Defining cognitive impairment in temporal lobe epilepsy with IC-CoDE phenotypes: A comparison of normative and ipsative approaches.

OBJECTIVE

The International Classification of Cognitive Disorders in Epilepsy (IC-CoDE) is used to classify cognitive phenotypes in individuals with epilepsy. This study compared the relative utility of ipsative versus normative approaches of IC-CoDE classification.

METHODS

1,010 patients with temporal lobe epilepsy (TLE) were classified into cognitive phenotypes (intact, single-domain, bi-domain, or generalized impairment) using both the established normative method and a novel ipsative method (cognitive impairment defined relative to premorbid IQ). Agreement between methods was quantified using quadratic weighted kappa (95% CI). Exploratory univariable ordinal logistic regression models compared associations between demographic/clinical variables and generalized impairment (vs. all others) across the two classification methods.

RESULTS

Agreement was fair (weighted Kappa = 0.22), and the ipsative method classified more patients as cognitively intact. Using the normative method, patients were more likely to be classified as having generalized impairment if they were younger (OR = 0.99; p = 0.017), had lower education (OR = 0.8; p < 0.001), were Black (OR = 2.62; p < 0.001), Hispanic/Latinx (OR = 2.15; p = 0.059), had longer epilepsy duration (OR = 1.01; p = 0.004), mesial temporal sclerosis (OR = 1.54; p = 0.003), an abnormal MRI (one lobe/hemisphere OR = 1.32, p = 0.046; bilateral OR = 2.00, p = 0.001), were taking more anti-seizure medications (OR = 1.48; p < 0.001), and did not have a tumor (OR = 0.57; p = 0.037) or vascular malformation (OR = 0.43; p = 0.041). Using the ipsative method, patients were more likely to be classified into the generalized impairment phenotype if they were younger (OR = 0.99; p = 0.045), had higher education (OR = 1.16; p < 0.001) or were Hispanic/Latinx (OR = 2.35; p = 0.044).

SIGNIFICANCE: The normative method may better reflect epilepsy severity factors in group studies, while the ipsative method may reduce misclassification in Black individuals. The utility of each method is likely dependent on the clinical or research context.