Surgery, implantable devices offer new hope for drug-resistant epilepsy

Editor's Note

Surgery and advanced implantable devices can give patients with drug-resistant epilepsy a far greater chance of seizure freedom than medication alone, UCSF News September 25 reports. While about 30 anti-seizure medications exist, one-third of patients do not achieve seizure control, and fewer than one in five seek care at comprehensive epilepsy centers that provide specialized diagnostics and treatments.

As detailed in the article, a significant share of patients who fail medication may not actually have epilepsy. Up to 25% of cases referred to specialized centers turn out to be psychogenic non-epileptic spells—psychological episodes often linked to trauma. These patients should not be treated with seizure drugs, which frequently cause dizziness, fatigue, incoordination, and cognitive fog.

For true drug-resistant epilepsy, surgery is often the most effective option. After multiple failed medications, the likelihood of becoming seizure-free with further drug therapy falls below 1%, compared with up to 80% with surgery. Even so, many patients still require medication after surgery, though often at lower doses. Surgical strategies vary depending on seizure type and location. Patients with focal seizures may undergo open procedures to remove seizure-producing tissue or minimally invasive laser ablation for hard-to-reach areas.

For others, implantable devices provide alternatives. Responsive neurostimulation (RNS) detects seizures at onset and delivers targeted electrical pulses, while deep brain stimulation reduces abnormal activity through intermittent impulses. Vagus nerve stimulation, the only device approved for children, sends seizure-blocking signals from a chest-implanted battery through the vagus nerve to the brain and avoids direct brain surgery.

Research is pushing the field further. Clinical trials are testing RNS in generalized epilepsy and Lennox-Gastaut syndrome, and UCSF is developing forecasting tools using RNS data to predict seizure likelihood within 24 hours, delivered via a smartphone app.