Long-term Prognosis and Psychosocial Outcomes after Surgery for MTLE

Authors


Address correspondence and reprint requests to Dr. S. Dupont at Unité d'Epileptologie Clinique Neurologique Paul Castaigne, Hôpital de la Salpêtrière, 47, boulevard de l'Hôpital, 75651 Paris cedex 13, France. E-mail: sophie.dupont@psl.ap-hop-paris.fr

Abstract

Summary: Purpose: To assess the seizure-freedom rates and self-perceived psychosocial changes associated with the long-term outcome of epilepsy surgery in patients with refractory medial temporal lobe epilepsy associated with hippocampal sclerosis.

Methods: A standard questionnaire was given to 183 patients who underwent surgery between 1988 and 2004, and 110 were completed.

Results: The mean duration of follow-up after surgery was 7 years, with a maximum of 17 years. The probability that patients were seizure-free after surgery was dependent on the definition of the seizure freedom. For the patients who were seizure-free since surgery (Engel's class Ia), the probability was 97.6% at 1 year after surgery, 85.2% at 2 years after surgery, 59.5% at 5 years after surgery, and 42.6% at 10 years after surgery. For the patients who still experienced rare disabling seizures after surgery but were seizure-free at least 1 year before the time of assessment, the probability was of 97.6% at 1 year after surgery, 95% at 2 years after surgery, 82.8% at 5 years after surgery, and 71.1% at 10 years after surgery.

The psychosocial long-term outcome, as measured by indices of driving, employment, familial and social relationships, and marital status, was similar to the psychosocial short-term outcome. It did not depend on seizure freedom or on follow-up time interval and was not influenced statistically by seizure frequency in cases of persisting seizures. Most but not all patients noticed a substantial overall improvement in their psychosocial condition; 48% drove (increased by 7%), 47% improved (14% worsened) in their employment status, and 68% improved (5% worsened) in their familial and social relationships. Overall, 91% of patients were satisfied with the surgery, and 92% did not regret their decision.

Conclusions: The results of this study suggest that temporal lobe surgery has real long-term benefits. Two specific conclusions emerge: (a) the long-term rates of freedom from seizure depend on how seizure freedom is defined, and (b) the psychosocial long-term outcome does not change dramatically over years and does not depend on seizure freedom.

Surgery is an effective treatment for medial temporal lobe epilepsy (MTLE) associated with hippocampal sclerosis (Engel, 1996). Studies on the short-term efficacy of surgery for MTLE have found that 60–85% of patients experience no seizures during the first year after surgery (Wiebe et al., 2001; Lowe et al., 2004). Two recent studies (McIntosh et al., 2004; Tellez-Zenteno et al., 2005) suggest that the long-term seizure-free rate after temporal lobe resective surgery is close to that reported in short-term studies. However, two points should be clarified in studies on the long-term effects of epilepsy surgery: (a) what is an appropriate definition of the long-term seizure-free rate? Some studies consider only patients who have never had a seizure since surgery, whereas others also include patients who are seizure-free at the time of evaluation; and (b) what are the long-term effects and the impact of surgery on social function of patients who have undergone surgery? Many patients with refractory epilepsy feel stigmatized by their condition, and this affects their professional, social, and marital status. Results of the few studies that have evaluated the long-term impact of epilepsy suggest that psychosocial problems tend to persist in later life of patients who have experienced the onset of epilepsy at school age (Shackleton et al., 2003), but that, compared with medical management, surgery may affect psychosocial outcomes positively (Jones et al., 2002).

In this study, we examined the long-term prognosis and the psychosocial outcomes in a consecutive and homogeneous series of patients who underwent surgery for refractory MTLE associated with hippocampal sclerosis. We focused on two measures: (a) seizure freedom, which is a relevant and consistently reported outcome criterion; and (b) the follow-up time interval, which is rarely considered in outcome studies. We evaluated seizure freedom and psychosocial outcome at extended periods of follow-up. We wished to determine whether different definitions of freedom from seizures—either no seizure since surgery or no seizure during the year preceding assessment and follow-up—and the frequency of persistent seizures significantly altered psychological and social indices of patients' lives, including their freedom to drive and to maintain a job as well as perceived social and marital status.

METHODS

Population

The study group was identified from the Epilepsy Data Base of the Epilepsy Unit of the Pitié-Salpêtrière Hospital in Paris. The target population consisted of all patients with refractory temporal lobe epilepsy associated with hippocampal sclerosis who underwent surgery between 1988 and 2004. All patients had received a similar presurgical evaluation (Adam et al., 1996), including medical, neurologic, and neuropsychological examinations, video-EEG monitoring, brain MRI, and, in some cases, FDG-PET examinations.

Criteria for inclusion criteria were (a) hippocampal sclerosis diagnosed by structural magnetic resonance imaging (MRI), (b) electroclinical presentation consistent with medial temporal lobe seizure onset, (c) surgical procedure consisting of either an anterior temporal lobectomy (ATL) or a selective amygdalohippocampectomy (SAH), and (d) postoperative follow-up of at least 1 year. We evaluated the second postoperative period when patients underwent two operations to control persistent seizures.

Collection of data

A standardized questionnaire was sent by post to all the patients fulfilling the selection criteria. A cover letter was added to the questionnaire, informing the patients of the study requirements before enrolling and giving them the option of electing not to participate.

The self-completion questionnaire asked for demographic data, epilepsy-related variables on postoperative status (seizure freedom, time of disappearance of seizures in relation to surgery, time to relapse, persistent seizures (type, description, frequency, timing), eventual reduction or discontinuation of antiepileptic drugs), and for patients' perceptions of the consequences of surgery on their daily lives (driving license and professional, familial, social, and marital status). When possible, missing data were completed from patients' medical records. Telephone contacts were made with nonresponding patients to enhance the proportion of responses.

Classification of seizure outcome

Seizure freedom outcome was assessed at the last follow-up with at least 1 year elapsing before the final evaluation and according to Engel's classification (Engel, 1987).

The first subgroup, Engel's class Ia, consisted of patients who reported no seizures after their surgery. The second subgroup, Engel's class I, included both seizure-free patients and those who have experienced simple partial seizures, or “brief auras” and “neighborhood” seizures and drug-withdrawal seizures. The third subgroup, Engel's class II patients, included patients who were not seizure-free but had a substantial improvement, exhibiting still only rare seizures. The fourth subgroup, Engel's class III–IV patients, included patients with frequent seizures and a truly unsatisfactory outcome.

We separated four categories of patients according to

  • – the seizure freedom with two definitions:
    • ○ patients who were completely seizure-free after surgery (Engel's class Ia patients)
    • ○ those that had been free from seizures for at least 1 year at the time of assessment
  • – the persistence of seizures and the importance of the reduction in seizure frequency:
    • ○ patients who had rare seizures (i.e., Engel's class II patients)
    • ○ patients with frequent and disabling seizures (i.e., Engel's class III–IV patients).

Statistical analysis

Patients were divided into subgroups according to seizure outcome and according to follow-up delay (1–5, 6–10, and 11–17 years).

Survival curves derived from Kaplan–Meier estimates were used to derive the probability of seizure-free survival, the probability to remain in outcome class I after surgery, and the probability that seizures would cease in patients who experienced seizures in the immediate postoperative period. Seizure-free survival and survival in class I was estimated for patients who were seizure-free in the immediate postoperative period.

Associations between postoperative outcomes and antiepileptic regimen or professional familial, social, and marital status were tested by χ2 or Fisher's exact tests. For each variable, three comparisons were done: class Ia versus Engel's class II, class Ia versus class III–IV, and class II versus class III–IV.

Some patients who were seizure-free at least 1 year before the time of assessment, belonged either to class Ia, to class II, or, in one case, to class III–IV patients; thus no comparison was possible between these groups. The Bonferroni method was used to address the multiple testing problem. Significance was established as p < 0.017. All analyses were performed with the SAS software version 8.2 (SAS Institute, Cary, NC, U.S.A.).

RESULTS

Demographic data

In our center, 183 patients underwent surgery for refractory TLE associated with hippocampal sclerosis between 1988 and 2004. Three of them were deceased at the time of this study. Although the cause of death was unknown in all cases, a sudden unexpected death in epilepsy (SUDEP) was suspected in two patients. The postal address and telephone number were incorrect for 25 patients.

Forty-five patients did not return the questionnaire after two phone calls: two patients refused to participate to the study because of lack of time or lack of desire to recall epilepsy, five patients consented to participate but did not return the questionnaire, and 38 patients were not directly contacted despite two phone calls and messages. The socioeconomic characteristics of these 45 patients (18 women/27 men; mean age, 38 ± 7.3 years; 18 right TLE and 27 left TLE; 35 anterior temporal lobectomies and 10 selective amygdalohippocampectomies; 2.44 ± 0.8 AEDs at the time of surgery; professional status, 29 employed patients, eight students, eight unemployed patients) did not differ from those that returned the questionnaire.

Thus 110 patients completed the questionnaire: 54 women and 56 men, with a mean age of 42 ± 9 years. The mean age at the time of surgery was 35 ± 10 years; the age at onset was 11 ± 9 years; and the mean duration of epilepsy was 24 ± 11 years. Fifty patients had left MTLE, and 60 had right MTLE. Twenty-two patients underwent a selective amygdalohippocampectomy (SAH), and 88 patients underwent a standard anterior temporal lobectomy (ATL). The mean duration of follow-up after surgery was 7 ± 4 years (range, 1–17 years). One patient underwent a second operation to control persistent seizures. In this case, we evaluated the second postoperative period only.

Forty-seven (43%) patients had a follow-up of between 1 and 5 years, 43 (39%) patients had a follow-up of between 6 and 10 years, and 20 (18%) patients had a follow-up of between 11 and 17 years.

Seizure outcome

At the time of assessment (June 2005), 53 (48%) had not experienced a seizure after surgery, and so corresponded to Engel's class Ia. Of a total of 110 patients, 78 (71%) corresponded to Engel's class I, 21 (19%) to Engel's class II, eight (7%) patients to Engel's class III, and three (3%) patients to Engel's class IV (Fig. 1).

Figure 1.

Engel's classification of patients.

Eighty-one (73%) patients had been seizure free at least 1 year before the time of assessment.

The year-to-year cross-sectional analysis of surgical outcome by using Engel's classification is presented in Fig. 2. From the total of 110 patients, 83 (75%) fit the criteria of Engel's class Ia at 1 year after surgery, 73 (66%) at 2 years after operation, 34 (54%) at 5 years after surgery, and 11 (41%) patients at 10 years after surgery. The probability of seizure-free (Ia) survival after surgery was 97.6% at 1 year, 85.2% at 2 years, 59.5% at 5 years, and 42.6% at 10 years after surgery. Classifying according to the criteria of Engel's class I patients gave relatively similar results over the first postoperative years in terms of probability of seizure-free survival after surgery—97.6% at 1 year and 95% at 2 years after surgery—but tended to remain stable over time—82.8% at 5 years and 71.1% at 10 years after surgery.

Figure 2.

Seizure-free survival since surgery.

The side of resection and the type of resection (SAH vs. ATL) were not predictive of outcome.

Seizure recurrence

Seizure semiology was often modified in patients who experienced seizures after surgery. In 57 patients who relapsed, semiology was altered in 42 cases, consisting exclusively of generalized tonic–clonic seizures in 22 (39%), exclusively as nocturnal seizures in 15 (26%), and of both (exclusive nocturnal generalized tonic–clonic seizures) in 12 (21%) patients.

In 37 patients, seizures recurred within 2 years after surgery, and 20 patients experienced a late seizure recurrence at 30–108 months after surgery. The frequency of seizures was significantly lower in patients with a late recurrence than in those with an early recurrence of seizures (p < 0.025) (Table 1).

Table 1. Outcome of late versus early relapsers
Time to relapse (yr)Moderate recurrence Sz/mo <1Severe recurrence Sz/mo ≥1p-Value
  1. Sz, seizure; Sz/mo, number of seizures per month after first relapse; moderate recurrence, a frequency of seizures <1/mo; severe recurrence, a frequency of seizure ≥1/mo.

≤2 yr (early sz recurrence)23140.025
>2 yr (late sz recurrence)182 

Seizure cessation

In contrast, 12 of 26 patients who experienced seizures in the immediate postoperative period noted a cessation of their seizures over time (Fig. 3). The median time until cessation of seizures was 6.2 years: at 6.2 years, 50% of patients who were not immediately seizure free noted a cessation of their seizures with this delay after the operation.

Figure 3.

Probability of seizure cessation over time.

Discontinuation of antiepileptic drugs

At the Salpêtrière Epilepsy Unit, patients take AEDs for at least 1 year after surgery. Reduction toward a cessation of AEDs is tailored to each individual; all patients progress through an extended period of AED tapering before medication ceases.

The mean number of AEDs prescribed was 2.4 ± 0.8 before surgery and 1.47 ± 1.1 after surgery. Patients often wished to discontinue AEDs as a reason for undergoing surgery (83 patients or 75.5% of our sample).

Discontinuation of AEDs according to the postoperative outcome

Of the 110 patients, 20% no longer take AEDs. All were seizure-free at least 1 year before assessment, and 18 had been seizure free since surgery (Ia).

Of those patients who continue to take AEDs, 35 (66% of the Ia subgroup) were seizure-free since surgery, and 59 (73% of the seizure-free since evaluation subgroup) were seizure-free during at least 1 year preceding assessment, and 32 (100% of the persistent seizures subgroup) had persistent seizures since surgery either with a rare frequency (class II) or with a higher frequency (classes III–IV).

AED doses had been reduced for 41% of the Ia patient group, for 43% of the seizure-free patients, and for 34% of patients with rare or frequent persistent seizures (38.1% of the rare-seizures subgroup and 27% of the frequent-seizures subgroup). Dosage was increased in seven patients (Table 2). In seven other patients, a relapse of seizures was correlated in time with AED discontinuation or withdrawal (Engel's class Id patients).

Table 2. Discontinuation of AEDs according to postoperative outcome and follow-up interval
AEDsSeizure-free patientsPersistent seizures
Totally seizure-free since surgery Ia patients Ia patients (n = 53)Seizure-free ≥1 yr before the time of assessment (n = 81)Rare seizures class II patients (n = 21)Frequent seizures classes III–IV patients (n = 11)
  1. aIa versus rare seizures: p < 0.0024.

  2. bIa versus frequent seizures: p < 0.0019.

Discontinuation1822 0 0
Reduction2235 8 3
Discontinuation or reduction4057  8a   3b
No change1321 9 5
Increase 0 3 4 3

A better outcome was associated with a higher rate of discontinuation or reduction of AEDs. The comparison between seizure-free patients (Ia patients) and those with rare or frequent persistent seizures with respect to the AED regimen (discontinuation or reduction vs. no change or increase) was statistically significant (p < 0.0024 and p < 0.0019, respectively) (Table 2). A lesser frequency of seizures led to a greater reduction of drugs, but the difference between the groups with rare seizures and frequent seizures was not significant.

Discontinuation of AEDs according to the follow-up time interval

A longer outcome was associated with a higher probability that AEDs had been discontinued, but the follow-up time interval did not statistically influence the results.

Driving license

Sixty-eight (62%) patients had a driving license at the time of assessment. Thirteen (11%) had obtained the license before the onset of epilepsy; 39 (35%) obtained it during their active epilepsy period; and 10 (9%) obtained a license after surgery. Of those who obtained a license to drive after surgery, two were seizure-free since surgery, five had been seizure-free for at least a year at the time of assessment, and five still experienced rare disabling seizures Forty-five (41%) patients drove regularly before, and 53 (48%), after surgery. The possibility of driving was a reason to undergo surgery in 58 (53%) patients. Reasons given by patients with seizures that drove before surgery included their professional obligations, a wish to be like other people, and the occurrence of an aura that let them stop driving if they sensed the onset of a seizure. Fifty-eight (53%) patients were aware of French legislation on epilepsy and driving that prohibits driving for patients who still have seizures and allows them to drive after a period of 1 year without seizures.

Among the patients driving after surgery, 28 (53% of the Ia subgroup) were Engel's class Ia, 43 (53% of the seizure free since evaluation subgroup) had been free of seizures for at least 1 year before assessment, and 12 (37.5% of the persistent seizures subgroup) patients had persistent seizures [10 (47.7%) of the rare seizures subgroup and two (18%) of the frequent seizures subgroup] (Table 3). The difference between the three groups (Ia vs. rare or frequent seizures, rare vs. frequent seizures) was not statistically significant (Table 3). As expected, patients with frequent seizures drove less than patients with rare seizures or with no seizures but surprisingly, 40% of patients who were seizure free since surgery (Engel's class Ia) did not have a driving license, and only 28 (52%) of those with a license stated that they drove regularly. The follow-up time interval did not statistically influence the results.

Table 3. Driving license and employment according to the postoperative outcome and follow-up interval
Driving and professional statusSeizure-free patientsPersistent seizures
Totally seizure-free since surgery Ia patients Ia patients (n = 53)Seizure-free ≥1 yr before the time of assessment (n = 81)Rare seizures class II patients (n = 21)Frequent seizures classes III–IV patients (n = 11)
Driving license325213 5
Active driving284310 2
Better professional status3143 6 2
Worse professional status 4 9 4 3

Employment

Sixty-five (61%) patients reported a change in their employment status after surgery. For 52 (47%) patients, professional status improved, and for 15 (14%) patients, it was reduced.

Of those patients reporting an improved professional status after surgery, 31 (58.5% of the Ia subgroup) were seizure free since surgery, and 43 (53.1% of the seizure free since evaluation subgroup) were seizure free for at least 1 year after assessment. The improvement consisted of finding a job for 18 previously unemployed patients, to a promotion for nine patients, to a new professional activity in nine patients, and to an improved performance for six patients.

Among patients who perceived a decline in their professional status, only four were seizure-free since surgery, and nine patients were seizure-free for at least 1 year after assessment. Three patients ceased employment, four entered invalidity, and three patients felt that their careers had stagnated.

The comparison between Ia seizure-free patients and patients with rare or frequent persistent seizures was not statistically significant, based on a Bonferroni comparison, although a trend was noted (Table 3).

The follow-up time interval did not statistically influence the results: employment status was not different for short- or long-term follow-up.

Social status

Nonfamily social relationships

Eight-one (74%) patients noted changes in relations with their nonfamily social relationships, an improvement for 75 (68%) patients and a decline for six (5%) patients.

In the group with improved relations, 32 (60.4% of the Ia subgroup) were seizure free since surgery, 53 (75.9% of the seizure-free since evaluation subgroup) were seizure-free at least 1 year before assessment, and 24 [75% of the persistent seizures subgroups; 17 (81%) of the rare seizures subgroup and seven (64%) of the frequent seizures subgroup] had persistent seizures) (Table 4). The difference was not statistically significant between the seizure-free patients (Ia) and the patients who exhibited rare or frequent persistent seizures, and between the rare and frequent seizures subgroups. The improvement was attributed to a decrease of anxiety in the nonfamily social relationships for 29% of patients, a greater independence for 13%, better communication in 13%, improved self-confidence in 12%, and improved mood in 9%.

Table 4. Social status according to the postoperative outcome and follow-up interval
Social statusSeizure-free patientsPersistent seizures
Totally seizure-free since surgery Ia patients Ia patients (n = 53)Seizure-free ≥1 yr before the time of assessment (n = 81)Rare seizures class II patients (n = 21)Frequent seizures classes III–IV patients (n = 11)
Nonfamily social relationships
 Improvement325317 7
 Worsening 4 4 3 0
Marital status
 Single1017 5 4
 Improvement213013 4
 Worsening 5 8 2 0
Familial status
 Improvement334815 4
 Worsening 0 0 2 0

Patients with worsened relations attributed changes to modifications in their own personality, misunderstandings of their nonfamily social relationships, and psychological problems. The follow-up time interval did not statistically influence the results.

Marital status

At the time of assessment, 25 (23%) patients were single, 46 (42%) reported improvements in relations within their marriage or couple, and 11 (10%) reported a worsening of relations. Of the latter group, six (5%) patients divorced, and five (4%) mentioned major conflicts after surgery.

The proportion of seizure-free patients and the seizure frequency did not statistically significantly influence the single status or changes in the marital status (Table 4). The follow-up time interval did not statistically influence the results.

Familial status

Of 66 (60%) patients who noted changes in relations with their family, the great majority reported an improvement (65 patients). Of these, 33 (62% of the Ia subgroup) patients were seizure-free since surgery, 48 (59% of the seizure-free since evaluation subgroup) were seizure-free for at least 1 year at the time of assessment, and 19 [59% of the persistent seizures subgroups: 15 (71%) of the rare seizures subgroup and four (37%) of the frequent seizures subgroup] had persistent seizures (Table 4). The difference between the seizure-free patients (Ia) and patients with rare or frequent persistent seizures was not statistically significant, nor was a significant difference noted between the patients with rare and frequent seizures.

Improvements in familial relations were attributed to a decrease of anxiety in 48% of cases, to a reduced dependence in 12% of cases, and to better communication in 12% of cases. The follow-up time interval did not statistically influence the results.

Satisfaction and regrets with surgery

One hundred (91%) patients were satisfied with surgery. Among these patients, 50 (94% of the Ia subgroup) were seizure-free since surgery, 76 (94% of the seizure-free since evaluation subgroup) were seizure-free for at least 1 year at time of assessment, and 27 (84% of the rare and frequent persistent seizures subgroups) had persistent seizures (Table 5). Among those patients whose seizures continued but were now less frequent (the 21 class II patients) or exclusively nocturnal (n = 15), the rates of satisfaction were excellent: 18 (86%) and 13 (87%), respectively, but not different from the other subgroups. The difference between the seizure-free patients and the patients who exhibited rare or frequent persistent seizures was not statistically significant.

Table 5. Satisfaction with surgery according to the postoperative outcome and follow-up interval
SatisfactionSeizure-free patientsPersistent seizures
Totally seizure-free since surgery Ia patients Ia patients (n = 53)Seizure-free ≥1 yr before the time of assessment (n = 81)Rare seizures class II patients (n = 21)Frequent seizures classes III–IV patients (n = 11)
Satisfaction507618 9
No satisfaction 3 5 3 2

Ten (9%) patients were dissatisfied: three were seizure free since surgery, and five were seizure-free over a year or longer at time of assessment. Dissatisfaction frequently resulted from a postoperative deficit (lateral hemianopsia in three, motor deficit in one, memory disturbances in one).

One-hundred one patients (92%) stated that they did not regret their decision. Of the nine patients with regrets, four were seizure-free, and five had persistent seizures. The follow-up time interval did not statistically influence the results.

DISCUSSION

In this study, we evaluated the impact of seizure freedom and duration of follow-up on epilepsy and psychosocial outcomes after temporal lobe surgery. Two major results emerged: (a) the epilepsy long-term outcome rates clearly depend on how seizure freedom is defined, and (b) the psychosocial long-term outcome is similar to that described in the short term and globally does not depend on seizure freedom or seizure frequency. The overall results of this study support the durability of the benefits of temporal lobe surgery because 91% of patients were satisfied with surgery whatever their epilepsy status or the duration of the follow-up.

Epilepsy outcome

Seizure-free patients

Long-term epilepsy outcome rates clearly depend on the definition of seizure freedom. Two classes of definition have been used. In one, patients must be free from seizures at the most recent follow-up. The other category uses Engel's system (Engel, 1987), which distinguishes two subgroups of seizure freedom. The first subgroup, Engel's class Ia, consists of patients who report no seizures after their surgery. The second subgroup, Engel's class I, includes both seizure-free patients and those who have experienced simple partial seizures, or “brief auras” and “neighborhood” seizures and drug-withdrawal seizures (McIntosh et al., 2004).

Here, we used both of these definitions of seizure freedom. Thus we focused on seizure freedom at the last follow-up with at least a year of seizure freedom before the final evaluation, and in separate analyses, we considered patients who were completely seizure-free. When analyses were restricted to Engel's class Ia patients (seizure free since surgery), the short-term seizure outcome was better than the long-term seizure outcome. Probabilities of seizure-free survival were 97.6%, 85.2%, and 42.6% at 1, 2, and 10 years after surgery, respectively. In contrast, seizure outcome appeared to be more stable with time, when we analyzed class I patients (similar to those who were seizure free for at least 1 year before assessment). Probabilities of seizure-free survival were 97.6%, 95%, and 71.1%, respectively at 1, 2, and 10 years after surgery.

Thus our results agree with previous original data and meta-analyses showing a proportion of long-term seizure-free patients (mostly class I patients) in the range of 61 to 66% with a gradual decrease in this proportion up to 15 years after surgery (Engel et al., 2003; McIntosh et al., 2004; Pagliolo et al., 2004; Tellez-Zenteno et al., 2005). Our results therefore highlight the importance of the definition of seizure freedom.

Seizure recurrence

Overall, 57 (52%) patients experienced a relapse. Among patients with at least one seizure since surgery, 42 (74%) patients reported a change in the type of seizure experienced. In 39% of these patients, seizures were exclusively generalized tonic–clonic events; in 26% of patients, they were exclusively nocturnal; and in 21%, they were both: generalized tonic–clonic seizures occurred exclusively at night. Few data have been published on the semiology of persistent seizures after surgery. Often, postoperative simple or complex partial seizures similar or close to those that were present before surgery are noted (Di Gennaro et al., 2004). Possibly the epileptogenic zone was not completely removed or perhaps a second epileptogenic zone was unmasked by the surgical removal of the primary one (Rosenow and Lüders, 2001). The appearance of exclusively tonic–clonic seizures that we describe has been observed less frequently and may be attributed to the effect of the surgery or to a genetic predisposition. The emergence of exclusive nocturnal seizures is more surprising. It was observed in 26% of our patients experiencing a relapse.

Patients who experienced one or more seizures after surgery could be divided into two main categories: those who relapsed after a delay of several years (late seizure recurrence) and patients who relapsed in the days or the months after surgery (early seizure relapse).

Thirty-seven patients experienced a relapse of seizures in the first 2 years after surgery, and 20 patients relapsed after 2 years (maximum interval of 108 months after surgery). Late seizure recurrence is crucial to patients, who typically regard themselves as cured after 1 or 2 postoperative years without seizures.

In a recent study,McIntosh et al. (2004) estimated at 25%, the probability that patients who had been free of seizures for 2 years would experience a seizure before 10 years after surgery. No risk factor for late recurrence could be identified. In another study that addressed specifically long-term seizure outcome in patients initially seizure free after epilepsy surgery (Yoon et al., 2003), a significant portion of patients remained vulnerable to relapse, even after years of seizure freedom. However, in our study, a longer seizure-free interval was associated with reduced likelihood for a severe recurrence (more than one seizure per month). Clinicians making presurgical evaluations should note the significant possibility of late seizure relapse while recognizing the higher probability of a favorable outcome. Recurrent seizures can be emotionally devastating to the patient and family, casting doubt on subsequent seizure control. Patients who are initially seizure free should be adequate informed of the small, but real, risk of a relapse.

In contrast, our data show that 50% of patients with a very early recurrence of seizures after their operation were seizure free at 6 years after surgery. This late cessation of seizures may result from new AED trials or may represent a late beneficial effect of surgery. It is a message of hope for the patients who are not immediately seizure free after surgery.

AEDs reduction and discontinuation

In our study, better outcomes were associated with a reduction in use or discontinuation of AEDs. These results agree with some recent findings (McIntosh et al., 2004) but differ from others (Schiller et al., 2000), in which patients who discontinued AEDs were found to have a significantly higher risk of recurrence. These differences may result from different discontinuation protocols and perhaps from the highly selected population studied here, which consisted exclusively of patients with MTLE with hippocampal sclerosis.

Psychosocial outcome

We found that psychosocial outcome was independent of seizure outcome. Recent reports conflict on relations between seizure freedom and improvements in psychosocial functioning. Some studies (Lowe et al., 2004; Reid et al., 2004) suggest that patients with persistent seizures had a poorer psychosocial profile; other studies (Jones et al., 2002) showed that freedom from seizures was not a prerequisite for an improved psychosocial outcome. Our study demonstrates that psychosocial outcome may improve both in seizure-free patients and in those who still experience seizures with no clear relation with seizure frequency. Globally, surgery had a significantly positive effect on employment as well as on social and marital status. Short-term and long-term psychosocial outcomes were rather similar. Nevertheless, some patients still experienced psychosocial difficulties, highlighting the burden of epilepsy, even after surgery.

Driving license

Driving is generally restricted and monitored in patients with refractory epilepsy. Despite these constraints, 41% of our patients drove regularly before surgery (Berg et al., 2000).

As expected, postsurgical outcome influenced the proportion of patients that drove regularly: better outcomes were associated with regular active driving. Surprisingly, however, the proportion of patients who drove after surgery did not increase significantly. We found that 48% of our patient sample drove regularly after surgery, whereas 41 did so before. Furthermore, more than half of the seizure-free patients, by either definition, did not drive after surgery. This point illustrates that fears and interdictions do not always disappear, even after successful surgery.

Employment

Fewer than half the patients in this study thought that their professional status improved after surgery. Postsurgical outcome had no significant effects on employment status, but a highest proportion of seizure-free patients noted an improvement of their professional status; 39% of the patients thought that their professional status was unchanged, and 14% considered that their employment status had worsened, no matter their seizure-outcome status. This point illustrates that employment difficulties continue for a sizable proportion of patients at long-term follow-up, regardless of seizure freedom. Because epilepsy is a chronic disorder that usually begins in childhood, it often has huge repercussions on school achievement, level of education, and professional orientation. Previous studies have highlighted the employment difficulties of epilepsy patients (Bishop and Allen, 2001). This study, as others (Jones et al., 2002; Shackleton et al., 2003) shows that whereas these difficulties may persist after surgery, freedom from seizures is likely to initiate a long-term and stable improvement in professional status.

Social status

A high proportion of patients reported better social or familial relationships after surgery. Overall, the improvement was attributed to a lower level of anxiety and better self-esteem. Neither freedom from seizures nor the duration of follow-up influenced these self-perceived psychosocial changes. Thus positive psychosocial outcomes are not limited to individuals who become seizure free, and some patients attain optimal psychosocial outcomes even without a complete suppression of seizures (Jones et al., 2002).

In total, 42% of the patients mentioned an improvement of their marital status, but 23% were still single, and 10% reported a divorce or a major conflict. Once again, neither seizure freedom nor the duration of follow-up significantly influenced the marital status of patients who underwent surgery. Clearly, a positive psychosocial change is not an inevitable outcome of epilepsy surgery.

Satisfaction

Whatever their epilepsy outcome and the duration of the follow-up, 91% of patients were satisfied with surgery, and 92% did not regret their decision,. Reasons for dissatisfaction with surgery included cognitive, visual, or motor deficits.

Limitations of the study

Although the study includes a large sample size, only 60% of the patients who underwent surgery between 1988 and 2004 responded to the survey, and this may represent a bias in participating subjects. Globally, the characteristics of the patients who did not return the questionnaire were similar to those of the patients who returned it.

Conclusions

The study demonstrates the long-term benefits of temporal lobe surgery. It describes significant and persistent gains in psychosocial function after surgery. Seizure freedom is important, but even without a complete absence of seizures, most patients report positive changes in psychosocial outcome. However, considerable psychosocial difficulties are still apparent at long-term follow-up in a substantial minority of patients whose epilepsy status is indistinguishable from the majority who showed an improvement in their psychosocial status and their seizure outcome. This group may require more directed rehabilitation to benefit from the reduced seizure frequency or freedom from seizures.

Acknowledgments

Acknowledgment:  We thank Dr. Miles for helpful comments on the text.

Ancillary