Nonconvulsive status epilepticus of frontal origin

Methods.

Inclusion criteria for NCSE of frontal origin were 1) impairment of cognitive functions with or without confusion for at least 60 minutes; 2) ictal positive motor manifestations restricted to slight turning of the head and/or eyes to one side, low-amplitude focal myoclonic jerks, and automatisms; 3) documentation of SE with continuous video EEG monitoring; 4) an ictal EEG pattern recorded consistently at onset from either the right, left, or both frontopolar or frontal electrodes (Fp1, Fp2, F3, F4, and Fz); and 5) if ictal discharges involved one or both of the anterior temporal/inferior frontal electrodes (F7 and F8) in addition to the frontal electrodes, a frontal origin was considered if ictal SPECT showed clear frontal uptake of the tracer after comparison with postictal SPECT.

Informed consent was first obtained from a first-degree relative of each patient. The same physician (P.T.) investigated all patients clinically. Most of the subjects had no previous seizures and we use the term “interictal” to denote the brief interval between the recurrent seizures of a single episode of NCSE. After termination of SE, video recordings were reviewed and neuropsychological findings during interictal phases were assessed retrospectively according to a qualitative scale developed for NCSE by Kudo et al.8 These were then compared with a postictal neuropsychological evaluation including the Wechsler Adult Intelligence Scale–Revised (WAIS-R). At that time additional informed consent was obtained from the patient.

EEGs were obtained with an Alvar 20-channel polygraph (Alvar, Montreuil, France) or with a Brain-Quick Micromed 24-channel polygraph (Micromed-France, Bordeaux, France). Scalp electrodes were placed according to the International 10-20 system.

Ictal SPECT was performed with continuous video EEG monitoring, and a neurologist supervised all patients continuously. Ten seconds after the beginning of a seizure, 740 MBq of freshly reconstituted ^99mTc hexamethyl propylene amine oxime (HMPAO) (Ceretec; Nycomed-Amersham International, Buckinghamshire, UK) was injected through a peripheral IV catheter in approximately 5 seconds. Using the same IV line, SE was treated 10 minutes later with a single benzodiazepine (BZ) injection, with either diazepam 2 mg/min until SE stopped or to a total dose of 25 mg, or clonazepam 0.2 mg/min until SE stopped or to a total dose of 3 mg. If IV BZ was not immediately effective, patients received a single 20 mg/kg phenytoin (PHT) infusion through an electric syringe pump at 50 mg/min.

SPECT acquisitions were carried out 45 to 60 minutes after the termination of SE. A conventional EEG was performed after SPECT acquisition to detect possible recurrence of SE.

Four patients (Patients 2, 5, 8, and 9) were studied with a rotating single-head gamma camera (General Electric 400 T; General Electric, Milwaukee, WI). A total of 128 projections of 25 seconds each were gathered over 360 degrees on 64 × 64 pixel matrices. Using the same imaging protocol, postictal SPECT was performed at least 5 days after the termination of SE. Patient 7 was investigated with a high-resolution rotating three-head gamma camera (Prism 3000 XP; Picker, Cleveland, OH) using 120 projections of 50 seconds on 128 × 128 pixel matrices. SPECT data were analyzed visually by two independent observers (O.M. and J.D.) using 6-mm–thick transverse slices. Each frontal lobe was divided into four areas of interest: frontopolar, dorsolateral frontal, mesial frontal, and orbital frontal. Studies of in Patient 9 also included three-dimensional (3D) analysis. We used a proportional stereotaxic atlas to compare the interictal and ictal studies, to localize brain structures, and to lateralize the epileptic focus. After normalization for count difference between ictal and postictal SPECT, a parametric 3D image was created, each voxel representing the percent of perfusion increase during the ictal phase. This 3D image was displayed after size normalization on top of a 3D digitized Talairach atlas using a color scale ranging from 0% in blue to 50% in white.

Hemogram, electrolytes, blood glucose, ammonia, proteins, liver function studies, brain CT, and MRI were performed in all patients. This protocol was approved by the local ethics committee. Follow-up after the initial episode of SE was available in 7 of the 10 patients.

Results.

Between January 1991 and December 1995 we identified 44 patients with adult-onset NCSE in the department of neurology of the Nice University Hospitals. Ten patients (23%) had NCSE of frontal origin, with two distinct electroclinical patterns (type 1 and type 2 NCSE of frontal origin, as described later). Clinical, EEG, MRI, and SPECT data are summarized in the table. Among the other patients, 18 (41%) had AS, 10 (13.5%) had CPSE of temporal origin, 2 (4.5%) had CPSE of occipital origin, and 4 (9%) had unclassifiable NCSE. ⇓

Type 1 NCSE of frontal origin.

In the first group of seven patients (type 1 NCSE of frontal origin; Patients 1 through 7) there was no clear impairment of consciousness. During interictal phases these patients were oriented and could carry out routine activities such as walking, eating, or washing themselves. They were able to give their address, telephone number, and name and age of their relatives. Simple commands, such as pointing to a part of the body or naming a familiar object, were carried out correctly. More complex tasks such as putting a sheet of paper into an envelope, subtracting serial sevens, and reproducing several simple patterns of alternative sequences were disturbed by impaired attention, slow responses, and perseveration. The clinical presentation in this type was characterized by continuous mood and behavior disturbance, with either a mild hypomanic state with affective disinhibition, enhanced word fluency, and familiarity (Patients 1, 2, 4, and 7) or, on the contrary, a state of affective indifference with blank facial expression, reduced word fluency, and lack of spontaneous activity and emotivity (Patients 3, 5, and 6). After SE, most patients could recall events that occurred during the episode.

Cyclic fluctuation of the symptoms occurred in four patients (Patients 2, 3, 5, and 6). The onset of seizures in these patients was marked by additional clinical signs: episodes of forced thinking (Patient 2) and slight head or eye deviation contralateral (Patients 2, 5, and 6) or ipsilateral (Patient 3) to the ictal discharge. Patient 3 was conscious of having seizures and had low-amplitude clonic jerks of the left corner of the mouth. Marked perseveration would occur during the planning of sequential tasks. For example, a short period of staring and unresponsiveness could be followed by repetitive drawing of the same simplified pattern when the patient was asked to write his address or to draw or copy rhythmic alternative sequences. Simple gestural automatisms such as picking at clothes, rubbing, or scratching movements were often noted, but oroalimentary or complex bipedal or bimanual automatisms were never observed. Patient 6 had prominent negative motor phenomena with a progressive loss of postural tone leading to total inability to perform any voluntary act. However, during interictal periods, he remembered perfectly what happened during the paralytic phase.

EEG background activity was normal in five patients (Patients 1, 2, 3, 4, and 6) with a 9 to 12-Hz reactive occipital alpha rhythm. Ictal patterns in these patients consisted of long recurrent discharges, beginning with characteristic unilateral, low-amplitude, frontopolar or frontocentral fast activity, followed by rhythmic spikes, spike-and-wave complexes, sharp waves, or slow waves. All patients showed a strictly unilateral spread of ictal activity that involved either the right (Patients 2, 3, and 6) or the left (Patients 1 and 4) frontal area (figure 1). The ictal pattern consisted of long bursts of rhythmic 1.5-Hz polyspike and slow wave activity in Patient 5, and rhythmic 1-Hz spike and wave activity in Patient 7. In both cases, phase reversal was noted at the onset of seizures over the right frontal electrode. By the end of the seizure, there was an ipsilateral spread to the anterior part of the hemisphere in four patients (Patients 1, 2, 5, and 7), but no consistent contralateral spread was ever observed. The mean length of seizures in this group was 159 seconds (range, 45 to 230 seconds), and the mean interval between attacks was 8.4 minutes (range, 2 to 15 minutes).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1. Ictal EEG in type 1 nonconvulsive status epilepticus of frontal origin (Patient 2). Low-voltage fast activity appears over the right frontal region and is followed by rhythmic sharp waves that spread over the ipsilateral hemisphere only. A 75-second interval has been deleted between the two samples. Total length of this seizure was 140 seconds.

Type 2 NCSE of frontal origin.

In three other patients (Patients 8, 9, and 10), NCSE was characterized by a confusional state with temporospatial disorientation, gross behavioral disturbances, and perseveration. It was difficult to assess the interictal level of consciousness in these patients because of their prominent distractibility, perseveration, and disturbed behavior. Cyclic fluctuation of the symptoms was present in all patients. Patient 9 was so aggressive that he required restraints. Slight eye and head deviation to the left at the beginning of seizures, followed by bilateral perioral myoclonia, was noted in Patient 8, whose seizures always ended with a broad smile. From the onset of the ictal EEG activity, these three patients had progressive reduction in contact and reactivity. By the end of the ictal period they were in a catatonic stupor, at times with simple gestural automatisms. Total unresponsiveness at the end of seizures occurred in Patients 8 and 10. All these patients were amnestic for the SE afterward.

Background EEG was abnormal. The posterior alpha rhythm was replaced by a diffuse unreactive theta or delta activity. Ictal patterns consisted of recurrent bifrontotemporal (Patients 8 and 9) or bifrontocentral (Patient 10) seizures (figure 2) with a right (Patient 9), left (Patient 10), or no (Patient 8) side predominance. Bilateral low-amplitude fast activity was gradually replaced in Patient 10 by slow, generalized low-amplitude polyspike and wave activity that persisted for as long as 10 minutes. The mean length of seizures in this group was 255 seconds (range, 35 to 540 seconds), and the mean interval between attacks was 11.3 minutes (range, 5 to 15 minutes).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2. Ictal EEG in type 2 nonconvulsive status epilepticus of frontal origin (Patient 8). The beginning of a seizure is marked by a slight deviation of the eyes and head to the left (arrow), followed by a rapid build-up of 3.5-Hz rhythmic sharp waves over both frontotemporal regions with slight right predominance. Total length of this seizure was 75 seconds. Ictal SPECT in this case showed right frontal hyperperfusion.

SPECT study.

Ictal and postictal ^99mTc HMPAO SPECT scans were performed in five patients (Patients 2, 5, 7, 8, and 9). No significant increase of ictal ^99mTc HMPAO uptake was observed in temporal regions. Compared with the postictal images, clear ^99mTc HMPAO uptake was localized to the right frontopolar area in Patients 2 and 7, and Patients 5 and 8 also showed unilateral hyperperfusion of the right mesial frontal area (Patient 5) and of the right dorsolateral frontal area (Patient 8).

Bilateral hyperperfusion of the whole anterior frontal area with clear anteromesial frontal maximum was found in Patient 9 (figure 3A). 3D analysis in this patient showed a 50% increase in perfusion that involved only the right anteromesial and orbitofrontal structures (figure 3B). Ictal SPECT in Patient 10 showed bilateral frontopolar and dorsolateral frontal hyperperfusion with a left-side predominance. However, this patient refused follow-up SPECT after resolution of SE. In all patients with a frontal lesion (Patients 2, 5, and 9), ictal HMPAO uptake enhanced the boundaries of the corresponding hypoperfused area later found on postictal SPECT.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3. (A) Ictal (top) and postictal (bottom) ^99mTc hexamethyl propylene amine oxime SPECT in Patient 9, performed with a single-head gamma camera. The ictal SPECT shows hyperperfusion of both frontal areas with a clear anteromesial frontal maximum. Postictal SPECT shows right frontal hypoperfusion corresponding on CT to an ethmoidal lymphoma invading the right basal frontal area. (B) Three-dimensional analysis of ictal and postictal SPECT in Patient 9 shows a 50% percent increase in perfusion, with prominent involvement of the right anteromesial and orbitofrontal structures. See text for details of the technique.

Discussion.

As defined by Niedermeyer et al.9 and by Rohr–Le Floch et al.,5 NCSE of frontal origin consists of prolonged periods of cognitive disturbance, at times with visible focal ictal signs, and a consistent epileptiform EEG pattern localized over one or both frontal regions at the onset of seizures. Since 1971 many terms have been used to describe this condition or closely related syndromes: borderline cases of petit mal status,10 prolonged psychic epileptic seizures,11 absence status with focal characteristics,9 acute prolonged confusion as a frontal-onset ictal state,12 nonconvulsive confusional frontal status,13 CPSE of frontal origin,4,14 acute confusional states with frontal origin,15 transitional petit mal status,16 and frontal status.5 We have used “NCSE of frontal origin,” because it is a topographic description and is clinically nonspecific, especially regarding impairment of consciousness.

Transient focal abnormalities related to the edema and hyperperfusion associated with the seizures are seen occasionally in imaging studies of partial NCSE. Hyperintensity in T2-weighted MRI images,17 and hyperperfusion on ictal SPECT7,18,19 have been reported. In our study, MRI showed no abnormalities that could be due to the seizures. In our 10 patients, NCSE was thus considered of probable frontal origin because of EEG or SPECT data, or both.

We observed two electroclinical varieties of NCSE of frontal lobe origin. Type 1 NCSE of frontal origin occurred in seven patients. Cyclic fluctuations of symptoms occurred in five of these patients. During interictal phases, consciousness was not clearly impaired and the term “confusional state” does not describe these patients accurately. Mood disturbance was the most prominent clinical feature, with either affective disinhibition resembling a fearful, hypomanic state, or affective indifference with lack of spontaneous activity and emotion. Interictal neurologic examination showed only subtle cognitive disturbances, such as impaired attention, perseveration, and difficulty in sequential planning of complex motor tasks. The EEG showed unilateral, well-localized frontopolar or frontocentral ictal discharges with little or no spread occurring in most patients on a normal background. Type 1 NCSE was thus best described as SPNCSE with affective or cognitive symptomatology rather than as CPSE.

Type 2 NCSE of frontal origin occurred in three other patients. They had a cyclic, severe confusional state with major behavioral disturbances. The EEG showed recurrent asymmetric discharges over both frontal, frontocentral, or frontotemporal areas with slow, unreactive background activity. The ictal pattern in Patient 10 developed from an initial frontopolar focus into a generalized pattern of 1.5-Hz polyspike and wave activity with left-side predominance reminiscent of AS. Despite bilateral EEG expression, a frontal origin of seizures was established by SPECT studies in two patients, and probably in one more, because increased hyperperfusion during a focal seizure is usually confined to the epileptogenic focus. In contrast to the first type, the electroclinical features of type 2 NCSE of frontal origin were more consistent with the classic picture of CPSE, with prominent confusion and markedly impaired responsiveness throughout, and postictal amnesia. The features of the postictal state following resolution of the NCSE were not useful in differentiating the two types. After cessation of type 1 NCSE, the return to clinical baseline was rapid and complete. Although recovery took somewhat longer after type 2 NCSE, these patients had also received much more medication.

We identified 55 cases of adequately documented NCSE of frontal origin published in 21 different papers from 1971 to 1997.4,5,7-15,18-27 Most of these satisfied our inclusion criteria. Ictal SPECT was reported in only three patients7,18,19 and also showed unilateral frontal hyperperfusion. A depth electrode study4 of eight patients with CPSE during presurgical evaluation for refractory partial epilepsy confirmed frontal lobe onset in four patients and suggested it in a fifth patient. The average patient age was 38 years (range, 13 to 84 years), and 59% of the patients were women. A total of 36% had no previous history of epilepsy, and imaging showed a focal frontal lesion in 35%. The disturbance of consciousness ranged from a slight subjective impairment of thinking and expression to a severe confusional state.

Subtle cognitive and behavioral disturbances without marked alteration of consciousness are frequent in NCSE of frontal origin. In the largest series,5 impairment of consciousness was found in only 10 of 18 patients (56%). As in type 1 NCSE of frontal origin, some patients were sly, ironic, or even elated whereas others were indifferent, irritable, perplexed, or even mute. Others have reported similar cases, with loss of self-initiated activity without overt confusion, with either affective disinhibition or indifference.4,8,9,11,12,18,20-22,27 Frontal ictal EEG abnormalities could be predominantly unilateral9,14,20 or more diffuse. Clinical features in type 1 NCSE of frontal origin were similar to those found in the “loss of psychic self-activation” syndrome described by Laplane et al.28 in patients with bilateral basal ganglial lesions. Although speculative, pathophysiologic mechanisms in type 1 NCSE of frontal origin may include an ictal disorganization of the circuits linking the prefrontal associative cortex and the basal ganglia.

Other patients8,15,19,24,25 were classified as CPSE because they had clear-cut alteration of consciousness with ictal epileptiform activity that involved both frontal regions simultaneously but asymmetrically. However, perhaps because alteration of consciousness is notoriously difficult to define and therefore to evaluate clinically,29 CPSE remains difficult to define precisely, as shown by the number of attempts.2,3 Delgado–Escueta and Treiman3 described two electroclinical types. The first, presumably related to primary or secondary disorganization of mesial temporal structures, is characterized by a cyclic disturbance of consciousness. Recurrent periods of staring, unresponsiveness, and stereotyped automatisms alternate regularly with a continuous confusional state with partial responsiveness and reactive automatisms. The second and most frequent type is a continuous confusional state without any marked cyclic fluctuation, often clinically indistinguishable from AS. It has been related to ictal disorganization of the frontal lobes without involvement of mesial temporal structures.3,4 Although of great conceptual interest, the validity of this phenomenologic subclassification is open to question. Most of our patients with type 1 NCSE of frontal origin showed marked cycling without overt confusion, and none of the patients’ episodes of type 2 NCSE of frontal origin was organized in any continuous fashion despite clear impairment of consciousness.

Several authors6,11,12,21,25 have emphasized the similarities between NCSE of frontal origin and AS, especially when it occurs in patients without known epilepsy. These similarities are probably not fortuitous. Several reports7-9,12,13,19 show the transformation from NCSE with frontopolar focal ictal discharges to AS with bilateral, synchronous, and asymmetric paroxysmal discharges. Secondary bilateral involvement and synchronization of focal ictal discharges, which may originate from epileptogenic zones in parasagittal areas of the frontal, parietal, or occipital lobe,30 are a more likely explanation. Also, minor structural changes, such as diffuse or frontally accentuated corticosubcortical atrophy, are found frequently on imaging studies in patients with de novo AS of late onset. Some authors7,21 think that this may be relevant. Epileptogenic factors might express themselves more easily in brains damaged by vascular, degenerative, or toxic factors. Secondary bilateral synchronicity may also explain the development of AS with “generalized” discharges during the course of a partial epilepsy, especially of frontal lobe origin.7,8

NCSE of frontal origin is a diagnostic challenge. It is rare, the symptoms are unusual, and the patients should be documented extensively. Unresponsiveness in our Patient 6 was a prominent, conscious negative ictal phenomenon, rarely described in NCSE.31 Careful examination was needed to avoid misinterpreting this ictal paralysis as impaired consciousness. Neuropsychologic examination is often necessary to document slight disturbances of cognitive function in NCSE, and thus to support the diagnosis.32

Emergency EEG is the cornerstone of diagnosis.33 To rule out AS and others types of partial NCSE, the EEG must be sufficiently long and detailed to record seizures from their onset and to record from anterior frontal regions.5 In most cases, nonepileptic conditions are diagnosed before the EEG. These have included prolonged postictal encephalopathy34; psychiatric illnesses such as depression,5 acute or interictal psychoses,35 and factitious disorders5,35; medication overdose or idiosyncratic reaction to antiepileptic drugs36; psychotropic drug withdrawal6,7,37; toxic or metabolic encephalopathy37; the amnesia-and-automatism syndrome produced by a short-acting BZ6,7; transient global amnesia5,32; and transient unresponsiveness in the elderly.38

In our series, six patients had gross abnormalities of the frontal lobes. Three had a brain tumor and three others had a history of surgery for nonmalignant frontal lesions. Focal frontal lesions have been reported previously in 35% of cases of NCSE of frontal origin.5,13,14,19,20,22,23,26 Detailed imaging is mandatory, especially when NCSE of frontal origin is the first indication of epilepsy. This occurred in six of our patients and in 36% of other reported cases.5,7,12,13,15,22-24

After IV BZ, there was initial failure to respond or early recurrence of SE in 8 of 10 patients. PHT seemed to be the drug of choice in all our BZ-resistant patients. But Patient 8, who had no obvious triggering factor, required pentobarbital anesthesia. Several factors may explain resistance to treatment. Partial NCSE is reportedly difficult to treat. Granner and Lee33 found that only 60% of partial NCSE responded to IV BZ, which nevertheless reduced the extent of the discharges and helped to circumscribe better the underlying epileptic focus. Most of our cases were related to a focal brain lesion or to a clear nonlesional precipitating factor of SE, such as nonketotic hyperglycemia or neurosyphilis. Refractory cases of SE have often been related to a definite etiology that may require specific treatment.2 Diagnostic delay in our series was approximately 48 hours, and the greater the delay, the greater the likelihood of initial treatment failure.2

No subjective or objective impairment of cognitive functions was found in our patients after the termination of their SE. However, unfavorable outcomes have occurred, and early diagnosis and vigorous treatment of all patients with partial NCSE is required. Delgado–Escueta and Treiman,3 and Engel et al.39 reported severe, prolonged amnesia that was permanent in one patient. Krumholz et al.40 reported 10 patients: 3 died, 4 had permanent memory impairment, and 3 had memory and cognitive disturbances that lasted more than 3 months. However, unlike our patients, most of these patients had CPSE of mesial temporal origin for more than 36 hours.