Abstracts

Rationale: Newer antiseizure medications (nASMs - approved after 1993) have similar efficacy in reducing seizure burden relative to older antiseizure medications (oASMs – carbamazepine, phenytoin, valproate, primidone, and phenobarbital), but are purported to have fewer side effects and improve quality of life (QoL). However, the degree to which use of nASMs rather than oASMs correlates with a higher QoL is largely unexplored. Methods: Using an Epic-based electronic medical record we analyzed patient-reported QoL in 3,257 patients seen at Dartmouth Hitchcock Medical Center between 2014 and 2019 (3.9 ± 3.7 visits per patient over 1.0 ± 1.4 years). QoL was assessed with a self-reported Liekert scale (“How would you rate your QoL,” from 0-10). Those patients with an ICD9 or ICD10 coded diagnosis of epilepsy taking ≥ 1 nASM or oASM (but not both nASM and oASM) were included. We collected information on number and type of antiseizure medication including benzodiazepine use; age; sex; insurance type; employment status; possession of an active driver’s license; diagnosis of depression, anxiety, PTSD, and personality disorder; psychogenic non-epileptic seizure diagnosis; Charlson comorbidity index; seizure frequency; whether seizures were self-reported as disabling; neurological disorders depression inventory for epilepsy (NDDIE) score; and patient perception of cognitive functioning per the cognitive sub-score of the quality of life in epilepsy scale (QOLIE-31), where higher scores indicate more complaints. Patients with complete data for at least one visit were included in the analysis (n=936 patients, number of visits ranged from 1 to 16). We determined the impact of these factors on patient reported QoL using ordinal mixed effects regression models to account for repeated sampling and time-dependent effects. We used stepwise elimination of parameters (with removal occurring for p≥0.2) to identify significant predictors. Results: Significant independent predictors of higher QoL were a comorbid diagnosis of anxiety (O.R. 0.421, 95% C.I. [0.307, 0.578], p<0.001, where an O.R. less than 1 indicates a detrimental association with QoL), disabling seizures (O.R. 0.686 [0.559, 0.842], p<0.001), age (O.R. 1.014 per year [1.006, 1.023], p=0.001), patient perception of cognitive functioning (O.R. 1.116 for each point,  [1.056, 1.178], p<0.001), being employed (O.R. 1.298 [1.026, 1.643], p=0.030), and having a driver’s license (O.R. 1.537[1.203, 1.964], p=0.001). NDDIE score was also a significant overall predictor of QoL (O.R. 0.643 for each point, 95% C.I. [0.596, 0.693], p<0.001), and modified the effect of taking a nASM as follows: taking a nASM was a significant predictor of QoL for patients with the lowest NDDIE score (O.R. 0.531, 95% C.I. [0.307, 0.917], p=0.023), but as NDDIE increased (suggesting more depression), the effect of taking a nASM attenuated by a factor of 1.074 for each point (95% C.I. [0.996, 1.159], p=0.064). Conclusions: In this large observational cohort, nASM use was significantly associated with lower self-reported QoL when compared to oASMs for those without significant depression, and there was no difference in subjects reporting many depressive symptoms. Prospective studies must confirm these findings given the possibility of selection bias inherent to observational studies. We also found that depression, anxiety, cognition, employment status, and driving autonomy influence QoL, confirming results from previous studies. This study supports the use of oASMs over nASMs given their lower cost, similar effectiveness as demonstrated by clinical trials, and apparent equivalence or superiority with regard to QoL. Funding: No funding