Cardiac injury in refractory status epilepticus


Address correspondence to Sara Hocker, Department of Neurology, Mayo W8-B, Mayo Clinic, 200 First Street SW, Rochester, MN 55905, U.S.A. E-mail:


Purpose:  We sought to describe the spectrum of cardiac injury in refractory status epilepticus (RSE).

Methods:  We reviewed all patients with RSE between 1999 and 2011 at Mayo Clinic. RSE was defined as generalized convulsive or nonconvulsive status epilepticus (SE) that continued despite initial therapies. Exclusion criteria were age <18 years, anoxic SE, psychogenic SE, simple partial SE, absence SE, and repeat RSE. Patients were divided into those with (transient left ventricular [LV] dysfunction; electrocardiography [ECG] abnormality—new T-wave inversion, ST elevation or ST depression, or QTc prolongation; and/or elevated blood troponin T levels) versus those without evidence of cardiac injury.

Key Findings:  We identified 59 consecutive patients with RSE. In 24 patients no cardiac-specific diagnostic studies were obtained. Twenty-two of the remaining 35 patients demonstrated markers of cardiac injury. General anesthesia was necessary for control of seizures in 31 of 35 patients for 10.5 ± 17.4 days. Twenty-three patients had troponin levels drawn at onset of SE, of which nine were abnormal. ECG findings at onset of SE included ST elevation (11.4%), ST depression (5.7%), new T-wave inversion (37.1%), and nonspecific ST changes (37.1%). Cardiac arrhythmias included ventricular tachycardia/fibrillation (11.4%), atrioventricular block (2.9%), atrial fibrillation/flutter (20.0%), sinus bradycardia (48.6%), and sinus tachycardia (65.7%). Intervention was required for cardiac arrhythmias in 42.9%. QTc was prolonged in 22.9% of patients. One patient met criteria for non–ST-elevation myocardial infarction (NSTEMI). Three of 14 patients evaluated with ECG during SE demonstrated reversible systolic dysfunction. In-hospital mortality was 34.3%. Outcome was worse in the group with markers of cardiac injury but the difference was not statistically significant (p = 0.14).

Significance:  Markers of cardiac injury are common in RSE and may be underrecognized in this population. These disturbances may require specific treatment and are often reversible. Routine performance of ECG and troponin followed by an echocardiography in those with repolarization abnormalities is probably justified. This was a biased sample of patients with severe RSE who had cardiac studies performed. The prevalence of findings in this study refers to this subgroup only.

Cardiac manifestations during seizures are frequent and typically benign (Nei et al., 2000; Zijlmans et al., 2002; Oppenheimer, 2006). Continued tonic–clonic seizure activity, termed convulsive status epilepticus (SE), produces cardiac dysfunction, myocyte damage (Painter et al., 1993; Legriel et al., 2008; Shimizu et al., 2008), and intense activation of the sympathetic nervous system, which can result in cardiac contractile dysfunction and susceptibility to arrhythmias (Ishiguro & Morgan, 2001; Shimizu et al., 2008). In an autopsy study, death from SE was significantly associated with the presence of cardiac contraction bands when compared to controls, presumably due to massive catecholamine release resulting from the seizures (Manno et al., 2005).

Acute neurologic insults have been associated with specific patterns of neurocardiogenic injury, also known as stress-induced cardiomyopathy, apical ballooning syndrome, or Takotsubo cardiomyopathy (TC); however, not all patients meet established criteria for these syndromes (Bybee et al., 2004; Bybee & Prasad, 2008). The full syndrome presents with reversible akinesis or dyskinesis of the left ventricular segment, associated with new electrocardiography (ECG) ST-segment or T-wave abnormalities mimicking acute myocardial infarction (MI) (Prasad et al., 2008). There are case reports of TC occurring after isolated convulsive seizures (Chin et al., 2005; Bosca et al., 2008; Lemke et al., 2008), partial nonconvulsive (Benyounes et al., 2011) as well as generalized convulsive SE (Sakuragi et al., 2007; Legriel et al., 2008; Shimizu et al., 2008).

SE is generally considered refractory when seizures continue despite initial therapies. The majority of these cases require anesthetic agents for control. Hypotension occurs frequently in this population and is generally attributed to the use of anesthetic agents; however, in some cases it may result directly from neurocardiogenic injury. Little is known about the effects on cardiac function of prolonged intermittent nonconvulsive seizure activity, which often occurs in these patients. The aim of this study was to describe the spectrum of cardiac injury in refractory status epilepticus (RSE) in order to determine whether ongoing surveillance for such injury in this population may be justified.


This study was a retrospective descriptive analysis of all consecutive adult patients treated for RSE at Saint Mary’s Hospital, Rochester, Minnesota, between September 2000 and August 2011. Cases were identified and clinical data were acquired through queries of the computerized electroencephalography (EEG) report system, and the medical record system. Included were patients ≥18 years with RSE defined as generalized convulsive status epilepticus (GCSE) or nonconvulsive status epilepticus (NCSE) (partial or generalized onset) unresponsive to treatment with two antiepileptic drugs and/or requiring anesthetic agents for seizure control and monitored with continuous EEG. Patients with anoxic, psychogenic, absence or simple partial SE, and repeat RSE were excluded. The type of SE recorded was defined by the initial presentation. If SE began with a witnessed generalized tonic–clonic seizure with continued convulsions or evolution into nonconvulsive status it was recorded as GCSE. NCSE was defined as changes in behavior and/or cognition from baseline associated with continuous epileptiform discharges on EEG (Meierkord & Holtkamp, 2007). Data abstracted included: (1) patient age; (2) gender; (3) cardiovascular comorbidities; (4) type of SE; (5) anesthetic agents required for seizure control; (6) outcome; (7) presence and type of arrhythmia; (8) hypotension and requirement for intervention; (9) pulmonary edema; (10) ECG findings including T-wave inversion, QTc, and ST segment changes, (11) blood troponin T levels; and (12) ECG findings. A cardiologist (AP) blinded to outcome data reviewed all ECG recordings and echocardiograms.

Patients were divided into two groups: those with (transient left ventricular [LV] dysfunction, new T-wave inversion, ST elevation or ST depression, QTc prolongation, and/or elevated troponin T levels), and without cardiac injury (no markers of injury). Arrhythmias were not considered directly representative of cardiac injury. QTc prolongation was defined as QTc ≥500 msec on a single ECG or serial increase in QTc ≥60 msec. Hypotension was defined as a systolic blood pressure <90 mm Hg or mean arterial pressure <50 mm Hg. Bradycardia was defined as heart rate <60 beats per minute (bpm) and tachycardia was defined as heart rate >100 bpm. Outcomes were determined by modified Rankin scale at hospital discharge.

Descriptive summaries were reported as mean ± standard deviation (SD) for continuous variables and frequencies and percentages for categorical variables. Categorical outcomes of interest were compared using Fisher’s exact test. All tests were two sided and p-values < 0.05 were considered statistically significant. The Mayo Clinic Institutional Review Board approved this study.


Fifty-nine patients with RSE were included after review of inclusion and exclusion criteria. In 24 patients, no ECG, troponin, or echocardiograms were obtained. Therefore, 35 patients were included in this analysis. Twenty-two (62.9%) of the 25 patients demonstrated markers of cardiac injury. Duration of general anesthesia among the 35 patients included in the analysis was 10.5 ± 17.4 days. Baseline patient characteristics are shown in Table 1.

Table 1.   Baseline patient characteristicsa
 All patients (n = 35)Cardiac injury (n = 22)No cardiac injury (n = 13)
  1. Data presented as n (%) unless otherwise indicated.

  2. aNo statistically significant findings.

Female15 (42.9)8 (36.4)7 (53.8)
Age (mean ± SD)50.5 ± 21.149.1 ± 21.652.8 ± 20.7
Coronary artery disease2 (5.7)1 (4.5)1 (7.7)
Hypertension17 (48.6)10 (45.5)7 (53.8)
Diabetes6 (17.1)4 (18.2)2 (15.4)
Hyperlipidemia9 (25.7)3 (13.6)6 (46.2)
Smoking8 (22.9)6 (27.3)2 (15.4)
Convulsive SE15 (42.9)9 (40.9)6 (46.2)
Anesthetic agents required31 (88.6)20 (90.9)11 (84.6)

Twenty-three patients had troponin levels drawn at onset of SE and they were elevated in nine (39.1%), with a mean level of 0.09 ± 0.2 ng/ml (normal <0.01 ng/ml). ECG recordings were obtained at onset of SE in all 35 patients. Findings included ST elevation in a regional distribution (n = 4, 11.4%), ST depression (n = 2, 5.7%), T-wave inversion (n = 13, 37.1%), nonspecific ST changes (n = 13, 37.1%), and normal or no new abnormality (n = 12, 34.3%). Of those with new T-wave inversion, eight demonstrated resolution on serial ECGs and five did not have repeat ECG studies performed. QTc was prolonged in eight patients (22.9%; Fig. 1).

Figure 1.

Mean QTc interval in cardiac injury and noncardiac injury groups.

Cardiac arrhythmias occurred during the course of SE in 32 patients (91.4%) and included sinus tachycardia (n = 23, 65.7%), sinus bradycardia (n = 17, 48.6%), atrial fibrillation/flutter (n = 7, 20.0%), ventricular tachycardia/fibrillation (n = 4, 11.4%), and atrioventricular block (n = 2, 5.1%). Intervention for the arrhythmia was required in 14 cases (40%). Ventricular arrhythmias were distributed evenly between the cardiac injury and no cardiac injury groups. QTc was increased in RSE patients associated with ventricular arrhythmias (520.3 ± 87.3 msec) compared with those without ventricular arrhythmias (455.9 ± 74.1 msec).

Echocardiograms were obtained in 14 patients. Findings included 10 with normal LV and right ventricular (RV) function; 2 with generalized LV and RV dysfunction, which resolved on subsequent studies; 1 with generalized LV dysfunction, which resolved on subsequent studies; and 1 with inferior regional wall motion abnormalities (RWMAs) and no subsequent study (Table 2).

Table 2.   Three patients with reversible systolic dysfunction
AgeSexDays from onset of SE to echocardiogramEchocardiographic findingsDays to normalizationECG findingsPeak troponin (ng/ml)
18F4EF 40%
Global hypokinesis
Mild right ventricle dysfunction
55QTc 472 msec
No major abnormalities
Not measured
47F12EF 48%
Generalized left ventricle dysfunction
37QTc 477 msec
New inferior T-wave inversion
Not measured
25M11EF 49%
Mild generalized hypokinesis
Mild right ventricle dysfunction
22QTc 577 msec
Nonspecific ST and T-wave abnormalities
Initial 0.03
Not measured

Pharmacologic treatment of RSE in the three patients with demonstrated reversible systolic dysfunction included (1) pentobarbital and isoflurane during initial and follow-up studies, (2) pentobarbital and isoflurane at initial and pentobarbital at follow-up, and (3) pentobarbital and midazolam at initial and midazolam at follow-up. The following anesthetic agents utilized at the time of echocardiogram in the 10 patients with normal systolic function were propofol in 4, pentobarbital in 4, pentobarbital and isoflurane in 1, and midazolam in 2. The frequency of anesthetic use was not significantly different between the group with and the group without markers of cardiac injury (p = 0.65). The patient with inferior RWMAs met criteria for non-ST elevation myocardial infarction (NSTEMI) (Thygesen et al., 2007), had ECG findings of prior inferior, and possibly anterolateral myocardial infarction (MI) and a history of coronary artery disease.

Hypotension occurred in 28 (80%) of the patients, was not more frequent in patients with cardiac injury (p = 0.22), and required pharmacologic intervention in 26 patients (92.9%).

In-hospital mortality was 34.3% and included 10 patients (45.5%) in the cardiac injury group and two patients (15.4%) without markers of cardiac injury. Mortality was higher in the group with markers of cardiac injury but was not significant (p = 0.14; Table 3). Death was due to unexpected cardiac arrest in two cases. Both presented with a wide complex ventricular rhythm while receiving high doses of propofol and may have been attributable to propofol infusion syndrome (Iyer et al., 2009).


The major findings in this study of patients with RSE are as follows: (1) markers of cardiac injury are present in nearly two thirds of cases; (2) transient systolic dysfunction may develop in this population and does not necessarily occur at the onset of SE; (3) ventricular arrhythmias are present in a significant minority of patients; and (4) the presence of markers of cardiac injury may portend an increased risk of mortality. These may be due to neurocardiogenic injury or from medications used to stop the seizures, in particular propofol.

Several recent studies have attempted to better define the brain–heart relationship in the setting of seizures (Schneider et al., 2010; Stöllberger et al., 2011; Dupuis et al., 2012), but this is the first study analyzing the frequency and impact of cardiac abnormalities, specifically in a consecutive series of patients with RSE. We observed a wide range of manifestations of cardiac injury including transient systolic dysfunction, QTc prolongation, transient T-wave inversion, ST changes, and troponin elevations.

Cardiac injury in RSE can occur from myocardial stunning due to excessive catecholamine release (stress-induced cardiomyopathy), myocardial ischemia, as well as from complications related to treatment, including heart failure from excess fluid administration or cardiovascular depression from anesthetic agents. Although it is clear that cardiac injury is common in this population, it is difficult to estimate the incidence of adrenergically driven neurocardiogenic injury due to the inability to retrospectively quantify the influence of medications on QTc, anesthetic agents on systolic function, or demand ischemia secondary to the seizures or other systemic complications (such as renal failure or sepsis) on troponin levels.

Although a formal diagnosis of adrenergically driven neurocardiogenic injury cannot be made in our patients, the finding of three cases of reversible systolic dysfunction thought to be consistent with atypical takotsubo cardiomyopathy (TC), is suggestive of this mechanism. In addition, the presence of troponin elevation in 39.1% of patients, QTc prolongation in 22.9%, and T-wave inversion in more than one third of cases indicates that cardiac abnormalities were common—and often reversible—in this cohort of patients with minimal prevalence of documented coronary disease.

The anesthetic agents propofol, pentobarbital, and isoflurane are known to have cardiovascular depressant effects including profound hypotension, and their potential contribution cannot be ignored. However, the systolic dysfunction was noted while patients were on stable doses of these agents, with two of the three remaining on high doses of barbiturate infusion when repeat echocardiogram demonstrated normalization of systolic function. In addition, 9 of the 10 patients with normal systolic function were receiving one or more of these agents at the time of the echocardiographic study.

The fact that systolic dysfunction in the three possible cases of TC was discovered between 4 and 12 days after the onset of SE is of interest. Studies of catecholamine levels in patients with SE demonstrated a quick rise within 30 min of the onset of seizure followed by a decrease over several hours (Simon et al., 1984; Meierkord et al., 1994), with a similar pattern seen in one case of TC associated with convulsive seizures (Shimizu et al., 2008). This would explain why the majority of cases of TC occur within 3 days of a convulsive event (Stöllberger et al., 2011). However, TC has been reported in a patient with partial nonconvulsive SE (Lemke et al., 2008) and the cases of transient systolic dysfunction seen in our series occurred when seizures were controlled based on continuous EEG. The mechanism in these cases is not well understood but may be related to autonomic instability or more prolonged catecholamine release despite electrographic seizure control. Alternatively, it is possible that seizures continued despite apparent resolution by scalp EEG.

Patients with markers of cardiac injury were not more likely to have cardiovascular risk factors or a convulsive seizure presentation than those without markers of cardiac injury. Scant previous studies assessing troponin levels after seizures have shown that increased troponin levels are not generally seen with uncomplicated seizures (Woodruff et al., 2003), but they may occur with more severe seizures presumably causing myocardial injury or demand coronary ischemia (Parvulescu-Codrea et al., 2006; Eskandarian et al., 2011). Troponin levels have not been previously evaluated in patients with RSE. In our series, mortality was higher in the cardiac injury group; however, the difference was not significant. Further research is needed to confirm this association and to determine if manifestations of cardiac injury in RSE represent a marker of severity of the status epilepticus or a factor that directly increases the mortality in these patients.

Our study has various limitations. Cardiac evaluation was highly variable and often incomplete. Therefore, our results may underestimate the true frequency of cardiac abnormalities in RSE. In addition, the serial ECG studies, troponins, and echocardiograms were not uniformly performed, which precluded us from comment on temporal trends. Our cohort was very sick, with nearly 89% of patients requiring anesthetic agents for control of seizures, and thus our findings may not be generalizable to all patients with SE who fail initial therapies.

Due to the complexity of patients with RSE requiring high-dose anesthetic agents, the presence of TC may be easily overlooked. Monitoring for this entity is important as it can predispose patients to acute hypoxemic respiratory failure secondary to pulmonary edema, cardiac arrhythmias, embolism, cardiogenic shock, and sudden death (Bybee et al., 2004; Akashi et al., 2008; Bosca et al., 2008). Development of cardiac thrombus (Wakabayashi et al., 2011) and left ventricular rupture (Stöllberger et al., 2009) has also been reported. The potential for these complications as well as the high frequency of markers of cardiac injury seen in our series supports the practice of systematic screening for such injury.


Markers of cardiac injury in RSE are common and may be underrecognized in this population of severe RSE in which cardiac studies are performed. Routine analysis of ECG and troponin levels followed by an echocardiogram in those with repolarization abnormalities is probably justified. We suggest repeating this evaluation during the course of RSE whenever new or worse hypotension occurs without clear explanation. Future research should study patients with RSE prospectively using a standardized protocol of cardiac evaluation and gathering sufficient information to evaluate the pathophysiology of cardiac changes in these patients.


None of the authors has any conflict of interest to disclose. 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.