SEARCH

SEARCH BY CITATION

Keywords:

  • DWI;
  • Status epilepticus;
  • Epilepsy;
  • Neuroimaging

Summary

  1. Top of page
  2. Summary
  3. Patients and Methods
  4. Results
  5. Discussion
  6. Acknowledgment
  7. References

Purpose: Diffusion-weighted magnetic resonance imaging (DWI) is used to detect changes in the distribution of water molecules in regions affected by various pathologies. Like other conditions, ictal epileptic activity, such as status epilepticus (SE), can cause regional vasogenic/cytotoxic edema that reflects hemodynamic and metabolic changes. This study describes the electroclinical and neuroimaging findings in 10 patients with partial SE whose DWI evaluation disclosed periictal changes related to sustained epileptic activity.

Patients and Methods: In this retrospective study we selected 10 patients with partial SE of different etiologies (six acute symptomatic SE; four with previous epilepsy and concomitant precipitating factors) who underwent video-EEG (electroencephalography) monitoring and a DWI study during the periictal phase. We analyzed ictal electroclinical features and DWI changes in the acute phase and during the follow-up period.

Results: DWI images revealed significant signal alterations in different brain regions depending on the location of ictal activity. DWI changes were highly concordant with the electroclinical findings in all 10 patients. As the SE resolved and the clinical conditions improved, DWI follow-up showed that the signal alterations gradually disappeared, thereby documenting their close relationship with ictal activity.

Conclusions: This study confirms the usefulness of DWI imaging in clinical practice for a more accurate definition of the hemodynamic/metabolic changes occurring during sustained epileptic activity.

Many neuroimaging techniques disclose in vivo changes in the region involved by epileptic activity during prolonged seizures or status epilepticus (SE). These changes are caused by factors such as increased metabolic activity, the consequent ictal hyperperfusion (often followed by postictal hypoperfusion), and some transient ultrastructural pathologic alterations, all of which reflect the sustained electrical activity of epileptic neurons.

Since ictal functional changes were first documented by angiography (Penfield, 1933), they have been investigated by means of other neuroimaging techniques that explore hemodynamics perfusion computed tomography (CT), perfusion magnetic resonance imaging (PWI), functional MRI (fMRI), single-proton emission computed tomography (SPECT)] and brain metabolism [positron emission tomography (PET) and spectroscopy-MRI].

By measuring the apparent diffusion coefficient (ADC), diffusion-weighted MRI (DWI) detects changes in the distribution of water molecules in regions affected by various forms of epileptic activity, such as partial SE (Cole, 2004). DWI is used to investigate many brain diseases (namely, acute ischemic stroke, brain tumors, infections, etc.), as it is sensitive to the molecular motion of water, or diffusivity, an intrinsic property of tissues (Grant et al., 2001). Changes in water diffusion in the epileptic focus are related to hemodynamic and metabolic changes in the brain tissue, which cause cytotoxic or vasogenic edema specifically detected by DWI (Yogarajah & Duncan, 2007).

This article describes the electroclinical and neuroimaging findings in 10 patients with symptomatic (or probably symptomatic) partial SE, whose DWI evaluation disclosed periictal changes related to sustained epileptic activity.

Patients and Methods

  1. Top of page
  2. Summary
  3. Patients and Methods
  4. Results
  5. Discussion
  6. Acknowledgment
  7. References

Inclusion criteria

We retrospectively examined patients admitted to our ward between 2002 and 2008 for SE or recurrent seizures who underwent DWI during the periictal period. The inclusion criteria were: (1) well-documented SE by video-EEG; (2) MRI scan showing signal changes in DWI sequences temporally related to the seizures; (3) MRI follow-up documenting the evolution of the signal changes; and (4) structural MRI showing no recent brain lesions; patients in whom DWI alterations were suspected to be related to underlying structural acute brain pathology in addition to SE, or patients for whom no follow-up MRI was available were excluded.

Patient population

Ten patients (six men, four women, mean age 68.1 years, range 16–90 years) were selected. Six of the 10 patients had no history of epilepsy and presented SE as the first manifestation of a symptomatic (or probably symptomatic) epileptic condition; four patients had a previous diagnosis of symptomatic partial epilepsy. The general characteristics of the patients, including their medical and structural MRI data, are shown in Table 1.

Table 1.   General characteristics, electroclinical and neuroimaging findings of the patients
PatientGeneral characteristicsSE electroclinical featuresNeuroimaging findings
Sex/ ageHistory of epilepsySeizure typeEtiology of epilepsySE etiologySE clinical featuresSE EEG findingsSE duration (days)SE therapyStructural MRILocation of DWI hyperintensityDWI/T2/ADC (acute phase) DWI/T2/ ADC (follow-up)
  1. SE duration defined as the interval between the onset of symptoms (identified on the basis of the information reported by the patients’ relatives or by medical staff) and video-EEG–documented SE resolution (defined as the disappearance of or significant reduction in epileptiform abnormalities).

  2. SPS, simple partial seizures; CPS, complex partial seizures; SG, secondary generalizations; TC, tonic–clonic; R, right; L, left; LVFA, low voltage fast activity; SW, spike-and-waves; PSW, polyspike-and-waves; BDZ, benzodiazepine; PHT, phenytoin; TPS, thiopental; VPA, valproate; PB, phenobarbital; LEV, levetiracetam.

116/MYesSPS, CPS, SGFragile X syndromeGastroenteritisR head/eyes deviation, swelling, clonic jerks in L limbs) [RIGHTWARDS ARROW] comaRecurrent seizures with LVFA in L posterior temporal regions [RIGHTWARDS ARROW] extensive spreading [RIGHTWARDS ARROW] PSW discharge in the L posterior regions triggering fast activity in the L frontal areas4BDZ, PHT, TPSMild cortical/ subcortical atrophy, asymmetry of lateral ventricles (L > R), signal alteration in L temporomesial structures (dysplasia? hippocampal sclerosis?) Extensive L cortical involvementDWI[UPWARDS ARROW]; T2[UPWARDS ARROW]; ADC[DOWNWARDS ARROW] (−35%) (2nd day)Complete resolution (20th day)
260/FYesSPS, CPS, SGIschemic stroke in L central region FeverClonic jerks in R body with predominant involvement of proximal and abdominal muscles, confusional stateSubcontinuous delta slow waves in L centrotemporal regions and rhythmic theta-delta activity and spikes in L centroparietal regions 8BDZ, VPA, PBL temporal strokeL temporoparietal cortex DWI[UPWARDS ARROW]; T2[UPWARDS ARROW]; ADC[DOWNWARDS ARROW] (−28%) (3rd day)Complete resolution (12th day)
390/FNoHyponatremiaClonic jerks at first in R, then in L upper limb [RIGHTWARDS ARROW] perioral muscles, sialorrhea and confusional statePeriodic delta activity and sharp waves in L fronto-temporal regions [RIGHTWARDS ARROW] sharp waves and SW activity in wide R hemispheric area6BDZ, LEVVascular encephalopathy, diffuse cerebral atrophyBilateral (sn > dx) perisylvian cortex and left thalamusDWI[UPWARDS ARROW]; T2[UPWARDS ARROW]; ADC[DOWNWARDS ARROW] (−22%) (6th day)Complete resolution (16th day)
472/MNoOperated subdural hematomaClonic jerks in R upper limb and aphasia [RIGHTWARDS ARROW] coma with ictal bradypneaSubcontinuous rhythmic spike activity and recurrent seizures with LVFA [RIGHTWARDS ARROW] rhythmic polyspike activity [RIGHTWARDS ARROW] periodic shape waves in L frontotemporal regions with spreading6PHT, BDZL hemispheric subdural hematoma L temporoparietal cortex DWI[UPWARDS ARROW]; T2[UPWARDS ARROW]; ADC[DOWNWARDS ARROW] (−36%) (4th day)Complete resolution (7th day)
582/FNoUnknownTC seizure [RIGHTWARDS ARROW] confusional state Persisting delta activity, periodic spikes in L frontocentrotemporal regions [RIGHTWARDS ARROW] recurrent paroxysms with LVFA and rhythmic theta activity in L temporal region4BDZVascular encephalopathy, cortical atrophyL hippocampal cortexDWI[UPWARDS ARROW]; T2[UPWARDS ARROW]; ADC[DOWNWARDS ARROW] (−20%) (3rd day)Complete resolution (9th day)
683/FNoUnknownClonic jerks in perioral regions and L limbs [RIGHTWARDS ARROW] confusional stateRecurrent seizures with LVFA in R frontocentral regions [RIGHTWARDS ARROW] sustained rhythmic polyspike activity in R hemisphere with contralateral transmission5PHTVascular encephalopathy, cortical/ subcortical atrophyR hemispheric cerebral cortex, left hemispheric cerebellar cortexDWI[UPWARDS ARROW]; T2[UPWARDS ARROW]; ADC[DOWNWARDS ARROW] (−35%) (1st day) Complete and gradual resolution (2nd day–7th day–15th day)
786/MNoHyperkaliemia (during acute pancreatitis) R head deviation followed by TC seizure [RIGHTWARDS ARROW] confusional stateGeneralized slow activity and periodic spike activity in L frontotemporal region4BDZ, PBCortical atrophy, chronic ischemic lesionsL hippocampal uncusDWI[UPWARDS ARROW]; T2[UPWARDS ARROW]; ADC[DOWNWARDS ARROW] (−20%) (2nd day)Complete resolution (30th day)
867/MYesSPS, CPS, SGHemorrhage in L frontal/ perisylvian regions Drug withdrawalMotor inhibition, akinetic mutism, clonic jerks in bilateral perioral region, L head deviation, oral and gestural automatisms, graspingLVFA in L temporal and frontocentral regions [RIGHTWARDS ARROW] recruiting rhythmic spikes [RIGHTWARDS ARROW] generalized 4 Hz SW activity 6BDZ, PHT, LEV, PBL perisylvian atrophy and gliosis L dorsolateral and polar frontal cortex DWI[UPWARDS ARROW]; T2[UPWARDS ARROW]; ADC[DOWNWARDS ARROW] (−26%) (2nd day)Complete resolution (25th day)
964/MYesSPSHemorrhage in R parietal region following vascular malformation rupture UnknownClonic jerks in L upper limbSubcontinuous theta-delta slow waves and sharp waves, spikes and SW in R frontal and centroparietal regions 9BDZR parietal hemorrhageR pre-central gyrusDWI[UPWARDS ARROW]; T2[UPWARDS ARROW]; ADC[DOWNWARDS ARROW] (−18%) (2nd day)Complete and gradual resolution (22nd day)
1061/MNoSPSUnknown, HIV infection as risk factor Motor inhibition, akinetic mutismRecruiting rhythmic spikes in R frontocentral regions [RIGHTWARDS ARROW] rhythmic theta activity [RIGHTWARDS ARROW] SW activity with contralateral transmission4LEVNormalL fronto-mesial cortexDWI[UPWARDS ARROW]; T2[UPWARDS ARROW]; ADC[DOWNWARDS ARROW] (−32%) (3rd day)Complete resolution (16th day)

Video-EEG study

According to the inclusion criteria, all patients underwent video-EEG monitoring with polygraphic channels (Telefactor System, 21 channels, International 10–20 System) during the SE and postictal period. During ictal activity, we clinically assessed all the patients and treated them while they were being monitored by video-EEG to better document their response to treatment. A satisfactory follow-up EEG performed after the SE was available for all patients.

MRI data acquisition and analysis

MR imaging was performed on a Gyroscan NT Intera Philips 1.5 Tesla system with a 5-mm slice thickness acquiring axial T1-weighted spin echo (TR\TE: 590\15), proton density and T2-weighted (TR 2800, TE 40/110), fluid-attenuated inversion recovery (FLAIR) (TR\TE: 6000\120, TI: 2000), and DWI (TR\TE: 4400\90) images. The ADC maps were obtained from the diffusion images. Regions of interest of a uniform shape and size (elliptic, 50 mm2) were chosen and drawn by a neuroradiologist. Diffusion gradients were applied along the three principal orthogonal axes using single-shot spin-echo echo-planar sequences with the following parameters: matrix, 224 × 224; field of view, 220 mm; section thickness, 5 mm; intersection gap, 1 mm; maximum gradient strength, 22 mT/m; acquisition time, 32 s; and b values, 0 and 1,000 s/mm2. ADC maps were also generated and analyzed directly on the images with b values of 0 s/mm2 obtained from the DWI sequence.

Results

  1. Top of page
  2. Summary
  3. Patients and Methods
  4. Results
  5. Discussion
  6. Acknowledgment
  7. References

Electroclinical findings

All of the selected cases presented a partial SE whose semiology partially changed during its evolution. The main clinical findings consisted of partial motor SE in five patients and a nonconvulsive partial SE in five patients. The duration of the status ranged from 10 h to 9 days. The electroclinical patterns of each patient are described in Table 1.

Ictal/periictal DWI and neuroimaging follow-up

DWI revealed significant signal alterations in different brain regions depending on the location of ictal activity. Analysis of the ADC value confirmed the significant reduction in these areas of increased signal. In almost all the cases, conventional MRI documented hyperintense signal alterations in the same regions on the long TR sequences, sometimes associated with sulcal enlargement. The signal alteration involved neocortical areas in eight patients, with and without extension to the underlying white matter, and was limited to the hippocampal regions in the remaining two patients. The subcortical structures (thalamus and contralateral cerebellar hemisphere) were involved in only two patients (in whom the altered signal involved the bilateral perisylvian cortex and a large hemispheric region). Unfortunately, MR perfusion imaging was performed in only one patient and documented hyperperfusion in the corresponding areas. The signal abnormalities observed in the acute phase largely resolved in all the patients and the DWI/FLAIR signals returned to normal between 9 and 30 days after the SE.

Electroclinical and DWI correlation

MRI showed areas of abnormal diffusion that were highly concordant with the clinical semiology and with the pathologic EEG activity in all patients in terms of location and lateralization. EEG ictal activity and DWI images of three cases are shown in Figures 1 and 2.

image

Figure 1.   Electroencephalography (EEG) pattern in three of the selected patients. Case 2: patient with symptomatic partial epilepsy (left ischemic stroke) and continuous clonic jerks involving the right abdomen and upper limb proximal muscles. Case 4: patient with subdural hematoma and postsurgical aphasia associated with continuous clonic jerks involving right face muscles. Case 5: patient with vascular encephalopathy and persisting confusional state after a generalized tonic–clonic seizure. Corresponding EEG ictal patterns are described in the table.

Download figure to PowerPoint

image

Figure 2.   Diffusion-weighted imaging (DWI) signal alterations documented in the acute phase and at follow-up in the same three patients. Images show DWI changes involving the left centroparietal (case 2), left temporoparietal (case 4), and left temporomesial areas (case 5) respectively.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Summary
  3. Patients and Methods
  4. Results
  5. Discussion
  6. Acknowledgment
  7. References

This study confirms the capability and reliability of DWI to demonstrate SE-related regional brain changes reflecting the hemodynamic and metabolic alterations observed during prolonged epileptic activity (Cole, 2004).

The spatiotemporal correlation observed between these changes and the electroclinical findings in all our patients indicates that these transient alterations are strongly dependent on the pathologic electrical activity. The changes rely on a complex pathophysiologic mechanism through which prolonged ictal activity, early hemodynamic changes, and impaired energy metabolism are associated with the consequent development of edema. The prolonged overactivation of epileptic neurons increases glucose and O2 requirements, thereby causing compensatory regional hyperperfusion. When the enhanced cerebral blood flood (rCBF) is no longer sufficient to prevent local hypoxia, pathophysiologic events (i.e., anaerobic glycolysis, production of lactic acid, reduction of the ATP, failure of Na+/K+-ATPase and the cell membrane) lead to cytotoxic edema and reduced extracellular volume (as documented by decreased ADC). Subsequently, the breakdown of the blood–brain barrier and possible cell death cause vasogenic edema and increased water diffusivity (Yogarajah & Duncan, 2007). Depending on whether the edema is prevalently cytotoxic or vasogenic, the predominant diffusion abnormalities in the epileptogenic area may result in a reduction (Szabo et al., 2005) or an increase in ADC (Scott et al., 2006), or both (Hong et al., 2004). The decreased ADC found in all patients in this study probably reflects the predominant cytotoxic component, which is closely related to the substantial reversibility of the alterations.

This reversibility seems to be confirmed, since none of these patients presented brain regional damage at the structural MRI follow-up. However, our observation is likely to be conditioned by the low number of patients with temporomesial structure involvement who are more frequently associated with SE-related fixed damage, as documented by various studies in animal models and a small number in humans (Farina et al., 2004).

Our results confirm those of other studies reporting the occurrence of DWI changes in cases of symptomatic (or probably symptomatic) epilepsies or SE (Lansberg et al., 1999; Kim et al., 2001; Szabo et al., 2005). These findings suggest that some insults, which may be either well-documented or “hidden,” make tissue more susceptible to DWI changes. In other words, a symptomatic etiology may yield more readily detectable DWI alterations than idiopathic conditions, in which self-limiting activity and intrinsic protective properties are presumed to be present.

This empirical observation is clearly limited by a substantial lack of DWI studies on idiopathic epilepsy and conditions. However, indirect data do come from specific conditions such as nonconvulsive SE in idiopathic generalized epilepsy, the benign nature of which is widely accepted if compared with clinically similar forms of cryptogenic/symptomatic SE (Shorvon & Walker, 2005).

Although no conclusions can be drawn from our study, it does suggest that some areas, such as the peri-rolandic cortex and the hippocampus, are more susceptible to DWI abnormalities than other areas (Hufnagel et al., 2003; Szabo et al., 2005). This increased regional susceptibility may result from a particular underlying vascular and structural organization (e.g., cell architecture) or a previous detectable (such as hippocampal sclerosis) or undetectable predisposing injury.

The fact that the neocortical structures in our study are more frequently involved than the temporomesial region is in contrast to previously published data (Scott et al., 2006).This apparent discrepancy is likely caused by differences in the selection criteria adopted, all of which must invariably tackle the intrinsic difficulty of designing DWI studies in uniform and strictly selected SE populations. From a speculative point of view, the complete reversibility of neocortical structure DWI alterations and lack of chronic damage as well as the more frequent reports of both transient and fixed MRI alterations in temporomesial regions suggest that the latter regions are more susceptible to seizure-induced neuronal damage.

This last point raises a strongly debated issue regarding DWI studies, that is, the potential role of this technique in predicting the development of seizure-induced brain damage and the consequent clinical implications. Indeed SE in animal models usually causes cerebral signal changes that first reflect cytotoxic/vasogenic edema and then degenerative phenomena, including neuronal cell loss. The brain areas damaged most in these models (i.e., the hippocampus, amygdala and pyriform cortex, and the neocortex) closely correspond to those in which a rCBF increase and an ADC maximal decrease are observed in the acute phase (Fabene et al., 2003; Engelhorn et al., 2006). As mentioned previously, the few clinical studies on this issue report permanent structural abnormalities after SE or seizures that mainly affect the hippocampus (Farina et al., 2004; Parmar et al., 2006), and rarely structures other than those in the temporomesial area (Lansberg et al., 1999; Bauer et al., 2006).

This regional susceptibility seems to be related not only to the previously mentioned structural vulnerability, but above all to an excitotoxic mechanism mediated by intrinsic neuronal seizure activity, which results in increased glutamate release from the presynaptic terminal, and to the widespread distribution of N-methyl-d-aspartate (NMDA)–type glutamate receptors, which are concentrated in the limbic system (Wang et al., 1996). Although data from animal studies have shown a close relationship between acute phase ADC changes and neuronal loss, the scanty results from clinical studies (Parmar et al., 2006; Scott et al., 2006) do not allow any conclusions to be drawn, and the predictive role of DWI in clinical practice thus remains purely speculative. Indeed, damage determined by substances used to induce SE cannot be ruled out in common animal models (Ogita et al., 2003) and the possibility that the brain damage is a preexisting condition causing seizures cannot be excluded in human studies.

Follow-up imaging in our study showed reversible DWI alterations (in all the cases consisting of decreased ADC) related to the complete resolution of the SE and the absence of long-term neurologic sequelae in all patients. Despite our small cohort, this study confirms the clinical usefulness of DWI as a means of more accurately defining the hemodynamic/metabolic changes occurring during sustained epileptic activities. Prospective studies are warranted to meet the lack of data on this particular application of DWI and to shed more light on its potential role in clinical practice.

Acknowledgment

  1. Top of page
  2. Summary
  3. Patients and Methods
  4. Results
  5. Discussion
  6. Acknowledgment
  7. References

Disclosure: The authors declare no conflicts of interest.

References

  1. Top of page
  2. Summary
  3. Patients and Methods
  4. Results
  5. Discussion
  6. Acknowledgment
  7. References