Copyright 2010 The Authors Journal compilation 2010 Blackwell Munksgaard
Acta Neurol Scand 2011: 123: 48–53 DOI: 10.1111/j.1600-0404.2010.01343.x
ACTA NEUROLOGICA SCANDINAVICA
Acute postoperative seizures after epilepsy surgery – a long-term outcome predictor? Alfstad KA˚, Lossius MI, Røste GK, Mowinckel P, Scheie D, Borota OC, Larsson PG, Nakken KO. Acute postoperative seizures after epilepsy surgery – a long-term outcome predictor? Acta Neurol Scand: 2011: 123: 48–53. 2010 The Authors Journal compilation 2010 Blackwell Munksgaard. Objectives – The prognostic value of acute postoperative seizures (APS) after epilepsy surgery is much debated. This study evaluated APS, defined as seizures in the first week post-surgery, as a predictor of long-term seizure outcome, and investigated the utility of other potential outcome predictors. Materials and methods – Medical records of 48 patients with temporal and extra-temporal epilepsy surgery were studied. Forty patients had lesional surgery. All had at least 2 year postoperative follow-up. Results – At 2 year follow-up, 25 patients (53%) were seizure free. Univariate analysis showed that APS (P = 0.048), using ‡six AEDs prior to surgery (P = 0.03), pathological postoperative EEG (P = 0.043) and female gender (P = 0.012) were associated with seizure recurrence. Conclusions – Univariate analysis indicate that APS, a high number of AEDs used prior to surgery, and pathological postoperative EEG are possible predictors of seizure recurrence after epilepsy surgery. Only gender retained significance in the multivariate analysis.
K. . Alfstad1, M. I. Lossius1, G. K. Røste2, P. Mowinckel3, D. Scheie4, O. C. Borota5, P. G. Larsson6, K. O. Nakken1 1
National Centre for Epilepsy, Rikshospitalet University Hospital, Oslo, Norway; 2Neurosurgical Department, Rikshospitalet University Hospital, Oslo, Norway; 3 Department of Pediatrics, Ullevl University Hospital, Oslo, Norway; 4Division of Pathology, Rikshospitalet University Hospital, Oslo, Norway; 5Department of Laboratory Medicine ⁄ Pathology, Ume University Hospital, Ume, Sweden; 6Department of Neurodiagnostics, National Centre for Epilepsy, Division of Clinical Neuroscience, Rikshospitalet University Hospital, Oslo, Norway Key words: epilepsy; postoperative seizures; seizure outcome; surgery Kristin . Alfstad, Department of Neurology ⁄ National Centre for Epilepsy, Rikshospitalet University Hospital, N-0027 Oslo, Norway Tel.: 47 23070000 Fax: 47 23070490 e-mail:
[email protected]. Accepted for publication January 20, 2010
Introduction
Reliable factors for accurately predicting long-term seizure outcome following epilepsy surgery are of major importance for ensuring the selection of appropriate candidates for this treatment option. It also enables the physician to present valuable information to the patient whether surgery is likely to be beneficial. Acute postoperative seizures (APS), here defined as seizures in the first week after surgery, are a distressing experience for any patient, and the question naturally arises if this indicates surgical failure. The prognostic value of these seizures has been examined in several studies. Most studies have evaluated the prognostic importance of APS in temporal lobe resections, and APS are mainly found to be a marker of later seizure recurrence (1–3). However, APS have also been considered to be associated with the surgical procedure per se, or a result of the Ôrunning down phenomenonÕ, i.e. early postoperative seizures with 48
later remission, and not necessarily a negative predictor of long-term outcome (4, 5). In extratemporal epilepsy surgery, data on the prognostic value of APS is rather conflicting (6–8). The aim of this study was to assess the occurrence and prognostic importance of APS, and also investigate other factors which might provide an indication of the long-term outcome following epilepsy surgery. Materials and methods
In Norway (4.8 million inhabitants), epilepsy surgery is centralized to one centre. Surgical candidates undergo preoperative evaluation at the National Centre for Epilepsy in Sandvika, and the surgery is performed at the Neurosurgical Department at Rikshospitalet University Hospital in Oslo. Resection specimens are sent to the Division of Pathology, Rikshospitalet University Hospital, for histological examination. A few days after
Outcome predictors after epilepsy surgery surgery, the patients are transferred to the National Centre for Epilepsy for postoperative care. The standardized preoperative evaluation includes a detailed clinical investigation, with emphasis on seizure frequency and semiology, interictal standard EEG, long term EEG video-monitoring with scalp electrodes, 1.5 Tesla MRI according to an epilepsy protocol, and neuropsychological testing, including Wada test. Long term EEG videomonitoring with subdural electrodes, ictal SPECT, PET, and functional MRI are obtained in selected cases. All candidates for surgery are discussed in detail at a weekly multidisciplinary epilepsy surgery conference. In addition to the information from the presurgical evaluation, the following data were obtained from the patientsÕ medical record: gender, age at onset of epilepsy, age at epilepsy surgery, preoperative seizure frequency (categorized as weekly or less than weekly), and seizure types (classified as simple partial and ⁄ or complex partial seizures with or without secondary generalized tonic–clonic seizures). The number of AEDs used prior to surgery was categorized as either being six or more, or less than six, as the patients had used on average 6.2 AEDs prior to surgery. Preoperative MRI was classified as either normal or with a morphological substrate in the lobe in question. Information on the surgical procedures and the occurrence of seizures in the first week after surgery (APS) were also obtained from the medical records. Post-surgical evaluation consisted of a thorough seizure history, a clinical examination, and an EEG recording during visits at our centre 3, 6, 12 and 24 months postoperatively. The 3-month postoperative EEG results were chosen as an end point for this study. According to standard protocol, this is a routine recording of 20 min duration. The EEGs were examined by an experienced neurophysiologist, and postoperative EEGs were regarded as pathological if epileptiform discharges, recognised as spikes and ⁄ or sharp waves or sharp rhythmic activity, were found. MRI was obtained 6 months postoperatively. We examined the medical records of 64 consecutive patients, 23 children (age < 14 years) and 41 adults, who had undergone epilepsy surgery in 2002 and 2003. Two children had undergone surgery in both 2002 and 2003, i.e. a total of 66 procedures were performed. The study was approved by the local ethics committee. In the same period, a total of 245 patients had been evaluated for surgery, i.e. 26% of those
evaluated were found eligible. Surgery had been performed at the Neurosurgical Department at Rikshospitalet on all patients except two who had undergone functional hemispherotomy at the Neurosurgical Department at Bonn University Hospital in Germany. Ten patients had previously had epilepsy surgery, of which eight were children. A total of 18 patients were excluded from the data set, four because of insufficient follow-up data, four because they had been re-operated on earlier than 2 years after the first operation, one patient who died 4 months after surgery for reasons unrelated to the operation, two patients with a very short time span between surgeries (surgery performed both in 2002 and 2003) and five patients because they had undergone palliative surgery (two patients with callosotomy due to Lennox-Gastaut syndrome and three patients with multiple subpial transections due to foci in eloquent cortex). The two patients who had hemispherotomy were also excluded, although considered a curative procedure still different from resective surgery. Of the 48 patients included in the study, 36 had temporal resections, 19 on the right and 17 on the left side. Eight patients had frontal resections, six on the right and two on the left side. Four patients had either parietal and ⁄ or occipital resections. In some selected cases the extent of the resection was guided by intraoperative electrocorticography. Histopathological examination of the resected specimens was re-evaluated by two neuropathologists, according to new diagnostic guidelines after the study start in 2005. Two years after surgery, the treatment results with respect to seizures were assessed according to Engel (4). For statistical purposes, outcome was defined as either seizure free, including nondisabling aura (Engel 1), or not seizure free (Engel 2–4). AED treatment remained unchanged for a minimum of 2 years postoperatively, unless the patient reported serious side effects or had disabling seizures (n = 6). Statistical analysis
StudentÕs t-tests and chi-square tests were used when testing for group differences for continuous and categorical variables, respectively. Odds ratios (OR) for the occurrence of APS were estimated with the use of univariate and multivariate logistic regression analysis, with seizure outcome as the dependent variable. Because of the lack of events, in the final analysis, the variables were restricted to the four most clinically and statistically important as indicated by the testing for group differences 49
Alfstad et al. Table 1 Treatment results according to Engel (n = 48) Engel classificationa 2 years postoperatively Class Class Class Class
1 2 3 4
Table 2 Possible predictors of seizure outcome (n = 48) No. of patients (%) 25 2 5 16
(53) (4) (11) (32)
a
Engel Classification (4). Class 1 = Seizure free including non-disabling aura for the last 2 years, or patients with a single seizure associated with discontinuation of AED. Class 2 = Rare non-disabling seizures (
