Objective: To develop a multivariate risk factor model for predicting postoperative verbal memory decline in an individual patient following dominant or nondominant anterior temporal lobectomy (ATL).
Methods: The authors studied 132 consecutive ATL patients who 1). were older than 16 years at surgery, 2). had estimated preoperative Full Scale IQ score of >69, 3) had unilateral language dominance based on the intracarotid amobarbital procedure (IAP), and 4) underwent neuropsychological testing at baseline and >or=6 months postoperatively (mean 1.2 years). Five potential risk factors for postoperative verbal memory decline were selected a priori that reflect the functional adequacy of the to-be-resected temporal lobe. These were 1). resection in the dominant hemisphere, 2). MRI findings other than exclusively unilateral mesial temporal sclerosis, intact preoperative 3). immediate and 4). delayed verbal memory function, and 5). intact IAP memory performance following injection contralateral to the seizure focus. Verbal memory decline was defined using two verbal memory tests and published reliable change indices.
Results: Thirty-eight percent of the sample declined reliably on one or both verbal memory measures. Logistic regression analysis demonstrated that all five risk factors were significantly and independently associated with outcome, with side of surgery having the strongest association (p < 0.0001) and preoperative immediate verbal memory the weakest (p < 0.05).
Conclusions: An individual patient's risk for postoperative verbal memory decline following dominant or nondominant ATL can be predicted using clinical data routinely available preoperatively (side of surgery, qualitative MRI, baseline memory testing, IAP performance). This information may be useful for preoperative patient counseling.