Cooper AD, Britton JW, Rabinstein AA. Functional and Cognitive Outcome in Prolonged Refractory Status Epilepticus. Arch Neurol. 2009;66(12):1505-1509. doi:10.1001/archneurol.2009.273
Copyright 2009 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2009
To determine the functional and cognitive outcomes of patients with prolonged refractory status epilepticus (PRSE) lasting 7 or more days despite the use of anesthetic agents for seizure suppression.
St Mary's Hospital, Mayo Clinic, Rochester, Minnesota.
Fourteen patients with PRSE.
Hospital follow-up interview.
Main Outcome Measures
Survival rate of PRSE and functional and cognitive outcome of surviving patients based on the modified Rankin Scale (mRS) and Telephone Interview for Cognitive Status (TICS).
Forty-three percent of patients (6 of 14) died during hospitalization for PRSE, and 57% (8 of 14) had died by the last follow-up. Of the 6 surviving patients, 4 showed improvement and 2 showed no change in mRS score (median mRS change, −1; range, 0 to −3). Owing to preexisting cognitive deficits, 1 patient could not complete the TICS. The 5 remaining patients scored a median of 34 on the TICS (range, 30-37; reference TICS score, ≥31; maximum TICS score, 41). Age, sex, PRSE duration, and etiology were not associated with chance of survival.
Despite the high mortality rate, survival with meaningful functional and cognitive recovery is possible after PRSE. Prolonged duration of status epilepticus alone should not be considered a reason to discontinue treatment.
Status epilepticus (SE) is a life-threatening medical emergency that affects 120 000 to 200 000 people per year in the United States.1 Status epilepticus that persists despite first- and second-line antiepileptic therapy is referred to as refractory status epilepticus (RSE) and occurs in about 31% to 43% of SE cases.2,3 In some cases, RSE may persist for weeks or months, and increasing SE duration has been associated with a worse prognosis.1,4 Little information exists regarding treatment response and prognosis of SE that lasts for several days or weeks.2,5- 7 In cases of prolonged refractory status epilepticus (PRSE), questions often arise about the futility of continued aggressive care and the potential for satisfactory neurologic outcome. In this study, we describe the treatment and functional outcome of PRSE that requires continuous anesthetic suppression for more than 1 week.
We reviewed the medical records of all patients undergoing continuous electroencephalographic monitoring in the neurological intensive care unit at Mayo Clinic from January 1999 through May 2008 to identify patients with PRSE. Status epilepticus was defined as seizure activity lasting more than 5 minutes or recurrent seizure activity lasting more than 5 minutes without regaining consciousness.8 Prolonged refractory status epilepticus was defined as SE in which treatment with anesthetic agents was required for 7 or more days to suppress SE. Duration of PRSE was calculated from the day when SE was first identified to the day when anesthetic drugs were successfully discontinued without recurrence of SE.9
Inclusion criteria required documented continuous seizure activity despite appropriate doses of nonanesthetic antiepileptic medication. If SE recurred within 5 days of anesthetic discontinuation and required reinstitution of anesthetics, the seizure-free days were included in the total duration of PRSE, as suggested by Holtkamp et al.5 Patients with epilepsia partialis continua, SE, or refractory myoclonus related to anoxic brain injury, and patients who did not have continuous electroencephalographic monitoring were excluded. Patients with anoxic brain injury were excluded because of previous studies that demonstrated poor outcome in this subset of patients with SE.10,11 Periodic lateralized epileptiform discharges alone, without clinical correlate, were not considered diagnostic of SE.
A comprehensive review of medical records and all clinical data was performed for patients who met the criteria. The data obtained included history of epilepsy, PRSE etiology, PRSE duration, number and type of antiepileptic drugs given, length of intensive care unit and hospital stays, presence of convulsive vs nonconvulsive seizures, neuroimaging results, and functional outcome. History of epilepsy was defined based on recommendations provided by Fisher et al.12 Encephalitis was defined as encephalopathy (depressed or altered level of consciousness lasting 24 hours or longer, lethargy, or change in personality) and 1 or more of the following symptoms: fever, focal neurologic findings, cerebrospinal fluid pleocytosis, or electroencephalography or neuroimaging findings consistent with encephalitis.13 Encephalitis was classified as infectious, noninfectious, or of unknown etiology.13 A convulsive seizure was defined as any seizure involving focal or generalized motor activity visualized by medical staff or documented in video electroencephalographic reports. Nonconvulsive seizures were defined as electrographic seizures occurring without the presence of motor activity. Neuroimage evaluation included a review of the final magnetic resonance image (computed tomography if no magnetic resonance image was performed) completed during each patient's PRSE episode.
Neurologic functional status prior to disease onset (premorbid status) and at the time of hospital discharge was assessed by medical record review using the modified Rankin scale (mRS).14 Functional status following hospital discharge was assessed using the mRS via a telephone interview and compared with the hospital discharge scores to evaluate for change in functional capacity. For patients presenting with a progressive neurologic disease with onset less than 1 month prior to their episode of PRSE, the premorbid mRS score was calculated by estimating function prior to the onset of their disease process, based on medical record review. If the progressive disease occurred for more than 1 month prior to PRSE, the mRS score 1 month prior to PRSE onset was calculated from the records. The hospital discharge mRS score was calculated on the day of dismissal. If a patient was unable to verbally communicate, the legal guardian was interviewed to determine the follow-up mRS score. In addition to assessing functional status, a structured telephone interview was used to assess cognitive outcome after hospital dismissal using the Telephone Interview for Cognitive Status (TICS).14,15 The time from hospital discharge to time of interview was calculated.
The JMP 7.0.1 software (SAS Institute Inc, Cary, North Carolina) was used for statistical analysis. Continuous data was compared using the Wilcoxon rank sum test, and the Fisher exact test was used for nominal data. Significance was set at P < .05.
This study was reviewed and approved by the Mayo Clinic Investigational Review Board. All patients included in the study, or their legal guardian, gave prior consent to participate in medical research.
Fourteen patients with PRSE were identified. Eight subjects (57%) were men, and the median age was 42 years (range, 18-69 years) (Table 1). Six patients had a history of epilepsy. In 1 patient with chronic posttraumatic epilepsy, the episode of PRSE was caused by acute encephalitis, the specific etiology of which was not determined. Three patients with a history of epilepsy had progressive symptomatic etiologies including 1 patient with a brain tumor, 1 with strokes due to a mitochondrial cytopathy, and 1 with an undefined familial progressive myoclonic form of epilepsy. One patient with epilepsy had a history of ischemic stroke and subarachnoid hemorrhage 9 years prior to her PRSE episode, and another patient with a history of epilepsy had symptomatic Lennox-Gastaut syndrome of unknown cause. The cause of PRSE in the patient with Lennox-Gastaut syndrome was uncertain but may have been medication noncompliance. The etiology of PRSE in the patients without a history of epilepsy included acute ischemic stroke (2 patients), subdural hematoma, thrombotic thrombocytopenic purpura, encephalitis (2 patients), neuropsychiatric systemic lupus erythematosus, and indeterminate cause (1 patient). Prolonged refractory status epilepticus etiologies classified according to the International League Against Epilepsy recommendations are displayed in Table 1.9 One patient developed only nonconvulsive seizures, while all other patients had both convulsive and nonconvulsive seizures during their course of PRSE.
All patients without contraindication received a benzodiazepine and fosphenytoin prior to initiation of anesthetic drugs. Three patients with phenytoin allergy received valproic acid and/or phenobarbital prior to anesthetic initiation. Patients received a median of 2 anesthetic drugs during treatment of PRSE (range, 1-5 drugs). Anesthetic drugs given included propofol, midazolam, pentobarbital, lidocaine, ketamine, and isoflurane, either as a single drug or in combination. In addition to anesthetic drugs, a median of 5 maintenance antiepileptic drugs (range, 4-6 drugs) were initiated per patient either singly or in combination in an effort to abort or maintain control of PRSE. These antiepileptic drugs included clonazepam, diazepam, felbamate, levetiracetam, lorazepam, phenobarbital, phenytoin, topiramate, valproic acid, and zonisamide. A temporal lobectomy was performed in 1 patient with unilateral frontal and temporal encephalomalacia and focal PRSE. This operation was unsuccessful in aborting PRSE, and the patient ultimately died. All but 1 patient required vasopressor use during anesthetic administration (Table 2, patient 9).
Forty-three percent of patients (6 of 14) died during hospitalization for PRSE and 57% (8 of 14 patients) had died by the last follow-up. Of the 6 patients who died during PRSE, 5 died after elective withdrawal of supportive care. Supportive care was withdrawn when prognosis was deemed poor in the presence of continued PRSE or lack of improvement in consciousness despite successful PRSE treatment. The sixth patient died of a cardiac arrhythmia. Two patients survived PRSE, but died after hospital discharge. One of these patients continued to have frequent seizures and died of pneumonia 3 months following hospital discharge. The other died of an unspecified cause 13 months after discharge.
The median duration of PRSE in the entire group was 18 days (range, 7-67 days), and 6 patients had PRSE for more than 2 weeks. The median durations of PRSE in survivors (19.5 days; range, 9-67 days) and nonsurvivors (18 days; range, 7-33 days) was not significantly different (P = .70). Of the 6 patients who survived to the time of interview, the median duration of PRSE was 33.5 days (range, 10-67 days). Survival of PRSE was not affected by age (odds ratio, 1.04; P = .22), sex (P > .99), or acute stroke or encephalitis as the PRSE etiology (P = .54 for acute stroke and P > .99 for encephalitis).
The median length of intensive care unit stay was 21.5 days (range, 7-97 days), and the median length of hospitalization was 31 days (range, 7-125 days). The difference in length of hospitalization for survivors (61 days) and nonsurvivors (19.5 days) was statistically significant (P = .01), as patients who died had shorter lengths of hospital stay.
All patients had medical complications during their hospitalization. These complications included pneumonia (9 patients), deep vein thrombosis (3 patients), pseudomembranous colitis (1 patient), critical illness myopathy (2 patients), critical illness neuropathy (1 patient), urinary tract infection (6 patients), sepsis (2 patients), rhabdomyolysis (1 patient), acute renal failure (1 patient), pseudomonas sinusitis (1 patient), ileus (3 patients), elevated liver enzymes (>3 times the reference value; 2 patients), line infection (1 patient), and Stevens-Johnson syndrome (1 patient) (Table 1).
Six of 8 patients who survived PRSE had worsening of their mRS score at hospital discharge compared with their premorbid score (median mRS change, 4; range, 0-4) (Table 2). The 2 patients without a decline were mentally retarded prior to onset of PRSE and had premorbid baseline mRS scores of 4 and 5. The median time from hospital discharge to telephone interview was 313 days (range, 131-2902 days). Hospital follow-up mRS scores were not available for the 2 patients who died prior to initiation of this study. Of the 6 patients who were interviewed following hospital discharge, 4 showed functional improvement on the mRS compared with their hospital dismissal scores and 2 showed no change (median mRS change, −1; range, 0 to −3). One of the 2 patients with no change in mRS score at last follow-up had a premorbid mRS score of 5 (severe disability) owing to preexisting mental retardation related to Lennox-Gastaut syndrome. This patient could not be evaluated with the TICS owing to preexisting cognitive deficits and inability to speak. The 5 remaining patients scored a median of 34 on the TICS (range, 30-37) (reference TICS score, ≥31; maximum TICS score, 41).
This study confirms that PRSE has high morbidity and mortality rates. All surviving patients in this study had poor functional outcome at hospital discharge based on dismissal mRS scores. However, improvement was seen in most survivors over time, and our results show that meaningful cognitive recovery is possible after successful treatment of PRSE. The study results suggest that functional improvement is possible after resolution of PRSE, and several months may be necessary to realize the full recovery potential of these patients. The range of time from hospital discharge to follow-up was large (131-2902 days) and may have affected the functional and cognitive outcome scores. Although only 2 of our 14 patients were independent at the time of last follow-up, it is possible that if the time to follow-up had been longer for some patients their scores may have improved more than our results show.
Owing to the retrospective study design, we do not have TICS scores at baseline or at hospital discharge to compare with the postdismissal TICS scores. A study by Adachi et al16 compared neuropsychological test scores in patients with epilepsy who had an episode of SE with a control population and did not find any significant change following SE. However, the patients in their study did not suffer from symptomatic etiologies of SE, unlike the patients in our study. In vivo and in vitro studies have demonstrated neuronal death associated with SE, predicting that patients with SE would be at risk of incurring long-term neurological deficits.17- 20 Several other clinical studies have looked at functional outcome in SE and have shown that about 10% of patients who survived SE have disabling neurologic deficits afterward.21- 24 It remains unclear if medical treatment of SE affects outcome or if outcome is determined solely by the etiology underlying SE.10,25,26
The hospital mortality rate of patients in our study was high (43%) when compared with studies on RSE, which show mortality rates ranging from 16% to 23%.2,3,25 The high mortality rate was likely owing to the prolonged nature of SE in our series in contrast to the typical duration in most RSE studies. All patients in this study were mechanically ventilated, which has also been associated with higher mortality rates.27 Duration of SE has been identified by other investigators to be a mortality risk factor.4,28 Although shorter duration of SE was associated with decreased chance of survival in our study, the shorter SE duration in fatal cases was likely owing to the decision to withdraw supportive care in these 5 patients.
Benzodiazepines and fosphenytoin were initially given to all patients who had no contraindication. Following administration of these 2 drugs, multiple other anesthetic agents and antiepileptic drugss were used. Owing to the retrospective nature of this study, we could not ascertain the precise reason for selection of the particular drugs used in each individual case. This finding is consistent with a survey of physicians regarding the treatment of RSE in which there was a lack of agreement on the choice of third- and fourth-line agents.29 At this time, no prospective randomized trials have been performed to compare the efficacy of the different treatment options available for RSE. One retrospective systematic review compared the reported outcomes of 3 anesthetic agents used in RSE but did not find a significant difference in mortality between them.30 Another retrospective study suggested that mortality may be higher with the use of propofol vs midazolam.31 A study of patients with RSE at our institution found complications due to propofol to be unacceptably high, prompting removal of propofol from our RSE treatment protocol.32 Other studies suggest that drug choice does not affect patient outcome.7,25 Algorithms for the treatment of RSE have been proposed.6 A randomized controlled study may be necessary to determine if a superior treatment of RSE exists.
Our results indicate that, although the risks of death and severe neurologic sequelae are high in cases of PRSE, survival with meaningful functional recovery is possible. Thus, prolonged duration (≥1 week) of SE is not a sufficient reason to discontinue therapy. Others have suggested that aggressive care should be continued in prolonged RSE when serial neuroimaging appears normal.33 Further studies on risk factors of poor outcome in PRSE may help clinicians and families decide when it is appropriate to continue aggressive care.
Correspondence: Alex D. Cooper, MD, 1570 N 115 St, Ste 14, Seattle, WA 98133 (email@example.com).
Accepted for Publication: June 22, 2009.
Author Contributions:Study concept and design: Cooper, Britton, and Rabinstein. Acquisition of data: Cooper and Rabinstein. Analysis and interpretation of data: Cooper, Britton, and Rabinstein. Drafting of the manuscript: Cooper, Britton, and Rabinstein. Critical revision of the manuscript for important intellectual content: Cooper, Britton, and Rabinstein. Statistical analysis: Britton and Rabinstein. Administrative, technical, and material support: Cooper, Britton, and Rabinstein. Study supervision: Britton and Rabinstein.
Financial Disclosure: None reported.
Additional Contributions: The authors thank Susanna Stevens, MS, biostatistician, for help with statistical analysis.