Seizure Medication Linked to Longer Survival in Diffuse Midline Glioma

A medication that children with brain tumors often already take for seizures was linked to longer survival in those with diffuse midline glioma, a group of tumors that form in the brain stem, thalamus or spinal cord, according to research published in Nature Medicine and led by investigators at Stanford Medicine. The drug, Keppra (levetiracetam), appeared to work by cutting the electrical connections that these tumors form with healthy nerve cells, and the same benefit did not show up in children whose tumors began elsewhere in the brain.

What did the Keppra study find in children with diffuse midline glioma?

Researchers reviewed records from 218 children treated for high-grade glioma at Stanford University between 1990 and 2020 and at the University of Michigan between 2012 and 2021. Among the 119 children in that group who had diffuse midline glioma, the 15 who had taken Keppra lived a median of 20.96 months, compared with 9.92 months among the 104 who had not.

The pattern did not hold for children whose high-grade gliomas started in the cerebral hemispheres rather than in midline structures. In that group, median survival was 24.0 months for the 40 children who had taken Keppra and 22.2 months for the 59 who had not, a difference that could have come about by chance.

Two additional groups of patients showed the same trend in diffuse midline glioma. At University Medical Center Hamburg-Eppendorf in Germany, six children who took Keppra lived a median of 24.0 months versus 11.0 months for 19 children who did not. At the University of California, San Francisco, five children who took Keppra lived a median of 15.5 months versus 12.2 months for 30 children who did not.

“This is very exciting for a disease that has had so few treatment options,” said Dr. Michelle Monje, the Milan Gambhir Professor in Pediatric Neuro-Oncology and a professor of neurology at Stanford Medicine.

Why did Keppra affect diffuse midline glioma and not other brain tumors?

Diffuse midline glioma cells build working connections, called synapses, with nearby healthy nerve cells and use the signals that travel across them to keep growing. A connection, known as a GABAergic synapse, is found in diffuse midline glioma but not in the hemispheric tumors the team studied, which is the difference the researchers traced the drug's effect to.

In mice carrying tumors grown from patient samples, Keppra slowed growth and extended survival in diffuse midline glioma but had no effect on proliferation in three separate models of hemispheric high-grade glioma. Recordings from individual tumor cells showed that Keppra weakened the GABAergic signals reaching those cells while leaving the same signals in healthy nerve cells unchanged.

The drug did not appear to act through the route that makes it work against seizures. Briviact (brivaracetam), which targets that route more precisely, had no effect on the tumor cells, and the drug still reduced tumor growth when the researchers disabled that route in the laboratory. Two other seizure medications, Zarontin (ethosuximide) and Dilantin (phenytoin), did not slow tumor growth at all.

What did the diffuse midline glioma research measure?

The clinical portion of the work measured overall survival, meaning how long patients lived, and compared children who had a history of Keppra use with those who had none. Tumor diagnoses were confirmed by a board-certified neuropathologist in cases where tissue was available. The patient groups predated the arrival of CAR-T cell therapy for these tumors.

In the animal experiments, mice received Keppra five days a week for four weeks, a schedule chosen to mirror how the drug is given to patients. Researchers tracked survival, tumor burden, how far the tumors spread and how quickly the tumor cells divided.

What are the limitations of the Keppra findings in diffuse midline glioma?

The clinical data came from records collected after the fact rather than from a trial designed to test the drug, and the number of children who had taken Keppra was small at every site. The researchers wrote that "prospective clinical trials are required to validate this conclusion."

They also raised the possibility that the children who received Keppra did so because they had seizures, and that tumors causing seizures may be the ones most responsive to the drug in the first place. Tumor location could matter as well, since the brain stem, thalamus and spinal cord carry more of the GABAergic input the drug appears to interrupt. An unrecognized factor in the patient records could account for part of the result.

Monje's team is launching a clinical trial of Keppra in patients with these tumors, and her laboratory is separately testing CAR-T cells in children with diffuse intrinsic pontine glioma and other diffuse midline gliomas.

References

1. Barron T, et al. "Levetiracetam therapeutically targets GABAergic synapses in diffuse midline glioma." Nature Medicine. Published: Sept. 17, 2026.
2. "Anti-seizure drug fights aggressive childhood brain tumors, Stanford Medicine-led study shows." Stanford Medicine. Posted: Sept. 17, 2026.