Abstracts

Rationale: The accurate diagnosis of Lennox-Gastaut Syndrome (LGS) at any stage of life is complex, and misdiagnosis can result in suboptimal treatment. To improve both identification and treatment of patients with LGS, the Refractory Epilepsy Screening Tool for LGS (REST-LGS) was developed by a working group of experts who care for patients with refractory epilepsy and intellectual and developmental disabilities. The inter-rater reliability (IRR) of the REST-LGS was measured as a first step in evaluating the validity of the tool for use in a clinical setting, and results suggest potential combinations of major and minor diagnostic criteria that are applicable among patients with LGS and drug-resistant epilepsy (DRE). Methods: Using de-identified records of patients >12 years with refractory epilepsy and ≥2 neurology clinic notes within the previous 2 years, potential major and minor diagnostic criteria were evaluated on individual case report forms (CRFs) by two raters at two large, diverse epilepsy centers: a specialist (epileptologist/epilepsy nurse practitioner) and a non-specialist (nurse, social worker, resident, or pre-medical student). A diagnosis of definite LGS required 4 major criteria to be met. IRR at each site was measured by Cohen’s κ coefficient. Upon data being unblinded (diagnosis of LGS or DRE revealed), a number of common major and minor diagnostic criteria on the CRF were determined by each patient at both sites. Results: Of 200 patient records evaluated (100 at each site), most patients (81% to 85%) met 1–3 major criteria. Gender distribution at each site was comparable, but patients from Site 1 were significantly younger than those from Site 2 (mean age at last visit ± SD: 24±12 vs 41±12 years; P< 0.0001). At Site 1, moderate agreement (κ = 0.40 to 0.60) was reached between the two raters’ judgments on whether patient records reflected persistent seizures despite trial of ≥2 AEDs; at Site 2, both raters agreed that all patients had persistent seizures. Moderate agreement to good agreement (κ = 0.40 to 0.80) was reached among raters at each site on the following variables: ≥2 seizure types; seizure onset ≤12 years; history of EEG with generalized SSW discharges   Conclusions: Overall, the first step in validation of the REST-LGS revealed moderate to very good agreement on criteria indicative of LGS among raters at two epilepsy centers. Results suggest the combined number of major and minor criteria that may be particularly indicative of a patient with LGS. Taken together, the REST-LGS may prove to be a valuable tool for both expert and non-expert in identifying patients who require further diagnostic evaluation for potential LGS. Funding: Lundbeck