SEATTLE—Anterior temporal lobectomy is known to be an effective treatment for many patients who have medically refractory epilepsy. However, one risk associated with this procedure is memory decline. “It has been estimated in a variety of studies that anywhere from 25% to 75% of left temporal lobectomy patients and 22% to 62% of right temporal lobectomy patients show a memory decline after surgery,” said Tara Lineweaver, PhD, of the Cleveland Clinic Foundation. Dr. Lineweaver and colleagues are conducting a study to examine the relative contributions made by neuropsychological assessment, magnetic resonance imaging (MRI) hippocampal volumetric analysis, and Wada test results in predicting memory outcomes after anterior temporal lobectomy using the most current measures and methods in these areas. The research is supported by a grant awarded by the Epilepsy Foundation of America. Dr. Lineweaver reported the preliminary results at the American Epilepsy Society’s 2002 Annual Meeting.

MODELS OF MEMORY OUTCOME

Based on the results of various studies, there are two models used to explain memory outcomes after temporal lobectomy, Dr. Lineweaver said. One is the cognitive reserve model, which proposes that the functional state of the mesial temporal lobe structures of the contralateral hemisphere (opposite the side of the resection) determines memory outcome. Patients with good functioning in these structures would be expected to perform better after surgery in terms of memory, whereas patients with poor functioning would be expected to have a poor outcome.

An alternative model is the functional adequacy model. “This posits that it is really how well the to-be-resected structures are functioning that determines memory outcome. So if you resect mesial temporal lobe structures that are functioning well because they’re epileptogenic, then you’d expect a poor memory outcome after surgery because you’re removing structures that were previously mediating memory. However, if the structures you resect aren’t functioning very well prior to surgery, then you wouldn’t expect to see much postsurgical change in memory.”

PINPOINTING POTENTIAL PREDICTORS

Dr. Lineweaver and colleagues identified 117 patients at the Cleveland Clinic who had undergone temporal lobectomy and had both baseline and follow-up neuropsychological testing available. “To date, 61 have had MRI volumetric analysis done. We selected only patients who were left hemisphere dominant for speech because we had only five patients who were right hemisphere dominant, and we lost another five patients because they had only unilateral Wada data instead of bilateral Wada data,” she reported. Of the resulting cohort of 51 patients, 78% were seizure free after surgery, and on average, patients underwent neuropsychological testing about four months prior to and six months after surgery.

The 51 patients were split into two groups: good memory outcome and poor memory outcome, using the 15th percentile as the cutoff score for determining group membership. “Once we split the patients into two groups, we conducted a series of exploratory logistic regression analyses to predict memory outcome on each of the four memory measures of the Wechsler Memory Scale–Third Edition,” elaborated Dr. Lineweaver. The group also looked at what percentage of patients was correctly classified once the prediction equations were applied.

Potential predictors included side of surgery, preoperative neuropsychological test scores (absolute test scores and discrepancy between visual and auditory memory); and volumetric analysis from MRI. When coding MRI results, “the first thing we looked for was asymmetric hippocampi,” Dr. Lineweaver said. “If patients had asymmetrical hippocampi, we classified them as having either ipsilateral or contralateral atrophy. If they did not demonstrate an asymmetry, then we classified them as having both ipsilateral and contralateral atrophy if both hippocampi were significantly small relative to normative standards, or as having neither if both of their hippocampi were within the normal range.”

Wada test results, including percent correct after ipsilateral injection and percent correct after contralateral injection, as well as an asymmetry score, were also considered as potential predictors, “because research has shown that can be a very sensitive indicator of memory outcome,” she added.

SOME MEMORABLE RESULTS

Side of surgery was not a significant predictor of memory outcome in any of the analyses. “This was a very surprising result based on the current literature, so we followed this up with some Chi-square analyses. We found a trend for a higher proportion of left temporal lobectomy patients rather than right temporal lobectomy patients to show a decline in auditory memory, but it did not reach significance—though that may be due to the small number of patients we had,” Dr. Lineweaver said. There was also no noticeable difference for left versus right anterior temporal lobectomy patients in visual memory outcome, she added.

What did predict memory outcome was contralateral atrophy on MRI. “Those patients who had contralateral atrophy on MRI tended to have a worse auditory memory outcome,” Dr. Lineweaver said. Also, in terms of the Wada test, the percent correct after ipsilateral injection was a significant predictor of outcome on three of the four memory test variables. “If patients had good memory performance after ipsilateral injection, they tended to show less decline in terms of memory test scores after surgery. If they had poor memory after these ipsilateral injections, suggesting that the contralateral hemisphere wasn’t functioning very well, then they were more likely to show a decline.”

Finally, neuropsychological memory test performance significantly predicted visual memory outcome. Those who had good visual memory at baseline were more likely to show a decline in visual memory, and those who had poor visual memory before they even went to surgery were less likely to show a decline. Taken together, “these factors improved prediction anywhere from 4% to 18% once MRI, neuropsychological, and Wada variables were taken into account. The results suggest that the Wada test does contribute to the prediction of memory outcome above and beyond side of surgery, MRI, and neuropsychological factors. That’s not what we expected to find, and it will be interesting to see, as we continue to increase our sample size, whether this will be a significant result or not,” Dr. Lineweaver observed.

However, MRI and neuropsychological factors also improved the prediction of memory outcome, she noted. “These results support both the cognitive reserve model in terms of the Wada and MRI results we found, and they support the functional adequacy model in terms of the neuropsychological testing measures predicting outcome.” She added that the group’s findings also suggest that results from some of the previous studies using older versions of memory tests may not necessarily generalize to those in current use.

LOOKING BACK, LOOKING AHEAD

In conclusion, Dr. Lineweaver remarked on several of the limitations associated with her study. “One is the small sample, and we’re working to correct that,” she said. “It’s also a biased sample, in that these patients are seen in a highly specialized tertiary care center at the Cleveland Clinic, and we see fairly complicated patients.” In addition, the test results were only for patients who did go through surgery, and many of the factors used to predict memory outcome were also used to select patients as candidates for surgery. Finally, “what we deemed an abnormal change in terms of psychometric measures may or may not correspond to a problem in the patient’s everyday life,” she said. “We are working on various things to try to address these problems.”

Among the prospective directions are continuing data collection for the present study, in order to increase the statistical power, and determining memory outcomes over a one- or two-year follow-up to judge the long-term effects of temporal lobectomy on memory.

Additionally, Dr. Lineweaver and her group are formally assessing subjective memory to determine whether there is a correlation between their statistical evidence of memory decline and experiential memory problems as perceived by the patients. “Another aspect to examine is seizure impact on memory over time,” she added in an interview. “Surgery affects memory outcome, but it might be that seizures themselves have a long-term effect on memory, and perhaps in a worse way.”

NR

—C. Justin Romano

Return to table of contents