Abstracts

Rationale: The memory component of the intracarotid amytal procedure (IAP) or Wada test was originally developed to prevent an amnestic outcome following unilateral anterior temporal lobectomy (ATL). In more recent years, the value of the IAP memory scores in predicting material specific memory outcome has become controversial. With the increasing potential of functional neuroimaging techniques to replace invasive diagnostic methods, the predictive validity of IAP memory data deserves further scrutiny. This study will determine the relative contribution of the IAP in the context of other standard presurgical variables in the prediction of verbal memory outcome following left ATL. Methods: The sample consisted of 25 patients with medically intractable epilepsy who were determined to be left language dominant on the IAP and subsequently underwent left ATL including resection of the hippocampus. All patients completed standard presurgical evaluations including long-term video-EEG monitoring, MRI, neuropsychological testing, and IAP. Predictors of verbal memory outcome were selected by review of existing literature and included the following presurgical variables: delayed recall of the Buschke Verbal Selective Reminding Test (VSRT), age of seizure onset (ASO), presence or absence of left hippocampal pathology on MRI (HPATH), and total percent correct IAP recognition memory score for the left injection (CIAP). Verbal memory outcome was determined by VSRT scores at 6-12 months post-op. The sample was classified as either unchanged (NC) or declined (DEC) based on regression-based norms and 90% reliable change index (RCI) for the delayed recall of the VSRT. Descriptive, non-parametric, and multivariate analyses were used to determine the strongest predictors of memory outcome following surgery. Results: In this sample, 28% showed clinically meaningful postsurgical verbal memory decline. Univariate analyses revealed statistically significant differences for ASO (p = .029), VSRT (p = .015), and HPATH (p = .045) between the two outcome groups, postoperatively. Follow-up stepwise logistic regression analysis showed baseline VSRT (p = .016) and then ASO (p = .029) to be significant predictors of verbal memory decline following ATL with an overall correct classification rate of 88% based on these two independent variables. Conclusions: The results of this study indicate that later age of seizure onset and higher presurgical verbal memory scores are the most important predictors of verbal memory decline following ATL. Although absence of left hippocampal pathology was significantly associated with postoperative memory declines, this variable did not account for any additional variance in the regression equation once ASO and VSRT were considered. Functional reserve of the right hippocampus as determined by left injection IAP total memory score was not found to be significantly related to verbal memory outcome by any analysis. These results may help to establish the value of the IAP in relation to non-invasive functional neuroimaging procedures currently under development.