Journal of Neurology https://doi.org/10.1007/s00415-022-11005-1 ORIGINAL COMMUNICATION Epileptic seizures in the emergency room: clinical and electroencephalographic findings associated with brain perfusion patterns on computed tomography JL Restrepo‑Vera1 · P. Coscojuela2 · E. Fonseca3 · M. Quintana3 · S. Sarria‑Estrada2 · E. Santamarina3 · L. Abraira3 · M. Sueiras4 · V. Thonon4 · J. Álvarez‑Sabin1 · M. Toledo3 · A. Rovira2 Received: 17 November 2021 / Revised: 29 January 2022 / Accepted: 31 January 2022 © The Author(s), under exclusive licence to Springer-Verlag GmbH Germany 2022 Abstract Background Diagnosis of epileptic seizures, particularly regarding status epilepticus (SE), may be challenging in an emer- gency room setting. The aim of the study was to study the diagnostic yield of perfusion computed tomography (pCT) in patients with single epileptic seizures and SE. Methods We retrospectively reviewed the records of patients who followed an acute ischemic stroke pathway during a 9-month period and who were finally diagnosed with a single epileptic seizure or SE. Perfusion maps were visually ana- lyzed for the presence of hyperperfusion and hypoperfusion. Clinical data, EEG patterns, and neuroimaging findings were compared. Results We included 47 patients: 20 (42.5%) with SE and 27 (57.5%) with single epileptic seizure. Of 18 patients who showed hyperperfusion on pCT, 12 were ultimately diagnosed with SE and eight had EEG findings compatible with an SE pattern. Focal hyperperfusion on pCT had a sensitivity of 60% (95% CI 36.4–80.2) and a specificity of 77.8% (95% CI 57.2–90.6) for predicting a final diagnosis of SE. The presence of cerebral cortical and thalamic hyperperfusion had a high specificity for predicting SE presence. Of note, 96% of patients without hyperperfusion on pCT did not show an SE pattern on early EEG. Conclusions In acute settings, detection by visual analysis of focal cerebral cortical hyperperfusion on pCT in patients with epileptic seizures, especially if accompanied by the highly specific feature of thalamic hyperperfusion, is suggestive of a diagnosis of SE and requires clinical and EEG confirmation. The absence of focal hyperperfusion makes a diagnosis of SE unlikely. Keywords Computed tomography · Epilepsy · Neuroimaging · Perfusion · Seizures · Status epilepticus · Thalamus Introduction * E. Fonseca e.fonseca@vhebron.net Epileptic seizures are a common reason for emergency room 1 (ER) visits [1]. Status epilepticus (SE), the most severe form Neurology Department, Vall d’Hebron University Hospital, Medicine Department, Universitat Autònoma de Barcelona, of epilepsy, is a life-threatening condition [2]. Its incidence Passeig Vall d’Hebron 119‑129, 08035 Barcelona, Spain is rising and is currently estimated at approximately 10–20 2 Neuroradiology Section, Radiology Department, Vall cases per 100,000 person-years [3]. Diagnosis of SE is chal- d’Hebron University Hospital, Vall d’Hebron Research lenging and requires clinical experience and paraclinical Institute, Passeig Vall d’Hebron 119‑129, 08035 Barcelona, diagnostic tools. Stroke and post-ictal epileptic phenomena Spain often mimic SE without prominent motor symptoms (non- 3 Epilepsy Unit, Neurology Department, Vall d’Hebron convulsive SE [NCSE]) [4]. In such cases, an electroenceph- University Hospital, Medicine Department, Universitat alogram (EEG) is necessary to confirm diagnosis, but even Autònoma de Barcelona, Passeig Vall d’Hebron 119‑129, 08035 Barcelona, Spain at tertiary hospitals that deal with neurological emergencies, 4 access to EEG can be delayed by 4 h or longer [5]. Seizure Neurophysiology Department, Vall d’Hebron University Hospital, Universitat Autònoma de Barcelona, Passeig Vall duration is one of the main modifiable prognostic factors in d’Hebron 119‑129, 08035 Barcelona, Spain 13 Vol.:(0123456789) Journal of Neurology SE and reductions in this time are closely correlated with a rotation time of 0.28 s, an image matrix of 512 × 512 to better outcomes [6, 7]. the convexity, and collimation of 32 × 1.2 mm. In total, 700 Brain perfusion computed tomography (pCT) is a widely images were obtained per patient and processed at a work- used neuroradiological imaging modality in ERs [8], and station to analyze cerebral blood volume (CBV), regional previous studies have shown its diagnostic value in patients cerebral blood flow volume (CBF), time to peak (TTP), and with acute neurological deficits of epileptic origin [9–11]. mean transit time (MTT). The images were reassessed by an Ictal abnormalities seen on pCT have several explana- experienced neuroradiologist (P.C.) blinded to the clinical tions, including upregulation of the neurovascular unit and data including final diagnosis and the findings of EEG. A blood–brain barrier disruption mediated by the hypergluta- qualitative visual analysis was carried out to identify areas matergic state of the brain during SE [12, 13]. There is thus showing perfusion differences in the various hemodynamic growing interest in pCT as a potential tool for the rapid iden- maps, focusing on cerebral cortex and thalamus. For the tification of epileptic seizures and SE in particular. The aim visual inspection of the raw pCT images, hyperperfusion of this study was to assess the diagnostic yield of emergency was defined as an increase in CBF and a decrease in TTP pCT in single epileptic seizures and SE in an ER setting. and MTT in cerebral cortex and/or thalamus. Hypoperfusion was defined as focally decreased CBF and increased TTP and MTT relative to the contralateral brain hemisphere in Methods all cases. Emergent EEG lasting ≥ 45 min was performed within This was a single-center retrospective study of patients 24 h of symptom onset. All EEGs were recorded on a 32- aged ≥ 18 years who underwent pCT following a non- and/or 64-channel digital EEG system (XLTEK, Deltamed). contrast CT and CT angiography of the brain for suspected An array of 21 electrodes was used beyond the 10–20 sys- acute stroke within 24 h of symptom onset in the ER and tem. The combination of referential (average) and bipo- who were subsequently diagnosed with a single epileptic lar EEG derivations with other signals (ECG and EMG) seizure or SE following additional tests. The study was allowed for the acquisition of polygraphic EEG recordings. conducted in a tertiary hospital that offers emergency EEG These were evaluated by an expert neurophysiologist (M.S. and pCT and was approved by the local ethics committee or V.T.), blinded to the pCT images. The EEG findings were (PR(AG)285/201). Data from all patients who followed the described using the standardized international nomenclature hospital’s acute stroke care pathway at any time between of the International Federation of Clinical Neurophysiology November 2019 and July 2020 were systematically obtained and the American Clinical Neurophysiology Society’s stand- from the hospital’s electronic database. Those with unclear ardized terminology for neurocritical care [15]. The Salz- seizures, a new ischemic or hemorrhagic lesion on brain burg criteria were used for the diagnosis of SE patterns on CT or magnetic resonance imaging (MRI), or cerebral arte- EEG [16]. Specific EEG features were categorized as non- rial occlusion or stenosis on angiographic sequences were epileptiform abnormalities (NEA), interictal epileptiform excluded. All patients were assessed by two neurologists on discharges (IED), single epileptic seizures (absence of con- arrival at the ER. Epileptic seizures were diagnosed based tinuous or fluctuating patterns suggestive of SE), and EEG on clinical features and EEG findings where available. SE patterns consistent with SE. was diagnosed based on clinical semiology and the EEG Descriptive and frequency statistics were obtained, and criteria from the 2015 Report of the International League comparisons made in IBM SPSS Statistics, version 22.0. Against Epilepsy Task Force on the Classification of Status Comparisons between the three perfusion pattern groups Epilepticus [14]. Hyperperfusion on pCT was not taken into (hyperperfusion, hypoperfusion, normal perfusion) were account to establish the diagnosis of SE. Clinical, EEG, and made using the Fisher exact test for categorical variables imaging data were retrospectively collected from the hos- and the Kruskal–Wallis test for numerical variables. Signifi- pital’s electronic database. A record was also made of all cance was set at p < 0.05. Sensitivity, specificity, and posi- intravenous anti-seizure drugs (ASDs) administered at any tive and negative predictive values were calculated to assess time between symptom onset and ancillary tests. the diagnostic value of pCT relative to final diagnosis and All patients underwent non-contrast CT, CT angiography, EEG findings. and pCT on a 128-slice scanner (Somatom Definition, Sie- mens Healthcare). The pCT study included dynamic acquisi- tion (70 kV and 150 mAs) with injection of 50 mL of non- Results ionic iodinated contrast material (Iohexol) at a concentration of 350 mg/mL (Omnipaque 350) using an injection pump to Of the 697 patients assessed during the study period, 47 obtain a constant flow of 5 mL/s. The acquisition conditions (6.74%) met the inclusion criteria. The selection pro- in all the imaging modalities were a slice thickness of 5 mm, cess and results of the imaging tests are shown in Fig. 1. 13 Journal of Neurology (57.5%) of the sample received a final diagnosis of a single epileptic seizure. No seizure clusters were observed. The clinical characteristics of the patients and EEG/pCT find- ings are summarized in Table 1. Brain pCT was performed a median of 3 h (interquartile range [IQR], 2–12 h) after the last time seen asympto- matic. The scan was abnormal in 24 patients (51%): focal hyperfusion (cerebral cortical hyperperfusion accom- panied or not by thalamic hyperperfusion) in 18 cases (38.3%) and focal hypoperfusion in six (12.7%). Thalamic hyperperfusion was identified in 8 (17%) of 47 patients, all of them with concurrent homolateral cortical hyperperfu- sion. No thalamic hypoperfusion was detected. Of the 18 patients with hyperperfusion, 12 (66.7%) were ultimately diagnosed with SE and six (33.3%) with a single epilep- tic seizure (Fig. 2a). All the patients in the hypoperfu- sion group were diagnosed with a single epileptic seizure. Three (50%) of them did not undergo brain MRI but pre- vious diagnosis of epilepsy in all of them, no evidence of acute lesion on CT and transient deficits were consid- ered sufficient evidence to reasonably rule out a transient ischemic stroke. The frequency of pCT alterations was not significantly influenced by time from symptom onset to pCT, seizure etiology, or intravenous administration of ASDs before neuroimaging studies (Table 1). In most cases, hyperperfusion was exclusively observed in the cortical gray matter, accompanied or not of homolateral thalamic hyperperfusion, with relative preservation of the underlying white matter (Fig. 3). Most cases of focal hypoperfusion involved both the cortical and subcortical regions. EEG was performed in 41 patients (87.2%). Median time to EEG was 16 h (IQR, 8.5–22.5 h) from last time seen asymptomatic and 8 h (IQR: 4–14.5) from pCT. EEG showed epileptiform abnormalities in 22/41 patients Fig. 1  Flow chart of patient inclusion/exclusion process and pCT/ EEG results. pCT, perfusion computed tomography; AIS, acute (53.7%): IEDs in 12, EEG patterns consistent with SE ischemic stroke; EEG, electroencephalogram; NEA, non-epileptic in nine, and a single epileptic seizure in one. Eight of abnormalities; IED, interictal epileptiform discharges; SES, single the nine patients with EEG findings consistent with SE epileptic seizure; SE, status epilepticus showed hyperfusion on the admission pCT. Patients with other EEG abnormalities were more likely to show normal The mean age of the patients was 72.8 ± 16.1 years and perfusion (Fig. 2a). 26 (55.3%) were women. Twenty-five patients (53.2%) Detection of thalamic hyperperfusion accompanying cor- underwent a brain MRI, while in the remaining 22 tical cerebral hyperperfusion was significantly associated to (46.8%), no follow-up imaging was considered necessary SE as final diagnosis and to an SE pattern on urgent EEG to clarify the diagnosis by the treating physician. Twenty (Table 2). patients (42.5%) received a final diagnosis of SE: nine Hyperperfusion on pCT had a sensitivity of 60% (95% (45%) patients suffered from NCSE with or without coma CI, 36.4–80.2) for the detection of SE as final diagnosis. along with EEG patterns consistent with SE and 11 (55%) Thalamic hyperperfusion on pCT was identified exclu- patients were clinically diagnosed with SE with prominent sively in patients with final diagnosis of SE. The rest of the motor symptoms; 10 of these patients had a non-ictal pat- corresponding values for predicting final diagnosis or the tern on EEG and one died shortly after CT before EEG presence of EEG patterns consistent with SE are shown in could be performed due to concurrent fatal pneumonia Table 3. Of note, 96% of patients without hyperperfusion on from SARS-CoV-2 infection. The remaining 27 patients pCT did not have an SE pattern on EEG. 13 Journal of Neurology Table 1  Demographic, clinical, and EEG features according to pCT findings Demographic, clinical, and EEG findings pCT findings Total (n = 47) Hypoperfusion Normal Hyperperfusion P value (n = 6, 12.7%) (n = 23; 48.9%) (n = 18; 38.3%) Age; years (mean ± SD) 72.8 (16.1) 72 (15.1) 78.4 (15.2) 65.5 (15.3) 0.046* Sex Female; n (%) 26 (55.3) 6 (100) 13 (56.5) 7 (38.9) 0.027 Male; n (%) 21 (44.7) 0 (0) 10 (44.5) 11 (61.1) History of epilepsy Yes; n (%) 11 (23.4) 3 (50) 2 (8.7) 6 (33.3) 0.047 No; n (%) 36 (76.6) 3 (50) 21 (91.3) 12 (66.6) Time from onset to pCT; hours (median, IQR) 3 (2–12) 7.5 (1–14.7) 3 (2–13) 4 (1.2–7.5) 0.9 ­ EG*; hours (median, IQR) Time from onset to E 16 (8.5–22.5) 15 (10.7–19.2) 16 (7.5–23.5) 15 (9.7–21.7) 0.99 Diagnosis SS; n (%) 27 (57.4) 6 (100) 15 (65.2) 6 (33.3) 0.01 SE; n (%) 20 (42.6) 0 (0) 8 (34.8) 12 (66.7) Seizure etiology Metabolic; n (%) 10 (21.3%) 1 (16.7) 3 (13) 6 (33.3) 0.607 Vascular; n (%) 10 (21.3%) 1 (16.7) 4 (17.4) 5 (27.8) Tumoral; n (%) 9 (19.1) 1(16.7) 5 (21.7) 3 (16.7) ASD withdrawal; n (%) 3 (6.4) 1 (16.7) 1 (4.3) 1 (5.6) PRES; n (%) 3 (6.4) 1 (16.7) 2 (8.7)) 0 (0) Unknown; n (%) 12 (25.5) 1 (16.7) 8 (34.8%) 3 (16.7) ASDs before pCT Yes; n (%) 15 (31.9) 3 (50) 9 (39.1) 3 (16.7) 0.194 No; n (%) 32 (68.1) 3 (50) 14 (60.9) 15 (83.3) EEG‡ NEA; n (%) 19 (46.3) 3 (75) 13 (61.9) 3 (18.8) 0.008 IED; n (%) 12 (29.3) 1 (25) 6 (28.6) 5 (31.3) SES; n (%) 1 (2.4) 0 (0) 1 (4.8) 0 (0) SE; n (%) 9 (22) 0 (0) 1 (4.8) 8 (50) EEG electroencephalography, pCT perfusion computed tomography, SD standard deviation, IQR interquartile range, SS singles seizure, SE status epilepticus, ASDs anti-seizure drugs, PRES posterior reversible encephalopathy syndrome, NEA non-epileptiform abnormalities, IED interictal epileptiform discharges, SES single epileptic seizure *Bold typed p values represent statistically significant differences between groups. ‡: 41 patients underwent EEG registry. Percentages are calculated according to these numbers. Discussion detection of SE as a final diagnosis; this diagnostic yield is similar to that described in retrospective studies with a In the present study of patients presenting with an epileptic comparable median time from pCT to EEG (e.g., 8 h or seizure at the ER of a tertiary hospital, focal hyperperfusion more) [10, 11, 18]. Simultaneous acquisitions of EEG and on early pCT was a useful marker of SE but not single epi- pCT that could not only study the perfusion signature of SE leptic seizures. Nearly two-thirds of the patients ultimately but also characterize the observed brain perfusion changes diagnosed with SE showed focal hyperperfusion on pCT, depending on the EEG pattern, frequency and/or morphol- while those diagnosed with a single epileptic seizure mostly ogy of the observed ictal discharges or rhythmic activity showed normal brain perfusion. These findings are consist- could provide greater insights into why over one-third of ent with previous reports of brain hemodynamics during SE the patients diagnosed with SE showed no abnormalities on and post-ictal states in similar settings [17]. pCT on our study. The authors of a retrospective study of However, a precise relationship between diagnosis of SE seven critically ill patients who underwent fluorodeoxyglu- and pCT pattern was not observed in our study. Hyperper- cose–positron emission tomography during EEG monitoring fusion had a 60% sensitivity and a 77.8% specificity for the observed a positive correlation between the frequency of 13 Journal of Neurology Fig. 2  Cumulative proportion of patients showing hyperperfusion pattern were almost exclusively observed in patients with hyperper- on pCT according to final diagnosis (A) and EEG patterns (B). The fusion on pCT and more than two-fifths of patients with IEDs had bars represent groups of patients. Approximately two of every three hyperperfusion. pCT, perfusion computed tomography; NEA, non- patients who received a final diagnosis of SE had hyperperfusion. In epileptiform abnormalities; IED, interictal epileptiform discharges; the group of patients diagnosed with a single epileptic seizure, 78% SES, single epileptic seizure; SE, status epilepticus showed no pCT abnormalities. EEG findings compatible with an SE lateralized periodic discharges and glucose metabolism [19]. higher ipsilateral relative CBF and CBV than patients with Taking in mind the unfeasibility of the latter in an ER setting, continuous or polymorphic slowing and normal EEG find- our group[9] compared prospectively in a previous study ings. In our study, however, most patients benefited from pCT images of patients with an EEG-confirmed diagnosis of early EEG in the ER, supporting the robustness of the elec- focal‐onset SE with post-ictal control patients and found that troclinical associations observed. Hyperperfusion on pCT pCT hyperperfusion identified four of the five confirmed SE had high sensitivity for predicting persistent continuous episodes. Similar results were reported by Hauf et al. [20] in seizure activity compatible with SE on urgent EEG. The a retrospective study of patients with confirmed NCSE who high negative predictive value observed suggests that the had previously undergone urgent pCT. Thus, although its likelihood of observing SE on early EEG is very low in presence is significantly suggestive of SE diagnosis when patients without hyperperfusion on pCT. In addition, in the compared to single epileptic seizures in the proper setting, present study accompanying thalamic hyperperfusion was the low detection rate for SE diagnosis obtained in the pre- found to be a specific imaging marker of SE. In the first sent study underlines the in-force preponderance of clinical place, no single isolated seizure episode, even accompa- assessment and EEG over pCT examination in the ER. On nied by cortical cerebral hyperperfusion, was associated to the other hand, these results underscore the dynamic nature increased perfusion in thalamus. Moreover, as reported first of hemodynamic perfusion changes associated with SE, on arterial spin labeling sequence on MRI by Ohtomo et al. previously demonstrated by human single-photon emission [24] and recently on pCT by Giovannini et al. [22], thalamic computed tomography [21]. hyperperfusion, although not sensitive, is highly specific of However, the findings of the present study also point to continuous sustained ictal activity compatible with SE on a significant relationship between focal hyperperfusion of EEG. Therefore, the present study suggests that early pCT the brain measurable by pCT and epileptiform discharges findings, together with other factors, such as clinical pres- on EEG in an acute setting, particularly in SE and sustained entation [25], could serve as an adjunctive diagnostic tool ictal activity on EEG. In a smaller study, Payabvash et al. for NCSE risk stratification; thus, probably avoiding delayed [23] found that patients with epileptiform interictal dis- management in acute settings, especially if accompanied by charges on follow-up EEG within 1 week of pCT showed the highly specific feature of thalamic hyperperfusion. 13 Journal of Neurology Fig. 3  A 57-year-old man experienced a focal motor- onset seizure that progressed to presumed non-convulsive status epilepticus without coma. Perfusion computed tomography performed within 2 h of symptom onset showed a focal increase in cerebral blood flow (A, arrow) together with a shortening of circulation times in the left frontal cortex (B and C, arrows). All the above findings were accompa- nied by relative preservation of subcortical white matter and hyperperfused homolat- eral thalamus (D, arrow). An electroencephalogram acquired within 2 h of neuroimaging (E, ten seconds-epoch longitudinal bipolar montage) showed focal electrographic status epilepticus in the left fronto-central region without motor prominent symp- toms. A second electroencepha- logram performed 20 h later (F) showed left fronto-central periodic lateralized discharges without modifiers Reliable prospective data have shown that pCT per- patterns—mainly normal brain perfusion—in patients formed within 80 min of the end of a seizure shows post- diagnosed with a single epileptic seizure. Most retro- ictal hypoperfusion in 80% of patients with epilepsy spective studies assessing pCT changes in patients with [26]. In our series, however, pCT revealed nonspecific single epileptic seizures in an acute setting have shown 13 Journal of Neurology Table 2  Association between thalamic hyperperfusion and final diag- The main limitations of our study are its small sample nosis and EEG pattern size and retrospective design, which inherently lead to Total Thalamic hyperperfusion P value* biases, such as variations in the time of pCT and EEG acqui- sition and the interpretation of clinical features. To solve Yes (n = 8) No (n = 39) this issue, other group’s studies included for analysis only Final diagnosis those patients that underwent an EEG examination close in SS; n (%) 27 (57.4) 0 (0) 27 (69.2) < 0.0001 time after the pCT [20, 22]. However, in the present study SE; n (%) 20 (42.6) 8 (100) 12 (30.8) patients were included if an EEG was performed within 24 h EEG‡ findings of symptom onset. Thus, probably capturing a more realistic NEA; n (%) 19 (46.3) 0 (0) 19 (55.9) 0.003 picture of regular delayed conventional EEG performance in IED; n (%) 12 (29.3) 2 (28.6) 10 (29.4) many ERs [5]. In addition, longer, continuous EEG record- SES; n (%) 1 (2.4) 0 (0) 1 (2.9) ings in the ER could improve the characterization of patients SE; n (%) 9 (22) 5 (71.4) 4 (11.8) along the ictal–interictal continuum and its neuroimaging correlate obtained on pCT. Another limitation of our study is SS singles seizure, SE status epilepticus, EEG electroencephalogra- phy, NEA non-epileptiform abnormalities, IED interictal epileptiform that we did not investigate the diagnostic specificity of pCT discharges, SES single epileptic seizure by comparing results between patients diagnosed with SE *Bold typed p values represent statistically significant differences and controls with other conditions likely to cause cerebral between groups. hyperemia, such as encephalitis [31]. In addition, in light of ‡: 41 patients underwent EEG registry. Percentages are calculated the presented results, further studies could assess pCT imag- according to these numbers ing usefulness to differentiate between SE, and other acute presenting non-ictal mimics, such as confusional or dissocia- similar results to ours [17]; thus, pCT seems not useful tive psychogenic states. Nonetheless, our study brings new regarding the diagnostic process of single epileptic sei- evidence and emphasizes the potential value of pCT in the zures. In addition, none of the patients with hypoperfu- assessment of epileptic disorders in ER settings. sion on pCT in our study were diagnosed with SE. Other authors, however, have reported “ictal hypoperfusion” in a small number of cases that could possibly be explained Conclusion by the presence of adjacent hyperperfused areas [27] or the acquisition of images during post-ictal periods [28]. Observation of focal brain hyperperfusion on pCT in an In brief, the current evidence suggests that SE should not appropriate clinical setting should raise suspicion of SE be contemplated in the differential diagnosis when pCT and lead to further clinical and EEG investigation, espe- shows isolated hypoperfusion in a patient with focal neu- cially if cerebral cortical hyperperfusion is accompanied by rological deficits in an ER setting. increased perfusion of homolateral thalamus. The absence of We decided to use a visual rather than a quantitative hyperperfusion rules out SE in most cases. pCT is not useful approach [29, 30] to analyze the perfusion maps, as a previ- in the ER for the detection of single epileptic seizures in ER ous study by our group showed that visual analysis appears due to its low diagnostic yield. Larger, prospective studies to be more sensitive for the detection of hyperperfusion [9], are needed to further study pCT as an adjunctive diagnostic and most importantly, is easily applicable to routine clinical modality for NCSE and EEG findings along the ictal–inter- practice. ictal continuum in emergency settings. Table 3  Predictive yield of Final diagnosis:status epilepticus (SE) EEG: SE pattern cerebral cortical and thalamic hyperperfusion on pCT for SE Cortical Thalamic Cortical Thalamic as a final diagnosis and an SE pattern on EEG Se, % (95% CI) 60 (36.4–80.2) 40 (20–63.6) 88.9 (50.7–99.4) 55.5 (25.6–84.6) Sp, % (95% CI) 77.8 (57.2–90.6) 100 (84.5–100) 75 (56.2–87.9) 93.7 (77.8–98.9) PPV, % (95% CI) 66.7 (41.1–85.6) 100 (59.7–100) 50 (20.7–69.4) 71.4 (30.2–94.9) NPV % (95% CI) 72.4 (52.0–86.5) 69.2 (52.2–82.4) 96 (77.7–99.8) 88.2 (71.6–96.1) EEG electroencephalogram, NPV negative predictive value, PPV positive predictive value, Se sensitivity, Sp specificity 13 Journal of Neurology Acknowledgements This research did not receive a grant from funding 9. González-Cuevas M, Coscojuela P, Santamarina E et al (2019) agencies in the public, commercial, or not-for-profit sector. The authors Usefulness of brain perfusion CT in focal-onset status epilepticus. thank Anne Murray for English language support. Epilepsia 60:1317–1324. https://​doi.​org/​10.​1111/​epi.​16063 10. Strambo D, Rey V, Rossetti AO et al (2018) Perfusion-CT imaging Author contributions All authors contributed to the study conception in epileptic seizures. J Neurol 265:2972–2979. https://​doi.​org/​10.​ and design. Material preparation, data collection and analysis were per- 1007/​s00415-​018-​9095-1 formed by Juan Luis Restrepo-Vera, Elena Fonseca, Pilar Coscojuela, 11. Van Cauwenberge MGA, Dekeyzer S, Nikoubashman O et al Manuel Quintana, and Manuel Toledo. The first draft of the manuscript (2018) Can perfusion CT unmask postictal stroke mimics? A was written by Juan Luis Restrepo-Vera and all authors commented on case-control study of 133 patients. Neurology 91:e1918–e1927. previous versions of the manuscript. All authors read and approved https://​doi.​org/​10.​1212/​WNL.​00000​00000​006501 the final manuscript. 12. Hogan-Cann AD, Lu P, Anderson CM (2019) Endothelial NMDA receptors mediate activity-dependent brain hemodynamic responses in mice. Proc Natl Acad Sci USA 116:10229–10231. Data availability The data that support the findings of this study are https://​doi.​org/​10.​1073/​pnas.​19026​47116 available from the corresponding author upon reasonable request. 13. Mehra A, Guérit S, Macrez R et al (2020) Nonionotropic action of endothelial NMDA receptors on blood-brain barrier permeability Declarations via Rho/ROCK-mediated phosphorylation of myosin. J Neurosci 40:1778–1787. https://​doi.​org/​10.​1523/​JNEUR​OSCI.​0969-​19.​ Conflicts of interest EF declares speaker fees from UCB Pharma, Es- 2019 teve laboratorios, Eisai Inc and Sanofi Genzyme. ES declares speaker 14. Trinka E, Cock H, Hesdorffer D et al (2015) A definition and fees from UCB Pharma, BIAL Pharmaceutical, EISAI Inc. Arvelle classification of status epilepticus–report of the ILAE task force and Esteve laboratorios. LA declares speaker fees from UCB Pharma, on classification of status epilepticus. Epilepsia 56:1515–1523. BIAL Pharmaceutical, EISAI Inc., Sanofi Genzyme and Esteve labora- https://​doi.​org/​10.​1111/​epi.​13121 torios. MT has received speaker honoraria from UCB Pharma, Esteve, 15. Hirsch LJ, LaRoche SM, Gaspard N et al (2013) American clinical BIAL Laboratories, and Sanofi. The other authors declare no financial neurophysiology society’s standardized critical care EEG termi- or other conflicts of interest. nology: 2012 version. J Clin Neurophysiol 30:1–27. https://​doi.​ org/​10.​1097/​WNP.​0b013​e3182​784729 Ethical approval We confirm that we have read the Journal’s position 16. Leitinger M, Beniczky S, Rohracher A et al (2015) Salzburg con- on issues concerning ethical publication and affirm that this report is sensus criteria for non-convulsive status epilepticus—approach to consistent with those guidelines. clinical application. Epilepsy Behav 49:158–163. https://​doi.​org/​ 10.​1016/j.​yebeh.​2015.​05.​007 17. Gugger JJ, Llinas RH, Kaplan PW (2020) The role of CT perfu- sion in the evaluation of seizures, the post-ictal state, and status epilepticus. Epilepsy Res 159:106256. https://​doi.​org/​10.​1016/j.​ References eplep​syres.​2019.​106256 18. López Ruiz R, Quintas S, Largo P et al (2019) Usefulness of mul- 1. Rizos T, Jüttler E, Sykora M et al (2011) Common disorders in tiparametric computerized tomography findings in the differential the neurological emergency room—experience at a tertiary care diagnosis of stroke mimics of epileptic origin: a preliminary study. hospital. Eur J Neurol 18:430–435. https://d​ oi.o​ rg/1​ 0.1​ 111/j.1​ 468-​ Neurologia 34:73–79. https://​doi.​org/​10.​1016/j.​nrl.​2016.​10.​010 1331.​2010.​03170.x 19. Subramaniam T, Jain A, Hall LT et al (2019) Lateralized periodic 2. Trinka E, Kälviäinen R (2017) 25 years of advances in the defi- discharges frequency correlates with glucose metabolism. Neu- nition, classification and treatment of status epilepticus. Seizure rology 92:e670–e674. https://​doi.​org/​10.​1212/​WNL.​00000​00000​ 44:65–73. https://​doi.​org/​10.​1016/j.​seizu​re.​2016.​11.​001 006903 3. Waterhouse EJ (2018) The epidemiology of status epilepticus. 20. Hauf M, Slotboom J, Nirkko A et al (2009) Cortical regional In: Drislane FW, Kaplan PW (eds) Status epilepticus, 2nd edn. hyperperfusion in nonconvulsive status epilepticus measured by Springer International Publishing, Cham, pp 15–30 dynamic brain perfusion CT. AJNR Am J Neuroradiol 30:693– 4. Moulin S, Leys D (2019) Stroke mimics and chameleons. Curr 698. https://​doi.​org/​10.​3174/​ajnr.​A1456 Opin Neurol 32:54–59. https://d​ oi.o​ rg/1​ 0.1​ 097/W ​ CO.0​ 00000​ 0000​ 21. Van Paesschen W (2004) Ictal SPECT. Epilepsia 45(Suppl 4):35– 000620 40. https://​doi.​org/​10.​1111/j.​0013-​9580.​2004.​04008.x 5. Vespa PM, Olson DM, John S et al (2020) Evaluating the clinical 22. Giovannini G, Malagoli M, Turchi G et al (2021) Cortical and impact of rapid response electroencephalography: the DECIDE thalamic hyper-perfusion in non-convulsive status epilepticus. multicenter prospective observational clinical study. Crit Care Relationship between perfusion CT patterns and Salzburg EEG Med 48:1249–1257. https://​doi.​org/​10.​1097/​CCM.​00000​00000​ criteria. Seizure 92:10–17. https://d​ oi.o​ rg/1​ 0.1​ 016/j.s​ eizur​ e.2​ 021.​ 004428 08.​002 6. Naylor DE, Liu H, Wasterlain CG (2005) Trafficking of GABA(A) 23. Payabvash S, Oswood M, Truwit C, Mckinney A (2015) Acute CT receptors, loss of inhibition, and a mechanism for pharmacoresist- perfusion changes in seizure patients presenting to the emergency ance in status epilepticus. J Neurosci 25:7724–7733. https://​doi.​ department with stroke-like symptoms: correlation with clinical org/​10.​1523/​JNEUR​OSCI.​4944-​04.​2005 and electroencephalography findings. Clin Radiol. https://d​ oi.o​ rg/​ 7. Sutter R, Kaplan PW, Rüegg S (2013) Outcome predictors for 10.​1016/j.​crad.​2015.​06.​078 status epilepticus–what really counts. Nat Rev Neurol 9:525–534 24. Ohtomo S, Otsubo H, Arai H et al (2021) Hyperperfusion in the 8. Powers WJ, Rabinstein AA, Ackerson T et al (2018) 2018 Guide- thalamus on arterial spin labelling indicates non-convulsive status lines for the early management of patients with acute ischemic epilepticus. Brain Commun. https://d​ oi.o​ rg/1​ 0.1​ 093/b​ rainc​ omms/​ stroke: a guideline for healthcare professionals from the American fcaa2​23 Heart Association/American Stroke Association. Stroke 49:e46– 25. Requena M, Fonseca E, Olivé M et al (2019) The ADAN scale: a e110. https://​doi.​org/​10.​1161/​STR.​00000​00000​000158 proposed scale for pre-hospital use to identify status epilepticus. Eur J Neurol 26:760-e55. https://​doi.​org/​10.​1111/​ene.​13885 13 Journal of Neurology 26. Li E, d’Esterre CD, Gaxiola-Valdez I et al (2019) CT perfusion 29. Manganotti P, Furlanis G, Ajčević M et al (2019) CT perfusion measurement of postictal hypoperfusion: localization of the sei- and EEG patterns in patients with acute isolated aphasia in sei- zure onset zone and patterns of spread. Neuroradiology 61:991– zure-related stroke mimics. Seizure 71:110–115. https://​doi.​org/​ 1010. https://​doi.​org/​10.​1007/​s00234-​019-​02227-8 10.​1016/j.​seizu​re.​2019.​07.​005 27. Lee HW, Hong SB, Tae WS (2000) Opposite ictal perfusion pat- 30. Austein F, Huhndorf M, Meyne J et al (2018) Advanced CT for terns of subtracted SPECT: Hyperperfusion and hypoperfusion. diagnosis of seizure-related stroke mimics. Eur Radiol 28:1791– Brain 123(10):2150–2159. https://​doi.​org/​10.​1093/​brain/​123.​10.​ 1800. https://​doi.​org/​10.​1007/​s00330-​017-​5174-4 2150 31. Li R, Shi P-A, Liu T-F et al (2019) Role of 3D pseudocontinuous 28. Wiest R, von Bredow F, Schindler K et al (2006) Detection arterial spin-labeling perfusion in the diagnosis and follow-up in of regional blood perfusion changes in epileptic seizures with patients with herpes simplex encephalitis. AJNR Am J Neurora- dynamic brain perfusion CT–a pilot study. Epilepsy Res 72:102– diol 40:1901–1907. https://​doi.​org/​10.​3174/​ajnr.​A6279 110. https://​doi.​org/​10.​1016/j.​eplep​syres.​2006.​07.​017 13