Changes in cognition after introduction or withdrawal of zonisamide versus topiramate in epilepsy patients: A retrospective study using Bayes statistics

Objective We aim to evaluate the impact of zonisamide (ZNS) compared to topiramate (TPM) on cognition in patients with epilepsy. Although the risk of cognitive side effects has been clearly demonstrated for TPM, comparable side effects in ZNS have been suggested but evidence from studies is inconclusive. Methods In this retrospective observational study, we analyzed patients' records from before and after introduction or withdrawal of ZNS vs TPM. Data were gathered during routine clinical care protocols. Standardized monitoring of executive functions (EpiTrack), verbal memory (short version of verbaler lern- und merkfahigkeitstest, VLMT), and subjective health (extended Adverse Events Profile; quality of life in epilepsy inventory, QOLIE-10) was performed in 73 patients when TPM (n = 45) or ZNS (n = 28) was introduced and 62 patients when TPM (n = 29) or ZNS (n = 33) was withdrawn. The data were analyzed using Bayes statistics that quantify evidence for or against an effect through Bayes factors (BFs). Results There was decisive evidence for a negative effect of adjunctive ZNS and TPM on executive function (BF = 965.08) and a positive effect of their withdrawal (BF = 429.51). The ZNS effect seemed smaller, although the difference was inconclusive. Verbal memory and subjective quality of life were not significantly affected. Subjectively, ZNS was connected to lower anxiety and fewer headaches, whereas TPM had a perceived effect on weight, fluent speech and comprehension, headaches, and balance. Significance This is the first study to provide objective evidence for a considerable negative effect of ZNS treatment on executive function in a naturalistic treatment setting. Comparable to the well-known TPM effect, cognition worsens with adjunction and recovers with withdrawal of ZNS. However, the majority of patients do not show a significant negative effect, suggesting disparate susceptibilities to adverse events. The findings emphasize the need for routine monitoring of cognitive side effects to identify early on those patients who are negatively affected by new AED.