SEARCH

SEARCH BY CITATION

Keywords:

  • Hippocampal sclerosis;
  • Epilepsy surgery;
  • Temporal lobe epilepsy;
  • Outcome;
  • Predictors

Summary

  1. Top of page
  2. Patients and Methods
  3. Results
  4. Discussion
  5. Acknowledgments
  6. References

Purpose: To study long-term postoperative course and identify predictors for postoperative seizure control in patients with medically intractable temporal lobe epilepsy (TLE) associated with hippocampal sclerosis (HS), diagnosed by magnetic resonance imaging (MRI), and ascertained histopathologically. To compare patients becoming seizure-free (i.e., cured from epilepsy) and patients experiencing prolonged seizure-free periods interposed with recurring seizures.

Methods: One hundred thirty-five patients (74 women) underwent complete evaluation for epilepsy surgery. The predictive value of duration of epilepsy, age at onset, age at surgery, gender, febrile convulsion history, ictal dystonic posturing, unilateral interictal electroencephalography (EEG) discharges (IED), preoperative secondarily generalized tonic–clonic seizures (SGTCS), and preoperative seizure frequency for short- and long-term postoperative seizure control were evaluated with two classification systems: Classification 1 (seizure-freedom with or without auras during 12-months before observation points) and the stringent classification 2 [International League Against Epilepsy (ILAE) Ia; absolute absence of seizures and auras after operation].

Results: Unilateral IED at year 1 and 2 (p = 0.037 and p = 0.034), male gender and low seizure frequency at year 2 (p = 0.013 and p = 0.046) were significant predictors for seizure freedom using classification 1. All variables (except male gender at year 2; p = 0.035) lost their predictive power, applying classification 2. The proportion of seizure-free patients remained stable between 70% to 79% with classification 1, but decreased from 64.4% at year 1 to 45.8% at year 5 with classification 2.

Discussion: Positive predictors of short-term outcome do not predict long-term outcome in patients with TLE associated with HS. Absolute freedom of seizures and auras cannot be predicted by conventional preoperative variables.

Temporal lobe epilepsy (TLE) with hippocampal sclerosis (HS), the most frequent epilepsy syndrome, is refractory to antiepileptic drugs in 58%–89% of patients (Semah et al., 1998; Stephen & Brodie, 2002). Epilepsy surgery was suggested to offer an extremely powerful treatment option for these patients providing 70%–80% seizure-freedom outcome (Engel, 1996). The postoperative clinical course, however, is variable. An immediate cessation of seizures, a reduction of seizures over time with a late cure, known as “running down” phenomenon, as well as a continuation of seizures can occur (Rasmussen, 1970; Salanova et al., 1996).

As data from pure cohorts of patients with TLE and HS are limited and generated controversial findings (Kilpatrick et al., 1999; Hennessy et al., 2001; Hardy et al., 2003; Janszky et al., 2005), we were interested to study the course of well-defined patients with TLE and HS after epilepsy surgery and to investigate possible predictors of postoperative seizure-outcome. Unilateral interictal electroencephalography (EEG) discharges (IED), epilepsy duration, older age at surgery, hand dystonic posturing, and the presence of secondarily generalized tonic–clonic seizures (SGTCS) were mentioned to be predictors for the postoperative outcome by some but not all investigators (Kilpatrick et al., 1999; Hennessy et al., 2001; Hardy et al., 2003; Janszky et al., 2005).

Absence of postoperative seizures during a limited period of 12 to 24 months prior to postoperative assessment independent of the occurrence of auras was a commonly used criterion of seizure freedom in previous studies (Kilpatrick et al., 1999; Hennessy et al., 2001; Hardy et al., 2003; McIntosh et al., 2004; Janszky et al., 2005; Spencer et al., 2005; Asztely et al., 2007). In addition to this definition, we analyzed our data using a stringent definition of seizure-freedom (i.e. completely seizure- and aura-free at any time during a 5-year-observation period after surgery). The ultimate goal of the study was to identify features that would be helpful for the preoperative counseling of surgical candidates.

Patients and Methods

  1. Top of page
  2. Patients and Methods
  3. Results
  4. Discussion
  5. Acknowledgments
  6. References

One hundred thirty-five patients underwent selective amygdalohippocampectomy (SAHE) or anterior temporal lobectomy (ATL) at our institution between 1995 and 2005. Surgery and the decision on the type of operation were made independently and on a case by case basis by one of us (T.C.). All subjects had to fulfil the following criteria: Age ≥15 years, presurgical evaluation in our epilepsy monitoring unit, complex partial seizures with and/or without SGTCS documented by prolonged video EEG monitoring, unilateral HS proven by magnetic resonance imaging (MRI) (dual pathology and bilateral HS were exclusion criteria), more than 1 year of postoperative follow-up.

The extensive presurgical evaluation included a detailed clinical history taken from the patients and their relatives during admission to our epilepsy monitoring unit. All patients underwent prolonged video EEG monitoring with electrode placement according to the Extended International 10-20 system including bilaterally placed sphenoidal electrodes. Seizure semiology, ictal and interictal EEG data were studied separately according to previously proposed criteria (Rosenow & Luders, 2001). Dystonic posturing was defined as sustained forced unnatural posturing of one upper extremity with a rotation component (Kotagal et al., 1989). Frequency and location of interictal spikes was assessed by visual analysis (10 min/h) over the entire recording time. Unilateral IED were defined as >90% of spikes occurring over the affected temporal lobe.

High-resolution MRI was performed on a 1.5 Tesla machine (Philips Gyroscan ACS-NT; Best the Netherlands). MRI diagnosis of HS was proven in all patients by postoperative histopathological examination. Criteria of HS were significant neuronal cell loss and gliosis in hippocampal subfields CA1, CA3, endofolium, and in the dendate gyrus associated with relative preservation of CA2 neurons (Margerison & Corsellis, 1966; Williamson et al., 1993; Briellmann et al., 2002).

The following variables were assessed for possible prognostic implications: Age at epilepsy onset, epilepsy duration prior to surgery, age at operation, gender, history of febrile convulsions, unilateral IED, preoperative occurrence of SGTCS, preoperative seizure frequency (we distinguished between more than one seizure per week, i.e., “high seizure frequency,” and less than or equal to one seizure per week, i.e., “low seizure frequency”), and ictal dystonic posturing. Antiepileptic treatment was evaluated preoperatively, as well as postoperatively at the time of the last control visit.

Postoperative clinical examinations were performed 1, 2, and 5 years after the operation. Calculations were performed with two distinct definitions of outcome: (1) Classification 1, no seizures with or without nondisabling auras during a 12 month-period prior to the assessment at 1, 2, and 5 years postoperatively and (2) classification 2, seizure-free outcome according International League Against Epilepsy (ILAE) class Ia (i.e., complete seizure freedom and absence of nondisabling auras at any time after surgery) (Wieser et al., 2001). Patients with seizures on attempted antiepileptic drug (AED) withdrawal were considered to be not seizure-free.

Statistical methods

First, we preformed a univariate logistic regression analysis in order to assess the prognostic significance of all variables. We calculated the odds ratios (OR) for a seizure-free outcome at 1, 2, and 5 years after the operation (significance level < 0.05).

Second, we performed a multivariate analysis in order to investigate the independent prognostic effect of all variables. We applied a stepwise logistic regression model using a forward selection procedure. We applied the Wald statistic to test whether the coefficients of the regression equation significantly differed from zero.

For all statistical analysis, we used the SPSS statistical program package (version 14.0; SPSS Inc., Chicago, IL, U.S.A.).

Results

  1. Top of page
  2. Patients and Methods
  3. Results
  4. Discussion
  5. Acknowledgments
  6. References

One hundred thirty-five consecutive patients (74 women, 61 men) with medically refractory TLE and HS were included in the study. The mean age at epilepsy onset was 11.4 years (range 1 to 37 years), and the mean epilepsy duration prior to surgery was 23.5 years (range 3 to 50 years). The mean age at operation was 35 years (range 15–52 years). Patient characteristics are presented in Table 1.

Table 1.  Demographic and baseline clinical characteristics of patients with TLE
  Number of patients (%)
  1. SGTCS, secondarily generalized tonic–clonic seizures; SAHE, selective amygdalohippocampectomy; ATL, anterior temporal lobectomy; IED, interictal EEG discharges.

Study population (total number) 135
SexMen61 (45.2)
Women74 (54.8)
Surgical proceduresSAHE105 (77.8) 
ATL30 (22.2)
Side of resectionRight63 (46.7)
Left72 (53.3)
AurasYes113 (83.7) 
No22 (16.3)
Seizure frequencyHigh (>one seizure/week)108 (80.0) 
Low (<one seizure/week)27 (20.0)
SGTCSYes99 (73.3)
No36 (26.7)
IEDUnilateral97 (71.9)
Bilateral38 (28.1)
Ictal dystonic posturingYes37 (27.4)
No98 (72.6)
Febrile convulsionsYes47 (34.8)
No88 (65.2)
Mean (years)Range
Mean age at surgery (years)34.915–52
Mean age at epilepsy onset (years)11.41–37
Mean duration of epilepsy (years)23.53–50

One hundred five patients (78%) underwent SAHE, and 30 (22%) had ATL. Sixty-three patients (47%) were operated on the right side, and left-sided operations were performed in the remaining 72 patients (53%).

The numbers of patients remaining seizure-free according to classification 1 throughout the observation period, as well as the time-dependent occurrence of seizures in the nonseizure-free patients are shown in Fig. 1. According to classification 1, 95 (70.4%) patients reported to be seizure-free at 1 year. Seventy-nine (71.2%) of 111 patients were seizure-free at the 2-year examination. There were 57 (79.1%) of 72 patients who were seizure-free 5 years after operation. Three patients were lost during the follow-up period, two patients committed suicide, and one patient died from sudden unexpected death of epilepsy. Fig. 2 shows the percentage of patients with or without seizures at the different time points after surgery and the percentage of patients changing between the seizure-free and nonseizure-free groups.

image

Figure 1. Postoperative outcome of 135 patients with TLE during different time periods after selective amygdalohippocampectomy or anterior temporal lobectomy, analyzed according to classification 1 (no seizures with or without nondisabling auras during a 12 month-period prior to the assessment at 1, 2, and 5 years postoperatively). Yellow bars represent seizure-free time intervals, red bars indicate time intervals in which seizures occurred.

Download figure to PowerPoint

image

Figure 2. Intergroup variability for seizure-free and nonseizure-free intervals after selective amygdalohippocampectomy or anterior temporal lobectomy analyzed according to classification 1 (no seizures with or without nondisabling auras during a 12-month period prior to the assessment at 1, 2, and 5 years postoperatively). Yellow bars represent percentage of patients without seizures, red bars represent percentage of patients with seizures at 1 year, 2, 5, and 10 years. Orange bars show percentage of patients without seizures despite reporting seizures in the precedent period, blue bars indicate percentages of patients with seizures, but seizure freedom in the precedent period.

Download figure to PowerPoint

According to classification 2 (i.e., ILAE class 1a) 87 patients (64.4%) were completely seizure-free 1 year after surgery. Sixty-two patients (55.8%) reported complete seizure- and aura-freedom at the 2-year examination. Thirty-three patients (45.8%) stayed completely seizure-free for 5 years after surgery (Fig. 3).

image

Figure 3. Outcome of patients with TLE after selective amygdalohippocampectomy or anterior temporal lobectomy according to two different classifications for seizure-free outcome. Classification 1, no seizures with or without nondisabling auras during a 12-month period prior to the assessment at 1, 2, and 5 years postoperatively; classification 2, absolute freedom from seizures and auras at any time after surgery.

Download figure to PowerPoint

According to classification 1, unilateral IED at years 1 and 2 and male gender at year 2 were positive predictors, whereas high preoperative seizure frequency at year 2 was a negative predictor for a seizure-free outcome. All variables (except for gender at year 2) lost their predictive power when the more rigorous classification 2 (ILAE Ia) was applied for data analysis. (Tables 2, 3 and 4)

Table 2.  Univariate analysis of variables at year 1 after epilepsy surgery for seizure-free outcome with or without nondisabling auras
Number of patientsClassification 1Classification 2
No seizures with/without auras during precedent 12 monthsNo seizures plus no auras at any time after surgery
Seizure-free 95 Mean (sd)Nonseizure-free 40 Mean (sd)p-valueOR (95% CI)Seizure-free 87 Mean (sd)Nonseizure-free 48 Mean (sd)p-valueOR (95% CI)
Age at epilepsy onset (years)11.2 (8.6) 11.9 (11.9)0.9891.00 (0.95–1.06)11.8 (8.6) 10.7 (10.7)0.4290.98 (0.93–1.03)
Duration of epilepsy (years)23.9 (11.0)22.6 (11.0)0.6340.99 (0.95–1.03)23.2 (10.7)24.0 (11.6)0.3771.02 (0.98–1.06)
Age at surgery (years)35.1 (8.2) 34.6 (9.7) 0.9891.00 (0.95–1.06)35.1 (8.1) 34.7 (9.6) 0.4290.98 (0.93–1.03)
 Seizure-free N of pat (%)Nonseizure-free N of pat (%)p-valueOR (95% CI)Seizure-free N of pat (%)Nonseizure-free N of pat (%)p-valueOR (95% CI)
  1. N of pat, number of patients.

Men46 (48.4)15 (37.5)0.3791.44 (0.64–3.22)41 (47.1)20 (41.7)0.6791.17 (0.55–2.51)
Surgical procedure SAHE76 (80.0)29 (72.5)0.6441.23 (0.51–2.99)70 (80.4)35 (72.9)0.3721.47 (0.63–3.44)
Febrile convulsions33 (34.7)14 (35.0)0.8770.94 (0.40–2.17)32 (36.9)15 (31.2)0.4291.38 (0.62–3.10)
Seizure frequency high73 (76.8)35 (87.5)0.0520.32 (0.10–1.01)66 (75.9)42 (87.5)0.0520.36 (0.13–1.01)
Unilateral IED73 (76.8)24 (60.0)0.0372.56 (1.05–6.19)65 (74.7)32 (66.7)0.2741.61 (0.68–3.81)
SGTCS69 (72.6)30 (75.0)0.7680.87 (0.35–2.16)66 (75.9)33 (68.7)0.3431.50 (0.65–3.40)
Ictal dystonic posturing29 (30.5) 8 (20.0)0.2291.78 (0.70–4.55)26 (29.9)11 (22.9)0.3041.57 (0.66–3.69)
 
Table 3.  Univariate analysis of variables at year 2 after epilepsy surgery for seizure-free outcome with or without nondisabling auras
Number of patientsClassification 1Classification 2
Seizure-free 79 Mean (sd)Nonseizure-free 32 Mean (sd)p-valueOR (95% CI)Seizure-free 62 Mean (sd)Nonseizure-free 49 Mean (sd)p-valueOR (95% CI)
Age at epilepsy onset (years)11.3 (8.6) 11.4 (11.3)0.3581.03 (0.96–1.10)11.6 (8.4)11.0 (10.6)0.4071.03 (0.96–1.09)
Duration of epilepsy (years)23.0 (10.8)24.8 (10.7)0.8001.01 (0.96–1.06)22.3 (9.9)25.1 (11.6)0.4491.02 (0.97–1.07)
Age at surgery (years)34.3 (8.3) 36.2 (8.3) 0.3581.03 (0.96–1.10)33.9 (7.7)36.0 (8.5) 0.4071.03 (0.96–1.09)
 
 Seizure-free N of pat (%)Nonseizure-free N of pat (%)p-valueOR (95% CI)Seizure-free N of pat (%)Nonseizure-free N of pat (%)p-valueOR (95% CI)
Men45 (57.0) 9 (28.1)0.0133.37 (1.28–8.82)36 (58.1)18 (36.7)0.0352.43 (1.06–5.80)
Surgical procedure SAHE76 (80.0)26 (81.3)0.9911.01 (0.33–3.03)50 (86.6)38 (77.5)0.3091.64 (0.63–4.25)
Febrile convulsions25 (31.6)15 (46.9)0.0830.43 (0.16–1.12)19 (36.6)21 (42.8)0.1600.54 (0.23–1.28)
Seizure frequency high59 (74.7)27 (84.4)0.0460.26 (0.07–0.98)45 (72.6)41 (83.7)0.0650.36 (0.12–1.07)
Unilateral IED64 (81.0)21 (65.6)0.034 3.38 (1.10–10.38)48 (77.4)37 (75.5)0.5211.40 (0.50–3.92)
SGTCS57 (72.2)25 (78.1)0.7650.84 (0.28–2.54)47 (75.8)35 (71.4)0.3691.54 (0.60–3.95)
Ictal dystonic posturing22 (27.8) 8 (25.0)0.8951.07 (0.38–3.04)18 (29.0)12 (24.5)0.6151.27 (0.50–3.19)
Table 4.  Univariate analysis of variables at year 5 after epilepsy surgery for seizure-free outcome with or without nondisabling auras
Number of patientsClassification 1Classification 2
Seizure-free 57 Mean (sd)Nonseizure-free 15 Mean (sd)p-valueOR (95% CI)Seizure-free 33 Mean (sd)Nonseizure-free 39 Mean (sd)p-valueOR (95% CI)
Age at epilepsy onset (years)11.4 (8.7)10.9 (7.9)0.6631.02 (0.92–1.13)10.2 (7.5)12.3 (9.1)0.2491.05 (0.96–1.15)
Duration of epilepsy (years)22.7 (9.4)26.5 (9.6)0.3761.04 (0.95–1.16)22.8 (9.0)24.1 (9.9)0.8311.01 (0.94–1.08)
Age at surgery (years)34.2 (7.9)37.3 (7.5)0.6631.02 (0.92–1.13)33.0 (7.7)36.4 (7.8)0.2491.05 (0.96–1.15)
 
 Seizure-free N of pat (%)Nonseizure-free N of pat (%)p-valueOR (95% CI)Seizure-free N of pat (%)Nonseizure-free N of pat (%)p-valueOR (95% CI)
Men31 (54.9) 6 (40.0)0.6131.40 (0.38–5.12)20 (60.6)17 (43.6)0.3621.64 (0.57–4.70) 
Surgical procedure SAHE48 (84.2)10 (66.7)0.1662.51 (0.69–9.24)28 (84.8)30 (76.9)0.2722.03 (0.57–7.18) 
Febrile convulsions24 (42.1) 6 (40.0)0.8160.85 (0.23–3.20)29 (87.9)17 (43.6)0.2770.55 (0.19–1.62) 
Seizure frequency high47 (82.5)11 (73.3)0.9481.05 (0.23–4.88)28 (84.8)30 (76.9)0.9801.02 (0.25–4.11) 
Unilateral IED45 (78.9)11 (73.3)0.8311.20 (0.23–6.22)29 (87.9)27 (69.2)0.1862.70 (0.62–11.75)
SGTCS47 (82.5)13 (86.7)0.9700.97 (0.17–5.54)29 (87.9)31 (79.5)0.1442.98 (0.69–12.89)
Ictal dystonic posturing17 (29.8) 3 (20.0)0.5731.53 (0.35–6.73)11 (33.3) 9 (23.1)0.3721.69 (0.53–5.34) 

For the prediction of seizure control after 2 years we performed a stepwise logistic regression analysis in those cases where more than one significant predictive variable could be identified in the univariate analysis. This applied only to seizure-free outcome after 2 years. We found that the significant variables in the univariate analysis (sex, seizure-frequency before the operation, unilateral IED) remained significant in the multivariate analysis excluding a dependence of these variables. The results of the multivariate stepwise logistic regression analysis are shown in Table 5. While the percentage of correctly predicting a seizure-free outcome after 2 years with the logistic regression equation was 96.2%, the percentage of correctly predicting a nonseizure-free outcome after 2 years was only 15.6%. The overall probability of predicting the outcome correctly accounted for 73%.

Table 5.  Results of the multivariate stepwise logistic regression analysis at 2 years after epilepsy surgery
VariableBS.E.WalddfSigExp (B)
  1. Sex: male, 1; female, 2; unilateral IED: yes, 1; no, 2; seizure frequency: high, 1; low, 2. IED, unilateral interictal electroencephalography discharges.

  2. The probability of being seizure-free after 2 years (p) can be calculated as follows:

    • image
  3. where z = constant + 1.270 × sex + 1.095 × IED - 1.215 × seizure frequency.

Sex 1.2700.4717.26810.0073.560
Unilateral IED 1.0950.5454.04010.0442.990
Seizure frequency-1.2150.6393.61510.0570.297
Constant-2.8301.0746.94510.0080.059

At the 1-year evaluation, all patients were on antiepileptic drugs. Eleven of 111 (10%) patients were off medication at 2 years and 23 of 72 patients (32%) at 5 years after operation. Twenty-one patients who stopped taking antiepileptic drugs were completely seizure- and aura-free. Two patients stopped the medication against physician's advice, despite having rare seizures.

Discussion

  1. Top of page
  2. Patients and Methods
  3. Results
  4. Discussion
  5. Acknowledgments
  6. References

Epilepsy surgery evolved to an effective treatment option for medically refractory patients suffering from TLE and can lead to freedom of seizures in about 70% of patients (Engel, 1996; Berg et al., 1998; Radhakrishnan et al., 1998; McIntosh et al., 2001; Wiebe et al., 2001). Applying classification 1, we found seizure freedom in 70.4% of 135 patients with TLE and HS at 1 year, 71.2% of 111 patients at 2 years, and 79.1% of 72 patients at 5 years after epilepsy surgery. We showed that unilateral IED at year 1, male gender and unilateral IED at year 2, were positive predictors, whereas a high seizure frequency at year 2 was a negative predictor. However, when a rigorous classification for permanent seizure freedom (classification 2, i.e., complete seizure freedom and absence of nondisabling auras at any time after surgery) was used, the number of successfully treated patients was reduced and decreased over time from 64.4% at year 1 to 45.8% at year 5 (Fig. 3). Furthermore, all variables lost their value in outcome prediction in this stringent, but for patients, more relevant analysis (Tables 2–4).

In our long-term study, the proportion of seizure-free patients remained stable between 70% to 79%, when seizure freedom was defined as absence of seizures with or without nondisabling auras in a 12-month period preceding the assessment. We applied this definition, commonly used in the literature for an initial analysis and defined it as classification 1. Using this approach, the postoperative course of our patients followed typical patterns. Ideally, patients became seizure-free after surgery. In a small subset of patients, surgery failed to abolish seizure activity. In a subgroup of our patients, the initial success was not sustained during long-term follow-up (Fig. 1). In other patients, an initial postoperative period with seizures did not necessarily predict an unfavorable long-term outcome, which is well known as “phenomenon of running down” (Rasmussen, 1970; Salanova et al., 1996). We were not able to identify a consistent clinical pattern for prediction of changes in seizure control in the long-term. Similar to others, we observed that a change from seizure freedom to the reoccurrence of seizures and vice versa could occur at any time point after surgery (Fig. 2) (Jeong et al., 2005). In some of our patients, surgery did not yield the anticipated, immediate success. Failure of surgery could be the result of a more complex epileptogenic network, involving temporal regions, the orbitofrontal cortex, the insula, the frontal and parietal operculum, and temporoparietoocciptal junction, but mimicking temporal lobe epilepsy (Ryvlin, 2003). For this multilobar epilepsy, characterized by typical temporal seizure semiology and electroclinical features, the term “temporal plus epilepsy” was proposed (Ryvlin, 2003; Ryvlin & Kahane, 2005). Alternatively, an independent contralateral temporal epileptogenic zone could be the cause for approximately 20% of surgical failures (Hennessy et al., 2000). One intention of our investigation was to study the postoperative outcome in this pure group of TLE patients with HS using a rigorous and clinically relevant definition of seizure freedom. Using our classification 2, which followed the ILAE Ia criteria (i.e., patients completely seizure-free since the operation), we found a lower number of successfully treated patients and a decrease of patients without recurrence of seizures and auras from 64.4% at year 1 to 45.8% at year 5 (Fig. 3). This finding compares well with the observation that <40% of patients can be considered cured 5 years after surgical treatment (Wieser & Hane, 2003).

In an effort to explain the apparent variability in outcomes between different patient cohorts and when different criteria for analysis are used, we investigated possible predictors for postoperative seizure freedom. For a number of factors, a predictive value for postoperative outcome has been reported in surgically treated patients with epilepsy of various etiologies (Salanova et al., 1996; Holmes et al., 1997; Radhakrishnan et al., 1998; McIntosh et al., 2001). We decided to study patients with TLE and HS selectively.

Using classification 1, we found unilateral IED at year 1 and 2 and male gender at year 2 to be predictors for a favorable outcome. A high seizure frequency negatively predicted outcome at year 2.

Our finding of unilateral IED as a positive predictor is in accordance with observations on patients with TLE and HS (Schulz et al., 2000; Hennessy et al., 2001; Abosch et al., 2002), but not all studies found a predictive value of unilateral IED for the postoperative outcome (Hardy et al., 2003; Janszky et al., 2005). While unilateral IED indicating strictly unitemporal disease represent an intuitively reasonable predictor for a favorable outcome, they failed to predict long-term outcome. This can be explained that the majority of surgical failures are not caused by contralateral seizures, but by an extended network within the ipsilateral hemisphere (Ryvlin, 2003; Ryvlin & Kahane, 2005).

A low seizure frequency was associated with a favorable 2-year postoperative outcome in our patients, a finding similar to previous observations (Clusmann et al., 2002). Other investigators did not find an influence of the number of seizures on postoperative outcome, however, definition of seizure frequency differed considerably between these studies (Kilpatrick et al., 1999; Janszky et al., 2005; Spencer et al., 2005). Some groups found an influence of age at surgery and epilepsy duration on the postoperative outcome (Jeong et al., 1999; Wieser et al., 2003; Yoon et al., 2003; Janszky et al., 2005; Schwartz et al., 2006). In agreement to other investigators we were not able to duplicate these results in our series of patients (Kilpatrick et al., 1999; Hennessy et al., 2001; Hardy et al., 2003).

Febrile convulsions, irrespective of the type, did not exert a significant prognostic influence in our series, as has been previously reported (Radhakrishnan et al., 1998; Kilpatrick et al., 1999; Hennessy et al., 2001). While a history of complex febrile convulsions is suggestive of a favorable influence on postoperative outcome (Abou-Khalil et al., 1993; Kanemoto et al., 1998; Janszky et al., 2003), we were not able to reproduce these findings in our study population. We undertook every effort to obtain an exact history by thoroughly interviewing the patient and her/his relatives. One has to appreciate, however, that in an adult population, an exact history of events in early childhood is difficult to obtain. After many years, it is often impossible to distinguish between simple and complex febrile convulsions. This may be an explanation for the diversity of results concerning the predictive value of complex febrile convulsions in previous studies.

SGTCS indicate a more rapid seizure spread. However, the presence of SGTCS did not have an impact on outcome at any time point in our series, other studies also failed to show an association between the presence and frequency of SGTCS with a worse outcome (Guldvog et al., 1994; Kilpatrick et al., 1999). In contrast, others reported that the presence of SGTCS had a negative predictive value (Hennessy et al., 2001; McIntosh et al., 2004; Janszky et al., 2005; Jeong et al., 2005). Hardy and coworkers mentioned a negative predictive value of a history of status epilepticus in patients with mesial temporal lobe epilepsy (MTLE), but an influence of the presence of SGTCS was not studied (Hardy et al., 2003).

We went on to study the importance of ictal dystonia as an ictal lateralizing signs in the presurgical evaluation for later outcome. Only one study examined the predictive value of ictal dystonia (Janszky et al., 2005). Janszky et al. demonstrated that ictal dystonia was a bad predictor for 2- and 3-year outcomes. As ictal dystonia is associated with widespread activation of temporal and frontal lobes, due to seizure propagation to the basal ganglia (Newton et al., 1992; Kuba et al., 2003), spreading of ictal activity could lead to an adverse outcome (Schulz et al., 2000). We were not able to identify ictal dystonia as useful predictors of outcome at any time point in our study.

Different surgical approaches resulted in similar outcomes in our series as had been reported elsewhere (Clusmann et al., 2002).

Most importantly, however, by implying classification 2 requiring permanent freedom of seizures and absolute absence of auras for data analysis, all variables lost their predictive power. The suggestion of male gender as a positive predictor for outcome at year 2 is difficult to interpret and should be tested in larger series of patients.

We acknowledge two limitations of our study. First the number of surgically treated patients with a long postoperative observation period is limited. This common problem to all studies published so far is associated with a reduction of statistical power especially for the analysis of long-term outcomes. In order to obtain definitive results, larger patient cohorts need to be evaluated over an even longer postoperative time course.

Secondly, the impact of postoperative AED treatment on postoperative seizure control could not be systematically assessed, as a significant number of patients were referred from outside centers for surgery, which were of course actively involved in postoperative drug treatment. At the same time, we believe that AED treatment plays an important role for postoperative seizure control.

Taken together, our data support the observations of others that surgery can be beneficial in many patients suffering from medically intractable TLE with HS. There are, however, no reliable predictors of seizure-free outcome, if a stringent classification is used. Most of the suggested variables do not predict outcome at all or lose their predictive power during a prolonged observation period. Taking this into consideration, the preoperative counseling of an individual patient with TLE and HS remains challenging.

Acknowledgments

  1. Top of page
  2. Patients and Methods
  3. Results
  4. Discussion
  5. Acknowledgments
  6. References

We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.

Conflict of interest: The authors report no conflict of interest.

References

  1. Top of page
  2. Patients and Methods
  3. Results
  4. Discussion
  5. Acknowledgments
  6. References