• Epilepsy surgery;
  • First seizure;
  • Outcome;
  • Refractory


  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Conclusions
  7. Disclosure
  8. References

Purpose:  Up to one-half of epilepsy surgery patients will have at least one seizure after surgery. We aim to characterize the prognosis following a first postoperative seizure, and provide criteria allowing early identification of recurrent refractory epilepsy.

Methods:  Analyzing 915 epilepsy surgery patients operated on between 1990 and 2007, we studied 276 who had ≥1 seizure beyond the immediate postoperative period. The probability of subsequent seizures was calculated using survival analysis. Patients were divided into seizure-free (no seizures for ≥1 year) and refractory (persistent seizures) and analyzed using multivariate regression analysis.

Results:  After a first seizure, 50% had a recurrence within 1 month and 77% within a year before the risk slowed down to additional 2–3% increments every two subsequent years. After a second seizure, 50% had a recurrence within 2 weeks, 78% within 2 months, and 83% within 6 months. Having both the first and second seizures within six postoperative months [odds ratio (OR) 4.04; 95% confidence interval (CI) 2.05–8.40; p = 0.0001], an unprovoked initial recurrence (OR 3.92; 95% CI 2.13–7.30; p < 0.0001), and ipsilateral spikes on a 6-months postoperative electroencephalography (EEG) (OR 2.05; 95% CI 1.10–3.88; p = 0.025) predicted a poorer outcome, with 95% of patients who had all three risk factors becoming refractory. All patients with cryptogenic epilepsy and recurrent seizures developed refractoriness.

Discussion:  Seizures will recur in most patients who present with their first postoperative event, with one-third eventually regaining seizure-freedom. Etiology and early and unprovoked postoperative seizures with epileptiform activity on EEG at six postoperative months may predict recurrent medical refractoriness.

Patients who undergo brain surgery for the treatment of medically intractable epilepsy anticipate a “cure.” However, up to one-half will have at least one seizure after surgery (McIntosh et al., 2004; Spencer et al., 2005; Jeha et al., 2006). Some undergo expensive reevaluations only to regain seizure-freedom without the need for further intervention(s), whereas others have recurrent seizures for years before the possibility of a reoperation is explored. No evidence-based guidelines exist for managing these patients.

Multiple factors may contribute to this uncertainty: first, the transition from recurrent but rather isolated breakthrough seizures to persistent intractable epilepsy after surgery is poorly defined. Several studies suggest that up to 30% of patients having seizures within the first 6–12 postoperative months eventually become seizure-free (Salanova et al., 1996; Ficker et al., 1999; Hennessy et al., 2000), but definite criteria allowing timely identification of the remaining 70% who will continue with persistent seizures remain elusive. Second, although multiple studies evaluated acute postoperative seizures occurring within 7–28 days of surgery (Garcia et al., 1991; Malla et al., 1998; Tigaran et al., 2003; Abou-Khalil, 2004; McIntosh et al., 2005), only few investigated later seizures (Ficker et al., 1999; Hennessy et al., 2000; Radhakrishnan et al., 2003). This resulted in the availability of rather limited information regarding patients who present with their first seizure beyond the immediate postoperative period. Third, surgical outcome series traditionally used either Engel class I (McIntosh et al., 2004; Sindou et al., 2006; Elsharkawy et al., 2008; Jehi et al., 2009), or seizure-freedom for 12–24 months (Hennessy et al., 2001; Janszky et al., 2003; Kelley & Theodore, 2005; Spencer et al., 2005) in the process of defining “seizure-freedom,” including patients who may have had “some disabling seizures after surgery, but [were] free of disabling seizures for at least 2 years” (Engel IC) or who had “convulsions with antiepileptic drug (AED) discontinuation” (Engel ID) in the same outcome group as patients who never had a seizure, on or off AEDs, after surgery. Conversely, series using the strict criterion of “complete seizure-freedom” since surgery (Yoon et al., 2003; Paglioli et al., 2004; Jeha et al., 2006) equate patients with a single postoperative seizure to those with long-term persistent epilepsy after surgery. This discrepancy in the definition of seizure-freedom or seizure recurrence resulted in the rather inadequate characterization of outcome in some patients after surgical resection, as the same group of patients achieving seizure-freedom after some postoperative seizures was either considered “seizure-free” or a “surgical failure” in various studies. Fourth, prior evaluations of this “running-down” phenomenon focused mainly on temporal lobe surgery (Bladin, 1987; Salanova et al., 1996, 1999b; Hennessy et al., 2000), therefore, limiting their applicability to an increasing number of extratemporal surgical resections.

This article aims to characterize the prognosis of patients presenting with their first seizure following epilepsy surgery, and provide some practical criteria allowing early identification of recurrent refractory epilepsy after surgery in a large cohort of patients operated on at a single comprehensive epilepsy surgery program. This may assist in avoiding unnecessary repeat presurgical evaluations, as well as in the timely performance of such evaluations in patients who truly need them.


  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Conclusions
  7. Disclosure
  8. References

Patient selection

We analyze the largest epilepsy surgery series published to date, evaluating 915 patients who underwent resective epilepsy surgery at the Cleveland Clinic Epilepsy Center between January 1, 1990 and December 31, 2007. In this study, we include all 276 patients who had at least one seizure anytime beyond the immediate postoperative period (>7 postoperative days) with at least 1 year of subsequent follow-up.

Preoperative patient evaluation protocol

Detailed clinical history was obtained in all patients at their initial outpatient visit to the Cleveland Clinic Epilepsy Center. All had a prolonged (5–7 days) scalp video–electroencephalography (EEG) monitoring where interictal discharges and ictal patterns were recorded and analyzed. In addition, seizures were captured and their semiology was analyzed (Luders et al., 1998). All had a high-resolution brain magnetic resonance imaging (MRI) using either 1.5 T or 3 T systems. Additional investigations included a detailed neuropsychological evaluation and fluorodeoxyglucose (FDG) positron emission tomography (PET) scans. Results of the initial evaluation were all discussed in a multidisciplinary patient management conference where recommendation was made to proceed with a resective surgery, insert a vagus nerve stimulator, perform additional studies [such as targeted MRI with surface coils, and ictal single photon emission computed tomography (SPECT)], or perform an invasive evaluation for further investigation.

Subdural electrodes (SDEs) were implanted for any of the following reasons: (1) to confirm the ictal onset zone if scalp EEG information was either nonlocalizing or inconsistent with other noninvasive investigations; (2) to perform functional mapping preoperatively if the ictal onset zone was thought to be close to eloquent cortex; and (3) whenever a multifocal epileptogenic process was suspected based on multifocal noninvasive EEG or imaging findings.

Surgery and pathologic subgroups

Patients underwent lesionectomy, lobectomy, selective amygdalohippocampectomy, or tailored neocortical resections based on SDE recordings. Resections were temporal (73%), frontal (17%), or in the posterior quadrant (9%).

Based on MRI and surgical pathology, patients were divided into the following etiologic subgroups: (1) malformations of cortical development (MCDs) visible on MRI (16% of cases); (2) MCDs with normal high-resolution MRI (7%); (3) hippocampal sclerosis (HS) visible on MRI (39% of cases); (4) HS with normal high-resolution MRI (3%); (5) tumor (17%); (6) cryptogenic with normal MRI and pathology (8%); (7) other causes including vascular malformations, trauma, and infection (10%). Patients with malignant brain tumors were not included in this study.

For the purposes of this study, a judgment was made as to the completeness of resection based on a combination of EEG and MRI data. In nonlesional cases, resection was considered complete if there was complete removal of the brain tissue covered by the electrodes of ictal onset and most frequent interictal abnormalities defined by invasive electrophysiologic evaluations. In lesional cases, complete removal of the MRI lesion was required. In patients with a lesional MRI who also had SDE or electrocorticography (ECoG), complete resection of electrophysiologic abnormalities was required in addition to removal of the MRI abnormality. The completeness of resection was considered unknown in nonlesional patients who did not undergo invasive EEG recordings (9 of 45 cases), and in the lesional patients for whom a postoperative MRI was either not obtained or missing (30 cases).

Postoperative follow-up

Timing and frequency of seizure recurrences were obtained from medical records review. As per our postsurgical follow-up protocol, all patients were seen at 6 weeks, 6 months (with a 2-h outpatient EEG), 1 year, and then yearly after surgery. Those with recurrent seizures were seen more frequently (every 3–6 months). In between clinic visits, patients were required to report breakthrough seizures. After January 2003, all patient phone calls were entered in the electronic medical records. Prior to that date, paper records of phone or mail contact were kept. When medical records indicated that a seizure occurred between two follow-up visits or phone contacts, the date of recurrence was taken as the midpoint of this period. With patients having the follow-up schedule detailed earlier, the period between documented contacts (visits or phone calls) never exceeded 6 months.


For the purposes of this study, we defined the “first postoperative seizure” as the first seizure occurring beyond the acute postoperative phase (>7 days). Isolated auras were not counted as seizures. We then recorded the dates of the second, third, and last seizures as detailed earlier. Multiple seizures occurring within 24 h were counted only once. A seizure was considered “provoked” if it occurred in the setting of an identifiable trigger such as AED withdrawal, sleep deprivation, fever, or other obvious systemic dysfunction.

Patients were then divided into two groups: seizure-free or refractory at last follow-up, using criteria similar to that of Kwan & Brodie (2000) in their evaluation of refractoriness following initiation of medical therapy. A patient was considered seizure-free if no seizures had occurred for at least 1 year prior to last follow-up. All those with persistent seizures were considered refractory. If a patient had multiple surgeries during the study period, he or she was entered only once and the “last follow-up” was considered 24 h prior to the first reoperation.

Statistical methods

Analysis of seizure recurrence risk

Kaplan-Meier survival analysis was used to calculate the probability of a subsequent seizure after one and then after two postoperative seizures.

Analysis of the predictors of refractoriness

Prior to modeling, the data were summarized with descriptive statistics for each variable including means, medians, and standard deviations for continuous variables and frequencies for categorical variables. For exploratory purposes, an initial univariate analysis was performed using Wilcoxon rank-sum, chi-square, and Fisher’s exact tests to compare seizure-free to refractory cases. Variables analyzed included baseline demographic characteristics (age at surgery, gender), clinical characteristics (age at seizure onset, side of epilepsy, preoperative and postoperative seizure frequency, and semiology including presence of auras and generalized tonic–clonic seizures, and preoperative and postoperative ictal and interictal EEG), surgical variables (such as lobe resected, type of surgery, completeness of resection), MRI, and histopathologic findings. These variables were chosen for their potential impact on seizure outcome following resective surgery. Potentially significant variables were then tested in a multivariate linear regression analysis model. Results were considered statistically significant at the 5% level.


  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Conclusions
  7. Disclosure
  8. References

Risk of any seizure recurrence, refractoriness, and seizure-freedom

The mean overall follow-up duration after surgery was 6.6 years (range 1.1–19 years; standard deviation 4.3) and mean follow-up after the first postoperative seizure was 5.6 years (range 1.1–18.6; standard deviation 4.1). The mean timing of the first postoperative seizure was 12.1 months (median 4.6; range 0.25–114.4 months). Forty patients (14%) did not have any further recurrences (mean follow-up duration of 4.1 years, median 3.0 years, range 1.2–16.4 years), 22 had a total of two seizures (mean follow-up duration of 5.4 years, median 3.7 years, range 2.0–17.0 years), and 17 patients had a total of three seizures by last follow-up (mean follow-up duration of 6.8 years, median 4.4 years, range 1.0–19 years). There was, therefore, a clear tendency in patients with longer follow-up to have more seizures by last follow-up, but this was not statistically significant (p = 0.09). The mean interval to achieve seizure-freedom (time from first to last postoperative seizure) was 17.8 (±3.1) months, and did not correlate with lobe or type of surgery (p = 0.4). Figure 1 illustrates the longitudinal risks of having a subsequent seizure following one (Fig. 1A) and two (Fig. 1B) postoperative seizures.


Figure 1.   Kaplan-Meier curve illustrating probability of having a subsequent seizure following one (A) or two (B) seizures occurring after epilepsy surgery. Graph shows 95% confidence intervals. n, number of patients at risk.

Download figure to PowerPoint

Ninety-five patients became seizure-free, with a mean duration of seizure-freedom of 4.9 ± 0.4 years (range 1.1–16.4 years) by last follow-up. The remaining 181 cases were having persistent seizures with a mean frequency of 14.4 ± 2.6 seizures per month. Figure 2 shows the distribution of refractory versus seizure-free cases after one (Fig. 2A), two (Fig. 2B), or three (Fig. 2C) seizures.


Figure 2.   Probability of having any subsequent seizure by last follow-up in patients who have had one (A), two (B), or three (C) postoperative seizures.

Download figure to PowerPoint

Given the possibility that patients with just one or rare yearly seizures might have been unfairly considered refractory using our strict definition of at least 1 year of seizure-freedom by last follow-up, we also evaluated the eventual seizure outcome of our “refractory” group using the Wieser criteria (Wieser et al., 2001). Eleven percent of the 181 patients in our “refractory group” were classified under class 3 (1–3 seizure-days per year), with the remainder having more frequent seizures: 26% were in outcome class 4 defined by four seizure days per year to 50% reduction of baseline seizure days, 27% in outcome class 5 defined by <50% reduction of baseline seizure days to 100% increase of baseline seizure days, and 36% doing significantly worse in outcome class 6 defined as more than 100% increase of baseline seizure days.

Postoperative management of antiepileptic drugs (AEDs)

In our center, postoperative AED management is individualized and no clear protocol is in place. However, most patients were usually kept on their AEDs for at least 2 years after surgery. In patients who remained seizure-free for 2 years and had no evidence of residual epileptogenicity (interictal spiking on postoperative routine EEG or residual lesion on MRI), it was at the epileptologist’s discretion to offer gradual AED withdrawal. In our cohort, an attempt for withdrawal or at least reduction of AEDs was done in 65 patients at some point during the study period. A successful discontinuation of AEDs was achieved in five patients following reoperation, so those were excluded from subsequent analysis of any interactions between AED use and eventual seizure outcome. Of the remaining 60 cases, AEDs were reduced in 35 patients and discontinued in 25. By the last follow-up, six of the patients for whom AEDs were discontinued were still off AEDs, whereas the others were back on treatment. The change in AED usage was planned in coordination with the treating physician in most cases, with only 15 patients taking themselves off AED on their own. In 3 of the 60 patients, the first postoperative seizure occurred more than a year after discontinuation of AEDs, so it was difficult to attribute seizure recurrence to AED withdrawal per se. In the remaining 57 patients, seizures recurred in the midst of the AED modification or soon thereafter, so it was considered in subsequent analyses that those patients had an initial recurrence that was “provoked” by AED withdrawal.

Planned and unplanned AED withdrawal trials were equally likely to correlate with a favorable eventual outcome (rate of eventual seizure freedom was 62% in planned and 80% in unplanned AED withdrawal trials; p = 0.19). However, there was a higher tendency for patients with recurrences occurring following complete AED discontinuation to regain seizure-freedom as opposed to those with breakthrough seizures following a simple dose reduction (80% vs. 57%, p = 0.06).

Predictors of refractoriness

Tables 1–3 show the characteristics of the overall cohort, and the refractory versus eventually seizure-free group. After multivariate linear regression analysis, the following variables predicted a higher risk of refractoriness: having both the first and second seizure recurrence within 6 months of surgery [odds ratio (OR) 4.04; 95% confidence interval (CI) 2.05–8.40; p = 0.0001], having un unprovoked initial recurrence (OR 3.92; 95% CI 2.13–7.30; p < 0.0001), and ipsilateral spikes on the routine six-month postoperative EEG (OR 2.05; 95% CI 1.10–3.88; p = 0.025) (overall model p-value <0.0001). Figure 3 details the outcome of our cohort as related to those three characteristics. In addition, all patients with cryptogenic epilepsy (as defined by normal MRI and nonspecific pathology) were refractory at last follow-up.

Table 1.   Univariate analysis of preoperative predictors of refractoriness
 Overall group (N = 276)Refractory (N = 181)Seizure-free (N = 95)
  1. p-Value is for Wilcoxon rank-sum, chi-square, or Fisher’s exact tests as appropriate, MRI, magnetic resonance imaging.

Baseline characteristics
 Lobe of first surgery (p < 0.001)
  Temporal lobe surgery203126 (62%)77 (38%)
  Frontal lobe surgery4841 (85%)7 (15%)
  Posterior quadrant resection2514 (56%)11 (44%)
 Gender (p = 0.38)
  Female13592 (68%)43 (32%)
  Male14189 (63%)52 (37%)
 Side of surgery (p = 0.61)
  Left154103 (67%)51 (33%)
  Right12278 (64%)44 (36%)
 Mean age at surgery (years) (p < 0.001)31.8 ± 0.8   29.5 ± 1.0 36.1 ± 1.3
 Age at surgery <18 years old (p = 0.002)
  Yes3833 (87%)5 (13%)
  No237148 (62%)89 (38%)
 Mean age at seizure onset (years) (p = 0.10)14.8 ± 0.8   13.8 ± 1.0 16.8 ± 1.4
 Mean monthly seizure frequency (p = 0.02)31.1 ± 4.1   37.9 ± 5.0 18.1 ± 6.9
 Etiology (p < 0.001)
  Tumor4727 (57%)20 (43%)
  Hippocampal sclerosis (nl MRI)73 (43%)4 (57%)
  Hippocampal sclerosis (abl MRI)10869 (64%)39 (36%)
  Malformation of cortical development (nl MRI)1812 (67%)6 (33%)
  Malformation of cortical development (abl MRI)4432 (73%)12 (27%)
  Cryptogenic2222 (100%)0
  Vascular91 (12%)8 (88%)
  Other2115 (71%)6 (29%)
 Cryptogenic epilepsy (p < 0.0001)
  Yes2222 (100%)0 (0%)
  No254159 (63%)95 (37%)
Initial seizure semiology
 Generalized tonic–clonic seizures (p = 0.36)
  Present226151 (67%)75 (33%)
  Absent5030 (60%)20 (40%)
 Multiple seizure semiologies (p = 0.36)
  Present3118 (58%)13 (42%)
  Absent245163 (67%)82 (33%)
 Initial MRI (p < 0.001)
  Normal4536 (80%)9 (20%)
  Unilateral abnormality203120 (59%)83 (41%)
  Bilateral abnormalities2623 (88%)3 (12%)
Initial electroencephalogram
 Contralateral temporal sharp waves (p = 0.011)
  Present5141 (80%)10 (20%)
  Absent225140 (62%)85 (38%)
 Contralateral seizure patterns (p = 0.31)
  Present266173 (65%)93 (35%)
  Absent108 (80%)2 (20%)
 Nonlocalizable ictal patterns (p = 0.19)
  Present1915 (79%)4 (21%)
  Absent257166 (65%)91 (35%)
 Multiple ictal patterns (p = 0.51)
  Present1511 (73%)4 (27%)
  Absent261170 (65%)91 (35%)
Subdural electrode usage
 Subdural electrodes used (p < 0.001)
  Yes9978 (79%)21 (21%)
  No175103 (59%)72 (41%)
 Reason to use subdural electrodes (p = 0.05)
  Functional mapping2214 (64%)8 (36%)
  Ictal localization7865 (83%)13 (17%)
Table 2.   Univariate analysis of postoperative predictors of refractoriness
 Overall group (N = 276)Refractory (N = 181)Seizure-free (N = 95)
  1. p-Value is for Wilcoxon rank-sum, chi-square, or Fisher’s exact tests as appropriate, EEG, electroencephalogram.

Completeness of resection (p = 0.004)
 Complete13780 (58%)57 (42%)
 Incomplete10076 (76%)24 (24%)
Mean timing of first recurrence (months) (p = 0.009)12.1 ± 1.2   9.9 ± 1.4 16.3 ± 2.0
Timing of first recurrence (p < 0.001)
 <6 months160118 (74%)42 (26%)
 ≥6 months11663 (54%)53 (46%)
Trigger for recurrence (p < 0.001)
 Antiepileptic drug withdrawal5720 (35%)37 (65%)
 Stress2912 (41%)17 (59%)
 None188148 (79%)40 (21%)
Ipsilateral spike on 6-month EEG (p < 0.001)
 Yes11592 (80%)23 (20%)
 No14278 (55%)64 (45%)
Contralateral spike on 6-month EEG (p = 0.03)
 Yes2924 (83%)5 (17%)
 No222146 (64%)82 (36%)
Table 3.   Analysis of further recurrences after first seizure
 Seizure-freeNot seizure-freep-value
  1. p-value is for Wilcoxon rank-sum, chi-square, or Fisher’s exact tests as appropriate.

Mean timing of first recurrence (months) (N = 276) 16.3 ± 2.0    9.9 ± 1.40.009
 Timing of first seizure
  <6 months42 (26%)118 (74%)<0.001
  ≥ 6 months53 (46%)63 (54%)
 Two seizures occurred within 6 months?
  Yes15 (14%)94 (86%)<0.001
  No80 (48%)87 (52%)
Mean timing of second seizure (months) (N = 237) 28.1 ± 3.4  15.2 ± 2.00.001
 Timing of second seizure
  <6 months15 (14%)94 (86%)<0.001
  ≥6 months44 (34%)84 (66%)
Mean interval to second recurrence (months) (N = 235)   8.8 ± 2.3    5.5 ± 1.30.21

Figure 3.   Distribution of cases with refractory postoperative seizures according to whether they fulfilled the three refractoriness risk factors identified in this study.

Download figure to PowerPoint

Variables that lost their significance after multivariate analysis were correlated with the identified independent outcome indicators. Patients with frontal lobe surgery were more likely to have ipsilateral spikes on their postoperative EEG (72% of frontal as opposed to 37% with temporal and 54% with posterior quadrant surgery; p = 0.0001), as were patients <18 years old (74% with residual spikes as opposed to 40% of older cases; p = 0.0003), and those with a higher baseline seizure frequency (p = 0.02). More unprovoked seizures occurred in patients with incomplete resection (78% in incomplete resection vs. 59% otherwise; p = 0.002), in those with contralateral preoperative interictal spiking (82% of cases with contralateral temporal sharp waves as opposed to 66% of those without; p = 0.02), in those with either a normal MRI or one showing bilateral abnormalities as opposed to those with a unilateral lesion (84%, 92%, and 62%, respectively; p = 0.0001), and in patients who needed SDE implantation preoperatively (p < 0.0001). Although 83% of patients who required the use of SDE implantation for ictal localization became refractory as opposed to 64% of those who needed grids for functional mapping (p < 0.001), this variable also lost its significance on multivariate analysis, probably because it was very tightly correlated with etiology: SDEs were implanted primarily for ictal localization in 100% of the cryptogenic cases as opposed to 76% otherwise (p = 0.01).


  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Conclusions
  7. Disclosure
  8. References

Our study investigates a large cohort of patients who underwent surgery for the treatment of medically intractable focal epilepsy, and identifies clinical and EEG predictors of drug-resistant postoperative seizure recurrence.

Seizure recurrence risk

Our data allow the following observations:

First, the odds of subsequent seizures following a recurrence are significant, with the period of highest risk being within few months of the initial event. After their first seizure, 50% (±3% survival standard error) of our patients had the second seizure within a month, 77% (±3%) had one within a year, with the risk slowing down to additional 2–3% increments every two subsequent years (Fig. 1A). These recurrence rates are significantly higher than those seen after a first seizure in newly diagnosed epilepsy (34–46% at 10 years) (Hauser et al., 1990; Krumholz et al., 2007; Arthur et al., 2008), and closer to those of previously reported epilepsy surgery cohorts (Wingkun et al., 1991; McIntosh et al., 2005). The 1-year recurrence risk in patients having acute postoperative seizures within 28 days following temporal lobectomy was 69–83% in one study (McIntosh et al., 2005), concordant with the 77% 1-year recurrence rate seen in our cohort. The likelihood of persistent seizures 6 years following an initial recurrence occurring within 1 year of surgery was 80% in another group (Wingkun et al., 1991), also concordant with the 87% recurrence risk of our patients by 6 years after a first seizure. As opposed to previous smaller cross-sectional reports, our study’s use of a longitudinal method of analysis provides longitudinal estimates of the recurrence risk, applicable throughout the postoperative follow-up period regardless of the timing of the first postoperative seizure. In addition, while confirming the belief that most recurrences occur early, our findings raise concerns that the risk of subsequent seizures persists, even as late as 10 years following the initial seizure.

Second, the recurrence rate increases dramatically after a second seizure, with 50% (±3%) having a third seizure within 2 weeks, 78% (±3%) within 2 months, and 83% (±2%) within 6 months. The risk then slows down to additional 2–3% increments every 6 months to stabilize at around 90% by 2–3 years (Fig. 1B). The mere occurrence of a second seizure may reflect a patient population with a higher epileptogenic potential and a lower threshold to develop more seizures, faster. This hypothesis is further supported by the higher proportion of patients eventually developing refractory epilepsy among those who had two as opposed to one postoperative seizure in our cohort (Fig. 2).

Third, a seizure—and even many seizures—occurring after surgery is not equivalent to recurrence of refractory epilepsy. Fourteen percent of our patients never had a second seizure, and 20–34% eventually became seizure-free (Fig. 2), whereas 11 percent of patients in our refractory group had a Wieser classification of 3 or less, signifying 1–3 seizures per year. These rates are comparable to previously reported odds of the “running-down” phenomenon defined as the late remission of postsurgical seizures (Rasmussen, 1970), occurring in 3.2 to 32% of temporal lobe surgery cases (Rasmussen, 1970; Bladin, 1987; Ficker et al., 1999; Wieser et al., 2003; Jeong et al., 2005; Kelemen et al., 2006). The frequency of seizures during the running-down interval in temporal lobe epilepsy was reported as up to several per month, with a seizure-free state usually achieved within 2 years (Elwes et al., 1985, 1991; Rougier et al., 1992; Salanova et al., 1996, 1999a,b; Hennessy et al., 2000). Our finding that eventually seizure-free patients may have had up to 10 postoperative seizures before achieving seizure-freedom within a mean of 17.8 months, irrespective of the resected lobe, suggests that this phenomenon is equally significant in extratemporal lobe epilepsy. The most accepted explanation is a dekindling effect, a process opposite to secondary epileptogenesis, where the induced synaptic dysfunction gradually declines in the surrounding epileptogenic cortex after the dominant pacemaker resection, and eventually “runs itself down” (Rasmussen, 1970; Bladin, 1987; Salanova et al., 1996). The main challenge has been the timely identification of patients who will “run down” as opposed to those with truly recurrent refractory epilepsy.

Postoperative management of antiepileptic drugs (AEDs)

Given the retrospective nature of this study, it is difficult to provide solid conclusions pertaining to an issue as delicate as AED use after surgery and its implications on eventual outcome. Within those constraints, our review suggests that seizures recurring after AED discontinuation may be easier to control than those recurring after simple dose reductions, a finding that may reflect underlying epilepsy processes of varying thresholds for generation of synchronized ictal discharges and refractoriness. Nevertheless, six of our patients with seizures following AED withdrawal eventually regained seizure-freedom and even successfully discontinued medical therapy. The small number of this interesting patient subgroup in our cohort limits the possibility of any meaningful statistical analyses or conclusions. Further prospective studies of postoperative AED management are needed to elucidate several remaining questions pertaining to this topic.

Predictors of refractoriness

Three easily assessed factors determined the risk of refractoriness in this study: 95% of our patients who had two unprovoked seizures within 6 months of surgery and showed ipsilateral spikes on their postoperative EEG were still having persistent seizures years later, as opposed to only 32% of those who had none of the three risk factors (i.e., patients with no or one provoked seizure within 6 months of surgery and no spikes on postoperative EEG). Approximately 40–50% of patients with one, and 70–80% of those with two poor prognostic indicators still had refractory seizures at last follow-up (Fig. 3). Cryptogenic epilepsy (as defined by normal MRI and subsequently nonspecific findings on pathologic examination of the resected cortex) was another independent indicator of poor eventual outcome, with 100% of cryptogenic epilepsy patients eventually redeveloping refractoriness following an initial breakthrough seizure.

Unprovoked initial recurrence

When the initial seizure was unprovoked, our patients were four times more likely to become refractory. This is consistent with a prior report finding a lower recurrence risk among patients when early postoperative seizures occurred in the presence of seizure precipitants (McIntosh et al., 2005), and another one documenting lower chances of an “excellent outcome” (27.8%) after temporal lobectomy when the initial recurrent seizure was unprovoked as opposed to provoked (72%) (Radhakrishnan et al., 2003).

A likely hypothesis is that cortical tissue requiring a trigger to develop an ictal event has a higher epileptogenic threshold than the brain tissue that is capable of synchronizing spontaneously. The correlation of unprovoked recurrences with cryptogenic epilepsy, normal imaging, and prior need for invasive EEG in our group further underscores a correlation with a poorly localized and diffuse epileptogenic zone.

Two seizures within six postoperative months

Having multiple seizures within the first postoperative year has been associated with worse long-term outcome in many (Garcia et al., 1991; Janszky et al., 2005) but not all studies of recurrence following temporal lobectomy (Wingkun et al., 1991). Similarly, a better prognosis has been suggested with late as opposed to early recurrences following temporal (Wingkun et al., 1991; Radhakrishnan et al., 2003; Schwartz et al., 2006) and frontal (Jeha et al., 2007) lobe surgery. Our cohort defines a specific cut-off for “multiple” and “early” seizures, predicting refractoriness as two or more seizures within 6 months of surgery (OR 4.04; 95% CI 2.05–8.40; p = 0.0001). Potential explanations include an uninterrupted baseline epileptogenic focus continuing to exhibit unabated seizures, as would occur following an incomplete resection; or alternatively, a highly epileptogenic brain where a secondary epileptogenic focus is quickly uncovered after the first resection. Either way, it is understandable to expect the postoperative epilepsy syndrome to be refractory.

Ipsilateral spikes on postoperative EEG

Many studies correlated spikes on a 6-month postoperative EEG with any postoperative seizure recurrence (Di Gennaro et al., 2004; Jeha et al., 2006, 2007; Jehi, 2008; Ghacibeh et al., 2009). A specific correlation with the “running-up” phenomenon and worsening long-term control in failed temporal lobectomies has been suggested (Hildebrandt et al., 2005; Kelemen et al., 2006). Our report confirms this correlation and highlights the significance of postoperative EEG as an easy noninvasive tool with potentially high prognostic implications.

Cryptogenic epilepsy etiology

Several studies have previously highlighted the poor long-term seizure outcome following resective epilepsy surgery in patients with cryptogenic epilepsy. A normal high-resolution MRI is one of the most significant poor prognostic indicators, especially following frontal lobe resections (Jeha et al., 2007). Nonspecific gliosis/normal histopathology has also been shown to predict late seizure recurrence (Yoon et al., 2003; McIntosh et al., 2004; Jeha et al., 2006). Our study confirms that patients who fulfill both conditions, that is, normal MRI and normal histopathology, represent an especially vulnerable patient population. Possible explanations include a diffuse/poorly localized and as such incompletely resected epileptogenicity at the time of surgery. Alternatively, this cryptogenic epilepsy may be due to more diffuse molecular or cellular epileptogenic mechanisms that have a low threshold to redevelop refractoriness in remaining brain tissue that was believed to be “uninvolved” at the time of surgical resection.


  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Conclusions
  7. Disclosure
  8. References

The majority of patients presenting with their first postoperative seizure will have at least one more seizure. The period of highest risk is within a few months of the initial event. Up to one-third of these patients will eventually become seizure-free. The risk for refractoriness can be assessed by easily ascertainable measures at six postoperative months: Patients who have two or more unprovoked seizures within that timeframe and show ipsilateral spikes on routine EEG are at highest risk of becoming refractory, as opposed to those for whom the first seizure is provoked, occurs later after surgery, and have no epileptiform activity on their 6 month postoperative EEG.


  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Conclusions
  7. Disclosure
  8. 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. None of the authors has any conflict of interest to disclose.


  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Conclusions
  7. Disclosure
  8. References
  • Abou-Khalil BW. (2004) Do immediate postoperative seizures predict surgical failure after frontal lobe epilepsy surgery? Epilepsy Curr 4:5152.
  • Arthur TM, DeGrauw TJ, Johnson CS, Perkins SM, Kalnin A, Austin JK, Dunn DW. (2008) Seizure recurrence risk following a first seizure in neurologically normal children. Epilepsia 49:19501954.
  • Bladin PF. (1987) Post-temporal lobectomy seizures. Clin Exp Neurol 24:7783.
  • Di Gennaro G, Quarato PP, Sebastiano F, Esposito V, Onorati P, Mascia A, Romanelli P, Grammaldo LG, Falco C, Scoppetta C, Eusebi F, Manfredi M, Cantore G. (2004) Postoperative EEG and seizure outcome in temporal lobe epilepsy surgery. Clin Neurophysiol 115:12121219.
  • Elsharkawy AE, Behne F, Oppel F, Pannek H, Schulz R, Hoppe M, Pahs G, Gyimesi C, Nayel M, Issa A, Ebner A. (2008) Long-term outcome of extratemporal epilepsy surgery among 154 adult patients. J Neurosurg 108:676686.
  • Elwes RD, Chesterman P, Reynolds EH. (1985) Prognosis after a first untreated tonic-clonic seizure. Lancet 2:752753.
  • Elwes RD, Dunn G, Binnie CD, Polkey CE. (1991) Outcome following resective surgery for temporal lobe epilepsy: a prospective follow up study of 102 consecutive cases. J Neurol Neurosurg Psychiatr 54:949952.
  • Ficker DM, So EL, Mosewich RK, Radhakrishnan K, Cascino GD, Sharbrough FW. (1999) Improvement and deterioration of seizure control during the postsurgical course of epilepsy surgery patients. Epilepsia 40:6267.
  • Garcia PA, Barbaro NM, Laxer KD. (1991) The prognostic value of postoperative seizures following epilepsy surgery. Neurology 41:15111512.
  • Ghacibeh GA, Smith JD, Roper SN, Gilmore R, Eisenschenk S. (2009) Seizure recurrence following epilepsy surgery: is post-operative EEG helpful? Seizure 18:193196.
  • Hauser WA, Rich SS, Annegers JF, Anderson VE. (1990) Seizure recurrence after a 1st unprovoked seizure: an extended follow-up. Neurology 40:11631170.
  • Hennessy MJ, Elwes RD, Binnie CD, Polkey CE. (2000) Failed surgery for epilepsy. A study of persistence and recurrence of seizures following temporal resection. Brain 123 Pt 12, 24452466.
  • Hennessy MJ, Elwes RD, Rabe-Hesketh S, Binnie CD, Polkey CE. (2001) Prognostic factors in the surgical treatment of medically intractable epilepsy associated with mesial temporal sclerosis. Acta Neurol Scand 103:344350.
  • Hildebrandt M, Schulz R, Hoppe M, May T, Ebner A. (2005) Postoperative routine EEG correlates with long-term seizure outcome after epilepsy surgery. Seizure 14:446451.
  • Janszky J, Schulz R, Ebner A. (2003) Clinical features and surgical outcome of medial temporal lobe epilepsy with a history of complex febrile convulsions. Epilepsy Res 55:18.
  • Janszky J, Pannek HW, Janszky I, Schulz R, Behne F, Hoppe M, Ebner A. (2005) Failed surgery for temporal lobe epilepsy: predictors of long-term seizure-free course. Epilepsy Res 64:3544.
  • Jeha LE, Najm IM, Bingaman WE, Khandwala F, Widdess-Walsh P, Morris HH, Dinner DS, Nair D, Foldvary-Schaeffer N, Prayson RA, Comair Y, O’Brien R, Bulacio J, Gupta A, Luders HO. (2006) Predictors of outcome after temporal lobectomy for the treatment of intractable epilepsy. Neurology 66:19381940.
  • Jeha LE, Najm I, Bingaman W, Dinner D, Widdess-Walsh P, Luders H. (2007) Surgical outcome and prognostic factors of frontal lobe epilepsy surgery. Brain 130:574584.
  • Jehi LE. (2008) Mesial temporal lobectomy: post-surgical seizure frequency. In LüdersHO (Ed) Textbook of epilepsy surgery. Informa, United Kingdom, pp. 12231235.
  • Jehi LE, O'Dwyer R, Najm I, Alexopoulos A, Bingaman W. (2009) A longitudinal study of surgical outcome and its determinants following posterior cortex epilepsy surgery. Epilepsia 50:20402052.
  • Jeong SW, Lee SK, Hong KS, Kim KK, Chung CK, Kim H. (2005) Prognostic factors for the surgery for mesial temporal lobe epilepsy: longitudinal analysis. Epilepsia 46:12731279.
  • Kelemen A, Barsi P, Eross L, Vajda J, Czirjak S, Borbely C, Rasonyi G, Halasz P. (2006) Long-term outcome after temporal lobe surgery – prediction of late worsening of seizure control. Seizure 15:4955.
  • Kelley K, Theodore WH. (2005) Prognosis 30 years after temporal lobectomy. Neurology 64:19741976.
  • Krumholz A, Wiebe S, Gronseth G, Shinnar S, Levisohn P, Ting T, Hopp J, Shafer P, Morris H, Seiden L, Barkley G, French J, Quality Standards Subcommittee of the American Academy of Neurology, American Epilepsy Society. (2007) Practice parameter: evaluating an apparent unprovoked first seizure in adults (an evidence-based review): report of the Quality Standards Subcommittee of the American Academy of Neurology and the American Epilepsy Society. Neurology 69, 19962007.
  • Kwan P, Brodie MJ. (2000) Early identification of refractory epilepsy. N Engl J Med 342:314319.
  • Luders H, Acharya J, Baumgartner C, Benbadis S, Bleasel A, Burgess R, Dinner DS, Ebner A, Foldvary N, Geller E, Hamer H, Holthausen H, Kotagal P, Morris H, Meencke HJ, Noachtar S, Rosenow F, Sakamoto A, Steinhoff BJ, Tuxhorn I, Wyllie E. (1998) Semiological seizure classification. Epilepsia 39:10061013.
  • Malla BR, O’Brien TJ, Cascino GD, So EL, Radhakrishnan K, Silbert P, Marsh WR. (1998) Acute postoperative seizures following anterior temporal lobectomy for intractable partial epilepsy. J Neurosurg 89:177182.
  • McIntosh AM, Kalnins RM, Mitchell LA, Fabinyi GC, Briellmann RS, Berkovic SF. (2004) Temporal lobectomy: long-term seizure outcome, late recurrence and risks for seizure recurrence. Brain 127:20182030.
  • McIntosh AM, Kalnins RM, Mitchell LA, Berkovic SF. (2005) Early seizures after temporal lobectomy predict subsequent seizure recurrence. Ann Neurol 57:283288.
  • Paglioli E, Palmini A, Paglioli E, Da Costa JC, Portuguez M, Martinez JV, Calcagnotto ME, Hoefel JR, Raupp S, Barbosa-Coutinho L. (2004) Survival analysis of the surgical outcome of temporal lobe epilepsy due to hippocampal sclerosis. Epilepsia 45:13831391.
  • Radhakrishnan K, So EL, Silbert PL, Cascino GD, Marsh WR, Cha RH, O’Brien PC. (2003) Prognostic implications of seizure recurrence in the first year after anterior temporal lobectomy. Epilepsia 44:7780.
  • Rasmussen T. (1970) Modern problems of pharmacopsychiatry. In NiedermeyerE (Ed) The neurosurgical treatment of focal epilepsy. Vol. 4. Krager, New York, pp. 306325.
  • Rougier A, Dartigues JF, Commenges D, Claverie B, Loiseau P, Cohadon F. (1992) A longitudinal assessment of seizure outcome and overall benefit from 100 cortectomies for epilepsy. J Neurol Neurosurg Psychiatr 55:762767.
  • Salanova V, Andermann F, Rasmussen T, Olivier A, Quesney L. (1996) The running down phenomenon in temporal lobe epilepsy. Brain 119(Pt 3), 989996.
  • Salanova V, Markand O, Worth R. (1999a) Longitudinal follow-up in 145 patients with medically refractory temporal lobe epilepsy treated surgically between 1984 and 1995. Epilepsia 40:14171423.
  • Salanova V, Rasmussen T, Andermann F. (1999b) The running down phenomenon in temporal lobe epilepsy. Adv Neurol 81:165169.
  • Schwartz TH, Jeha L, Tanner A, Bingaman W, Sperling MR. (2006) Late seizures in patients initially seizure free after epilepsy surgery. Epilepsia 47:567573.
  • Sindou M, Guenot M, Isnard J, Ryvlin P, Fischer C, Mauguiere F. (2006) Temporo-mesial epilepsy surgery: outcome and complications in 100 consecutive adult patients. Acta Neurochir (Wien) 148:3945.
  • Spencer SS, Berg AT, Vickrey BG, Sperling MR, Bazil CW, Shinnar S, Langfitt JT, Walczak TS, Pacia SV, Multicenter Study of Epilepsy Surgery. (2005) Predicting long-term seizure outcome after resective epilepsy surgery: the multicenter study. Neurology 65, 912918.
  • Tigaran S, Cascino GD, McClelland RL, So EL, Richard Marsh W. (2003) Acute postoperative seizures after frontal lobe cortical resection for intractable partial epilepsy. Epilepsia 44:831835.
  • Wieser HG, Blume WT, Fish D, Goldensohn E, Hufnagel A, King D, Sperling MR, Luders H, Pedley TA, Commission on Neurosurgery of the International League Against Epilepsy (ILAE). (2001) ILAE commission report. proposal for a new classification of outcome with respect to epileptic seizures following epilepsy surgery. Epilepsia 42:282286.
  • Wieser HG, Ortega M, Friedman A, Yonekawa Y. (2003) Long-term seizure outcomes following amygdalohippocampectomy. J Neurosurg 98:751763.
  • Wingkun EC, Awad IA, Luders H, Awad CA. (1991) Natural history of recurrent seizures after resective surgery for epilepsy. Epilepsia 32:851856.
  • Yoon HH, Kwon HL, Mattson RH, Spencer DD, Spencer SS. (2003) Long-term seizure outcome in patients initially seizure-free after resective epilepsy surgery. Neurology 61:445450.