LGI1-Antibody Encephalitis: Tapering After a Strong First-Line Response
A textbook-strong response to first-line immunotherapy for LGI1-antibody encephalitis still leaves his memory only partially recovered — a residual deficit the literature ties directly to relapse risk, right as the team reaches the standard point to begin tapering.
Frank O., 67, cuts dovetails by hand — furniture-making being the exacting second trade he taught himself across six years of retirement, after four decades on a machine shop floor. Eight weeks ago he first noticed his right hand jerking while marking a cut — brief, repetitive, a couple of seconds each time — and shrugged it off as a muscle twitch, the same way he'd shrugged off worse from decades of shop work. Over the following two weeks he grew noticeably forgetful, repeating questions and losing track of measurements mid-project, and a workup for the memory complaints turned up a sodium critically low enough to prompt further testing. LGI1 antibody came back positive in serum. He was started on high-dose IV methylprednisolone plus a course of IVIG within days of the result; the facial and hand jerking — recognized by then as faciobrachial dystonic seizures — stopped within the first week, the kind of early, near-complete response the seizure-specific literature actually predicts once immunotherapy starts.
His memory is a different story. Six weeks out, he still loses track of conversations mid-sentence and cannot reliably recall what he did that morning, well short of his baseline. Thompson and colleagues’ cohort of 103 patients with this exact seizure type found that what protects cognition is stopping the jerking before impairment sets in: of the 58 patients who became cognitively impaired, 57 did so while their seizures were still active. Frank’s stopped within a week — but only once immunotherapy started, and his memory had already gone during the weeks of active jerking before anyone knew what they were treating. The fast response that reads like good news is not, on that data, the thing that would have protected him. And Campetella and colleagues’ more recent cohort found that residual cognitive dysfunction after the initial episode was itself associated with a markedly higher hazard of relapse. So the deficit those early weeks left behind is the one variable the literature flags twice over, which is precisely the tension the team now has to weigh against the standard taper schedule in front of them.
Six weeks in, at the start of the taper
His faciobrachial jerking stopped in the first week — exactly what the seizure-specific immunotherapy data predicts — but his memory has not come back the same way, and that gap is itself a real, named risk marker. Campetella and colleagues’ recent single-center cohort of 210 patients found that residual cognitive dysfunction after the initial episode carried a hazard ratio of 13.8 for relapse. I’d be honest about that number, though: the confidence interval runs from 1.5 to 129.5, so what it establishes is that the risk is genuinely up, not that it is fourteenfold. Their model also never tested seizure response as a variable, so I can’t tell you his fast FBDS cessation offsets it. I’d rather add rituximab now, while he’s doing well, than wait for that risk to become an actual second episode before we act on it.
You’re right that the cognitive-impairment marker is real — but Alkabie and Budhram’s series of patients managed on prolonged corticosteroids alone, without any maintenance immunotherapy at all, still reached good outcomes in most cases. Eighteen patients, so I won’t oversell it — but it is the closest thing we have to the comparison you’d want.
“Wait for that risk to become an actual episode” treats rituximab as a free option — it isn’t, not in a sixty-seven-year-old already six weeks into a steroid course; we’d be adding a second immunosuppressive drug’s infection and infusion risk to protect against an outcome that, on the same data you’re citing, still doesn’t happen to most people even without it.
The part neither of you has actually touched is the taper itself. Right now it’s set to run the standard course toward discontinuation over the next few months — that’s the one variable in front of us today that’s genuinely still open, and it’s less drastic than adding a new drug in either direction. Extend it, reassess his cognition explicitly at a set interval rather than by feel, and let that reassessment — not a decision made today on a six-week snapshot — be what actually decides whether rituximab gets added later.
Agreed: the oral corticosteroid taper is extended beyond its originally planned timeline, with an explicit cognitive reassessment scheduled at ten weeks rather than judged informally along the way; rituximab is held pending that reassessment rather than started today or ruled out.
Not agreed: the neurologist remains uneasy waiting given the real relapse-risk marker already in hand, and would have preferred adding rituximab now rather than at a later checkpoint; the geriatrician thinks even the extended taper is more caution than the data strictly requires given how well he has already done. Both positions were left standing rather than resolved into the plan.