- Top of page
Summary: Purpose: Previous studies have established the importance of the insular cortex and temporal lobe in cardiovascular autonomic modulation. Some investigators, based on the results of cortical stimulation response, functional imaging, EEG recordings of seizures, and lesional studies, have suggested that cardiac sympathetic and parasympathetic function may be lateralized, with sympathetic representation lateralized to the right insula, and parasympathetic, to the left. These studies have suggested that ictal bradycardia is most commonly a manifestation of activation of the left temporal and insular cortex. However, the evidence for this is inconsistent. We sought to assess critically the predictable value of ictal bradycardia for seizure localization and lateralization.
Methods: In this study, we reviewed the localization of seizure activity in 13 consecutive patients with ictal bradycardia diagnosed during prolonged video-EEG monitoring at Mayo Clinic Rochester. The localization of electrographic seizure activity at seizure onset and bradycardia onset was identified in all patients. In addition, we performed a comprehensive review of the ictal bradycardia literature focusing on localization of seizure activity in ictal bradycardia cases.
Results: All occurrences of ictal bradycardia in the 13 identified patients were associated with temporal lobe–onset seizures. However, no consistent lateralization of seizure activity was found at onset of seizure activity or at onset of bradycardia in this population. Seizure activity was bilateral at bradycardia onset in nine of 13 patients. The results from the literature review also showed that a predominance of patients had bilateral activity at bradycardia onset; however, more of the ictal bradycardia cases from the literature had left hemispheric localization of seizure onset.
Conclusions: Ictal bradycardia most often occurs in association with bilateral hemispheric seizure activity and is not a consistent lateralizing sign in localizing seizure onset. Our data do not support the existence of a strictly unilateral parasympathetic cardiomotor representation in the left hemisphere, as has been suggested.
Since Russell (1) reported the disappearance of a patient's pulse during a seizure in 1906, it has been recognized that seizures can influence autonomic cardiovascular function. Cardiac rhythm changes occur in the majority of epileptic seizures. The most common cardiac arrhythmia seen during epileptic seizures is sinus tachycardia, which occurs in >90% of seizures and is usually of no consequence (2,3). However, the effect of seizures on cardiac rhythm is not always benign. Potentially dangerous arrhythmias have been described in association with seizures. Therefore the influence of seizures on cardiovascular autonomic function may have great clinical significance in some cases (3–5). Sudden unexplained death in epilepsy patients (SUDEP) occurs at a rate 24 times greater than that in the general population (6). One potentially fatal arrhythmia seen in association with seizures is bradycardia. Patients with bradycardia and asystole during an epileptic seizure are said to have the ictal bradycardia syndrome (7). Once regarded as rare, significant bradycardia prompting cardiac pacemaker insertion was noted in four (21%) of 19 epilepsy patients in one study in which long-loop ECG monitoring was performed over a several-month period (8).
The focus of many studies in the ictal bradycardia literature is the lobar localization of seizure onset, with inferences relating these observations to the functional representation of central cardiac autonomic control (5,7,9–17). An emphasis also has been placed on the significance of seizure lateralization in these cases, in light of assertions from human and animal stimulation studies as to the relative importance of the right and left hemisphere in parasympathetic and sympathetic cardiovascular modulation (18). Most reports consist of results in one to five patients in which an ictal bradycardia event was captured in the course of combined EEG-electrocardiogram (ECG) monitoring. Many of these reports are summarized by Tinuper et al. (5). When localizing information was available in these studies, the analysis consisted primarily of the localization of seizure onset, but the localization of seizure activity at bradycardia onset is rarely mentioned. In terms of evaluating the relative importance of the left and right hemispheres in the parasympathetic and sympathetic modulation of heart rate, the distribution and localization of seizure activity at bradycardia onset may be of more importance than seizure onset.
We report the localization and lateralization of seizure activity at seizure onset and bradycardia onset in 13 patients with ictal bradycardia diagnosed during continuous video-EEG monitoring at our institution. This is the largest single-center series of ictal bradycardia cases to date. One purpose of this study was to evaluate whether ictal bradycardia is a reliable localizing sign in the clinical evaluation of epilepsy. The second purpose was to review the lateralization of seizure activity associated with bradycardia in light of the literature suggesting a lateralization of sympathetic and parasympathetic cardiovascular representation in the human cortex.
- Top of page
A query was performed of all adult and pediatric electronic EEG reports of studies performed at Mayo Clinic Rochester on patients undergoing prolonged EEG monitoring from January 1990 to December 2004. A search was performed for the terms “bradycardia,”“syncope,”“bradyarrhythmia,”“sinus arrest,” or “asystole.” Bradycardia was defined as an R-R interval of >2.0 s or activation of cardiac pacing in patients previously implanted with a cardiac pacemaker. Ictal bradycardia was defined as the occurrence of bradycardia during a recorded seizure. Thirteen patients meeting this definition were included. Five cases were excluded: four patients had bradycardia that was not associated with recorded seizure activity, and one had bradycardia during a seizure that was also associated with apnea.
All patients were evaluated with computer-assisted continuous 30-channel scalp EEG recordings. The International 10-20 system was used for electrode placement. A single-channel recording was used to record cardiac rhythm continuously in all cases. Multiple montages were reviewed to facilitate seizure localization and detection. One patient also underwent intracranial monitoring with bilateral medial temporal depth electrodes. In this patient, single eight-contact depth electrodes were implanted stereotactically via a posterior approach in the amygdala and hippocampus in a longitudinal orientation in both hemispheres. The localization and time of seizure onset was determined by identification and localization of the earliest rhythmic EEG change from baseline seen in association with a clinical seizure. The localization of seizure activity at the onset of bradycardia was determined by the localization of rhythmic EEG activity at the time in which the ECG met the criteria for bradycardia, as defined in this study. Two patients were monitored with pulse oximetry. Those patients taking anticonvulsant medications at the time of evaluation underwent a gradual dosage reduction to help precipitate seizure activity. All patients underwent neurologic examination and neuroimaging; and most underwent cardiologic evaluation.
In patients meeting these criteria, all clinical and electrophysiological data from the Mayo Clinic were reviewed by the authors. Seizure onset was defined as the time of onset of the earliest electrographic seizure activity on EEG during a clinical seizure. The onset of bradycardia was defined as the time of onset of bradycardia during a recorded seizure. Bradycardia duration was defined as the duration from the time of bradycardia onset to the time that the bradyarrhythmia failed to meet the criteria of bradycardia as defined in this study. Ictal tachycardia was defined as development of a heart rate >120 beats/min after onset of seizure activity on the EEG. The patient or the immediate family members were contacted to obtain follow-up information as indicated. All patients consented to participate in this study. This study received approval by the Internal Review Board of Mayo Clinic Rochester.
A review of the available literature on ictal bradycardia also was performed. Studies were selected for review if data on localization and lateralization of seizure activity were available. Articles without adequate information on seizure localization and articles discussing ictal bradycardia in association with apnea were excluded.
- Top of page
Thirteen patients were identified of 6,168 patients who underwent video-EEG monitoring at our institution during the study period (see Table 1). Ictal bradycardia was present in 29 (48%) of 60 of the recorded seizures in these patients. Of the 60 recorded seizures, 58 were partial seizures, and two were secondarily generalized seizures. All 29 seizures associated with ictal bradycardia were complex partial seizures. See Table 2 for ictal behaviors during seizures associated with bradycardia. Five patients had some seizures associated with ictal bradycardia and some without. In two patients, there was no difference in the ictal behavioral or EEG characteristics between these seizure types. In the other three patients, decerebrate posturing occurred in two patients during ictal asystole, and loss of tone with brief myoclonic jerks occurred in the other, while the seizures without associated bradycardia were manifested by behavioral arrest, confusion and automatisms but not with extensor posturing or loss of tone. The mean age at diagnosis of ictal bradycardia was 45.7 years (range, 17–73 years). The mean age at seizure onset was 35.8 years (range, 8–69 years). Twelve patients were right-handed, and in one patient we did not have information on handedness. The mean seizure-disorder duration at diagnosis was 10.5 years (range, 9 days to 57 years). Ten of 13 patients were taking anticonvulsant medication at the time of diagnosis, and three were not taking anticonvulsant medication, as their diagnosis of epilepsy had not been established before evaluation with prolonged video-EEG monitoring. Two patients had previously been implanted with a permanent pacemaker based on a diagnosis of cardiogenic bradycardia, as detected on continuous ECG without EEG monitoring at other institutions. In our cohort, two patients were taking carbamazepine (CBZ), two were taking valproic acid (VPA), four were taking phenytoin (PHT), and nine were taking newer antiepileptic medications (AEDs). The most commonly used AED in our patients was levetiracetam (LEV), which was used by six patients. The two patients monitored with pulse oximetry did not show oxygen desaturation during their clinical events.
Table 1. Ictal bradycardia: patient characteristics
|Pt||Age/sex||Dur||Radiological findings||Total Sz||IB||IT|
| 1||67F||9 d.||MRI: subacute L subinsular infarct, coiled LICA aneurysm||1||1||0|
| 2||72F||3 y.||MRI: normal||1||1||0|
| 3||35F||5 y.||CT with contrast: normal||8||8||0|
| 4||54F||8 m.||MRI: normal||1||1||0|
| 5||52F||14 y.||MRI: left temporal glioma||1||1||0|
| 6||73M||8 y.||MRI: right caudate lacunar infarct||3||3||0|
| 7||25M||5 y.||MRI: bilateral mesial temporal sclerosis||18 ||4||0|
| 8||40M||8 y.||MRI: bilateral mesial temporal sclerosis||4||4||4|
| 9||17M||18 m.||MRI: normal||2||1||0|
|10||28M||12 y.||MRI: heterotropia of right occipital horn||6||1||2|
|11||33F||4 y.||MRI: normal||9||1||2|
|12||29F||21 y.||MRI: normal||2||2||0|
|13||70F||57 y.||MRI: non-enhancing right temporal mass||4||1||0|
| ||Total =||60 ||29 ||8|
Table 2. Ictal bradycardia: EEG and ECG findings
|Pt||BHR (bpm)||Ictal behavior||EEG @ seizure onset||EEG @ IB onset||Time to IB (seconds)||Dur of BC (seconds)|
| 2||66||ba, oa, hd||LT||BT||16||23|
| 3||74||ba, anxiety||RT||RT||25||25|
| 5||54||loc, falling limp, jerks||LT||LT||17||19|
| 7a||96||ba; buep; and hd that||3BT||BT||19||15|
| 7b||84||stopped after PPM||1LT||BT||38|| 4|
| 8||77||ba, oa, buep||2 LT; 2 ID||2 LT; 2 ID||34||36|
| 9||72||ba, buep||RT||BT||24||28|
|10||84||ba, oa, luep||RT||BT||111 ||26|
|11||78||ba, o/ha, blinking||LT||BT||10||10|
|12||63||ba, falling limp, buep fall, urinary||RT||BT||35||23|
|71.2|| ||Mean =|| 28.3|| 23.2|
Twelve patients underwent MRI imaging, and one patient had a computed tomography (CT) head scan because of the presence of a permanent pacemaker (see Table 1). Six patients had an abnormality on neuroimaging of potential epileptogenic significance. The lesions seen on MRI included bilateral mesial temporal sclerosis (two), low-grade glioma (two) (see Fig. 1), unilateral periventricular nodular heterotopia (one), and a coiled left internal carotid bifurcation aneurysm with associated subacute left subinsular infarction (one). The presence of ictal bradycardia was not associated with the lateralization of these lesions. The localization of the observed radiologic abnormalities were: right temporal lobe (one), right temporoparietal region (one), left temporal lobe (one), left subinsular region (one), and bilateral medial temporal regions (two).
Figure 1. EEG in a 20-year-old man with ictal bradycardia (patient 7): Intracranial monitoring. The EEG is displayed by using a Cz reference montage with the left mesial temporal depth derivations displayed in the uppermost section labeled with “LTD-RF,” and the left scalp derivations below these labeled with odd numbers. The ECG monitor is displayed in the central channel. The right hemispheric scalp derivations are displayed in the bottom half with even numbers, and the right mesial temporal depth derivation in the lowest section labeled “RTD-RF.” Seizure onset over the left mesial temporal depth electrode (straight arrow) is shown in the left panel. The EEG activity at the time of pacemaker activation (angled arrow) is displayed in the right panel.
Download figure to PowerPoint
EEG and ECG findings
The localization of seizure activity at seizure onset and bradycardia onset are summarized in Table 2. Seizure onset was localized over the temporal lobes in all recorded seizures associated with bradycardia in 12 patients. In one patient, two seizures were localized to the temporal lobe, and in two seizures, localization was indeterminate. Seizure onset was lateralized to the right hemisphere in seven patients, the left hemisphere in five, and was nonlateralized at seizure onset in one. Of all 29 seizures associated with bradycardia, seizure onset was identified on the left in seven seizures, the right in 17, was bitemporal in three, and indeterminate in two. The EEG of patient 7 during intracranial monitoring is depicted at the onset of the seizure and at the onset of bradycardia in Fig. 1. In the seven patients without a potentially epileptogenic lesion on imaging, the seizure activity did not consistently lateralize to a particular hemisphere, and all but one of them had bilateral activity at the onset of bradycardia.
Seizure localization at bradycardia onset was bilateral in nine patients, lateralized to the left in three, and to the right in one. Of all 29 seizures associated with bradycardia, seizure activity at bradycardia onset was bitemporal in 15 seizures, localized to the left in four, the right in eight, and was indeterminate in two. The mean latency from seizure onset to bradycardia onset was 28.3 s (range, 10–111 s). The mean duration of bradycardia was 23.2 s (range, 4–36 seconds). No patient required resuscitation.
Bradycardia was not the only arrhythmia seen in these patients. Ictal tachyarrhythmias were identified in three patients. Two of the patients had sinus tachycardia, and in one, atrial fibrillation developed. Two patients with tachyarrhythmias had left temporal seizure onset, and one had right temporal. The onset of the tachyarrhythmia preceded spread to the other hemisphere in all of these seizures.
All but one of our patients were evaluated with a 12-lead ECG, and this was normal in six (see Table 1). One patient had atrial fibrillation, and another had a right bundle branch block. Four patients had mild baseline sinus bradycardia related to β-blocker use. None of the six patients who underwent echocardiograms had significant abnormalities.
Review of the ictal bradycardia literature
The localization and lateralization of seizure activity at seizure onset and bradycardia onset in the literature in comparison to our present study are summarized in Fig. 2. One hundred seven cases of ictal bradycardia were found in the literature, of which 81 were diagnosed during simultaneous EEG and ECG recordings during a bradycardic event (4,5,12–17,19–50). Of these 81 cases, adequate information was found to allow lobar localization of seizure onset in 65. Seizure onset was localized to the temporal lobe in 36, the frontal lobe in 13, the frontotemporal lobes in 15, and the occipital lobe in one. Information regarding seizure-onset lateralization was present in 56 cases. Seizure onset was lateralized to the left hemisphere in 35 and the right hemisphere in 19. Bilateral ictal onset was found in two. Only 22 cases had information on EEG lateralization at the onset of bradycardia. Of these, 12 showed bilateral hemispheric electrographic seizure activity at bradycardia onset, six showed left hemispheric seizure activity, and four showed right hemispheric seizure activity.
Figure 2. Lateralization of EEG activity at seizure onset and bradycardia onset from the present study and the review of the literature.A: The total number of cases identified from the literature with simultaneous EEG and ECG monitoring that had information on lateralization of seizure onset. B: Lateralization of seizure onset in the 13 cases identified in the present study.C: Total number of cases identified from the literature with simultaneous EEG and ECG monitoring that had information on lateralization of EEG activity at bradycardia onset.D: Lateralization of EEG activity of all 13 cases in the present study at bradycardia onset.
Download figure to PowerPoint
- Top of page
Three conclusions can be drawn from this study. First, it is clear that ictal bradycardia is seen primarily in association with seizures involving the temporal lobes. Second, ictal bradycardia is not a reliable lateralizing sign in the localization of seizure origin in patients with partial epilepsy. Third, ictal bradycardia most often occurs in association with bilateral hemispheric seizure activity. Therefore our data do not support the presence of a specific parasympathetic cardiac representation in the left hemisphere, as has been suggested (18).
All of the episodes of ictal bradycardia in our series were seen in association with temporal lobe seizures, consistent with the majority of reports in the ictal bradycardia literature (5). It is clear that the presence of bradycardia during a seizure is of localizing value, suggesting temporal lobe onset in patients with partial epilepsy. However, caution should be used in extrapolating data from our population to the normal population, as all of our patients had temporal lobe epilepsy, and many of them had temporal lesions. Previous lines of research have demonstrated that the temporal lobe influences cardiovascular function including heart rate, and diseases of the temporal lobe may affect cardiovascular autonomic function. For example, infarction of the temporal lobe and insula has been associated with decreased heart-rate variability (51,52). Abnormalities on power-spectral analysis of heart-rate variability have been identified in the temporal lobe epilepsy population in other studies (53–57). Changes in cardiac rhythm have been demonstrated in association with electrical stimulation of the insular and temporal cortices by other investigators (18,44). It should be noted that not all investigators have been able to demonstrate cardiovascular effects in response to disturbances of function in the temporal lobe and insula (58–60).
Although useful for lobar localization, our data suggest that ictal bradycardia is not a consistent lateralizing sign. In our study, ictal bradycardia was seen in right temporal onset seizures in seven patients and left temporal in five. In addition, the lateralization of intracranial lesions that likely bore an etiologic relation to the seizures in six patients was not consistent. The collective literature also does not show ictal bradycardia to be of reliable consistency in terms of seizure lateralization. It also is important to note that a consistent cardiac response was not seen from one seizure to the next in the same patient with an otherwise identical ictal EEG pattern. In some patients, tachycardia was seen in some seizures, and bradycardia, in other seizures. Interestingly, in these cases, the ictal EEG findings were otherwise identical. Similar findings have been described by other authors (8). These findings provide further evidence that the presence of particular changes in heart rate do not reliably lateralize seizure onset, and that cortical regulation of heart rate does not appear to be precisely lateralized in humans. The output of the cardiovagal system relies on a balance between the sympathetic and parasympathetic system. This study reemphasizes the complexity of the interaction between cortical excitation and heart-rate changes.
Our data indicate that bradycardia is more likely to occur in the setting of bilateral hemispheric seizure activity and provide no evidence to support that parasympathetic cardiac representation is lateralized in the cerebral cortex. The distribution of seizure activity at the onset of bradycardia is not consistently discussed in the ictal bradycardia literature but may bear a more direct relation to the location of cortical regions involved in the bradycardic response than does the localization of seizure onset. Other investigators have suggested that the parasympathetic and sympathetic central modulation regions in the cortex are lateralized. Oppenheimer et al. (18), based on a series of unilateral stimulation studies of the insula in five patients undergoing epilepsy surgery, demonstrated differences in cardiovascular response on stimulation of the left and right insular cortex. These investigators concluded that parasympathetic cardiovascular responses appeared to be lateralized to the left insular cortex, and sympathetic responses to the right. This conclusion was based on a mean increase in heart rate of 6 beats/min on stimulation of what were called “tachycardia” sites (most of which were present in the right insula) and a mean decrease of 5 beats/min in stimulation of “bradycardia” sites (usually present in the left insula). In interpreting these data, it is important to keep in mind that unilateral insular stimulation may give rise to contralateral insular activation via transcallosal connections between the right and left insula (61). The presence of contralateral insular activation cannot be excluded in this study, as contralateral electrocorticography was not performed.
Ictal bradycardia was not common in the patients undergoing prolonged video-EEG monitoring at our institution and was discovered in <1% of these patients. It should be noted that all EEG reports of all patients undergoing prolonged video-EEG monitoring were screened for the presence of bradycardia, including those patients undergoing evaluation for symptomatic events other than epilepsy. This study was not intended or designed to measure the prevalence of ictal bradycardia in epilepsy patients. We also used strict criteria for defining ictal bradycardia, which were not always used in some of the series reporting a higher prevalence. For instance, Leutmezer et al. (38) found ictal bradycardia in two (1.4%) of 145 seizures. Both of these patients would not have met our criteria with minimal heart rates of 46 and 51 beats/min. It is possible that our selection method led to underreporting some patients with milder ictal bradycardia, but this is unlikely to have biased our results on the lateralization of seizure activity.
Some limitations should be kept in mind when using scalp EEG in the cerebral localization of clinical phenomena. EEG may fail to detect seizure activity in brain regions such as the medial temporal and insular cortices, in some cases because of their distance from the scalp or if the generated potential fields are oriented tangential to the electrode recording surface (62,63). Accordingly, the possibility of concurrent contralateral temporal activation cannot be entirely excluded in our patients with unilateral electrographic seizure activity on scalp EEG at bradycardia onset or in those cases reported in the literature. In our study, the patient that underwent monitoring with bilateral mesial temporal depth electrodes clearly did not have heart-rate changes until bilateral activation occurred. Similar findings have been reported by others who observed ictal bradycardia during intracranial monitoring (12,50). The inability to exclude concurrent contralateral activity renders it impossible to establish the existence of a lateralized cardiac parasympathetic center on the basis of scalp EEG. Another limitation of scalp EEG is that the potential contribution to the bradycardic response from activated subcortical regions cannot be excluded. Microelectrode stimulation and recording studies, ictal HMPAO-SPECT, and analysis of c-fos activation during induced seizures have shown that hypothalamic and brainstem nuclear activation is routinely present during seizures (64–66).
The relative importance of the right and left hemisphere in the parasympathetic and sympathetic modulation of the cardiovascular system also has been evaluated by assessing the effects of cortical infarction and lateralized amobarbital infusion on cardiac autonomic modulation. These studies have not yielded consistent results. A reduction in parasympathetic cardiovascular function was found in patients after right hemispheric stroke in one study, suggesting lateralization of parasympathetic modulation to the right (67,68). In other studies, effects on both sympathetic and parasympathetic function were seen with both right- and left-sided cerebral infarction (51,69). A few investigators noted a clearer association of ECG changes, such as QT prolongation and complex arrhythmia, with right hemispheric stroke compared with left (52,70). However, other investigators have shown a closer association of sudden death with left cerebral infarcts (71). Still others confirmed the association between cardiac autonomic dysregulation and the presence of a cerebral infarction but were unable to demonstrate a difference between left- and right-sided lesions (72). The intracarotid amobarbital (Wada) test also has been used to evaluate the lateralization of parasympathetic and sympathetic function, with some investigators showing evidence of lateralization (73,74), whereas others did not (75,76). Faced with these variable results and the findings of the present study and in the ictal bradycardia literature, a consistent model of lateralized cortical cardiovascular sympathetic and parasympathetic representation cannot now be constructed.