No. 2 ·October 2019

eISSN 2635-5280 eISSN

October 2019 October ·

No. 2 No. · Vol. 2 Vol.

Journal of Neurointensive Care Vol.2 · No.2 · September 2019 eISSN 2635-5280 http://www.e-jnic.org

Vol. 2 · No. 2 · October 2019

Aims and Scope Journal of Neurointensive Care (J Neurointensive Care, JNIC) is the official journal of the Korean Neurointensive Care Society and is published biannually (the last day of April and October). It is a peer reviewed, open access journal aimed at publishing all aspects of neurointensive care medicine, such as stroke, brain and spine trauma, perioperative neurosurgical intensive care, neuro-pediatric severe anormaly, CNS infection, seizure, myelitis and etc. It is intended for all neurointensive care providers as neurosurgeons, neurologists, anesthesiologists, emergency physicians, and critical care nurses treating patients with urgent neurologic disorders.

Open Access This is an open-access article distributed under the terms of the Creative Commons Attribution Non- Commercial License (http://creativecommons.org/licenses/by-nc/4.0) which permits unrestricted noncommercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

Subscription information The Korean Neurointensive Care Society will send J Neurointensive Care for free to some important individuals and institutions. Full text PDF files are also available at the official website (http://www.e-jnic.org). To order a subscription to J Neurointensive Care, please contact out editorial office.

Publisher: Sang Hyung Lee

Editors-in-Chief: Dong-Hyuk Park, Sang Gu Lee

Editorial Office Department of Neurosurgery, Korea University College of Medicine 73, Inchon-ro, Seongbuk-gu, Seoul 02841, Korea Tel: +82-2-920-6833 Fax: +82-2-929-0629 E-mail: [email protected]

Printing Office M2community Co. 8th FL, DreamTower, 66 Seongsui-ro, Seongdong-gu, Seoul 04784, Korea Tel: +82-2-2190-7300 Fax: +82-2-2190-7333 E-mail: [email protected]

Copyright © 2019 by Korean Neurointensive Care Society ∞This paper meets the requirements of KS X ISO 9706, ISO 9706-1994 and ANSI/NISO Z39.48-1992 (Permanence of Paper) eISSN 2635-5280 http://www.e-jnic.org

Editorial Board Editors-in-Chief Dong-Hyuk Park Korea University, Korea Sang Gu Lee Gachon University, Korea

Associate Editor Jun Seok W. Hur Korea University, Korea

Editorial Board Jin Hwan Cheong Hanyang University, Korea Won-Sang Cho Seoul National University, Korea Kyu-Sun Choi Hanyang University, Korea Joon ho Chung Yonsei University, Korea Eun Jin Ha Seoul National University, Korea Kyung Sool Jang The Catholic University of Korea, Korea Ju Ho Jeong Dongguk University, Korea Kwang Wook Jo The Catholic University of Korea, Korea Sung Pil Joo Chonnam National University, Korea Chang-Hyun Kim Keimyung University, Korea Tae Gon Kim CHA University, Korea Young Jin Kim Dankook University, Korea Young Woo Kim The Catholic University of Korea, Korea Young Zoon Kim Sungkyunkwan University, Korea Doo-Sik Kong Sungkyunkwan University, Korea Hyon-Jo Kwon Chungnam National University, Korea Soon Chan Kwon Ulsan University, Korea Jong-Young Lee Hallym University, Korea Sang Weon Lee Pusan National University, Korea Sung Ho Lee Kyung-Hee University, Korea Taek Kyun Nam Chung-Ang University, Korea Cheol Wan Park Gachon University, Korea Jun Bum Park Ulsan University, Korea Keun Young Park Yonsei University, Korea Young Seok Park Chonbuk National University, Korea Jeong-Am Ryu Sungkyunkwan University, Korea Sang Woo Song Konkuk University, Korea Chan Jong Yoo Gachon University, Korea Vol. 2 · No. 2 · October 2019 CONTENTS

Original Articles 39 Change in Nutrition Strategy after Nutritional Support Team Activities for Cerebrovascular Disease Patients in the Intensive Care Unit Jin Lee, Sung-Tae Kim, Won Hee Lee, Keun Soo Lee, Sung Hwa Paeng, Se Young Pyo, Young Gyun Jeong, Moo Seong Kim, Yong Tae Jung

45 Endovascular Treatment of Intracranial Aneurysms Associated with Brain Arteriovenous Malformations Seonah Choi, Joonho Chung

52 Preliminary Experience of Lvis Blue in the Internal Carotid Artery for The Treatment Of Wide-Necked Intracranial Aneurysms Jeong Hwa Kim, Chang Ki Jang, Jae Whan Lee, Keun Young Park, Joonho Chung

58 Radiologic Factors for Predicting Dynamic Spinal Cord Compression in Conventional Eervical MRI Kwang-Ui Hong, Jong-myung Jung, Seung-Jae Hyun, Ki-Jeong Kim, Tae-Ahn Jahng

64 Safety and Feasibility of Percutaneous Dilatational Tracheostomy Performed by a Neurointensivist Compared with Conventional Surgical Tracheostomy in Neurosurgery Intensive Care Unit John Kwon, Yong Oh Kim, Jeong-Am Ryu

70 Crescent Sign Following Enterprise-Stent assisted Coil Embolization of Distal Internal Carotid Artery Aneurysms Young-Jin Kim, Je Young Yeon, Jong-Soo Kim, Seung-Chyul Hong

77 Acute Acalculous Cholecystitis in Neurological Patients; Clinical Review, Risk Factors, and Possible Mechanism See Won Um, Hak Cheol Ko, Seung Hwan Lee, Hee Sup Shin, Jun Seok Koh

Case Reports 82 Delayed Occurrence of CSF Oculorrhea after Mild Blunt Traumatic Head Injury Sang Bin Choi, Sang Pyung Lee, Jin Wook Baek, Tae Joon Park

87 Endovascular Treatment of Ruptured Dissecting Aneurysms on the P2-3 Segment of the Posterior Cerebral Artery: A Report of Two Cases Seonah Choi, Joonho Chung Vol. 2 · No. 2 · October 2019 CONTENTS

93 Internal Carotid Artery Dissection with Traumatic Pseudoaneurysm Formation after Penetrating Head Injury Jeong Sik Ham, Jang Hun Kim, Won-Ki Yoon

99 Critical Care in Patient with Neuromuscular Cervicothoracic Kyphosis Kwang-Ui Hong, Jong-myung Jung, Seung-Jae Hyun, Ki-Jeong Kim, Tae-Ahn Jahng eISSN 2635-5280 Original Article J Neurointensive Care 2019;2(2):39-44 https://doi.org/10.32587/jnic.2019.00150

Change in Nutrition Strategy after Nutritional Support Team Activities for Cerebrovascular Disease Patients in the Intensive Care Unit Jin Lee, Sung-Tae Kim, Won Hee Lee, Keun Soo Lee, Sung Hwa Paeng, Se Young Pyo, Young Gyun Jeong, Moo Seong Kim, Yong Tae Jung Department of Neurosurgery, School of Medicine, Inje University Busan Paik Hospital, Busan, Korea

Received: June 24, 2019 Accepted: July 8, 2019 Objective This study was done to introduce a nutrition support strategy and analyze the effect of nutritional Corresponding Author: support team (NST) activities in a single-center intensive care unit. Sung-Tae Kim, M.D., Ph.D. Methods Department of Neurosurgery Medicine, School of Medicine, Between January and June 2017, acute severe cerebrovascular disease (CVA) patients who admit- Inje University Busan paik ted to the intensive care unit and referred to the NST, were enrolled and assigned to the NST Hospital, Bok Ji Ro 75, Busan Jin- group. Demographics, initial GCS score, needed calorie intake (NCI), type of nutrition, time of Ku, Busan, Korea switching from parenteral nutrition (PN) to enteral nutrition (EN), NCI satisfaction rate, inci- Tel:+82-51-890-6144 dence of feeding intolerance (FI), and clinical outcomes were analyzed, and compared with those Fax:+82-51-898-4244 in the control group, retrospectively. E-mail: [email protected] Results There were 38 patients in the NST group and 39 in the control group. The initial GCS score was 7.03±2.19 in the NST group and 6.82±2.19 in the control group. The average NCI was 1557.9 kcal in the NST group and 1635.9 kcal in the control group. In all patients, the start of PN was possible within 24 hours after admission. The EN start rate was slightly lower in the NST group. EN was initiated earlier in the NST group than in the control group. The NCI satisfaction rate at discharge was higher in the NST group. No statistically significant differences were observed in the rate of good clinical outcomes and incidence of FI between the two groups. Conclusion NST activity enabled the earlier start of EN and achievement of a high NCI satisfaction rate without increasing nutrition support related risks.

Keywords: Nutritional support; Cerebrovascular disorder; Intensive care unit

INTRODUCTION previous studies4,7,8,13-15), we defined it as the presence of a gastric residual volume of 100 ml>24 h on any day or gastrointestinal (GI) Appropriate nutritional support is key to the recovery of patients dysfunction such as diarrhea, abdominal distention, or signs of GI who are acutely ill or have undergone major surgery. Previous stud- bleeding in the form of melena, hematochezia, hematemesis and ies have focused to a certain degree on nutrition in the intensive bleeding via the L-tube. The team attempted to initiate PN within care unit, and guidelines exist for the provision of nutritional sup- 24 hours from admission and then start EN, with 600 kcal as the port in such settings2,18). However, nutritional support strategies starting dose, as soon as possible after an intracranial pressure cri- should be formulated based on each disease type. Cerebrovascular sis3,16). EN and PN were provided in combination until EN covered disease (CVA) is unique in that it is characterized by increased in- about 70% of a patient’s NCI. When FI was noticed, the team ac- tracranial pressure, which may lead to gut-brain axis dysfunc- tively recommended GI motility drug use7), changed the EN for- tion10,22) and normal gastrointestinal tract function. Therefore, it is mula, continuously implemented feeding techniques, and used oth- mandatory to consider such disease characteristics in the provision er nutritional interventions. Once stable oral feeding was estab- of nutritional support. However, few studies have focused on nutri- lished, a 100% NCI satisfaction rate was considered achieved. tional support for CVA11,21). Recently, our institute established a nutritional support team Data collection (NST) that includes neurosurgeons among others, for the active Data were collected equally from both the NST and control management of CVA patients. Significant changes have been ob- groups. Basic data on sex, age, weight, height, length of hospital served in the nutritional support strategies employed after the per- stay and main diagnosis were collected. Information on the initial formance of NST activities. The purpose of this study was to intro- GCS score, NCI, type of nutrition in the form of PN and EN, time duce our nutrition support strategy and the nutrition patterns used taken for PN and EN initiation, pattern of the increase in EN, NCI for CVA patients, and analyze the effect of NST activities in a sin- satisfaction rate, and clinical outcomes was analyzed. Data on FI gle-center intensive care unit. and surgical interventions (craniotomy, stereotactic surgery, ex- tra-ventricular drainage, neurointervention for aneurysm or arte- MATERIALS AND METHODS rio-venous malformation) were also collected. Clinical outcomes were measured using the modified Rankin scale (mRS). Study population This retrospective study was approved by the Institutional Re- Statistical analysis view Board (Approval number, 19-0043).This analysis included Statistical analyses were performed using SPSS for Windows patients with new-onset severe CVA (initial Glasgow Coma Scale (version 25; SPSS, Chicago, IL, USA). In the statistical analyses, [GCS] score 4 to 12), who were admitted to the intensive care unit clinical outcomes were dichotomized into ‘good’ (mRS 0 to 4) and of our institution. Acute ischemic stroke patients who underwent ‘poor’ (mRS 5 to 6). Factors related to nutrition were evaluated. only medical treatment or intra-arterial thrombectomy were ex- Univariate analysis was performed with age, sex, initial GCS score, cluded. As NST activity was initiated only in 2015, the early years length of hospital stay, weight, height, body mass index (BMI), sur- of patient data were not used to consider NST activity optimiza- gical intervention, NCI, PN start rate, EN start rate, EN-related pa- tion. Between January and June 2017, patients who were referred rameters, NCI satisfaction rate on discharge, and clinical outcomes to the NST were enrolled and assigned to the NST group (39 pa- as factors using Pearson’s chi-square test, t-test, Kruskal-Wallis test, tients). The control group included patients with the same inclu- and Fisher’s exact test. Statistical significance was set at p< 0.05 for sion criteria, but who were enrolled from the period between Janu- a 95% confidence interval. ary and June 2014 (38 patients). RESULTS Nutrition support team and strategy The NST comprised neurosurgeons, general surgeons, dedicated The demographic characteristics of the patients are shown in nurses, pharmacists, and clinical dietitians. NST rounds were con- Table 1. There were 38 patients (F/M=17/21, mean age ducted twice a week, and the physicians alternately attended the 63.37±13.59 y) in the NST group and 39 patients (F/M= 21/18, round once a week. The needed calorie intake (NCI) was calculat- mean age 65.51±16.78 y) in the control group, and the BMI val- ed via the Harris-Benedict equation with the ideal weight20). There ues in the two groups were 22.48±3.35 kg/m2 and 23.43±4.38 is no consistent definition for feeding intolerance (FI). Based on kg/m2, respectively. The initial GCS score was 7.03±2.19 in the

40 www.e-jnic.org Nutritional Support Team Activities Jin Lee et al.

Table 1. Demographic characteristics of the nutrition support team group and control group Group p-value NST (38) Control (39) Sex, F/M 17/21 21/18 0.424* Age (y) 63.37±13.59 65.51±16.78 0.540† Weight (kg) 60.87±10.05 63.67±14.86 0.338† Height (cm) 164.42±7.94 171.77±48.28 0.357† BMI (kg/m2) 22.48±3.35 23.43±4.38 0.291† GCS score 7.03±2.19 6.82±2.19 0.681† Length of hospital stay 33.16±28.77 35.51±27.62 0.715† Surgical intervention (%) 31 (81.6) 36 (92.3) 0.192‡ NST: nutritional support team; GCS: Glasgow coma scale; F: female; M: male; BMI: body mass index. *Pearson’s chi-square test. †t-test. ‡Fisher’s exact test.

NST group and 6.82±2.19 in the control group. Thirty-one pa- NST Control tients (81.6%) in the NST group and 36 (92.3%) in the control Enrolled 38 39 group received surgical intervention, and their lengths of hospital Stroke category stay were 33.16±28.77 and 35.51±27.62 days, respectively. There Cerebral infarction 2 7 was no statistically significant difference between the NST group S-ICH (hypertensive, AVM) 23 20 and control group in each parameter. Fig. 1 shows the nutrition flow of the two groups. Stroke was di- Cerebral aneurysm rupture 13 12 vided into three categories. Patients who underwent decompressive craniectomy due to large territory infarction were categorized as cerebral infarction patients, while those with spontaneous intra- PN start 38 39 cerebral hematoma (S-ICH) due to hypertensive causes or arteriovenous malformation ruptures were categorized as S-ICH patients. Patients with spontaneous subarachnoid hemorrhage due to aneurysm rupture were categorized as cerebral aneurysm rupture pa- EN start 29 (oral 3) 35 (oral 3) tients. PN was initiated in all patients in both groups. EN was started in 29 patients in the NST group and 35 patients in the control group. Twenty-six of the 29 patients in the NST group and 15 of Calorie increase after EN 26 15 the 35 patients in the control group had an EN calorie increase. Six patients in both groups had FI. Twenty-six patients in the NST group and 24 in the control group were eventually discharged while under EN. Feeding intolerance 6 6 Table 2 shows the statistical findings of the feeding-related variables. There was no statistically significant difference in terms of the NCI between the two groups (1,557.89±228.55 vs Feeding type on discharge ± 1,635.90 447.50 kcal, p= 0.338). In all patients, PN could be ini- EN (oral) 26 (7) 24 (6) tiated within 24 hours after admission. The EN start rate was lower PN 1 in the NST group (76.3% vs 89.7%, p= 0.138). However, EN was initiated earlier in the NST group than in the control group Fig. 1. Nutrition flow in the two groups of patients with cerebrovascular (4.83±1.28 days vs 7.14±5.49 days after admission, p= 0.02). In disease. NST, nutritional support team; S-ICH, spontaneous intracerebral addition, the interval between the start of EN and increase in hemorrhage, AVM, arteriovenous malformation; PN, parenteral nutrition; EN, enteral nutrition. calories was shorter in the NST group than the control group. (4.91±3.97 vs 17.50±16.69 days, p= 0.028). www.e-jnic.org 41 Nutritional Support Team Activities Jin Lee et al.

Table 2. Statistically significant differences in the feeding-related variables between the nutritional support team group and control group Group p-value NST (38) Control (39) Needed calorie intake 1,557.89±228.55 1,635.90±447.50 0.338* PN start (%) 38(100) 39(100) EN start (%) 29 (76.3) 35 (89.7) 0.138† EN start day† 4.83±1.28 7.14±5.49 0.020* EN start oral feeding 3 3 Calories at EN initiation 685.7±269 714.3±376 0.736* Patients with EN calorie increase (%)‡ 26 (89.7) 15 (42.9) 0.000† Interval between EN start and calorie increase § 4.91±3.97 17.50±16.69 0.028|| Needed calorie intake satisfaction rate on discharge¶ 81.4%±25.9 56.1%±28.4 0.001* Feeding intolerance (%)** 6 (20.7) 6 (17.1) 0.757† NST: nutritional support team; EN: enteral nutrition; PN: parenteral nutrition. *t-test. †Pearson’s chi-square test. ‡included in case of oral feeding. §Number of patients in each group: NST=23, Control=12. ||Kruskal-Wallis test. ¶Number in each group: NST=26, Control=25. **Number in each group: NST=29, Control=35.

Table 3. Clinical outcomes in the nutrition support team group and control group Group mRS at 30 days p-value NST (38) Control (39) 1 1 1 2 1 0 3 3 2 4 9 9 5 13 13 6 (no EN) 11 (9) 14 (4) Good clinical outcomes mRS≤ 4 14/38 12/39 0.573* (%) -36.8 -30.8 mRS: modified Rankin scale; NST: nutritional support team; EN: enteral nutrition; PN: parenteral nutrition. *Pearson’s chi-square test.

The EN starting calorie value was similar in both groups mographics and basic data, allowing for the easy comparison of the (685.7±269 vs 714.3 ± 376 kcal, p= 0.736). The NCI satisfaction effects of the nutritional feeding activities. In both groups, PN rate on discharge was higher in the NST group than the control could be initiated within 24 hours after admission. However, on group (81.4%±25.9 vs 56.1%±28.4, p= 0.001). The incidence of average, EN was initiated faster in the NST group. Furthermore, FI was similar in two groups (20.7% vs 17.1%, p= 0.757). the EN dose was initially increased later in the control group than Table 3 shows the differences in the clinical outcomes between the NST group, suggesting that the final NCI satisfaction rate on the two groups. The NST group had a higher rate of good clinical discharge may have been lower in the control group. Such differ- outcomes (mRS≤ 4) on day 30 after admission (36.8%, 14 pa- ences can be attributed to the effects of the NST activities. The EN tients) than the control group (30.8%, 12 patients); however, sta- calorie increase in the control group was decided on by the neuro- tistical significance was not observed (p= 0.573). surgeon in charge. However, different neurosurgeons may have had different nutritional feeding strategies and may also have reacted to DISCUSSION FI differently. In the NST group, owing to the presence of stan- dardized strategies and goals, better clinical outcomes may have In our study, the NST and control groups had similar patient de- been possible.

42 www.e-jnic.org Nutritional Support Team Activities Jin Lee et al.

In the NST group, the percentage of patients in whom EN was enhance the role of the NST in the future, diverse nutritional strat- started was lower than that in the control group. This could be at- egies should be applied according to a patient’s condition and diag- tributed to the patients’ overall condition on admission. EN was nostic category. intentionally not initiated in patients who were diagnosed as hav- A multicenter randomized controlled study Optimizing early ing a low chance of neurological recovery or those in a coma. All enteral nutrition in severe stroke (OPENS) is being performed in such patients died. The rate of good clinical outcomes was higher China21). The results of that study and other similar studies must in the NST group; however, the difference between the groups was be considered in the formulation of further nutritional support not statistically significant. This could be because the study had a strategies for the ICU. small sample size and patients’ outcomes do not solely depend on A limitation of this study is its retrospective design. Therefore, nutrition. The clinical outcomes themselves should be further ana- the presence of selection bias and protocol deviations cannot be lyzed in future studies to establish if malnutrition can lead to the ruled out. Furthermore, the sample size was small; therefore, the more rapid deterioration of health in CVA6,9). Although statistically results of the statistical analysis are limited. In addition, this study significant differences in terms of clinical outcomes were not ob- was single-center design and enrolled only people from Korea. For served in our retrospective study, faster EN initiation, high NCI the establishment of nutrition-related guidelines for CVA patients, satisfaction rate and low FI incidence rate in the NST group indi- well-controlled prospective studies with a large sample size must cate that the administered NST activities led patients with severe be conducted. CVA in the right nutritional strategic direction5,6,17). Despite the NST activities, the NST group patients’ nutritional CONCLUSION strategies did not adhere to those mentioned in previous guidelines for ICU patients2,18). To reduce the rate of FI, length of ventilation NST activities enabled the earlier start of EN and the achieve- and rate of infections, hypocaloric EN was provided and EN was ment of a high NCI satisfaction rate, without an increase in the risk initiated earlier in the early stages of ICU care in our NST group18). of FI. Well-planned prospective studies are needed to reveal the However, it still took an average of 9 days or longer after admission clinical benefit of active nutritional support for CVA patients. for half the NCI to be reached, and about 80% of the NCI was eventually obtained via EN in the NST group at discharge. We may NOTES have taken FI too seriously compared to other models, leading to such results7,15). Compared to another study, the prevalence of FI Conflict of interest (17-20%) in our groups showed similar prevalence of FI in neurologic patients, which is lower than that noticed in cardiovascular, No potential conflict of interest relevant to this article was re- gastrointestinal, trauma and sepsis ICU patients7). The rate of FI ported. seems to be influenced by a patient’s condition or diagnostic category. The similar results, in terms of FI prevalence, observed in the ACKNOWLEDGEMENTS patients with neurologic impairment may be due to the gut-brain axis; further research should focus on this aspect1). Even among This work was supported by 2018 Inje University Busan Paik stroke patients, large territory infarction patients and hemorrhagic Hospital research grant. stroke patients with slow clinical recovery may need different nutrition strategies compared to stroke patients who undergo in- REFERENCES tra-arterial thrombectomy with earlier clinical recovery. The Harris-Benedict equation was used to estimate the energy 1. Carabotti M, Scirocco A, Maselli MA, Severi C. The gut-brain expenditure in our patients. However, indirect calorimetry seems axis: interactions between enteric microbiota, central and enter- to be better for the estimation of energy expenditure in critically ill ic nervous systems. Ann Gastroenterol 2015;28:203–209. patients12); however, at present, its use is limited owing to the asso- 2. Carney N, Totten AM, O’Reilly C, Ullman JS, Hawryluk GW, ciated cost. In clinical practice, it is becoming increasingly common Bell MJ, et al. Guidelines for the Management of Severe Traumat- for diverse units comprising hospitalists and intensivists to work ic Brain Injury, Fourth Edition. Neurosurgery 2017;80:6–15. with physicians for treatment administration19). The NST can be 3. Cook AM, Peppard A, Magnuson B. Nutrition considerations considered such a unit, as it strives to establish better treatment in traumatic brain injury. Nutr Clin Pract 2008;23:608–620. policies and achieve improved clinical outcomes. Furthermore, to 4. Deane AM, Fraser RJ, Chapman MJ. Prokinetic drugs for feed www.e-jnic.org 43 Nutritional Support Team Activities Jin Lee et al.

intolerance in critical illness: current and potential therapies. nal symptoms in intensive care patients. Acta Anaesthesiol Crit Care Resusc 2009;11:132–143. Scand 2009;53:318–324. 5. Giner M, Laviano A, Meguid MM, Gleason JR. In 1995 a cor- 14. Reintam A, Parm P, Kitus R, Starkopf J, Kern H. Gastrointesti- relation between malnutrition and poor outcome in critically ill nal failure score in critically ill patients: a prospective observa- patients still exists. Nutrition 1996;12:23–29. tional study. Crit Care 2008;12:R90. 6. Gramlich L, Kichian K, Pinilla J, Rodych NJ, Dhaliwal R, Hey- 15. Reintam Blaser A, Starkopf L, Deane AM, Poeze M, Starkopf J. land DK. Does enteral nutrition compared to parenteral nutri- Comparison of different definitions of feeding intolerance: A tion result in better outcomes in critically ill adult patients? A retrospective observational study. Clin Nutr 2015;34:956–961. systematic review of the literature. Nutrition 2004;20:843–848. 16. Ren G. Application of early nutrition support in neurosurgical 7. Gungabissoon U, Hacquoil K, Bains C, Irizarry M, Dukes G, coma patients. Indian J Surg 2015;77:1205–1208. Williamson R, et al. Prevalence, risk factors, clinical conse- 17. Schroeder D, Gillanders L, Mahr K, Hill GL. Effects of immedi- quences, and treatment of enteral feed intolerance during criti- ate postoperative enteral nutrition on body composition, mus- cal illness. JPEN J Parenter Enteral Nutr 2015;39:441–448. cle function, and wound healing. JPEN J Parenter Enteral Nutr 8. Mentec H, Dupont H, Bocchetti M, Cani P, Ponche F, Bleich- 1991;15:376–383. ner G. Upper digestive intolerance during enteral nutrition in 18. Singer P, Blaser AR, Berger MM, Alhazzani W, Calder PC, Casa- critically ill patients: frequency, risk factors, and complications. er MP, et al. ESPEN guideline on clinical nutrition in the inten- Crit Care Med 2001;29:1955–1961. sive care unit. Clin Nutr 2019;38:48–79. 9. Oertel MF, Hauenschild A, Gruenschlaeger J, Mueller B, Schar- 19. Wise KR, Akopov VA, Williams BR, Ido MS, Leeper KV, brodt W, Boeker DK. Parenteral and enteral nutrition in the Dressler DD. Hospitalists and intensivists in the medical ICU: a management of neurosurgical patients in the intensive care unit. prospective observational study comparing mortality and length J Clin Neurosci 2009;16:1161–1167. of stay between two staffing models. J Hosp Med 2012;7:183– 10. Ott L, Young B, Phillips R, McClain C, Adams L, Dempsey R, 189. et al. Altered gastric emptying in the head-injured patient: rela- 20. Yatabe T. Strategies for optimal calorie administration in critical- tionship to feeding intolerance. J Neurosurg 1991;74:738–742. ly ill patients. J Intensive Care 2019;7:15. 11. Patejdl R, Kastner M, Kolbaske S, Wittstock M. Clinical nutri- 21. Yuan F, Yang F, Zhang W, Jia Y, Ma Y, Qu Y, et al. Optimizing early tion and gastrointestinal dysfunction in critically ill stroke pa- enteral nutrition in severe stroke (OPENS): protocol for a multi- tients. Neurol Res 2017;39:959–964. centre randomized controlled trial. BMC Neurol 2019;19:24. 12. Picolo MF, Lago AF, Menegueti MG, Nicolini EA, Basile-Filho 22. Zhu CS, Grandhi R, Patterson TT, Nicholson SE. A review of A, Nunes AA, et al. Harris-Benedict equation and resting ener- traumatic brain injury and the gut microbiome: insights into gy expenditure estimates in critically ill ventilator patients. Am J novel mechanisms of secondary brain injury and promising tar- Crit Care 2016;25:e21–29. gets for neuroprotection. Brain Sci 2018;8:113. 13. Reintam A, Parm P, Kitus R, Kern H, Starkopf J. Gastrointesti-

44 www.e-jnic.org eISSN 2635-5280 Original Article J Neurointensive Care 2019;2(2):45-51 https://doi.org/10.32587/jnic.2019.00171

Endovascular Treatment of Intracranial Aneurysms Associated with Brain Arteriovenous Malformations Boeun Lee1, Jae Whan Lee1, Keun Young Park1, Dong Joon Kim2, Byung Moon Kim2, Joonho Chung1,3 1Department of Neurosurgery, Severance Hospital, Yonsei University College of Medicine, Seoul, Korea 2Department of Radiology, Severance Hospital, Yonsei University College of Medicine, Seoul, Korea 3Severance Institute for Vascular and Metabolic Research, Yonsei University College of Medicine, Seoul, Korea

Received: June 12, 2019 Accepted: August 19, 2019 Objective Brain arteriovenous malformation (AVM) is frequently associated with the presence of intracrani- Corresponding Author: al aneurysms, which increase hemorrhage rates resulting in unfavorable outcomes. We report our Joonho Chung, M.D., Ph.D. experience with endovascular treatment (EVT) of intracranial aneurysms associated with AVM. Department of Neurosurgery, Severance Hospital Severance Methods Institute for Vascular and Between March 2010 and February 2017, we treated 274 patients with AVMs. Among them, 27 Metabolic Research Yonsei (9.9%) patients with intracranial aneurysms associated with AVM were treated endovascularly. University College of Medicine Clinical and radiographic data for these patients were retrospectively reviewed. 50-1, Yonsei-ro, Seodaemun-gu, Results Seoul 03722, Korea We treated 32 aneurysms (10 intranidal and 22 proximal aneurysms) in 27 patients. The most fre- Tel: +82-2-2228-2150 Fax: +82-2-393-9979 quent presenting symptoms were hemorrhage (59.3%) caused by an AVM nidus in 7 cases and by E-mail: [email protected] aneurysm rupture in 9 cases. Of 22 proximal flow-related aneurysms, we performed a simple coiling technique in 16 aneurysms and 6 parent artery occlusions. For 10 intranidal aneurysms, we performed targeted embolization with liquid embolic materials, such as n-Butyl cyanoacrylate or Onyx. There were three procedure-related complications and only one patient suffered transient hemiparesis. Favorable outcome (modified Rankin Scale, mRS 0-2) was achieved in 18 (66.7%) patients and unfavorable outcome (mRS 3-6) in 9 (33.3%) patients, including one death. Conclusion When facing intracranial aneurysms associated with AVMs, EVT might be a useful treatment option for proximal and intranidal aneurysms.

Keywords: Brain arteriovenous malformation; Endovascular treatment; Flow-related aneurysm; Intranidal aneurysm

INTRODUCTION cular lesions that consist of abnormal connections between cerebral arteries and veins without an associated capillary bed. AVM Brain arteriovenous malformations (AVMs) are complex vas- patients can present with hemorrhage, seizure, neurological defi-

Copyright © 2019 The Korean Neurointensive Care Society This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited. www.e-jnic.org 45 EVT for aneurysms associated with brain AVMs Boeun Lee et al. cit, or headache, and AVMs often remain unrecognized until they Envoy (Codman Neurovascular, Miami Lakes, FL, USA) was posi- become symptomatic. With a prevalence that is likely less than tioned in the ICA or VA. Intravenous systemic heparin (50 U/kg) 10 per 100000, they are relatively rare, but can nevertheless have was administered while placing the guiding catheter. A microcath- devastating consequences for patients who harbor them. Once eter was advanced super-selectively to the AVM’s small feeding ar- an AVM ruptures, the risk of additional hemorrhage over the first teries. For intranidal aneurysms, n-Butyl cyanoacrylate (NBCA) year increases to 6% to 18%3-7), before returning to 2% to 4% or Onyx were used to occlude the aneurysms and the nidus simul- per year, which is the hemorrhage rate of previously diagnosed taneously with the intention of reducing hemorrhagic complica- unruptured AVMs. An increased risk for hemorrhage has been tions that could arise during the procedure interval. For proximal reported in previous studies when an AVM is associated with an flow-related aneurysms, a variety of techniques were performed as aneurysm. The association of intracranial aneurysms with AVMs needed: simple coiling, multiple-microcatheter, balloon remodel- has been well-established. Depending on the case series, aneu- ing, or the stent-assisted technique as a reconstruction technique, rysms are present in 3-58% of patients with AVMs8,10,13,16). and deconstruction techniques, such as parent artery occlusion. Over the last decade, outstanding progress in endovascular Procedure-related events were defined as any events during the techniques targeting intracerebral aneurysm repairs has high- procedure that were symptomatic or asymptomatic. Clinical out- lighted the need for preventive AVMs treatments in the context comes were assessed using the modified Rankin Scale (mRS) of associated aneurysms, which is a growing issue. Treating either and evaluated at the time of discharge and out-patient follow-up AVM or aneurysms is challenging. Some experts have recom- by an independent investigator who was not involved in patient mended early treatment of aneurysms, but some have recom- care. Angiographic and clinical follow-ups were performed at 6 to mended treatment of the symptomatic lesion first. In terms of 12 months in all patients. treating both lesions simultaneously, EVT may be preferable. Thus, we report our experience with endovascular treatment RESULTS (EVT) for intracranial aneurysms associated with AVMs. Demographic and morphologic characteristics of the 27 pa- MATERIALS AND METHODS tients with intracranial aneurysms associated with AVMs are summarized in Table 1. There were 9 women (33.3%) and 18 This retrospective study was approved by the Institutional Re- view Board and the requirement for informed consent was Table 1. Demographic and morphological characteristics for 27 patients waived. Between March 2010 and February 2017, we treated 274 with cerebral aneurysms associated with arteriovenous malformations. patients with brain AVMs. Among them, a total of 27 (9.9%) pa- Aneurysms associated with arteriovenous malformations (N = 27) tients who met the following criteria were included in this study: Age at presentation, y (mean±SD) 38.7±12.6 1) patients diagnosed with intracranial aneurysms associated Female, n (%) 9 (33.3) with AVMs by computed tomography (CT), magnetic resonance Presentations, n (%) imaging (MRI), or digital subtraction angiography (DSA); and Hemorrhage 16 (59.3) 2) aneurysms that were treated by EVT. Cerebral aneurysms as- Non-hemorrhage 11 (40.7) sociated with AVMs were defined and divided into two catego- Locations, n (%) ries; intranidal aneurysms (i.e., aneurysms located within or in Supratentorial 23 (85.2) the immediate vicinity of the AVM nidus) and proximal flow-re- Infratentorial 4 (14.8) lated aneurysms (i.e., aneurysms arising along the course of arter- AVM size, n (%) ies that eventually supply the AVM). Clinical and radiographic < 3 cm 7 (25.9) 3-6 cm 17 (63.0) data for these patients were retrospectively reviewed. All patients > 6 cm 3 (11.1) underwent DSA with 3-dimensional rotational angiography to Venous drainage pattern, n (%) evaluate vascular architecture and the location of their lesions. Deep venous drainage only 4 (14.8) Treatment decisions were made by agreement of microvascular Any deep venous drainage 5 (18.5) neurosurgeons and neurointerventionists. All procedures were Associated arterial aneurysm, n (%) performed under general anesthesia with a standard approach Intranidal 5 (18.5) from the common femoral artery. A 6-Fr femoral sheath was insert- Proximal 17 (63.0) ed into the common femoral artery. A 5-Fr or 6-Fr guiding catheter Intranidal & Proximal 5 (18.5)

46 www.e-jnic.org EVT for aneurysms associated with brain AVMs Boeun Lee et al. men (66.7%). Patient ages ranged from 18 to 61 years, with a occlusion of the AVM nidus let us perform additional microsur- mean age of 38.7 years. Four (14.8%) AVMs were located in an gical nidus removal. The nidus filled with Onyx guided us to de- infratentorial location and 23 (85.2%) were supratentorial. Six- lineate the margin of the AVM nidus and helped us remove the teen (59.3%) patients initially presented with hemorrhage. The nidus completely (Fig. 1E). At one-month follow-up, a CT scan remaining 11 patients (40.7%) initially presented with symptoms revealed a post-treatment coil mass and empty space of the nidus or signs unrelated to hemorrhage. Among the patients who pre- (Fig. 1F). However, the patient’s clinical outcome was an unfa- sented with hemorrhage, bleeding focus was determined by the vorable, mRS of 4 at her discharge. location and distribution of the hematoma on enhanced CT scan and this was correlated with DSA. Bleeding was caused by the DISCUSSION AVM nidus in 7 cases and by aneurysm rupture in 9 cases. Eleven patients had no history of hemorrhage. It has been established that hemorrhage is the most devastat- Five patients had AVMs with multiple arterial aneurysms, thus ing AVM complication and the annual risk of rupture, regardless there was a total of 32 aneurysms in 27 patients. The associated an- of its characteristics, is 4% on average1). However, AVM may have eurysms were divided into 22 (68.8%) proximal flow-related aneu- different behaviors due to the following factors: elapsed time af- rysms and 10 (31.3%) intranidal aneurysms. For 22 proximal ter diagnosis, presence of deep venous drainage, associated aneu- flow-related aneurysms, we performed a simple coiling technique in rysms, AVM location, size, male sex, age, and other variables spe- 16 aneurysms and parent artery occlusion (PAO) in 6. For 10 in- cific to each patient. The hemorrhage risk if there are associated tranidal aneurysms, we performed targeted embolization with liq- aneurysms has been reported as 7% per year and most hemor- uid embolic materials (i.e., NBCA or Onyx). Ten patients under- rhages are equally distributed between aneurysms and AVMs, went microsurgical nidus removal and 12 patients underwent gam- which means an aneurysm may be the site of rupture1). The re- ma knife surgery after embolization. The other five AVMs were ported rate of intracranial aneurysms associated with AVMs var- completely occluded by EVT simultaneously with the aneurysms. ies widely in previously published data sets, depending on the There were three procedure-related complications and only one definition of what represents an associated aneurysm. Intracrani- patient suffered transient hemiparesis. Favorable outcome (mRS al aneurysms associated with AVMs may be caused by hemody- 0-2) was achieved in 18 (69.2%) and unfavorable outcome (mRS namic stress related to the AVM or they may be congenital14,20). 3-6) in 8 (30.8%) patients, including one death at discharge. The hemodynamic stress secondary to increased AVM blood flow plays a significant role in provoking this coexistence of le- Case Illustrations sions. Patients with subarachnoid hemorrhage, intraparenchymal Case 1 hemorrhage, or both and an angiographically demonstrated A 54-year-old female was admitted to our institute with deep AVM with an associated aneurysm, should have therapy directed drowsy consciousness and right hemiparesis. Brain CT scan re- first toward the lesion that is suspected to be the source of the vealed diffuse subarachnoid hemorrhage in the basal cistern, hemorrhage. In the present study, 32 aneurysms associated with quadrigeminal cistern, and intraparenchymal hemorrhage in the AVMs in 27 patients were treated by EVT with a relatively low left pons-to-cerebellum and the left occipital lobe (Fig. 1A). DSA complication rate. Twenty-two proximal flow-related aneurysms revealed a proximal flow-related aneurysm in the P3 segment of were treated by simple coiling or PAO, whereas 10 intranidal an- the left posterior cerebral artery (PCA) and a 2.7 x 2.5 mm sized eurysms were treated by targeted embolization with liquid em- AVM nidus in the left occipital lobe fed by the left middle cere- bolic materials. Favorable outcome was achieved in 70% of 27 bral artery (MCA), left anterior cerebral artery (ACA), and the patients including 16 patients with hemorrhage. EVT may be left PCA (Fig. 1B). It drained into cortical veins connected to the considered a useful treatment for aneurysms associated with superior sagittal sinus. An intraparenchymal hemorrhage in the AVMs. However, our results should not alter the fundamental left pons implied a P3 segment aneurysm rupture and the aneu- role of EVT in the management of AMVs, especially as an adju- rysm showed a fusiform-like wide-neck involving 360-degrees of vant to microsurgery and radiosurgery. the PCA. PAO with coiling was performed to occlude the rup- The vascular architecture of an AVM is an important consider- tured aneurysm and one of the feeding arteries into the AVM ni- ation when planning treatment. Along with a thorough clinical dus (Fig. 1C). Then, the left MCA and ACA were super-selected examination, all AVM patients need detailed preoperative radio- by Apollo detachable-tip microcatheters (Medtronic, Irvine, graphic clarification of their AVM’s anatomy, architecture, and CA) and Onyx embolization was performed (Fig. 1D). Partial associated aneurysms obtained with MRI and high-frame rate www.e-jnic.org 47 EVT for aneurysms associated with brain AVMs Boeun Lee et al.

C D

Fig. 1. A 54-year-old female was admitted to our institute with deep drowsy consciousness and right hemiparesis. (A) Brain computed tomography scan revealed diffuse subarachnoid hemorrhage in the basal cistern and quadrigeminal cistern and intraparenchymal hemorrhage in the left pons-to-cerebellum and left occipital lobe. (B) Digital subtraction angiography revealed a proximal flow-related aneurysm on the P3 segment of the left posterior cerebral artery (PCA) and a 2.7 x 2.5 mm sized arteriovenous malformation (AVM) nidus in the left occipital lobe fed by the left middle cerebral artery (MCA), left anterior cerebral artery (ACA), and the left PCA. It drained into the cortical veins-to-superior sagittal sinus. (C) Parent artery occlusion with coiling was performed to occlude the ruptured aneurysm and one of the feeding arteries to the AVM nidus. (D) Left MCA and ACA were super-selected by detachable-tip microcatheters and Onyx embolization was performed. (E) Partial occlusion of the AVM nidus let us perform additional microsurgical nidus removal. The nidus filled with Onyx guided us to delineate the margin of the AVM nidus and helped us remove the nidus completely. (F) One-months follow-up CT scan revealed a post-treatment coil mass and empty space of the nidus. White circle indicates a ruptured aneurysm associated AVM. Black circle indicates the AVM nidus. White arrow indicates the P3 segment occlusion by coiling. Rectangle indicates the partially occluded nidus by Onyx embolization. (Continued to the next page)

48 www.e-jnic.org EVT for aneurysms associated with brain AVMs Boeun Lee et al.

E F

Fig. 1. (Continued from the previous page) A 54-year-old female was admitted to our institute with deep drowsy consciousness and right hemiparesis. (A) Brain computed tomography scan revealed diffuse subarachnoid hemorrhage in the basal cistern and quadrigeminal cistern and intraparenchymal hemorrhage in the left pons-to-cerebellum and left occipital lobe. (B) Digital subtraction angiography revealed a proximal flow- related aneurysm on the P3 segment of the left posterior cerebral artery (PCA) and a 2.7 x 2.5 mm sized arteriovenous malformation (AVM) nidus in the left occipital lobe fed by the left middle cerebral artery (MCA), left anterior cerebral artery (ACA), and the left PCA. It drained into the cortical veins-to-superior sagittal sinus. (C) Parent artery occlusion with coiling was performed to occlude the ruptured aneurysm and one of the feeding arteries to the AVM nidus. (D) Left MCA and ACA were super-selected by detachable-tip microcatheters and Onyx embolization was performed. (E) Partial occlusion of the AVM nidus let us perform additional microsurgical nidus removal. The nidus filled with Onyx guided us to delineate the margin of the AVM nidus and helped us remove the nidus completely. (F) One-months follow-up CT scan revealed a post- treatment coil mass and empty space of the nidus. White circle indicates a ruptured aneurysm associated AVM. Black circle indicates the AVM nidus. White arrow indicates the P3 segment occlusion by coiling. Rectangle indicates the partially occluded nidus by Onyx embolization.

DSA. The angiography must delineate the characteristics of the eloquent18). Recently, AVM EVT has been used as a curative op- AVM, including its arterial feeders, venous drainage, internal an- tion that is mainly done in an adjuvant manner, although these ad- gioarchitecture, and associated vascular lesions, and it must de- vances have made stand-alone embolization a possibility in many fine the collateral circulation and venous drainage pathways of patients. EVT of aneurysm-associated AVMs is beneficial before the normal brain tissue. Treatment decisions should be consid- microsurgery or radiosurgery. Thus, physician experience with ered to maximize clinical outcomes and to minimize risks, and treating AVMs includes being capable of performing microsurgery can be made regarding the best management approach by com- and radiosurgery in addition to EVT, with concordant facilities. paring the natural history of the lesion with known associated in- Anatomic cures can be accomplished by a combination of two tervention-related morbidity and mortality11). However, treat- or more modalities, one of which is usually EVT. A multimodality ment guidelines for brain AVMs are still controversial despite treatment plan can minimize risk and improve patient outcomes. outstanding advances in diagnostic and therapeutic resources. As necessary integration of different treatment modalities, such as Driven largely by innovations in devices and techniques, the microsurgery, radiosurgery, and EVT, increases cure rates tend to field of neurointerventional surgery is constantly changing. As be higher. Especially, explosive progress in endovascular devices these changes have occurred across the board, EVT for brain and techniques provides better quality options for EVT performed AVMs has seen corresponding advances. Innovations defined by alone as well as an adjuvant treatment before microsurgery or a the improvements seen in newer and better embolic materials, concurrent therapy with radiosurgery. Preoperative embolization microcatheters (flow-directed microcatheters, Onyx comparable reduced mean nidus volume by 74% to 84%19). However, pre- or and dual lumen balloon microcatheters, and detachable-tip mi- post-radiosurgery embolization to where less likely respond to ra- crocatheters), and other devices (distal access intermediate cath- diation, to high-flow fistulas or to associated aneurysms consists of eters) have resulted in safer and more effective embolization of conflicting evidence. As physicians have gained more experience cerebral AVMs2). EVT can be applied to AVM treatment in vari- with EVT, curative embolization rate is increasing12,15,17). ous ways: preoperative, intraoperative, pre-radiosurgery, curative, There are several limitations to this study. This study was a ret- palliative, or post-radiosurgery9,18). rospective study with a small case series so that the effects of pos- Thus, goals should be defined prior to each procedure. Addi- sible selection bias cannot be excluded. In addition, it was diffi- tionally, traumatic super-selective microcatheterization is a key. In cult to have a true control group for evaluating the safety and effi- the setting of AVM EVT, all AVM locations should be regarded as cacy of EVT with associated aneurysms. Furthermore, a longer www.e-jnic.org 49 EVT for aneurysms associated with brain AVMs Boeun Lee et al. follow-up period is needed to evaluate the procedure’s true safety 5. Flores BC, Klinger DR, Rickert KL, Barnett SL, Welch BG, White and efficacy. JA, et al. Management of intracranial aneurysms associated with arteriovenous malformations. Neurosurg Focus 2014;37:E11. CONCLUSION 6. Garcia-Monaco R, Rodesch G, Alvarez H, Iizuka Y, Hui F, Las- jaunias P. Pseudoaneurysms within ruptured intracranial arte- When facing intracranial aneurysms associated with AVMs, riovenous malformations: diagnosis and early endovascular EVT might be a useful treatment option for proximal flow-relat- management. AJNR Am J Neuroradiol 1993;14:315–321. ed and intranidal aneurysms. This procedure could cure aneu- 7. Gross BA, Du R. Natural history of cerebral arteriovenous mal- rysms and AVMs simultaneously, or be good for adjuvant treat- formations: a meta analysis. J Neurosurg 2013;118:437–443. ment before microsurgical nidus removal or gamma knife surgery 8. Lv X, Wu Z, Li Y, Yang X, Jiang C, Sun Y, et al. Endovascular for AVMs. treatment of cerebral aneurysms associated with arteriovenous malformatioins. Eur J Radiol 2012;81:1296–1298. NOTES 9. Martin NA, Khanna R, Doberstein C, Bentson J. Therapeutic embolization of arteriovenous malformations: the case for and Conflict of interest against. Clin Neurosurg 2000;46:295–318. We declare that we have no conflict of interest. 10. Meisel HJ, Mansmann U, Alvarez H, Rodesch G, Brock M, Las- jaunias P. Cerebral arteriovenous malformations and associated Ethical approval aneurysms: analysis of 305 cases from a series of 662 patients. all human subjects study procedures were in accordance with Neurosurgery 2000;46:793–802. the ethical standards of our Institutional Review Board and the 11. Novakovic RL, Lazzaro MA, Castonguay AC, Zaidat OO. The 1964 Declaration of Helsinki and its later amendments or com- diagnosis and management of brain arteriovenous malforma- parable ethical standards. tions. Neurol Clin 2013;31:749–763. 12. Pierot L, Cognard C, Herbreteau D, Fransen H, van Rooij WJ, Informed consent Boccardi E, et al. Endovascular treatment of brain arteriovenous In this retrospective study, the requirement for informed con- malformations using a liquid embolic agent: results of a prospec- sent was waived. tive, multicenter study (BRAVO). Eur Radiol 2013;23:2838– 2845. Funding 13. Piotin M, Ross IB, Weill A, Kothimbakam R, Moret J. Intracra- No funding was received for this study. nial arterial aneurysms associated with arteriovenous malformations: endovascular treatment. Radiology 2001;220:506–513. REFERENCES 14. Redekop G, TerBrugge K, Montanera W, Willinsky R. Arterial aneurysms associated with cerebral arteriovenous malforma- 1. Brown RD, Wiebers DO, Forbes GS. Unruptured intracranial tions: classification, incidence, and risk of hemorrhage. J Neuro- aneurysms and arteriovenous malformations: frequency of in- surg 1998;89:539–546. tracranial hemorrhage and relationship of lesions. J Neurosurg 15. Renieri L, Consoli A, Scarpini G, Grazzini G, Nappini S, Mangi- 1990;73:859–863. afico S. Double arterial catheterization technique for emboliza- 2. Crowley RW, Ducruet AF, McDougall CG, Albuquerque FC. tion of brain arteriovenous malformations with onyx. Neuro- Endovascular advances for brain arteriovenous malformations. surgery 2013;72:92–98. Neurosurgery 2014;74:S74–S82. 16. Stapf C, Mast H, Sciacca RR, Choi JH, Khaw AV, Connolly ES, 3. da Costa L, Thines L, Dehdashti AR, Wallace MC, Willinsky et al. Predictors of hemorrhage in patients with untreated brain RA, Tymianski M, et al. Managementandclinical outcome of arteriovenous malformation. Neurology 2006;66:1350–1355. posterior fossa arteriovenous malformations: report on a sin- 17. Strauss I, Frolov V, Buchbut D, Gonen L, Maimon S. Critical ap- gle-centre 15-year experience. J Neurol Neurosurg Psychiatry praisal of endovascular treatment of brain arteriovenous malfor- 2009;80:376–379. mation using Onyx in a series of 92 consecutive patients. Acta 4. Elhammady MS, Aziz-Sultan MA, Heros RC. The manage- Neurochir (Wien) 2013;155:611–617. ment of cerebral arteriovenous malformations associated with 18. Valavanis A, Yasargil MG. The endovascular treatment of brain aneurysms. World Neurosurg 2013;80:e123–9. arteriovenous malformations. Adv Tech Stand Neurosurg 1998;

50 www.e-jnic.org EVT for aneurysms associated with brain AVMs Boeun Lee et al.

24:131–214. 1304. 19. van Rooij WJ, Jacobs S, Sluzewski M, van der Pol B, Beute GN, 20. Westphal M, Grzyska U. Clinical significance of pedicle aneu- Sprengers ME. Curative embolization of brain arteriovenous rysms on feeding vessels, especially those located in infratentori- malformations with onyx: patient selection, embolization tech- al arteriovenous malformations. J Neurosurg 2000;92:995– nique, and results. AJNR Am J Neuroradiol 2012;33:1299– 1001.

www.e-jnic.org 51 eISSN 2635-5280 Original Article J Neurointensive Care 2019;2(2):52-57 https://doi.org/10.32587/jnic.2019.00178

Preliminary Experience of Lvis Blue in the Internal Carotid Artery for The Treatment of Wide-Necked Intracranial Aneurysms Jeong Hwa Kim1, Chang Ki Jang1, Jae Whan Lee1, Keun Young Park1, Joonho Chung1,2 1Department of Neurosurgery, Severance Hospital, Yonsei University College of Medicine, Seoul, Korea 2Severance Institute for Vascular and Metabolic Research, Yonsei University College of Medicine, Seoul, Korea

Received: August 8, 2019 Accepted: August 23, 2019 Objective A new Low-profile Visualized Intraluminal Support device (LVIS Blue) is a braided stent that pro- Corresponding Author: vides a higher degree of metal coverage and increasingly used for the stent-assisted coiling of intra- Joonho Chung, M.D., Ph.D. cranial aneurysms. In the present study, our preliminary experience of using LVIS Blue in the in- Department of Neurosurgery, ternal carotid artery (ICA) for the treatment of wide-necked intracranial aneurysms was reported. Stroke Center, Severance Hospital Severance Institute for Vascular Methods and Metabolic Research Yonsei This retrospective review included patients with ICA aneurysms prospectively collected from our University College of Medicine database who were treated using LVIS Blue at our institution from September 2017 to December 50-1, Yonsei-ro, Seodaemun-gu, 2018. Angiographic results, clinical outcomes, and technical issues were evaluated. Seoul, 03722, Korea Results Tel: +82-2-2228-2150 Fax: +82-2-393-9979 A total of 22 patients with unruptured wide-necked ICA aneurysms were treated using LVIS Blue. E-mail: ns.joonho.chung@gmail. Seven aneurysms were located in the superior hypophyseal artery, 4 in posterior communicating com artery, 4 in ICA dorsal wall, 3 in ICA bifurcation, 2 in ophthalmic artery, 1 in ICA ventral wall, and 1 in ICA lateral wall. During the procedure, suboptimal positioning and foreshortening of the stent occurred in 1 patient each. Immediate postembolization angiography showed complete obliteration in 13 patients (59.1%), residual neck in 4 (18.2%), and residual aneurysm in 5 patients (22.7%). Follow-up angiographic results showed complete obliteration in 19 patients (86.4%). In-stent stenosis (> 50% stenosis) was not observed on follow-up angiography. In addition, morbidity or mortality did not occur. Conclusion The LVIS Blue may be feasible and safe for stent-assisted coiling in ICA aneurysms. Foreshorten- ing, suboptimal deployment, and difficulty to open proximal ends may be considerable technical issues in the tortuous ICA.

Keywords: Intracranial aneurysm; Wide-necked aneurysm; LVIS Blue; Stent-assisted coiling

52 www.e-jnic.org LVIS Blue stent in the ICA Jeong Hwa Kim et al.

INTRODUCTION a standard approach from the common femoral artery. A 6-Fr or 7-Fr guiding catheter was positioned in the ICA. Intravenous Stent-assisted coil embolization (SAC) has been accepted as a systemic heparin (50 U/kg) was administered while placing the feasible option for occluding the wide-necked complex aneu- guiding catheter. An activated clotting time was maintained at rysms with preservation of the parent artery, associated with flow 2-2.5 times that of baseline during endovascular treatment. A remodeling which promote the complete thrombosis of the coiled Headway-21 (MicroVention) or a Prowler Select Plus (Codman aneurysm3,5,12,15). The Low-profile Visualized Intraluminal Sup- Neurovascular) were used to deliver the LVIS Blue stent. The port device (LVIS; MicroVention, Inc., Tustin, CA) is a self-ex- jailing or semi-jailing techniques were primarily performed, and panding, compatible, and widely used closed cell-designed stent. if failed, the through the strut (cell-through) technique was im- The small cell structure (0.8 mm) and high metal coverage (ap- plemented. proximately 23%) allow the achievement of greater complete or Patients took a daily dose of 75 mg of clopidogrel and 100 mg near-complete occlusion rate of intracranial aneurysm compared of aspirin for more than 7 days before the SAC procedure. For pa- with the currently used microstents8). The LVIS Blue, a new mod- tients who had insufficient premedication or when the SAC was ification of LVIS, is a braided stent that provides a higher degree planned immediately after the diagnostic DSA, loading dose of as- of metal coverage (approximately 28%) and is increasingly used pirin and clopidogrel 300 mg each was administered the day be- for SAC. The LVIS Blue showed improved wall apposition and fore treatment. Platelet function testing was routinely performed aneurysm neck coverage11) and has a good crossing profile for mi- for all patients. After the procedure, patients were prescribed 75 crocatheters, better wall apposition and less perforator coverage mg of clopidogrel daily for 3 months and 100 mg of aspirin daily than a flow redirection endoluminal device (FRED)14). for at least 12 months. However, clinical data regarding the safety and effectiveness of LVIS Blue are insufficient. Thus, in the present study, preliminary Clinical and Radiographic Outcomes experience of using LVIS Blue in the internal carotid artery (ICA) Angiographic and clinical outcomes were retrospectively re- for the treatment of wide-necked intracranial aneurysms was reviewed. Follow-up angiography was performed in all patients be- ported. tween 6 and 24 months (mean 13.4 months) after the procedure, and clinical follow-up was performed for 6 – 24 months (mean METHODS 16.5 months). The angiographic results were evaluated by 2 independent investigators and categorized according to the modified Patient Selection Raymond-Roy classification (Class I: complete obliteration, Class This retrospective study was approved by our Institutional Re- II: residual neck, Class IIIa: residual aneurysm, contrast within coil view Board, and the requirement for informed consent was interstices, Class IIIb: residual aneurysm, contrast along aneurysm waived. Endovascular treatment of 197 aneurysms (176 unrup- wall). Clinical outcomes were assessed using the modified Rankin tured and 21 ruptured) in 129 patients was performed from Sep- Scale (mRS) and evaluated at the time of discharge and out-patient tember 2017 to December 2018. Among 114 aneurysms treated follow-up by an independent investigator who was not involved in with SAC, 22 cases (all wide-necked unruptured aneurysms in the patient care. In-stent stenosis was defined as > 50% stenosis on ICA) were treated with LVIS Blue stents. follow-up angiography. After the procedure, skull X-ray was per- All 22 patients underwent digital subtraction angiography formed and diffusion magnetic resonance imaging (diffusion (DSA) and rotational angiography (Philips Allura FD20 Clarity MRI) was conducted at postoperative day 1. Procedural complica- System, Philips Medical Systems, Best, The Netherlands) with tion was defined as acute infarct or hemorrhage that impaired the three-dimensional reconstructions to characterize aneurysm and clinical outcome and required prolonged hospital care. parent artery anatomy (Allura 3D-RA workstation, Philips Medi- cal Systems). Based on DSA results, the treatment decision was RESULTS made by consensus among microvascular neurosurgeons and neurointerventionists. Wide-necked aneurysms were defined as Table 1 shows the data and outcomes of the 22 patients (5 having a neck width ≥ 4 mm or a dome-to-neck ratio < 2. males and 17 females, age range 37 – 77 years, mean 54.1 years). No patient had a history of subarachnoid hemorrhage. All aneu- Endovascular Treatment rysms were located at the intracranial portion of the ICA with All procedures were performed under general anesthesia with wide-necked saccular type morphology. The most common site www.e-jnic.org 53 LVIS Blue stent in the ICA Jeong Hwa Kim et al.

Table 1. Baseline characteristics and outcomes of 22 patients with 22 graphic results showed complete obliteration in 19 patients aneurysms in the internal carotid artery (86.4%). All 4 patients with residual necks and 2 patients with re- N=22 sidual aneurysms on the immediate postembolization angiogra- Age (year, mean±SD) 54.1±11.1 phy improved and showed complete obliteration on the follow-up Female (n, %) 17 (77.3) angiography. In-stent stenosis (> 50% stenosis) was not observed Hypertension (n, %) 9 (40.9) on follow-up angiography. With respect to clinical outcomes, Diabetes (n, %) 5 (22.7) morbidity or mortality did not occur. Dyslipidemia (n, %) 3 (13.6) Smoking (n, %) 3 (13.6) Aneurysm size (mm, mean±SD) 6.24±2.37 DISCUSSION Neck size (mm, mean±SD) 4.54±1.56 Location (n, %) In this preliminary study, LVIS Blue was used for the treatment Superior hypophyseal artery 7 (31.8) of patients with unruptured, wide-necked saccular aneurysm in Ophthalmic artery 2 (9.1) the ICA. Angiographic and clinical outcomes showed the LVIS Dorsal wall 4 (18.2) Blue is a feasible and safe device with favorable occlusion rate. Ventral wall 1 (4.5) Several technical issues related to the mechanical characteristics of Lateral wall 1 (4.5) the LVIS stent were also elucidated during the procedure, as re- Posterior communicating artery 4 (18.2) ported in previous studies. ICA bifurcation 3 (13.6) In recent years, the treatment of intracranial aneurysms focused Technical considerations of stents (n, %) on the reconstruction of the parent artery and aneurysmal neck. Suboptimal positioning 1 (4.5) Foreshortening 1 (4.5) Decrease of intra-aneurysmal flow can be achieved by inserting Initial angiographic results (n, %) stents or flow diverters, with or without coil compaction. Stents are Complete obliteration 13 (59.1) believed to redirect the flow in the parent artery and slow down the Residual Neck 4 (18.2) aneurysmal flow leading to thrombosis and consequently stabiliz- Residual Aneurysm 5 (22.7) ing the aneurysm1). For wide-necked aneurysms, preservation of Follow-up angiographic results (n, %) the parent artery is the technical challenge in the endovascular pro- Complete obliteration 19 (86.4) cedure. The use of self-expandable stents or flow diverters has pro- Residual Neck 0 (0) vided a significant advance in the treatment with high rates of com- Residual Aneurysm 3 (3) plete occlusion and low rates of recurrence at follow-up9). Clinical outcomes (n, %) The LVIS is a hybrid closed-cell stent made of nickel titanium Morbidity 0 (0) (nitinol) with flared ends and 2 radiopaque helices of tantalum Mortality 0 (0) strands to assist full-length visualization. The device has a high ICA: internal carotid artery; SD: standard deviation. metal-to-surface coverage intended to help promote neo-endo- thelization. The sliding design of its cells are feasible for crossing was superior hypophyseal artery (n=7) followed by posterior the struts with a microcatheter. According to the systematic re- communicating artery (n= 4), dorsal wall (n= 4), ICA bifurcation view of the LVIS device by Zhang et al., the immediate angio- (n= 3), ophthalmic artery (n= 2), ventral wall (n= 1), and lateral graphic results in the literature review demonstrated relatively wall (n= 1). Aneurysm size ranged from 4.94–11.31 mm (mean lower rates of complete occlusion, progression to total occlusion 6.24±2.37 mm) and neck diameter from 2.86–7.76 mm (mean was impressive with a high rate of complete occlusion (84.3%), 4.54±1.56 mm). Complications related to delivering or deploy- and the thromboembolic complications were negligible, with a ing the LVIS Blue stent did not occur. However, suboptimal posi- rate of 4.9%18). tioning (Fig. 1) and foreshortening (Fig. 2) of the stent occurred Regarding technical issues, poor wall apposition due to incom- in 1 patient each. plete segmental expansion after application of the LVIS was re- At the end of the procedure, parent vessel patency was con- ported in several studies2,4,7,16). Cho et al. reported the poor wall firmed with no evidence of associated flow decrease in the cov- apposition occurred in 9.1% of patients, with consumption of ered vessel. Immediate postembolization angiography showed twist or distortion of LVIS stent at the tortuous segment or acute complete obliteration in 13 patients (59.1%), residual neck in 4 curve of the ICA4). (18.2%), and residual aneurysm in 5 (22.7%). Follow-up angio- The new low-profile visualized intraluminal support device,

54 www.e-jnic.org LVIS Blue stent in the ICA Jeong Hwa Kim et al.

A B

Fig. 1. A 63-year-old female with a left posterior communicating artery aneurysm. The aneurysm is a mitten-shaped saccular sac with a daughter sac. The dome-to-neck ratio is 1.43 with the height of 4.6 mm and neck size of 3.2 mm. (A) A 4.5 x 23 mm LVIS Blue stent was deployed and the distal tips of LVIS Blue stent barely covered the aneurysm neck due to an unstable delivery of microcatheter in the ICA curve. (B) Despite suboptimal positioning of the stent, immediate postembolization and follow-up angiographic results showed complete obliteration of the aneurysm. Black arrows indicate distal tips of the stent.

A B

Fig. 2. A 47-year-old male with a dorsal wall aneurysm in the internal carotid artery. The dome-to-neck ratio is 1.35 with a height of 6.9 mm and neck size of 5.1 mm. (A) A 4.5 x 23 mm LVIS Blue was deployed and the jailing technique for coiling was performed. During the procedure, the distal tips of the LVIS Blue showed foreshortening due to unconstrained segment of the aneurysm neck. (B) Immediate postembolization angiography showed residual neck of the aneurysm. Black arrows indicate initial positioning of the stent distal tips. A white arrow indicates final positioning of the stent distal tips after foreshortening. www.e-jnic.org 55 LVIS Blue stent in the ICA Jeong Hwa Kim et al.

LVIS Blue, is a stent with 16 nitinol single wire-braids which is to its unconstrained state, generally resulting in a lower metal cov- compatible with a 0.021-inch microcatheter and recommended erage surface area. This observation correlated with the previously for a parent vessel 2.5–4.5 mm in size. The metal coverage of the published results6,14). To overcome this limitation, Lim et al. used LVIS Blue is up to 28%, higher than the previous LVIS (approxi- an LVIS Blue stent within an Enterprise stent for patients with mately 23%) and lower than the flow redirection endoluminal de- vertebral artery (VA) dissecting aneurysms, which resulted in suc- vice (FRED, 30–35%). In the present study, although immediate cessful obliteration while maintaining the VA patency13). angiographic results demonstrated a complete occlusion rate of The present study had several limitations including the retro- 59%, progression to total occlusion was impressive with complete spective data collection and the small number of patients only obliteration (86.4%) at the 1-year follow-up; morbidity or mortal- with an unruptured aneurysm. The follow-up period was short, ity did not occur. This result showed better angiographic and clin- less than 2-years, and longer follow-up is needed to assess the sta- ical outcomes than in previous studies on LVIS or LVIS Blue11,18). bility of adequate obliteration. Due to the lack of studies on LVIS Chung et al. provided informative physical properties of LVIS Blue stent, comparisons with different devices and laboratory ex- and LVIS Blue in laboratory studies and suggested the smaller an- periments to improve technical defects of LVIS Blue stent should gle of the cell design was modified in LVIS Blue compared with be performed in additional clinical trials. LVIS, leading to a better ability to open the device and greater wall apposition in tortuous arteries and higher aneurysmal neck cover- CONCLUSION age. The LVIS Blue properties may play an important role in de- creasing the development of stent thrombosis and thromboem- The LVIS Blue may be feasible and safe for stent-assisted coiling bolism, which occur due to incomplete stent apposition6). In an- of ICA aneurysms. Foreshortening, suboptimal deployment, and other study, the technical performance based on bench-top and difficulty to open proximal ends may be considerable technical is- cadaver evaluations was compared between the LVIS Blue and sues in the tortuous ICA. Larger cohort studies with longer fol- the FRED, several advantages of the LVIS Blue over the FRED low-up are necessary to support the results from this study. were observed. The most important finding was the feasibility for trans-cell approach and lesser perforator coverage with similar in- NOTES ner neck coverage of FRED at the curved vessel models. In addition, the LVIS Blue stent is delivered through lower profile micro- Conflict of interest catheters (0.021-inch) than the FRED stent (0.027-inch), leading The authors declare no conflicts of interest. to easier deployment of the stent into more distal vessels and use as an adjunctive method14). Ethical approval The delayed improvement in aneurysm occlusion after SAC is All procedures performed in the studies involving human par- considered secondary to a ‘partial’ flow-diverting effect of the ticipants were in accordance with the ethical standards of our In- modern coil-adjunctive stent17). Jankowitz BT et al. performed a stitutional Review Board with the 1964 Helsinki Declaration and computational flow analysis using 3 aneurysm models, in which its later amendments or comparable ethical standards. Pipeline flow diverter was compared with high porosity (ATLAS, Enterprise) and moderate-to-low porosity (LVIS Blue) stents. Informed consent The study results demonstrated flow diversion only partially rep- In this retrospective study, the requirement for informed con- resents a threshold in device design that encompasses metal sur- sent was waived. face coverage10). This result indicates the LVIS Blue may be beneficial for complete obliteration of an aneurysm due not only to a Funding higher occlusion rate when using coils inside the aneurysm, but No funding was received for this study. also its flow diversion effect. Several technical issues were elucidated in the present study. In REFERENCES the curved segment of the artery, stent torsion renders opening the proximal tip of the LVIS Blue stent difficult. In addition, the 1. Augsburger L, Farhat M, Reymond P, Fonck E, Kulcsar Z, Ster- oversized LVIS Blue stent may lead to the development of a tran- giopulos N, et al. Effect of flow diverter porosity on intraaneu- sitional zone, where the stent exits the parent artery to the aneurysmal blood flow. Clinical Neuroradiology 2009;19:204–214. rysm neck and the device transitions from a constrained diameter 2. Behme D, Weber A, Kowoll A, Berlis A, Burke TH, Weber W.

56 www.e-jnic.org LVIS Blue stent in the ICA Jeong Hwa Kim et al.

Low-profile Visualized Intraluminal Support device (LVIS Jr) coil-adjunctive intracranial stents. J Neurointerv Surg 2019; as a novel tool in the treatment of wide-necked intracranial an- 11:908–911. eurysms: initial experience in 32 cases. J Neurointerv Surg 11. Koch MJ, Stapleton CJ, Raymond SB, Williams S, Leslie-Mazwi 2015;7:281–285. TM, Rabinov JD, et al. LVIS Blue as a low porosity stent and 3. Brinjikji W, Murad MH, Lanzino G, Cloft HJ, Kallmes DF. En- coil adjuvant. J Neurointerv Surg 2018;10:682–686. dovascular treatment of intracranial aneurysms with flow divert- 12. Lawson MF, Newman WC, Chi YY, Mocco JD, Hoh BL. ers: a meta-analysis. Stroke 2013;44:442–447. Stent-associated flow remodeling causes further occlusion of in- 4. Cho YD, Sohn CH, Kang HS, Kim JE, Cho WS, Hwang G, et al. completely coiled aneurysms. Neurosurgery 2011;69:598–603; Coil embolization of intracranial saccular aneurysms using the discussion 603-594. Low-profile Visualized Intraluminal Support (LVIS) device. 13. Lim YC, Shin YS, Chung J. Flow Diversion via LVIS Blue Stent Neuroradiology 2014;56:543–551. within Enterprise Stent in Patients with Vertebral Artery Dis- 5. Chung J, Lim YC, Suh SH, Shim YS, Kim YB, Joo JY, et al. secting Aneurysm. World Neurosurg 2018;117:203–207. Stent-assisted coil embolization of ruptured wide-necked aneu- 14. Matsuda Y, Chung J, Keigher K, Lopes D. A comparison be- rysms in the acute period: incidence of and risk factors for tween the new Low-profile Visualized Intraluminal Support periprocedural complications. J Neurosurg 2014;121:4–11. (LVIS Blue) stent and the Flow Redirection Endoluminal De- 6. Chung J, Matsuda Y, Nelson J, Keigher K, Lopes DK. A new vice (FRED) in bench-top and cadaver studies. J Neurointerv low-profile visualized intraluminal support (LVIS) device, LVIS Surg 2018;10:274–278. Blue: laboratory comparison between old and new LVIS. Neu- 15. Nelson PK, Levy DI. Balloon-assisted coil embolization of wide- rol Res 2018;40:78–85. necked aneurysms of the internal carotid artery: medium-term 7. Fiorella D, Arthur A, Boulos A, Diaz O, Jabbour P, Pride L, et al. angiographic and clinical follow-up in 22 patients. AJNR Am J Final results of the US humanitarian device exemption study of Neuroradiol 2001;22:19–26. the low-profile visualized intraluminal support (LVIS) device. J 16. Poncyljusz W, Bilinski P, Safranow K, Baron J, Zbroszczyk M, Ja- Neurointerv Surg 2016;8:894–897. worski M, et al. The LVIS/LVIS Jr. stents in the treatment of 8. Ge H, Lv X, Yang X, He H, Jin H, Li Y. LVIS Stent Versus Enter- wide-neck intracranial aneurysms: multicentre registry. J Neu- prise Stent for the Treatment of Unruptured Intracranial Aneu- rointerv Surg 2015;7:524–529. rysms. World Neurosurg 2016;91:365–370. 17. Wang J, Vargas J, Spiotta A, Chaudry I, Truner RD, Lena J, et al. 9. Geyik S, Yavuz K, Yurttutan N, Saatci I, Cekirge HS. Stent-as- Stent-assisted coiling of cerebral aneurysms: a single-center clin- sisted coiling in endovascular treatment of 500 consecutive ce- ical and angiographic analysis. J Neurointerv Surg 2018;10: rebral aneurysms with long-term follow-up. AJNR Am J Neu- 687–692. roradiol 2013;34:2157–2162. 18. Zhang X, Zhong J, Gao H, Xu F, Bambakidis NC. Endovascular 10. Jankowitz BT, Gross BA, Seshadhri S, Girdhar G, Jadhav A, Jo- treatment of intracranial aneurysms with the LVIS device: a sys- vin TG, et al. Hemodynamic differences between Pipeline and tematic review. J Neurointerv Surg 2017;9:553–557.

www.e-jnic.org 57 eISSN 2635-5280 Original Article J Neurointensive Care 2019;2(2):58-63 https://doi.org/10.32587/jnic.2019.00192

Radiologic Factors for Predicting Dynamic Spinal Cord Compression in Conventional Cervical MRI Kwang-Ui Hong, Jong-myung Jung, Seung-Jae Hyun, Ki-Jeong Kim, Tae-Ahn Jahng Department of Neurosurgery, Spine Center, Seoul National University Bundang Hospital, Seoul National University College of Medicine, Seongnam, Korea

Received: August 23, 2019 Accepted: September 10, 2019 Objective Conventional cervical MRI is the gold standard exam for diagnosis of cervical myelopathy, but Corresponding Author: cannot detect dynamic cord compression. This study aims to evaluate radiologic factors suggest- Ki-Jeong Kim, M.D., Ph.D. ing dynamic spinal cord compression. Department of Neurosurgery, Methods Spine Center, Seoul National We retrospectively reviewed the patients who examined dynamic MRI in addition to conventional University Bundang Hospital, Seoul National University College MRI. A total of 50 patients were included in this study, comprised of 36 in the group P (aggrava- of Medicine, 82 Gumi-ro tion of spinal cord compression on dynamic MRI) and 14 in the group N (non-aggravation of spi- 173beon-gil, Bundang-gu, nal cord compression on dynamic MRI). Radiologic factors were compared. Seongnam 13620, Korea Results Tel: +82-31-787-7166 The following factors were analyzed: age, cervical canal diameter (CCD), spinal cord diameter Fax: +82-31-787-4097 (SCD), subarachnoid space (SAS), cervical lordosis (CL), cervical ROM, anterior length of cervi- E-mail: [email protected] cal cord (ALCC), posterior length of cervical cord (PLCC), length of anterior column (LAC), and length of posterior column (LPC), Significant differences were found in age, CCD, SAS, CL, LPC (p<0.05). In ROC curves, age (AUC 0.813, cut-off value 54), CCD (AUC 0.858, cut-off value 10.32), and SAS (AUC 0.884, cut-off value 3.25) were a significant factor in predicting deterioration of spinal cord compression during the dynamic posture. Conclusion Dynamic MRI is more useful for the diagnosis of dynamic spinal cord compression in the following conditions: age is 54 years or older, CCD average is less than 10.32 mm, SAS average is less than 3.25 mm, CL is -1.98˚ or LPC is less than 105.45 mm.

Keywords: Dynamic MRI; Cervical spine; Cord compression; Risk factor

INTRODUCTION namic factors which aggravates central canal stenosis1). Conven- tional MRI is a gold standard exam for diagnosis of cervical my- In the previous literature, the pathophysiology of cervical com- elopathy, but cannot detect dynamic cord compression. Dynamic pressive myelopathy is related to static cord compression and dy- MRI is used to identify dynamic cord compression in cervical

58 www.e-jnic.org Factors for predicting cord compression Kwang-Ui Hong et al. myelopathy patients whose cord compression was not found with at each intervertebral disc level were measured at the extension of conventional MRI. Risk factors of cervical myelopathy such as the midline of the intervertebral disc (Fig. 2). CCD and SCD on bulging disc and ligament flavum buckling which cause dynamic C1 level were measured at the line connecting C1 anterior arch cord compression have been reported in the literature2). However, and posterior arch. Subarachnoid space (SAS) was calculated with Dynamic MRI after a routine MRI examination is still expensive the following formula: and require additional time and patient who take dynamic MRI is 3) placed in slightly uncomfortable positions . This study aims to SASlevel = CCDlevel - SCDlevel evaluate radiologic factors suggesting dynamic spinal cord compression. Anterior and posterior length of cervical cord (ALCC and PLCC) was defined as a line along the anterior and posterior spi- METHODS nal cord from line connecting C1 anterior and posterior arch to the extension of the midline of C7 intervertebral disc. Length of Patients anterior column (LAC) was measured from apex of odontoid Between January 2014 and December 2016, 89 patients exam- process to the midpoint of C7 lower endplate. Length of posterior ined dynamic MRI due to neck pain or neurological symptoms column (LPC) was measured from C1 posterior arch to lower (radiating pain to upper extremities, tingling sensation of the up- end of C7 lamina. Cervical lordosis was defined as the Cobb’s an- per limb, weakness of upper and lower extremities and gait distur- gle between C2 lower endplate and C7 lower endplate. Cervical bance) (Fig. 1). Thirty-nine patients were excluded since multi- lordosis (CL) on flexion and extension position was measured on level ossification of the posterior longitudinal ligament, cervical dynamic MRI. Cervical range of motion (ROM) was calculated spinal cord tumor, history of trauma, history of spine surgery. A with the following formula: total of 50 patients were included in this study, comprised of 36 in the group P (aggravation of spinal cord compression on dynamic Cervical ROM = CLextension- CLflexion MRI) and 14 in the group N (non-aggravation of spinal cord compression on dynamic MRI). All measurements were performed using the INFINITT pro-

Radiologic measurements Central canal diameter (CCD) and spinal cord diameter (SCD)

Dynamic MRI: 89 patients

Previous cervical spine surgery: 17 HISTORY OF TRAUMA: 9 Severe OPLL: 8 Sinal cord tumor: 2 No neutral position MRI: 3

50 patients

Group P: 36 patients Group N: 14 patients Fig. 2. Measurement of radiologic parameters at MRI. CCD, central canal diameter; SCD, spinal cord diameter; ALCC, anterior length of cervical cord; PLCC, posterior length of cervical cord; LAC, length Fig. 1. flow chart of the study participants. of the anterior column; LPC, length of the posterior column. www.e-jnic.org 59 Factors for predicting cord compression Kwang-Ui Hong et al.

Table 1. Demographics of the patients group P group N p-value Age (years) 62 (42-79) 45 (18-69) <0.001 Sex (male : female) 20:16 5:9 0.345 CCD (mm) 9.43 (7.51-11.64) 11.25 (9.52-14.10) <0.001 CCD at C1 (mm) 13.58 (10.14-18.43) 14.79 (12.34-18.39) 0.041 SCD (mm) 6.61 (4.96-8.08) 6.83 (6.10-7.65) 0.087 SCD at C1(mm) 7.68 (5.69-9.65) 8.02 (6.80-9.07) 0.360 SAS (mm) 2.90 (1.68-4.43) 4.00 (3.26-6.50) <0.001 SAS at C1 (mm) 5.54 (3.07-9.24) 6.87 (4.30-9.92) 0.004 ALCC (mm) 105.50 (87.56-125.80) 110.91 (94.96-123.35) 0.114 PLCC (mm) 99.94 (81.69-119.87) 107.76 (88.17-120.56) 0.099 LAC (mm) 113.14 (94.23-123.93) 117.27 (105.17-126.84) 0.290 LPC (mm) 96.46 (79.7-116.15) 106.51 (83.73-118.38) 0.037 CL (°) 11.16 (-22.30-32.28) -2.11 (-17.34-20.65) 0.014 Cervical ROM (°) 32.69 (12.19-56.36) 32.35 (16.86-52.99) 0.557 The values represent the means and rages. Boldface type indicates statistical significance. Group P: dynamic cord compression positive; Group N: dynamic cord compression negative; CCD: central canal diameter; SCD: spinal cord diameter; SAS: subarachnoid space; CL: cervical lordosis; ALCC: anterior length of cervical cord; PLCC: posterior length of cervical cord; LAC: length of anterior column; LPC: length of posterior column.

gram (INFINITT Healthcare Co. Ltd., Seoul, South Korea). All 106.51; p= 0.037). CL of group P was 11.16° (lordotic curve), the radiological assessments were evaluated by two independent but that of group N was -2.11° (kyphotic curve). spine surgeons blinded to the study information, and average pa- A ROC curve was generated. A cutoff point of 54-year-old pre- rameters were used in the analysis. dicts the occurrence of spinal cord compression in dynamic MRI with 73.3% specificity and 85.7% sensitivity (Fig. 3A). The ROC Statistical analysis curve demonstrated an AUC of 0.813 (p<0.001). A cutoff point The statistical analyses were performed using the Student t-test of 10.319 mm in CCD predicts the occurrence of spinal cord com- for continuous variables and the Chi-square test or Fisher exact pression in dynamic MRI with 80.0% specificity and 77.1% sensi- test for categorical variables. A receiver operating characteristics tivity (Fig. 3B). The ROC curve demonstrated an AUC of 0.858 (ROC) curve analysis was performed to determine the cut-off (p<0.001). A cutoff point of CCD at C1 level was 13.89 and AUC value of the radiologic factor. The accuracy of the exam was evalu- was 0.759 (p=0.004). A cutoff point of 3.254 mm in SAS predicts ated by calculating the area under a curve (AUC). All statistical the occurrence of spinal cord compression in dynamic MRI with analyses were performed using R (Version 3.6.1, open-source 100.0% specificity and 71.4% sensitivity (Fig. 3C). The ROC software, www.r-project.org). curve demonstrated an AUC of 0.884 (p<0.001). A cutoff point of SAS at C1 level was 6.32 and AUC was 0.762 (p=0.004). A cut- RESULTS off point of -1.98 degrees in CL predicts the occurrence of spinal cord compression in dynamic MRI with 53.3% specificity and Demographic data of both groups are shown in Table 1. There 88.6% sensitivity (Fig. 3D). The ROC curve demonstrated an were no significant statistical differences in sex, SCD, ALCC, AUC of 0.72 (p=0.014). A cutoff point of 105.45 mm in LPC pre- PLCC, LAC, and cervical ROM. The mean age of the group P dicts the occurrence of spinal cord compression in dynamic MRI was 62, and that of group N was 45 (p< 0.001). The mean CCD with 53.3% specificity and 80.0% sensitivity (Fig. 3E). The ROC and CCD at the C1 level of group P showed significantly narrow- curve demonstrated an AUC of 0.688 (p=0.037). er than those of the group N (9.43 vs. 11.25, 13.58 vs. 14.79, respectively). The mean SAS and SAS at the C1 level of group P DISCUSSION also showed significantly narrower than those of the group N (2.90 vs. 4.00, 5.54 vs. 6.87, respectively). The mean LPC of Conventional MRI is the gold standard exam for diagnosis of group P was significantly shorter than that of group N (96.46 vs. cervical myelopathy because it supplies more practical informa-

60 www.e-jnic.org Factors for predicting cord compression Kwang-Ui Hong et al.

A B

C D

Fig. 3. Receiver operating characteristics (ROC) curves for predicting dynamic cord compression. (A) age, (B) central canal diameter, (C) subarachnoid space, (D) cervical lordosis, (E) length E of the posterior column.

www.e-jnic.org 61 Factors for predicting cord compression Kwang-Ui Hong et al. tion about spinal cord and central canal stenosis, which changes present study, there was a significant difference in age between depending on the ligament and disc. Conventional cervical MRI group P and group N. It was supposed that the difference between is typically performed with the patient in a supine position and two groups correlates with disc degeneration according to age. neck in a neutral position during examination. Thus, conventional The sagittal diameter of the spinal cord is nearly constant in MRI may not represent the physiologic neck motion and dynam- adults, averaging about 8 mm from C3 to C712). Therefore, pa- ic cord compression. Previous papers suggest that pathophysiolo- tients with congenitally narrow central canals may be more sus- gy of cervical myelopathy is related to static spinal cord compres- ceptible to spinal cord compression with relatively minor patho- sion and dynamic cord compression caused by dynamic factors logical changes such as disc bulging, osteophyte, folding of liga- such as vertebral column instability and aggravation of the spinal mentum flavum. Previous literature provides that a ratio of the canal diameter in different postures1,4). Since Dynamic MRI is congenital narrowness of the central canal in cervical myelopathy performed in neck flexion and extension position with placing patient was 66~72%, and the most frequent anomaly was a short custom-built positioning sponges under the head and neck, it pro- lamina9). Hayashi et al. proposed that developmental stenosis de- vides additional information on dynamic compression. Dynamic fined as AP diameter of CSF area less than 13 mm at pedicle level MRI seems to be more sensitive in detecting dynamic cord com- is risk factor of missed dynamic stenosis2). Chen et al. observed pression. Tykocki et al. reported that 50% of patients in their study that out of all the patients they reviewed, 31% had functional cord were classified to a high grade in Muhle scale (grade 2 and 3) in impingement in extension posture, whereas only 3% had it in flex- extension, but only 35% in neutral position5,6). They also reported ion posture. They considered that sagittal canal diameter of C7 that neutral MRI might underestimate the degree of cervical cord less than 10 mm was severe central canal stenosis and the chance compression, because grade 0 in Muhle scale was three times of demonstrating dynamic cord compression on extension posi- more often in neutral position than in extension position5,6). In tion increased to 79%13). Greenberg suggested that in the adults, study of Zeitoun et al., 22.5% of grade 3 cord compression levels spinal cord compression always occurred when the space available in extension position were grade 1 in neutral position, and the dif- for the cord(SAC) was 14 mm or less, possible occurred when the ference was most significant from C3-4 level to C5-6 level7). In SAC was 15-17 mm and never happened when the SAC was 18 preoperative study of flexion-extension MRI conducted by Zhang mm or more at Dens level14). In present study, CCD had signifi- et al., functional cord impingement (grade 3 in Muhle scale) in- cant difference between group P and group N, and dynamic cord creased to 12% in flexion, 34% in neutral, and 74% in extension compression was significantly demonstrated using dynamic MRI position8). They also reported that the ratio of grade 3 to grade 1 when CCD at C1 was 13.89 mm or less and CCD average was nearly doubled comparing to neutral MRI. When dynamic MRI 10.32 mm or less. There was significant difference in SAS be- was used for planning the operation, the number of levels requir- tween the two groups; the chance of demonstrating dynamic cord ing surgery was significantly increased8). Therefore, this exam can compression using MRI was increased when SAS average was be useful for planning the surgery of cervical myelopathy patients. 3.25 mm or less. As a result, patients with congenital central canal However, only T2 weighted sequence in sagittal image with/with- narrowness and insufficient subarachnoid space were more sus- out axial image is included in dynamic cervical MRI since these ceptible to dynamic cord compression. images can be taken rapidly to avoid motion artifact from the pa- A few previous studies also reported that the cervical canal wid- tient placed in uncomfortable position. Dynamic MRI also re- ened by 10-15% on flexion position and narrowed by 10-25% on quires patients who underwent conventional MRI to spend addi- extension position6,8). Disc bulging and folding of ligamentum fla- tional time and cost3,9). vum on extension position cause spinal cord compression4,9,10). It Age-related disc degeneration causes changes in the facet joint; was also reported that sagittal SCD decreased in flexion position these lead to segmental instability and aggravation of physiologic and increased in the extension position11,15). In other words, spinal motion. Segmental instability and angular motion are related to cord elongate, narrow and unfold in neck flexion position and dynamic stenosis by thickening ligamentum flavum4,10). Ligamen- shorten, thicken, and fold in neck extension position. These tum flavum of kyphotic segments may become thicker than that changes reflect that the spinal cord becomes susceptible to injury of the lordotic segment because the kyphotic segment may have in the presence of cervical stenosis. In this study, SCD had no sig- more room for extension. Segmental kyphosis also causes spinal nificant difference between group P and group N. We thought cord tethering with reduced cord mobility due to spinal stenosis, that the difference of SCD according to posture is more important which cannot distribute axial strain throughout the cord leading than SCD in static state. to the accumulation of distracting forces and axonal injury2,11). In The limitation of the present study was its retrospective design,

62 www.e-jnic.org Factors for predicting cord compression Kwang-Ui Hong et al. small sample size. The criteria for performing dynamic MRI had 5. Muhle C, Metzner J, Weinert D, Falliner A, Brinkmann G, Meh- not been established. Further research for additional factors such dorn MH, et al. Classification system based on kinematic MR as the severity of cord compression, disc degeneration, segmental imaging in cervical spondylitic myelopathy. AJNR Am J Neu- and cervical angle is needed in a large number of patients with roradiol 1998;19:1763–1771. long-term follow-up. 6. Tykocki T, du Plessis J, Wynne-Jones G. Correlation between the severity of myelopathy and cervical morphometric parame- CONCLUSION ters on dynamic magnetic resonance imaging. Acta Neurochir (Wien) 2018;160:1251–1258. Dynamic MRI is more useful for the diagnosis of dynamic spi- 7. Zeitoun D, El Hajj F, Sariali E, Catonné Y, Pascal-Moussellard nal cord compression in the following conditions: age is 54 years H. Evaluation of spinal cord compression and hyperintense in- or older, CCD average is less than 10.32 mm, SAS average is less tramedullary lesions on T2-weighted sequences in patients with than 3.25 mm, CL is -1.98˚ or LPC is less than 105.45 mm. cervical spondylotic myelopathy using flexion-extension MRI protocol. Spine J 2015;15:668–674. NOTES 8. Zhang L, Zeitoun D, Rangel A, Lazennec JY, Catonné Y, Pas- cal-Moussellard H. Preoperative evaluation of the cervical spon- Conflict of interest dylotic myelopathy with flexion-extension magnetic resonance No potential conflict of interest relevant to this article was re- imaging: about a prospective study of fifty patients. Spine (Phila ported. Pa 1976) 2011;36:E1134–9. 9. Dalbayrak S, Yaman O, Firidin MN, Yilmaz T, Yilmaz M. The ACKNOWLEDGEMENTS contribution of cervical dynamic magnetic resonance imaging to the surgical treatment of cervical spondylotic myelopathy. None. Turk Neurosurg 2015;25:36–42. 10. Muhle C, Weinert D, Falliner A, Wiskirchen J, Metzner J, Baum- REFERENCES er M, et al. Dynamic changes of the spinal canal in patients with cervical spondylosis at flexion and extension using magnetic 1. Miyazaki M, Notani N, Ishihara T, Kanezaki S, Tsumura H. Sur- resonance imaging. Invest Radiol 1998;33:444–449. gical outcomes after laminoplasty for cervical spondylotic my- 11. Panjabi M, White A 3rd. Biomechanics of nonacute cervical spi- elopathy: A focus on the dynamic factors and signal intensity nal cord trauma. Spine (Phila Pa 1976) 1988;13:838–842. changes in the intramedullary spinal cord on MRI. Clin Neurol 12. Anderson PA, Steinmetz MP, Eck JC. Head and neck injuries in Neurosurg 2017;162:108–114. athletesed. Rosement, IL: AAOS; 2006. 2. Hayashi T, Wang JC, Suzuki A, Takahashi S, Scott TP, Phan K, 13. Chen CJ, Hsu HL, Niu CC, Chen TY, Chen MC, Tseng YC, et et al. Risk factors for missed dynamic canal stenosis in the cervi- al. Cervical degenerative disease at flexion-extension MR imag- cal spine. Spine (Phila Pa 1976) 2014;39:812–819. ing: prediction criteria. Radiology 2003;227:136–142. 3. Nigro L, Donnarumma P, Tarantino R, Marika Rullo, Antonio 14. Greenberg AD. Atlanto-axial dislocations. Brain 1968;91:655– Santoro, Roberto Delfini. Static and dynamic cervical MRI: two 684. useful exams in cervical myelopathy. J Spine Surg 2017;3:212– 15. Kuwazawa Y, Pope MH, Bashir W, Takahashi K, Smith FW. The 216. length of the cervical cord: effects of postural changes in healthy 4. Jha SC, Miyazaki M, Tsumura H. Kinetic change of spinal cord volunteers using positional magnetic resonance imaging. Spine compression on flexion-extension magnetic resonance imaging (Phila Pa 1976) 2006;31:E579–83. in cervical spine. Clin Neurol Neurosurg 2018;174:86–91.

www.e-jnic.org 63 eISSN 2635-5280 Original Article J Neurointensive Care 2019;2(2):64-69 https://doi.org/10.32587/jnic.2019.00213

Safety and Feasibility of Percutaneous Dilatational Tracheostomy Performed by a Neurointensivist Compared with Conventional Surgical Tracheostomy in Neurosurgery Intensive Care Unit John Kwon1, Yong Oh Kim2, Jeong-Am Ryu1,2 1Department of Neurosurgery, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Korea 2Department of Critical Care Medicine, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Korea

Received: August 28, 2019 Accepted: September 10, 2019 Objective To evaluate safety and feasibility of percutaneous dilatational tracheostomy (PDT) performed by a Corresponding Author: neurointensivist compared with conventional surgical tracheostomy (CST) in neurosurgery patients Jeong-Am Ryu, M.D., Ph.D. and neurocritically ill patients. Department of Critical Care Methods Medicine and Department of This was a retrospective and observational study of adult patients who underwent tracheostomy in Neurosurgery, Samsung Medical neurosurgical intensive care unit (ICU) from January 2015 to December 2017. The primary end- Center, Sungkyunkwan University point was procedure-induced complications. Secondary endpoints were initial success of tracheosto- School of Medicine, 81 Irwon-ro, my and procedure time. Gangnam-gu, Seoul 06351, Korea Results Tel: +82-2-3410-6399 Fax: +82-2-2148-7088 A total of 118 patients underwent tracheostomy during the study period. Elective surgery of brain tu- E-mail: [email protected] mor (33.1%) and intracranial hemorrhage (20.3%) were the most common reasons for ICU admission. Prolonged intubation (42.4%) and airway protection or prevent risk of aspiration (25.4%) were the most common reasons for tracheostomy. There was no significant difference in initial success rate of tracheostomy between the two groups (p=0.110). However, procedural time was lengthier in CST than that in PDT (39.0 [30.0–60.0] min vs. 15.0 [11.0–23.0] min, p<0.001). Procedure-induced complications were more common in patients who underwent CST compared to those in patients who underwent PDT (26.3% vs. 11.5%, p=0.039). Although moderate or major bleeding occurred in five patients who underwent CST, only one patient had moderate bleeding in PDT. Especially, there were two respiratory arrests during procedure in CST. In addition, two VAPs and seven wound infections occurred in CST. Conclusions PDT performed by a neurointensivist may be safe and feasible compared to CST in neurosurgery patients and neurocritically ill patients.

Keywords: Percutaneous dilatational tracheostomy; Neurointensivist; Neurosurgery intensive care unit

64 www.e-jnic.org PDT by a Neurointensivist John Kwon et al.

INTRODUCTION cheostomy and procedure time were investigated. Procedure time was defined as time from sterilization to connection of the trache- Percutaneous dilatational tracheostomy (PDT) has been increas- ostomy tube with the mechanical ventilator after successful trache- ingly used because of its easy placement with lower rates of clinically ostomy11,12). Primary endpoint was procedure-induced complica- significant bleeding and wound infection in critically ill patients17). tion. We investigated complications at insertion or during mainte- Tracheostomy may be performed in neurosurgery or neurocritially nance such as insertional injury, bleeding or hematoma, fracture of ill patients due to various reasons such as prolonged intubation, air- tracheal ring, cuff perforation, accidental decannulation, surgical way protection, or to prevent risk of aspiration caused by brain inju- conversion, hypoxemia, stomal ulcer, and ventilator-associated ries12,17). In patients with brain injuries, tracheostomy has to be more pneumonia (VAP). Regarding associated post-procedural bleeding vigilantly performed17). Neurointensivists are specialists focusing on complications, major bleeding was defined as bleeding that re- management of brain injured patients16). Therefore, PDT per- quired cauterization, surgical treatment, or additional blood trans- formed by a neurointensivist may have many theoretical advantages fusion11,12). Minor bleeding was defined as bleeding from the inci- due to easy placement at bedside, low risk of complications, and sion site that required dressing more than three times a day or epi- specialized management for brain injured patients17). nephrine for local hemostasis. VAP was defined if patients received However, there are limited reports of PDT in neurocritially ill clinical diagnosis more than 48 hours after tracheotomy and re- patients12,17). In addition, there has been no report about complica- ceived antimicrobial therapy with a new radiographical infiltrate tions associated with fiberoptic bronchoscopy during PDT. In pa- and two of three clinical criteria: purulent secretions, fever, and tients with head injuries, fiberoptic bronchoscopy can lead to intra- leukocytosis or leukopenia13,17). Secondary endpoints were initial cranial hypertension10). Therefore, the purpose of this study was to success of tracheostomy and procedure time. investigate the safety and feasibility of PDT performed by a neurointensivist compared with conventional surgical tracheostomy Procedure (CST) in neurosurgery patients and neurocritically ill patients. In this study, a neurointensivist or a neurosurgeon determined tracheostomy. Tracheostomy was considered if a patient was ex- METHODS pected to be on prolonged mechanical ventilation, need air way protection, or show decreased sedation due to neurocritically ill Study population and design and/or general critically conditions6). We considered optimal This was a retrospective, single-center, observational study of candidates to be patients without (1) uncontrolled increased in- adult patients admitted to the neurosurgical intensive care unit tracranial hypertension, or impending brainstem herniation; (2) (ICU) at Samsung Medical Center from January 2015 to De- acute phase of stroke or delayed cerebral ischemia in subarach- cember 2017. This study was approved by the Institutional Re- noid hemorrhage; (3) refractory status epilepticus; (4) fraction view Board of Samsung Medical Center (approval number: SMC of inspired oxygen (FiO2) greater than 0.6, positive end-expirato-

2018-09-011). The requirement for informed consent was ry pressure greater than 10 cm H2O or plateau pressure greater waived due to its retrospective nature. We included adult patients than 40 cm H2O; (5) requiring high-dose inotropes or vasopres- admitted to the neurosurgical ICU who underwent tracheosto- sors; (6) uncontrolled dysrrhythmia; (7) severe acidosis; (8) ac- my during the study period. Those who were hospitalized for tive bleeding; (9) active infection at the tracheostomy site or more than 14 days after tracheostomy were selected. Of these pa- gross neck deformity6). In PDT performed by a neurointensivist, tients, we excluded patients under age 18, those who did not patients received sedation and pain control with propofol or mid- have brain injury, those who were discharged before 14 days after azolam, intravenous analgesics with fentanyl and/or remifentanil, tracheostomy, and those who had insufficient medical records. and topical analgesic with 1.5% lidocaine and 1:200,000 epi- Additionally, patients were excluded if they were admitted to de- nephrine17). Especially, at least 30 minutes before the procedure, partments other than neurosurgery. a continuously high dose infusion of propofol and bolus administration of osmotic agent were performed for patients with intra- Definitions and outcomes cranial hypertension. Mechanical ventilation was controlled and

We retrospectively reviewed all neurosurgery patients and neu- FiO2 was increased to 1.0. Before administration of neuromuscu- rocritcally ill patients admitted to the neurosurgery ICU who un- lar blockade, the goal of sedation was to achieve Richmond Agi- derwent tracheostomy during the study period. Regarding data tation Sedation Scale goal of negative 4 to 56,18). Anatomic land- associated with the tracheostomy procedure, initial success of tra- marks were identified by palpation. We also used ultrasonography www.e-jnic.org 65 PDT by a Neurointensivist John Kwon et al. to identify and avoid large blood vessels near the tracheostomy Statistical analyses site6). PDT was performed using Ciaglia Blue Rhino® (Cook Medi- All data are presented as medians and interquartile ranges cal Inc., Bloomington, IN, USA) by a neurointensivist at bedside. (IQRs) for continuous variables and numbers (percentages) for Patients were kept in supine position with hyperextension of the categorical variables. Data were compared using Mann-Whitney U neck during procedure. About 10 mm vertical incision was made test for continuous variables and Chi-squared test or Fisher’s exact at the inferior edge of the cricoid cartilage5,11). Pretracheal soft tis- test for categorical variables. All tests were two-sided and p-values sues were bluntly dissected with a mosquito clamp if needed. Un- < 0.05 were considered statistically significant. Data were analyzed derlying trachea was then identified using a needle and introducer using IBM SPSS statistics version 20 (IBM, Armonk, NY, USA). sheath under bronchoscopic visualization. After a J-tipped Selding- er wire was inserted through the introducer sheath into the trachea, RESULTS the introducer sheath was then removed and dilation of the trachea and soft tissue was performed initially with a short dilator followed Baseline characteristics and procedural by a curved conical dilator5,11). Finally, the tracheostomy tube was characteristics loaded onto an introducer dilator and gently inserted into the tra- A total of 118 patients were analyzed in this study. Among these chea over the guidewire through the dilated stoma and secured in patients, 61 underwent PDT by a neurointensivist and 57 patients place under bronchoscopic visualization9). underwent CST (36 by neurosurgeons, 21 by otolaryngologists) CST was performed by a neurosurgeon or an otolaryngologist at during the study. CST was performed in 4 (1 by neurosurgeons, 3 bedside or surgical room. In tracheostomy performed by a neuro- by otolaryngologists) patients in the operating room. Median age of intensivist or neurosurgeon, T-cannula was changed by neurosur- patients was 61 (IQR 49–70) years. Of 118 patients, 51 (43.2%) gery residents at the first seven day. However, in CST performed were males. Elective surgery of brain tumor (33.1%) and intracrani- by an otolaryngologist, T-cannula was changed by an otolaryngol- al hemorrhage (20.3%) were the most common reasons for ICU ogist at the first three day after CST. admission. There was no significant difference in age, gender, body mass index, comorbidities, Glasgow Coma Scale, or APACHE II

Table 1. Baseline characteristics PDT (n=61) CST (n=57) p-value Age (yr) — median (IQR) 59.0 (47.0–70.0) 64.0 (55.0–70.0) 0.297 Gender, male — no. of patients (%) 28 (45.9) 23 (40.4) 0.673 BMI (kg/m2) — median (IQR) 23.9 (20.9–25.7) 23.5 (20.2–25.5) 0.368 Obese (BMI >30 kg/m2) — no. of patients (%) 4 (6.6) 4 (7.0) 0.999 Comorbidities — no. of patients (%) Hypertension 33 (54.1) 35 (61.4) 0.538 Diabetes mellitus 20 (32.8) 13 (22.8) 0.316 Dyslipidemia 11 (18.0) 18 (31.6) 0.135 Malignancy 17 (27.9) 8 (14.0) 0.107 Coronary artery disease 4 (6.6) 5 (8.8) 0.737 Chronic kidney disease 3 (4.9) 4 (7.0) 0.710 Chronic liver disease 3 (4.9) 3 (5.3) 0.999 Reason for ICU admission — No. of patients (%) Brain tumor 29 (47.5) 10 (17.5) 0.014 Intracranial hemorrhage 10 (16.4) 14 (24.6) Traumatic brain injury 10 (16.4) 13 (22.8) Subarachnoid hemorrhage 7 (11.5) 17 (29.8) Cerebral infarction 2 (3.3) 3 (5.3) Other 3(4.9) 0 (0) GCS on ICU admission — median (IQR) 7.0 (4.0–12.0) 7.0 (4.0–8.0) 0.412 APACHE II score on ICU admission — median (IQR) 26.0 (24.0–30.0) 28.0 (25.0–30.0) 0.117 PDT: Percutaneous dilatational tracheostomy; CST: conventional surgical tracheostomy; IQR: interquartile range; BMI: body mass index; ICU: intensive care unit; CNS: central nerve system; GCS: Glasgow Coma Scale; APACHE: Acute Physiology and Chronic Health Evaluation.

66 www.e-jnic.org PDT by a Neurointensivist John Kwon et al.

Table 2. Procedural characteristics at the time of tracheostomy PDT (n=61) CST (n=57) p-value Reason for tracheostomy—No. of patients (%) 0.51 Difficult ventilator weaning or prolonged intubation 29 (47.5) 21 (36.8) Airway protection or prevent risk of aspiration 14 (23.0) 16 (28.1) Reduction of sedative 10 (16.4) 14 (24.6) Airway toilet 5 (8.2) 2 (3.5) Difficult airway 3 (4.9) 4 (7.0) Use of antiplatelet agent—No. of patients (%) 3 (4.9) 1 (1.8) 0.619 Use of hyperosmolar agent during tracheotomy—No. of patients (%) 0 (0) 2 (0) 0.231 GCS on procedure—median (IQR) 7.0 (6.0–12.0) 7.0 (6.0–10.0) 0.967 Invasive ICP monitoring—No. of patients (%) 33 (54.1) 32 (56.1) 0.97 Duration of mechanical ventilation before tracheostomy (days)— 9.0 (4.0–14.0) 8.0 (5.0–11.0) 0.49 median (IQR) Inserted site—No. of patients (%) 0.177 1st tracheal membrane 13 (21.3) 21 (36.8) 2nd tracheal membrane 36 (59.0) 27 (47.4) 3rd tracheal membrane 12 (19.7) 9 (15.8) Lab results on the day of tracheostomy—median (IQR) Platelet count (×103/μl) 228 (164–288) 220 (165–265) 0.579 PT(INR) 1.1 (1.1–1.2) 1.1 (1.1–1.2) 0.333 aPTT (sec) 36.4 (34.0–40.9) 36.1 (33.1–41.4) 0.605 Ventilator setting—median (IQR)

FiO2 (%) 35.0 (30.0–40.0) 35.0 (30.0–40.0) 0.638

PEEP (cmH2O) 5.0 (5.0–5.0) 5.0 (5.0–5.0) 0.638 Vasopressor requirement—No. of patients (%) 9 (14.8) 9 (15.8) 0.876 PDT: Percutaneous dilatational tracheostomy; CST: conventional surgical tracheostomy; GCS: Glasgow Coma Scale; IQR: interquartile range; ICP: intracranial pressure; INR: international normalized ratio; aPTT: activated partial thromboplastin time; FiO2: Fraction of inspired oxygen; PEEP: positive end-expiratory pressure. score on ICU admission between the two groups except for reason tious complication in patients who underwent PDT, two VAPs and for admission among baseline characteristics (Table 1). Difficult seven wound infections occurred in patients who underwent CST. ventilator weaning or prolonged intubation (42.4%) and airway However, there was no significant difference in mortality or length protection or prevent risk of aspiration (25.4%) were the most com- of stay in the ICU or hospital between the two groups. There was mon reasons for tracheostomy. There were no significant differenc- no procedure-induced intracranial hypertension (intracranial pres- es in laboratory results of platelet count or coagulation on the day of sure [ICP] > 20 mmHg) during tracheostomy in either group. tracheostomy between the two groups (Table 2). Clinical outcomes were shown in Table 3

Clinical outcomes DISCUSSION There was no significant difference in initial success rate of tracheostomy between the two groups (p= 0.110). However, pro- In this study, we investigated the safety and feasibility of PDT cedural time was lengthier in CST than that in PDT (39.0 [30.0– performed by a neurointensivist at bedside compared with CST 60.0] min vs. 15.0 [11.0–23.0] min, p<0.001). Procedure-in- in neurocritically ill patients. This study had the following major duced complications were more common in patients who under- findings: 1) Procedure-induced complications were more com- went CST compared to those in patients who underwent PDT mon in the CST group than those in the PDT group; 2) There was (26.3% vs. 11.5%, p=0.039). Although moderate or major bleed- no significant difference in initial success rate of tracheostomy being occurred in five patients who underwent CST, only one patient tween the two groups; 3) Procedural time was lengthier in CST had moderate bleeding in PDT. Especially, there were two respira- than that in PDT; 4) However, there were no significant differenc- tory arrests during procedure in CST. Although there was no infec- es in mortalities or length of stay in the ICU or hospital between www.e-jnic.org 67 PDT by a Neurointensivist John Kwon et al.

Table 3. Clinical outcomes PDT (n=61) CST (n=57) p-value Procedural data Initial success of tracheostomy — No. of patients (%) 61 (100) 54 (94.7) 0.110 Procedure time (min) — median (IQR) 15.0 (11.0–23.0) 39.0 (30.0–60.0) <0.001 Complication — No. of patients (%) 7 (11.5) 15 (26.3) 0.039 Bleeding 4 (6.6) 9 (15.8) 0.109 Moderate or major bleeding 1 (1.6) 5 (8.8) 0.105 Minor bleeding 3 (4.9) 4 (7.0) 0.710 Infection 0 (0) 3 (5.3) 0.110 Ventilator-associated pneumonia 0 (0) 2 (3.5) 0.231 Wound infection within 7days 0 (0) 1 (1.8) 0.483 Fracture of tracheal ring 3 (4.9) 0 (0) 0.244 Lost airway & respiratory arrest 0 (0) 2 (3.5) 0.231 Subcutaneous emphysema 0 (0) 1 (1.8) 0.483 Outcomes ICU mortality — No. of patients (%) 2 (3.3) 0 (0) 0.496 Hospital mortality — No. of patients (%) 6 (9.8) 3 (5.3) 0.493 Length of stay in ICU (days) — median (IQR) 16.0 (10.0–22.0) 13.0 (10.0–17.0) 0.279 Length of stay in hospital (days) — median (IQR) 49.0 (26.0–73.0) 49.0 (36.0–73.0) 0.550 PDT: Percutaneous dilatational tracheostomy; CST: conventional surgical tracheostomy; IQR: interquartile range; ICU: intensive care unit. the two groups. emic encephalopathy, and neuromuscular respiratory failure16). PDT is a relatively safe procedure that can be easily learned by Tracheostomy is generally performed in patients with these neuro- physicians with less experience in surgery than CST21). A recent critically ill problems1,2,7,8,15,20). Neurointensivists may be more vigi- study has shown that there are no significant differences in clinical lant in recognizing and managing subtle dangers such as hypoven- outcomes or complications between trainee-led PDT and experi- tilation, head-down positioning, hypoxia, hypotension, and elevat- enced intensivist-led PDT11). Therefore, PDT performed by sub- ed ICP that might occur in brain injured patients during the proce- specialty trainees of critical care medicine is a safe and feasible pro- dure14,19,20). Therefore, a neurointensivist may have advantages of cedure11). In addition, PDT can be performed by an intensivist at performing the procedure for managing neurocritically ill patients. bedside in critically ill patients due to risk of transport3,4,11,12,17,21). In This study has several limitations. First, it was a retrospective re- fact, PDT has been recognized as a reliable alternative to CST21). view of medical records. Second, tracheostomy was determined by Recently, PDT has been increasingly used because of its lower a neurointensivist or a neurosurgeon. It was not protocol-based. In complications such as bleeding and wound infections17). In this addition, the non-randomized nature of registry data might have study, procedure-induced complications were much lower in pa- resulted in selection bias. Third, there was no procedure-induced tients who underwent PDT than those in patients who underwent intracranial hypertension during tracheostomy. However, the mea- CST. Especially, there was no fatal complication in patients who surement of ICP during procedure might not be accurate. Finally, underwent PDT. There has been no report about complications our study had limited statistical power due to its small sample size. associated with fiberoptic bronchoscopy during PDT. In patients Although it provides valuable insight, prospective large-scale stud- with head injuries, fiberoptic bronchoscopy can lead to intracranial ies are needed to evaluate the usefulness of PDT performed by a hypertension10). Sustained increased ICP may lead to poor out- neurointensivist for patients with severe brain injuries to obtain ev- come in neurocritically ill patients9,10). However, there was no pro- idence-based conclusions. cedure-induced intracranial hypertension during tracheostomy in either group of the present study. CONCLUSION Why is it important for a neurointensivist to perform a PDT in brain injured patients? Neurointensivists are specialists focusing on In this study, procedure-induced complications were fewer in management of patients with acute neurologic conditions, includ- patients who underwent PDT compared to those in patients who ing traumatic brain injury, stroke, status epilepticus, hypoxic-isch- underwent CST. In addition, procedural time was shorter in

68 www.e-jnic.org PDT by a Neurointensivist John Kwon et al.

PDT compared to CST. Therefore, PDT performed by a neuro- with spontaneous supratentorial hemorrhage. Cerebrovasc Dis intensivist may be safe and feasible compared to CST in neuro- 2006;21:159–165. surgery patients and neurocritically ill patients. 9. Jeon SB, Koh Y, Choi HA, Lee K. Critical care for patients with massive ischemic stroke. J Stroke 2014;16:146–160. NOTES 10. Kerwin AJ, Croce MA, Timmons SD, Maxwell RA, Malhotra AK, Fabian TC. Effects of fiberoptic bronchoscopy on intracra- Conflict of interest nial pressure in patients with brain injury: a prospective clinical The authors declare that they have no competing interests. study. J Trauma 2000;48:878–882; discussion 882-873. 11. Lee D, Chung CR, Park SB, Ryu J-A, Cho J, Yang JH, et al. Safe- Informed consent ty and Feasibility of Percutaneous Dilatational Tracheostomy Informed consent was obtained from each participant includ- Performed by Intensive Care Trainee. Korean J Crit Care Med ed in this study. 2014;29:64–69. 12. Lee DH, Jeong J-H. Safety and Feasibility of Percutaneous Dila- ACKNOWLEDGEMENTS tational Tracheostomy in the Neurocritical Care Unit. J Neuro- crit Care 2018;11:32–38. We would like to thank Hye Jung Kim, the nursing director of 13. Michael SN, Donald EC. Guidelines for the management of the neurosurgical intensive care unit, for providing excellent ad- adults with hospital-acquired, ventilator-associated, and health- vice and engaging in fruitful discussions. We would also like to care-associated pneumonia. Am J Respir Crit Care Med thank all nurses of the neurosurgery intensive care unit at Sam- 2005;171:388–416. sung Medical Center. 14. Milanchi S, Magner D, Wilson MT, Mirocha J, Margulies DR. Percutaneous tracheostomy in neurosurgical patients with in- REFERENCES tracranial pressure monitoring is safe. J Trauma 2008;65:73–79. 15. Rabinstein AA, Wijdicks EF. Outcome of survivors of acute 1. Ahmed N, Kuo YH. Early versus late tracheostomy in patients stroke who require prolonged ventilatory assistance and trache- with severe traumatic head injury. Surg Infect (Larchmt) ostomy. Cerebrovasc Dis 2004;18:325–331. 2007;8:343–347. 16. Ryu JA, Yang JH, Chung CR, Suh GY, Hong SC. Impact of 2. Bouderka MA, Fakhir B, Bouaggad A, Hmamouchi B, Hamou- Neurointensivist Co-management on the Clinical Outcomes of di D, Harti A. Early tracheostomy versus prolonged endotrache- Patients Admitted to a Neurosurgical Intensive Care Unit. J Ko- al intubation in severe head injury. J Trauma 2004;57:251–254. rean Med Sci 2017;32:1024–1030. 3. Cabrini L, Monti G, Landoni G, Biondi-Zoccai G, Boroli F, 17. Seder DB, Lee K, Rahman C, Rossan-Raghunath N, Fernandez Mamo D, et al. Percutaneous tracheostomy, a systematic review. L, Rincon F, et al. Safety and feasibility of percutaneous trache- Acta Anaesthesiol Scand 2012;56:270–281. ostomy performed by neurointensivists. Neurocrit Care 4. Cho JH. Is Percutaneous Dilatational Tracheostomy Safe to 2009;10:264–268. Perform in the Intensive Care Unit? Korean J Crit Care Med 18. Sessler CN, Gosnell MS, Grap MJ, Brophy GM, O’Neal PV, 2014;29:57–58. Keane KA, et al. The Richmond Agitation-Sedation Scale: va- 5. Ciaglia P. Percutaneous tracheostomy. Crit Care Med 1992; lidity and reliability in adult intensive care unit patients. Am J 20:904. Respir Crit Care Med 2002;166:1338–1344. 6. Giri PC, Bellinghausen Stewart A, Dinh VA, Chrissian AA, 19. Stocchetti N, Parma A, Lamperti M, Songa V, Tognini L. Neuro- Nguyen HB. Developing a percutaneous dilatational tracheos- physiological consequences of three tracheostomy techniques: tomy service by medical intensivists: experience at one academ- a randomized study in neurosurgical patients. J Neurosurg An- ic institution. J Crit Care 2015;30:321–326. esthesiol 2000;12:307–313. 7. Goettler CE, Fugo JR, Bard MR, Newell MA, Sagraves SG, 20. Stocchetti N, Parma A, Songa V, Colombo A, Lamperti M, Tog- Toschlog EA, et al. Predicting the need for early tracheostomy: nini L. Early translaryngeal tracheostomy in patients with severe a multifactorial analysis of 992 intubated trauma patients. J brain damage. Intensive Care Med 2000;26:1101–1107. Trauma 2006;60:991–996. 21. Yoo H, Lim SY, Park C-M, Suh GY, Jeon K. Safety and Feasibili- 8. Huttner HB, Kohrmann M, Berger C, Georgiadis D, Schwab S. ty of Percutaneous Tracheostomy Performed by Medical Inten- Predictive factors for tracheostomy in neurocritical care patients sivists. Korean J Crit Care Med 2011;26:261–266. www.e-jnic.org 69 eISSN 2635-5280 Original Article J Neurointensive Care 2019;2(2):70-76 https://doi.org/10.32587/jnic.2019.00206

Crescent Sign Following Enterprise-Stent assisted Coil Embolization of Distal Internal Carotid Artery Aneurysms Young-Jin Kim, Je Young Yeon, Jong-Soo Kim, Seung-Chyul Hong Department of Neurosurgery, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Korea

Received: August 28, 2019 Accepted: September 10, 2019 Objective We aimed to determine the “significant crescent signs” associated with delayed thromboembolic Corresponding Author: events (DTE) after changes in antiplatelet medication and to evaluate their incidence and natural Je Young Yeon, M.D., Ph.D. course. Department of Neurosurgery, Methods Samsung Medical Center, Between April 2009 and April 2011, 105 consecutive patients underwent stent-assisted coil em- Sungkyunkwan University School of Medicine, 81 Irwon-ro, bolization for unruptured intracranial aneurysms (n=107) in a single tertiary hospital. Sixty-three Gangnam-gu, Seoul 06351, cases fulfilled the inclusion criteria and were included in this study. Crescent sign was evaluated Korea. through follow-up magnetic resonance angiography (MRA) immediately (<48 hours), 6 to 12 Tel: +82-2-2148-9878 months, 1 to 2 years, and 2 to 4 years and correlated with the occurrence of DTE. Fax: +82-2-3410-0048 Results E-mail: [email protected] Crescent signs were identified in 26 (41.3%) of 63 cases on immediate follow-up MRA. During the mean radiological follow-up of 53 months, the crescent signs (larger than 5 mm in length) persisted for more than 1 year in 11 (42.3%) of 26 cases. The crescent signs on follow-up MRA showed decreased size in 10 of 26 cases and disappeared in 5 cases. DTE occurred in 6 (9.5%) of 63 cases. In these 6 cases, all DTE occurred when the antiplatelet regimen was changed or stopped. Five of these six cases exhibited a crescent sign measuring over 5 mm in length, at the time of stroke. Conclusions Crescent signs may be associated with DTE. When crescent signs are larger than 5 mm in length and continuously identified on follow-up MRA, continued dual antiplatelet treatment is required.

Keywords: Intracranial Aneurysm; Self Expandable Metal Stent; Therapeutic Embolization; An- tiplatelet Drugs

INTRODUCTION aneurysm management in the past decade1-3). Stent-assisted coiling has showed a favorable occlusion rate of aneurysms previously Stent-assisted coiling has broadened the scope of endovascular considered uncoilable with protection of the parent vessel1,3,4). The

70 www.e-jnic.org Crescent sign after coil embolization Young-Jin Kim et al.

Enterprise Stent (Cordis Neurovascular, Miami, Fl.) is a widely unit, all patients were evaluated with a follow-up MRA between 6 used self-expandable, closed-cell type and represents a significant and 12 months, 1 and 2 years, and 2 and 4 years. The crescent sign improvement in stent-assisted coiling of wide neck aneurysms. was measured by measuring the length of the gap between the However, despite development of stent-assisted procedures, in- stent and vessel wall on immediate MRA and follow-up images complete stent apposition (ISA) and interruption of antiplatelet (Fig. 1). therapy can increase the risk of thromboembolic complications5,6). A DTE was defined as a symptomatic newly developed infarc- Previous studies have reported that periprocedural thromboem- tion on MRI more than 3 months after treatment, which involved bolic complications after Enterprise stent-assisted embolization vascular territory distal to the stented segment in the absence of may be the result of ISA that was associated with the use of a stent other possible sources of thromboembolism. in a tortuous parent vessel. The shape of an orphaned area of semilunar flow created by ISA is visible on 3T MRA and looks like a Antiplatelet management crescent, which is why they are referred to as “crescent signs.” The All patients received dual-antiplatelet therapy (DAT) consisting crescent signs appear to be a risk factor for delayed thromboembol- of aspirin (100 mg/day) and clopidogrel (75 mg/day for 7 days) ic events (DTE) and resolve spontaneously in only a small minori- before the procedure. After the procedure, all patients were main- ty of cases6-8). tained on DAT for a few months, followed by mono-antiplatelet The optimal antiplatelet therapy protocol for prevention of therapy. thromboembolic complications after stent-assisted embolization of cerebral aneurysm has not been established. However, few stud- Statistical analysis ies reported that DTE were associated with cessation of antiplatelet All analyses were performed using SPSS 20.0 version. Means therapy5,9,10). and standard deviations were provided for continuous variables. The objectives of this study were to identify “significant crescent Student’s t-tests and Fisher’s exact tests were used for analysis of signs” associated with DTE after changes in antiplatelet medication the variables. A p-value of < 0.05 was considered statistically signif- and to evaluate their incidence and natural course. icant.

METHODS RESULTS

Patient selection Among the 105 patients (107 aneurysms), 63 cases were treated There were 105 consecutive patients with 107 aneurysms treat- using Enterprise stent-assisted coil embolization and were included with Enterprise stent-assisted embolization for intracranial an- ed in this study. The patients consisted of 51 women (81%, eurysms at our institution between April 2009 and April 2011. Pa- 55.7±7.4 years) and 12 men (19%, 57.6±10.2 years). Thirty-one tients were included in this study if they met the following criteria: patients’ aneurysms were located on the paraclinoid segment 1) immediate (< 48 hrs) and follow-up MRA was performed; 2) (49.2%), while 22 (34.9%) aneurysms were located on the dorsal distal internal carotid artery (ICA) aneurysms were present; and 3) wall of the ICA. Aneurysms on the origin of the posterior commu- after discharge, the patients received antiplatelet medication. Sev- nicating artery were found in 7 cases (11.1%). Each aneurysm enty-five cases met the inclusion criteria. Exclusion criteria were as (1.6%) was located on the origin of the ophthalmic and anterior follows: 1) use of multiple Enterprise stents (n= 5); 2) cavernous choroidal artery or distal ICA (superior to anterior choroidal ar- ICA aneurysm (n= 2); and 3) clinico-radiological follow-up short- tery). The mean aspect ratio (height/neck) of aneurysms was 1.3 er than 6 months after embolization (n= 5). Finally, a total of 63 (range: 0.6 to 3.1) and the mean “dome to neck” ratio was 1.2 cases were eligible for this study. Patients’ age and sex and the loca- (range: 0.7 to 2.3). The locations of the treated aneurysms are list- tion and size (aspect and dome-to-neck ratio) of the aneurysm ed in Table 1. were analyzed in addition to comorbid conditions including hypertension, diabetes mellitus, hyperlipidemia, coronary artery dis- Onset of DTE and antiplatelet status ease, prior stroke history, and smoking history. During the follow-up period, DTE occurred in 6 (9.5%) of 63 cases (median: 10 months, range: 5 to 14 months). All 6 cases de- Image assessment and follow-up veloped DTE in association with antiplatelet medication changes. During the admission period, all patients underwent brain MRI Three of the six cases occurred in patients who were not on anti- and MRA within 48 hours after the procedure. In the out-patient platelet therapy (temporarily stopped for surgery, Fig. 2), and the www.e-jnic.org 71 Crescent sign after coil embolization Young-Jin Kim et al.

A B

C D

Fig. 1. A 65-year-old woman with a wide-neck left paraclinoid aneurysm. A: 3D-AP reconstructed image of aneurysm. B: DSA (digital subtraction angiography) after stent-assisted coil embolization. C: Immediate post-treatment MRA shows a crescent sign measuring 7 mm in length on the inner vessel curve (yellow circle). D: The crescent sign is not present on the follow-up MRA performed 1 year after treatment (yellow circle). other 3 events occurred in patients who changed DAT to mo- We also analyzed patient factors between those who experienced no-antiplatelet therapy (MAT). The characteristics of patients DTE and the remaining patients. Baseline characteristics are com- with DTE are listed in Table 2. pared in Table 3. The incidence of DTE was higher in patients with a history of prior intracranial hemorrhage or infarction (p= 0.002). Correlation between patients’ factors and DTE There was no significant difference in the other patient clinical fac- occurrence tors: age (p= 0.144), hypertension (p= 0.372), diabetes mellitus

72 www.e-jnic.org Crescent sign after coil embolization Young-Jin Kim et al.

Table 1. Locations of treated aneurysms (p=0.182), hyperlipidemia (p=1.000), smoking history Location No. of Aneurysms (%) (p= 0.404), and coronary artery disease (p= 1.000). In the DTE ICA bifurcation 0 (0%) group, 5 cases exhibited a crescent sign measuring over 5 mm in Distal ICA (superior to anterior choroidal artery) 1 (1.6%) length at the time of stroke. Thus, we hypothesized that crescent Dorsal wall 22 (34.9%) signs over 5 mm may have clinical implications associated with Anterior choroidal segment 1 (1.6%) DTE, and we evaluated the incidence of DTE between groups Posterior communicating artery 7 (11.1%) based on crescent signs measuring over 5 mm. The results indicat- Ophthalmic artery 1 (1.6%) ed that a persistent crescent sign measuring over 5 mm, which we Paraclinoid segment 31 (49.2%) named as a “significant crescent sign,” has a strong correlation with DTE occurrence (p= 0.001).

A B C

D E F

Fig. 2. A 55-year-old woman with a right posterior communicating artery aneurysm. Immediate post-procedural MRA reveals a crescent sign measuring about 12 mm on the medial aspect of the Enterprise stent. The patient received DAT (aspirin + clopidogrel), and brain MRA performed at 12 months after the procedure shows the same size of the crescent sign. The patient discontinued clopidogrel 5 days for knee arthroplasty at 14 months after embolization. After knee arthroplasty, the patient developed delayed thromboembolic events (DTE) (mild dysarthria and left arm weakness). A: Pre-procedural 3D reconstructed image of the aneurysm. B: DSA after stent-assisted coil embolization. C: MRA performed at 12 months after the procedure shows the crescent sign measuring 12 mm (white arrow). D: Acute infarction in the right hemisphere (apparent diffusion coefficient map obtained at the time of occurrence of DTE). E: MRA shows complete occlusion of the right ICA (yellow circle). F: The crescent sign persists on the MRA performed 5 years after the DTE (white arrow). www.e-jnic.org 73 Crescent sign after coil embolization Young-Jin Kim et al.

Table 2. Characteristics of patients with ischemic complication after change in antiplatelet regimen Time interval to No. Age (yr), Sex Aneurysm Location Event (month) Antiplatelet medication 1 57/F Paraclinoid 10 Aspirin+Clopidogrel for 9 months Aspirin only for 1 month Stroke → → 2 64/F Paraclinoid 11 Aspirin+Clopidogrel for 7 months Clopidogrel only for 4 months Stroke → → 3 69/M Paraclinoid 14 Aspirin+Clopidogrel for 4 months Clopidogrel only for 10 months Stroke → → 4 55/F PCoA* 14 Aspirin+Clopidogrel for 1 month Clopidogrel only for 13 months Temporary stop for 5 days (knee arthroplasty) → Stroke → → 5 57/F Dorsal wall, Lt. ICA 9 Aspirin+Clopidogrel for 8 months à Clopidogrel only for 1 months Temporary stop for 5 days (Lung cancer) Stroke → → 6 57/F Dorsal wall, Rt. ICA 5 Aspirin+Clopidogrel for 4 months à Clopidogrel only for 1 month Temporary stop for 5 days (Lung cancer) Stroke → → No. 5 and 6 indicate both sides of the same patient.

Table 3. Univariate analysis of clinical chracteristics in two groups (DTE vs. Non-DTE) DTE (n=6) Non DTE (n=57) p-value Age 59.8±7.9 55.6±8.1 0.144 Hypertension 3 17 0.372 Diabetes mellitus 1 1 0.182 Hyperlipidemia 1 9 1.000 Smoking history 1 4 0.404 Prior stroke history 3 1 0.002 Coronary artery disease 0 3 1.000 Crescent sign> 5 mm 5 6 0.001 DTE: delayed thromboembolic events.

Table 4. Correlation between the crescent sign and DTE occurrence with regard to change in antiplatelet regimen 1. Dual → mono antiplatelet switch (n=61) DTE Non-DTE Significant crescent sign > 5 mm (n=17) 5 12 p=0.005 No/minimal crescent sign < 5 mm (n=44) 1 43

2. Temporary medication stop (n=13) DTE Non-DTE Significant crescent sign > 5 mm (n=3) 3 0 p=0.003 No/minimal crescent sign < 5 mm (n=10) 0 10 DTE: delayed thromboembolic events.

Incidence and natural course of the crescent sign and (persistent minimal crescent). In the other 5 cases, the crescent antiplatelet medication signs disappeared or were not detectable on follow-up MRA. Crescent signs measuring over 5 mm were identified in 26 In a total of 63 cases, patients had varied duration of DAT with (41.2%) of 63 cases on immediate post-procedural MRA (< 48 an average of 6.1 months (range: 1 to 32 months). We analyzed the hrs). The locations of crescent signs were the anterior genu of the correlation between a significant crescent sign and incidence of carotid siphon (n= 21), between the ophthalmic artery and com- DTE when antiplatelet therapy was changed or stopped (Table 4). municating segment (n= 3), neck remnant of aneurysm (n= 1), If a crescent sign measuring over 5 mm existed at the time of and untreated aneurysm within stent (n= 1). During the mean ra- changing antiplatelet regimen (DAT to MAT), the incidence of diological follow-up of 53 months, crescent signs measuring over 5 DTE was significantly increased compared to the group without mm persisted for 1 to 6 years in 11 (42.3%) of the 26 cases. In 10 crescent signs or that having minimal crescent signs <5mm of the remaining 15 cases, the crescent signs were smaller in size (p= 0.005). When the patient stopped antiplatelet therapy tempo-

74 www.e-jnic.org Crescent sign after coil embolization Young-Jin Kim et al. rarily for any other reasons, if the patient had a significant crescent gration; and 3) angle remodeling of a deployed stent6). In Heller et sign, the incidence of DTE was also increased compared to other al.’s and our studies, however, only small crescent signs were re- groups (p= 0.003). solved spontaneously. Considering that the crescent sign and incidence of DTE have a significant correlation, continuous follow-up DISCUSSION seems necessary once the crescent sign is identified after stent-assisted coiling. The development of intracranial self-expanding stents has The post-procedural antiplatelet therapy has been accepted as broadened the applicable range of endovascular treatment to intra- the standard of care for prophylaxis against in-stent stenosis or cranial aneurysms that are wide-necked, large, or giant-sized. The thromboembolic events. Yamada et al.10) reported that the use of purpose of using a stent in embolization is prevention of coil pro- DAT with aspirin and clopidogrel for elective coil embolization of trusion in addition to a diversion effect of the hemodynamic blood unruptured cerebral aneurysms has been shown to reduce the rate flow to diminish intra-aneurysmal flow1,3,4). of symptomatic thromboembolic complications. Mocco et al.5) re- The Enterprise stent was introduced in May 2007. Recently, several ported that all 7 (3%) DTE cases were concomitant to cessation of studies of short- and long-term results for the Enterprise stent were re- DAT. Rossen et al.9) reviewed 154 patients who underwent aneu- ported, demonstrating excellent occlusion rate and device navigabili- rysm treatment with stent techniques; despite all patients taking ty, as well as low morbidity and mortality rates2,4, 5,11). Especially, a aspirin at the time of the ischemic events, 6 (5%) ischemic events long-term report showed that 81% of angiograms demonstrated developed after clopidogrel discontinuation. In our study, 6 (9.5%) greater than 90% aneurysm occlusion rates and only 8.3% required of 63 cases experienced ischemic events and all 6 cases developed retreatment. These results are similar to data reported in a large series DTE in association with antiplatelet medication changes (3 cases– of aneurysm coil embolization with and without stent assistance11). MAT change, 3 cases–stopped temporarily). Regarding our results, However, concern remains over the deployment of potentially especially in patients with a persistent significant crescent sign on thrombogenic sources into the intracerebral artery when ISA exists follow-up MRA, changing the regimen or cessation of antiplatelet after stent-assisted coiling. Although ISA has been more reported therapy could increase the risk of ischemic events. However, there in cardiology literature12-14), recently, many studies have reported remain questions regarding the duration of dual-antiplatelet thera- an association between ISA and ischemic events in the neurosurgi- py and the optimal time to switch to MAT. In the cardiovascular cal fields. The presence of ISA can be detected by its characteristic field, increasing thromboembolic events related to a drug-eluting shape on MRA: a semilunar orphaned area of ICA, which looks stent after cessation of clopidogrel have led to the recommendation like a crescent sign5). to maintain the antiplatelet therapy for 6 to 12 months15,16). A ret- Heller et al. reported that in ISA of the Enterprise stent-mediated rospective study of 395 patients with 403 intracranial aneurysms coiling of aneurysms of 50 patients, delayed ischemic events occurred treated with stent-assisted coil placement reports that longer-term in 8 (16%) of 50 cases, and all cases involved patients with ISA6). In dual antiplatelet therapy for more than 9 months and a late switch our study, DTE occurred in 6 cases (9.5%) of 63 cases, and 5 of 6 cas- to mono-antiplatelet therapy were found to significantly reduce es had a significant crescent sign at the time of the event. The signifi- delayed ischemic stroke17). To establish the optimal duration and cant crescent sign was also statistically associated with DTE method of antiplatelet therapy for the prevention of in-stent steno- (p=0.001). Thus, we suggest that significant crescent signs (over 5 sis in patients with intracranial stents, especially when the ISA is mm) are an important factor in predicting the probability of DTE. present, further research is needed. Considering the natural course of crescent signs, some crescent signs were observed consistently on follow-up images. Heller et Limitation al.6) reported that ISA was identified in postoperative imaging stud- As this study was a retrospective analysis, there is a possibility of ies in 22 (44%) of 50 patients, with 19 (86%) of 22 crescent signs selection bias. A relatively small number of patients compared with persisting and 3 (14%) of 22 crescent signs resolving on subse- previous studies may be another limitation. The crescent sign was quent serial imaging. In this study, initially, significant crescent measured only in length in this study. We believe that various use- signs were identified in 26 (41.2%) of 63 cases, and 11 (42.3%) of ful methods can be employed to evaluate the crescent sign. the 26 cases showed persistent significant crescent signs on their last follow-up MRA. Resolution of the crescent sign on serial imag- CONCLUSION ing could be explained by 3 possibilities as follows: 1) in-stent stenosis or thrombosis of the orphaned lumen; 2) delayed stent mi- After stent-assisted embolization, the crescent sign showed www.e-jnic.org 75 Crescent sign after coil embolization Young-Jin Kim et al. strong association with the incidence of DTE. Crescent signs mea- stent-mediated coil embolization of aneurysms. J Neurosurg suring over 5 mm in length may have more important clinical im- 2011;115:624–632. plications. Antiplatelet strategies need to be more customized after 8. Heller RS, Malek AM. Parent vessel size and curvature strongly Enterprise stent-assisted embolization due to an increased risk of influence risk of incomplete stent apposition in enterprise in- developing DTE, especially with the presence of persistent signifi- tracranial aneurysm stent coiling. AJNR Am J Neuroradiol cant crescent signs. Further research is warranted to determine the 2011;32:1714–1720. optimal antiplatelet regimen and duration in patients who undergo 9. Rossen JD, Chalouhi N, Wassef SN, Thomas J, Abel TJ, Jab- intracranial stent-assisted embolization with ISA. bour PM, et al. Incidence of cerebral ischemic events after discontinuation of clopidogrel in patients with intracranial aneu- NOTES rysms treated with stent-assisted techniques. J Neurosurg 2012;117:929–933. Conflict of interest 10. Yamada NK, Cross DT, Pilgram TK, Moran CJ, Derdeyn CP, No potential conflict of interest relevant to this article was re- Dacey RG. Effect of antiplatelet therapy on thromboembolic ported. complications of elective coil embolization of cerebral aneurysms. AJNR Am J Neuroradiol 2007;28:1778–1782. REFERENCES 11. Fargen KM, Hoh BL, Welch BG, Pride GL, Lanzino G, Boulos AS, et al. Long-term results of enterprise stent-assisted coiling 1. Akpek S, Arat A, Morsi H, Klucznick RP, Strother CM, Mawad of cerebral aneurysms. Neurosurgery 2012;71:239–244; dis- ME. Self-expandable stent-assisted coiling of wide-necked in- cussion 244. tracranial aneurysms: a single-center experience. AJNR Am J 12. Attizzani GF, Capodanno D, Ohno Y, Tamburino C. Mecha- Neuroradiol 2005;26:1223–1231. nisms, pathophysiology, and clinical aspects of incomplete 2. Higashida RT, Halbach VV, Dowd CF, Juravsky L, Meagher S. stent apposition. J Am Coll Cardiol 2014;63:1355–1367. Initial clinical experience with a new self-expanding nitinol 13. Foin N, Gutierrez-Chico JL, Nakatani S, Torii R, Bourantas stent for the treatment of intracranial cerebral aneurysms: the CV, Sen S, et al. Incomplete stent apposition causes high shear Cordis Enterprise stent. AJNR Am J Neuroradiol flow disturbances and delay in neointimal coverage as a func- 2005;26:1751–1756. tion of strut to wall detachment distance: implications for the 3. Lylyk P, Ferrario A, Pasbon B, Miranda C, Doroszuk G. Buenos management of incomplete stent apposition. Circ Cardiovasc Aires experience with the Neuroform self-expanding stent for Interv 2014;7:180–189. the treatment of intracranial aneurysms. J Neurosurg 2005; 14. Kume T, Waseda K, Ako J, Sakata K, Yamasaki M, Shimohama 102:235–241. T, et al. Intravascular ultrasound assessment of postprocedural 4. Weber W, Bendszus M, Kis B, Boulanger T, Solymosi L, Kuhne incomplete stent apposition. J Invasive Cardiol 2012;24:13– D. A new self-expanding nitinol stent (Enterprise) for the treat- 16. ment of wide-necked intracranial aneurysms: initial clinical 15. Moussa ID, Colombo A. Antiplatelet therapy discontinuation and angiographic results in 31 aneurysms. Neuroradiology following drug-eluting stent placement: dangers, reasons, and 2007;49:555–561. management recommendations. Catheter Cardiovasc Interv 5. Mocco J, Fargen KM, Albuquerque FC, Bendok BR, Boulos 2009;74:1047–1054. AS, Carpenter JS, et al. Delayed thrombosis or stenosis follow- 16. Maeng M, Thim T, Chisholm G, Johansen M, Schmidt M, ing enterprise-assisted stent-coiling: is it safe? Midterm results Sørensen HT, et al. TCT-55 Major Adverse Cardiac Events As- of the interstate collaboration of enterprise stent coiling. Neu- sociated With Discontinuation Of Clopidogrel Treatment rosurgery 2011;69:908–913; discussion 913-904. Within The First Year After Coronary Stent Implantation In 6. Heller R, Calnan DR, Lanfranchi M, Madan N, Malek AM. In- Western Denmark. Journal of the American College of Cardi- complete stent apposition in Enterprise stent-mediated coiling ology 2012;60:B17–B17. of aneurysms: persistence over time and risk of delayed isch- 17. Hwang G, Kim JG, Song KS, Lee YJ, Villavicencio JB, Setiawan emic events. J Neurosurg 2013;118:1014–1022. NS, et al. Delayed Ischemic Stroke after Stent-assisted Coil 7. Heller RS, Miele WR, Do-Dai DD, Malek AM. Crescent sign Placement in Cerebral Aneurysm: Characteristics and Optimal on magnetic resonance angiography revealing incomplete stent Duration of Preventative Dual Antiplatelet Therapy. Radiolo- apposition: correlation with diffusion-weighted changes in gy: Imaging Cancer 2014;273:194–201.

76 www.e-jnic.org eISSN 2635-5280 Original Article J Neurointensive Care 2019;2(2):77-81 https://doi.org/10.32587/jnic.2019.00199

Acute Acalculous Cholecystitis in Neurological Patients; Clinical Review, Risk Factors, and Possible Mechanism See Won Um, Hak Cheol Ko, Seung Hwan Lee, Hee Sup Shin, Jun Seok Koh Department of Neurosurgery, Kyung Hee University Hospital at Gangdong, College of Medicine, Kyung Hee University, Seoul, Korea

Received: August 26, 2019 Accepted: September 10, 2019 Backgrounds Acute acalculous cholecystitis (AAC) has rarely been reported and has not been fully investigated Corresponding Author: in patients with severe neurological conditions that can cause changes in consciousness. Delays in Hak Cheol Ko, M.D., Ph.D. diagnosis of AAC may increase the severity of this condition. The aim of this study was to investi- Department of Neurosurgery, gate the clinical features and possible mechanisms of AAC in neurological patients. Kyung Hee University Hospital at Gangdong, 892, Dongnam-ro, Methods Gangdong-gu, Seoul 05278, Among the patients with neurological conditions admitted to our hospital between March 2007 Korea and September 2018, the medical records of 52 patients who underwent cholecystectomy after be- Tel: +82-2-440-8402 ing diagnosed with cholecystitis were reviewed retrospectively. Data regarding age, sex, neurological Fax: +82-2-440-8404 conditions, Glasgow Coma Scale (GCS) on admission, past medical history, time from admission E-mail: [email protected] to the onset of AAC, initial symptoms or signs of AAC, and fasting period were investigated. Results The mean age of the 52 patients (36 male, 16 female) was 67.1±14.4. The mean GCS was 10. The neurological conditions were intracerebral hemorrhage, subarachnoid hemorrhage, subdural hematoma, intraventricular hemorrhage, epidural hematoma, cerebral infarction, cerebral abscess, and hypoxic brain damage. The mean time interval between the onset of AAC and admission was 22.5 days and the mean fasting period was 8.1 days. Conclusion AAC may be a significant complication in patients with neurological conditions. It is important to identify symptoms and signs of AAC, accurately diagnose the AAC, particularly in long-term stay at intensive care unit (ICU), a long period of fasting, state of hypotension or hypoperfusion, high positive end-expiratory pressure (PEEP) ventilation.

Keywords: Cholecystitis; Neurological Intensive Care Unit; Mechanism

INTRODUCTION mechanism of inflammation and ischemia in the gallbladder (GB) wall1), stagnant pooling of bile juice, and the lithogenicity of bile Acute cholecystitis is known to be caused by the pathologic juice in a complex manner. The incidence of acute cholecystitis is

Copyright © 2019 The Korean Neurointensive Care Society This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited. www.e-jnic.org 77 Acute Acalculous Cholecystitis in Neurological Patients See Won Um et al. especially high after a major surgery or serious illness. Acute chole- in the 70s with the age group in their 70s showing the highest prev- cystitis may occur in the presence of a GB stone, however acute alence. There were 36 males and 16 females with the males show- acalculous cholecystitis (AAC) is reported to occur more often2). ing 1.89 times higher incidence than the females. For the underly- Hypoperfusion and bile stasis are the key pathogenesis of AAC, ing disease, there were HTN in 23 patients, diabetics mellitus which could be aggravated by hypotension, dehydration, and the (DM) in 10 patients. The mean Glasgow Coma Scale (GCS) was usage of vasoactive drugs. Bile stasis is caused by the use of opioids, 10 which implies neurological deterioration with drowsy to stu- fasting with total parenteral nutrition (TPN), and mechanical ven- porous mentality being the most prevalent. Also, Intracerebral tilation with positive end-expiratory pressure (PEEP)3). hemorrhage was the most common neurological diagnosis among AAC is reportedly a complication of a disease with high mortali- brain insults with subarachnoid hemorrhage, subdural hemor- ty unless treatment, long-term and short-term mortality may reach rhage, and epidural hemorrhage in order of descent. 59% and 35%, respectively4). The incidence of AAC is rare in pa- The clinical factors related to AAC are summarized in Table 2. tients with severe neurological conditions with alterations in con- The mean time from admission to onset of AAC was 22.5 days and sciousness, such as cerebrovascular disease (CVD) or severe brain length of hospital stay showed an average of 77.9 days. The mean trauma. However, diagnosing and treating AAC are important in fasting period was 8.1 days. The initial symptoms and signs of avoiding high mortality. Therefore, it is important to recognize the AAC were elevation of liver function test (LFT), which was also symptoms and signs of AAC, make an accurate diagnosis, and pro- the most common symptoms and signs in 27 patients; there were vide expedient treatment. Neurological patients are accompanied right upper quadrant (RUQ) pain or tenderness in 4 patients, fever, by severe neurological symptoms and signs, such as loss of con- in 2 patients, C-reactive protein (CRP) elevation in 5 patients, and sciousness, paraplegia, and aphasia, which can increase the severity incidentally diagnosed in 10 patients. The 10 incidentally diagnosis of the symptoms of AAC. of AAC was done through abdominal sonography or CT after sus- Nevertheless, abdominal evaluations are not routinely done in picions of acute kidney injury, cystitis, pneumonia, and pleural ef- patients with CVD or brain trauma and there are no established fusion. The asymptomatic 10 patients received cholecystectomy clinical characteristics of AAC in the intensive care unit (ICU). The aim of this study is to investigate the clinical features of AAC and the possible mechanisms of AAC in neurological patients. Table 1. Clinical characteristics of 52 patients Characteristics Sex (number) METHODS Male 36 Among the patients with neurological conditions admitted to Female 16 Age (mean) 67.1 our hospital between March 2007 and September 2018, the medi- Past medical history cal records of 52 patients who underwent cholecystectomy after Hypertension 23 being diagnosed with cholecystitis were reviewed retrospectively. Diabetes mellitus 10 Data regarding age, sex, neurological conditions, mental status on Cerebrovascular disease 2 admission, past medical history, time from admission to the onset Heart disease 5 of AAC, initial symptoms or signs of AAC, and fasting period were Alcoholics 4 investigated. Glasgow Coma Scale on admission AAC was diagnosed through abdominal sonography or comput- ≥ 13 26 ed tomography (CT), which showed no gallstones. The date of di- 7-12 4 agnosis was defined as the day in which AAC was diagnosed < 6 22 Neurological conditions through the abdominal imaging. Subarachnoid hemorrhage 12 Intracerebral hemorrhage 19 RESULTS Subdural hematoma 11 Epidural hematoma 3 The clinical characteristics of total 52 patients who underwent Intraventricular hemorrhage 8 cholecystectomy is summarized in Table 1. The average age was Cerebral infarction 3 67.1 years with broad age spectrum spanning from 17 to 90 years- Hypoxic brain damage 1 old, only one 17 years-old, no one in the 20s, 17 in the 60s, and 19 Cerebral abscess 2

78 www.e-jnic.org Acute Acalculous Cholecystitis in Neurological Patients See Won Um et al. after discussion among board certified gastroenterologists or gen- GCS in the neurosurgical treatment and conservative management eral surgeons. Of the patients who were treated for neurological group were 9.4 and 12.1, respectively. diseases, there were 35 patients who have received neurosurgical operative treatment and 17 patients who have received conserva- DISCUSSION tive management. Of which the group received neurosurgical operative treatment had a higher prevalence of AAC. The prevalence of cholecystitis in GB stone patients after sur- The patients who have received cholecystectomy after being di- gery is reported to be similar between male and female. However, agnosed with AAC were divided into the neurosurgical treatment over 80% of AAC after other operations unrelated to trauma are group and the conservative management group (Table 3). The male patients aged 50 years old and over5). This indicates that mean time from admission to onset of AAC in the neurosurgical emergent operation or an elective operation complicated by hypo- treatment group was 27.3 days, which was 14.7 days longer than tension and blood loss is pronounced in male patients2). The inci- the conservative management group (12.6 days). The mean fasting dence of AAC in this study showed a 1.89 times higher rate of inci- period in the neurosurgical treatment group was 9.1 days and the dence in males than in females as noted. This could be explained that in the conservative management group was 6.1 days. Mean that the rate of comorbidities is higher in the male patients6), however sex alone cannot be considered an independent factor for cholecystectomy in neurological patients. Nevertheless, special at- Table 2. Clinical factors associated with acute acalculous cholecystitis tention could be paid to the occurrence of AAC in male patients Clinical factors considering what is reported in previous studies and this study. Mean time from admission to onset of AAC (day) 22.5 Most of the neurological deteriorated patients admitted to the 7) Hospital stay (day) 77.9 ICU requires PEEP, which can be associated with bile stasis .

Mean fasting period (day) 8.1 7-10cm ventilation in H2O PEEP can raise hepatic venous pres- 7) Initial symptoms or signs (number) sure, which causes a decrease in portal perfusion , in turn can RUQ pain or tenderness 8 cause bile stasis and increases the possibility of gallbladder mucosal Fever 2 injury. There was no statistically significant difference among the Elevation of LFT 27 patients in this study as 4-5cm ventilation of H2O PEEP was ap- Elevation of CRP 5 plied, but it is recommended to always keep in mind the possibility Diagnosis incidentally 10 of bile stasis in patients requiring 7-10cm high PEEP. Neurological treatment The average duration in ICU stay was 13.3 days with mean fast- Operative 35 8) Conservative 17 ing period 7.3 days. Yang et al. reported that successive fasting AAC: acute acalculous cholecystitis; CRP: c-reactive protein; LFT: liver time as a risk factor of acute cholecystitis in aneurysmal subarach- function test; RUQ: right upper quadrant. noid hemorrhage patients. Successive fasting time reported was

Table 3. Clinical factors associated with acute acalculous cholecystitis Variables Neurosurgical treatment (n=35) Conservative management (n=17) Mean time from admission to onset of AAC (day) 27.3 12.6 Mean fasting period (day) 9.1 6.1 Mean GCS on admission 9.4 12.1 Neurological conditions Subarachnoid hemorrhage 10 2 Intracerebral hemorrhage 11 8 Subdural hematoma 9 2 Epidural hematoma 1 2 Intraventricular hemorrhage 7 1 Cerebral infarction 1 2 Hypoxic brain damage - 1 Cerebral abscess 2 - AAC: acute acalculous cholecystitis; GCS: Glasgow coma scale. www.e-jnic.org 79 Acute Acalculous Cholecystitis in Neurological Patients See Won Um et al.

5.38±2.78 days and the pathogenesis of acute cholecystitis includ- NOTES ed bile stasis as a result of fasting time as its mechanism. Loss of enteric stimulation during parenteral nutrition impairs the gallblad- Conflict of interest der motor function and causes GB “sludge” formation. Long-term No potential conflict of interest relevant to this article was re- TPN can increase the incidence of AAC by up to 30%7), GB ported. “sludge” formation can occur in about 50% of patients treated with TPN for 4 weeks and in 100% of patients after 6 weeks9). Auto- REFERENCES nomic nerve dysfunction in neurological patients may cause bile stasis. Sevastos et al.10) was proved to cause abnormalities in the 1. Naidech AM, Bendok BR, Bernstein RA, Alberts MJ, Batjer flow control of bile and pancreatic juice through increase in vagal HH, Watts CM, et al. Fever burden and functional recovery af- paresis and catecholamine secretion following the damage of the ter subarachnoid hemorrhage. Neurosurgery 2008;63:212– brain center with the cessation of oral feeding. 217; discussion 217-8. Gallbladder ischemia is pathogenesis of AAC. Orlando et al.11) 2. Barie PS, Eachempati SR. Acute acalculous cholecystitis. Curr reported that gallbladder ischemia and bile stasis leading to hy- Gastroenterol Rep 2003;5:302–309. poperfusion. Intraluminal pressure is increased by bile stasis, which 3. Kim SJ, Lee SJ, Lee SH, Lee JH, Chang JH, Ryu YJ. Clinical results in a decrease in gallbladder perfusion pressure. In this hy- characteristics of patients with newly developed acute cholecys- pothesis, bacterial invasion of ischemic tissue is a secondary phe- titis after admission to the intensive care unit. Aust Crit Care nomenon, acute bacterial and viral diseases, such as hepatitis, infec- 2019;32:223–228. tion of upper respiratory tract, are important precipitating fac- 4. Glenn F, Becker CG. Acute acalculous cholecystitis. An increas- tors11-14). Gallbladder ischemia increases mucosal phospholipase ing entity. Ann Surg 1982;195:131–136. A2 and superoxide dismutase activities, and mucosal lipid peroxide 5. Delemos AS, Friedman LS. Systemic causes of cholestasis. Clin content. Whereas gallstone-related disease is associated with arteri- Liver Dis 2013;17:301–317. al dilation and extensive venous filling, but AAC is associated with 6. Nikfarjam M, Harnaen E, Tufail F, Muralidharan V, Fink MA, arterial occlusion and minimal to absent venous filling15). Starkey G, et al. Sex differences and outcomes of management Another mechanism closely related to the pathogenesis of AAC of acute cholecystitis. Surg Laparosc Endosc Percutan Tech is systemic inflammatory response. The response to systemic inju- 2013;23:61–65. ry and subsequent release of inflammatory mediator, such as factor 7. Kouroumalis E, Hopwood D, Ross PE, Milne G, Bouchier IA. XII, platelet-activating factor, induce the inflammatory response of Gallbladder epithelial acid hydrolases in human cholecystitis. J the gallbladder5). The neurological patients admitted to the ICU is Pathol 1983;139:179–191. prone to lung problems, including pneumonia, atelectasis, pleural 8. Yang NR, Hong KS, Seo EK. Acute cholecystitis as a cause of effusion, catheter-related infections, operations-related infection, fever in aneurysmal subarachnoid hemorrhage. Korean J Crit drug fever, inflammation due to hemorrhage without infectious Care Med 2017;32:190–196. causes, pseudomembranous colitis, bacteremia, and urinary tract 9. Messing B, Bories C, Kunstlinger F, Bernier JJ. Does total paren- infection (UTI), which could be related to the systemic inflamma- teral nutrition induce gallbladder sludge formation and lithiasis? tory response. Pneumonia and UTI took up more than half of the Gastroenterol 1983;84:1012–1019. infections in the ICU. Systemic inflammatory response can cause 10. Sevastos N, Savvas SP, Rafailidis PI, Manesis EK. Cholestasis in cholestasis, gallbladder ischemia, therefore could affect the inci- acute stroke: an investigation on its prevalence and etiology. dence of AAC in neurological patients2). Scand J Gastroenterol 2005;40:862–866. 11. Orlando R, Gleason E, Drezner AD. Acute acalculous cholecys- CONCLUSION titis in the critically ill patient. Am J Surg 1983;145:472–476. 12. Souza LJ, Braga LC, Rocha Nde S, Tavares RR. Acute acalculous AAC may be a significant complication in patients with neuro- cholecystitis in a teenager with hepatitis a virus infection: a case logical conditions. It is important to identify symptoms and signs report. Braz J Infect Dis 2009;13:74–76. of AAC, accurately diagnose the AAC, particularly in long-term 13. Imamoglu M, Sarihan H, Sari A, Ahmetoglu A. Acute acalcu- stay at ICU, a long period of fasting, state of hypotension or hy- lous cholecystitis in children: Diagnosis and treatment. J Pediatr poperfusion, PEEP ventilation. Surg 2002;37:36–39. 14. Parithivel VS, Gerst PH, Banerjee S, Parikh V, Albu E. Acute

80 www.e-jnic.org Acute Acalculous Cholecystitis in Neurological Patients See Won Um et al.

acalculous cholecystitis in young patients without predisposing lous Cholecystitis in critically ill patients: risk factors, diagnosis factors. Am Surg 1999;65:366–368. and treatment strategies. J Pancreas 2016;17:580–586. 15. de Oliveira Júnior SA, Lemos TEV, de Medeiros Junior AC, Freire AD, de Carvalho Garcia C, e Silva R, et al. Acute Acalcu-

www.e-jnic.org 81 eISSN 2635-5280 Case Report J Neurointensive Care 2019;2(2):82-86 https://doi.org/10.32587/jnic.2019.00143

Delayed Occurrence of CSF Oculorrhea after Mild Blunt Traumatic Head Injury Sang Bin Choi, Sang Pyung Lee, Jin Wook Baek, Tae Joon Park Department of Neurosurgery, Cheju Halla General Hospital, Jeju, Korea

Received: June 3, 2019 Accepted: July 8, 2019 Cerebrospinal fluid (CSF) oculorrhea is very rare after traumatic head injury. CSF fistula can cause hazardous complications including infection of central nervous system(CNS). A 40-year-old male Corresponding Author: visited with epiphora occurring every morning. He underwent traumatic head injury about a year Tae Joon Park, M.D. and a half ago. Intrathecal Gadolinium-Enhanced Magnetic Resonance (MR) Cisternography was Department of Neurosurgery, performed to diagnose his CSF oculorrhea. The delayed CSF oculorrhea was successfully treat- Cheju Halla General Hospital, 65, ment with continuous lumbar drainage. This case report discusses the diagnosis and treatment of Doryeong-ro, Jeju 63127, Korea the very rare case of the delayed CSF oculorrhea after mild blunt traumatic head injury. Tel: +82-64-740-5000 Fax: +82-64-740-5655 Keywords: CSF oculorrhea; Delayed; CSF fistula; Traumatic head injury E-mail: [email protected]

INTRODUCTION from excessive lacrimation caused by orbital soft tissue trauma or epiphora caused by lacrimal outflow obstruction2,8). Cerebrospinal fluid (CSF) fistula is an uncommon but well-doc- Therefore, any patient who presents with excessive posttrau- umented phenomenon which occur after head trauma, typically matic drainage through the orbit should be suspected to have CSF manifesting as otorrhea or rhinorrhea2,5). CSF rhinorrhea and otor- oculorrhea9). rhea are relatively common complications of head injuries to com- In this report, the diagnosis and management of one rare case pare with CSF oculorrhea5). Although it is rare, CSF could enter of delayed CSF oculorrhea after mild blunt cranio-orbital trauma the orbit following a anterior cranial fossa injury and CSF can exit is discussed. via the eye, mimicking tear formation10). The paucity of reported CSF oculorrhea cases, especially com- CASE REPORT pared with the frequency of other types of CSF fistulas developing after traumatic head injury, suggests that cranio-orbital fistulas are On October 5, 2018, a 40-years-old male who underwent trau- rare and may be underdiagnosed complications8). matic head injury about a year and a half ago, visited our outpa- CSF fistula can cause hazardous complications including infec- tient department. The patient complained of epiphora occurring tion of central nervous system (CNS), intracranial hypotension, every morning. In glucose stick test at local medical center, blood seizures, etc1,8). Therefore, early detection and treatment is re- sugar level has been checked 110 mg/dL, and tear's sugar had quired. But CSF oculorrhea is very rare1,3,5,6,9) and it is difficult to been checked 68mg/dL. Under suspicion of CSF leakage, the pa- diagnose early, because it need to differentiate CSF oculorrhea tient was admitted for a thorough examination.

82 www.e-jnic.org Delayed CSF oculorrhea after mild head trauma Sang Bin Choi et al.

In his past history, there was a car accident against truck on the DISCUSSION highway on February 25, 2017. The traumatic epidural hematoma(EDH) and subdural hematoma(SDH) with a linear skull frac- CSF leakage typically occurs at four anatomical locations, which tures (lateral wall of left sided orbit, temporal, and sphenoid bone) is the frontal sinus, the cribriform plate and ethmoid roof, the sphe- was revealed in his craniofacial computerized tomography(CT) noid sinus, and the petrous bone8,9), and it occurs due to the defect scanning (Fig. 1) but the amount of these hematoma was only of anatomical barriers (the skin or mucosa of the air sinuses, the scanty and conservative management was done. periosteum, the skull bones, the dura mater, and the arachnoid Although the patient suffered from left sided facial paralysis and membrane)8,9). When CSF leakage occurred after head trauma, otorrhagia, due to left sided temporal bone fracture. These symp- two thirds of the cases occurred within the first 48hours, and the toms were improved gradually and in the end, there was no CSF most cases, the leakage detected within 3months8). leak at the time of discharge. CSF otorrhea (CSF leaks through the temporal bone or mastoid Based on the patient's past history and the glucose level of the fracture associated with dural tears into the middle ear), and CSF eye drop, the CSF oculorrhea was suspected strongly. To confirm rhinorrhea (CSF leaks through fracture of cribriform plate associ- CSF leak, Intrathecal Gadolinium-Enhanced Magnetic Resonance ated with dural tears into the paranasal sinuses and nasal cavity)2) (MR) Cisternography (Fig. 2) was performed and CSF collection occur commonly, because of a relatively high-pressure gradient be- was identified in inferior part of both orbits and nasopharynx with tween the intracranial space and the external ear canal or paranasal unidentified fistula site in MRI findings. In spite of conservative sinuses2). But CSF oculorrhea is rare, because of the relative low treatment with absolute bed rest, CSF oculorrhea was persisted, pressure gradient between the intracranial subarachnoid space and and we performed continuous lumbar drainage and it was main- the intraorbital space2). tained for two weeks. In the history of oculorrhea, a case of profuse CSF drainage The lumbar drainage catheter removed in two weeks and there through the orbit following avulsion of the eye was documented was no more CSF leakage during the hospitalization and there first by Walsh and Hoyt9). And tear like-CSF leakage have been de- were no complications such as cerebral meningitis either. After he scribed by Joshi KK, Crockard HA7). The CSF leakage through the discharged, the patient visited our outpatient department 2 weeks orbit was termed ‘oculorrhea’ by Salame K et al.9). and 6 weeks later and there was no CSF leakage any more. In the reported 16 case of the oculorrhea2,4,8,9,10)(Table 1). It oc-

A B

Fig. 1. Craniofacial computed tomography. A. Axial view shows fracture involving right ethmoid bone(arrow) and lateral orbital wall fracture(arrowhead). B. Coronal view show comminuted fracture involving right frontal bone extent into orbital roof(arrow) and maxillary sinus(arrowhead). www.e-jnic.org 83 Delayed CSF oculorrhea after mild head trauma Sang Bin Choi et al.

A B

C D

Fig. 2. Intrathecal Gadolinium-Enhanced MR Cisternography. Axial A and Coronal B T1W1 immediately following the intrathecal demonstrate early partial enhancement (arrow). Axial C and Coronal D T1W1 following a 24h delay shows retention of contrast media in inferior part of both orbits and nasopharynx (arrow). curred in the relative young ages by severe traumatic head injury and the prognosis of patients should be poor1,8). Therefore we within a week after traumatic head injury. should be suspect CSF oculorrhea in patients with eye drops who In our case, the patient was a 40 aged-man and he underwent has the past history of skull base fracture especially frontal or orbit- mild blunt head trauma compare with other reported cases. In ad- al roof even it is only mild blunt injury8), and if the CSF oculorrhea dition, it was detected in delayed period after trauma. is suspected, the diagnosis should be done immediately thorough The delayed CSF oculorrhea can be explained by the persistent examination and active management is required9). fistula between the cranium and orbit10), but to establish the diag- In the diagnosis of CSF oculorrhea, the analysis of the fluid for nosis of cranio-orbital fistula can be challenging because of the dif- glucose concentration may be helpful4,8,9) and Immunoassay for ficulty in differentiating oculorrhea from excessive lacrimation beta 2–transferrin is more specific and requires only a small caused by orbital soft tissue trauma or epiphora caused by lacrimal amount of fluid8,9). CT cisternography with iopamidol followed by outflow obstruction2). thin-sectioned coronal-view CT scanning through the bone win- Delayed CSF oculorrhea should be concerned for all patients dow is the best method for localization of the fistula8,9). As in our with excess tearing after anterior skull fractures8). Because it is diffi- patient, Intrathecal Gadolinium-Enhanced MR Cisternography cult to know exactly when it happened. If the CSF oculorrhea de- can perform in the evaluation of CSF leakage1,3,5). tection is missed, the number of complications could be happen In most case, the initial management of CSF leak is conservative

84 www.e-jnic.org Delayed CSF oculorrhea after mild head trauma Sang Bin Choi et al.

Table 1. CSF Oculorrhea case series list, clinical description Author Year Age Sex Mechanism of injury Ocular sign on admission Onset Bard 1963 36Y M stabbing wound Laceration immediate Barker-Grifﬁth 2007 14Y M potatogun injury Multiple lacerations and ecchymosis of lids immediate Bongartz 1981 2Y M falling injury Swelling of right eyelid immediate Civelek 2006 7Y M stabbing wound Proptosis; diplopia, orbital cellulitis; periorbital abscess immediate Dryden 1986 4Y M Motor vehicle Right inferior oblique overreaction; meningeal signs Immediate Garza-Mercado 1982 20Y M assault Edema and ecchymosis of eyelids, limited extraocular movements Immediate Joshi 1978 8M F Motor vehicle Unconscious mentality mydriasis, fixed pupil, laceration 5days later Joshi 1978 5Y undisclosed N/A Immediate Kjer 1954 3Y M stabbing wound Unconscious mentality, Clonic spasm, upper lid edema, fixed pupil Immediate Rao 1999 78Y M stabbing wound Meningeal signs, Immediate Rha 2012 56Y M Motor vehicle Unconscious mentality papilledema, proptosis, mydriasis Immediate Salame 2000 20Y F Trauma: sports Periorbital swelling and ecchymosis, laceration, mydriasis Immediate Pease 2012 22Y M Motor vehicle Laceration, diminished ocular movements Immediate Alexandra 2014 34Y M Motor vehicle periorbital swelling, limited extraocular movement, decreased sensation of V2 3days later Jain 2016 52Y F Post-surgery; third nerve palsy immediate frontotemporal craniotomy Sheth 2018 16Y M Motor vehicle extensive laceration, proptosis 3days later treatment8,9). However, the rate of complications can increase in cas- REFERENCES es of delayed CSF oculorrhea, so active treatment is needed. It is important to preventing CSF leak through cranio-ocular fistula and to 1. Antonelli V, Cremonini AM, Campobassi A, Pascarella R, Zofrea reduce at risk for infection within the central nervous system2,8,9). G, Servadei F. Traumatic encephalocele related to orbital roof We suggest that CSF diversion is proper method for the manage- fractures: report of six cases and literature review. Surg Neurol ment of a traumatic cranio-orbital fistula6). CSF diversion can be 2002;57:117–125. performed continuous lumbar drainage for up to 10 days10). In spite 2. Apkarian AO, Hervey-Jumper SL, Trobe JD. Cerebrospinal fluid of continuous lumbar drainage, if the CSF continues to leak, a surgi- leak presenting as oculorrhea after blunt orbitocranial trauma. J cal approach can be taken to primarily repair the dural defect8,10). Neuroophthalmol 2014;34:271–273. 3. Chandra N, Ojha BK, Chandwani V, Srivastava C, Singh SK, CONCLUSION Chandra A. A rare case of posttraumatic eyelid swelling: cerebrospinal fluid blepharocele. J Neurosurg Pediatr 2013;11:242–244. CSF leakage requires early diagnosis and early treatment, be- 4. Dryden RM, Wulc AE. Pseudoepiphora from cerebrospinal fluid cause it can cause serious complications4). CSF oculorrhea is very leak: case report. Br J Ophthalmol 1986;70:570–574. rare1,3,5,6,9). But, CSF leakage like tears can be occur in patients 5. Govindaraju V, Bharathi R. Post-traumatic blepharocele in an which is middle-aged man with trauma history even mild blunt adult. Sultan Qaboos Univ Med J 2013;13:E479–481. head injury especially in the morning. We should be alert in early 6. Jain S, Patel RM, Hage Z, Lim J, Lee S, Amin-Hanjani S, et al. Ce- diagnosis of CSF leakage and immediate and adequate treatment is rebrospinal Fluid Leak Presenting as Epiphora. Ophthalmic Plast needed4). Reconstr Surg 2016;32:e26–28. 7. Joshi KK, Crockard HA. Traumatic cerebrospinal fluid fistula NOTES simulating tears. Case report. J Neurosurg 1978;49:121–123. 8. Pease M, Marquez Y, Tuchman A, Markarian A, Zada G. Diagno- Conflict of interest sis and surgical management of traumatic cerebrospinal fluid oculorrhea: case report and systematic review of the literature. J Neu- No potential conflict of interest relevant to this article was re- rol Surg Rep 2013;74:57–66. ported. 9. Salame K, Segev Y, Fliss DM, Ouaknine GE. Diagnosis and management of posttraumatic oculorrhea. Neurosurg Focus 2000;9:e3. www.e-jnic.org 85 Delayed CSF oculorrhea after mild head trauma Sang Bin Choi et al.

10. Sheth AA, Ngo V, Lam M. Traumatic cerebrospinal fluid oculorrhea managed with an external ventricular drain. J Surg Case Rep 2018;2018:rjy215.

86 www.e-jnic.org eISSN 2635-5280 Case Report J Neurointensive Care 2019;2(2):87-92 https://doi.org/10.32587/jnic.2019.00157

Endovascular Treatment of Ruptured Dissecting Aneurysms on the P2-3 Segment of the Posterior Cerebral Artery: A Report of Two Cases Seonah Choi1, Joonho Chung1,2 1Department of Neurosurgery, Severance Hospital, Yonsei University College of Medicine, Seoul, Korea 2Severance Institute for Vascular and Metabolic Research, Yonsei University College of Medicine, Seoul, Korea

Received: July 16, 2019 Accepted: July 31, 2019 Intracranial ruptured dissecting aneurysms can have fatal outcomes. Among them, posterior cerebral artery dissecting aneurysm with hemorrhage is uncommon but usually fatal due to its deep Corresponding Author: location and proximity to surrounding critical structures. Most previously published reports on Joonho Chung, M.D., Ph.D. this type of aneurysm have concluded that the neurologic outcome can be improved with the use Department of Neurosurgery, of certain surgical or endovascular approaches. However, in severe cases with a poor initial clinical Stroke Center, Severance Hospital status, patient improvement should not be expected. We report two cases of P2-3 segment rup- Yonsei University College of tured dissecting aneurysms that were treated by external ventricular drainage for control of in- Medicine 50-1, Yonsei-ro, creased intracranial pressure followed by one-stage endovascular treatment for the aneurysm. Seodaemun-gu, Seoul, 03722, Korea Tel: +82-2-2228-2150 Keywords: Dissecting aneurysm; Endovascular treatment; Parent artery occlusion; Posterior ce- Fax: +82-2-393-9979 rebral artery E-mail: ns.joonho.chung@gmail. com

INTRODUCTION stent-assisted coiling or using flow diverter devices as a reconstruction technique, and deconstruction techniques, such as par- Posterior cerebral artery (PCA) dissecting aneurysm is rare and ent artery occlusion. In the case of a ruptured dissecting aneu- typically presents with ischemic symptoms in the PCA territory rysm, either deconstruction or reconstruction is feasible, and their and subarachnoid hemorrhage10,15). Once aneurysmal rupture oc- strengths and weaknesses are well known among neurosurgeons curs along with subarachnoid hemorrhage, a poor outcome can and neurointerventionists. As a result, the optimal approach for be expected such as a high modified Rankin Scale (mRS) score or each case may vary according to anatomical and morphological a low overall survival rate; therefore, immediate and proper treat- features of each aneurysm. However, devastating results (mRS ment is necessary. Due to aneurysmal location and associated cru- 3-6) are very likely despite the best treatments. We report 2 cases cial anatomy, initial neurological status is unfavorable in most of of P2-3 segment dissecting aneurysms that were treated with ex- these patients. In these cases, endovascular treatment (EVT) is ternal ventricular drainage (EVD) for control of increased intra- usually adopted for fast intervention of the aneurysm’s natural cranial pressure (ICP) followed by EVT but still experienced course. EVT includes simple coiling, balloon remodeling, morbidity and mortality.

CASE REPORTS giography showed a wide neck right P2 aneurysm suspicious for dissection. The patient underwent EVD prior to digital subtrac- Case 1 tion angiography (DSA) due to acute hydrocephalus and poor A 50-year-old female patient with no previous medical prob- Hunt-Hess grade. DSA revealed a right PCA dissecting aneurysm lems presented to the emergency department with deep stupor (Fig. 1B and C), and one-stage stent-assisted coil embolization mental status and was admitted. Her initial Hunt-Hess grade was (Enterprise stent 4.5 mm x 28 mm) was carried out. The aneu- 4. Her computed tomography (CT) scan revealed intraventricular rysm was completely obliterated, and the PCA was patent (Fig. hemorrhage and intraparenchymal hemorrhage (Fig. 1A). CT an- 1D). The patient remained vegetative (modified Rankin Scale 5)

A B

C D

Fig. 1. A 50-year-old female with an initial Hunt-Hess grade 4. (A) Computed tomography obtained in the emergency department shows intraventricular, intraparenchymal and subarachnoid hemorrhages. (B) A right P2 segment dissecting aneurysm was noted on three- dimensional reconstruction image and (C) on digital subtraction angiography. (D) The aneurysm was completely occluded by stent-assisted coil embolization.

88 www.e-jnic.org Ruptured dissecting aneurysm on P2-3 Seonah Choi et al. at the 24-month clinical follow-up, and a repeat DSA was not per- mus intraparenchymal hemorrhage combined with intraventricu- formed. lar and subarachnoid hemorrhage and increased basal cistern at- tenuation (Fig. 2A). Her initial Hunt-Hess grade was 5, and her Case 2 modified Fisher grade was 4. We performed EVD and catheteriza- A 64-year-old female patient presented to the emergency de- tion for intraparenchymal hemorrhage. DSA showed a left P2-3 partment with semicomatous mental status and pupil dilatation. dissecting aneurysm greater than 20 mm in size, and a distal PCA Her CT scan revealed a large amount of basal ganglia and thala- was noted at the end of the aneurysm (Fig. 2B). Parent artery oc-

A B

C D

Fig. 2. A 64-year-old female with a initial Hunt-Hess grade 5. (A) Computed tomography reveals a large amount of intraparenchymal, intraventricular and subarachnoid hemorrhage, including basal cisterns. (B) Digital subtraction angiography reveals a left P2-3 segment dissecting aneurysm; a distal PCA was noted at the end of the aneurysm. (C) Parent artery occlusion was performed immediately by coil embolization. www.e-jnic.org 89 Ruptured dissecting aneurysm on P2-3 Seonah Choi et al. clusion was performed immediately by coil embolization (Fig. 2C). syndrome, mixed connective tissue disease, fibromuscular dyspla- However, the patient died 10 days after the procedure due to in- sia, homocystinuria, polycystic kidney disease, and trauma15,16). creased ICP and brainstem dysfunction. However, our two cases had no previous associated medical histo- ries or medications. DISCUSSION The most common presentation is subarachnoid hemorrhage with headache and neurological deficits that defer depending on PCA aneurysms are uncommon, with an incidence < 1 %. the involved segments of the PCA2,10,16). In general, PCA aneu- Most PCA aneurysms are non-saccular type10). More than 80- rysms arise mostly on P1 or at the P1-2 junction8,9), but dissecting 90% of all intracranial dissections occur in the posterior circula- aneurysms usually involve P2 or P2-3 segments10,16); therefore, as- tion, but almost all are located near the vertebral and basilar arter- sociated neurologic deficits include memory impairment, oculo- ies and have a male predominance13,15). In contrast, PCA dissec- motor palsy, visual disturbance (homonymous hemianopsia) and tion is extremely rare, occurs at a younger age, and has a female sometimes hemiparesis due to mass effect2). Acute hydrocephalus predominance; only a few cases have been reported10,15). Among should be a concern when presented because it aggravates mental these reported cases, only a few ruptured P2-3 dissecting aneu- deterioration. Our cases, both suffered mental status deteriora- rysms are documented. The suggested causes of this type of dis- tion, so focal neurologic deficits could not be examined. The nat- section include syphilis, migraine, cystic medial necrosis, Marfan ural history of dissecting aneurysm is not yet well described but

Table 1. List of P2-3 dissecting aneurysm ruptured including the present case Authors Age Sex Intial HHG PCA segments Treatment Outcomes (mRS) Berger and Wilson (1984) [1] 21 F 2 P2 Surgical clipping(vessel occlusion) 2 Lanzinski et al. (2000) [9] 48 F 3 P2-3 Parent artery occlusion 1 17 M 4 P2-3 Parent artery occlusion 0 13 F 4 P2 Parent artery occlusion 0 Kitazawa et al. (2001) [7] 43 M 1 P2 surgical clipping (vessel occlusion) not mentioned Hallacq et al. (2002) [5] 20 M 2 P2 Coil embolization 0 60 M 2 P2 Coil embolization 0 49 F 3 P2 Coil embolization 1 47 M 2 P2 Coil embolization 1 Hamada et al. (2005) [6] 33 F 4 P2 Surgical clipping (vessel occlusion) 0 Lv et al. (2009) [12] 37 F 4 P2 Parent artery occlusion 0 49 M 4 P2 Parent artery occlusion 0 43 M 4 P2 Parent artery occlusion 0 48 M 3 P2 Parent artery occlusion 0 58 F 4 P2 Parent artery occlusion 1 31 M 2 P2 Parent artery occlusion 0 4 M 2 P2 Parent artery occlusion 0 5 M 2 P2 Parent artery occlusion 0 Taqi et al. (2011) [16] 23 F 4 P2 Stent-assisted coiling not mentioned 48 F 2 P2 Stent-assisted coiling not mentioned 5 F 2 P2-3 Parent artery occlusion not mentioned 45 F 4 P2 Parent artery occlusion not mentioned 25 M 4 P3 Parent artery occlusion not mentioned 54 F 2 P2-3 Stent-assisted coiling not mentioned Krishnan et al. (2014) [8] 36 F 4 P2 Surgical clipping (vessel occlusion) 2 Osman et al. (2017) [14] 49 F 2 P2 Stent-assisted coiling 0 Cunegatto-Braga et al. (2018) [3] 4 M 2 P2 Flow diverter device 0 Choi and Chung 50 F 4 P2-3 Stent-assisted coiling 5 (the present cases) 64 F 5 P2-3 Parent artery occlusion 6

90 www.e-jnic.org Ruptured dissecting aneurysm on P2-3 Seonah Choi et al. notorious for poor outcome of 30-70% re-bleeding risk if not CONCLUSION treated13). Diagnosis of dissecting aneurysm is usually made based upon Patients with ruptured P2-3 segment dissecting aneurysms DSA findings of “pearl and string (a fusiform dilatation associated might be expected to experience unfavorable outcomes due to the with proximal or distal narrowing),” “double lumen (visualization aneurysm’s deep location within surrounding critical structures. of two channels),” and “fusiform dilatation”13,16). CT angiography However, EVD for ICP control followed by one-stage EVT for may be an alternative option, but it may not reveal an aneurysm if the aneurysm could be a safe and feasible treatment option. a mass effect exists, as we encountered with Case 2. Magnetic resonance angiography can also demonstrate the location and length NOTES of the dissection, but it has drawbacks of motion artifacts and limitations due to the small caliber of the vessels in this area15). There- Conflict of interest fore, DSA is the gold standard. we declare that we have no conflict of interest. Due to its hard-to-access location, typical presence of numer- ous perforators, and limitations in clipping dissecting aneu- Ethical approval rysms4), endovascular approaches have been attempted for PCA all procedures performed in the studies involving human partic- aneurysm treatment. The primary goal is isolation of the aneu- ipants were in accordance with the ethical standards of our Insti- rysms and preservation of the parent artery10). If the situation al- tutional Review Board with the 1964 Helsinki Declaration and its lows, simple coiling, balloon remodeling, and stent-assisted coil- later amendments or comparable ethical standards. ing may also be performed to achieve the goal10). However, parent artery occlusion is inevitable despite the risk of PCA territory in- Informed consent farction. Usually, occlusion of the P2 segment or the distal seg- in this retrospective study, the requirement for informed con- ment of the PCA is tolerable due to purulent collateral circula- sent was waived. tions; if not, a combined distal bypass could be a rescue treat- ment10). For the last several years, flow diverters have been used in ACKNOWLEDGEMENTS various vessels even smaller than the PCA3,11), and this could be an option for dissecting PCA aneurysms. however, the required No funding was received for this study. accompanying anticoagulant/antiplatelet therapy could worsen the patient’s situation, so treatment decisions should involve great REFERENCES care3,10,13,15). Previous articles have shown that, with proper management, al- 1. Berger MS, Wilson CB. Intracranial dissecting aneurysms of the most all P2-3 segment dissecting aneurysm cases had favorable posterior circulation. Report of six cases and review of the litera- clinical outcomes (Table 1)1,3,5-9,12,14,16). The difference of the pres- ture. J Neurosurg 1984;61:882–894. ent cases from those previously published is clinical outcome 2. Ciceri EF, Klucznik RP, Grossman RG, Rose JE, Mawad ME. An- (mRS 5 and 6) despite complete occlusion of the aneurysm eurysms of the posterior cerebral artery: classification and endo- during the acute period. Considering that isolated PCA dissec- vascular treatment. AJNR Am J Neuroradiol 2001;22:27–34. tions are known to present with ischemic symptoms and still have 3. Cunegatto-Braga M, Hogan B, Aguilar-Salinas P, Bejer AD, Hanel favorable outcomes15), our cases could be classified as unusual RA. Pipeline embolization device flow diversion for a dissecting among PCA dissections. The patients’ initial clinical status (such ruptured posterior cerebral artery aneurysm in a pediatric patient. as Hunt-Hess grade), amount of hemorrhage, and presence of World Neurosurg 2018;117:255–260. brainstem injuries likely had a greater impact than other factors on 4. Goehre F, Jahromi BR, Lehecka M, Lehto H, Kivisaari R, An- their clinical outcomes. Unfortunately, we were unable to achieve drade-Barazarte H, et al. Posterior Cerebral Artery Aneurysms: good outcomes in our cases, but we believe and recommend that Treatment and Outcome Analysis in 121 Patients. World Neuro- a combined surgical and endovascular treatment approach should surg 2016;92:521–532. be considered first to control both increased ICP and future 5. Hallacq P, Piotin M, Moret J. Endovascular occlusion of the pos- re-bleeding. terior cerebral artery for the treatment of P2 segment aneurysms: retrospective review of a 10-year series. AJNR Am J Neuroradiol 2002;23:1128–1136. www.e-jnic.org 91 Ruptured dissecting aneurysm on P2-3 Seonah Choi et al.

6. Hamada J, Morioka M, Yano S, Todaka T, Kai Y, Kuratsu J. Clini- 1043. cal features of aneurysms of the posterior cerebral artery: a 15- 12. Lv X, Li Y, Jiang C, Yang X, Wu Z. Parent vessel occlusion for P2 year experience with 21 cases. Neurosurgery 2005;56:662–670. dissecting aneurysms of the posterior cerebral artery. Surg Neu- 7. Kitazawa K, Tanaka Y, Muraoka S, Okudera H, Orz Y, Kyoshima rol 2009;71:319–325. K, et al. Specific characteristics and management strategies of ce- 13. Oran I, Cinar C, Yağci B, Tarhan S, Kiroğlu Y, Serter S. Ruptured rebral artery aneurysms: report of eleven cases. J Clin Neurosci dissecting aneurysms arising from non-vertebral arteries of the 2001;8:23–26. posterior circulation: endovascular treatment perspective. Di- 8. Krishnan P, Kartikueyan R, Thamatapu E, Chowdhury SR. Dis- agn Interv Radiol 2009;15:159–165. secting aneurysm of the posterior cerebral artery: Defining the 14. Osman MW, Kadziolka K, Peirot L. Optional endovascular role of deliberate surgical proximal artery occlusion. J Neurosci therapy of dissecting posterior cerebral artery aneurysm. Inter- Rural Pract 2014;5:443–445. ven Neurol 2017;6:219–228. 9. Lazinski D, Willinsky RA, TerBrugge K, Montanera W. Dissect- 15. Sherman P, Oka M, Aldrich E, Jordan L. Isolated posterior cere- ing aneurysms of the posterior cerebral artery: angioarchitecture bral artery dissection: Report of three cases. AJNR Am J Neu- and a review of the literature. Neuroradiology 2000;42:128–133. roradiol 2006;27:648–652. 10. Liu L, He H, Jiang C, Lv X, Li Y. Deliberate parent artery occlu- 16. Taqi MA, Lazzaro MA, Pandya DJ, Badruddin A, Zaidat OO. sion for non-saccular posterior cerebral artery aneurysms. Interv Dissecting aneurysms of posterior cerebral artery: clinical pre- Neuroradiol 2011l;17:159–168. sentation, angiographic findings, treatment, and outcome. Front 11. Lozupone E, Piano M, Valvassori L, Quilici L, Pero G, Visconti Neurol 2011;2:38. E, et al. Flow diverter devices in ruptured intracranial aneurysms: a single-center experience. J Neurosurg 2018;128:1037–

92 www.e-jnic.org eISSN 2635-5280 Case Report J Neurointensive Care 2019;2(2):93-98 https://doi.org/10.32587/jnic.2019.00164

Internal Carotid Artery Dissection with Traumatic Pseudoaneurysm Formation after Penetrating Head Injury Jeong Sik Ham1, Jang Hun Kim1,2,3, Won-Ki Yoon3 1Department of Neurosurgery, The Armed Forces Capital Hospital, Seongnam, Korea 2Trauma Center, The Armed Forces Capital Hospital, Seongnam, Korea 3Department of Neurosurgery, Korea University Guro Hospital, Seoul, Korea

Received: July 29, 2019 Accepted: August 8, 2019 Traumatic pseudoaneurysm of the intracranial arteries is a devastating event leading to high mortality. Its prevalence is rare, and associated with blunt or penetrating head injuries. Here, we de- Corresponding Author: scribe a case of a 46-year-old woman who was transferred to our hospital following a stab-induced Jang Hun Kim, M.D. head trauma. Owing to multiple assaults by a sharp knife, the patient presented with an injured in- Trauma Center, The Armed tracranial internal carotid artery that formed dissecting pseudoaneurysms. Stupor mentality was Forces Capital Hospital 81 noted, and brain computed tomography revealed traumatic subarachnoid hemorrhage, and intra- Saemaeul-ro, 177 beon-gil, cerebral hemorrhage in a “knife shape”. Diagnostic angiography showed an active extravasation in Bundang-gu, Seongnam-si, the middle cerebral artery and distal internal carotid artery with two pseudoaneurysms. Immedi- Gyeonggi-do 13574 Korea ate trapping was planned, and embolization was performed with detachable coils. The patient was Tel: +82-10-3607-6745 transferred to the operation room and emergent decompressive craniectomy with wide durotomy E-mail: [email protected] was performed. Unfortunately, the patient expired few days after the operation despite maximal neurocritical care.

Keywords: Internal carotid artery dissection; Penetrating head injury; Pseudoaneurysm

INTRODUCTION cular injury in TBI patients, which can lead to devastating results. Therefore, in a case of TBI with an accompanying vascular injury, Trauma-related pseudoaneurysms of intracranial arteries are ex- care and attention is needed to avoid poor prognosis. tremely rare, and are known to be associated with blunt or pene- Here, we describe a rare case of a woman presenting with a trating head injuries2). The reported prevalence of traumatic cere- stab-induced penetrating injury to her cranium; she was assaulted brovascular injury is 0.8% among all traumatic brain injury (TBI) by a sharp knife that directly injured her intracranial internal ca- patients10). The clinical presentation shows a wide range of signs rotid artery (ICA) and led to the formation of dissecting pseudo- and symptoms, including headache, loss of consciousness, altered aneurysms. mentality, and subarachnoid hemorrhage (SAH) in brain computed tomography (CT), and such neurological deficits may ap- CASE REPORT pear either immediately, or months to years later1). Due to the aforementioned reasons, surgeons sometimes fail to identify vas- A 46-year-old female visited emergency center (Korea Univer-

4 3

2 5

7 6

Fig. 1. Seven stab injuries are noted: (1) right lower eyelid, (2) left posterior triangle of lateral neck, (3) frontal, (4) lower temporal, (5) upper temporal, (6) parietal, and (7) occipital area of head, respectively.

A B

Fig. 2. Initial brain CT reveals left temporal bone fracture, SAH on both sylvian fissures, IVH in the third ventricle, and traumatic “knife-shaped” ICH in the left hemisphere (A). Bone setting of the brain CT presents a comminuted skull fracture of the left temporal bone of the size of the knife (B). CT, computed tomography; SAH, subarachnoid hemorrhage; IVH, intraventricular hemorrhage; ICH, intracerebral hematoma.

94 www.e-jnic.org Traumatic Intracranial Pseudoaneurysm Jeong Sik Ham et al. gency center and she was transferred to the interventional suite the left ICA and attempted angiography again. Angiograms immediately. The femoral artery was punctured within 10 min- showed an active extravasation in the left first segment of the mid- utes after CT evaluation. dle cerebral artery (MCA) and distal ICA. Multiple pseudoaneu- First, diagnostic angiography was performed (Fig. 3). A 7 Fr rysms were noted. sheath and 5 Fr guide catheter (Glidecath Headhunter 1; Terumo, Considering that an emergent decompressive craniectomy Somerset, NJ) was inserted into the right ICA. Angiograms re- should be performed due to malignant brain swelling with in- vealed collateral blood flow running from the right to left via the creased intracranial pressure (IICP), immediate trapping of left anterior communicating artery. With left common carotid artery ICA was planned. A microcatheter (Prowler Select Plus; Codman, (CCA) angiography, the left ICA was not observed. ICA dissec- Raynham, MA) with a microwire (Synchro-2; Stryker, Fremont, tion and occlusion was suspected. CA) was advanced to the left distal ICA. In total, 6 EA detachable After diagnostic angiography, we inserted an intermediate cath- coils were embolized. We also added 33% cyanoacrylate glue for eter (Envoy Distal Access; DePuy Synthes, Raynham, MA) into additional hemostasis to the embolized coils. On the final angio-

Fig. 3. Right ICA angiogram reveals normal vasculature of the right intracerebral arteries. Collateral blood flows from the right to left hemisphere through the A-comm artery (A). Left CCA angiogram shows preservation of all the ECA branches except STA. However, the ICA is not visible on the proximal portion and we suspect ICA injury somewhere in the distal part (B). ICA, internal carotid artery; A-comm, anterior communicating; CCA, common carotid artery; STA, superficial temporal artery. www.e-jnic.org 95 Traumatic Intracranial Pseudoaneurysm Jeong Sik Ham et al.

B C

Fig. 4. Left ICA angiogram notes that two large pseudoaneurysms were formed in the anterior and lateral directions (A). An intermediate catheter was inserted into the left ICA and a microcatheter was navigated to the traumatic pseudoaneurysm. Six coils were embolized in the pseudoaneurysm as well as in the distal ICA (B). During the final angiography, no leakage of contrast was noted (C). ICA, internal carotid artery. gram, no evidence of acute extravasation was noted (Fig. 4). The rocritical care, she expired 3 days after operation. whole procedural time (puncture to closure) was 48 minutes. She was transferred to the operation room and general anesthe- DISCUSSION sia was induced. Emergent decompressive craniectomy with wide durotomy was performed. However, the skin incision was made Penetrating injuries to the cranium are devastating events which 40 minutes after closing the femoral puncture site. In operation result in high rates of mortality and morbidity5). Initial GCS score findings, severe brain swelling was observed and intraparenchy- and head trauma severity are important prognostic factors4). When mal hematoma was noted. Wide craniectomy with ICH removal a patient experiences IICP, emergent decompression is readily nec- was rapidly carried out. Duroplasty with artificial dura was fol- essary. Moreover, if a cerebral arterial injury was accompanied, it lowed and closure of skin was performed after the closed suction would be a greater challenge for the surgeon to manage it. A hasty system was placed (Fig. 5). Postoperatively, she was comatose, craniotomy without managing vessel injury might worsen the situ- and both her pupils were dilated and fixed. Despite intensive neu- ation, and therefore, an integrated approach combining surgical

96 www.e-jnic.org Traumatic Intracranial Pseudoaneurysm Jeong Sik Ham et al.

A B

Fig. 5. Decompressive craniectomy with wide durotomy was followed (A). Postoperative brain CT reveals a black brain with severe brain swellings (B). and endovascular treatment is needed, particularly in cases of ICA Unfortunately, decompressive craniectomy was delayed almost dissection with pseudoaneurysm, as in this report. two hours because of vascular interventional therapies, and the A majority of the traumatic intracranial aneurysms have been transfer from the interventional suite to the operation room. In reported as a result of direct vascular injury secondary to penetrat- terms of neurointensive care, a hybrid operation setting might re- ing injury or comminuted skull fracture7,8). Blunt-trauma-related duce delays6); in this case, the delays might have worsened the pa- pseudoaneurysms rarely form as a result of shearing forces causing tient’s neurological status. Therefore, we believe that interdisci- torsion and overstretching of the arteries at the time of trauma7,8), plinary approaches and hybrid preparation would have otherwise, though the frequency of the cause of traumatic aneurysms may led to the patient’s survival, and are recommended in such cases of depend on local factors3). TBI with accompanying vascular injuries. In this case, the direct penetrating injury dissected the distal ICA and tore the vessel lumen. Histologically, trauma-related an- CONCLUSION eurysms have been categorized as true, false, and dissecting aneu- rysms2). Among these types, false aneurysms resulting from the Here, we describe a rare case of a woman presenting with a destruction of three layers of the vessel wall are induced primarily stab-induced penetrating injury to her cranium; she was assaulted by head trauma. Traumatic “pseudo- or false aneurysms” are con- by a sharp knife that directly injured her intracranial ICA and led sequently formed, which are thought to be sensitive to rupture to the formation of dissecting pseudoaneurysms. Left ICA trap- compared to true aneurysms. ping with coil embolization and followed decompressive craniec- Traumatic aneurysms should be suspected in cases of acute tomy was immediately performed, however, the patient was ex- neurological deterioration following a penetrating head injury, pired few days later. and patients should undergo immediate CT scanning9). If the scan demonstrates SAH in a pattern suggestive of aneurysmal SAH, an NOTES angiogram should follow to identify any vascular injury7,8). In this case, the CT feature was quite different from that of usual aneu- Conflict of interest rysmal SAH. Comminuted skull fracture of temporal bone was No potential conflict of interest relevant to this article was re- observed, and ICH was presented at the skull base in a “knife ported. shape” with accompanying SAH. At the tip of the knife, the circle of Willis was located and we could easily identify the intracranial ACKNOWLEDGEMENTS vascular injury. As expected, distal ICA was dissected, and immediate trapping was followed. None. www.e-jnic.org 97 Traumatic Intracranial Pseudoaneurysm Jeong Sik Ham et al.

REFERENCES 6. Kinoshita T, Hayashi M, Yamakawa K, Watanabe A, Yoshimura J, Hamasaki T, et al. Effect of the Hybrid Emergency Room Sys- 1. Asari S, Nakamura S, Yamada O, Beck H, Sugatani H. Traumatic tem on Functional Outcome in Patients with Severe Traumatic aneurysm of peripheral cerebral arteries. Report of two cases. J Brain Injury. World Neurosurg 2018;118:e792–e799. Neurosurg 1977;46:795–803. 7. Nishioka T, Maeda Y, Tomogane Y, Nakano A, Arita N. Unex- 2. Burton C, Velasco F, Dorman J. Traumatic aneurysm of a pe- pected delayed rupture of the vertebral-posterior inferior cere- ripheral cerebral artery. Review and case report. J Neurosurg bellar artery aneurysms following closed head injury. Acta Neu- 1968;28:468–474. rochir (Wien) 2002;144:839–845; discussion 845. 3. Cohen JE, Gomori JM, Segal R, Spivak A, Margolin E, Sviri G, 8. Paiva WS, Andrade AF, Sterman Neto H, de Amorim RL, Cal- et al. Results of endovascular treatment of traumatic intracranial das JG, Teixeira MJ. Traumatic pseudoaneurysm of the superior aneurysms. Neurosurgery 2008;63:476–485; discussion 485- cerebellar artery. J Trauma Acute Care Surg 2012;72:E115. 476. 9. Schwartz RB, Tice HM, Hooten SM, Hsu L, Stieg PE. Evalua- 4. Gutierrez-Gonzalez R, Boto GR, Rivero-Garvia M, Perez-Zamar- tion of cerebral aneurysms with helical CT: correlation with ron A, Gomez G. Penetrating brain injury by drill bit. Clin Neurol conventional angiography and MR angiography. Radiology Neurosurg 2008;110:207–210. 1994;192:717–722. 5. Kazim SF, Shamim MS, Tahir MZ, Enam SA, Waheed S. Man- 10. Tunthanathip T, Phuenpathom N, Saehaeng S, Oearsakul T, Sa- agement of penetrating brain injury. J Emerg Trauma Shock karunchai I, Kaewborisutsakul A. Traumatic cerebrovascular in- 2011;4:395–402. jury: Prevalence and risk factors. Am J Emerg Med 2019.

98 www.e-jnic.org eISSN 2635-5280 Case Report J Neurointensive Care 2019;2(2):99-102 https://doi.org/10.32587/jnic.2019.00185

Critical Care in Patient with Neuromuscular Cervicothoracic Kyphosis Kwang-Ui Hong, Jong-myung Jung, Seung-Jae Hyun, Ki-Jeong Kim, Tae-Ahn Jahng Department of Neurosurgery, Spine Center, Seoul National University Bundang Hospital, Seoul National University College of Medicine, Seongnam, Korea

Received: August 9, 2019 Accepted: August 28, 2019 A 10-year-old boy had a neuromuscular cervicothoracic kyphosis and kyphotic deformity got worse as he grew. He underwent posterior spinal fusion from T6 to pelvis two years ago. However, Corresponding Author: kyphosis progressed gradually, and difficulty occurred in breathing with a ventilator. We perform Seung-Jae Hyun, M.D., Ph.D. deformity correction with vertebral column resection at T5 and posterior fixation from T2 to T9 Department of Neurosurgery, and posterior onlay fusion. Surgical correction is offered to stop the kyphosis progression, and fi- Spine Center, Seoul National nally to maintain the airway. After surgery, the patient transferred to an intensive care unit for re- University Bundang Hospital, spiratory care. The patient’s breathing was much better than before surgery, and the patient was Seoul National University College transferred to the general ward. We report the importance of postoperative care in spinal deformi- of Medicine, 82 Gumi-ro ty patient with respiratory distress. 173beon-gil, Bundang-gu, Seongnam 13620, Korea Tel: +82-31-787-7164 Keywords: Neuromuscular kyphosis; Spinal deformity surgery; Pulmonary complications; In- Fax: +82-31-787-4097 tensive care E-mail: [email protected]

INTRODUCTION CLINICAL PRESENTATION

Spinal deformity surgery in children with neuromuscular spinal A 10-year-old boy had a neuromuscular cervicothoracic kypho- deformity (NSD) is associated with a high morbidity rate1,4,8-10). sis. The kyphotic deformity became worse as he grew. He became Respiratory complications following spinal deformity surgery in paraplegic, with complete loss of bowel and bladder sensation. patients with NSD have been reported from 15.6% to 28.2%2,4,8,9). Therefore, he underwent posterior spinal fusion from T6 to pelvis Respiratory complications may cause further impairment of respi- two years ago. However, kyphosis progressed gradually, and his re- ratory function, leading to prolonged intubation time, and the need spiratory function also gradually worsened. He underwent trache- for re-intubation. ostomy and breathed using a ventilator. The patient was unable to Postoperative intensive care may reduce postoperative morbidi- eat enough food by mouth. So, the gastrostomy tube was inserted. ty and allow fast recovery. We report a case of severe neuromuscu- However, kyphosis progressed gradually, and difficulty occurred in lar cervicothoracic kyphosis treated with deformity correction and breathing with a ventilator (Fig. 1). The patient suffered from postoperative intensive care. pneumonia repeatedly, and his nutrition gradually worsened. Before surgery, the patient’s pulmonary physician was contacted,

DISCUSSION

Morbidity following spinal deformity surgery for NSD is high, especially in respiratory complications1,4,8-10). Therefore, perioperative intensive lung care is essential for improving surgical outcome. The previous study suggests a perioperative protocol for optimizing respiratory function7). Preoperatively, it is recom- A B mended to start regimen, including percussion vest therapy, cough assist, an increase in nebulizer treatments, and frequent suction- Fig. 1. (A) Preoperative whole spine anterior-posterior radiograph. (B) ing. Postoperatively, patients without a history of pneumonia or a Preoperative whole spine lateral radiograph. tracheostomy follow a protocol that includes humidified oxygen

A B

Fig. 2. (A) Intraoperative photographs were taken after head fixation. (B) Intraoperative photographs were taken after vertebral column resection at T5 level and posterior spinal fixation from T2 to T9.

100 www.e-jnic.org Patient with neuromuscular kyphosis Kwang-Ui Hong et al.

showed no specific findings, and pneumonia did not occur. Post- operative three days, the patient was transferred to a general ward. In this case, we performed aggressive lung care, including prophylactic antibiotics, before and after surgery. Given the rarity of this deformity, the optimal regimen for perioperative lung care remains unclear. However, respiratory complications can be reduced by active lung care after spinal deformity correction has been achieved in NSD patients.

CONCLUSION

A tracheostomy tube present or a history of pneumonia are risk factors of postoperative respiratory complications in NMD patients. We suggest that intensive lung care should be performed even if there were no respiratory events during the deformity correction surgery.

A B NOTES

Fig. 3. (A) Postoperative whole spine anterior-posterior radiograph. Conflict of interest (B) Postoperative whole spine lateral radiograph. No potential conflict of interest relevant to this article was reported. and nebulized albuterol treatment four times per day. ACKNOWLEDGEMENTS NSD patients with tracheostomies were more likely to experience respiratory complications, although these preventative and None. aggressive regimens were applied preoperatively and postoperatively. The presence of tracheostomy is associated with negative REFERENCES outcomes after treatment of various conditions3,5). History of pneumonia is a clinically important independent predictor of in- 1. Barsdorf AI, Sproule DM, Kaufmann P. Scoliosis surgery in creased risk of respiratory complications7). The association be- children with neuromuscular disease: findings from the US Na- tween tracheostomy presence and postoperative pneumonia has tional Inpatient Sample, 1997 to 2003. Arch Neurol 2010;67: been reported3,6). The previous study reported the association of 231–235. gastrostomy tube and respiratory complications7). The presence of 2. Bendon AA, George KA, Patel D. Perioperative complications a gastrostomy was a risk factor of respiratory complications after and outcomes in children with cerebral palsy undergoing scoli- spinal deformity surgery in NSD patients. The presence of trache- osis surgery. Paediatr Anaesth 2016;26:970–975. ostomy, history of pneumonia, and the presence of gastrostomy 3. Birkeland AC, Rosko AJ, Beesley L, Bellile E, Chinn SB, Shu- may lead to worsened surgical outcomes. man AG, et al. Preoperative Tracheostomy Is Associated with In our case, surgical correction was offered to stop the deformity Poor Disease-Free Survival in Recurrent Laryngeal Cancer. progression and to maintain the function of breathing. Before sur- Otolaryngol Head Neck Surg 2017;157:432–438. gery, bronchoscopy was done to evaluate lung condition. After the 4. Brooks JT, Sponseller PD. What’s New in the Management of operation, the patient was transferred to intensive care unit even Neuromuscular Scoliosis. J Pediatr Orthop 2016;36:627–633. though there were no respiratory events during the surgery. Fre- 5. Hayashi T, Fujiwara Y, Sakai H, Maeda T, Ueta T, Shiba K. Risk quent chest percussion and suctioning, cough assist, and nebulizer factors for severe dysphagia in acute cervical spinal cord injury. treatments were applied. However, the patient developed a fever Spinal Cord 2017;55:940–943. (38.6°C) on postoperative two days. Tazocin (piperacillin/tazo- 6. Li L, Yuan W, Zhang S, Wang K, Ruan H. Analysis of Risk Fac- bactam) was administered as a prophylactic antibiotic. Chest X-ray tors for Pneumonia in 482 Patients Undergoing Oral Cancer www.e-jnic.org 101 Patient with neuromuscular kyphosis Kwang-Ui Hong et al.

Surgery With Tracheotomy. J Oral Maxillofac Surg 2016;74: 9. Rumalla K, Yarbrough CK, Pugely AJ, Koester L, Dorward IG. 415–419. Spinal fusion for pediatric neuromuscular scoliosis: national 7. Luhmann SJ, Furdock R. Preoperative Variables Associated trends, complications, and in-hospital outcomes. J Neurosurg With Respiratory Complications After Pediatric Neuromuscu- Spine 2016;25:500–508. lar Spine Deformity Surgery. Spine Deform 2019;7:107–111. 10. Sharma S, Wu C, Andersen T, Wang Y, Hansen ES, Bunger CE. 8. Pesenti S, Blondel B, Peltier E, Launay F, Fuentes S, Bollini G, et Prevalence of complications in neuromuscular scoliosis surgery: al. Experience in Perioperative Management of Patients Under- a literature meta-analysis from the past 15 years. Eur Spine J going Posterior Spine Fusion for Neuromuscular Scoliosis. 2013;22:1230–1249. Biomed Res Int 2016;2016:3053056.

102 www.e-jnic.org Instructions for Authors

General Information (http://e-jnic.org/authors/authors.php). 3. The list of the authors in the manuscript should include only 1. Journal of Neurointensive Care (JNIC) is the official journal of the those who were directly involved in the process of the work. Korean Neurointensive Care Society and published biannually Authors can refer to the guideline by Harvard University in ((the last day of April and October). This Journal publishes 1999 to find details on authorship(http://www.hms.harvard. important papers covering the whole field of neurosurgical edu/integrity/authorship.html). intensive care unit, including studies in neuroscience, neurology, 4. The editorial board will make a decision on the approval for and molecular biology. Studies on rare cases and technical notes publication of the submitted manuscripts, and can request any of special instruments or equipment that might be useful to the further corrections, revisions, and deletions to the article text if field of neurosurgical intensive care are also acceptable. Drawing necessary. upon the expertise of an interdisciplinary team of physicians from 5. The price for all work requiring review, publishing, and re- neurosurgery, neurology, anesthesiology, critical care, and nursing printing of the paper will be determined by the editorial board. backgrounds, (JNIC) covers all aspects neurosurgical intensivists need to be aware of in order to provide optimal patient care. Manuscript Preparation 2. It should be assured that authors must not simultaneously submit an identical or similar paper for publication elsewhere. All manuscripts should be written in English using in 11 points Multiple publication is acceptable only in the case of meeting the Arial font and double-spaced. criteria of Uniform Requirements for Manuscripts Submitted to Biomedical Journals (Ann Intern Med 108: 258-265, 1988). 1. Publication type Manuscripts must be prepared in accordance with Uniform JNIC publishes editorials, reviews, original articles including requirements for Manuscripts submitted to Biomedical Journal clinical and laboratory research, case reports, letters to the developed by International Committee of Medical Journal editor and etc.Please review the below article type specifications Editors (February 2006). including the required article lengths, illustrations, table 3.All manuscripts must be written in English. Authors should limits and reference counts. The word count excludes the title minimize the use of English abbreviations. Spell out all page, abstract, tables, acknowledgements, contributions and abbreviations at first occurrence, and then introduce them by references. Manuscripts should be as succinct as possible.Any placing the abbreviation in parenthesis after the term being article longer than these limits should be discussed with the abbreviated. Abbreviation should be avoided as possible as one editor. can. When it is used, full expression of the abbreviation following abbreviated word in parentheses should be given at first use. a) Editorials All units should be given in metric system (The International Editorials are invited perspectives dealing with very active fields System of Units: SI units). of research, hot interest, fresh insights, and debates.

Submission of Manuscript Word count: up to 1,000 words Tables and figures: at editorial discretion 1. Authors are requested to submit their papers electronically by References: up to 10, ideally 5 using online manuscript submission system(http://submit. e-jnic.org). This site will guide authors stepwise through the b) Review articles submission process. The authors and topics for review articles will be selected by 2. Upon submission of a manuscript, authors should upload the editorial board and review articles should also undergo the author checklist and copyright transfer agreement form review process. Manuscripts include titlepage, unstructured www.e-jnic.org i Instructions for Authors JNIC Journal of Neurointensive Care abstract and keyword, main text (introduction, manuscript f) Special article body, conclusion), conflict of interest, acknowledgements (if Special articles are devoted to providing updated reports by necessary), references, tables, figurelegends, and figures. specialists in various fields or significant issues (e.g., history of the field) for the members of the society. The authors and Abstract: 200 words topics of special drafts will be assigned and specially requested Word count: up to 3,000 words by the editorial board.The authors’ views in special drafts will be Tables and figures: up to 7 respected as much as possible. References: 40 g) Other Publication Types c) Original articles Other publication types may be accepted. The recommended Original articles should contain the results of clinical or basic format should be discussed with the Editorial Board. research and should be sufficiently well documented to be acceptable to critical readers. The manuscript for an original 2. Manuscript format article should be organized in the following sequence: title page, Authors should refer to “Uniform Requirements for structured abstract and keywords, main text (introduction, Manuscripts Submitted to Biomedical Journals” (http:// methods, results, discussion, conclusion), conflict of interest, www.icmje.org/about-icmje/faqs/icmje-recommendations/). acknowledgements (if necessary), references, tables, figure The article should be organized in the order of title, abstract legends, figures, and supplementary data. (Objective, Methods, Results, and Conclusion sections should be included in laboratory investigation or clinical article but are Abstract: Structured, 250 words not necessary in other types of studies), key words,introduction, Word count: up to 3,500 words materials and methods, results, discussion, conclusions, Tables and figures: up to 7 references, tables, and figures or illustrations. In case reports, References: 30 materials and methods and results can be replaced with cases. d) Case reports 1) Title page Case reports will be published only in exceptional The title page should be composed of external and internal title circumstances, when they illustrate a rare occurrence of clinical pages. importance. These manuscripts should be organized in the a) The external title page should contain the article title, and following sequence: title page, unstructuredabstract and full names of all authorswith their institutional affiliations keywords, introduction, case report(s), discussion, conclusion, in English. The type of manuscript (Original Article, Case acknowledgments, references, tables, figure legends, and figures. Report or Case series, Technical report, Letter to editor, etc.) should be also addressed. When the work includes multiple Abstract: 150 words authors with different affiliations, the institution where the Word count: 1,500 words research was mainly conducted should be spelled out first, Tables and figures: up to 5 then be followed by foot notes in superscript Arabic numerals References: 10 beside the authors’ names to describe their affiliation in a consecutive order of the numbers. Then, mark the running e) Letters to the editor head as not to exceed 50 characters in English. The external Authors can submit a sound critic or opinion for the specific title page should also contain the address, TEL. and FAX. article published in the journal, topic of general interest numbers, and e-mail address of the corresponding author at regarding neurosurgical intensive care, personal view on a the bottom of the page, as well as information on the previous specific scientific issue, departmental announcements or presentation of the manuscript in conferences and funding changes, or other information of the clinical fields. resources, if necessary. b) The internal title page should only contain the article title Word count: 1,000 words in English. The internal title page must not contain any Tables and figures: up to 2 information on the names and affiliations of the authors. References: 10 2) Abstract and Keywords All manuscripts must contain an abstract. A list of Key Words, ii www.e-jnic.org Instructions for Authors Journal of Neurointensive Care JNIC with a maximum of six items, should be included at the end of Method of statistical analyses and criteria of significance level the abstract. The selection of Key Words should be based on should be described. In Case Reports, case history or case Medical Subject Heading (MeSH) of Index Medicus and the description replace the Materials and Methods section as well as Web site (http://www.nlm.nih.gov/mesh/MBrowser.html). Results section. The abstract should include brief descriptions on the objective, Please inform us the approved number of IRB when you submit methods, results, and conclusion as well as a detailed description the manuscript. of the data. An abstract containing 250 words or less is required for original articles, 200 words for review articles and 150 words 5) Results for case. Abstracts for LaboratoryInvestigation and Clinical The authors should describe logically their results of observations Article should begin with the statement of the paper’s purpose and analyses performed using methodology given in the previous and end with conclusions. Abstracts for other types of papers section and provide actual data. For biometric measurements should begin with a brief and clear statement of the paper’s in which considerable amount of stochastic variation exists purposeand be followed by appropriate details that support the a statistical treatment should be used in principle. The result conclusions of the paper. section should include sorely the findings of the current study, and not refer to previous reports. While an effort should be made 3) Introduction to avoid overlapping descriptions by Tables and by main text, The introduction should address the purpose of the article important trends and points in the Table should be described in concisely and include background reports mainly relevant to the the text. Experimental results should be described using Arabic purpose of the paper (detailed review of the literature should be numbers and the SI unit system. addressed in the discussion section). 6) Discussion 4) Materials and Methods Discussions about the findings of the research and interpretations Materials and Methods section should include sufficient in relation to other studies are made. It is necessary to emphasize details of the design, objects, and methods of the article in the new and critical findings of the study, not to repeat the results order, as well as the data analysis strategies and control of of the study presented in the previous sections. The meaning bias in the study. Enough details need to be addressed in the and limitation of observed facts should be described, and methodology section of an experimental study so that it can the conclusion should be related to the objective of the study be further replicated by others. When reporting experiments only when it is supported by the results of the research. It is with human subjects, the authors should indicate whether they encouraged for the authors to use subheadings in the discussion received an approval from the Institutional Review Board for section so that the readers can follow the logical flow of the the study. When reporting experiments with animal subjects, authors’ thought. the authors should indicate whether the handling of the animals was supervised by the research board of the affiliated institution 7) Conclusion or a similar one. Photographs disclosing patients must be The conclusion section should include a concise statement of accompanied by a signed release form from the patient or family the major findings of the study in accordance with the study permitting publication. We endorse the principles embodied in purpose. the ‘Declaration of Helsinki’ and expect that all investigations involving human materials have been performed in accordance 8) References with these principles. For animal experiment, ‘the Guiding a) Only references cited in text must appear in the reference list Principles in the Care and Use of Animals’ approved by the and marked in the form of superscript at the end of the sentences American Physiological Society must be observed. Explanation they were used in text (example: reference11,15,18)). of the experimental methods should be concise and sufficient b) All references should bealphabetized by the first author’s last for repetition by other qualified investigators. Procedures that name. have been published previously should not be described in c) When a work has six or less authors, cite the names of all authors. detail. However, new or significant modifications of previously When a work has over six authors, cite the first six authors’ name published procedures need full descriptions. The sources of followed by “et al.”Abbreviations for journal titles should be special chemicals or preparations should be given along with congruent with the style of IndexMedicus. A journal title with their location (name of company, city and state, and country). one word does not need to be written out in abbreviation. The www.e-jnic.org iii Instructions for Authors JNIC Journal of Neurointensive Care styles of references are as follows: figure number as the file name (i.e. Fig1.jpg). When a figure is composed of more than 2 parts, authors should combine Journal the figure in the correct orientation. Separate files without 1. Dávalos A, Pereira VM, Chapot R, BonaféA, Andersson T, embedded labels should be submitted only if the Editorial board Gralla J, et al. Retrospective multicenter study of solitaire FR requests them after the peer review. Authors should submit for revascularization in the treatment of acute ischemic stroke. figures in black and white if they want them to be printed in Stroke 2012;43:2699–2705. black and white. Authors are responsible for any additional costs of producing color figures. Website The files should have following resolutions for printing: line art 1. World Health Organization, The International Spinal Cord at 300 dpi, combination half-tones at 300 dpi, and half-tones Society. International perspectives on spinal cord injury. Geneva, (gray scale or color without type or lettering) at 300 dpi. If the CH: World Health Organization, 2013(http://apps.who.int/ quality of the photographs is considered as inappropriate for iris/bitstream/10665/94192/1/WHO_NMH_VIP_13.03_ printing, re-submission of them can be requested by the journal. eng.pdf?ua=1) [Accessed October 1, 2018] Tables, graphs, figures, and photographs should be used only when necessary. Book 1. Conover WJ: Practical Nonparametric Statistics, ed 2. New Publication York: Jon Wiley & Sons; 1971. p. 216-218. Once a manuscript is accepted for publication by the journal, Article in a Book it will be sent to the press, and page proofs will be sent to 1. Ojemann RG: Surgical management of bacterial intracranial authors. Authors must respond to the page proofs as soon as aneurysms in Schmideck HH, Sweet HH (eds): Operative possible after making necessary corrections ofmisspellings, and Neurosurgical Techniques. Indications, Methods and Results, the location of the photographs, figures or tables. Authors can ed 2. Orlando: Grune& Stratton; 1988. p. 997-1001 make corrections for only typing errorsand are not allowed to make any author alteration or substantive changes of the text. 9) Tables, figures, and illustrations Proofs must be returned to the press within 72 hours of receipt. Tables and figure legends should be included below the No response from the authors within this time frame will lead references pages at the end of the paper, but figures should be the publication of the proof read without corrections, and the submitted separately fromthe text of paper. Table should be editorial board is not responsible for any mistakes or errors simple and should not duplicate information in figures. Title all occurring in this process. tables and number them with Arabic numerals in the order of their citation. Type each table on a separate sheet. Describe all Post-Publication Discussion and abbreviations. Each column should have an appropriate heading, Corrections and if numerical measurements are given, the unit should be added to column heading. The significance of results should The post-publication discussion is available through letter to the be indicated by appropriate statistical analysis. Table footnotes editor. If any readers have a concern on any articles published, should be indicated with superscript markings. When remarks they can submit letter to the editor on the articles. If there are used to explain items of the table, the markers should be founds any errors or mistakes in the article, it can be corrected given in the order of *, †, ‡, §, ||. through errata, corrigenda, or retraction. Each figure should be submitted as a separate file, with the

iv www.e-jnic.org Author Checklist

Title of article:

Author(s):

Author(s) should check the following items under the heading of ‘Authors’. The spaces under the heading of ‘Editorial Staff’ are reserved for editorial office. Please leave no marking at the spaces under the heading of ‘Editorial Staff’.

1. Mandatory components of a manuscript 1) Manuscript layouts □ Yes / □ No 2) Manuscript should be prepared in following sequences; a) Original Article: title page, inner title page, abstract and keywords, introduction, materials and methods,results, discussion,conclusion, conflicts of interest*, acknowledgment, references, tables, figure legends, figures and supplementary data. □ Yes / □ No b) Case Report: title page, inner title page, abstract andkeywords,introduction, case report, discussion, conclusion, references, tables, figure legends and figures. □ Yes / □ No c) Review Article: title page, inner title page, abstract and key words, introduction, manuscript body, conclusion, conflict of interest*, acknowledgment, references, tables, figure legends and figures. *If applicable □ Yes / □ No

2. Word count 1) The word count of abstract and manuscript are within limits for the publication type. □ Yes / □ No 2) Please indicate the total word number at the title page. □ Yes / □ No 3) All pages are numbered in sequence, starting with the title page. □ Yes / □ No

3. Title page The title page must contain all of the followings; clear title, name and affiliation of all authors, information of the corresponding author (address, telephone number, fax number, and e-mail address), type of article and running head. □ Yes / □ No

4. Inner title page Only title of the manuscript is listed. Any information on the authors and corresponding author is not shown on the inner title page □ Yes / □ No

5. Abstract 1)Abstract should have four parts, which are objective, methods, results, and conclusion. Please indicate the word number after Key words. □ Yes / □ No 2)Recommended words from Medical Subject Heading Terms database at http://www.ncbi.nlm.nih.gov/mesh/MBrowser.html are used for Key word and the words are written within a maximum of six. □ Yes / □ No

6. Text 1) Text is written in 11 points Arial fonts with double line spacing and 3 cm as bottom and left and right margins on A4 paper □ Yes / □ No 2) Every word except name of a place, a personal name, and a proper noun is written in lower case. Also, some words such as names of virus, Latin and p values are written in italics. Comma as thousand separators must be placed for indicating numbers. □ Yes / □ No www.e-jnic.org v Author Checklist JNIC Journal of Neurointensive Care

7. Figures and tables Cite in numerical order as they are first mentioned in the text □ Yes / □ No

8. References 1) References are described followed by the rules of the Journal of JNIC □ Yes / □ No 2) It is recommended that references may contain at least one of the articles of JNIC. If it may not, please specify the reason in the cover letter. □ Yes / □ No

9. Tables 1) Table style should follow the conventional rules as suggested in Instructions to Authors. □ Yes / □ No 2) Tables should be self-explanatory, and the data must not be duplicated in the text or figures □ Yes / □ No

10. Figures and Figure legends Each figure should be submitted as a separate file, with the figure number as the file name (i.e. Fig1. jpg). When a figure is composed of more than 2 parts, authors should combine the figure in the correct orientation. Separate files without embedded labels should be submitted only if the Editorial board requests them after the peer review. Authors should submit figures in black and white if they want them to be printed in black and white. Authors are responsible for any additional costs of producing color figures. □ Yes / □ No

11. Figure resolution Author must guarantee the quality of figures. It should be noted that the manuscript could be rejected if print-suitable high-quality figures are not provided at the initial phase of submission. □ Yes / □ No

12. Figure legend Figure legend should be self-explanatory. Abbreviations should not be used, and the present tense must be used for the description. Appropriate description of dyeing method and magnification for histological figure should be provided □ Yes / □ No

13. Submission All authors should sign on the transfer of copyright and agreement and corresponding author should indicate that he (she) takes full responsibility of authorship from all authors. □ Yes / □ No

Date:

Editorial office use only

□ suitable for review process □ a qualified consent for submission acceptance □ needs author’s correction

Date:

vi www.e-jnic.org Copyright Transfer Agreement

The author(s) submit my/our manuscript with the following title

in consideration of the Editorial Board of the Journal of Neurointensive Care reviewing, editing and publishing. This manuscript contains page(s), Figure(s), Picture(s), Table(s).

I/we hereby transfer, assign and otherwise convey to the Korean Neurointensive Care Society upon acceptance of the manuscript for publication by the Journal of Neurointensive Care all copyright. I/we have all rights except copyright. I/we can use part or all of the contents of the manuscript under written agreement of the Korean Neurointensive Care Society. In case that I/we use materials from the manuscript I/we will clarify the reference.

I/we certify that the contents of the manuscript, in all or in part, has not been published and is not being considered for publication elsewhere, unless otherwise specified herein. I certify that I have made a substantial contribution to the medical/scientific/intellectual content of the manuscript and on that basis agree to be named as an author. I approve the manuscript for publication and will take public responsibility for its content.

Each of the undersigned is an author of the manuscript and all authors are named on this document.

Author’s name Author’s signature Date

vii