Seizure identification by clinical description in temporal lobe epilepsy: how accurate are we?

OBJECTIVE: To determine the accuracy of the clinical history performed by epileptologists in the identification of seizures in patients with suspected temporal lobe epilepsy. METHODS: The clinical and EEG telemetry (EEGT) monitoring data of 88 patients with suspected refractory temporal lobe seizures referred for evaluation of epilepsy surgery were prospectively evaluated. All clinical events obtained by history in these patients were adjudicated as being a seizure or not by two blinded (without access to EEG data) independent epileptologists. Each clinical event was then matched with the corresponding clinical event recorded with EEG telemetry in the epilepsy monitoring unit (gold standard). Sensitivity, specificity, overall accuracy, predictive value, and interrater agreement for the clinical assessment were obtained. RESULTS: Of 357 clinically different events, 175 (49%) were reproduced in the epilepsy monitoring unit. Only 10 events were misidentified by history as being a seizure or not, resulting in an overall clinical accuracy of 94%. Epileptologists' sensitivity for seizure identification was 96% (95% CI 92, 98%) but specificity was only 50% (95% CI 22, 79%). Accuracy for complex partial seizures and generalized seizures was higher than for simple partial seizures (SPS). Misidentification occurred only with SPS and nonepileptic events. Agreement beyond chance among epileptologists was good. CONCLUSION: In this selected group of patients with temporal lobe epilepsy, seizure identification by clinical history is highly accurate. Epileptologists rarely miss seizures (high sensitivity) but more often overcall nonepileptic events as seizures (low specificity).

Remacemide hydrochloride as an add-on therapy in epilepsy: a randomized, placebo-controlled trial of three dose levels (300, 600 and 800 mg/day) in a B.I.D. regimen.

Remacemide hydrochloride is a low-affinity, non-competitive NMDA receptor channel blocker under investigation for the treatment of epilepsy. This double-blind, placebo-controlled, multicentre study assessed the safety and efficacy of adjunctive remacemide hydrochloride or placebo, in adult patients with refractory epilepsy who were already taking up to three antiepileptic drugs (including an enzyme-inducer). Patients (n= 262) were randomized to one of three doses of remacemide hydrochloride (300, 600 or 800 mg/day) or placebo, in a B.I.D. regimen, for up to 14 weeks. Plasma concentrations of carbamazepine (CBZ) and phenytoin (PHT) were controlled throughout. Patients recorded their seizures on a diary card. There was an increase in the percentage of responders (defined as a reduction in seizure frequency from baseline > or = 50 %), from 15 % (9/60) with placebo, to 30 % (18/60) in the 800 mg/day group. A pairwise comparison between remacemide hydrochloride 800 mg/day and placebo was statistically significant (P = 0.049). Most reported adverse events (mainly CNS and gastrointestinal) were mild or moderate in severity and dose-dependent. Adjunctive remacemide hydrochloride treatment was associated with a higher, dose-related responder rate compared with placebo. The difference reached significance at the highest dose tested (800 mg/day). Remacemide hydrochloride was well tolerated.

A randomized, controlled trial of surgery for temporal-lobe epilepsy.

BACKGROUND: Randomized trials of surgery for epilepsy have not been conducted, because of the difficulties involved in designing and implementing feasible studies. The lack of data supporting the therapeutic usefulness of surgery precludes making strong recommendations for patients with epilepsy. We conducted a randomized, controlled trial to assess the efficacy and safety of surgery for temporal-lobe epilepsy. METHODS: Eighty patients with temporal-lobe epilepsy were randomly assigned to surgery (40 patients) or treatment with antiepileptic drugs for one year (40 patients). Optimal medical therapy and primary outcomes were assessed by epileptologists who were unaware of the patients' treatment assignments. The primary outcome was freedom from seizures that impair awareness of self and surroundings. Secondary outcomes were the frequency and severity of seizures, the quality of life, disability, and death. RESULTS: At one year, the cumulative proportion of patients who were free of seizures impairing awareness was 58 percent in the surgical group and 8 percent in the medical group (P<0.001). The patients in the surgical group had fewer seizures impairing awareness and a significantly better quality of life (P<0.001 for both comparisons) than the patients in the medical group. Four patients (10 percent) had adverse effects of surgery. One patient in the medical group died. CONCLUSIONS: In temporal-lobe epilepsy, surgery is superior to prolonged medical therapy. Randomized trials of surgery for epilepsy are feasible and appear to yield precise estimates of treatment effects.

Frontal lobe seizure propagation: scalp and subdural EEG studies.

PURPOSE: To study propagation properties of focally originating frontal lobe seizures: principally direction and promptness of ictal spread. METHOD: Forty-eight scalp EEG-recorded and 17 subdurally recorded seizures in two separate groups of patients were visually scrutinized. RESULTS: Initial propagation was directed more commonly to contiguous frontal cortex than to the opposite hemisphere in both studies: 58% and 86% for scalp and subdural recordings, respectively. Eighteen (38%) of scalp EEG seizures propagated within 5-10 s of apparent onset, whereas 16 (33%) did so after 11-20 s; no evidence of propagation could be discerned in the remaining 14 (29%). Of subdurally recorded attacks, only four (24%) propagated in the first 4 s; six (35%) did so after a delay of 5-10 s; six (35%) either failed to propagate or did so after > or =20-s latency. Latency to initial propagation was between 5 and 20 s in 41 (63%) of the 65 seizures in both studies. Ictal activity remained within the frontal lobe of origin for > or =10 s in 39 (81%) of scalp-recorded seizures and 11 (65%) of subdural seizures. It remained so confined for > or =15 s in 26 (54%) and nine (53%), respectively. Twenty-four (50%) of scalp-recorded seizures displayed evidence of opposite hemisphere (bisynchronous or other contralateral) involvement, occurring 5-10 s after onset in eight (17%) seizures and 10-20 s after onset in 16 (33%). Similarly, the subdural study documented spread to the opposite hemisphere in 11 (65%) of seizures with a latency ranging from 1 to 45 s. CONCLUSIONS: Frontal lobe seizures in this study propagated less promptly than reputed, and initial spread occurred more commonly to the frontal lobe of origin than to the opposite hemisphere. Such properties underlie the good lateralizing value of clinical semiology of frontal lobe seizures.

Temporal epileptogenesis: localizing value of scalp and subdural interictal and ictal EEG data.

PURPOSE: To determine the value of scalp epileptiform EEG data and subdural interictal spikes in localizing temporal epileptogenesis among patients requiring invasive recordings. For this delineation, we related such factors to site of subdural seizure origin in 27 consecutive patients. METHODS: Patients with temporal lobe epilepsy whose non-invasive lateralizing data were inconclusive and therefore required subdural electroencephalography were studied. All patients had (a) 24-h scalp telemetered EEGs, (b) adequate bitemporal subdural placements with an inferomesial line extending from a posterior burr hole anteriorly to <2.5 cm from anterior uncus and a lateral line reaching within 2.5 cm of the temporal tip, and (c) > or =2 subdurally recorded seizures. RESULTS: Three hundred one (96%) of 314 subdurally recorded clinical seizures involving all 27 patients arose from a discrete focus; 266 (85%) arose from mesial temporal regions, which was the origin of the majority of seizures in 24 (89%) patients. The majority of subdural seizures arose ipsilateral to the majority of scalp EEG spikes in 22 (81%) of 27, and most subdural seizures of 15 (75%) of 20 arose ipsilateral to scalp seizures. Lateralization of interictal subdural spikes correlated with that of subdural seizures in 74-92% of patients, depending on the method of spike compilation: for example, most subdural seizures arose from the same lobe of most consistent principal temporal spikes in 92% of patients. These indices of epileptogenesis also appeared more commonly on the side of effective (> or =90% improvement) temporal lobectomy than contralaterally in the following proportions: most consistent principal subdural spikes, 86% of patients ipsilateral vs. 9% contralateral; scalp-recorded clinical seizures, 55% vs. 18%; scalp EEG spikes, 45% vs. 9%. CONCLUSIONS: Even among patients whose scalp data are sufficiently complex to require invasive recording for clarification, lateralization of temporal scalp interictal and ictal epileptiform activity and subdural interictal spikes should be included when assessing the side of temporal epileptogenesis.

Current trends in electroencephalography.

Several recent articles re-emphasize the value of clinical electrophysiology: in localizing epileptogenesis, predicting effectiveness of epilepsy surgery, and disclosing a mechanism of benign Rolandic epilepsy of childhood.A review of the role of EEG in the diagnosis of epilepsy indicated that epileptiform activity will appear in 50% of initial awake recordings of adults with epilepsy and in 85% of subjects undergoing two recordings. This contrasts with the appearance of spikes in only 4 of 1000 normal persons. Several studies focused on the value of electroencephalography in extratemporal epilepsy: 62% of patients with neocortical epilepsy had at least one localizing ictal EEG; occipital and temporal neocortical seizures were localized in a greater proportion than frontal or parietal attacks. Interictal spikes, if unifocal, always arose from the epileptogenic region in a study of their seizure localizing value. Such congruence augured for better seizure control by focal resection in two studies reviewed herein. Studies indicating the value of interictal temporal lobe spikes and scalp-recorded seizures in lateralising a temporal seizure focus are reviewed. One study found EEG to be slightly more reliable for lateralization of temporal epileptogenesis than MRI. In patients with benign Rolandic seizures, enhanced motor evoked potentials (MEPs) were obtained from transcranial magnetic stimulation when this was applied 50-80 msec after electrical stimulation of the thumb whereas this interval inhibited the MEP in normal subjects. This suggests that afferent cutaneous input abnormally and synchronously activates a large population of sensory neurons; such activation is subsequently transmitted to the motor cortex to produce the focal spikes in this condition.Finally, advances in non-invasive technology have redefined and limited the need for invasive monitoring in children with intractable seizure disorders.

Pathogenesis of Lennox-Gastaut syndrome: considerations and hypotheses.

No other epilepsy more substantially combines cryptogenicity and intractability than does Lennox-Gastaut Syndrome. Although antecedent neurological conditions have preceded LGS in some patients, others with similar ills fail to develop this syndrome, and precise charting of the pathway between any presumed aetiology and the epilepsy has eluded researchers. Despite these frustrations, LGS has elicited a host of publications over the years, possibly because it combines features of many other epileptic disorders and thereby occupies a crossroads position among them. Lacking a comprehensive experimental model of LGS, this article combines relevant data from several clinical and basic sources in order to formulate a concept of pathogenesis. Data are presented to suggest that the occurrence of factors enhancing excitability during a vulnerable period of cortical and thalamic development may permanently imprint a bilateral, diffuse epileptogenic system upon the mammalian brain.

Pediatric candidates for temporal lobe epilepsy surgery.

Temporal lobectomy is an under-valued and effective therapy for selected children with temporal lobe seizures. This article details three components of the identification and evaluation of such children. The ictal symptoms and signs remains an essential component. This includes an aura suggesting limbic system involvement, and the nature of automatisms and other motor phenomena which may have lateralising value. Electroencephalography continues to accurately identify a principally epileptogenic focus in children and adults. The value of high quality magnetic resonance imaging is emphasized. Invasive studies are required in only a minority of cases. Several studies have revealed that 73-100% of carefully selected children receive benefit from this procedure whereas significant neurosurgical complications rarely occur.

Pattern-evoked potential latencies from central and peripheral visual fields.

Median P100 and N70 latencies for peripheral field (>8 degrees) TV-generated pattern visual-evoked potentials were 6 and 8 milliseconds less than for central field patterns in subjects with normal pattern visual-evoked potentials. These differences held for patient subgroups with P100 latency maxima from 140 to 100 milliseconds, and for left and right eye data separately compiled. Latencies for central field stimulation exceeded those for peripheral field stimulation in 233 (85%) of 274 eyes for the N70 potential and in 210 (77%) of 274 eyes for the P100 potential. Such data suggest that the faster conducting peripheral visual system, conveying location and motion information, prepares the occipital cortex for the later arriving central data conveying more intricate details.

Auras, seizure focus, and psychiatric disorders.

OBJECTIVE: The goal of this study was to assess the types of auras and their associations with the focus of seizures in 144 consecutive adult patients with partial epilepsy refractory to medical treatment. The association of frequency and type of aura with a psychiatric disorder was also studied. BACKGROUND: The nature of the aura, its association with the seizure focus, and its relation with emotions and behavioral dysfunction have received relatively little attention, and remain controversial. METHODS: Patients were admitted to an epilepsy investigation unit for detailed observation of their seizures and type of aura. Patients underwent standard electroencephalographic telemetry, including implanted subdural electrodes, if necessary, to delineate the seizure focus. The auras were classified into nine categories. All patients had an independent clinical psychiatric interview for a Diagnostic and Statistical Manual, Third Edition-Revised (DSM-III-R)-based diagnosis. RESULTS: Of the 144 patients studied, 111 (77.1%) had one or more symptoms during the aura and 33 (22.9%) had no aura. The most frequent aura types were viscerosensory (47 cases [32.6%]), experiential (44 cases [30.6%]), and cephalic (21 cases [14.6%]). A temporal lobe focus was present in 116 (80.5%) patients. Only experiential auras were significantly associated with a temporal lobe seizure focus (33 of 44 cases). A DSM-III-R psychiatric diagnosis was present in 74 (51.4%) patients. Of the 116 patients with a temporal lobe focus, 60 (50.2%) had a DSM-III-R psychiatric diagnosis. Patients with two or more auras were significantly more likely to have a psychiatric diagnosis (70.2%) compared with those with one aura (39.1%) or no aura (48.5%). A significant association between experiential auras and psychiatric morbidity was seen. CONCLUSIONS: Auras occurred in a majority of patients with a treatment-refractory partial epilepsy who were hospitalized for assessment of their seizure focus for possible surgical candidacy. Two or more auras were associated with a psychiatric diagnosis as was the presence of an experiential aura. Experiential auras were most common in temporal lobe epilepsy. Approximately half the patients, irrespective of seizure focus, had an associated psychiatric morbidity.

Stereotactic amygdalohippocampotomy for the treatment of medial temporal lobe epilepsy.

PURPOSE: This study was carried out to assess the safety and efficacy of stereotactic ablation of the amygdala and hippocampus for the treatment of medial temporal lobe epilepsy. METHODS: Twenty-two stereotactic amygdalohippocampotomies were performed in 19 patients with unilateral temporal lobe seizures by using magnetic resonance imaging (MRI) localization for target planning and radiofrequency techniques for lesion production. Seizure frequency was assessed at 3-monthly follow-up visits. Two lesion groups were defined. In group I, four to 11 (mean, 6.4) discrete lesions were made, encompassing the amygdala and anterior 13-21 mm (mean, 16.8 mm) of the hippocampus. In group II, a large number of confluent lesions were made (mean, 26.0; range, 12-54) encompassing the amygdala and anterior 15-34 mm (mean, 21.5 mm) of the hippocampus. MRI scanning was carried out 24 h and 6-9 months after surgery. RESULTS: In five group I patients, one (20%) experienced a favorable seizure outcome. Of 15 group II patients, one of whom had previously undergone limited lesioning and was also analyzed as part of group I, nine (60%) experienced a favorable seizure outcome, with two seizure free. MRI scans at 6- to 9-months' follow-up disclosed discrete areas of atrophy in the amygdala and hippocampus, interspersed with preserved brain in the group I patients. More uniform and complete destruction of amygdala and hippocampus was evident in group II patients. All lesions were confined to the amygdala and hippocampus, sparing the parahippocampal gyrus (PHG). CONCLUSIONS: The extensive amygdalohippocampal ablation in group II patients improved seizure outcome compared with more limited ablation in group I, but these results were not so good as those from temporal lobectomy in a similar patient group. When considered together with the results of selective amygdalohippocampectomy, and temporal resections that spare hippocampus or amygdala (all producing similar outcomes, and all involving resection of the entorhinal cortex), this study suggests a pivotal role of the entorhinal cortex in temporal epileptogenesis.

The periodicity of lateralized epileptiform discharges.

OBJECTIVES: Although the term periodic lateralized epileptiform discharges (PLEDs) was introduced in 1964, there has been little discussion about the definition of periodicity, whether the degree of periodicity has clinical significance, and whether electroencephalographers can accurately assess periodicity. The objectives of this study were to determine if a mathematical descriptor of periodicity could analyse PLEDs and whether such has clinical relevance, to correlate the degree of periodicity with underlying etiology and to assess the accuracy of sight recognition of periodically occurring EEG events. METHODS: A 6 year retrospective analysis of EEGs on patients with PLEDs was undertaken. The coefficient of variation of inter-complex intervals was calculated for all records. The relative periodicity of each record was rated by electroencephalographers. Corresponding hospital records were reviewed with regard to clinical and pathological information. RESULTS: In cases of acute viral encephalitis, discharges were more periodic than in other etiologies (P = 0.013). Age, clinical state, and time of last seizure did not influence periodicity but 40 (70%) of 57 recordings with PLEDs were done within 2 days of a seizure. A history of seizures was present in 42 (89%) of 47 patients. Electroencephalographers' assessment of periodicity correlated well with measured periodicity (r = -0.49, P = 0.0001). However, EEGers were more likely to assess records as periodic if: discharges occurred frequently (r = -0.48, P = 0.0001), or the complexes were visually assessed to have a uniform morphology (r = 0.31, P = 0.02), even though neither of these factors influenced the periodicity of discharges. CONCLUSIONS: PLEDs were objectively measured to be more periodic when they were associated with acute viral encephalitis than with other etiologies. Nonetheless, the considerable overlap of values among the several presenting disorders diminished specificity. Electroencephalographers were able to recognize periodically occurring discharges. However, they were strongly influenced by discharge frequency and their perception of the uniformity of discharge morphology.

Referral patterns of family physicians may allow population-based incidence studies of childhood epilepsy.

PURPOSE: To evaluate the burden of illness of childhood epilepsy on patient, care giver, and society, representative incidence cohorts must be followed longitudinally. Case ascertainment through pediatricians and neurologists would be a valid method if family physicians refered all new cases of childhood epilepsy. The study objective was to determine whether family physicians' referral patterns in Southwestern Ontario make it possible to conduct a population-based incidence study of childhood epilepsy by sampling only from specialists' practices. METHODS: Of the 1,718 family physicians practicing in Southwestern Ontario, a systematic sample participated in a mailed survey. Case simulations describing seven types of childhood seizures were presented to physicians with instructions to respond as to whether they would investigate/manage without referral; refer to a specialist only if problems occurred; or refer to a specialist always. RESULTS: Of 214 family physicians, 185 (86.4%) returned completed surveys; 86% would not refer a child with a febrile seizure. Referral to a specialist would be made always by 93% of family physicians for patients with status epilepticus, 95% for worsening partial epilepsy, 82% for a first, brief, generalized clonic seizure, 80% for absence epilepsy, and 99% for neonatal seizures. Only 50% of family physicians would always refer a neurodevelopmentally abnormal child with generalized clonic epilepsy, but a further 37% would refer if problems occurred. CONCLUSIONS: It is feasible to recruit a representative population-based cohort of recently diagnosed patients for epidemiologic studies of childhood epilepsy by surveying pediatricians and neurologists. These survey results could be used to adjust estimates of incidence obtained through specialists' practices for the bias in case ascertainment that may result from this practical method.

Hemispheric epilepsy.

This report documents clinical and EEG features of 13 patients whose most frequent EEG abnormality was unilateral interictal spike-waves that occurred consistently over the same hemisphere on multiple recordings. These 13 patients were selected from a 25000-patient database encompassing 25 years. The hemisphere involved was the left in 8 (62%) of the 13 patients. Bisynchronous spike-waves also appeared on a majority of recordings in 10 (77%) patients. Focal spikes occurred in a majority of recordings in 7 (54%) patients; these usually predominated in the frontal or frontal temporal regions and were always ipsilateral to hemispheric spike-waves. Their field distribution often merged with regional or hemispheric spike-waves. EEG evidence of ictal origin ipsilateral to hemispheric spike-waves appeared in 7 of 8 patients whose seizures were recorded; ictal potentials never predominated contralaterally. All patients had seizure disorders beginning in childhood or adolescence. Eleven (85%) had more than one type of seizure. Principal ictal types contained no focal phenomena in 10 (77%) patients. Bilateral motor seizures were tonic-clonic (12 patients), myoclonic (3), tonic (3) and atonic (1). Absence-like attacks occurred in 10 patients: only staring, automatisms and loss of awareness occurred in 6, with aurae in 3, and with motor phenomena in 3. Features of some seizures in 9 patients (69%) suggested a focal or unilateral origin or involvement, but these consistently occurred contralaterally to principal spike-waves in only 4 patients. At a median follow-up of 19 years, 11 patients (85 %) continued to have seizures; the 2 seizure-free patients required two anti-epileptic drugs each. History, neurological examination and neuro-imaging failed to disclose an aetiology or structural lesion in any patient. Eleven patients (85%) had normal intelligence. This newly documented syndrome of hemispheric epilepsy therefore comprises four components: (i) diffuse unilateral spike-waves on EEG which consistently appear over the same hemisphere in a given patient; (ii) a persistent, usually intractable seizure disorder with generalized ictal features in all patients and focal in most; (iii) onset of seizures in childhood or adolescence; and (iv) no apparent aetiology or related structural abnormality of the CNS.

MR in temporal lobe epilepsy: analysis with pathologic confirmation.

PURPOSE: We evaluated the MR findings in patients with temporal lobe epilepsy to determine the predictive value of MR imaging in assessing patient outcome. METHODS: MR studies from 186 of 274 consecutive patients who underwent temporal lobectomy for intractable epilepsy were reviewed retrospectively. Images were interpreted by an experienced neuroradiologist, who was blinded to the side of seizure activity and to pathologic findings. RESULTS: MR imaging exhibited 93% sensitivity and 83% specificity in detecting hippocampal/amygdalar abnormalities (n = 121), and 97% sensitivity and 97% specificity in detecting abnormalities in the rest of the temporal lobe (n = 60). Abnormal high signal of the hippocampus on T2-weighted images had a sensitivity of 93% and specificity of 74% in predicting mesial temporal sclerosis (n = 115). The presence of hippocampal atrophy on MR correlated with the duration of seizures. Sensitivity and specificity of MR imaging in detecting temporal lobe tumors (n = 42) were 83% and 97%, respectively, based on abnormal signal and mass effect. After surgery, 63% of patients were seizure free and 28% had a significant reduction of seizure frequency at an average of 24 months (range, 12 to 78 months) after surgery. Patients with a single lesion in the anterior temporal lobe or hippocampus/amygdala had a better outcome than patients with multiple lesions (n = 22). Interrater agreement varied from 0.4 to 0.93, with best agreement for tumors or abnormal hippocampal signal on T2-weighted images. CONCLUSION: MR imaging is highly sensitive in detecting and locating abnormalities in the temporal lobe and the hippocampus/amygdala in patients with temporal lobe epilepsy. Hippocampal atrophy appears to correspond to the duration of seizure disorder.

Late lateralizing and localizing EEG features of scalp-recorded temporal lobe seizures.

The authors studied the lateralizing value for temporal epileptogenesis of focal or hemispheric EEG ictal features that first appear > or = 5 seconds after hemispheric or diffuse EEG changes or after an artifact-obscured onset, by correlation with side of lobectomy abolishing or producing > or = 90% improvement of complex partial seizures (CPS). One hundred forty-eight seizures in 39 patients rendered seizure-free (SF) by lobectomy lacked early localizing features of which 91 (61%) seizures in 37 (95%) patients had late lateralizing or localizing phenomena. Eighty-seven (96%) of 91 late localizing or lateralizing seizures occurred ipsilateral to lobectomy, involving 35 (95%) of 37 patients. Late contralateral and shifting phenomena occurred in a total of 4 seizures of 2 patients. Among 17 non-SF but > or = 90% improved patients, 30 (48%) of 63 seizures lacking early localizing features showed late localizing or lateralizing phenomena. Of these, twenty (67%) involving 12 (71%) patients occurred ipsilateral to lobectomy. Among the remaining 5 (29%) patients, a total of 5 seizures (17%) were contralateral while 5 (17%) exhibited shifting laterality. These data indicate that seizures with late (> or = 5 seconds) localizing or lateralizing features cannot be discounted when assessing laterality of temporal epileptogenesis but cannot be used as the sole electrographic criterion.

Gabapentin monotherapy: I. An 8-day, double-blind, dose-controlled, multicenter study in hospitalized patients with refractory complex partial or secondarily generalized seizures. The US Gabapentin Study Group 88/89.

We evaluated the efficacy and safety of gabapentin administered as monotherapy in an 8-day, randomized, double-blind, dose-controlled, parallel-group, multicenter study comparing dosages of 300 and 3,600 mg/d gabapentin in 82 hospitalized patients whose antiepileptic medications had been discontinued for seizure monitoring. Seizures under study were complex partial seizures with or without secondary generalization. Patients exited the study if they experienced a protocol-defined exit event indicating lack of efficacy. Time to exit was significantly longer (p = 0.0001) and completion rate was significantly higher (53% versus 17%; p = 0.002) for patients receiving 3,600 mg/d gabapentin. Gabapentin was well tolerated by patients in both dosage groups, and no patients exited the study due to adverse events, despite rapid initiation of full dose within 24 hours. These results demonstrate that gabapentin has anticonvulsant activity and is well tolerated when administered as monotherapy in patients with refractory partial seizures.