EEG requirements may not capture full LGS spectrum
Patients have differences in seizure patterns, brain activity
Written by |
LGS patients show differences in disease presentation. (Photo by iStock)
People who meet the clinical criteria for Lennox-Gastaut syndrome (LGS) may have substantial differences in their seizure patterns, brain activity, and developmental impairment, and some severely affected patients may not meet all formal electroencephalogram (EEG) criteria used to define the condition, a study suggested.
Among 105 people with clinically defined LGS, 65.7% met the complete clinical and EEG criteria established by the International League Against Epilepsy (ILAE). Yet those who did not meet all of the EEG requirements were generally similar in developmental and motor impairment, seizure burden, brain MRI findings, and treatment resistance.
A separate, data-driven analysis identified three partially overlapping LGS profiles. The researchers said these profiles may help capture differences across the LGS spectrum more effectively and inform clinical trial design.
The study, “Refining diagnostic boundaries and electroclinical profiles of Lennox–Gastaut syndrome through unsupervised clustering,” was published in the journal Epilepsia.
EEG features can change over time
LGS typically begins in childhood. LGS symptoms include several types of seizures, intellectual disability, and characteristic patterns of abnormal electrical activity in the brain, as measured by EEG.
Current diagnostic criteria, as supported by the ILAE, combine clinical and EEG characteristics. The EEG criteria require evidence of two characteristic patterns of abnormal brain activity called slow spike-and-wave discharges (SSWD) and generalized paroxysmal fast activity (GPFA).
But these EEG features can change over time. SSWD, for example, may disappear in adulthood. Sleep EEG recordings, which are essential to detect GPFA, may be unavailable in routine clinical practice, particularly for older patients. As a result, people who otherwise have a typical LGS presentation may not have documented evidence of both EEG abnormalities.
Researchers in Italy set out to determine how often people with clinically defined LGS meet the complete ILAE clinical and EEG criteria, and whether an unbiased analysis of clinical, EEG, and brain-imaging data could identify distinct profiles among patients with clinically defined LGS.
They performed a retrospective analysis of patients followed at two epilepsy centers in Rome between January 2020 and July 2025. All met accepted clinical criteria for LGS, had long-term clinical information, and at least one sleep EEG available for review.
More than half (60.9%) of the 105 patients were women, with a median age of 24.
Overall, 69 of the 105 patients (65.7%) had documented evidence of both SSWD and GPFA and therefore met the complete electroclinical definition of LGS. The other 36 met the clinical criteria, but not all EEG requirements.
The two groups did not significantly differ in age at seizure onset, degree of intellectual and motor impairment, MRI abnormalities, or seizure burden.
Researchers then used a data-analysis method known as unsupervised clustering to try to identify distinct clinical profiles among LGS patients. The analysis involved 97 adolescents and adults and considered seizure types, EEG findings, developmental and motor impairment, MRI results, and other clinical characteristics.
The analysis identified three clusters.
Cluster 1, comprising 41 people, represented the most severely impaired group. Seizures began at a median age of 5 months, 95.1% of patients had severe intellectual disability, and 85.4% had severe motor impairment. Yet only 43.9% of patients in this cluster met the complete electroclinical LGS criteria.
Cluster 2, consisting of 42 patients, also represented a severe, early-onset form of LGS, with a high prevalence of infantile seizure onset (88.1%) and intellectual disability (92.9%), although severe motor impairment was less frequent (34.1%). In this cluster, 73.8% of patients met the complete electroclinical LGS criteria.
Cluster 3, comprising 14 patients, represented a later-onset LGS profile. Severe intellectual disability was less common (35.7%), and none of the patients had severe motor impairment. Nearly all patients in this cluster (92.9%) met the complete electroclinical LGS criteria.
The researchers stressed that the three clinical profiles are exploratory and should not be considered distinct biological subtypes of LGS. Still, the findings suggest that the clinical diversity of LGS may not be fully captured by rigid EEG-based definitions.
“In this context, the identification of different phenotypic profiles within the LGS spectrum may be beneficial for the design, implementation, and interpretation of randomized clinical trials in this rare patient population,” the researchers wrote.
