC and liver-function monitoring thereafter — not because the schedule guarantees safety, and not because it lowers the underlying incidence, but because earlier detection is the only real tool available, explained to his parents as exactly that.
And give them a plain-language list of specific warning signs — unusual bruising, persistent fever, unexplained fatigue — to watch for between scheduled labs, since a family that knows what to look for is a real, if imperfect, second layer of monitoring on top of the scheduled bloodwork itself.
Agreed: felbamate started with baseline labs already obtained, weekly CBC for the first month and a defined ongoing interval after, and explicit, plain-language warning-sign education given to his parents before the first dose — the risk stated directly, not softened by the monitoring plan's existence.
Not agreed: how the informed-consent conversation itself should be framed for an 8-year-old who will, in time, need to understand his own treatment — the Epileptologist favored age-appropriate honesty now; the Hematologist preferred waiting until he's old enough to process the specific risk without added anxiety, with his parents as the primary consenting party for now. Left as an open parenting-and-care question, not a clinical one the team resolved.