Axial fluid-attenuated inversion recovery magnetic resonance imaging scan demonstrates an object map from which ventricular and brain areas are calculated. The plane is just superior to the caudate head.
The diagonal line represents the linear correlation of the duration of anesthetic agent use and the development of brain atrophy. Data points represent the 19 individual patients. See the Methods section for an explanation of how ΔVBR is calculated.
The MRI plane selected is immediately superior to the caudate head. A woman in her 20s presented with super-refratory status epilepticus (SRSE) due to autoimmune encephalitis. A, The initial MRI scan was obtained 6 days after the onset of SRSE. B, The follow-up MRI scan was obtained 128 days later, after treatment with anesthetic agents for 76 days. The follow-up scan shows diffuse atrophy with widened sulci and increased ventricular caliber. The change in vertricular brain ratio (ΔVBR) in this patient was 66.2%. See the Methods section for an explanation of how ΔVBR is calculated.
The MRI planes (left, midbrain; middle, third ventricle; and right, caudate head) were selected to best demonstrate development of atrophy. The images were taken from a man in his 20s with super-refratory status epilepticus. A, The initial MRI scans were obtained on admission. B, The follow-up MRI scans were obtained 102 days later, when the cause of seizures remained cryptogenic despite an exhaustive evaluation. The change in ventricular brain ratio (ΔVBR) in this patient was 124%. The follow-up scans show diffuse atrophy with widened sulci and increased ventricular caliber. Increased FLAIR signal hyperintensity was also noted involving the corticospinal tracts. See the Methods section for an explanation of how ΔVBR is calculated.
Customize your JAMA Network experience by selecting one or more topics from the list below.
Identify all potential conflicts of interest that might be relevant to your comment.
Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.
Err on the side of full disclosure.
If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.
Not all submitted comments are published. Please see our commenting policy for details.
Hocker S, Nagarajan E, Rabinstein AA, Hanson D, Britton JW. Progressive Brain Atrophy in Super-refractory Status Epilepticus. JAMA Neurol. 2016;73(10):1201–1207. doi:10.1001/jamaneurol.2016.1572
Copyright 2016 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.
Does brain atrophy develop in patients with super-refractory status epilepticus (SRSE) despite seizure control with anesthetic agents?
In this medical record review of 19 patients undergoing brain magnetic resonance imaging at prespecified points, measurements of ventricular brain ratio (VBR) were taken at disease onset and follow-up, and change in VBR (ΔVBR) was calculated as a percentage of the starting measure. Median ΔVBR was 23.3%, and a significant correlation between duration of anesthetic agent use and ΔVBR was found.
Atrophy develops during SRSE despite seizure control, and current management of SRSE may contribute.
Prolonged seizures in super-refractory status epilepticus (SRSE) have been shown to cause neuronal death and reorganization, and visual inspection in individual case studies has demonstrated progressive cortical and subcortical atrophy. At present, magnetic resonance imaging (MRI) studies that evaluate brain atrophy in SRSE are lacking.
To document and quantify the development of atrophy over time in SRSE.
Design, Setting, and Participants
This retrospective medical record review included all patients with SRSE who were admitted to a tertiary referral campus of the Mayo Clinic Hospital with SRSE from January 1, 2001, to December 31, 2013. Patients with (1) an initial MRI scan performed within 2 weeks of SRSE onset, (2) a second MRI scan within 6 months of SRSE resolution, and (3) a minimum duration of 1 week between MRI scans were included. The ventricular brain ratio (VBR) was measured on T2-weighted fluid-attenuated inversion recovery (FLAIR) images at disease onset and during follow-up. Measurements were performed on axial FLAIR images with section thickness of less than 5 mm. The plane immediately superior to the caudate head was chosen for analysis. The hypothesis that atrophy develops during SRSE despite seizure control (electroencephalogram background suppression with anesthetic drugs) was tested. Data were analyzed from June 1 to December 31, 2015.
Main Outcomes and Measures
Change in VBR (ΔVBR) as a percentage of the starting measure.
Nineteen patients met the inclusion criteria; these included 10 men (53%) and 9 women (47%) with a median age of 41 (interquartile range [IQR], 25-68) years. Anesthetic agents were required for a median of 13 (IQR, 5-37) days. Initial MRI was performed a median of 2 (IQR, 1-7.5) days from the onset of SRSE, and the second MRI was performed a median of 11 (IQR, 5-15.5) days from the resolution of SRSE, with a median of 40 (IQR, 15-65) days between MRI scans. Median ΔVBR was 23.3% (IQR, 10.5%-70.3%). A significant correlation between the duration of anesthetic agent use and ΔVBR was found (Spearman r = 0.54; P = .02).
Conclusions and Relevance
Atrophy developed in all patients with SRSE who underwent serial imaging, despite administration of agents for seizure control. The degree of atrophy appears to be related to the duration of SRSE.
Quiz Ref IDSuper-refractory status epilepticus (SRSE) is defined as SE that continues or recurs 24 hours or more after the onset of anesthetic therapy.1 Mortality ranges from 30% to 50%, and half of survivors have a poor functional outcome despite seizure control.2,3
Quiz Ref IDThe pathophysiologic features of neuronal damage in patients with SE have been well documented in human and animal models.4-10 Increased excitation and release of excitatory neurotransmitters such as glutamate result in calcium influx, leading to activation of intracellular cascades and osmotic swelling, each contributing to neuronal loss.5-8,10 Cytotoxic and vasogenic edema, hyperperfusion of the epileptic region, and alteration of the leptomeningeal and blood-brain barriers during seizures have all been demonstrated on neuroimaging.11 In the hours, days, and weeks after prolonged seizures, long-term changes in gene expression and chronic microhemorrhages due to an altered blood-brain barrier result in seizure-induced neuronal death and neuronal reorganization.5 These findings have been confirmed by histopathologic studies in patients with refractory seizures,5-7,12 which showed the development of generalized and focal atrophy. Neuron-specific enolase levels have been shown to be highly elevated after SE, suggesting neuronal damage and associated parenchymal loss.13 Individual case studies14,15 in patients with prolonged SRSE have also demonstrated progressive cortical and subcortical atrophy by visual inspection.
Brain magnetic resonance imaging (MRI) has been used as a tool for elucidation of the etiology of SE and less commonly to evaluate the development of hippocampal signal changes or atrophy.4,11,16-20 Other MRI abnormalities reported in SE include diffusion-weighted or T2-weighted cortical hyperintensities with a corresponding low apparent diffusion coefficient that do not correspond to vascular territories.4,11,17,21 These cortical changes have been shown to be reversible on follow-up imaging, suggesting the presence of cytotoxic and vasogenic edema, hyperperfusion, and alteration of the blood-brain-barrier. Other investigators22,23 have suggested that permanent structural changes may occur. Although multiple case studies of patients with RSE who required prolonged anesthesia have demonstrated the development of regional brain atrophy,14,15 none have systematically examined MRI abnormalities in RSE or SRSE; thus, whether development of atrophy is common or to what degree it occurs remains unknown.
Ventricular brain ratio (VBR) is a frequently used measure of cerebral atrophy calculated as the area of the lateral ventricles divided by the brain area in a given plane. The VBR has been used previously to examine the development of brain atrophy in neurologic and psychiatric patient populations.24-27 Measuring the degree of parenchymal loss using VBR has been found to be a predictive factor for the development of cognitive dysfunction in schizophrenia and for transition from mild cognitive impairment to dementia.24,25 Generalized brain atrophy has also been documented using VBR in patients with epilepsy.27 In a study of 40 patients with acute encephalitis in which VBR was used to document parenchymal volume loss, SE was associated with greater loss of brain parenchyma.26
In this study, we aimed to document and quantify the development of diffuse brain atrophy in consecutive patients with SRSE who underwent repeated brain MRI at specified points. A secondary objective was to define clinical variables correlated with a greater change in brain volume. Documenting the development of atrophy in SRSE would be an important step in furthering our understanding of the balance between treatment aggressiveness and seizure control.
This retrospective medical record review included all patients who were admitted with SRSE to Mayo Clinic Hospital, St Marys Campus, Rochester, Minnesota, from January 1, 2001, to December 30, 2013. Patients were identified through queries of our computerized electroencephalography (EEG) reporting system and the electronic medical record. This study was reviewed and approved by institutional review board of the Mayo Clinic, which did not require informed consent for this retrospective review with minimal risk.
Patients 18 years or older with (1) SRSE, (2) an initial MRI scan obtained within 2 weeks of onset of SRSE, (3) a second MRI scan obtained within 6 months of the resolution of SRSE, and (4) a minimum of 1 week between scans were included. If more than 1 scan met criteria, the one closest to the onset of SRSE or resolution of SRSE was selected. Exclusion criteria consisted of an anoxic-ischemic cause, epilepsia partialis continua, and absence of SE. All patients underwent continuous EEG monitoring from the time of diagnosis of SE to the time of resolution of SE, transition to palliative care, or patient death. We defined SRSE as seizures (confirmed by continuous EEG) that continued for at least 24 hours after initiation of general anesthetic therapy, including cases in which the SE recurred with reduction or withdrawal of anesthetic therapy.2 Patients with nonconvulsive seizures were included if they had a clear acute change in the level of consciousness, followed by witnessed clinical seizures along with epileptiform activity on continuous EEG.
The VBR was calculated by dividing the area of the lateral ventricles by the total area of the brain.26,27 The change in VBR (ΔVBR) is calculated by subtracting the VBR at disease onset from the VBR at follow-up and dividing the difference by the VBR at disease onset. The ΔVBR represents the change in VBR at follow-up compared with the acute disease stage as a percentage of the starting value; a higher value signifies greater parenchymal loss.26
All patients underwent initial and follow-up brain MRI at the Mayo Clinic. The measurements were performed on axial fluid-attenuated inversion recovery (FLAIR) images with a section thickness of less than 5 mm. FLAIR images were used because they were of higher quality than the corresponding T1-weighted images. The FLAIR images were imported into Analyze software (version 12.0) developed by the Mayo Clinic Biomedical Imaging Resource Core.28 To obtain corresponding image planes in the initial and follow-up MRI volumes, we chose the plane immediately superior to the caudate head as shown in Figure 1, with an orientation as close as possible to the same angle and height in both scans, taking into account the different position of the head. The areas of the lateral ventricles and of the whole cortical brain structure were measured by a semiautomated region-growing technique using an initial seed point and threshold range to guide the automated boundary detection for the regions of interest in the Analyze software. Using previously developed methods, we defined the margin of the brain region at the inner outline of the subarachnoid space and traced wide sulci, including the minimum possible volume of subarachnoid space.26,29 Measurements were performed separately by two of us (S.H. and E.N.). The mean of the 2 measurements were used for calculating VBR. Plane orientation was calibrated to prevent error using methods previously described.26 A ΔVBR greater than 50% was considered severe atrophy.
Additional collected variables included age, sex, cause of SRSE, duration of anesthetic agent use in days, type and number of anesthetic agents used, hospital length of stay, and the modified Rankin Scale (MRS) score at disease onset, hospital discharge, and last available follow-up.30 Duration of anesthetic agent use was recorded as a surrogate for the duration of SRSE. Modified Rankin Scale scores of 0 to 2 were considered a good outcome, and scores of 3 to 6 were considered a poor outcome.
Data were analyzed from June 1 to December 31, 2015. Statistical analysis was performed using SPSS software (version 16.0; SPSS, Inc). We summarized qualitative variables as frequencies and percentages and checked quantitative variables for normality. We summarized the normally distributed quantitative variables as means and SDs and the nonnormally distributed variables as median (interquartile range [IQR]). We used the Wilcoxon signed rank test to compare paired measurements of ΔVBR and MRS scores at different points. The Spearman rank correlation coefficient was used to test for correlation of different variables with ΔVBR. Results of all statistical tests were interpreted for significance at a cutoff of .05.
Forty-two patients 18 years or older had SRSE during the study period, among whom 19 met our inclusion criteria. Lack of MRI during the prespecified points was the reason for exclusion of the remaining 23 patients. Quiz Ref IDThe median patient age was 41 (IQR, 25-68) years; 10 patients (53%) were men and 9 (47%) were women. Causes of SRSE are listed in the Table. Anesthetic agents were required for a median of 13 (IQR, 5-37) days. Patients received a median of 2.4 (IQR, 2-4) different anesthetic drugs. Five patients received only 1 anesthetic agent, whereas 14 received 2 or more anesthetic drugs. Midazolam hydrochloride was used by 17 of 19 patients (89%); propofol, by 9 (47%); pentobarbital sodium, by 8 (42%); ketamine hydrochloride, by 5 (26%); phenobarbital, by 2 (11%); and isoflurane, by 1 (5%). Hospital length of stay was a median of 40 (IQR, 18-80) days. The initial MRI was performed a median of 2 (IQR, 1-7.5) days after the onset of SRSE, and the follow-up MRI was performed a median of 11 (IQR, 5-15.5) days from the resolution of SRSE, with a median of 40 (IQR, 15-65) days between scans. Median VBR of the initial MRI scan was 0.06 (IQR, 0.04-0.07) and of the follow-up MRI scan was 0.08 (IQR, 0.07-0.12). The VBR of the initial MRI scan was significantly different from the VBR of the follow-up MRI scan (z = −3.82; P < .001). Median ΔVBR was 23.3% (IQR, 10.5%-70.3%).
We found significant positive correlations between ΔVBR and the duration of anesthetic agent use (ρ = 0.54; P = .02) (Figure 2) and the duration of hospital stay (ρ = 0.64; P = .003). We found a significant negative correlation between ΔVBR and age (ρ = −0.49; P = .04). Mean VBR of the initial MRI scan was significantly higher in patients 50 years or older when compared with the younger patients (ρ = 0.74; P < .001).
The median MRS score was 1 (IQR, 0-4) before admission, 5 (IQR, 1-6) at hospital discharge, and 6 (IQR, 3-6) at the last available follow-up. Overall the median MRS score was 4 (IQR 3-6). The MRS score increased significantly from admission to discharge (z = −3.84; P < .001). No significant correlation was observed between the ΔVBR and the MRS score taken at admission (ρ = 0.13; P = .59), discharge (ρ = 0.13; P = .61), and follow-up (ρ = 0.13; P = .66) and did not change after excluding patients who underwent transition to palliative care. In patients with a ΔVBR greater than 50%, the MRS scores at last available follow-up were 1 (n = 1), 3 (n = 1), 4, (n = 1), 5 (n = 1), and 6 (n = 2). Among the 19 patients, 6 (32%) underwent transition to palliative care during their hospital stay. At the 1-year follow-up, 8 patients had died, 5 were lost to or unavailable for follow-up, and 6 were alive. The MRS scores among survivors were 1 (n = 1), 2 (n = 2), 3 (n = 1), and 4 (n = 2). All patients who were lost to or unavailable for follow-up at 1 year were alive at 3 to 6 months, with MRS scores of 3 (n = 1), 4 (n = 3), and 5 (n = 1).
Examples of 2 individual patients are shown to illustrate the atrophy (Figure 3 and Figure 4). Figure 3 shows the initial and follow-up MRI of 1 patient using the plane selected for analysis just superior to the caudate head. Figure 4 shows 3 planes chosen to better illustrate the diffuse cerebral atrophy and compares initial and follow-up MRIs.
Quiz Ref IDIn our cohort of patients with very prolonged SRSE, measurable brain atrophy developed in all patients; however, we found a wide variation in the degree of atrophy between patients. Atrophy was more likely to develop in patients with a prolonged hospitalization and longer duration of anesthetic treatment, but a positive ΔVBR was less likely to develop in elderly patients. Surprisingly, the development of atrophy did not correlate with functional outcome.
Multiple studies have demonstrated the development of focal atrophy in the hippocampi, subcortical white mater, basal ganglia, corpus callosum, and cerebellum after SE.4,11,17-23 Our study is, to our knowledge, the first to document the development of diffuse brain atrophy despite control of SRSE using quantitative imaging. This finding suggests that parenchymal loss is ongoing, even after control of seizures with anesthetic drugs. The correlation between duration of anesthetic drug requirement and degree of global brain atrophy further strengthens the validity of the association. The reason for the progressive brain atrophy is not known but may be owing to reversible cerebral edema, ongoing subclinical seizures that are not detected by scalp EEG, direct effects of anesthetic medication therapy,31,32 or disuse of therapy.
In our practice we use brain MRI for evaluation of the cause or to assess the severity of the cause of SRSE and repeat the MRI as clinically indicated. In very prolonged cases of SRSE, the development of prominent brain atrophy has raised concerns about the ultimate outcome; however, this concern is only theoretical and may not be justified. In fact, 1 study demonstrated “reversible” brain atrophy after prolonged SE in 2 patients with N-methyl-d-aspartate receptor encephalitis, although the degree of atrophy in those cases was not quantified.33 Our results showing that the development of atrophy was not associated with functional outcome also suggest that atrophy per se may not determine poor recovery and, therefore, should not be used as the sole argument to withdraw life-sustaining measures.
We found that the duration of anesthetic therapy and hospital stay are major predictive factors for brain volume loss and long-term clinical outcome. We used duration of anesthetic therapy as a surrogate for duration of SRSE. However, we cannot discount a potential direct toxic effect of high-dose anesthetic agents on neurons or the possibility of neuronal loss from the underlying disuse. Three recent retrospective studies in the United States and Switzerland34-36 have shown an independent association between anesthetic drugs and worse outcome after correction for all known confounders excepting refractoriness. Anesthetic agents have been associated with neuronal degeneration in preclinical studies in animals during the neonatal period and with cognitive impairment that persisted into adulthood.37 Proof is lacking in humans that prolonged exposure to anesthesia causes neuronal degeneration; however, various mechanisms for neuronal degeneration with anesthesia have been postulated.
Suppression of synaptic signaling may lead to apoptosis in postsynaptic neurons via intrinsic and extrinsic apoptotic pathways. The intrinsic pathway involves activation of mitochondrial-dependent pathways with a release of cytochrome c and activation of proapoptotic factors, causing DNA cleavage and cell death. The extrinsic pathway is usually triggered by external stimuli such as cytokines and tumor necrosis factor. Another possibility is that alteration of cell signaling prevents the processing of the neurotrophin brain-derived neurotrophic factor, which is responsible for cell growth. Alteration of this signaling molecule leads to cell death. Experimental studies have shown that prior hypoxic neuronal injury increases susceptibility to further apoptosis when exposed to anesthetic agents and that the rate of apoptosis increases with prolonged or repeated exposure to volatile and intravenous anesthetic agents.37,38 Patients with SE are susceptible to hypoxemia owing to a number of different mechanisms, including mucous plugging, atelectasis, pulmonary edema, apnea, and hypoventilation.
Age was negatively correlated with the development of atrophy in our study. This correlation is intuitive because older patients are more likely to have atrophy at baseline and younger patients have more brain to lose. The higher VBR on the initial MRI scan in older patients further substantiates this hypothesis.
No correlation was observed between the degree of development of atrophy and functional outcome at different time points. In fact, one-third of patients who developed severe brain atrophy recovered to a good functional outcome. Although this outcome is somewhat reassuring, the small number of patients and inability to control for other variables does not discount the possible influence of the development of atrophy on long-term functional outcomes. In addition, the MRS score does not measure cognitive outcomes directly.
The strengths of this study are its relevance to current debates about treatment aggressiveness in SRSE and the use of standardized measurements to quantify the degree of atrophy. We also required that 2 measurements were performed blindly by different investigators and our measurements were very similar, providing validity to the measurement tool. Our study also has limitations. The study was insufficiently powered to determine whether differences in degree of atrophy were based on the cause, the presence of complications during SE such as hypotension, or the selection of anticonvulsant. We also could not assess the effect of the different types and doses of anesthetic agents on atrophy development. Some selection bias may have occurred because only 45% of patients with SRSE during the study period underwent brain MRI during the prespecified points. In addition, because this study was exploratory only, we did not use quantitative volumetric analysis.
We were able to systematically demonstrate the development of atrophy in a consecutive series of patients with SRSE using already developed methods of quantification. Although we chose 1 cutoff to standardize the measurements, we saw the atrophy diffusely in all our patients. Quiz Ref IDDevelopment of atrophy was associated with the duration of anesthetic use and hospital stay but not with functional outcomes. Future studies should focus on which areas are most affected, assess the association of atrophy with relevant clinical variables, and importantly, examine the influence of atrophy development on long-term cognitive function in survivors of SRSE.
Corresponding Author: Sara Hocker, MD, Department of Neurology, Mayo Clinic, Mayo Bldg W8-B, 200 First St SW, Rochester, MN 55905 (firstname.lastname@example.org).
Accepted for Publication: April 8, 2016.
Published Online: August 15, 2016. doi:10.1001/jamaneurol.2016.1572
Author Contributions: Dr Hocker had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Hocker, Rabinstein, Hanson, Britton.
Acquisition, analysis, or interpretation of data: Hocker, Nagarajan.
Drafting of the manuscript: Hocker, Nagarajan.
Critical revision of the manuscript for important intellectual content: Hocker, Rabinstein, Hanson, Britton.
Statistical analysis: Hocker, Nagarajan.
Administrative, technical, or material support: Hocker, Hanson, Britton.
Study supervision: Hocker.
Conflict of Interest Disclosures: Dr Hocker reports serving as a coinvestigator and consultant for a clinical trial involving allopregnanolone for super-refractory status epilepticus sponsored by SAGE Therapeutics, outside the submitted work. Mr Hanson reports receiving personal fees from AnalyzeDirect, outside the submitted work. Dr Britton reports serving as a coinvestigator for clinical trials involving use of cannabidiol in Lennox-Gastaut and Dravet syndrome sponsored by GW Pharma, outside the submitted work. No other disclosures were reported.
Create a personal account or sign in to: