Epilepsy is a chronic disorder; therefore, it is important to assess the long-term prognosis for possible irreversible effects after surgical treatment of epilepsy. Whether or not seizures resolve after surgery is undoubtedly an important parameter in assessing the efficacy of surgery, and whether or not antiepileptic drugs (AEDs) must be continued after surgery and when to discontinue them are issues that need to be communicated and agreed upon between physicians and patients.
I. Frequency and risk factors for post-surgical seizure recurrence
Previous studies have confirmed the short-term (1-5 years) efficacy and safety after resection of epileptic lesions. In a systematic review and Meta-analysis of 32 studies that included 2250 surgical patients by Engel et al. the results showed that 65% of patients were seizure-free, 21% improved, and 14% had no change compared to pre-surgery after anterior medial temporal lobe resection.Edwards et al. (The prognosis was assessed as Engel classification grade 1 and found that seizure absence was more frequent in those with complete resection of the MCD on one side. MRI examinations showing megalencephalic or bilateral MCD changes on one hemisphere often suggested post-surgical seizures.
Arzimanoglou et al. performed a retrospective analysis of the surgical criteria and prognosis of 20 patients with Sturge-Weber syndrome (SWS) presenting as intractable epilepsy admitted to 2 centers from 1972-1990. The age at onset of epilepsy in the enrolled patients ranged from 2 months to 12 years and from 8 months to 34 years at the time of surgery. Among them, one corpus callosotomy, five hemisphere resections, and 14 cortical resections were performed, and almost all of them were effective; 13 cases were seizure-free after surgery, especially the five children with previous mild hemiparesis who had hemisphere resections had their seizures completely disappeared after surgery. Because of the relatively poor outcome of frontal lobe epilepsy surgery, Janszky et al. retrospectively analyzed the clinical data of 61 patients who underwent preoperative evaluation and frontal lobe resection in an attempt to identify factors that might predict the outcome of surgery. The mean age of this group of patients was 19.2 years, with a follow-up of 0.5-5 years after surgery; the histopathology was MCD (57.4%), tumor (16.4%), and other lesions (26.2%), respectively. Comparing the 30 patients without seizures after surgery with the 31 patients with seizures, three preoperative and two postoperative variables, i.e., extensive epileptiform discharges, extensive slow waves, use of intracranial electrodes, incomplete resection of abnormal MRI areas, and postoperative epileptiform discharges, were found to be associated with poor patient prognosis; the only preoperative factor associated with seizure-free outcome was the absence of generalized EEG signs. Multivariate analysis showed that lack of generalized EEG signs could be an independent predictor of prognosis; somatosensory aura, secondary generalized seizures, and absence of MRI abnormalities were all additional independent risk factors for poor surgical outcome. The value of EEG in predicting surgical outcome in frontal lobe epilepsy has been emphasized, and the absence of extensive EEG abnormalities is most suggestive of absence of seizures.
Yoon et al. retrospectively analyzed the likelihood of and risk factors for subsequent seizure recurrence in 175 patients who underwent epilepsy surgery from 1972 to 1992 in the first year after surgery. The prognostic indicators set included the risk of recurrence, the presence of aura in seizure-free patients, and the frequency of seizures in recurring patients. At a mean follow-up of 8.4 years, 63% were consistently recurrence-free, and their seizure-free rates at 3, 5 and 10 years after surgery were (83±6)%, (72±7)% and (56±9)%, respectively. After correction for age at surgery, preoperative epilepsy duration and site of resection, no pathological abnormality was associated with an increased risk of recurrence compared with medial temporal lobe sclerosis or other lesions, with a risk ratio (HR) of 2.38. Among seizure-free patients, a preoperative duration of ≥20 years was associated with an increased risk of postoperative aura (HR 3.55); recurrent patients had ≤1 seizure per year; no pathological abnormality, earlier recurrence was associated with higher frequency of seizures after surgery. This study showed that patients who remained seizure-free in the first year after surgery decreased from 63% to 56% after 10 years, but half of the patients with recurrence had at most 1 seizure per year; a long duration of disease before surgery and no pathological abnormality were found to predict the possibility of reoccurrence.
Janszky et al. observed and analyzed 86 patients with temporal lobe epilepsy who still had seizures within 6 months after surgery and up to 2 years later and predictive factors, and found that 1/3 of the patients still had seizures 2 to 6 months after surgery and were seizure-free after 2 years. Their results concluded that factors such as abnormal changes in pre-surgical MRI presentation and secondary generalized seizures were associated with continued seizures after surgery; whereas occasional seizures after surgery, focal temporal lobe spikes on one side, and no secondary generalized seizures before surgery suggested a seizure-free prognosis.
Tonini et al. conducted a Meta-analysis by Medline search of 47 papers published from 1984-2001 on epilepsy surgery including assessment of post-surgical seizure outcome in a total of 3511 patients (175 of whom were children). The inclusion criteria for this study were that the sample size of each literature should not be less than 30 cases, 90% had MRI data and at least 1 year of follow-up after surgery, and the main abnormalities in MRI and pathology were medial temporal lobe sclerosis, tumor, and MCD. analysis showed that the advantage ratios (OR) for febrile seizures, medial temporal lobe sclerosis, tumor, abnormal MRI, EEG and/or MRI consistency, and extensive surgical resection 0.48, 0.47, 0.58, 0.44, 0.52, and 0.24, respectively, strongly predicted post-surgical seizure remission (positive prediction); whereas post-surgical epileptic discharges (OR 2.41) and intracranial electrode monitoring (OR 2.72) predicted poor outcome (negative prediction); neuronal migration defects (OR 1.51), central nervous system infection (OR 1.37), vascular lesions (OR 1.51), interictal spikes (OR 0.55), and resected side (OR 1.7) did not affect the likelihood of seizure remission after surgery.
Tellez-Zenteno et al. searched 76 publications on long-term prognosis (>5 years) of patients (7343 cases) with different epilepsy surgery modalities included in Medline, Index Medicus and Cochrane databases since 1991 and performed Meta-analysis. The results showed a lack of randomization in these studies, with only 6 studies having a control group. Those who underwent focal resection were subdivided into temporal lobe surgery (3895 cases), combined temporal-external temporal lobe surgery (2334 cases), frontal lobe surgery (486 cases), combined external temporal lobe surgery (169 cases), hemispherectomy (169 sides), parietal lobe surgery (82 cases), and occipital lobe surgery (35 cases), with 66%, 59%, 27%, 34%, 61%, 46% and 46%; the seizure-free probability in the nonfocal resection group was 35% for corpus callosotomy and 16% for multiple subchondral transverse fiber dissection (MST). Univariate analysis of those who underwent temporal lobe surgery showed the highest long-term seizure-free rate after temporal lobe tumor surgery (76%) and the lowest long-term seizure-free rate in those who were >50 years old at the time of surgery (41%), operated before 1980 (54%), or followed up for more than 10 years (45%). Factors such as age at the time of temporal lobectomy, duration of follow-up, and the use of a different prognostic classification system (e.g., Engel prognostic classification or other classification methods) were significantly associated with patient prognosis, and the long-term (≥5 years) seizure-free rate after surgery was similar to that reported in short-term (1-5 years) controlled studies (63.2%). Univariate analysis of patients who underwent combined temporal lobe – extratemporal lobe surgery showed that the OR for seizure-free patients was significantly higher in patients with vascular malformations, in children, at more than 10 years of follow-up, and in those operated on after 1980; conversely, it was significantly lower in those with MCD. The seizure-free rate in children after hemispherectomy was 70-80% and remained around 60% even 5 years after surgery. In the linear regression analysis model, only the age of surgery (P=0.001) and the prognostic classification method (P=0.018) were significant predictors of the low seizure-free rate reported for earlier epilepsy surgery and the higher seizure-free rate defined by the Engel method. Poor prognosis is often associated with incomplete resection of the epileptogenic lesion, and epileptic discharges tend to spread widely and rapidly because of the proximity of the lesion to key functional cortical areas of the brain or the excessive extent of the lesion in the frontal lobe. Heterogeneity may explain differences in prognosis, such as etiology, involvement of only the frontal lobe and/or adjacent structures by the epileptogenic lesion and surgery, and the completeness of surgical resection. A Meta-analysis of MST or MST combined with cortical resection for epilepsy including 211 patients from 6 centers showed a >95% reduction in seizures in 62%-71% and 68%-87% of patients, respectively; however, in terms of long-term outcome, the lowest seizure-free rate was observed after MST (16%), and callosotomy was unsatisfactory (34 (34%).
Several studies have shown that both AED dose reduction and medication discontinuation are associated with seizure recurrence. Regardless of the etiology, clinicians and patients should be aware that although seizure-free status can be maintained for many years after surgery, there is a tendency for the seizure-free rate to decrease over time. Schiller et al. retrospectively evaluated the frequency and risk factors for seizure recurrence after drug discontinuation in patients with refractory epilepsy treated surgically from 1989 to 1993. 84 of 210 patients had complete drug discontinuation after surgery, 96 had a reduction in drug dose and/or type, and 30 had no change in medication. The Kaplan-Meier Survival Assessment patients had a 14% and 36% reoccurrence rate at the end of year 2 and year 5, respectively, after complete discontinuation of medication; 47 of those who reduced their medication reached monotherapy, 13 had reoccurrence, and none had final discontinuation. The reoccurrence rate was 9 and 14% at the end of the second and fifth year after the last drug reduction, respectively; the reoccurrence rate was 3% and 7% during the same observation period for those with no change in treatment medication; the latter two groups were not statistically significant. Those without abnormalities on preoperative MRI were more likely to relapse after drug discontinuation than those with focal pathological changes, but no statistical significance was obtained, and the length of the seizure-free period of postoperative drug treatment did not affect the relapse rate. This study showed that intraoperative electrocorticography, extent of surgical resection, postoperative EEG and seizure-free duration were not predictive factors for relapse after drug discontinuation.
In the 143 patients observed by Van Veelen et al, the relapse rate was 30% after reducing the dose or stopping the AED after 2 years, but there was still a 17% relapse rate after continuing the drug, for a total relapse rate of 47%. Recent studies have also demonstrated the fact that patients who are seizure-free soon after surgery (within the year after surgery) can have a relapse rate of 44% 10 years after surgery. Moreover, half of the relapsed patients reduced their dose or stopped taking their medication on the advice of their physicians, suggesting that in clinical practice, physicians may have overestimated the efficacy of surgery. Therefore, randomized controlled clinical trials are needed to further determine the efficacy of surgery.
Schmidt et al. defined post-surgical cure as being seizure-free for at least 5 years after surgery and specifically looked at cure rates after surgery for medication-refractory temporal lobe epilepsy. This study included 13 retrospective and 5 prospective nonrandomized clinical trials published since 1980, enrolling a total of 1658 patients. The results showed that 1 in 4 adults and 1 in 3 children or adolescents were seizure-free after stopping medication 5 years after surgery, a cure rate of 25%; 66% of those who continued to take medication after surgery were seizure-free; 2 years after stopping medication after surgery were seizure-free, and seizure-free status could be stable for 5 or even 15 years. However, 55% of patients without disabling seizures did not completely stop taking medication 5 years after surgery in this study, so it was not possible to determine their actual cure rate and did not provide a predictive feature of surgical cure. Five to 10 years after anterior temporal lobectomy or 10 years after amygdala-hippocampal resection, 34%-35% of patients were completely seizure-free.
In another study, the same authors retrospectively analyzed the results of six clinical trials aimed at understanding how many seizure-free patients after surgery might relapse after planned discontinuation of medication. The results showed that the mean relapse rate was 33.8% in adults and 20% in children at 1 to 5 years of follow-up. Recurrence increased at 1-3 years of follow-up and occurred mostly within 3 years of stopping the AED; 1 in 3 patients who converted to seizure-free after epilepsy surgery had a reoccurrence after stopping the medication. Seizure recurrence was not influenced by the duration of AED treatment after surgery, and delayed discontinuation after 1-2 years of complete seizure freedom did not appear to provide additional benefit, and occasional seizures or aura after surgery did not preclude successful discontinuation at the end. 90% or more of patients with seizure recurrence resumed previous AED treatment to achieve seizure control.
In a retrospective analysis of 88 patients with intractable temporal lobe epilepsy who discontinued their AEDs after successful temporal lobectomy, the prognosis for seizures after complete discontinuation of AEDs was better than for seizures after AED reduction, with more successful discontinuation occurring in patients who were younger and had a shorter duration of disease at the time of surgery. The above findings provide more valuable information for treatment decisions after epilepsy surgery.
II. Mechanisms of post-surgical re-seizures
Dreeks et al. concluded that epilepsy surgery is the conventional treatment for intractable focal canker sores, and that seizure recurrence after resection depends on the type and location of the underlying pathological changes. Although the prognosis for those with independent temporal lobe epilepsy is good (60-90% seizure-free), epilepsy surgery in other brain tissue regions rarely achieves such a high success rate. Those who fail surgery usually have a seizure recurrence 6-12 months after surgery. These post-surgical seizures can be classified into 3 types i.e. habitual, non-habitual and adjacent seizures, with different pathological mechanisms. Understanding these mechanisms will undoubtedly be an important guide for the selection of post-surgical treatment options.
1. Habitual seizures
The seizure symptoms after surgery are the same as those before surgery, and the same clinical seizures are produced because the tissue around the adjacent previously epileptogenic lesion is not completely removed and becomes a new epileptogenic area or mature lesion. The EEG at the time of seizure may still show the same epileptic origin as before surgery, but the waveform may be altered, possibly because of the postoperative skull defect and partial removal of tissue. Habitual seizures can occur immediately after surgery, or months or years after surgery, usually within the first year of surgery; however, they rarely occur after a long period of seizure absence (which may last more than 5-10 years). Habitual seizures within the first year after surgery are more likely to become persistent or intractable, while habitual seizures occurring after the first year or even more than 10 years are easily controlled with medication and do not become drug-resistant epilepsy.
In 1983, Rasmussen proposed the phenomenon of “running down” to explain the gradual disappearance of habitual seizures, as the surgical removal of a major part of the epileptogenic region may facilitate the natural evolution of epilepsy. Later, Salanova et al. found that the epileptogenic region was smaller in those without seizures compared to those with “running down” after surgery. These studies suggest that there may be residual epileptogenic regions with a higher threshold of epileptic discharges that are ultimately insufficient to cause persistent clinical seizures, presumably due to maturation of inhibitory loops or disruption of excitatory connections.
2. Non-habitual seizures
In some patients, the seizure symptoms after surgery are different from the preoperative habitual seizures and change to a new clinical type, called nonhabitual seizures. The mechanisms that lead to non-habitual seizures are: (1) The persistent epileptogenic focus and the functional brain area that produces symptoms are removed, changing the symptoms of epileptic seizures. For example, right foot clonic seizures continue to have clonic seizures in the right upper arm after removal of the primary motor cortex (M1) foot area. (2) The epileptogenic focus or epileptic seizure spreading pathway is selectively partially removed, which changes the clinical presentation of the patient. Epileptiform discharges can propagate through different loops, and this mechanism may also explain why post-surgical patients experience aura after stopping medication without the tendency to develop epileptiform seizures with motor symptoms or loss of consciousness. The typical abdominal aura is particularly common after surgery in patients with temporal lobe epilepsy, and it is hypothesized that some of the epileptogenic foci that are not removed still constitute epileptic discharges that cause abdominal aura when they spread to the insula, but such epileptic discharges do not cause autonomic symptoms and loss of consciousness. At the same time, secondary generalized seizures are relatively increased in temporal lobectomized individuals, probably because standard temporal lobectomy does not usually remove selective contact brain areas such as insula and septum in the amygdala-hippocampus and pontocerebrum, and thus epileptic discharges preferentially pass through these structures after temporal lobectomy leading to a relatively higher proportion of secondary generalization. (3) Exposure or maturation of other epileptogenic foci after surgery may result in a new seizure type. Defining the extent of the epileptogenic foci is a tricky issue because, for one, it is more extensive than the actual site of initiation of epileptic discharges; and, for another, it takes time for the maturation of the epileptogenic foci to cause seizures, e.g., MCD, although congenital, can be resting for more than 10 years before epileptogenesis. Therefore, it can be assumed that the formation of new epileptogenic foci after surgery may reflect the maturation process of the epileptogenic focus. The surgical scar may also become a new epileptogenic focus and produce new clinical seizure symptoms.
3. Adjacent seizures
Used to indicate focal simple motor seizures immediately after temporal lobe surgery (1-2 weeks), often without loss of consciousness. 20-58% of patients experience such seizures, either due to surgical irritation, edema, bleeding, or infection, or due to decreased AED blood levels, the latter of which has been found not to be a common factor in seizures in recent studies. Adjacent seizures must be distinguished from acute postoperative seizures, which are all types of seizures in the first week after surgery, with an incidence of 20%-49% and are associated with a poor prognosis, but do not preclude later conversion to seizure-free, which can still occur in 33%-51% of patients. In general, adjacent seizures are less harmful. Malla et al. reported that 75% of seizures eventually disappeared in patients with focal motor seizure aura and generalized tonic-clonic seizures after surgery. However, the same acute postoperative seizures as habitual seizures (46%-85%) have a poor prognosis. Most studies concluded that acute postoperative seizures occurring in the first 24 h of week 1 or late in week 1 did not differ significantly in final outcome; only one clinical observation reported that seizures with only 1 seizure or limited to day 1 after surgery were more likely to become seizure-free than multiple seizures or seizures late in week 1.
In conclusion, most patients are seizure-free after surgery or develop habitual seizures originating around the resected cortex, the latter occurring even 10 years after surgery, and occasionally a gradual decrease in the number of seizures is observed, showing a “tapering” phenomenon. However, most acute postoperative seizures, especially early habitual seizures, are associated with a poor prognosis. Isolated aura symptoms or secondary generalized seizures are more common after surgery, and nonhabitual seizures may also be observed. Therefore, it is clearly important to collect comprehensive information about the relevance of surgical outcome before surgery, to carefully evaluate and study the risk factors and mechanisms of reoccurrence after surgery in each patient, and to continue and choose to give appropriate AED treatment, both in order to ensure surgical efficacy and to help patients overcome the disease and regain their self-confidence.