Epileptic seizures can be associated with a variety of autonomic phenomena, including autonomic auras and partial seizures (Nagaraddi & Lüders, 2008), but also as part of the periictal phenomenology of generalized tonic–clonic seizures (GTCS) (Gastaut & Broughton, 1972). Recently, both GTCS and periictal autonomic dysregulation have come under scrutiny in the quest for precise agonal pathophysiologic mechanisms leading to sudden unexpected death in epilepsy (SUDEP). These mechanisms remain unknown, although there is consensus that multimodal approaches—incorporating cardiovascular, respiratory, autonomic, biochemical, and genetic factors—will be pivotal to success. Because most SUDEP patients die in the periictal period, the epilepsy monitoring unit (EMU) provides a unique study environment in a high SUDEP risk population. “Polygraphy” using several simultaneous physiologic measurements (electroencephalography [EEG], 3-channel electrocardiography [ECG], pulse oximetry, respiration, and continuous noninvasive blood pressure [BP]) is part of the study protocol in the Prevention and Risk Identification of SUDEP Mortality (PRISM) Project (National Institutes of Health/National Institute of Neurological Disorders and Stroke [NIH/NINDS]-NS076965-01) and reveals some interesting insights. Herein, we describe a patient with a GTCS in whom these measurements were obtained.
Periictal autonomic dysregulation is best studied using a “polygraphic” approach: electroencephalography ([EEG]), 3-channel electrocardiography [ECG], pulse oximetry, respiration, and continuous noninvasive blood pressure [BP]), which may help elucidate agonal pathophysiologic mechanisms leading to sudden unexpected death in epilepsy (SUDEP). A number of autonomic phenomena have been described in generalized tonic–clonic seizures (GTCS), the most common seizure type associated with SUDEP, including decreased heart rate variability, cardiac arrhythmias, and changes in skin conductance. Postictal generalized EEG suppression (PGES) has been identified as a potential risk marker of SUDEP, and PGES has been found to correlate with post-GTCS autonomic dysregulation in some patients. Herein, we describe a patient with a GTCS in whom polygraphic measurements were obtained, including continuous noninvasive blood pressure recordings. Significant postictal hypotension lasting >60 s was found, which closely correlated with PGES duration. Similar EEG changes are well described in hypotensive patients with vasovagal syncope and a similar vasodepressor phenomenon, and consequent cerebral hypoperfusion may account for the PGES observed in some patients after a GTCS. This further raises the possibility that profound, prolonged, and irrecoverable hypotension may comprise one potential SUDEP mechanism.
We recorded video-EEG and 3-channel ECG with a Nihon Kohden (Tokyo, Japan) Neurofax EEG-1100A system. Oximetry was recorded with an Oximax N-600X machine (Covidien PLC, Dublin Ireland), respiratory movements with Ambu Sleepmate, Ambu Ltd, Copenhagen, Denmark) abdominal and thoracic belts, and continuous, beat-to-beat noninvasive BP recordings with CNAP (CNSystems Medizintechnik AG, Graz, Austria). EEG files were converted to the European Data Format (EDF), output verified as identical to the original files and reviewed and analyzed in a MATLAB-based EDF viewer (MathWorks, Natick, MA, U.S.A.). After institutional review board approval, prior written informed consent for study participation was obtained.
Our patient was an 18-year-old woman with intractable, previously uncharacterized juvenile myoclonic epilepsy and with no cardiovascular, respiratory, or endocrine comorbidities. A GTCS arising from sleep was recorded while on a reduced dose of valproic acid 100 mg b.i.d. GTCS progression was typical with onset of generalization, tonic phase, and vibratory or “jittery phase,” as described by Gastaut, and a clonic phase. The postictal phase was characterized by stupor, and the EEG showed postictal generalized EEG suppression (PGES) followed by gradual clinical and EEG recovery to baseline. As shown in Fig. 1A–C, systolic BP (SBP), diastolic BP (DBP), and mean arterial BP (MAP) increased from baseline (MAP = 81 mmHg [SBP = 98 mmHg, DBP =69 mmHg]) to a maximum in the middle of the clonic phase (MAP = 94 mmHg [SBP = 209 mmHg, DBP =36 mmHg]). Five seconds after onset of version, BP signal was lost for 12 s and restored in the middle of the jittery phase. In the suppression phase, prolonged hypotension (>60 s) was seen, with BP dropping to its lowest ebb of MAP = 41 mmHg (SBP = 48 mmHg, DBP = 34 mmHg) and concurrent diminution of pulse pressure signal. Onset of recovery of MAP coincided closely with the end of postictal generalized EEG suppression (PGES).
Periictal autonomic dysregulation has been well described, most notably by Gastaut, using pulse rate, pupillometry, electrodermography, cystometry, and conventional sphygmomanometric measurements (Gastaut & Broughton, 1972). More recent technology using wrist sensors for electrodermal activity (Poh et al., 2012) and heart rate variability algorithms (Surges et al., 2009) confirm significant sympathetic and parasympathetic changes in the ictal and postictal periods.
BP changes in epilepsy and seizures have been noted in several contexts. Gastaut measured BP periictally but noncontinuously. Continuous intraradial invasive recordings in three acutely unwell patients with GTCS noted either a pattern of transient increase in MAP or an M-shaped waveform, where an initial increase in MAP was followed by a fall and then a secondary overshoot of BP at seizure end. The authors did not precisely correlate seizure phases with BP, although an association between fall in MAP and the tonic phase was observed in one of the patients (Magnaes & Nornes, 1974). Electroconvulsive therapy induces a transient initial drop in BP as recorded by continuous, beat-to-beat BP monitoring and in pulse rate (the parasympathetic phase) followed by a sharp rise in both parameters (the sympathetic phase) (Geersing et al., 2011). Electrical stimulation of insular cortex, cingulate gyrus, prefrontal cortex, and amygdalar nucleus in humans and animals have been reported to show either increase or decrease in BP without clear lateralizing value (Sevcencu & Struijk, 2010). However, continuous, beat-to-beat BP recordings in spontaneous epileptic seizures have not previously been carried out successfully because of several constraints. These include the previous absence of appropriate technology for continuous noninvasive BP, the logistics of equipment cost, and relatively stringent technician oversight required in the EMU. Polygraphic correlation of BP with oximetry, pulse rate, respiration, ECG, and EEG has therefore never previously been studied. The CNAP machine, which is used in our study, is standard equipment in many autonomic laboratories and in the operating room, where continuous noninvasive arterial BP correlates closely with simultaneous invasive arterial BP measurements (Hahn et al., 2012). The device is based on the Peñáz method, which detects vascular unloading by a photoplethysmograph. It uses beat-to-beat finger BP measurements and calibrates these with brachial cuff readings, correcting for hydrostatic differences (Penaz, 1973) (Chung et al., 2013). The automatic detection and correction of dispersed light reduces artifact, even during motion (Fortin et al., 2006). Hand movements in a vertical plane relative to the heart can produce a physiologic dip in BP recordings, although this was not relevant in our patient, who was inert, supine, and had her hand in the same horizontal plane as the heart in the postictal period. Conditions affecting peripheral perfusion such as vascular disease, cold temperature, and vasoactive medications may underestimate BP. We were able to adapt use to the EMU with appropriate protocols.
In our GTCS patient, the significant hypotension observed in the postictal period is of major interest and differs significantly from previous BP observations in GTCS (Magnaes & Nornes, 1974). The absence of bradycardia suggests a vasodepressor mechanism due to peripheral vasodilation rather than a vagally driven cardioinhibitory mechanism. In contrast to the parasympathetic nervous system, the sympathetic system mediates vasoconstriction (postsynaptic α1- and α2-adrenergic receptors) and vasodilation (β-adrenergic receptors). The latter are not directly activated by sympathetic nerves but rather by circulating catecholamines (Thomas, 2011). The sudden drop in BP in this case may reflect an abrupt cessation of sympathetic drive to α1- and α2-adrenergic receptors; a β-adrenergic mechanism in the postictal period seems less likely. It is conceivable that in some patients, severe, prolonged, and irrecoverable hypotension and failure of BP homeostasis may comprise a mechanism for SUDEP. Hypotension due to a defective neurocardiac reflex has also been observed in a parallel phenomenon, the sudden infant death syndrome (Ledwidge et al., 1998). The observation of an apparent correlation between PGES and BP in our patient (Fig. 1A–C) is interesting and may explain the PGES seen in some patients after a GTCS (Lhatoo et al., 2010). Prolonged vasovagal syncope and attendant hypotension can produce a similar EEG picture due to cerebral hypoperfusion (Brenner, 1997), where initial slowing of background rhythms is followed by high amplitude delta activity for a few seconds and then subsequent generalized flattening of the EEG, returning to normal in the reverse sequence. With the tilt-table test, neither the “flat EEG” nor loss of consciousness resolve immediately with return to the supine position, both of which occur after a further few seconds, possibly representing a reperfusion period (Ammirati et al., 1998). However, in our patient, the EEG did not return to normal baseline as rapidly as with syncope, possibly due to the postictal EEG slowing associated with postictal stupor. Nevertheless, the relationship between hypotension and PGES in our case is likely secondary to hypotensive cerebral hypoperfusion, and the BP changes do not appear artifactual. PGES has been associated with GTCS-induced postictal autonomic dysregulation in some patients (Poh et al., 2012), and there may be other, as yet undetermined, mechanisms that drive PGES. Neither respiration nor oxygenation (Fig. 1) appeared to correlate closely with BP and PGES. Hypoxia and hypercapnia do not usually cause major fluctuations in BP. Acute hypoxia can produce both a generalized, dose-dependent increase in sympathetic vasoconstrictor outflow caused by activation of the carotid chemoreceptor reflex, increased levels of angiotensin II and endothelin-I, and vasodilation mediated by epinephrine, atrial natriuretic peptide, and red blood cell–generated ATP (Tamisier et al., 2005). The balance of these two opposing trends is slightly tipped toward the latter, resulting in a mild or no change in systemic BP (Morgan, 2007). Hypercarbia produces similar but shorter lived sympathetic activation (Xie et al., 2001). In our case, carbon dioxide capnography was not used and the extent of hypercarbia, if any, is not known.
In summary, postictal hypotension may occur after GTCS and appears to closely correlate with PGES, suggesting significant hypotension and cerebral hypoperfusion as a possible cause of this phenomenon. Severe hypotension and failure of BP homeostasis may comprise one potential SUDEP, mechanism and PGES may identify patients at risk.
This study was supported in part by NINDS grant NS076965-01 – The Prevention and Risk Identification of SUDEP Mortality (PRISM) Project.
None of the authors has any conflict of interest to declare. We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.