Cortical Blood, No Seizure: Does She Need an AED Anyway
A retired kindergarten teacher's lobar hemorrhage touches the cortex, which is exactly the anatomy generations of training taught to fear seizures from. She hasn't had one, and the current guideline doesn't want her treated as if she has.
N.R., a 63-year-old retired kindergarten teacher, was pulling weeds when her husband noticed her left arm hanging strangely and her words coming out wrong. Imaging found a 22mL lobar hemorrhage in the left parietal lobe, cortical involvement in a pattern suggestive of cerebral amyloid angiopathy rather than the deep hypertensive bleeds more classically associated with that anatomy. She has had no witnessed convulsion, but her exam has fluctuated intermittently — brief unresponsive spells alternating with periods of following commands — fluctuation concerning enough on its own that continuous EEG monitoring was started, and concerning enough to raise, unprompted, the question generations of neurosurgical training answered the same reflexive way: cortical blood means seizure risk, and seizure risk means prophylaxis.
The 2022 AHA/ASA intracerebral hemorrhage guideline update states that reflex plainly, and directly against it: routine prophylactic antiepileptic therapy is not recommended, even with cortical involvement, because the retrospective and observational evidence behind the older cortical-blood teaching never demonstrated a genuine seizure-prevention benefit large enough to offset the drug's own cost — fatigue, mood effects, and expense accrued over a hospitalization and often well beyond it. What the same guideline does recommend, and what her fluctuating exam actually calls for, is continuous EEG monitoring specifically because it can distinguish a subclinical seizure from the encephalopathy, sedation, or simple metabolic fluctuation that produces an identical bedside picture — a targeted answer to exactly the uncertainty her exam has raised, rather than a blanket drug given because the anatomy alone once seemed to justify it.
The amyloid angiopathy pattern on her imaging is itself worth naming honestly to her husband, who has asked directly whether this bleed means she's at risk for another one: the lobar location and her age both fit the picture, a conversation the team has started having in parallel with today's narrower seizure question, since the two are related but not the same decision. Before this admission she lived independently, drove herself to the garden center three mornings a week, and had no history of any cognitive decline — the baseline against which her current fluctuating exam is actually being measured.
What the guideline actually says versus what training says
I trained on cortical involvement as a reason to start prophylaxis, and her exam is fluctuating in a way that looks seizure-like to me even without EEG confirmation yet. I understand the current guideline moved away from routine prophylaxis, but I'm not comfortable just watching an exam like this without treating.
I hear that, and the teaching you trained on wasn't unreasonable for its time — it just never held up against the actual outcome data once someone looked. The 2022 guideline reviewed exactly this question and recommends against routine prophylaxis even with cortical blood, because the seizure-prevention benefit was never large enough to outweigh the drug's own cost across a population. Her fluctuating exam is a real finding, but it's exactly what continuous EEG is designed to sort out — seizure versus encephalopathy versus something metabolic — rather than a reason to skip that step and treat blind.
Let's make the plan concrete rather than open-ended, since 'just monitor' can feel like doing nothing at the bedside. If continuous EEG captures any epileptiform activity, clinical or subclinical, levetiracetam starts immediately — that threshold is set now, not decided fresh in the moment. Until then, no AED, and her fluctuating exam gets worked up for other causes in parallel rather than assumed to be seizure by default.
Continuous EEG monitoring ran for 48 hours with no epileptiform activity captured; her fluctuating exam was ultimately attributed to intermittent oversedation from as-needed pain medication, resolved once dosing was adjusted. No AED was started.
The neurosurgeon's initial instinct wasn't borne out by the EEG data, but the team didn't treat that as a mark against the concern itself — the pre-agreed trigger meant the disagreement never had to be relitigated at the bedside once the actual cause of her fluctuation was found by other means. Continuous monitoring was credited directly with avoiding an antiepileptic course she never needed.