Surgical Treatment for Refractory Temporal ... - Wiley Online Library

Epilepsia, 47(8):1364–1372, 2006 Blackwell Publishing, Inc.  C 2006 International League Against Epilepsy

Surgical Treatment for Refractory Temporal Lobe Epilepsy in the Elderly: Seizure Outcome and Neuropsychological Sequels Compared with a Younger Cohort ∗ Athanasios Grivas, ∗ Johannes Schramm, ∗ Thomas Kral, ∗ Marec von Lehe, †Christoph Helmstaedter, †Christian E. Elger, and ∗ Hans Clusmann Departments of ∗ Neurosurgery and †Epileptology, University Bonn Medical Center, Bonn, Germany

Summary: Purpose: Surgical treatment of refractory temporal lobe epilepsy (TLE) is promising for selected patients, but only little experience has been acquired in operating on older patients, especially with limited resections. We intend to delineate clinical and surgical factors influencing outcome in patients older than 50 years at operation and to compare the results with those of a younger patient cohort. Methods: Fifty-two patients older than 50 years were operated on for intractable mesial or combined mesiolateral TLE between 1991 and 2002. The mean age at operation was 55 years, and the mean duration of epilepsy was 33 years. Forty selective amygdalohippocampectomies (33 for hippocampal sclerosis, seven for removal of a mesiotemporal lesion), five lateral temporal lesionectomies plus amygdalohippocampectomy, and seven anterior temporal lobectomies were performed. Eleven (21%) patients had undergone invasive presurgical video-EEG monitoring. The mean follow-up period was 33 months. We compared the results with those of a younger cohort operated on in the same time period. Results: Thirty-seven older patients attained complete seizure control (71% class I), and 10 patients had only rare postoperative seizures (19% class II). Four patients improved >75% (8%

class III), and one patient did not improve (2% class IV). The same rate of seizure control was attained by 11 patients older than 60 years at surgery. These results were not significantly different from those in a younger patient group. A trend toward better seizure control was noted in 16 patients with an epilepsy duration of 12 months (mean, 34 months) for seizure-outcome analyses. Complete categorized neuropsychological data sets were available for 34 patients from the older and for 359 patients from the younger group. Presurgical evaluation All patients were investigated according to the Bonn protocol for presurgical evaluation of refractory epilepsy (11,12) with long-term scalp video-EEG recordings, magnetic resonance imaging (MRI), and neuropsychological assessment. The evaluation of clinical variables included a history of febrile seizures, the age at onset, and the duration of epilepsy. Data were grouped for further analysis: age at onset 75% (13.3% class III), and 16 patients had no worthwhile improvement (5.0% class IV), resulting in 18.3% “unsatisfactory” seizure control. Epilepsia, Vol. 47, No. 8, 2006

FIG. 1. Seizure outcome with respect to the two complementary age groups: the study group with 52 patients aged 50 years or older at surgery and the control group with 321 patients younger than 50 years at operation. Seizure outcome is given in accordance with the four main classes of the Engel classification: class I, seizure free; class II, two or fewer seizures/year; class III, improvement ≥75%; class IV, improvement 60 yr Age at onset ≤20 yr >20 yr Epilepsy duration ≤30 yr >30 yr Febrile seizures With Without or unclear Seizure frequency ≤5/mo >5/mo Type of operation AH-HS AH-Les ATL LX+AH Side of operation Right Left Histopathology HS Develop. tumors Glial tumors Other diagnoses

52 22 30 52 41 11 49 24 25 49 16 33 49 4 45 49 20 29 52 33 7 7 5 52 29 23 52 31 3 4 14

Engel I, II

Engel III, IV

21 (95.5%) 26 (86.7%)

1 (4.5%) 4 (13.3%)

37 (90.2%) 10 (90.9%)

4 (9.8%) 1 (9.1%)

20 (83.3%) 24 (96%)

4 (16.7%) 1 (4%)

p Value 0.381 0.470 0.189 0.158

16 (100%) 28 (84.4%)

5 (10.2%)

3 (75%) 41 (91.1%)

1 (25%) 4 (8.9%)

0.070 20 (100%) 24 (82.8)

5 (17.2%) 0.847

30 (90.9%) 7 (100%) 6 (85.7%) 4 (80%)

3 (9.1%)

25 (86.2%) 22 (95.7%)

4 (13.8%) 1 (4.3%)

29 (93.5%) 3 (100%) 3 (75.0%) 12 (85.7%)

2 (6.5%)

1 (14.3%) 1 (20%) 0.368 0.548 1 (25%) 2 (14.3%)

AH-HS, Amygdalohippocampectomy for hippocampal sclerosis; AH-Les, amygdalohippocampectomy for mesiotemporal lesion; ATL, anterior temporal lobectomy; LX+AH, lesionectomy +amygdalohippocampectomy; HS, hippocampal sclerosis.

Seizure onset before 20 years was associated with a trend only to less-satisfactory seizure control (83% vs. 96%; p = 0.162) (Table 1). Seizure duration of 30 years attained 84.8% satisfactory seizure control (p = 0.158). Thus positive trends, but no significant differences, were found for later seizure onset and shorter duration of the epileptic disorder. Information on juvenile febrile seizures were provided by only four patients, so that statistical evaluation was not performed. Neuroimaging Forty-one (79%) MRI diagnoses were concordant with the histologic diagnoses; seven (13.5%) were unclear; and four (7.5%) were not correct. Satisfactory seizure outcomes were not significantly different, with 92.6% associated with correct MRI diagnoses, compared with 82.2% with unclear or wrong MRI diagnoses.

1367

Invasive procedures Presurgical invasive electroclinical investigation was performed in 11 (21.1%) patients. In seven of these cases, both hippocampal depth and subdural electrodes were implanted, whereas hippocampal electrodes alone were implanted in three cases, and subdural strip electrodes alone in one case. All patients who underwent invasive monitoring achieved a satisfactory seizure outcome. Resection types The majority of patients (45 of 52) underwent limited resections, and only few patients were thought to require two-thirds anterior lobectomies for combined mesiolateral TLEs. Seizure outcome was satisfactory in 90% of patients after AH for HS (n = 33, AH-HS), and in all patients with AHs for a mesiotemporal lesion (n = 7, AH-Les). Satisfactory seizure control was also attained in six (85.7%) of seven patients with anterior temporal lobectomies, and in four (80%) of five patients with temporal lesionectomy plus AH (n = 5, LX+AH patients). Results with respect to seizure control were not different with different resection types (p = 0.847). Complications Surgical complications were found in four (7.7%) patients: one medically treated pulmonary embolism in a patient with a medical history of deep venous thrombosis, one pneumonia, and two cases of early postoperative intracerebral hemorrhages, which had to be evacuated. One of these patients had a middle cerebral artery (MCA) infarction. No deep wound infections were found, and no deaths occurred in this series. Temporary deficits were observed in 10 (19.2%) patients: three (5.8%) cases with cranial nerve deficits and postoperative diplopia, which had resolved during a period of 12 months postoperatively: two (3.8%) cases with mild expressive dysphasia resolved within a few weeks; two (3.8%) patients developed mild hemiparesis as a result either of postoperative intracerebral hemorrhage or of a contralateral thalamic infarction in another patient. The clinical deficits resolved within months in both cases; and three (5.8%) patients showed a prolonged postoperative organic psychological syndrome, which lasted for 1–4 weeks. Persisting neurologic deficits were noted in two (3.8%) patients: one had a mild dysphasia and one had a mild dysphasia associated with a moderate hemiparesis as a result of a left MCA infarction and postoperative hemorrhage. Both patients with persisting deficits were able to take care of themselves, became ambulatory, and were independent. Three (5.9%) patients had a complete contralateral homonymous hemianopia. Two of these patients had undergone selective AH via a transcortical route. Epilepsia, Vol. 47, No. 8, 2006

1368

A. GRIVAS ET AL.

The analysis of the complications encountered for the younger control group revealed 0.7% permanent mild motor deficits or dysphasia and a 1.4% rate of complete hemianopia. One young patient died, and one remained in a vegetative state after AH, probably because of pseudohypoxic brain swelling (23). However, it must be mentioned that after completion of this study, a 70-year-old woman died of a severe postoperative hemorrhage. Taken together, the overall rate of relevant permanent neurologic sequelae, although mostly mild, was higher in the older compared with the younger patient group (p = 0.02). Histopathology A diagnosis of HS was established in 31 patients. Developmental tumors were found in three specimens [two dysembryoplastic neuroepithelial tumors (DNTs), and one ganglioglioma], other tumors in four cases [one oligodendroglioma, World Health Organization (WHO) grade II, one oligoastrocytoma WHO grade II, one astrocytoma WHO grade I, and one astrocytoma WHO grade III]. In 14 cases, other diagnoses were confirmed; of these five were cavernous hemangiomas, and eight cases with temporal gliosis. In the HS group, 22 (71.0%) patients were in class I, and seven (22.5%) patients were assigned to class II. All patients with developmental tumors showed complete seizure relief (class I). Within the glial tumor group, three of four patients reached class I (75%), and the residual other diagnoses were associated with 64.3% class I and 21.4% class II seizure relief. These differences are overall not significant (Table 1). Neuropsychological findings In 34 of 52 older patients, complete pre- and postoperative (3 and/or 12 months) neuropsychological data sets were available. In the younger patient group, 359 patients had complete neuropsychological data sets available and were thus included for comparison with the older patient group. In the older group and with regard to the factor “attention,” 19 (55.9%) patients achieved an average preoperative scoring (categories 2 and 3). On testing of verbal memory, only seven (20.6%) patients gained average results (categories 2 and 3), whereas in 27 (79.4%) cases, the results were below or far below average (categories 0 and 1). The preoperative visual memory testing revealed that 28 (82.4%) of the subjects scored below average (categories 0 and 1), and only six (17.6%) patients reached an average level (categories 2 and 3). None of the older patients attained above-average results for verbal and visual memory testing. For detailed results and the respective data in the younger control group, which showed a significantly better profile in all tested items, see Table 2. The postoperative neuropsychological changes are summarized in Table 3. In the older group, postoperative attentional performance improved significantly in 11 Epilepsia, Vol. 47, No. 8, 2006

(32.4%) patients, remained unchanged in 14 (41.2%) patients, and deteriorated significantly in nine (26.5%) patients. Younger patients show significantly stronger benefits after surgery regarding postoperative attentional performance (p = 0.016). With regard to verbal memory, postoperative improvement was observed in 20.6% of the older sample (seven patients) and deterioration in 29.4% (10 patients), whereas 50.0% (17 patients) showed no significant postoperative changes. Respective changes were similar in the younger control group (p = 0.680), but the different preoperative performance levels must be recognized. Visual memory function improved in 20.6% of the older sample (seven patients), remained unchanged in 47.1% (16 patients), and deteriorated in 32.4% (11 patients). Also in this respect, postoperative changes were not significantly different between the age groups, but with the lower baseline in older patients (see Table 2) kept in mind. In our older sample, we did not detect a significant effect of the side of surgery on the postoperative changes in memory performances (p = 0.78), which is putatively due to the relatively small number of older patients in the subgroups and the use of categorized neuropsychological data. In contrast, a significant impact of left-sided operations on verbal performance was found in the larger young control group, with 43.1% of the sample showing deterioration and 9.9% improvements, compared with right-sided procedures with 28.7% deterioration and 22.5% improvement (p = 0.001). The numbers of patients operated on with the different surgical approaches were too small to draw definite conclusions regarding neuropsychological postoperative effects, but we noted, although not significantly, that postoperative improvements in verbal and visual memory were hardly seen (n = 1) after temporal lobectomies, and generally in 20–27% after the more-limited approaches (NS). Subgroup analysis in patients older than 60 years Eleven patients (five men, six women) were older than 60 years at surgery (mean, 64; range, 60–71 years). Seizure outcome was not different from the complementary older patient group of 41 patients aged 50 to 60 years, with ∼90% satisfactory seizure control (nine Engel class I, one Engel class II). The mean age at onset was at 28 years (range, 1–70 years). The average epilepsy duration was 36 years (range, 1–62 years) in this subgroup. The only patient with unsatisfactory seizure control (Engel class III) was the one with seizure onset at the youngest age (during the first year of life, initially with febrile seizures) and the longest epilepsy duration of 62 years. In this oldest subgroup, four patients underwent AH for HS; three, AH for a mesiotemporal lesion; three, extended lesionectomies and AH; and one patient had a standard anterior temporal lobectomy. All but one patient had surgery on the right side. With only 11 patients in this group and all

TLE SURGERY IN THE ELDERLY

1369

TABLE 2. Categorized preoperative neuropsychological performances in the older (n = 34) and the younger patient group (n = 359) Below-average categories 0 + 1 Attention Older >50 yr >60 yr Younger Verbal memory Older >50 yr >60 yr Younger Visual memory Older >50 yr >60 yr Younger

Average categories 2 + 3

Above-average categories 4 + 5

13 (38.2%) 4 (50%) 84 (23.4%)

19 (55.9%) 4 (50%) 250 (69.7%)

2 (5.9%) − 25 (7.0%)

27 (79.4%) 4 (50%) 209 (58.3%)

7 (20.6%) 4 (50%) 140 (39.0%)

− − 10 (2.8%)

28 (82.4%) 6 (75%) 188 (52.4%)

6 (17.6%) 2 (25%) 162 (45.1%)

− − 9 (2.5%)

p Value 0.028

0.002

0.005

Data are provided for the subgroup of eight patients older than 60 years at surgery. The p value is used for comparison of older (>50 years) versus younger patient group. Note that all patients from the subgroup older than 60 years had right-sided TLE.

but one experiencing satisfactory seizure relief, a further correlation was not intended. No permanent morbidity occurred in this oldest subgroup; one patient showed a temporary deficit due to a cranial nerve palsy with diplopia. Another patient had a prolonged postoperative organic psychosyndrome. However, as mentioned earlier, in the later course, one 70-yearold patient died after a severe postoperative hemorrhage. Neuropsychological information was available for eight patients, all with surgery on the right. Preoperative performing was as expected, but obviously not strikingly different from the complementary older patient group. However, in the oldest group, none of the tested features scored above average. Postoperatively, results were surprisingly poor for verbal memory function, although the patients exclusively had surgery on the right. For detailed results, see Tables 3 and 4. An analysis for correlation of neuropsychological impairments and assessment of different aspects of quality of life was not performed.

DISCUSSION The management of elderly people with intractable epilepsy is thought to be more complicated because of alterations in physiology, drug absorption and elimination, the presence of concomitant diseases, and drug interactions with other medication (1,24). Temporal lobectomy in TLE has been suggested to be effective and reasonably safe in this age group, but experience is limited. So far, only sparse information is available on the postoperative effects with respect to cognitive outcome and suspected memory decline (3–5,25). In the present study, we investigated the effectiveness of surgical treatment in patients older than 50 years, with a particular focus on limited resections, seizure control, and

neuropsychological sequelae. An additional analysis was performed for a subgroup of patients older than 60 years at surgery. We compared the results to a control group of patients younger than 50 years, who underwent surgery in the same time period and with identical paradigms. Age groups Complete seizure relief was achieved in 71% of older patients. Together with 19% class II outcomes, this results in 90% “satisfactory” seizure control. This rate was not different between the two complementary older subgroups (>50 or >60 years). This rate is somewhat higher compared with outcome results from former studies, which vary between 65% and 83% in combined classes I and II outcomes (2–5). Comparing the seizure outcome to the control group of patients younger than 50 years at surgery, we found no difference in class I outcomes (71.2% and 72.0%, respectively). Some gradual differences were observed regarding class II, III, and IV outcomes, but these differences were not statistically significant (Fig. 1). Jensen (26) reported no significant differences between different age groups. An early report by Falconer (27) described a lower percentage of satisfactory results with increasing age. Similar findings were reported in other studies (3–5). Boling et al. (2) reported a trend that patients older than 50 years were less likely to achieve complete seizure relief with limited resections, which we did not confirm here, because complete seizure relief was merely identical in younger and older patients in the present study. Their results compare with our data with regard to the remarkable rate of class II outcomes (19.2% and 22%, respectively). A recent study by Janszky et al. (10) suggested that the negative predictors “age and epilepsy duration” may exert their effects only with respect to long-term outcomes of 3–5 years, which lay beyond the scope of the present study. Epilepsia, Vol. 47, No. 8, 2006

1370

A. GRIVAS ET AL. TABLE 3. Postoperative changes in neuropsychological performance in the older patient group (n = 34) and the control group with younger patients (n = 359)

Attention Older >50 yrs >60 yrs Younger Verbal Memory Older >50 yrs >60 yrs Younger Visual Memory Older >50 yrs >60 yrs Younger

Deterioration

State result

Improvement

9 (26.5%) 3 (37.5%) 36 (10.0%)

14 (41.2%) 2 (25%) 180 (50.1%)

11 (32.4%) 3 (37.5%) 143 (39.9%)

10 (29.4%) 6 (75%) 129 (36.0%)

17 (50%) 2 (25%) 172 (47.9%)

7 (20.6%) − 58 (16.2%)

11 (32.4%) 3 (37.5%) 101 (28.1%)

16 (47.1%) 5 (62.5%) 165 (46.0%)

7 (20.6%) − 93 (25.9%)

p value 0.016

0.680

0.761

Only significant individual changes (i.e., transit to another category) were considered. Additional data are provided for the subgroup of eight patients older than 60 years. The p value is for comparison of older (>50 years) versus younger patient groups. Note that all patients from the subgroup older than 60 years had right-sided TLE.

Clinical data Only four patients reported a history of febrile seizures. We believe this number is too small, but this kind of information may easily be lost during ≥5 decades. A trend of better outcome (96% “satisfactory” seizure control) was reached in those patients who had their onset of habitual seizures during adulthood compared with patients who experienced the first seizures during childhood or adolescence (83% “satisfactory” outcome). However, the difference was not significant. Similar findings were obtained in another study from our institution for the treatment of exclusively lesional mesial TLE (19). Early manifestation of seizures may be due to an overall increased seizure susceptibility or more widespread and potentially undetected structural and functional abnormalities. A seizure duration of 20 seizures per month were significantly less Epilepsia, Vol. 47, No. 8, 2006

likely to become seizure free after surgery. In a previous study, we described a steadier relation between increasing seizure frequency and poorer outcome (11). Taken together, seizure frequency and epilepsy duration, although not found to be significant in this study, seem to play some role with respect to seizure control, probably because of differences in individual seizure susceptibility. Operative strategy and histopathology Generally, resections were planned according to the presence and the extent of a temporal lobe lesion or evidence of hippocampal sclerosis, as demonstrated by preoperative MR investigation. Complete lesionectomies were intended in all cases. Resections were modified according to the electroclinical findings, obtained with video-EEG monitoring. In 11 patients, long-term implantation of electrodes was indicated, and all patients achieved satisfactory seizure relief. This strategy implies that unaffected temporal lobe tissue should be spared, if possible. Thus only seven patients underwent classic anterior temporal lobectomies. The change in paradigm away from lobar resections toward individualized limited resections has been described in detail before (11,19). The rationale is a potential cognitive benefit, but with identical rates of seizure control. We did not determine any statistically relevant differences in seizure control with different surgical approaches (Table 1). Developmental tumors (DNT and ganglioglioma) and HS were associated with best seizure control rates (100% and 93.5%, respectively), without significant differences to the other groups of diagnoses. This finding confirms previously published data of different subgroups (11,29–31). Complications Concerns on operating in the elderly also arise from the fact that surgical risks may be critically elevated. While

TLE SURGERY IN THE ELDERLY seizure control was excellent in patients older than 50 or 60 years in this study, complications occurred significantly more frequently in the older patient group, compared with a younger control group. Permanent morbidity occurred in two patients who developed dysphasia, and one of whom had a moderate hemiparesis due to MCA infarction after intraoperative bleeding from the vessel, and constitute 3.8% of the operated-on cases in the older group. However, no mortality or severe permanent neurologic morbidity was found, as all patients became ambulatory and were able to care for themselves. It is interesting to note that no permanent morbidity occurred in the subgroup older than 60 years. However, as mentioned earlier, one 70-year-old patient died postoperatively after completion of this study. The overall incidence of complications in the younger group was significantly lower. The complication rates encountered are comparable to those of other reports showing permanent speech disturbances after temporal lobe surgery of between 0 and 5% (22,25,32). Polkey (33) estimated a risk of hemiparesis in