Abstracts

Rationale: Topiramate (TPM), a broad-spectrum anti-seizure drug, is one of many commonly prescribed medications that can cause severe cognitive impairments. The frequency of these deficits has been reported to be as high as 44% (Lee et al., 2006). Although the incidence of side effects appears to be related to age and degree of drug exposure (Marino et al., 2012), these factors do not account for all variability in patient susceptibility to TPM’s negative impact. The aims of the present study were (i) to test the hypothesis that variability in working memory capacity (WMC) contributes to variation in responses to TPM, and (ii) to investigate the relationship among WMC, cognitive impairment and TPM plasma levels. Methods: 29 healthy participants completed a randomized, double blind, crossover study of TPM, placebo (PBO), and lorazepam (LZP), which served as a comparator drug. Three parallel groups of subjects received a single dose of TPM (100, 150, or 200mg), LZP (2mg), and PBO at three separate sessions, then completed a modified Sternberg working memory task 4 hours after drug administration. Strings of 1, 3, or 5 nonsense syllables were presented for 1.5 sec and, after a 3 sec retention interval, subjects were asked to indicate whether a probe stimulus matched the previously presented string. Reaction time (RT) and accuracy (ACC) data were recorded, and the extent of drug-induced impairment was calculated by comparing each treatment to PBO ((drug-PBO)/PBO), thus normalizing across individual accuracy rates. WMC was assessed by calculating averaged Cowan’s k scores (Cowan, 2001) from data collected during the PBO session; TPM plasma levels were measured using a liquid chromatography/mass spectrometry assay (Subramanian et al., 2006). Bonferroni-adjusted post-hoc t-tests were used to compare across treatments. Pearson’s correlations were calculated to examine relationships between performance and WMC, and performance and plasma levels. Results: TPM and LZP decreased ACC for all memory loads compared to PBO (all p < .005). LZP also increased RT compared to TPM and PBO (all p < .001). The magnitude of TPM-induced ACC declines was correlated with WMC, with high capacity subjects showing more severe performance deficits (r=-.46,p=.01). The size of these effects was also correlated with TPM plasma levels (r=-.52,p=.005), and a median split on WMC showed that these effects were driven by high capacity subjects (r=-.51, p=.047).  Conclusions: We found that, while LZP-induced deficits seem to be a result of generalized cognitive slowing, TPM appears to directly impact the working memory system. Individuals with high WMC are particularly at risk at high plasma concentrations, challenging the assumption that efficient cognitive function confers an advantage in overcoming drug-induced deficits. These results support the hypothesis that variability in WMC plays a role in patient responses to TPM, and suggest that WMC is a clinically relevant variable that should be taken into account when starting patients on TPM treatment regimens. Funding: This research was funded by R01 NS076665 (PI: Marino)