Comparative effectiveness of VNS versus continued pharmacologic treatment of Lennox-Gastaut Syndrome: reducing emergency seizure care and improving functional outcomes


Lennox Gastaut Syndrome (LGS) is a rare developmental-epileptic encephalopathy (DEE), with recurrent disabling seizures refractory to medical treatment. Seizure onset is typically in the first two years of life, but the syndrome is often not fully recognized until children are 3-5 years-old. LGS constitutes 3-4% of all childhood-onset epilepsy and 1-2% of all epilepsy. In the United States, an estimated 13,400 children under the age of 18 and 34,300 adults 18 years and older have LGS, for a total of 48,000 people affected by this disorder. LGS is characterized by life-long refractory seizures, severe disabilities, developmental delay, cognitive impairments, and prolonged seizures that frequently necessitate the use of rescue medications and hospitalizations. Young people with LGS are 10 times as likely as healthy peers to die both from seizure-related and non-seizure related causes.6 There is no entirely effective treatment or cure. Treatments are generally palliative and aimed at reducing the number of severe seizures (prolonged or likely to cause injury). These strategies include the use of multiple anti-seizure medications, dietary therapies, and Vagus Nerve Stimulators (VNS). 

Study Aims:

(1) Using the PEDSnet data, we will identify children in the network who carry the diagnosis of LGS per their ICD-10 code (G40.81x) and construct retrospective cohort to be followed from the time of the first ICD-10 LGS code beginning in Oct 1, 2015 to June 1 2021 (see diagnosis validation below). Medical and surgical cohorts of pediatric patients with LGS will be identified based on published algorithms of medical intractability criteria, and further matched by propensity score for analysis.

(2) For each LGS patient, VNS and medical, we will define the index date for comparison, and determine the seizure-related ED/PICU utilization the year prior to the index date (baseline), number of rescue medication prescriptions, number of new medications that were prescribed, and number of outpatient, ED, inpatient, ICU visits for both medical and surgical cohorts. We will determine the same outcomes measures for the 6 months after surgery (recovery and adjustment period) and the 12 months beyond that (comparison period) for both medical and surgical cohorts. Simple analyses will examine the before-after differences in VNS versus non-VNS patients using a Wilcoxon Rank-Sum test. Multivariable logistic regression with adjustment for hospital site as a random variable will allow adjustment for patient age, medications, and other factors.

(3) To test hypotheses regarding evaluation and access to surgical VNS therapy and earlier versus later use of VNS, we will perform a cross-sectional study at 4 sites, Lurie, Seattle, Colorado, and Nationwide. Several parent-reported measures already trialed by one of the investigators (ATB) in this population and acknowledged by parents as appropriate and relevant to the level of their children's function will be outcome measures. Independent variables will be use of VNS, age at VNS, duration of VNS, etiology, and seizure burden.

(4) We will perform two collaborative projects with providers and with parents to enhance shared decision-making surrounding VNS. Each will utilize a mixed methods approach (qualitative to quantitative). Focus groups with (a) parents at the 4 sites and (b) physicians at those sites will identify important themes surround the use or avoidance of VNS. For each group, based on the focus group data, systematic surveys will be designed and implemented through LGSF (for parents) and through professional societies (for physicians).