Choosing an AED Around a Hemodialysis Schedule, Not Just Around Renal Clearance
A single patient with new-onset epilepsy who is also on hemodialysis three times a week. Renal clearance is the obvious variable — dialyzability and dosing logistics turn out to matter just as much.
F.M., a 67-year-old woman, has end-stage renal disease from long-standing diabetic nephropathy and has been on hemodialysis three times a week, Monday, Wednesday, and Friday, for the past two years. She had a witnessed focal-to-bilateral seizure at her dialysis center last week, her first ever, and workup found no acute structural cause — age-related and vascular changes on MRI consistent with her known cardiovascular risk factors, nothing acute. She lives with her daughter, who drives her to every dialysis session and has become, in her own words, the family's default medical coordinator.
The obvious variable in choosing her first AED is renal clearance, and on that axis alone, levetiracetam looks straightforward — primarily renally cleared, requiring a reduced maintenance dose in ESRD, which is a manageable, well-established adjustment. But renal clearance isn't the only thing hemodialysis changes about a drug's behavior. Levetiracetam is also significantly removed by dialysis itself, water-soluble and minimally protein-bound, meaning a meaningful fraction of her dose is filtered out during each session — which means either a supplemental post-dialysis dose or accepting subtherapeutic levels on dialysis days specifically, a real, recurring three-times-a-week complication that a once-daily dosing mental model doesn't naturally anticipate.
Valproate, by contrast, is hepatically metabolized and highly protein-bound, which makes it far less affected by dialysis directly — but ESRD itself alters protein binding, increasing the free (active) fraction of valproate at any given total drug level, which complicates interpreting a standard total valproate level the way it would be interpreted in a patient with normal renal function.
Choosing a first drug around the dialysis schedule
On renal clearance alone, levetiracetam is the straightforward choice — roughly two-thirds is excreted unchanged in urine, it isn't protein bound, and the ESRD dose adjustment is well established. But renal clearance isn't the whole picture: levetiracetam is roughly 50% dialyzable, and its half-life drops to around three hours during a session against about 25 hours between them. The Keppra labeling and the standard renal dosing references (Aronoff et al., 2007) both handle this the same way — 500 to 1,000mg every 24 hours plus a 250 to 500mg supplemental dose after each session. Her Monday, Wednesday, Friday schedule means that supplement has to actually happen three times a week, or the regimen is subtherapeutic by design.
You're right that the dialysis-removal problem is real, not a minor dosing footnote, and I don't want to wave it off. But valproate isn't a clean alternative just because it avoids that specific problem. Valproate needs no dose adjustment in renal impairment at all — that's the part that looks attractive — but reduced protein binding in uremia raises its free fraction, and the standard references say plainly that monitoring total valproate concentrations alone can be misleading in these patients. A level can read reassuringly therapeutic while her active, unbound drug is running higher than the number suggests.
Presenting levetiracetam's dialysis-removal issue as the deciding problem understates that valproate trades one monitoring complication for a different one — not clearance itself, but whether the number we're checking still means what it normally means.
Both problems are real and both are manageable, which is actually the more useful way to frame this than picking the drug with the fewer complications on paper. Levetiracetam with a structured, written post-dialysis supplemental-dose protocol — given directly to her, her daughter, and the dialysis center's own nursing staff — solves the dialyzability problem concretely, since the fix is a scheduling and communication fix, not a pharmacologic one.
Valproate's free-fraction issue is harder to solve cleanly, since it requires either free-level monitoring most labs don't run routinely or accepting real interpretive uncertainty from a standard total level — that's a genuine, ongoing ambiguity, not a one-time logistics fix. Given the choice between a solvable coordination problem and a persistent interpretive one, I'd start with levetiracetam and the supplemental-dose protocol.
Agreed: levetiracetam started at an ESRD-adjusted maintenance dose, with a written supplemental post-dialysis dosing protocol distributed to F.M., her daughter, and the dialysis center directly — the actual mechanism for making the three-times-weekly complication reliably manageable rather than a recurring guess.
Not agreed: whether valproate should be held in explicit reserve as a named second-line option if levetiracetam's psychiatric/behavioral profile proves poorly tolerated in her, given her age, or whether that decision should wait until an actual problem, if any, appears. Left open rather than pre-decided.