PJG v30n02

30 | 1 | 23 Highlights in this issue: • Meta-Analysis: laparoscopic endoscopic cooperative surgery in upper GI lesions • Treatment of anastomotic leaks following esophagectomy • Prediction of liver decompensation prior to transarterial chemoembolization UESA DE ENDO G CIEDADE PORTU SCOPIA DIGEST IVA SO • SPED • A Partner Publication by

GE – Portuguese Journal of Gastroenterology Director Guilherme Macedo, MD, PhD – São João Hospital Centre, Porto, Portugal Adjunct Directors Jorge Canena, MD, PhD – CUF Infante Santo Hospital, Lisbon, Portugal José Presa, MD – Trás-os-Montes e Alto Douro Hospital Centre, Vila Real, Portugal Editor-in-Chief Diogo Libânio, MD, PhD – Portuguese Oncology Institute of Porto, Porto, Portugal Co-Editors Miguel Areia, MD, PhD – Portuguese Oncology Institute of Coimbra, Coimbra, Portugal Luís Maia, MD – Porto Hospital Centre, Porto, Portugal Carolina Palmela, MD – Beatriz Ângelo Hospital, Loures, Portugal Eduardo Rodrigues Pinto, MD, PhD – São João Hospital Centre, Porto, Portugal Editorial Board Andreia Albuquerque, MD, PhD – St. James’s University Hospital, Leeds, UK Nuno Almeida, MD, PhD – Coimbra Hospital and University Centre, Coimbra, Portugal Pedro Amaro, MD – Coimbra Hospital and University Centre, Coimbra, Portugal Jorge Amil Dias, MD – São João Hospital Centre, Porto, Portugal Marianna Arvanitaki, MD, PhD – Erasmus Hospital, Brussels, Belgium Pedro Barreiro, MD – Western Lisbon Hospital Centre, Lisbon, Portugal Miguel Bispo, MD – Champalimaud Foundation, Lisbon, Portugal Raf Bisschops, MD, PhD – University Hospitals Leuven, KU Leuven, Leuven, Belgium James Buxbaum, MD – University of Southern California, Los Angeles, USA Ana Caldeira, MD – Amato Lusitano Hospital, Castelo Branco, Portugal Jorge Canena, MD, PhD – CUF Infante Santo Hospital, Lisbon, Portugal Marco Carbone, MD, PhD – University of Milano-Bicocca, Milan, Italy Helder Cardoso, MD – São João Hospital Centre, Porto, Portugal F. Castro Poças, MD, PhD – Porto Hospital Centre, Porto, Portugal Helena Cortez-Pinto, MD, PhD – Hospital Santa Maria, Lisbon, Portugal José Cotter, MD, PhD – Nossa Senhora da Oliveira Hospital, Guimarães, Portugal (Continued on next page) Printed in Portugal by Printipo: [email protected]

(Continued) Marília Cravo, MD, PhD – Luz Hospital, Lisbon, Portugal Isabelle Cremers, MD – Setúbal Hospital Centre, Setúbal, Portugal Jacques Devière, MD, PhD – Université Libre de Bruxelles, Hôpital Erasme, Brussels, Belgium Mário Dinis Ribeiro, MD, PhD – Portuguese Oncology Institute of Porto, Porto, Portugal Daniela Dobru, MD, PhD – University of Medicine and Pharmacy, Târgu Mureș, Romania Sandra Faias, MD, PhD – Portuguese Oncology Institute of Lisbon, Lisbon, Portugal Pedro Figueiredo, MD, PhD – Coimbra Hospital and University Centre, Coimbra, Portugal Paulo Freire, MD, PhD – Coimbra Hospital and University Centre, Coimbra, Portugal Lorenzo Fuccio, MD, PhD – S. Orsola-Malpighi University Hospital, Bologna, Italy Alessandro Fugazza, MD – Humanitas Clinical and Research Centre – IRCCS, Rozzano, Italy Federica Furfaro, MD – Humanitas Clinical and Research Centre – IRCCS, Rozzano, Italy Cesare Hassan, MD, PhD – Nuovo Regina Margherita Hospital, Rome, Italy Konstantinos Katsanos, MD, PhD – University of Ioannina School of Health Sciences, Ioannina, Greece Arjun Koch, MD, PhD – Erasmus MC University Medical Centre, Rotterdam, Netherlands Roman Kuvaev, MD, PhD – Yaroslavl Regional Cancer Hospital, Yaroslavl, Russia Luis Lopes, MD, PhD – Alto Minho Local Health Unit, Viana do Castelo, Portugal Susana Lopes, MD, PhD – São João Hospital Centre, Porto, Portugal Mariana Machado, MD, PhD – Vila Franca de Xira Hospital, Vila Franca de Xira, Portugal Tadateru Maehata, MD, PhD – St. Marianna University School of Medicine, Kawasaki, Japan Vítor Magno, MD – Dr. Nélio Mendonça Hospital, Funchal, Portugal Fernando Magro, MD, PhD – São João Hospital Centre, Porto, Portugal Tato Marinho, MD, PhD – Northern Lisbon Hospital Centre, Lisbon, Portugal Dileep Mangira, MD, PhD – Western Health, Melbourne, VIC, Australia Ricardo Marcos Pinto, MD, PhD – Porto Hospital Centre, Porto, Portugal Diogo Moura, MD, PhD – Hospital das Clínicas, Porto Alegre, Brazil Pedro Moutinho Ribeiro, MD – São João Hospital Centre, Porto, Portugal Kerri Novak, MD – Calgary Division of Gastroenterology and Hepatology, Calgary, AB, Canada Nuno Nunes, MD – Dívino Espírito Santo Hospital, Ponta Delgada, Portugal Oliver Pech, MD, PhD – Krankenhaus Barmherzige Brüder, Regensburg, Germany Isabel Pedroto, MD, PhD – Porto Hospital Centre, Porto, Portugal Enrique Perez-Cuadrado, MD, PhD – European Hospital Georges Pompidou, Paris, France Pedro Pimentel-Nunes, MD, PhD – Portuguese Oncology Institute of Porto, Porto, Portugal Rolando Pinho, MD – Vila Nova de Gaia/Espinho Hospital Centre, Vila Nova de Gaia, Portugal (Continued on next page) © 2023 S. Karger AG, Basel www.karger.com [email protected]

(Continued) Francisco Portela, MD – Coimbra Hospital and University Centre, Coimbra, Portugal José Pedro Rodrigues, MD – Central Lisbon Hospital and University Centre, Lisbon, Portugal Susana Rodrigues, MD, PhD – Bern University Hospital, Bern, Switzerland Carla Rolanda, MD, PhD – Braga Hospital, Braga, Portugal Bruno Rosa, MD – Nossa Senhora da Oliveira Hospital, Guimarães, Portugal Daniel Sifrim, MD, PhD – Queen Mary University of London, London, UK Elisa Soares, MD – Coimbra Hospital and University Centre, Coimbra, Portugal João Bruno Soares, MD – Braga Hospital, Braga, Portugal Luís Tomé, MD, PhD – Coimbra Hospital and University Centre, Coimbra, Portugal Joana Torres, MD, PhD – Beatriz Ângelo Hospital, Loures, Portugal Monica Velosa, MD – Queen Mary University of London, London, UK José Velosa, MD, PhD – Lusíadas Hospital, Lisbon, Portugal © 2023 S. Karger AG, Basel www.karger.com [email protected]

Journal Information Guidelines for Authors Aims and Scope The GE Portuguese Journal of Gastroenterology (formerly Jornal Português de Gastrenterologia), We strongly encourage authors to read founded in 1994, is the official publication of Sociedade Portuguesa de Gastrenterologia the Guidelines for Authors at (Portuguese Society of Gastroenterology), Sociedade Portuguesa de Endoscopia Digestiva www.karger.com/pjg _guidelines prior to (Portuguese Society of Digestive Endoscopy) and Associação Portuguesa para o Estudo do submitting an article Fígado (Portuguese Association for the Study of the Liver). The journal publishes clinical and basic research articles on Gastroenterology, Digestive En- doscopy, Hepatology and related topics. Review articles, clinical case studies, images, letters to the editor and other articles such as recommendations or papers on gastroenterology clinical practice are also considered. Only articles written in English are accepted. Journal Contact For questions or comments, please contact the persons responsible who can be found at http://www.karger.com/Journal/Contact/272027 Price per printed issue: Free of charge ISSN Print Edition: 2341–4545 Disclaimer: The statements, opinions and data con- Print run: 100 ISSN Online Edition: 2387–1954 tained in this publication are solely those of the indi- ERC-No.: 117866 vidual authors and contributors and not of the pub- Editor address: Rua Abranches Ferrão, nº 10–14º, Journal Homepage: www.karger.com/pjg lisher and the editor(s). The appearance of advertise- PT–1600-001 Lisbon (Portugal) Bibliographic Indices: This journal is regularly listed ments in the journal is not a warranty, endorsement, Printed in: PT–2735-197 Cacém (Portugal) in bibliographic services, including PMC, PubMed, or approval of the products or services advertised or Web of Science, SciELO Citation Index, Google Scholar, of their effectiveness, quality or safety. The publisher DOAJ, Scopus, and WorldCat. and the editor(s) disclaim responsibility for any injury to persons or property resulting from any ideas, Publication Data: GE Port J Gastroenterol is pub- methods, instructions or products referred to in the lished 6 times a year. Volume 30 with 6 issues ap- content or advertisements. pears in 2023. Copyright: © 2023 Portuguese Society of Gastroen- terology (VAT number PT501759050). Published by S. Karger AG, Basel (Switzerland). All rights reserved. No part of this publication may be translated into other languages, reproduced or uti- lized in any form or by any means, electronic or me- chanical, including photocopying, recording, micro- copying, or by any information storage and retrieval system, without permission in writing from the pub- lisher. © 2023 S. Karger AG, Basel www.karger.com [email protected]

Vol. 30, No. 1, 2023 Contents Editorial 1 Metastatic Malignant Melanoma of the Gastrointestinal Tract: Too Dark to be Seen? Chaves, J.; Libânio, D. (Porto) Review Article 4 Efficacy and Safety of Laparoscopic Endoscopic Cooperative Surgery in Upper Gastrointestinal Lesions: A Systematic Review and Meta-Analysis de Brito, S.O.; Libânio, D.; Pinto, C.M.M.; Araújo Teixeira, J.P.P.O.; Araújo Teixeira, J.P.M. (Porto) Research Articles 20 Positioning Aeromonas Infection in Inflammatory Bowel Disease: A Retrospective Analysis Pereira Guedes, T.; Alves Silva, J.; Neves, S.; Falcão, D.; Costa, P.; Lago, P.; Pedroto, I.; Salgado, M. (Porto) 29 Development of a Model to Predict Liver Decompensation prior to Transarterial Chemoembolization Refractoriness in Patients with Intermediate-Stage Hepatocellular Carcinoma Ferreira-Silva, J.; Costa-Moreira, P.; Cardoso, H.; Liberal, R.; Pereira, P.; Macedo, G. (Porto) 38 Anastomotic Leaks following Esophagectomy for Esophageal and Gastroesophageal Junction Cancer: The Key Is the Multidisciplinary Management Ortigão, R.; Pereira, B.; Silva, R.; Pimentel-Nunes, P.; Bastos, P.; Abreu de Sousa, J.; Faria, F.; Dinis-Ribeiro, M.; Libânio, D. (Porto) 49 Comparison between Two Types of 22-Gauge Fine-Needle Biopsy for Solid Pancreatic Tumors Jaurrieta-Rico, C.; Picazo-Ferrera, K.; Aguilar-Solis, R.; Escobedo-Paredes, D.; Bandala-Jaques, A.; Chavez-Gomez, V.; Hernandez-Guerrero, A.; Alonso-Larraga, J.O. (Mexico City) Clinical Case Studies 57 Isolated Intracardiac Metastasis: The First Sign of Hepatocellular Carcinoma Paixão, A.; Silva, R.; Lopes, N.; Carvalho, S.; Carrola, P.; Presa Ramos, J. (Vila Real) 61 Percutaneous Endoscopic Gastrostomy Placement under NIV in Amyotrophic Lateral Sclerosis with Severe Ventilatory Dysfunction: A Safe and Effective Procedure Gaspar, R.; Ramalho, R.; Coelho, R.; Andrade, P.; Goncalves, M.R.; Macedo, G. (Porto) 68 IgG4-Related Esophageal Disease Presenting as Esophagitis with Chronic Strictures Correia, C.; Moreira, H.; Almeida, N.; Soares, M.; Cipriano, A.; Figueiredo, P. (Coimbra) Cover illustration Endoscopic image of a gastric melanoma metastasis, with a dark coloration area of 5mm in one of the edges. From Pinto et al., pp. 79-82. © 2023 S. Karger AG, Basel

Endoscopic Snapshots 73 Gastrointestinal Metastatic Melanoma: The Key for Diagnosis Soares Santos, D.F.; Costa, M.; Carvalho, P.; Santos, R.M.; Carvalho, A. (Coimbra) 76 Band Ligation-Assisted Forceps Scissor Transection of a Unique Pedunculated Colorectal Lesion with Stalk Varices Zimmer, V. (Neunkirchen/Homburg); Heinrich, C. (Saarbrücken) Images in Gastroenterology and Hepatology 79 Gastric Metastatic Melanoma Mimicking a Hyperplastic Lesion Pinto, C.M.M.; Rodriguez, M.; Souto Moura, M.; Afonso, M.; Bastos, P.; Dinis Ribeiro, M. (Porto) © 2023 S. Karger AG, Basel

Editorial Received: August 14, 2022 Accepted: August 18, 2022 GE Port J Gastroenterol 2023;30:1–3 Published online: October 28, 2022 DOI: 10.1159/000527210 Metastatic Malignant Melanoma of the Gastrointestinal Tract: Too Dark to be Seen? Jéssica Chavesa Diogo Libânioa, b aGastroenterology Department, Portuguese Institute of Oncology, Porto, Portugal; bMEDCIDS- Department of Community Medicine, Health Information and Decision, Faculty of Medicine, University of Porto, Porto, Portugal Keywords who presents with a set of non-specific symptoms, in- Melanoma · Metastatic · Gastrointestinal tract cluding GI symptoms. The initial imaging study was neg- ative for malignant disease and endoscopy with biopsies Melanoma maligno metastático do trato of dark-coloured polypoid lesions allowed the diagnosis gastrointestinal: demasiado escuro para ser visto? of gastric metastasis from MM, which is a rare finding in metastatic MM. The prognosis, due to the patient’s co- Palavras Chave morbidities which rendered her unfit for chemotherapy, Melanoma · Metastases · Gastrointestinal was poor. This case highlights the role of endoscopy (the key to solve the mystery) in the diagnosis and manage- Malignant melanoma (MM) is the most common ment of this patient. cause of mortality due to skin cancer worldwide and its incidence is increasing [1]. The majority of MM are from The second case by Pinto et al. [5] highpoints the fun- cutaneous origin, and most gastrointestinal (GI) tract damental role of histology in conjugation with endoscop- melanomas are a result of metastasis, although MM can, ic findings. A more distracted eye could have easily missed less frequently, arise primarily from GI origin [2]. the darker area found in endoscopy or misinterpreted it as a non-significant lesion, and tissue acquisition in adja- Metastasis of MM in the GI tract is common (estimat- cent areas possibly lead to the initial misdiagnosis. Repeat ed in up to 60% of all patients with advanced disease), but endoscopy and biopsies proved to be the right choice of in practice only a small proportion are clinically signifi- action, and this should be considered when clinical his- cant. Indeed, only about 1–5% are clinically diagnosed tory, endoscopic and histological findings do not match. antemortem [2, 3]. In this issue of GE – Portuguese Jour- This case is also a reminder to never forget the previous nal of Gastroenterology, two case reports are published medical history of a patient, namely, of previous malig- reporting interesting GI involvement of MM, highlight- nant disease, as it might just be the clue needed for final ing the sometimes difficult task of diagnosing them. diagnosis. Firstly, Soares-Santos et al. [4] described a case of an MM is among the most common carcinomas to me- elderly woman with no previous history of melanoma tastasize to GI tract and can be spread throughout. Even so, it appears to have particular affinity to the small bow- [email protected] © 2022 The Author(s). Correspondence to: www.karger.com/pjg Published by S. Karger AG, Basel Jéssica Chaves, [email protected] is is an Open Access article licensed under the Creative Commons Attribution-NonCommercial-4.0 International License (CC BY-NC) (http://www.karger.com/Services/OpenAccessLicense), applicable to the online version of the article only. Usage and distribution for com- mercial purposes requires written permission.

el, specially to the jejunum and ileum [3, 6]. At a molecu- clinical practice than post-mortem, suggesting that most lar level, the greatest expression of CCL25 in the small of the times metastasis is asymptomatic [3, 11]. Thus, if bowel, which is a ligand to CCR9 expressed in the mela- metastatic disease is already present, endoscopic and his- noma cell surface, may somehow explain the typical tological diagnosis of MM metastasis of the GI tract (atypical) metastasis to this part of the GI tract [3, 7]. should only be pursued if it modifies management of the patient. On the other hand, primary GI melanomas can arise from various GI segments, more commonly from the anal In cases of melanoma of unknown primary (that cor- canal, rectum, and oesophagus and accounts for a minor- responds to about 3% of all cases of MM), i.e., cases in ity of MM, with an estimated incidence of 0.58 cases per which, according to Das Gupta criteria [14], cutaneous, million people. They are more frequently encountered in ophthalmologic, anal, and genital melanoma have been elderly women and tend to be more aggressive and diag- excluded, the true value of endoscopic evaluation is dif- nosed at an advance stage – 36% versus 4% comparing to ficult to establish and more recent consensus argues that cutaneous melanoma, respectively [8, 9]. A primary GI it may not be useful to search for the primary tumour in melanoma might be suspected in the absence of prior his- mucosal membranes, eyes, or other organs [15, 16]. In tory of cutaneous melanoma or if the lesion is isolated case presented by Soares-Santos et al. [4], the symptoms without other extraintestinal metastasis, and it can be in- presented by the patient motivated the endoscopic study ferred histologically if a precursor lesion is present in tis- and lead to the diagnosis of metastatic MM. sue sample [10]. Prognosis of MM has dramatically been transformed Patients with metastatic MM of the GI tract may expe- since the introduction of new therapeutical targets. Be- rience generalized non-specific GI symptoms such as ab- fore the introduction of target agents, such as BRAF in- dominal pain or constipation, depending primarily on hibitors and immunotherapy, MM in advanced stage had the place affected. Cases of GI occlusion and active bleed- a median survival time of 6.2 months, with only 25.5% of ing have also been described [2, 3]. Clinical diagnosis of the patients alive at 1 year [17]. In the era of immune GI primary melanoma or secondary involvement can be checkpoint inhibitors and targeted BRAF/MEK inhibi- challenging, especially if symptoms are mild and non- tors, the clinical management of metastatic MM has for- specific. The time between primary excision and meta- tunately changed. Most immune checkpoint inhibitors static disease can also be a confounding factor, since most are now being used in the treatment of metastatic MM metastases are diagnosed within the first 3 years, but there with or without surgery, improving overall survival. are some cases reporting metastatic disease 15 years after Nivolumab, for instance, had a 1-year survival rate of 73% initial treatment [11]. in patients with non-operable or metastatic MM, with a good safety profile [18]. Immune-related adverse events Imaging studies such as computed tomography or that can urge with this therapy, and may occur in almost positron emission tomography (PET) may be useful in every organ, are usually mild and treatable [3, 18]. identifying sites of possible metastatic melanoma and can be ordered during follow-up, particularly in advanced Surgery also plays a role in the management of these disease. Nevertheless, mainly for computed tomography patients, and so a multidisciplinary approach is recom- scan, the sensitivity for detecting metastases is about 60– mended. An increase in quality of life and survival is, like- 70% [3]. wise, seen in patients undergoing resection of GI metas- tases of MM. Despite this fact, the decision to recommend Endoscopic evaluation, as seen in the 2 case reports a surgical procedure must take into account patients’ co- explored in this issue, is an irreplaceable tool to obtain a morbidities, age, and melanoma disease burden [19]. diagnosis and can, with the exception of videocapsule en- doscopy, acquire tissue for histological appraisal, which In conclusion, even if metastatic MM of GI tract is not is vital in confirming the diagnosis [2, 3, 12]. Endoscopic an uncommon condition, its clinical diagnosis is far from appearance is variable and metastatic lesions might be optimal. The GI tract may be just too dark to be seen (po- misleading. Polypoid or excavated lesions may be ob- tentially due to non-specific symptoms and the need for served, and even though colour could be helpful, they invasive procedures), or the lesions may be just too may present themselves as amelanotic, so biopsy of sus- “white” to be deceptive. A high clinical suspicion must be pected lesions should be performed [3, 12, 13], as seen in present in patients presenting with GI symptoms and his- the case reported by Pinto et al. [5]. However, as stated tory of MM. Treatment options are increasing, so is the previously, it is important to note that metastatic mela- survival of these patients. noma to the GI tract is much less frequently diagnosed in 2 GE Port J Gastroenterol 2023;30:1–3 Chaves/Libânio DOI: 10.1159/000527210

Conflict of Interest Statement Author Contributions The authors have no conflicts of interest to declare. Jéssica Chaves performed the literature search and wrote the Funding Sources manuscript. Diogo Libânio reviewed the manuscript and made No funding was used for the development of this work. critical corrections. Data Availability Statement All data generated or analysed during this study are included in this article. Further enquiries can be directed to the correspond- ing author. References 1 Dzwierzynski WW. Melanoma risk factors 7 Hadjinicolaou AV, Hadjittofi C, Athanaso- 15 Garbe C, Amaral T, Peris K, Hauschild A, poulos PG, Shah R, Ala AA. Primary small Arenberger P, Basset-Seguin N, et al. Euro- and prevention. Clin Plast Surg. 2021;48(4): bowel melanomas: fact or myth? Ann Transl pean consensus-based interdisciplinary Med. 2016;4(6):113. guideline for melanoma. Part 1: diagnostics 543–50. – update 2022. Eur J Cancer. 2022;170:236– 8 Zheng Y, Cong C, Su C, Sun Y, Xing L. Epide- 55. 2 Sinagra E, Sciumè C. Ileal melanoma, a rare miology and survival outcomes of primary gastrointestinal melanoma: a SEER-based 16 Sirvan SS, İhsan Eren H, Kurt Yazar S, Günenç cause of small bowel obstruction: report of a population study. Int J Clin Oncol. 2020; AC, Yeşilada AK, Irmak F, et al. Approach to 25(11):1951–9. patients with malignant melanoma of un- case, and short literature review. Curr Radio- known primary origin. Sisli Etfal Hastan Ttip 9 Kahl AR, Gao X, Chioreso C, Goffredo P, Bul. 2019;53(2):125–31. pharm. 2020;13(1):56–62. Hassan I, Charlton ME, et al. Presentation, management, and prognosis of primary gas- 17 Korn EL, Liu P-Y, Lee SJ, Chapman J-AW, 3 Kohoutova D, Worku D, Aziz H, Teare J, trointestinal melanoma: a population-based Niedzwiecki D, Suman VJ, et al. Meta-analy- study. J Surg Res. 2021;260:46–55. sis of phase II cooperative group trials in met- Weir J, Larkin J. Malignant melanoma of the astatic stage IV melanoma to determine pro- 10 Syed HR, Shekar S, Aravantagi A. Melanoma gression-free and overall survival bench- gastrointestinal tract: symptoms, diagnosis, and the gastrointestinal (GI) tract: maintain- marks for future phase II trials. J Clin Oncol. ing a high index of suspicion. Cureus. 2021; 2008;26(4):527–34. and current treatment options. Cells. 2021; 13(2):e13408. 18 Long GV, Atkinson V, Ascierto PA, Robert C, 10(2):327. 11 Farshad S, Keeney S, Halalau A, Ghaith G. A Hassel JC, Rutkowski P, et al. Effect of case of gastric metastatic melanoma 15 years nivolumab on health-related quality of life in 4 Soares Santos DF, Costa M, Carvalho P, San- after the initial diagnosis of cutaneous mela- patients with treatment-naïve advanced mel- noma. Case Rep Gastrointest Med. 2018; anoma: results from the phase III CheckMate tos RM, Carvalho A. Gastrointestinal meta- 2018:7684964. 066 study. Ann Oncol. 2016;27(10):1940–6. static melanoma: the key for diagnosis. GE 12 Urbain D, Aerts M, Reynaert H, Mana F, 19 Sanki A, Scolyer RA, Thompson JF. Surgery Neyns B. Small-bowel metastasis of malig- for melanoma metastases of the gastrointesti- Port J Gastroenterol. 2021. https://www.karg- nant melanoma: video capsule endoscopy ap- nal tract: indications and results. Eur J Surg pearance. Endoscopy. 2010;42(Suppl 2):E185. Oncol. 2009;35(3):313–9. er.com/DOI/10.1159/000518000. 13 Ozturk O, Basar O, Koklu S, Yuksel O, Purnak 5 Pinto CMM, Rodriguez M, Souto Moura M, T, Sokmensuer C. An unusual presentation of malignant melanoma: amelanotic gastric me- Afonso M, Bastos P, Dinis Ribeiro M. Gastric tastasis. Am J Gastroenterol. 2015;110(3):476. metastatic melanoma mimicking a hyperplas- 14 DasGupta T, Dowden L, Berg TW. Malignant melanoma of unknown primary origin. Surg tic lesion. GE Port J Gastroenterol. 2021. Gynec Obst. 1963;117:341–5. https: //www.karger.com/ DOI/10.1159/000520211. 6 Patel JK, Didolkar MS, Pickren JW, Moore RH. Metastatic pattern of malignant melano- ma. A study of 216 autopsy cases. Am J Surg. 1978;135(6):807–10. Metastatic Melanoma of the GE Port J Gastroenterol 2023;30:1–3 3 Gastrointestinal Tract DOI: 10.1159/000527210

Review Article Received: March 11, 2022 Accepted: May 23, 2022 GE Port J Gastroenterol 2023;30:4–19 Published online: November 7, 2022 DOI: 10.1159/000526644 Efficacy and Safety of Laparoscopic Endoscopic Cooperative Surgery in Upper Gastrointestinal Lesions: A Systematic Review and Meta-Analysis Sara Oliveira de Britoa Diogo Libâniob, c Cláudia Martins Marques Pintob João Pedro Pinho Osório de Araújo Teixeirad, e João Paulo Meireles de Araújo Teixeirad, e aFaculty of Medicine, University of Porto, Porto, Portugal; bGastroenterology Department, Portuguese Oncology Institute of Porto, Porto, Portugal; cCINTESIS/Biostatistics and Medical Informatics Department, Porto Faculty of Medicine, Porto, Portugal; dGeneral Surgery, São João University Hospital Center, Porto, Portugal; eSurgery Department, Faculty of Medicine, University of Porto, Porto, Portugal Keywords studies/1,336 patients (all retrospective cohorts). No signifi- Laparoscopic endoscopic cooperative surgery · Endoscopic cant differences were found between LECS and preexisting submucosal dissection · Laparoscopic resection · techniques (endoscopic submucosal dissection (ESD)/lapa- Subepithelial lesions · Meta-analysis roscopy) regarding any outcomes. However, there was a trend to shorter hospitalization time, longer procedure dura- Abstract tion, and fewer adverse events in LECS versus Laparoscopy Background and Aims: Laparoscopic and endoscopic coop- and ESD. R0 tended to be higher in the LECS group. Hospi- erative surgery (LECS) combines advantages of endoscopy talization time was significantly shorter in gastric versus EGJ and laparoscopy in order to resect upper gastrointestinal le- lesions (mean 7.3 vs. 13.7 days, 95% CI: 6.6–7.9 vs. 8.9–19.3). sions. Our aim was to evaluate the efficacy and safety of LECS There were no significant differences in conversion rate, ad- in patients with EGJ (esophagogastric junction), gastric and verse events, or mean procedural time according to location. duodenal lesions, as well as to compare LECS with pure en- There was a trend to higher conversion rate and longer pro- doscopic and pure laparoscopic procedures. Methods: cedure durations in EGJ and higher rate of adverse events in PubMed, Scopus, and ISI Web of Knowledge were searched. duodenal lesions. Conclusion: LECS is a valid, safe, and effec- Efficacy (R0, recurrence) and safety (conversion rate, proce- tive treatment option in patients with EGJ, gastric, and duo- dure and hospitalization time, adverse events, mortality) denal lesions, although existing studies are retrospective outcomes were extracted and pooled (odds ratio or mean and prone to selection bias. Prospective studies are needed difference) using a random-effects model. Study quality was to assess if LECS is superior to established techniques. Key assessed with Newcastle-Ottawa Scale and heterogeneity by Messages: LECS is safe and effective in the treatment of up- Cochran’s Q test and I2 . Subgroup analysis according to loca- per gastrointestinal lesions, but there is no evidence of su- tion was performed. Results: This meta-analysis included 24 periority over established techniques. © 2022 The Author(s). Published by S. Karger AG, Basel [email protected] © 2022 The Author(s). Correspondence to: www.karger.com/pjg Published by S. Karger AG, Basel Sara Oliveira de Brito, [email protected] is is an Open Access article licensed under the Creative Commons Attribution-NonCommercial-4.0 International License (CC BY-NC) (http://www.karger.com/Services/OpenAccessLicense), applicable to the online version of the article only. Usage and distribution for com- mercial purposes requires written permission.

Eficácia e segurança da Cirurgia Laparoscópica e mento seguro e eficaz para lesões gastrointestinais Endoscópica Cooperativa em lesões gastrointestinais superiores: revisão sistemática e meta-análise superiores, mas sem evidência de superioridade face às técnicas existentes. © 2022 The Author(s). Published by S. Karger AG, Basel Palavras Chave Introduction Cirurgia cooperativa Laparoscópica e endoscópica · disseção endoscópica da submucosa · resseção Laparoscopic and endoscopic cooperative surgery laparoscópica · Lesões subepiteliais · Meta-análise (LECS) is a procedure, which combines the advantages of endoscopy and laparoscopy. It was proposed by Hiki et Resumo al. [1] in 2008 as a technique to resect gastric subepithe- Introdução e objetivos: A Cirurgia cooperativa lapa- lial lesions (SELs). Before LECS was developed, SELs were roscópica e endoscópica (LECS) combina vantagens da generally treated by laparoscopic wedge resection (LWR). endoscopia e laparoscopia na resseção de lesões gastro- However, gastric SELs may not be recognized from out- intestinais superiores. O nosso objetivo é avaliar a eficá- side of the stomach wall, making it difficult to accurately cia e segurança da LECS em pacientes com lesões na jun- determine resection lines through LWR [2–4]. This can ção esofagogástrica (EGJ), estômago e duodeno, e com- lead to incomplete or excessive resection, which may lead parar a LECS com procedimentos puramente to increased recurrence or postoperative alterations of the endoscópicos e laparoscópicos. Métodos: PubMed, Sco- stomach with gastric stasis [2, 4]. pus, ISI Web of Knowledge foram pesquisadas. Dados so- bre eficácia (R0, recorrência) e segurança (taxa de con- The first proposed technique was classical LECS that versão, duração do procedimento e hospitalização, recor- consists of endoscopic confirmation of the incision lines, rência, eventos adversos, mortalidade) foram colhidos e followed by an endoscopic mucosal incision, while the agrupados (odds ratio ou média das diferenças), usando seromuscular layer is incised laparoscopically. At the end, modelo de efeitos randomizados. Qualidade dos estudos the incision line is sutured laparoscopically and the spec- foi avaliada pela Escala Newcastle-Ottawa e heterogene- imen is removed transabdominally [5, 6]. Classical LECS idade pelos testes Q da Cochran e I2. Foi realizada análise was proposed for SELs without ulceration, regardless of de subgrupos, consoante a localização. Resultados: Esta location. Its main benefits are allowing complete resec- meta-análise incluiu 24 estudos/1336 pacientes (todos tion with minimal margins, preserving gastric motility coortes retrospetivos). Não encontramos diferenças sig- and postoperative quality of life. Specially in esophago- nificativas entre LECS e as técnicas pré-existentes (Dis- gastric junction SELs, LECS could avoid total or proximal seção endoscópica da submucosa (ESD)/Laparoscopia) gastrectomy [4]. However, classical LECS also has some em nenhum aspeto. Porém, encontramos uma tendência limitations: possible peritoneal contamination with tu- para hospitalização mais curta, procedimentos mais lon- mor cells or gastric juice (due to opening of the gastric gos e menos efeitos adversos na LECS versus Laparosco- wall) and requirement of advanced endoscopic and lapa- pia e ESD. R0 tende a ser maior no grupo LECS. Hospital- roscopic skills [7, 8]. ização foi significativamente menor em lesões gástricas versus EGJ (média 7.3 vs. 13.7 dias, 95% CI: 6.6–7.9 vs. 8.9– To overcome these disadvantages and expand LECS 19.3). Não encontramos diferenças significativas na taxa for treatment of SELs with ulceration and gastric epithe- de conversão, eventos adversos nem tempo médio de lial neoplasms, some modifications were developed, such procedimento. Porém encontramos uma tendência para as inverted LECS, combination of laparoscopic and endo- taxas de conversão maiores e procedimentos mais lon- scopic approaches to neoplasia with nonexposure tech- gos na EGJ e maior taxa de eventos adversos no duodeno. nique (CLEAN-NET), nonexposed endoscopic wall-in- Conclusão: LECS é um tratamento válido, seguro e eficaz version surgery (NEWS), and closed-LECS [4, 6, 7, 9]. em pacientes com lesões na EGJ, estômago e duodeno, apesar dos estudos retrospetivos existentes estarem pro- Inverted LECS decreases the risk of intraperitoneal pensos a viés de seleção. São necessários estudos pros- seeding by inverting the mass into the gastric lumen. Al- petivos para avaliar a superioridade da LECS face às téc- though this risk is not null, there is also intentional per- nicas existentes. Mensagens-chave: LECS é um trata- foration. In addition to classical LECS, it allows the resec- tion of masses with less than 5 cm regardless of the loca- tion [6, 7, 10]. Efficacy and Safety of Laparoscopic GE Port J Gastroenterol 2023;30:4–19 5 Endoscopic Cooperative Surgery DOI: 10.1159/000526644

CLEAN-NET, NEWS, and closed-LECS are only suit- endoscopic cooperative surgery” OR LECS OR D-LECS OR “non- able for SELs lesser than 3 cm. CLEAN-NET includes exposed endoscopic wall-inversion surgery” OR “non-exposed en- eversion of the mass and nonexposed full-thickness re- doscopic wall-inversion surgery” OR “Nonexposed wall-inversion section after seromuscular incision, preserving the conti- surgery” OR “non-exposed wall-inversion surgery” OR CLEAN- nuity of the mucosa, that works as a barrier. The risk of NET OR “combination of laparoscopic and endoscopic approach- mucosal laceration in lesions superior to 3 cm justifies the es for neoplasia with non-exposure technique” OR “Closed-LECS” size limitation [9, 11]. NEWS stands on a “suture first and OR “closed laparoscopic and endoscopic cooperative surgery” OR then cut” rule, including a full-thickness resection tech- “non-exposure endoscopic-laparoscopic cooperative surgery” OR nique without intentional perforation [8, 12, 13]. The “inverted LECS”). Queries for other databases were adapted from previous 2 types of modified LECS are not indicated for this query. No time or language restrictions were made in this lesions located at the EGJ or pyloric ring. On closed-LECS phase. and NEWS, lesions are retrieved by the transoral route [3, 4, 7, 9, 14, 15]. After removal of duplicates, two authors (J.T., S.B.) indepen- dently screened all titles and abstracts to exclude irrelevant studies. Endoscopic submucosal dissection (ESD) is the first- The full text of selected and relevant studies was then evaluated line treatment for gastric epithelial lesions without deep independently by the same two researchers according to the inclu- submucosal invasion, allowing en bloc resection, inde- sion criteria described below. This phase was performed with pendently of size. Despite this, in SELs the risk of perfora- Rayyan online platform [21]. Disagreements were solved by con- tion is higher since the resection plan is deeper and is as- sensus between the authors or with the intervention of a third re- sociated with long operation times and long learning viewer (D.L.) when required. curves [6, 7, 16]. Modified LECS procedures can also have a role in SELs located in the deep submucosa/muscle lay- Inclusion criteria were: (1) retrospective or prospective, case- er and in some early gastric cancers that would be techni- control, or cohort studies and clinical trials; (2) including patients cally difficult to treat with ESD. Specifically in the duode- submitted to laparoscopic and endoscopic cooperative surgery due num, where ESD is associated with a high perforation risk to esophagogastric, gastric and/or duodenal lesions (single-arm (20–30%), LECS is an attractive option since it might be studies as well as comparative studies with competing techniques safer than ESD and conventional surgery [4–6]. were included); (3) evaluating at least one of the following efficacy or safety outcomes: R0/complete resection; need for conversion; According to previous studies, LECS is a safe and fea- procedure time; hospitalization time; recurrence; adverse events; sible procedure, with a complication rate lesser than 5%, and mortality. Exclusion criteria were: (1) case reports, reviews, but there are no studies demonstrating if LECS is in fact letters to editor, surveys, and animal studies; (2) language other better than ESD or conventional surgery, neither showing than English/Portuguese/Spanish/Italian/French; (3) studies pub- which is the better LECS procedure, among the classical lished only in abstract form; (4) studies including less than 10 pa- and the modified ones [2, 4, 6, 17–19]. tients; and (5) studies with patient overlap with other included studies (in this case, the most informative reference was used). The aim of this systematic review and meta-analysis was to evaluate efficacy and safety outcomes of LECS for Data Extraction and Quality Evaluation gastric, EGJ, and duodenal lesions and to compare LECS Data extraction was performed by S.B. and D.L. Data extraction with competing techniques (ESD and conventional sur- forms included (1) author, publication year, country, study period, gery) whenever possible. study design, setting (2) population characteristics: numbers of participants, tumor location, histological subtypes; (3) type of re- Methods section techniques: LECS (NEWS, CLEAN-NET; closed-LECS), LWR, ESD, laparoscopic resection; and (4) outcomes: R0, need for This systematic review and meta-analysis followed the princi- conversion, procedure time (minutes), hospitalization time (days), ples set in the Preferred Reporting. Items for Systematic Reviews recurrence, adverse events, and mortality. Data regarding costs and Meta-Analyses: the PRISMA Statement [20]. and quality of life were also extracted whenever possible. Quality evaluation was performed using the Newcastle-Ottawa Study Search and Selection Assessment Scale adapted, for cohort studies, by S.B. and D.L. [22]. Studies were identified through scanning of 3 electronic data- Newcastle-Ottawa Scale ranges from 0 to 9 points in double-arm bases (MEDLINE through PubMed, Scopus, and ISI Web of studies and from 0 to 6 points in single-arm studies. No specific Knowledge), with the last search performed on March 8, 2021. The value is assigned to high or low quality, although higher scores in- search query for PubMed was (gastric OR stomach OR duodenal dicate higher quality and greater methodological aspects. OR duodenum OR “esophagogastric junction”) AND (“laparo- scopic and endoscopic cooperative surgery” OR “Laparoscopic- Data Synthesis and Statistical Analysis Raw data regarding each outcome (number of events and total) were collected in order to calculate outcome prevalence and stan- dard error. Effect measures included odds ratio (OR) for categori- cal variables and mean difference (MD) for continuous variables. For continuous outcomes, in some studies, median and range were transformed into mean and standard deviation through the meth- ods proposed by the Cochrane collaboration and Hozo et al. [23, 24]. Meta-analysis was performed with Review Manager 5.4 soft- 6 GE Port J Gastroenterol 2023;30:4–19 de Brito/Libânio/Pinto/Araújo Teixeira/ Araújo Teixeira DOI: 10.1159/000526644

Identification Records identified through Additional records identified database searching through other sources Pubmed (n = 180) (n = 0) Scopus (n = 651) Web Of Science (n = 163) Total (n = 994) Records after duplicates removed (n = 690) Screening Records screened Records excluded (n = 690) (n = 453) Eligibility Full-text articles assessed Full-text articles excluded (n = 213), for eligibility with reasons: Included (n = 237) Case reports (n = 87) Studies included in Reviews (n = 7) qualitative synthesis Not LECS (n = 40) (n = 24) Less than 10 patients (n = 18) Foreign language (n = 8) Studies included in quantitative synthesis Abstracts without full text (n = 15) Editorials (n = 2) (meta-analysis) (n = 24) Original articles (n = 28) Not 1 outcome defined (n = 2) Patients overlap (n = 6) Fig. 1. Flowchart of included studies. LECS, laparoscopic and endoscopic cooperative surgery. ware [25], using a random-effect model (when at least 3 studies Results were available for each outcome) [26]. Heterogeneity was evalu- Study Selection ated with Cochran Q test and I2, being significant heterogeneity A total of 994 studies were identified through data base defined as p < 0.05 or I2 >40%, respectively. Subgroup analysis ac- search. After removing the duplicates, 690 studies were screened regarding title and abstract and 453 were con- cording to lesion was performed for 4 outcomes (need for conver- sidered irrelevant. Therefore, the full text of 237 studies was assessed for eligibility by applying inclusion and ex- sion, procedure time, hospitalization time, and adverse events). clusion criteria. Of those, 24 were included in this system- atic review and all were included in meta-analysis. Study Pooled mean for continuous variables and prevalence for categor- flowchart is shown in Figure 1, according to the PRISMA statement [20]. ical variables were calculated with OpenMetaAnalyst and Meta- XL, using a random-effect model [26, 27]. Publication bias was planned if ≥10 studies were included in comparative analysis for the primary outcomes (procedure time and adverse events). Efficacy and Safety of Laparoscopic GE Port J Gastroenterol 2023;30:4–19 7 Endoscopic Cooperative Surgery DOI: 10.1159/000526644

8 GE Port J Gastroenterol 2023;30:4–19 Table 1. General characteristics of included studies DOI: 10.1159/000526644 Study and year Country Type of study N Techniq Esophagogastric junction Japan Retrospective 21 LECS (N Aoyama et al. [42], 2020 Japan Single-center 5 CLEAN- Japan 20 LECS Hoteya et al. [19], 2014 Retrospective Czech Republic Single-center 23 LECS Ri et al. [48], 2020 and Austria Retrospective 69 NEWS ( Stomach Single-center Hajer et al. [39], 2019 15 CLEAN- Retrospective 28 (N = 19 Two-center 69 LECS (N Kanehira et al. [44], 2020 Japan Retrospective 42 Single-center 71 NEWS 60 Mahawongkajit and Thailand Retrospective 45 CLEAN- Chanswangphuvana [45], 2020 Japan Single-center 20 Laparos Mitsui et al. [34], 2018 Japan 48 (N = 44 Retrospective 101 NEWS Okubo et al. [47], 2020 Single-center 10 194 LECS (N Retrospective 22 LWR (N Single-center LECS (N de Brito/Libânio/Pinto/Araújo Teixeira/ Aoyama et al. [41], 2020 Japan Retrospective LECS (N Araújo Teixeira Shoji et al. [37], 2018 Japan Single-center Balde et al. [30], 2016 China LECS Cao et al. [38], 2019 China Retrospective Hoteya et al. [19], 2014 Japan Single-center LECS (N Komatsu et al. [32], 2016 Japan Kang et al. [28], 2013 China Retrospective LECS Kikuchi et al. [14], 2017 Japan Single-center Ri et al. [48], 2020 Japan Closed- Waseda et al. [29], 2014 Japan Retrospective Single-center LECS Retrospective LECS Single-center Retrospective Single-center Retrospective Single-center Retrospective Single-center Retrospective Single-center Retrospective Single-center

que Histological type Newcastle N = 13); NEWS (N = 4); 18 Leiomyomas, 2 GISTs, 1 Ectopic pancreas 6* -NET (N = 4) 4 Leiomyoma, 1 GIST 6 Not available (all SEL) 7 (N = 11); LWR (N = 12) NEWS: 6 GISTs, 1 Submucosal lipoma, 1 Leiomyoma, 1 Endocrine tumor, 6 1 Vanek’s tumor, and 1 Ectopic pancreatic; LWR: 7 GISTs, 1 Leiomyoma, -NET (N = 50); LWR 2 Ectopic pancreatic tissue, 1 Endometriosis, 1 Hyperplasiogenic polyp 6 9) N = 10); NEWS (N = 5) CLEAN-NET: 25 GIST; 10 Schwannoma; 6 Leiomyoma; 3 Ectopic pancreas; 3 5* Granuloma; 1 Carcinoid tumor; 1 Glomus tumor; 1 Cyst; LWR: 11 GIST; 3* -NET (N = 25); 7 Leiomyoma; 1 Schwannoma 7 scopic distal gastrectomy 4) LECS: 9 GIST; 1 Leiomyoma NEWS: 3 GIST; 1 Schwannoma; 1 Ectopic pancreas N = 14); NEWS (N = 26); 28 GIST N = 31) N = 30); ESD (N = 30); Not available N = 25) ESD (N = 20) 24 GIST; 6 Ectopic pancreas; 5 Leiomyoma; 3 Neurinoma; 2 Granuloma; 6* N = 33) LWR (N = 15) 1 Hemangioma; 1 Gastritis cystica profunda; 1 Desmoid fibromatosis 7 LECS: 6 GIST; 7 Leiomyoma; 1 other; NEWS: 16 GIST; 2 Leiomyoma; 8 others; 9 -LECS LWR: 27 GIST; 1 Leiomyoma; 3 others 5 60 GIST 6 7 45 GIST 6* 5* 15 GIST; 1 Leiomyoma; 1 Schwannoma; 1 Cavernous hemangioma; 2 Aberrant 7 pancreas 6* Not available 78 GIST; 13 Leiomyoma; 3 Ectopic pancreas; 3 NET; 2 Lipoma; 1 Glomus tumor; 1 Inflammatory pseudotumor Not available Not available (all SEL) 10 GIST; 1 NET; 3 Aberrant pancreas; 3 Leiomyoma; 2 Glomus tumor; 2 Schwannoma; 1 Granuloma

Efficacy and Safety of Laparoscopic Table 1 (continued) Country Type of study N Techniq Endoscopic Cooperative Surgery Japan 46 LECS (N Study and year Japan Retrospective 20 Ojima et al. [36], 2018 China Single-center 91 LECS Tsujimoto et al. [17], 2012 Yin et al. [33], 2017 Japan Retrospective LECS (N Duodenum Japan Single-center resectio Kanaji et al. [43], 2020 Japan Ichikawa et al. [31], 2016 Retrospective 20 D-LECS Nunobe et al. [46], 2020 Single-center D-LECS 12 (13 D-LECS Retrospective lesions) Single-center 206 Retrospective Single-center Retrospective Multicenter Ojima et al. [35], 2018 Japan Retrospective 68 D-LECS Single-center D-ESD ( Yanagimoto et al. [40], 2019 Japan Retrospective 10 D-LECS Thailand Single-center 1 NEWS Mahawongkajit and GE Port J Gastroenterol 2023;30:4–19 Chanswangphuvana [45], 2020 Retrospective DOI: 10.1159/000526644 Single-center N, total number of cases; M-A, meta-analysis; LECS, laparoscopic and endoscopic cooperative approaches to neoplasia with non-exposure technique; GIST, gastrointestinal stromal tumor; SE laparoscopic and endoscopic cooperative surgery; D-LECS, laparoscopic and endoscopic cooper Quality Assessment Form for Cohort Studies ranges from 0 to 6 points. 9

que Histological type Newcastle N = 21); EIGS (N = 25) 46 GIST 7 N = 15); Laparoscopic 16 GIST, 1 Inflammation for parasite; 1 Leiomyoma; 1 Glomus tumor; 1 Aberrant 5* on (N = 30); ESD (N = 46) pancreas 9 S S 91 GIST S 16 Adenocarcinoma; 2 NET; 1 Adenoma; 1 Ectopic pancreas 5* S (N = 18) 5* (N = 50) 10 Adenocarcinoma; 2 NET; 1 Adenoma 5* S 79 Adenoma; 70 Adenocarcinoma; 36 NET; 7 GIST; 6 Hyperplasia of Brunner’s 5 gland; 5 Hyperplastic polyp; 1 Ectopic pancreas, 1 Schwannoma; 1 Ulcer formation and malformation of vessels 6* D-LECS: 7 Intramucosal adenocarcinoma; 1 Adenoma; 6 NET; 4 GIST 5* D-ESD: 14 Intramucosal adenocarcinoma; 27 Adenoma; 6 NET; 1 Hyperplastic polyp; 2 Lipoma 6 Adenocarcinoma; 3 Adenoma; 1 Neuroendocrine tumor 1 Neuroendocrine tumor e surgery; NEWS, nonexposed endoscopic wall-inversion surgery; CLEAN-NET, combination of laparoscopic and endoscopic EL, subepithelial lesions; LWR, laparoscopic wedge resection; ESD, endoscopic submucosal dissection; closed-LECS, closed rative surgery for duodenal neoplasm. *These are single-arm studies. Therefore, the score potential using Newcastle-Ottawa

a b c d Fig. 2. a–g Forest plots of several outcomes e according to surgical technique. a Forest f plot of procedure time according to surgi- g cal technique (Laparoscopic Techniques vs. LECS). b Forest plot of hospitalization time according to surgical technique (Lap- aroscopic Techniques vs. LECS). c Forest plot of adverse event according to the sur- gical technique (Laparoscopic Techniques vs. LECS). d Forest plot of R0 according to the surgical technique (ESD vs. LECS). e Forest plot of procedure time according to the surgical technique (ESD vs. LECS). f Forest plot of hospitalization time accord- ing to the surgical technique (ESD vs. LECS). g Forest plot of adverse event ac- cording to the surgical technique (ESD vs. LECS). LECS, laparoscopic and endoscopic cooperative surgery; ESD, endoscopic sub- mucosal dissection; SD, standard devia- tion. 10 GE Port J Gastroenterol 2023;30:4–19 de Brito/Libânio/Pinto/Araújo Teixeira/ Araújo Teixeira DOI: 10.1159/000526644

Table 2. Comparative studies Efficacy and Safety of Laparoscopic Study Location N His Endoscopic Cooperative Surgery EGJ versus Non-EGJ ver Hoteya et al. [19], 2014 EGJ 5 100 Ri et al. [48], 2020 Non-EGJ 20 100 EGJ 20 100 LECS versus Laparoscopic Techniques Non EGJ 194 100 Hajer et al. [39], 2019 NEWS 11 100 Kanehira et al. [44], 2020 92% LWR 12 100 Shoji et al. [37], 2018 100 CLEAN-NET 50 100 Komatsu et al. [32], 2016 100 LWR 19 NR Okubo et al. [47], 2020 NR LECS (LECS and NEWS) 40 NR Yin et al. [33], 2017 NR LWR 31 100 LECS versus ESD 100 Balde et al. [30], 2016 LECS 33 100 Cao et al. [38], 2019 LWR 15 100 100 GE Port J Gastroenterol 2023;30:4–19 Yin et al. [33], 2017 CLEAN-NET 25 100 DOI: 10.1159/000526644 100 Duodenal LECS versus ESD Laparoscopic distal gastrectomy 44 100 Ojima et al. [35], 2018 LECS 15 56% 16% Laparoscopic resection* 30 LECS 30 ESD 30 LECS 25 ESD 20 LECS 15 ESD 46 LECS 18 ESD 50 11 N, total number of cases; SEL, subepithelial lesions; EL, epithelial lesions; EGJ, eso surgery; NEWS, nonexposed endoscopic wall-inversion surgery; LWR, laparoscopic nonexposure technique; ESD, endoscopic submucosal dissection. *This includes LWR

stology SEL Ro Need for Procedure time, Hospitalization Recurrence Adverse Mortality rsus EL (%) conversion events min time, days 0% versus 0% 5 0 196±48.6 13.2±3.7 0 0 NR 0% versus 0% 20 0 145.9±36.5 9.9±1.4 0 0 NR 0% versus 0% 20 8 320.25±102.25 20.5±12.5 NR 4 NR 0% versus 0% 194 2 195.25±86.321 31.75±30.3442 NR 8 NR 0% versus 0% 11 NR 95.5±14.4309 7.15±1.4649 0 4 0 % versus 8% 11 NR 64.25±14.7302 6.75±2.3229 0 1 2 0% versus 0% 50 0 105.4±42.5 6.1±0.25 0 0 0 0% versus 0% 19 0 58.1±13.75 6±0.00001 0 0 0 0% versus 0% 40 NR NR NR 0 0 NR 0% versus 0% 31 NR NR NR 0 7 NR R 33 NR 220±82.5 10±5.75 0 0 NR R 15 NR 129±82.0008 21.25±12.4457 2 3 NR R NR NR 325.8±17.72 NR NR NR NR R NR NR 369.3±13.35 NR NR NR NR 0% versus 0% 15 0 65.33±20.57 6.33±2.53 0 2 0 0% versus 0% 30 0 81.67±22.49 6.53±2 1 3 0 0% versus 0% 30 0 96.5±15.075 6±1.25 0 1 NR 0% versus 0% 27 0 41.5±8.375 6±0.325 2 8 NR 0% versus 0% 25 0 80.76±13.86 3.44±1 0 0 0 0% versus 0% 18 0 84.05±15.33 7.85±1.18 0 10 0 0% versus 0% 15 0 65.33±20.57 6.33±2.53 0 2 0 0% versus 0% 45 0 32.96±11.76 5.83±2.69 0 7 0 % versus 44% 18 0 132.75±57.5 7±1.5 NR 0 0 % versus 84% 26 0 41±28 3±12 NR 14 0 ophagogastric junction; NR, data not reported on the study; LECS, laparoscopic and endoscopic cooperative wedge resection; CLEAN-NET, combination of laparoscopic and endoscopic approaches to neoplasia with R, proximal, and distal gastrectomy.

Study Characteristics and Quality Evaluation procedure time, hospitalization time, and adverse events The main study characteristics are shown in Table 1 were not significantly different between the two groups [14, 17, 19, 28–48]. Most of the studies (23) were from (shown in Fig. 2d–g), although there was a trend to high- Asia (96%) and 22 studies (92%) were single-center. All er complete resection rate in LECS’s group (100% vs. 94%, studies were retrospective. The median Newcastle-Otta- OR: 4.01, 95% CI: 0.67–24.22, I2 = 0%), longer procedure wa Scale score of included studies was 6 (7 in double-arm time in LECS’ group (MD 28.1, 95% CI: −8.81 to 65.00, I2 studies and 5 in single-arm studies). A total of 1,337 le- = 98%), shorter hospitalization time in LECS’ group (MD sions, from 1,336 patients, were included in the analysis, −1.33, 95% CI: −4.65 to 1.99, I2 = 98%), and lower adverse of which 46 (3.4%) were located in the EGJ, 974 (72.9%) events rate in LECS’ group (4,3% vs. 26%, OR: 0.15, 95% in the stomach, and 317 (23.7%) in the duodenum. Half CI: 0.02–1.34, I2 = 67%). The conversion rate was 0% in of the studies (12) were single-arm (of which 1 study was both groups. Recurrence was noticed in 2 of the 96 cases about lesions on the EGJ, 6 on the stomach, 4 on the duo- (2.08%) using ESD [30] and in 0 of the 70 cases using denum, and 1 had lesions on both the stomach and the LECS. No deaths were reported in 106 cases (shown in duodenum [45]). The other 12 studies were comparative. Table 2). One study reported data regarding costs [37] and another one regarding quality of life [47]. Duodenal LECS versus ESD One study compared the outcomes of LECS and ESD Comparative Studies in duodenal lesions [35]. R0 was 100% in LECS’ group, LECS versus Laparoscopic Techniques but only 52% in ESD group, corresponding to 26 among Six studies compared the outcomes of LECS and lapa- the 50 cases. There was no need for conversion in any roscopic techniques (4 LECS vs. LWR [32, 37, 39, 44]; 1 studies. Procedure duration and hospitalization time LECS vs. laparoscopic gastrectomy [47]; 1 LECS vs. LWR were higher in LECS’ group. Recurrence was not report- and gastrectomy [33]) in gastric lesions. In meta-analysis, ed. ESD’s group accounted for a total of 14 adverse events procedure time, hospitalization time, and adverse events (28%), whereas LECS had 0 adverse events among its 18 were not significantly different between the two groups cases (0%). No deaths were reported (shown in Table 2). (shown in Fig. 2a–c), although there was a trend to longer Meta-analysis was not performed due to a low number of procedure time in LECS’ group (MD 18.3 min, 95% CI: studies. −23.1 to 60.3, I2 = 98%), shorter hospitalization time in LECS’ group (MD −0.37, 95% CI: −1.8 to 1.0, I2 = 74%), Single-Arm Studies regarding LECS and lower adverse events rate in LECS’ group (4.0% vs. Half of all included studies were noncomparative 13.1%, OR: 0.44, 95% CI: 0.04–4.65, I2 = 72%). Despite the LECS studies. Twenty-four studies provided data for the tendency to longer procedure time in LECS’ group, in the calculation of pooled efficacy and safety outcomes. This 2 studies that compared LECS with Laparoscopic gastrec- is shown in Table 3. tomy [33, 47], procedure time was higher in laparoscopic techniques’ group (shown in Table  2). R0 was 100% in EGJ LECS group and 99.6% (1/256) in LWR/laparoscopy’s Three studies provided data on 46 EGJ lesions [19, 42, group. R0 was not achieved in 1 case using LWR [39]. The 48]. R0 resection rate was 100% and pooled conversion conversion rate was 0% in both groups. Recurrence oc- rate was 11.3% (95% CI: 0–47.1%, I2 = 87%). The 8 con- curred in 3 of 107 cases (2.8%) in laparoscopic techniques’ version cases occurred in the same study [48]. Pooled group [32, 33] and in 0 of 149 cases (0%) in LECS’ group. mean procedural time was 246 min (95% CI: 185–307, I2 Two deaths were reported among the 61 cases of the lap- = 89%), and pooled mean hospitalization time was 13.7 aroscopic group [39], and 0 among the 76 cases of the days (95% CI: 8.0–19.3, I2 = 91%). Two studies reported LECS’ group (shown in Table  2). Both reported deaths recurrence rates [19, 42] and it was 0% (0 in 26 cases). were not related to the oncological disease neither its Pooled adverse events’ rate was 7.1% (95% CI: 0–24.7%, treatment. Meta-analysis was not possible in these last I2 = 67%), with all 4 adverse events occurring in the same outcomes. study [48]. Mortality was 0% (0 in 21 cases). LECS versus ESD Gastric Three studies compared the outcomes of LECS and Eighteen studies provided data on 702 gastric lesions ESD in gastric lesions [30, 33, 38]. In meta-analysis, R0, [14, 17, 19, 28–30, 32–34, 36–39, 41, 44, 45, 47, 48]. R0 12 GE Port J Gastroenterol 2023;30:4–19 de Brito/Libânio/Pinto/Araújo Teixeira/ Araújo Teixeira DOI: 10.1159/000526644

Table 3. Single-arm studies Study N Histology SEL/EL Ro Need for Procedure time, Hospitalization Recurrence Adverse Mortality (%) conversion min time, days events EGJ 21 100% versus 0% 21 0 225±48 9.0±1.9 0 00 Aoyama et al. [42], 2020 5 100% versus 0% 5 0 196.0±48.6 13.2±3.7 0 0 NR Hoteya et al. [19], 2014 20 100% versus 0% 20 8 320.25±102.25 20,5±12,5 NR 4 NR Ri et al. [48], 2020 Stomach 11 100% versus 0% 11 NR 95.5±14.4309 7.15±1.4649 0 40 Hajer et al. [39], 2019 0 105.4±42.5 6,1±0.25 0 00 Kanehira et al. [44], 2020 50 100% versus 0% 50 2 193.107±59.726 NR NR 1 NR Mitsui et al. [34], 2018 NR 325.80±17.72 NR NR NR NR Okubo et al. [47], 2020 28 100% versus 0% NR NR NR NR 0 00 Mahawongkajit and Chanswangphuvana [45], 2020 NR 198±14.25 7.0±0.425 0 10 Aoyama et al. [41], 2020 25 NR NR NR NR NR 0 0 NR Shoji et al. [37], 2018 0 96.5±15.075 6.0±1.25 0 1 NR Balde et al. [30], 2016 15 100% versus 0% 15 0 80.76±13.86 3.44±1.00 0 00 Cao et al. [38], 2019 0 145.9±36.5 9.9±1.4 0 0 NR Hoteya et al. [19], 2014 42 100% versus 0% 42 NR 220±82.5 10±5.75 0 0 NR Komatsu et al. [32], 2016 0 113±36 NR 0 20 Kang et al. [28], 2013 40 100% versus 0% 40 1 253±45 9.2±1.5 0 2 NR Kikuchi et al. [14], 2017 2 181±49.8333 7±17.5 NR 8 NR Ri et al. [48], 2020 30 100% versus 0% 30 0 NR NR 0 2 NR Waseda et al. [29], 2014 NR 151±53.5 8.5±2.5 1 10 Ojima et al. [36], 2018 25 100% versus 0% 25 0 157.5±68.4 11.6±9.5 0 00 Tsujimoto et al. [17], 2012 0 65.33±20.57 6.33±2.53 0 20 Yin et al. [33], 2017 20 100% versus 0% 20 Duodenum 0 00 Mahawongkajit and Chanswangphuvana [45], 2020 33 NR 33 0 20 Kanaji et al. [43], 2020 0 4 NR Ichikawa et al. [31], 2016 101 100% versus 0% 101 11 38 NR Nunobe et al. [46], 2020 0 00 Ojima et al. [35], 2018 10 NR NR 0 2 NR Yanagimoto et al. [40], 2019 194 100% versus 0% 194 22 100% versus 0% 22 21 100% versus 0% 21 20 100% versus 0% 20 15 100% versus 0% 15 1 100% versus 0% 1 261 NR 0 20 15% versus 85% 19 236.25±56.75 12.5±2.5 0 12 15% versus 85% NR 358.5±101.0511 18.5±12.1552 0 206 25% versus 75% 196 180±90.1667 9±12 0 18 56% versus 44% 18 132.75±57.5 7±1.5 NR 10 10% versus 90% 10 256.35±60.6412 8.75±2.0052 0 N, total number of patients; SEL, subepithelial lesions; EL, epithelial lesions; EGJ, esophagogastric junction; NR, data not reported on the study. resection rate was 100%. Pooled conversion rate was 1.0% adverse events reported in 677 cases. Mortality was re- (95% CI: 0.4–2.2%, I2 = 0%). The 5 conversion cases oc- ported only in 9 studies and it was 0% (0 in 300 cases). curred in 3 different studies [14, 34, 48]. Pooled mean Duodenal procedural time was 158 min (95% CI: 118–199, I2 = Six studies provided data on 267 duodenal lesions [31, 35, 40, 43, 45, 46]. R0 resection rate was 95.7%. Incom- 100%), and pooled mean hospitalization time was 7.3 plete resection was reported in 11 of 255 cases (4.3%), days (95% CI: 6.6–8.0, I2 = 98%). Fifteen studies reported which included 2 studies: Kanaji et al. [43] and Nunobe et al. [46] . Pooled conversion rate was 4.3% (95% CI: recurrence rates and it was 0.22% (1 in 455 cases). This 2.2–7.1; I2 = 0%). Need for conversion was reported in 11 case was reported by Ojima et al. [36], in an 85-year-old patient with a 90-mm lesion. Pooled adverse events’ rate was 3.5% (95% CI: 1.8–6.0; I2 = 44%), with a total of 24 Efficacy and Safety of Laparoscopic GE Port J Gastroenterol 2023;30:4–19 13 Endoscopic Cooperative Surgery DOI: 10.1159/000526644

Need for conversion by Tumor location Study or Subgroup Prev (95% CI) % Weight Gastric 0,000 ( 0,000, 0,034) 7,1 [44] Kanehira E 2020 0,071 ( 0,001, 0,204) 5,5 [34] Mitsui T 2018 0,000 ( 0,000, 0,057) 5,7 [30] Balde AI 2016 0,000 ( 0,000, 0,068) 5,2 [38] Cao L 2019 0,000 ( 0,000, 0,085) 4,6 0,000 ( 0,000, 0,017) 8,8 [19] Hoteya S 2014 (gastric) 0,100 ( 0,000, 0,381) 2,9 [28] Kang WM 2013 0,010 ( 0,000, 0,031) 9,8 [14] Kikuchi S 2017 0,000 ( 0,000, 0,077) 4,9 [48] Ri M 2020 (Gastric) 0,000 ( 0,000, 0,085) 4,6 [29]Waseda Y 2014 0,000 ( 0,000, 0,112) 3,9 [17]Tsujimoto H 2012 0,010 ( 0,004, 0,022) 63,2 [33] Yin X 2017 0,000 ( 0,000, 0,081) 4,7 Gastric subgroup 0,000 ( 0,000, 0,320) 1,8 Q=8,77, p=0,55, I2=0% 0,400 ( 0,194, 0,625) 4,6 EGJ 0,113 ( 0,000, 0,471) 11,1 [42] Aoyama J 2020 [19] Hoteya S 2014 (EGGEJ) 0,000 ( 0,000, 1,000) 0,5 [48] Ri M 2020 (EGGEJ) 0,000 ( 0,000, 0,085) 4,6 0,000 ( 0,000, 0,140) 3,3 EGGEJ subgroup Q=14,86, p=0,00, I2=87% 0,053 ( 0,026, 0,089) 9,9 0,000 ( 0,000, 0,094) 4,3 Duodenal 0,000 ( 0,000, 0,166) 2,9 [45] Mahawongkajit P 2020 0,043 ( 0,022, 0,071) 25,7 [43] Kanaji S 2020 [31] Ichikawa D 2016 0,025 ( 0,009, 0,048) 100,0 [46] Nunobe S 2020 0 1 [35]Ojima T 2018 [40] Yanagimoto Y 2019 Prevalence Duodenal subgroup Q=2,69, p=0,75, I2=0% Overall Q=43,04, p=0,00, I2=56% a Adverse events by Tumor location Study or Subgroup Prev (95% CI) % Weight Gastric 0,364 ( 0,101, 0,673) 2,9 0,000 ( 0,000, 0,034) 5,0 [39] Hajer J 2019 0,000 ( 0,000, 0,112) 3,4 [44] Kanehira E 2020 0,036 ( 0,000, 0,147) 4,3 [45] Mahawongkajit P 2020 (Gastric) 0,024 ( 0,000, 0,099) 4,8 [34] Mitsui T 2018 0,000 ( 0,000, 0,043) 4,7 0,033 ( 0,000, 0,138) 4,4 [41] Aoyama J 2020 (Gastric) [37] Shoji Y 2018 0,000 ( 0,000, 0,068) 4,1 [30] Balde AI 2016 0,000 ( 0,000, 0,085) 3,8 [38] Cao L 2019 0,000 ( 0,000, 0,052) 4,5 0,020 ( 0,000, 0,059) 5,6 [19] Hoteya S 2014 (Gastric) [32]Komatsu S 2016 0,200 ( 0,006, 0,513) 2,7 [28] Kang WM 2013 0,041 ( 0,017, 0,074) 6,0 0,091 ( 0,002, 0,255) 3,9 [14] Kikuchi S 2017 0,048 ( 0,000, 0,193) 3,9 [48] Ri M 2020 (Gastric) 0,000 ( 0,000, 0,085) 3,8 [29] Waseda Y 2014 0,133 ( 0,003, 0,362) 3,4 [36] Ojima T 2018 (Gastric) [17] Tsujimoto H 2012 [33]Yin X 2017 Gastric subgroup 0,035 ( 0,018, 0,060) 71,2 Q=28,69, p=0,03, I2=44% 0,000 ( 0,000, 0,081) 3,9 EGGEJJ 0,000 ( 0,000, 0,320) 1,8 [42] Aoyama J 2020 ((GEEGJJ)) 0,200 ( 0,049, 0,408) 3,8 [19] Hoteya S 2014 ((GEEGJ) [48] Ri M 2020 ((GEEGJ) GEEGJJ subgroup 0,071 ( 0,000, 0,247) 9,5 Q=5,97, p=0,05, I2=67% Duodenal 0,100 ( 0,002, 0,279) 3,8 [43] Kanaji S 2020 0,308 ( 0,082, 0,589) 3,1 [31] Ichikawa D 2016 0,184 ( 0,134, 0,241) 6,0 [46] Nunobe S 2020 0,000 ( 0,000, 0,094) 3,6 [35] Ojima T 2018 (Duodenal) 0,200 ( 0,006, 0,513) 2,7 [40] Yanagimoto Y 2019 Duodenal subgroup 0,148 ( 0,058, 0,256) 19,3 Q=9,79, p=0,04, I2=59% Overall 0,059 ( 0,031, 0,095) 100,0 Q=88,49, p=0,00, I2=73% 0 0,05 0,1 0,15 0,2 0,25 0,3 0,35 0,4 0,45 0,5 0,55 0,6 0,65 0,7 b Prevalence Fig. 3. a–d Rates of conversion and adverse events according to according to the location. c Forest plots of mean hospitalization location (gastric, EGJ, and duodenal). Forest plots of mean hospi- time according to the location (gastric, EGJ, and duodenal). d For- talization time and mean procedure time according to the location est plots of mean procedure time according to the location (gastric, (gastric, EGJ and duodenal). a Rates of conversion according to the EGJ, and duodenal). EGJ, esophagogastric junction. location (gastric, EGJ, and duodenal). b Rates of adverse events (Figure continued on next page.) 14 GE Port J Gastroenterol 2023;30:4–19 de Brito/Libânio/Pinto/Araújo Teixeira/ Araújo Teixeira DOI: 10.1159/000526644

c d GE Port J Gastroenterol 2023;30:4–19 15 3 DOI: 10.1159/000526644 Efficacy and Safety of Laparoscopic Endoscopic Cooperative Surgery

cases, in Nunobe et al. [46] . Pooled mean procedural time increased the operative time, resulting in higher person- was 229 min (95% CI: 176–281, I2 = 94%), and pooled nel and anesthetic expenses. mean hospitalization time was 10.1 days (95% CI: 7.5– 12.7, I2 = 95%). Recurrence rate was 0% (0 in 249 cases). Quality of Life Pooled adverse events’ rate was 14.8% (95% CI: 5.8–25.6, Okubo et al. [47] evaluated postoperative quality of life I2 = 59%), with a total of 46 adverse events reported in 267 after local resection (CLEAN-NET) and distal gastrecto- cases. Mortality was reported only in 3 studies [35, 43, 45] my, using the Postgastrectomy Syndrome Assessment and it was 0% (0 in 39 cases). Scale (PGSAS-45) questionnaire and endoscopic evalua- tion at 1, 6, and 12 months after surgery. Authors report- Overall, pooled conversion rate was 2.5% (95% CI: ed significantly endoscopic differences 12 months after 0.9–4.8%, I2 = 56%). Although there were no statistically gastrectomy, with less esophageal reflux and residual gas- significant differences in the conversion rate according to tritis in the group submitted to CLEAN-NET. CLEAN- the location, there was a trend to higher conversion in NET subgroup also presented better clinical symptoms EGJ lesions (11.3%, 95% CI: 0–47.1%, I2 = 87%), followed 12 months after procedure, reporting less indigestion, less by duodenal lesions (4.3%, 95% CI: 2.2–7.1%, I2 = 0%) dissatisfaction during meals, less dissatisfaction for daily and gastric lesions (1.0%, 95% CI: 0.4–2.2%, I2 = 0%). This life, and more amount of food ingested per meal, resulting is shown in Figure 3. Pooled adverse events’ rate was 5.9% in better nutritional status and body weight ratio. (95% CI: 3.1–9.5, I2 = 73%). Although there were no sta- tistically significant differences in adverse events’ rate ac- Histology cording to the location, there was a trend to higher rate of Three studies reported histological evaluation of the adverse events in duodenal lesions (14.8%, 95% CI: 5.8– surgical specimens in the EGJ [19, 42, 48]. All lesions were 25.6, I2 = 59%), followed by EGJ lesions (7.1%, 95% CI: subepithelial (total of 46 lesions). Regarding gastric le- 0–24.7, I2 = 67%) and gastric lesions (3.5%, 95% CI: 1.8–6, sions, 15 studies reported histological evaluation of 847 I2 = 67%). This is shown in Figure 3b. Pooled mean pro- surgical specimens [17, 19, 28–30, 33, 34, 36–39, 41, 44, cedural time was 185 min (95% CI: 153–217, I2 = 100%). 45, 48]. One lesion (0.1%) was epithelial [39] and the oth- Mean procedural time was also not significantly different er 846 lesions (99.9%) were subepithelial. Six studies re- according to the location. However, the same trend was ported histological evaluation of retrieved lesions in the verified: higher procedural time in EGJ lesions (246 min, duodenum, making a total of 318 lesions [31, 35, 40, 43, 95% CI: 184–307, I2 = 100%), followed by duodenal le- 45, 46]. Seventy-six lesions (24%) were subepithelial and sions (229 min, 95% CI: 176–281, I2 = 94%) and gastric 242 (76%) were epithelial lesions. The location with high- lesions (158 min, 95% CI: 118–199, I2 = 100%). This is er number of epithelial retrieved lesions was the duode- shown in Figure 3c. Pooled mean hospitalization time num. This is shown in Tables 2 and 3. was 8.3 days (95% CI: 7.6–8.9, I2 = 98%). Hospitalization time was significantly shorter in gastric lesions versus Discussion EGJ lesions (mean 7.3 vs. 13.7 days, 95% CIs of 6.6–7.9 and 8.9–19.3, respectively). There was higher hospitaliza- To the best of our knowledge, this is the first system- tion time in EGJ lesions (13.7 days, 95% CI: 8–19.3, I2 = atic review and meta-analysis comparing LECS with lap- 91%), followed by duodenal lesions (10.1 days, 95% CI: aroscopic techniques (LWR and laparoscopic gastrecto- 7.5–12.8, I2 = 95%) and gastric lesions (7.3 days, 95% CI: my) and endoscopic techniques (ESD). We found that 6.6–8, I2 = 98%). This is shown in Figure 3d. LECS is an effective and safe therapy for upper GI SELs, with high rates of R0 and low adverse events rates, short- Costs er hospitalization time, and longer procedure time. Ad- Shoji et al. [37] analyzed the operative costs of 3 tech- ditionally, we found that, to date, there is no clear evi- niques: LWR, LECS, and NEWS. NEWS was associated dence of the benefit of LECS over ESD or LWR/laparos- with a significantly lower mean total cost, followed by copy namely in procedural/hospitalization time nor in LECS, and being LWR the most expensive technique. The R0/adverse events, in gastric lesions. major cause pointed by the authors for these differences was the cost of suturing devices, such as laparoscopic lin- Regarding gastric lesions, no significant differences ear staplers, which were less used in the NEWS’ group (p were found between LECS and preexisting techniques < 0.001). Authors also referred that, although a hand- (ESD or LWR/laparoscopy) regarding any outcomes. sewn technique used on LECS and NEWS was cheaper, it 16 GE Port J Gastroenterol 2023;30:4–19 de Brito/Libânio/Pinto/Araújo Teixeira/ Araújo Teixeira DOI: 10.1159/000526644

This might be explained by the high heterogeneity and field, accessing the narrow duodenal lumen, holding the low number of comparative studies. However, we could endoscope on position, and maneuvering it in such lim- observe some trends in our results: LECS was associated ited space [31, 43, 49, 50]. These difficulties can also ex- with higher R0, shorter hospitalization time, and longer plain the higher adverse events’ rates in the duodenum, procedure duration, comparing to ESD. Comparing to comparing with other gastrointestinal locations. laparoscopic techniques, hospitalization time was lower and procedure duration was higher when LECS was the Both procedure and hospitalization durations tended chosen procedure. Although analyzing separately LWR to be longer in the EGJ group, followed by the duodenum. and laparoscopic gastrectomy’s procedure duration, we This can be explained by the higher rate of conversion in noticed that LECS was shorter than those procedures us- these locations, as it will inevitably prolong the duration ing laparoscopic gastrectomies and longer than those us- of the procedure and may lead to more complex tech- ing LWR. LECS was associated with fewer adverse events niques, such as proximal gastrectomies. EGJ lesions may comparing with both laparoscopic and endoscopic pro- be difficult to completely resect without excessive remov- cedures. Hajer et al. [39] stood out from the other 4 stud- al of surrounding tissues, increasing the conversion rate. ies [32, 33, 37, 44], which compared adverse events be- Additionally, it also requires hand-suturing, which con- tween LECS and LWR/laparoscopy, having a marked de- tributes to longer duration of the procedure [51]. viation to the direction that favors LWR/laparoscopy. It is important to mention that it is a two-center study in Some other characteristics were approached. LWR which the two compared subgroups were from different was associated with higher operative costs than LECS and hospitals and from different countries. Given that, we its modified procedures [37].CLEAN-NET was associat- cannot assure that both groups have the same character- ed with better postoperative quality of life than distal gas- istics [39]. trectomy [47]. Regarding duodenal lesions, ESD was associated with In another meta-analysis on this topic, Cai et al. [52] nonsignificantly higher rates of incomplete local resec- compared LECS with ESD and included Ojima et al. [36] tion and adverse events, while LECS was associated with in their meta-analysis. The procedure used by Ojima et al. nonsignificantly longer procedure duration and hospital [36] in the ESD arm (EIGS) did not fit in our definition stay. of ESD, as it required the opening of the abdominal and gastric walls in order to deliver the endoscope and surgi- Hospitalization time was significantly shorter in gas- cal instruments. Despite these differences, Cai et al. [52] tric lesions compared to EGJ lesions. There were no sta- reported higher incidence of complications and lower tistically significant differences in conversion rate, mean procedure time in ESD. Our results tended to the same procedural time, or adverse events’ rate according to the conclusions. location. However, there was a trend to higher conversion rate and longer procedure duration in EGJ lesions and This is the first systematic review and meta-analysis higher rate of adverse events in duodenal lesions. comparing LECS with laparoscopic and endoscopic tech- niques. We evaluated the safety and efficacy of LECS on The conversion rate tended to be higher in the EGJ three different locations: EGJ, stomach, and duodenum. subgroup because Ri et al. [48] reported 8 cases who need- ed conversion due to esophageal invasion and large de- This systematic review and meta-analysis has some fects after lesion resection (more than half of the circum- limitations. First, all studies were observational and ret- ference of the EGJ). This might explain the higher need rospective reports (mainly single-center), as no random- for conversion presented by this study. ized controlled trials exist on this subject, making them prone to selection bias. Although median quality of exist- Two studies that reported the need for conversion in ing studies is good (6), the risk of bias is not null. Second, gastric lesions [14, 34] used 2 modified procedures: high heterogeneity was found in some outcomes, proba- NEWS and closed-LECS, which include transoral remov- bly due to large variations among techniques. Despite be- al of the lesion. Several lesions with more than 30 mm ing inspired by the same technique, modified LECS pro- were reported, which require conversion to achieve ade- cedures have some differences. Moreover, our sample size quate specimen retrieval. Therefore, this may partially ex- was relatively small: we analyzed 1,337 lesions in this sys- plain the higher rates of conversion in these studies. tematic review and most studies had a small number of cases, which can decrease their precision. Lastly, the R0 was nonsignificantly lower in duodenal procedures number of relevant studies comparing LECS on EGJ and (95.7% vs. 100%). This can be explained by several techni- other locations as well as comparing duodenal LECS with cal difficulties such as maintaining an adequate vision ESD was insufficient to perform meta-analysis. Efficacy and Safety of Laparoscopic GE Port J Gastroenterol 2023;30:4–19 17 Endoscopic Cooperative Surgery DOI: 10.1159/000526644

Conclusion Funding Sources LECS can be a valid, safe, and effective treatment op- This article did not require funding. tion in patients with EGJ, gastric, and duodenal lesions. However, we consider that treatment choice must be in- Author Contributions dividualized, taking in account the experience of the cen- ter and the clinical expertise of the medical team involved. Sara Oliveira de Brito and Diogo Libânio conceived and de- Prospective studies are needed to confirm if LECS is su- signed the study. Sara Oliveira de Brito and João Pedro Pinho Os- perior to other established techniques. ório de Araújo Teixeira undertook the bibliographic research and, independently, screened the articles. Sara Brito, Diogo Libânio, Statement of Ethics and João Pedro Pinho Osório de Araújo Teixeira selected for eli- gibility, determined inclusion, and reviewed the articles included. An ethics statement is not applicable because this study is based Sara Oliveira de Brito drafted the manuscript. Diogo Libânio, exclusively on the published literature. Cláudia Martins Marques Pinto, and João Paulo Meireles de Araú- jo Teixeira revised the draft and approved the final manuscript. Conflict of Interest Statement Data Availability Statement The authors have no conflicts of interest to declare. All data generated and analyzed during this study are included in this article. Further enquiries can be directed to the correspond- ing author. References 1 Hiki N, Yamamoto Y, Fukunaga T, Yamagu- 9 Kim HH, Uedo N. Hybrid NOTES: combined 15 Fujishiro M, Furukawa K, Yamamura T, Na- chi T, Nunobe S, Tokunaga M, et al. Laparo- laparo-endoscopic full-thickness resection kamura M, Honda T, Maeda O, et al. Nonex- scopic and endoscopic cooperative surgery techniques. Gastrointest Endosc Clin N Am. posed wall-inversion surgery as a novel local for gastrointestinal stromal tumor dissection. 2016 Apr;26(2):335–73. resection method for neoplasms in the gastro- Surg Endosc. 2008 Jul;22(7):1729–35. intestinal tract. Nagoya J Med Sci. 2020 May; 10 Nunobe S, Hiki N, Gotoda T, Murao T, Har- 82(2):175–82. 2 Namikawa T, Hanazaki K. Laparoscopic en- uma K, Matsumoto H, et al. Successful appli- doscopic cooperative surgery as a minimally cation of laparoscopic and endoscopic coop- 16 Wong VWY, Goto O, Gregersen H, Chiu PWY. invasive treatment for gastric submucosal tu- erative surgery (LECS) for a lateral-spreading Endoscopic treatment of subepithelial lesions mor. World J Gastrointest Endosc. 2015 Oct mucosal gastric cancer. Gastric Cancer. 2012 of the gastrointestinal tract. Curr Treat Options 10;7(14):1150–6. Jul;15(3):338–42. Gastroenterol. 2017 Dec;15(4):603–17. 3 Matsuda T, Nunobe S, Ohashi M, Hiki N. 11 Inoue H, Ikeda H, Hosoya T, Yoshida A, 17 Tsujimoto H, Yaguchi Y, Kumano I, Takaha- Laparoscopic endoscopic cooperative surgery Onimaru M, Suzuki M, et al. Endoscopic mu- ta R, Ono S, Hase K. Successful gastric submu- LECS for the upper gastrointestinal tract. cosal resection, endoscopic submucosal dis- cosal tumor resection using laparoscopic and Transl Gastroenterol Hepatol. 2017;2:40. section, and beyond: full-layer resection for endoscopic cooperative surgery. World J gastric cancer with nonexposure technique Surg. 2012 Feb;36(2):327–30. 4 Hiki N, Nunobe S. Laparoscopic endoscopic (CLEAN-NET). Surg Oncol Clin N Am. 2012 cooperative surgery (LECS) for the gastroin- Jan;21(1):129–40. 18 Qiu WQ, Zhuang J, Wang M, Liu H, Shen ZY, testinal tract: updated indications. Ann Gas- Xue HB, et al. Minimally invasive treatment troenterol Surg. 2019 May;3(3):239–46. 12 Goto O, Mitsui T, Fujishiro M, Wada I, Shi- of laparoscopic and endoscopic cooperative mizu N, Seto Y, et al. New method of endo- surgery for patients with gastric gastrointesti- 5 Hiki N, Nunobe S, Matsuda T, Hirasawa T, scopic full-thickness resection: a pilot study of nal stromal tumors. J Dig Dis. 2013 Sep;14(9): Yamamoto Y, Yamaguchi T. Laparoscopic non-exposed endoscopic wall-inversion sur- 469–73. endoscopic cooperative surgery. Dig Endosc. gery in an ex vivo porcine model. Gastric Can- 2015 Jan;27(2):197–204. cer. 2011 Jun;14(2):183–7. 19 Hoteya S, Haruta S, Shinohara H, Yamada A, Furuhata T, Yamashita S, et al. Feasibility and 6 Kato M, Yahagi N. Advanced endoscopic 13 Goto O, Takeuchi H, Kawakubo H, Matsuda safety of laparoscopic and endoscopic coop- treatment of gastric and duodenal neoplasms: S, Kato F, Sasaki M, et al. Feasibility of non- erative surgery for gastric submucosal tu- beyond standard EMR and ESD. Am J Gastro- exposed endoscopic wall-inversion surgery mors, including esophagogastric junction tu- enterol. 2018;113(10):1423–6. with sentinel node basin dissection as a new mors. Dig Endosc. 2014 Jul;26(4):538–44. surgical method for early gastric cancer: a 7 Goto O, Takeuchi H, Kitagawa Y, Yahagi N. porcine survival study. Gastric Cancer. 2015 20 Moher D, Liberati A, Tetzlaff J, Altman DG; Endoscopic submucosal dissection (esd) and Apr;18(2):440–5. PRISMA Group. Preferred reporting items related techniques as precursors of “new for systematic reviews and meta-analyses: the notes” resection methods for gastric neo- 14 Kikuchi S, Nishizaki M, Kuroda S, Tanabe S, PRISMA statement. PLoS Med. 2009 Jul 21; plasms. Gastrointest Endosc Clin N Am. 2016 Noma K, Kagawa S, et al. Nonexposure lapa- 6(7):e1000097. Apr;26(2):313–22. roscopic and endoscopic cooperative surgery (closed laparoscopic and endoscopic cooper- 21 Ouzzani M, Hammady H, Fedorowicz Z, El- 8 Tsai TC, Meireles OR. Combined surgical and ative surgery) for gastric submucosal tumor. magarmid A. Rayyan -a web and mobile app endoscopic approaches to full-thickness re- Gastric Cancer. 2017 May;20(3):553–7. for systematic reviews. Systematic Rev. 2016; section. Tech Gastrointest EndoscTechniques 5(1):210. Gastrointest Endosc. 2019 Jan;21(1):26–32. 18 GE Port J Gastroenterol 2023;30:4–19 de Brito/Libânio/Pinto/Araújo Teixeira/ Araújo Teixeira DOI: 10.1159/000526644

22 Wells GA, Shea B, O’Connell D, Peterson J, 33 Yin X, Yin Y, Chen H, Shen C, Tang S, Cai Z, 43 Kanaji S, Morita Y, Yamazaki Y, Otowa Y, Welch V, Losos M, et al. The Newcastle-Otta- et al. Comparison analysis of three different Takao T, Tanaka S, et al. Feasibility of laparo- wa Scale (NOS) for assessing the quality of types of minimally invasive procedures for scopic endoscopic cooperative surgery for nonrandomised studies in meta-analyses. gastrointestinal stromal tumors ≤5 cm. J Lap- non-ampullary superficial duodenal neo- Oxford; 2000. Available from: http://www. aroendosc Adv Surg Tech A. 2018 Jan;28(1): plasms: single-arm confirmatory trial. Dig ohri.ca/programs/clinical_epidemiology/ox- 58–64. Endosc. 2020 Mar;33(3):373–80. ford.asp. 34 Mitsui T, Yamashita H, Aikou S, Niimi K, Fu- 44 Kanehira E, Kanehira AK, Tanida T, Taka- 23 Hozo SP, Djulbegovic B, Hozo I. Estimating jishiro M, Seto Y. Non-exposed endoscopic hashi K, Obana Y, Sasaki K. CLEAN-NET: a the mean and variance from the median, wall-inversion surgery for gastrointestinal modified laparoendoscopic wedge resection range, and the size of a sample. BMC Med Res stromal tumor. Transl Gastroenterol Hepatol. of the stomach to minimize the sacrifice of in- Methodol. 2005;5(1):113–0. 2018;3:17. nocent gastric wall. Surg Endosc. 2020 Jan; 34(1):290–7. 24 Higgins JP, Thomas J, Chandler J, Cumpston 35 Ojima T, Nakamori M, Nakamura M, Hayata M, Li T, Page MJ, et al. Cochrane handbook K, Katsuda M, Takifuji K, et al. Laparoscopic 45 Mahawongkajit P, Chanswangphuvana P. for systematic reviews of interventions. John and endoscopic cooperative surgery versus Laparoscopy-assisted endoscopic full-thick- Wiley & Sons; 2019. Available from: www. endoscopic submucosal dissection for the ness resection of upper gastrointestinal sub- training.cochrane.org/handbook. treatment of low-risk tumors of the duode- epithelial tumors: a single-center early experi- num. J Gastrointest Surg. 2018 May;22(5): ence. Mol Clin Oncol. 2020 May;12(5):461–7. 25 Manager R. VReview Manager Version 5.4, 935–40. The Cochrane Collaboration. Denmark: The 46 Nunobe S, Ri M, Yamazaki K, Uraoka M, Nordic Cochrane Centre Copenhagen; 2020. 36 Ojima T, Nakamura M, Nakamori M, Taki- Ohata K, Kitazono I, et al. Safety and feasibil- Available from: https://training.cochrane. fuji K, Hayata K, Katsuda M, et al. Laparo- ity of laparoscopic and endoscopic coopera- org/online-learning/core-software-co- scopic and endoscopic cooperative surgery is tive surgery for duodenal neoplasm: a retro- chrane-reviews/revman. a feasible treatment procedure for intralumi- spective multicenter study. Endoscopy. 2021 nal gastric gastrointestinal stromal tumors Oct;53(10):1065–8. 26 Viechtbauer W. Conducting meta-analyses in compared to endoscopic intragastric surgery. R with the metafor package. J Stat Stat Softw. Surg Endosc. 2018 Jan;32(1):351–7. 47 Okubo K, Arigami T, Matsushita D, Sasaki K, 2010;36(3):1–48. Kijima T, Noda M, et al. Evaluation of post- 37 Shoji Y, Takeuchi H, Goto O, Tokizawa K, operative quality of life by PGSAS-45 follow- 27 Wallace BC, Dahabreh IJ, Trikalinos TA, Lau Nakamura R, Takahashi T, et al. Optimal ing local gastrectomy based on the sentinel J, Trow P, Schmid CH. Closing the gap be- minimally invasive surgical procedure for lymph node concept in early gastric cancer. tween methodologists and end-users: R as a gastric submucosal tumors. Gastric Cancer. Gastric Cancer. 2020 Jul;23(4):746–53. computational back-end. J Stat Softw. 2012; 2018 May;21(3):508–15. 49(5):1–15. 48 Ri M, Nunobe S, Makuuchi R, Ida S, Kumagai 38 Cao L, Zheng KM, Wang HL, Zhao YJ, Yang K, Ohashi M, et al. Is laparoscopic and endo- 28 Kang WM, Yu JC, Ma ZQ, Zhao ZR, Meng ZD, Li W. Laparoscopic and endoscopic co- scopic cooperative surgery (LECS) for gastric QB, Ye X. Laparoscopic-endoscopic coopera- operative dissection for small gastric gastro- subepithelial tumor at the esophagogastric tive surgery for gastric submucosal tumors. intestinal stromal tumor without causing in- junction safe? Asian J Endosc Surg. 2020 Apr; World J Gastroenterol. 2013 Sep 14;19(34): jury to the mucosa. Gastroenterol Res Pract. 14(2):223–31. 5720–6. 2019 Dec;2019:7376903. 49 Matsumoto S, Yoshida Y. Selection of appro- 29 Waseda Y, Doyama H, Inaki N, Nakanishi H, 39 Hajer J, Havlůj L, Kocián P, Klimbacher G, priate endoscopic therapies for duodenal tu- Yoshida N, Tsuji S, et al. Does laparoscopic Shamiyeh A, Gürlich R, et al. Comparison of mors: an open-label study, single-center ex- and endoscopic cooperative surgery for gas- nonexposed endoscopic wall-inversion sur- perience. World J Gastroenterol. 2014 Jul 14; tric submucosal tumors preser ve residual gery with endoscopic-navigated laparoscopic 20(26):8624–30. gastric motility? Results of a retrospective sin- wedge resection for gastric submucosal tu- gle-center study. PLoS One. 2014; 9(6): mours: results of a two-centre study. Gastro- 50 Marques J, Baldaque-Silva F, Pereira P, Ar- e101337. enterol Res Pract. 2019;2019:7573031. nelo U, Yahagi N, Macedo G. Endoscopic mu- cosal resection and endoscopic submucosal 30 Balde AI, Chen T, Hu Y, Redondo N JD, Liu 40 Yanagimoto Y, Omori T, Jeong-Ho M, dissection in the treatment of sporadic non- H, Gong W, et al. Safety analysis of laparo- Shinno N, Yamamoto K, Takeuchi Y, et al. ampullary duodenal adenomatous polyps. scopic endoscopic cooperative surgery versus Feasibility and safety of a novel laparoscopic World J Gastrointest Endosc. 2015 Jun 25; endoscopic submucosal dissection for select- and endoscopic cooperative surgery tech- 7(7):720–7. ed gastric gastrointestinal stromal tumors: a nique for superficial duodenal tumor resec- propensity score-matched study. Surg En- tion: how i do it. J Gastrointest Surg. 2019 Oct; 51 Matsuda T, Hiki N, Nunobe S, Aikou S, Hira- dosc. 2016 Feb;31(2):843–51. 23(10):2068–74. sawa T, Yamamoto Y, et al. Feasibility of lap- aroscopic and endoscopic cooperative sur- 31 Ichikawa D, Komatsu S, Dohi O, Naito Y, Ko- 41 Aoyama J, Goto O, Kawakubo H, Mayanagi S, gery for gastric submucosal tumors (with vid- suga T, Kamada K, et al. Laparoscopic and en- Fukuda K, Irino T, et al. Clinical outcomes of eo). Gastrointest Endosc. 2016 Jul;84(1): doscopic co-operative surgery for non-am- non-exposed endoscopic wall-inversion sur- 47–52. pullary duodenal tumors. World J Gastroen- gery for gastric submucosal tumors: long- terol. 2016 Dec 21;22(47):10424–31. term follow-up and functional results. Gastric 52 Cai XL, Li XY, Liang C, Xu Y, Zhang MZ, Yu Cancer. 2020 Jan;23(1):154–9. WM, et al. Endoscopic or laparoscopic resec- 32 Komatsu S, Ichikawa D, Kosuga T, Okamoto tion for small gastrointestinal stromal tu- K, Konishi H, Shiozaki A, et al. Clinical im- 42 Aoyama J, Kawakubo H, Matsuda S, Ma- mors: a cumulative meta-analysis. Chin Med pact of laparoscopy and endoscopy coopera- yanagi S, Fukuda K, Irino T, et al. Clinical out- J. 2020 Nov 20;133(22):2731–42. tive surgery (LECS) on gastric submucosal tu- comes of laparoscopic and endoscopic coop- mor after its standardization. Anticancer Res. erative surgery for submucosal tumors on the 2016 Jun;36(6):3041–7. esophagogastric junction: a retrospective sin- gle-center analysis. Gastric Cancer. 2020 Nov; 23(6):1084–90. Efficacy and Safety of Laparoscopic GE Port J Gastroenterol 2023;30:4–19 19 Endoscopic Cooperative Surgery DOI: 10.1159/000526644

Research Article Received: May 25, 2021 Accepted: October 14, 2021 GE Port J Gastroenterol 2023;30:20–28 Published online: November 25, 2021 DOI: 10.1159/000520272 Positioning Aeromonas Infection in Inflammatory Bowel Disease: A Retrospective Analysis Tiago Pereira Guedesa Joana Alves Silvaa Sara Nevesb Daniela Falcãoa Paula Costac Paula Lagoa Isabel Pedrotoa Marta Salgadoa aGastroenterology Department, Centro Hospitalar Universitário do Porto, Porto, Portugal; bInfectious Diseases Department, Centro Hospitalar Universitário do Porto, Porto, Portugal; cMicrobiology Department, Centro Hospitalar Universitário do Porto, Porto, Portugal Keywords ila/caviae. Bacterial co-isolation was found in 4 non-IBD and Aeromonas · Inflammatory bowel disease · Crohn’s disease · Ulcerative colitis · GI infections · Immunology · histological findings of cytomegalovirus in 2 IBD patients. Microbiology and inflammatory bowel disease Non-IBD patients presented more frequently with fever and Abstract Background and Aim: Aeromonas are Gram-negative rods IBD patients with bloody diarrhea and abdominal pain. There known to cause a spectrum of diseases. Inflammatory bowel disease (IBD) is an idiopathic complex condition resulting was higher tendency for severe infection rate in IBD patients from interaction of multiple factors. Aeromonas infection in association with IBD is still largely unknown. We aim to look with higher antimicrobial therapy use. Steroids were exclu- for the significance of Aeromonas infection and for signifi- cant differences between IBD and non-IBD patients. Meth- sively used in the IBD group. From IBD, 4 patients had the ods: A retrospective observational analysis was performed of all patients positive for Aeromonas in stool cultures, dur- diagnosis of ulcerative colitis and 9 of Crohn’s disease with ing a 10-year period, from a tertiary and university hospital. Results: Fifty patients were included, 56% male with a mean colonic involvement. Of these patients, 5 received IBD diag- age of 42.1 years. Thirty-eight (76%) were non-IBD and 12 (24%) IBD patients. IBD patients were more frequently under nosis after the acute episode of Aeromonas infection. Con- immunosuppressors. Two patients were asymptomatic and 44% developed mild, 44% moderate, and 16.7% severe in- clusions: Clinical presentation of Aeromonas infection differs fection. The main strains isolated were Aeromonas hydroph- between IBD and non-IBD patients. Non-IBD patients had milder severity of infection with less use of antibiotics. Aeromonas infection seems to greatly contribute to IBD manifestation. © 2021 Sociedade Portuguesa de Gastrenterologia Published by S. Karger AG, Basel Tiago Pereira Guedes and Joana Alves Silva contributed equally to this work (co-first authors). [email protected] © 2021 Sociedade Portuguesa de Gastrenterologia Correspondence to: www.karger.com/pjg Published by S. Karger AG, Basel Tiago Pereira Guedes, [email protected] is is an Open Access article licensed under the Creative Commons Attribution-NonCommercial-4.0 International License (CC BY-NC) (http://www.karger.com/Services/OpenAccessLicense), applicable to the online version of the article only. Usage and distribution for com- mercial purposes requires written permission.

Infeção por Aeromonas e doença inflamatória Introduction intestinal: uma análise retrospetiva The Aeromonas genus, belonging to the Aeromonada- Palavras Chave ceae family, comprises facultatively anaerobic Gram-neg- Aeromonas · Doença inflamatória intestinal · Doença de ative bacteria widely present in aquatic environments, Crohn · Colite ulcerosa · Infeção gastrointestinal · soil, and food goods (such as meat, shellfish, and dairy Imunologia · Microbiologia e doença inflamatória products), making the gastrointestinal tract an under- intestinal standable reaching point for Aeromonas [1, 2]. However, there has been discordant information concerning its role Resumo as a gastrointestinal pathogen, since its isolation in stool Introdução: A etiologia da Doença Inflamatória Intestinal from asymptomatic individuals is not uncommon [3]. Of (DII) é complexa e resultante da interação de diversos fa- the 17 phenospecies of the genus Aeromonas, the most tores, nomeadamente microbiológicos. A infeção por frequently isolated from human feces are A. hydrophila, Aeromonas caracteriza-se por um espectro alargado de A. caviae, A. veronii, and A. trota [4–6]. The most com- manifestações clínicas. O papel da infeção por Aeromon- mon manifestation of Aeromonas infection is diarrhea, as na DII não está caracterizado. Objetivos: Avaliar o sig- usually acute and self-limited, but other presentations nificado da infeção por Aeromonas na DII e as diferenças such as bloody diarrhea and abdominal pain, or chronic com a infeção em doentes não-DII. Métodos: avaliação more indolent diarrhea, have been described [7]. Al- retrospetiva e observacional de todos os doentes com though most of the infections occur in immunocompro- isolamento microbiológico de Aeromonas em amostras mised patients, it can also cause disease in healthy indi- fecais num período de 10 anos, num hospital terciário. Re- viduals [8]. sultados: foram avaliados 50 doentes, 56% do sexo mas- culino, com idade média de 42.1 anos. Doze (24%) com Although the ultimate cause of inflammatory bowel diagnóstico de DII e trinta e oito (76%) não-DII. Os doentes disease (IBD) is still unknown, there are many compo- com DII encontravam-se mais frequentemente sob imu- nents interacting in its network of pathogenic mecha- nossupressão. Dois doentes foram assintomáticos, 44% nisms like environmental factors, genetic susceptibility, desenvolveram doença ligeira, 44% moderada e 16.7% dysregulated immune response, and microbiological fac- severa, havendo maior tendência para infeção severa nos tors [9, 10]. Regarding the latter, both alterations in intes- DII. Os doentes não-DII apresentaram mais frequente- tinal microbiota and infections by external agents might mente febre e os DII diarreia sanguinolenta e dor abdom- play a role in IBD onset and its flares [11, 12]. Dysbiosis inal. O uso de antimicrobianos foi superior no grupo DII e can also result from commensal flora that, although nor- a utilização de corticoesteroides foi exclusiva nestes mal in speciation, possess more subtle virulence factors doentes. Isolamento concomitante de outros agentes mi- such as enteroadherence, or the lack of diversity of the crobiológicos ocorreu em 4 doentes não-DII e 2 com DII fecal microbiome [13, 14]. tinham histologia compatível com infeção por Citomega- lovírus. Da população DII, 4 eram Colite Ulcerosa e 9 Doen- Considering the specific bacteria related to IBD onset, ça de Crohn com envolvimento cólico. Destes, 5 rece- Campylobacter is probably the one with more data, while beram o diagnóstico após a infeção por Aeromonas. Con- the heterogeneity of the studies does not allow to draw clusão: A apresentação clínica da infeção por Aeromonas consistent conclusions [15]. Although data are mainly foi distinta entre as populações DII e não-DII, sendo que based on case report studies, some authors have suggest- os doentes DII apresentaram doença mais severa e maior ed that Aeromonas can be a trigger of flares in IBD [16– utilização de antimicrobianos. A infeção na DII ocorreu es- 18] and as well as a trigger to develop de novo chronic sencialmente em doentes com envolvimento cólico. colitis in patients with no previous history of IBD [7, 19]. To our knowledge, only 2 recent studies involving IBD © 2021 Sociedade Portuguesa de Gastrenterologia patients have tried to correlate the diagnosis and severity Published by S. Karger AG, Basel of the disease with Aeromonas infection [7, 8]. Our aim was to further look for the significance of Aeromonas infection as well as for significant differences between IBD and non-IBD patients. Positioning Aeromonas Infection and IBD GE Port J Gastroenterol 2023;30:20–28 21 DOI: 10.1159/000520272

Materials and Methods Results Patients and Clinical Information Baseline Characteristics A retrospective analysis of data was performed of inpatients Stool isolation of Aeromonas was identified in 53 pa- and outpatients with a positive stool culture for Aeromonas, be- tients during the 10-year period of the study. Fifty pa- tween January 2009 to January 2019, in a tertiary and university tients were included and analyzed. Three were excluded hospital. Patients of all ages were included. Clinical data were ob- due to the absence of minimal clinical information. tained retrospectively from the electronic medical records. Our population consisted mainly of individuals of Baseline data included age at Aeromonas fecal detection, sex, male sex (54%) with a mean age of 42.1 years (1–89 years comorbidities (namely the presence of IBD and transplantation old). Fourteen (28%) patients were under 18 years old. status), ongoing drugs (namely immunosuppressants), recent Baseline characteristics are presented in Table 1. travels, and previous abdominal surgery. Clinical characteristics at Twelve patients (24%) had the diagnosis of IBD, with the time of Aeromonas detection were also analyzed: the presence 7 (14%) having it diagnosed previously and 5 (10%) after of symptoms attributed to gastrointestinal infection such as nau- Aeromonas fecal isolation. Five patients had ongoing che- sea, vomiting, diarrhea, number of bowel movements, abdominal motherapy treatment during Aeromonas isolation and 2 pain, and fever; endoscopy findings, if performed; presence of his- had advanced liver disease. Thirteen patients (26%) were topathological features of cytomegalovirus (CMV) infection in co- immunosuppressed considering therapeutics as calci- lon biopsies; other bacterial co-identification in stool cultures; neurin inhibitors, AZA, systemic steroids, and biological treatment applied (namely antibiotics and steroids); disease sever- therapy as infliximab. Six patients (12%) were previously ity and/or death related to the infection. submitted to organ transplantation. At the time of The indication for treatment was done according to the physi- Aeromonas isolation, 4 patients were under azathioprine cian’s criteria. Disease severity was defined as: (a) mild self-limited (AZA), 6 under topical or systemic steroids, and 6 under gastroenteritis, as a gastrointestinal infection that resolved sponta- calcineurin inhibitors. Only 1 IBD patient was under neously without the need for antibiotics; (b) moderate gastroen- combined therapy with infliximab and AZA. Only 1 non- teritis, as a gastrointestinal infection that required antibiotic treat- IBD patient was under AZA for autoimmune hepatitis. ment; and (c) severe gastroenteritis, defined by a gastrointestinal Also, the only IBD patient under calcineurin inhibitors infection complicated with septicemia and/or renal impairment. was a liver transplant recipient. Death related to the infection was defined as the occurrence of No patient had recent travel history. Twelve patients death over the duration of infection and/or antibiotic therapy. (24%) presented a history of previous abdominal surgery, Stool samples were processed for bacterial culture using GN with 3 having a previous hemicolectomy, 1 a cholecystec- broth and selective media: Macconkey agar, Macconkey agar with tomy, 3 liver transplants, 1 reno-pancreatic transplant, sorbitol, Salmonella-Shigella agar, CIN (Cefsulodin-Irgasan-No- and 2 gastric bypasses. vobiocin) agar, and Campylosel agar. In certain patients, a blood agar with an ampicillin disc was also used. The culture media were Clinical Presentation in IBD and Non-IBD Patients incubated at 35  ° C and observed daily for 48 h, except for the Cam- The clinical and microbiological characteristics at the pylosel media. Suspected colonies of Aeromonas spp. in CIN media time of Aeromonas isolation are shown in Table 2. From and those that grew close to the ampicillin disc and, simultane- the overall sample, 48/50 (96%) patients were found to ously, showed a positive oxidase reaction, were identified. The have symptoms that could be attributed to the presence identification was made in automated systems: phenotypical of Aeromonas. Only 1 IBD and 1 non-IBD patient were methods (Vitek2® – bioMérieux) or by mass spectrometry tech- considered asymptomatic: concerning the IBD patient, nique (MALDI-TOF MS® – bioMérieux). the stool culture was required in the course of the diagno- sis of IBD, as part of a protocol required at the first con- IBD and Non-IBD Patients sultation. This patient was sent to IBD consultation due Data from all patients was analyzed, and two groups (IBD and to chronic diarrhea and abdominal pain in the past year non-IBD patients) were created for comparison. The IBD group but was asymptomatic at the time of consultation. Con- included all patients with the diagnosis of IBD previously and after cerning the non-IBD patient, the stool culture was re- Aeromonas isolation. Individuals for whom both ulcerative colitis quired in the course of the postrenal transplantation pro- (UC) and Crohn’s disease (CD) were recorded on successive dates tocol. were categorized according to the latest. The most frequent strains isolated in both groups were A. hydrophila/caviae (distinction not possible with the Statistical Analysis Descriptive analysis was performed considering absolute and rel- ative frequencies for categorical variables or mean and standard de- viation for quantitative variables. Association between qualitative variables was performed based on the χ2 test (with Yates’ correction) or Fisher’ exact test (when assumptions could not be verified in 2 × 2 tables). Group comparison according to quantitative variables was performed using the independent-sample t test or Mann-Whitney test (variable with relevant skewness). In all analyses, a significance level of 0.05 was considered and SPSS version 26 was used. 22 GE Port J Gastroenterol 2023;30:20–28 Pereira Guedes/Alves Silva/Neves/Falcão/ Costa/Lago/Pedroto/Salgado DOI: 10.1159/000520272

Table 1. Baseline characteristics Total IBD Non-IBD p value n (%) or mean (SD) n (%) or mean (SD) n (%) or mean (SD) Total 50 (100.0) 12 (24.0) 38 (76.0) – Age, years 42.1 (28.6) 40.2 (23.5) 42.7 (30.3) 0.794c Female sex 23 (46.0) 4 (33.3) 19 (50.0) 0.498a IBD diagnosis 7 (14.0) 7 (58.3) – – Previous to Aeromonas isolation 5 (10.0) 5 (41.7) – – Posterior to Aeromonas isolation 11 (28.9) 1.000b Comorbidities at the moment of Aeromonas fecal isolation 6 (15.8) 1.000b 3 (7.9) 1.000b Arterial hypertension 14 (28.0) 3 (25.0) 3 (7.9) 1.000b 1 (8.3) 2 (5.3) 1.000b Dyslipidemia 7 (14.0) 1 (8.3) 5 (13.2) 0.319b 0 (0.0) 5 (13.2) 1.000b Diabetes 4 (8.0) 0 (0.0) 0 (0.0) 13 (34.2) 0.735b Chronic renal failure 3 (6.0) 1 (8.3) 5 (13.2) 0.319b 6 (18.8) 0.073b Cirrhosis 2 (4.0) 1 (2.6) 0.038b Cancer 5 (10.0) 2 (5.3) 0.024b 5 (13.2) 0.294b Transplant 6 (12.0) 0 (0.0) 0.240b 11 (28.9) 0.248b Ongoing drugs at the moment of Aeromonas fecal isolation PPI 18 (36.0) 5 (41.7) Chemotherapy 5 (10.0) 0 (0.0) Immunosuppressed patients 13 (26.0) 7 (58.3) Immunosuppressant drugs 4 (8.0) 3 (25.0) AZA 6 (12.0) 4 (33.3) Topical or systemic steroids 6 (12.0) 1 (8.3) Calcineurin inhibitors 1 (2.0) 1 (8.3) Biologic therapy (anti-TNF) 12 (24.0) 1 (8.3) Previous abdominal surgery AZA, azathioprine; IBD, Inflammatory bowel disease; PPI, proton pump inhibitors; SD, standard deviation; a χ2 test; b Fisher Test; c Independent sample Student’s t test. routinely used methods). Symptoms presented at the IBD) needed to be admitted to the hospital for manage- time of Aeromonas isolation differ among both groups, ment of severe infection and symptomatic control in with IBD patients presenting more frequently with bloody moderate cases. diarrhea (83.3 vs. 10.5%, p < 0.001) and abdominal pain (75.0 vs. 34.2%, p = 0.032). The presence of fever was Concerning the treatment of the acute episode, IBD more common among patients without IBD (0.0 vs. patients were more frequently submitted to antibiotics 34.2%, p = 0.022). IBD patients were more frequently un- (83.3 vs. 36.8%, p = 0.013), and quinolones were the der AZA (25.0 vs. 2.6%, p = 0.038) and systemic or topical most frequently used class in both (36% overall). Only steroids (33.3 vs. 5.3%, p = 0.024) when compared with IBD patients were treated with systemic or topical ste- non-IBD patients. Immunosuppressive therapy with cal- roids (33.3 vs. 0.0%, p = 0.002), and all of these patients cineurin inhibitors was more often used in non-IBD pa- were also under antibiotics. Four (8%) Aeromonas-re- tients (8.3 vs. 13.2%, p = 0.294). sistant strains were identified (3 [7.9%] in non-IBD vs. 1 [8.3%] in IBD). Other fecal bacteria were identified The majority of patients presented a mild-moderate only in 4 non-IBD patients (10.5%): 3 children, two Aeromonas infection with a proportionally higher ten- 1-year-olds and one 8-year-old, presented Campylo- dency for severe episodes in IBD patients (27.3 vs. bacter jejuni, and a 73-year-old patient presented also 13.5%, p > 0.05). There were 8 (16.7%) severe cases, 7 of with Salmonella. All patients progressed favorably with which concerned immunocompromised patients. no deaths observed. Nineteen patients (41.7% were IBD and 36.8% non- Positioning Aeromonas Infection and IBD GE Port J Gastroenterol 2023;30:20–28 23 DOI: 10.1159/000520272

Table 2. Clinical presentation features in IBD and non-IBD patients Total IBD Non-IBD p value n (%) or median (SD) or n (%) or median (SD) or n (%) or median (SD) median (IQR) median (IQR) or median (IQR) Aeromonas strain 32 (64.0) 9 (75.0) 23 (60.5) – Hydrophila/caviae 12 (24.0) 1 (8.3) 11 (28.9) – Hydrophila 3 (6.0) 1 (8.3) 2 (5.3) Aeromonas spp. 3 (6.0) 1 (8.3) 2 (5.3) 0.426b Aeromonas veronii 48 (96.0) 11 (91.7) 37 (97.4) 0.240b 46 (92.0) 10 (83.3) 36 (94.7) <0.001b Symptoms 14 (8.0) 10 (83.3) 4 (10.5) 0.032a Diarrhea 22 (44.0) 9 (75.0) 13 (34.2) 0.717a Bloody diarrhea 21 (42.0) 4 (33.3) 17 (44.7) 0.022b Abdominal pain 13 (26.0) 0 (0.0) 13 (34.2) – Nausea and vomiting Fever 18 (36.0) 2 (16.7) 16 (42.1) 0.582a 24 (48.0) 9 (75.0) 15 (39.5) Bowel movements/day 4 (8.0) 0 (0.0) 4 (10.5) 1.000a <5 5–10 21 (44.0) 4 (36.4) 17 (45.9) 0.013a >10 21 (44.0) 5 (45.5) 16 (43.2) – 8 (16.7) 3 (27.3) 5 (13.5) Severity 19 (38.0) 5 (41.7) 14 (36.8) 0.002b Mild – Moderate 24 (48.0) 9 (75.0) 14 (36.8) Severe 18 (36.0) 9 (75.0) 9 (23.7) – Hospital admission 3 (6.0) 1 (8.3) 2 (5.3) – 1 (2.0) 0 (0.0) 1 (2.6) Treatment 1 (2.0) 0 (0.0) 1 (2.6) Antibiotics 4 (8.0) 4 (33.3) 0 (0.0) Quinolone Carbapenem plus quinolone 18 (36.0) 1 (8.3) 17 (44.7) Carbapenem plus vancomycin 6 (12.0) 5 (41.7) 1 (2.6) Amoxicillin and clavulanic acid 2 (4.0) 0 (0.0) 2 (5.3) Systemic or topical steroids 4 (8.0) 1 (8.3) 3 (7.9) 0 (0.0) – – Other treatments No treatment Mesalazine Probiotic Resistant strains identified Dead IBD, inflammatory bowel disease; IQR, interquartile range; SD, standard deviation; a χ2 test; b Fisher test; c Independent sample Student’s t test. IBD Diagnosis and Aeromonas Infection infliximab and AZA (patient 4, Table 3), and the remain- Clinical characteristics of IBD patients and Aeromonas ing were under AZA monotherapy at the time of Aeromo- infection are presented in Table 3. From the IBD patient nas isolation. UC patients were under mesalamine. One subgroup, 7 had the diagnosis of IBD previously and 5 patient (patient 9, Table 3) was under tacrolimus due to a after the Aeromonas isolation. From the latter, only 1 had previous liver transplant. Five patients were treated with the infection and subsequent diagnosis of IBD at pediatric antibiotics, and 2 of them simultaneously with systemic age (patient 8, Table 3). steroids. One patient was treated only with topical ste- From those with a previously established diagnosis of roids. Three patients needed hospital admission, with one IBD, 3 were UC patients and 4 CD patients (Table 3). requiring ganciclovir treatment due to concomitant his- Only 1 patient with CD was under combined therapy with tological features of CMV on colonic biopsy (patient 5, 24 GE Port J Gastroenterol 2023;30:20–28 Pereira Guedes/Alves Silva/Neves/Falcão/ Costa/Lago/Pedroto/Salgado DOI: 10.1159/000520272

Positioning Aeromonas Infection and IBD Table 3. Aeromonas isolation in IBD patients Patient Age, CD, Ongoing IBD De novo Main symptoms 1 sex (MC)/UC treatment IBD 24, F CD AZA No Bloody diarrhea (A2L1B2) 2 28, F AZA No Diarrhea, abdominal CD Topical steroids Yes pain 3 36, F (A2L1B1) None No Diarrhea, abdominal 4 76, M CD Infliximab pain (A2L2B1) AZA Abdominal pain, CD, nausea, vomiting (A3L1+L4 B2) 5 16, M CD AZA No Bloody diarrhea, 6 64, F (A1L2B1) abdominal pain 7 31, M CD None Yes Bloody diarrhea, 8 15, M (A3L2B) weight loss 9 27, M CD None Yes Bloody diarrhea 10 62, M (A2L2B1) 11 12, F CD Mesalamine, Yes Bloody diarrhea, A1L2B1 abdominal pain GE Port J Gastroenterol 2023;30:20–28 UC systemic steroids Diarrhea, abdominal DOI: 10.1159/000520272 pain, weight loss UC Mesalamine No Bloody diarrhea UC Mesalamine one Yes Bloody diarrhea month previous No Mesalamine 12 62, M UC Mesalamine No – CD, Crohn disease; F, female; IBD, inflammatory bowel disease; UC, ulcerative col 25

Endoscopic/imaging features Treatment Hospitalization Aeromonas strain – Topical steroids No Aeromonas hy- No drophila/caviae l– Ciprofloxacin No Yes Aeromonas hy- l Colonic mucosa with conges- Mesalamine drophila/caviae tion and aphtoid erosions Aeromonas hy- Ileal and jejunal segmental Imipenem, systemic ste- drophila stenosis with transmural hy- roids perenhancement and small Aeromonas hy- loop dilation drophila/caviae Friable and congestive mucosa Ciprofloxacin, systemic Yes Aeromonas hy- with erosions steroids, ganciclovir Yes drophila/caviae No Aeromonas spp Colonic mucosa with deep Ciprofloxacin, systemic Yes ulcers on colon steroids, ganciclovir Yes Aeromonas hy- No drophila/caviae Decreased vascular pattern, Mesalamine No Aeromonas hy- erosions No drophila/caviae Aeromonas sobria Decreased vascular pattern, Ciprofloxacin, aphtoid erosions mesalamine Aeromonas hy- drophila/caviae l Mayo endoscopic subscore 2 Ciprofloxacin Aeromonas hy- drophila Mayo endoscopic subscore 2 Ciprofloxacin Aeromonas spp – Ciprofloxacin, mesalamine – No additional litis; M, male; MC, Montreal classification.

Table 3). The only patient under infliximab presented The reports of the most predominant clinical species with intestinal sub-occlusion with a quinolone-resistant of Aeromonas have changed over the years. Improved A. hydrophila/caviae isolation. All patients progressed fa- molecular methods led to the conclusion that 95.4% of the vorably, and none needed to escalate IBD therapy after strains associated with the clinical disease were A. caviae the Aeromonas acute episode. Seven patients were sub- (37.3%), A. veronii (23.5%), A. dhakensis (21.5%), and A. mitted to endoscopic evaluation. However, histological hydrophila (13.1%) [21]. In our study, the 3 identified data was compatible with IBD, with no other specific fea- species are concordant with the most associated with clin- tures. Immediately after the acute episode, only clinical ical disease in other studies [7, 8]. Nevertheless, since A. and analytical data were used to assess the remission, dhakensis cannot be identified by the biochemical meth- namely the symptoms reported by the patients, CRP, and ods and mass spectrometry used in our laboratory, we calprotectin. cannot exclude that there might be a misidentification as A. caviae, A. hydrophila, or A. veronii. Although our data Concerning the 5 patients with a diagnosis of IBD after reinforce the predilection of this finding in non-IBD pa- the episode of Aeromonas isolation, 4 of them were diag- tients (96.2%), we have interestingly shown higher isola- nosed with colonic CD and 1 with UC. Two of them were tion of A. hydrophila. This could be explained by the high- admitted to the hospital with 1 showing simultaneous er number of patients under 18 years old and different histological features of CMV infection on colonic biopsy geographical factors. It should, however, be kept in mind (patient 6, Table 3), being treated with antibiotics, sys- that the frequency of each strain differs according to the temic steroids, and ganciclovir. The other 3 received the country. In our study, no relation was found between the IBD diagnosis in an ambulatory setting due to the persis- strains and severity of the clinical manifestations. tence of chronic diarrhea in which Aeromonas was iso- lated in the first stool culture requested for the workup. When compared with a recent similar study performed From this subgroup of patients with de novo IBD diagno- in a tertiary hospital in Spain (n = 98) [8], the mean age sis, 2 required posterior escalation therapy to infliximab of our population was lower (42 vs. 62 years-old) as well (patients 6 and 10, Table 3). as the proportion of female sex (46 vs. 51%). This differ- ence can be explained mainly by the fact that only adult Discussion/Conclusion patients were involved in the Spanish study. On the other hand, we presented a higher percentage of isolates in pa- The pathogenic role of Aeromonas in human entero- tients with previous IBD diagnosis (14 vs. 11%) with very colitis is still controversial as is the association of Aeromo- similar cases of CD (8 vs. 7 patients) and an equal number nas infection with IBD development or as a flare trigger. of UC patients (n = 4). To our knowledge, a few case reports have been pub- lished, with only 2 similar case series available [7, 8]. The main comorbidities found in our group of patients were malignancies (10%), pharmacological immunosup- The global prevalence of Aeromonas gastrointestinal in- pression (34%), and liver diseases (4%). Aeromonas is fection ranges from 2 to 88% and carriage status in healthy known to cause gastrointestinal symptoms in both im- individuals from 1 to 45% [7]. Stool isolation rates differ munocompetent and immunocompromised individuals depending, among others, on geography, food habits, and with predisposing factors reported, such as diabetes, he- isolation methods [20]. The prevalence of infection in de- matological malignancies, or hepatobiliary diseases [20]. veloped countries ranges from 0.8 to 7.4% and carriage rate Overall, the severity of the infection in our population from 0 to 4% [20], indicating a possible higher prevalence was mainly mild-moderate with 17% severe. The percent- in developing countries. Although no specific data exists age of severe infection was greater in IBD patients (27.3 for Portugal, we had a relatively low number of Aeromonas vs. 13.5%), also reported by others [7]. Of the 5 non-IBD stool isolation considering the period covered. The evolu- patients that presented with severe infection, 2 were un- tion of Aeromonas detection methods through the years dergoing chemotherapy, 2 had previous abdominal sur- may have a role in it, since the majority (56%) of our iso- gery, and the last was an HIV patient with no antiretrovi- lates were identified between 2014 and 2019, coinciding ral therapy adherence. The reported mortality rate can with the introduction of mass spectrometry methods. In reach 30% in the set of Aeromonas’ bacteremia [21]. For- this study (n = 50), 2 patients were asymptomatic carriers tunately, there were no reported deaths attributed to the (4%), concordant with the reported carriage rate in healthy infection in our study. individuals in developed countries [20]. The antibiotic susceptibility profile for Aeromonas does not appear to have changed substantially [22]. In the 26 GE Port J Gastroenterol 2023;30:20–28 Pereira Guedes/Alves Silva/Neves/Falcão/ Costa/Lago/Pedroto/Salgado DOI: 10.1159/000520272

present study, only 4 isolates were resistant to antibiotics ulcerative proctitis, who presented a severe colitis refrac- usually active against aeromonads, according to EUCAST tory to steroid therapy with favorable evolution under an- breakpoints [23]. Two patients had a strain resistant to tibiotic therapy directed to the isolated Aeromonas [16]. ciprofloxacin and cotrimoxazole, another patient to co- On the other hand, other reports had postulated a puta- trimoxazole, and the last one to ciprofloxacin. There was tive role of Aeromonas infection for the development of no identified resistance to carbapenems, a fortunate sign de novo IBD diagnosis [7, 12]. It remains under discus- regardless of the reports describing an increased Aeromo- sion if the infection prompts IBD development or natu- nas resistance to the latter [24]. Resistant strains were rally unmasks an underlying disease with previous sub- more common in non-IBD patients (75 vs. 25%). This clinical activity. The mechanisms inherent to these inter- could be explained by the fact that the mean age of non- relationships are unknown, but Aeromonas-associated IBD patients was higher, which usually implies a signifi- intestinal dysbiosis could possibly lead to reduced bacte- cantly higher exposure to antibiotics and previous hospi- rial diversity and, in those genetically susceptible, result tal admissions. in IBD onset. In our population, we also found concomi- tant CMV infection (Table 3, patient 6). Giving the epi- In our analysis, IBD patients represented 24% of our demiology of CMV as a flare-causing pathogen in IBD, population with a predominance of male sex and a ten- the endoscopic features (deep ulcers on colon) and the dency for younger age when compared to non-IBD pa- significant prevalence of Aeromonas carriage status, it is tients. In the previous series, IBD patients represented plausible that the Aeromonas isolated was not responsible 11–14% of all Aeromonas-positive patients with the same for causing the disease. age trend [7, 8]. Symptomatic Aeromonas infection was seen in almost all patients (96%). Distinctive symptoms In our population, we observed the subsequent diag- between the two groups were the presence of bloody diar- nosis of IBD in 5/12 patients after Aeromonas infection. rhea and abdominal pain which was significantly more Lobatón et al. [7] also described 2 cases of diarrhea and common in IBD patients. This was in line with previous abdominal pain with Aeromonas detection simultaneous- findings suggesting that a colonic involvement in the vast ly to CD diagnosis. Similarly, more than 30 years ago, an majority of IBD might explain the high rates of bloody association was proposed between Aeromonas infection diarrhea in these patients [7]. and the new onset of 3 cases of UC [19]. A higher proportion of IBD patients presented a se- Although it represents one of the few studies relating vere clinical infection which followed the trend of previ- Aeromonas infection and IBD patients, there are some ous studies [7]. Three IBD patients presented a more limitations to address. The retrospective character of the severe disease: one was under tacrolimus after liver study, data retrieval from a unicentric microbiology da- transplant due to primary sclerosing cholangitis; the tabase, and the loss of 3 patients due to insufficient clini- other patient was under combined therapy with inflix- cal information might bias the interpretation of our re- imab and AZA; finally, a 64-year-old female with de sults. It is important to mention that only 7 IBD patients novo CD and severe endoscopic features with concomi- with flares were submitted to endoscopic evaluation, tant CMV colitis. Almost half of non-IBD patients pres- which, to some extent, could compromise the conclusion. ent a mild disease, which is a very similar result com- Also, the analysis of the time between Aeromonas isola- pared to the 44% previously reported [7, 16]. IBD popu- tion and the first signs of IBD would be important in the lation significantly needed more antibiotic therapy and evaluation of the possible causality between the two dis- steroids, following the trend of higher severity of infec- eases in future larger studies. It is important to mention tion. This might be explained by the lower threshold to the heterogeneity of the control group, mainly the inclu- antibiotic therapy in patients with IBD, namely in im- sion of pediatric patients that could compromise the munocompromised, regardless of clinical, analytical, comparisons that were made. The reduced sample and and endoscopic features. the bias to request stool culture mainly in symptomatic patients do not allow us to infer about the overall and eco- Regarding the place of Aeromonas infection in the nat- logical scenario of Aeromonas identification in our popu- ural history of the disease, the scarcely reported literature lation. on the subject poses it as a potential trigger to flare and for the de novo IBD diagnosis. In our study, from those In conclusion, Aeromonas infection appears to play an with IBD, 7/12 were diagnosed previously to Aeromonas important role in IBD activity. This infection might con- isolation. A role for Aeromonas as a trigger to IBD flare tribute as one more piece in the interactome unsolved was previously reported in a patient diagnosed with mild puzzle for IBD. Our results reinforce that Aeromonas in- Positioning Aeromonas Infection and IBD GE Port J Gastroenterol 2023;30:20–28 27 DOI: 10.1159/000520272

fection can be a trigger for IBD flare or de novo IBD di- Conflict of Interest Statement agnosis, supporting the importance of fecal culture analy- The authors have no conflicts of interest to declare. sis. Our results allying with the lack of data on Aeromo- nas infection and IBD might indicate an overlook of this Funding Sources infection. Statement of Ethics No funding. The authors declare that the procedures followed were in ac- Author Contributions cordance with the regulations of the relevant clinical research eth- ics committee and with those of the Code of Ethics of the World T.P.G. and J.A.S. equally contributed to the manuscript. T.P.G. Medical Association (Declaration of Helsinki). and M.S. conceived and designed the study. T.P.G., J.A.S., S.N., D.F. collected, analyzed, and interpreted data. P.C., P.L., I.P., and The authors declare that they have followed the protocols of M.S. performed critical review of the manuscript. their work center on the publication of patient data. The authors declare that no patient data appear in this article. References 1 Zhou Y, Yu L, Nan Z, Zhang P, Kan B, Yan D, 9 de Souza HS, Fiocchi C, Iliopoulos D. The 17 de Sola Earle C, Montiel Quezel-Guerraz N, et al. Taxonomy, virulence genes and antimi- IBD interactome: an integrated view of aetiol- Hidalgo Rojas L, Sánchez Cantos A, García crobial resistance of Aeromonas isolated from ogy, pathogenesis and therapy. Nat Rev Gas- Alegría J. Gastroenteritis aguda grave por extra-intestinal and intestinal infections. troenterol Hepatol. 2017 Dec;14(12):739–49. Aeromonas en paciente colectomizado por BMC Infect Dis. 2019 Feb;19(1):158. enfermedad de Crohn. Rev Esp Enferm Dig. 10 Baumgart DC, Carding SR. Inflammatory 1997 Jan;89(1):48–50. 2 Bhowmick UD, Bhattacharjee S. Bacteriologi- bowel disease: cause and immunobiology. cal, clinical and virulence aspects of aeromo- Lancet. 2007 May;369(9573):1627–40. 18 Ahishali E, Pinarbasi B, Akyuz F, Ibrisim D, nas-associated diseases in humans. Pol J Mi- Kaymakoglu S, Mungan Z. A case of Aeromo- crobiol. 2018 Jun;67(2):137–49. 11 Orel R, Kamhi Trop T. Intestinal microbiota, nas hydrophila enteritis in the course of ulcer- probiotics and prebiotics in inflammatory ative colitis. Eur J Intern Med. 2007 Sep; 3 Morgan DR, Johnson PC, DuPont HL, Sat- bowel disease. World J Gastroenterol. 2014 18(5):430–1. terwhite TK, Wood LV. Lack of correlation Sep;20(33):11505–24. between known virulence properties of 19 Willoughby JM, Rahman AF, Gregory MM. Aeromonas hydrophila and enteropatho- 12 Lobatón T, Domènech E. Bacterial Intestinal Chronic colitis after Aeromonas infection. genicity for humans. Infect Immun. 1985 Oct; Superinfections in Inflammatory Bowel Dis- Gut. 1989 May;30(5):686–90. 50(1):62–5. eases Beyond Clostridum difficile. Inflamm Bowel Dis. 2016 Jul;22(7):1755–62. 20 Igbinosa IH, Igumbor EU, Aghdasi F, Tom M, 4 Tomás JM. The main Aeromonas pathogenic Okoh AI. Emerging Aeromonas species infec- factors. ISRN Microbiol. 2012 Sep;2012: 13 Barnich N, Carvalho FA, Glasser AL, Darcha tions and their significance in public health. 256261. C, Jantscheff P, Allez M, et al. CEACAM6 acts ScientificWorldJournal. 2012;2012:625023. as a receptor for adherent-invasive E. coli, 5 Ko WC, Lee HC, Chuang YC, Liu CC, Wu JJ. supporting ileal mucosa colonization in 21 Fernández-Bravo A, Figueras MJ. An Update Clinical features and therapeutic implications Crohn disease. J Clin Invest. 2007 Jun;117(6): on the Genus Aeromonas: Taxonomy, Epide- of 104 episodes of monomicrobial Aeromo- 1566–74. miology, and Pathogenicity. Microorgan- nas bacteraemia. J Infect. 2000 May;40(3):267– isms. 2020 Jan;8(1):129. 73. 14 Nishikawa J, Kudo T, Sakata S, Benno Y, Sug- iyama T. Diversity of mucosa-associated mi- 22 Janda JM, Abbott SL. The genus Aeromonas: 6 von Graevenitz A. The role of Aeromonas in crobiota in active and inactive ulcerative coli- taxonomy, pathogenicity, and infection. Clin diarrhea: a review. Infection. 2007 Apr;35(2): tis. Scand J Gastroenterol. 2009;44(2):180–6. Microbiol Rev. 2010 Jan;23(1):35–73. 59–64. 15 Castaño-Rodríguez N, Kaakoush NO, Lee 23 EUCAST. Clinical breakpoints and dosing of 7 Lobatón T, Hoffman I, Vermeire S, Ferrante WS, Mitchell HM. Dual role of Helicobacter antibiotics [Internet]. [cited 2021 Feb 27]. M, Verhaegen J, Van Assche G. Aeromonas and Campylobacter species in IBD: a system- Available from: https://www.eucast.org/clini- species: an opportunistic enteropathogen in atic review and meta-analysis. Gut. 2017 Feb; cal_breakpoints/ patients with inflammatory bowel diseases? A 66(2):235–49. single center cohort study. Inflamm Bowel 24 Sinclair HA, Heney C, Sidjabat HE, George Dis. 2015 Jan;21(1):71–8. 16 Baliellas Comellas C, Xiol Quingles X, Castel- NM, Bergh H, Anuj SN, et al. Genotypic and lote Alonso J, Torras Torra J. [Severe occur- phenotypic identification of Aeromonas spe- 8 Elorza A, Rodríguez-Lago I, Martínez P, Hi- rence of idiopathic ulcerative colitis caused by cies and CphA-mediated carbapenem resis- dalgo A, Aguirre U, Cabriada JL. Gastrointes- Aeromonas sobria]. Rev Esp Enferm Dig. tance in Queensland, Australia. Diagn Micro- tinal infection with Aeromonas: incidence 1991 Jun;79(6):442–3. biol Infect Dis. 2016 May;85(1):98–101. and relationship to inflammatory bowel dis- ease. Gastroenterol Hepatol. 2020 Dec;43(10): 614–9. 28 GE Port J Gastroenterol 2023;30:20–28 Pereira Guedes/Alves Silva/Neves/Falcão/ Costa/Lago/Pedroto/Salgado DOI: 10.1159/000520272

Research Article Received: August 5, 2021 Accepted: October 8, 2021 GE Port J Gastroenterol 2023;30:29–37 Published online: November 30, 2021 DOI: 10.1159/000520530 Development of a Model to Predict Liver Decompensation prior to Transarterial Chemoembolization Refractoriness in Patients with Intermediate-Stage Hepatocellular Carcinoma Joel Ferreira-Silvaa Pedro Costa-Moreiraa, b Helder Cardosoa, b Rodrigo Liberala, b Pedro Pereiraa, b Guilherme Macedoa, b aGastroenterology Department, Hospital de São João, Porto, Portugal; bFaculty of Medicine of the University of Porto, Porto, Portugal Keywords nine patients were included. Objective response rate (ORR) Hepatocellular carcinoma · Transarterial chemoembolization · Child-Pugh class to initial TACE was assessed as present in 59 (63.4%) and as Abstract absent in 34 (36.6%) patients. Liver decompensated before Introduction: Transarterial chemoembolization (TACE) is the first-line treatment for patients with intermediate-stage he- TACE refractoriness in 51 (51.5%) patients, and the median patocellular carcinoma (HCC). For patients without an ade- quate response, current finding suggests that treatment time to liver decompensation was 14 (IQR 8–20) months af- with molecular target agents, approved for advanced stage, might present benefits. However, this requires a preserved ter first TACE. In multivariate analysis, beyond up-to-7 crite- liver function. This study aims to evaluate possible predictors of early deterioration of hepatic reserve, prior to TACE refrac- ria (HR 2.4, p = 0.031), albumin <35 mg/dL (HR 3.5, p < 0.001) toriness, in a cohort of patients treated with TACE. Methods: Retrospective analysis of 99 patients with Child-Pugh class A and absence of ORR (HR 2.4, p = 0.020) were associated with and intermediate-stage HCC who underwent TACE as the first-line treatment. All patients were submitted to a bio- decreased overall survival free of liver decompensation. chemical and medical evaluation prior to initial TACE and every month afterward. Response to initial TACE was evalu- Moreover, beyond up-to-7 criteria, albumin <35 mg/dL and ated at 1 month. The time to Child-Pugh class deterioration before TACE refractoriness was assessed. Results: Ninety- absence of ORR associated negatively with 6-month survival free of liver decompensation. Our model created using those variables was able to predict liver decompensation at 6 months with an AUROC of 0.701 (p = 0.02). Conclusions: The absence of ORR after initial TACE, beyond up-to-7 criteria and albumin <35 mg/dL, was a predictive factor for early liv- er decompensation before TACE refractoriness in our popu- lation. Such patients might benefit from treatment escala- tion to systemic therapy, in monotherapy or in combination with TACE. © 2021 Sociedade Portuguesa de Gastrenterologia Published by S. Karger AG, Basel [email protected] © 2021 Sociedade Portuguesa de Gastrenterologia Correspondence to: www.karger.com/pjg Published by S. Karger AG, Basel Joel Ferreira-Silva, [email protected] is is an Open Access article licensed under the Creative Commons Attribution-NonCommercial-4.0 International License (CC BY-NC) (http://www.karger.com/Services/OpenAccessLicense), applicable to the online version of the article only. Usage and distribution for com- mercial purposes requires written permission.

Desenvolvimento de um modelo preditor de lação. Esses doentes podem beneficiar do escalonamen- descompensação hepática pré refratariedade a to do tratamento para a terapia sistêmica, em monote- quimioembolização transarterial em doentes com rapia ou em combinação com TACE. carcinoma hepatocelular em estadio intermediário © 2021 Sociedade Portuguesa de Gastrenterologia Publicado por S. Karger AG, Basel Palavras Chave Introduction Carcinoma hepatocellular · Quimioembolização transarterial · Classe Child-Pugh Transarterial chemoembolization (TACE) is the first-line treatment for patients with non-resectable he- Resumo patocellular carcinoma (HCC) in the absence of de- Introdução: A quimioembolização transarterial (TACE) é compensated cirrhosis, large tumor size (>10 cm), co- o tratamento de primeira linha para doentes com carci- morbidities, portal vein thrombosis, or extrahepatic noma hepatocelular (HCC) em estadio intermédio. Em spread [1–3]. The survival benefit of TACE in patients doentes sem resposta adequada, a evidência atual sug- with intermediate-stage HCC is well established [4]. ere que o tratamento com agentes de alvo molecular, However, this therapy frequently loses its therapeutic aprovado para estágio avançado, pode apresentar efficacy over time, despite repeated procedures, leading benefícios. Porém, isso requer função hepática preser- to TACE refractoriness [5]. vada. O objetivo deste estudo é avaliar possíveis predi- tores de deterioração precoce da reserva hepática, antes For patients without an adequate response after da refratariedade ao TACE, em uma coorte de doentes TACE, the concept of therapeutic stage migration can be tratados com TACE. Métodos: Análise retrospectiva de considered [6]. Systemic therapy (ST) has been demon- noventa e nove doentes com Child-Pugh classe A e HCC strated to be effective in patients after TACE failure [7]. em estadio intermédio que foram submetidos a TACE Furthermore, recent studies suggest that, in patients with como tratamento de primeira linha. Todos os doentes HCC and TACE refractoriness, ST might improve the foram submetidos a uma avaliação bioquímica e médica prognosis, when compared with repeated TACE proce- antes do TACE inicial e a cada mês após. A resposta ao dures [8–10]. Moreover, several studies demonstrate a TACE inicial foi avaliada em 1 mês. O tempo para a dete- benefit of systemic treatment as an adjuvant to TACE rioração da classe Child-Pugh antes da refratariedade a [11]. As such, ST can be considered in patients with in- TACE foi avaliado. Resultados: Noventa e nove doentes termediate-stage HCC [1, 7]. However, ST, similar to lo- foram incluídos. A resposta radiológica objetiva (ORR) coregional therapy, requires preserved liver function, ao TACE inicial foi avaliada como presente em 59 (63.4%) and patients that develop early deterioration of liver e ausente em 34 (36.6%) doentes. Descompensação function following TACE are deprived of ST as a thera- hepática ocorreu, antes da refratariedade a TACE, em 51 peutic option [12]. To complicate matters, up to 25% of (51.5%) doentes e o tempo médio para a descompensa- patients with TACE refractoriness develop Child-Pugh ção hepática foi de 14 (IQR 8–20) meses, após o primeiro B/C after initial TACE [8, 9]. Furthermore, it has previ- TACE. Na análise multivariada, além dos critérios up-to-7 ously been described that repeated TACE procedures are (HR 2,4, p = 0.031), albumina <35 mg/dL (HR 3,5, p < associated with deterioration of hepatic reserve, even 0.001) e ausência de ORR (HR 2,4, p = 0.020) foram asso- though recent studies might challenge the relevance of ciados a diminuição da sobrevida livre de descompensa- this association [11, 13]. Defining predictors of early liv- ção hepática. Além disso, a sobrevida de 6 meses livre de er decompensation in patients with HCC proposed to descompensação hepática apresentou associação, além TACE could potentially help in better treatment selec- dos critérios up-to-7, albumina <35 mg/dL e ausência de tion in such patients. Previous works fail to propose pre- ORR. Foi criado um modelo com essas variáveis, capaz dictors of early deterioration of liver function before de prever a descompensação hepática com AUROC de TACE refractoriness. 0,701 (p = 0.02). Conclusões: A ausência de ORR após TACE inicial, além dos critérios up-to-7 e albumina <35 This study aims to evaluate possible predictors of over- mg/dL foram fatores preditivos para descompensação all and 6-month survival free of liver decompensation be- hepática antes da refratariedade a TACE na nossa popu- fore TACE refractoriness in a cohort of Western patients with intermediate-stage HCC submitted to TACE. 30 GE Port J Gastroenterol 2023;30:29–37 Ferreira-Silva et al. DOI: 10.1159/000520530

Fig. 1. Flowchart regarding patient selec- Patients with intermediate stage HCC and Child- tion. HCC, hepatocellular carcinoma; Pugh A cirrhosis treated with TACE (n = 161) TACE, transarterial chemoembolization. Excluded: • Other treatment previous or concomitant with TACE (n = 31) • Treatment with conventional TACE (n = 12) • Non-appropriate follow-up (n = 10) • Comorbidities of other serious diseases (n = 9) Studied patients (n = 99) Material and Methods Definition of TACE Responder, Liver Decompensation, and TACE Refractoriness Study Design ORR was defined as (i) present in patients with complete re- We performed a retrospective study of all consecutive patients sponse or partial response and as (ii) absent in patients with stable with intermediate-stage HCC submitted to drug-eluting polyvinyl disease or progressive disease in accordance with mRECIST crite- alcohol microspheres TACE (DEM-TACE), as proposed by a mul- ria. Liver decompensation was defined as irreversible exacerbation tidisciplinary board, in a tertiary center between January 2010 and from Child-Pugh class A to Child-Pugh class B or C after first October 2020. This study was carried out in compliance with the TACE. TACE refractoriness, in accordance with the Japan Society ethical principles outlined in the Declaration of Helsinki and was of Hepatology, was defined as (i) ˰2 TACE procedures with each approved by the Ethical Committee of our center. The requirement reevaluation showing progression in tumor number or insufficient for written informed consent was waived. Inclusion criteria where: response of the treated tumors (viable lesions >50%) or (ii) appear- (i) radiological or histological evidence of HCC in accordance with ance of extrahepatic spread or vascular invasion [17]. the diagnosis criteria of American Association for the Study of Liv- er Diseases guidelines, (ii) intermediate-stage HCC in accordance Outcomes with Barcelona Clinic Liver Cancer staging system, (iii) Child-Pugh The primary outcome was overall survival free of liver decom- class A cirrhosis before DEM-TACE, and (iii) treatment with pensation, defined as the time interval from the first TACE to liv- TACE. Exclusion criteria were: (i) comorbidities of other serious er decompensation. Secondary outcomes were 6-month survival diseases (American Society of Anesthesiologists physical status >2), free of liver decompensation, overall survival free of TACE refrac- (ii) lack of appropriate follow-up (clinical evaluation or radiologi- toriness (time interval from the first TACE to TACE refractori- cal tumor evaluation within 3 months after initial DEM-TACE), ness), and overall survival (time from the first TACE to death). (iii) other treatment previous or concomitant with TACE, and (iv) treatment with conventional TACE [14]. All patients were submit- Statistical Analysis ted to a biochemical and medical evaluation before initial TACE. All variables were presented as categorical variables, with counts and percentages. Overall survival free of liver decompensation was Transarterial Chemoembolization Procedure and Subsequent defined as the time interval between initial TACE and development Follow-Up of liver decompensation. Patients who survived without liver de- After catheter insertion in the celiac artery, guided by angiog- compensation at the last follow-up date (December 10, 2020) or raphy, a microcatheter injects contrast into the common hepatic were lost to follow-up were censored. Patients with TACE refrac- artery, identifying the arteries that feed the tumor. The microcath- toriness before liver decompensation were censored at the time of eter is then advanced as far as possible in the segmental or subseg- TACE refractoriness. Overall survival free of liver decompensation mental branches feeding the tumor. Drug-eluting microspheres rates were estimated using the Kaplan-Meier method and com- loaded with doxorubicin are injected into the tumor’s supply arter- pared using the log-rank test. Multivariable analysis, using Cox- ies until blood flow is obstructed. The selection of doxorubicin regression, was performed for the analysis of the following out- dose is determined by our institution’s protocol. comes: overall and 6-month survival free of liver decompensation, Objective response rate (ORR) to TACE was evaluated 1 month overall survival free of TACE refractoriness, and overall survival. after each procedure with magnetic resonance or computer to- All variables statistically significant in the previous univariable mography scan, in accordance with modified response evaluation analysis were included in this analysis. Multiple imputations had criteria in solid tumors (mRECIST) [15]. Subsequently, every 3 been performed with 5 independent draws for missing values. The months imaging was performed and intrahepatic recurrences or interactions between predictors were also tested. presence of viable tumors dictated additional treatment, TACE, or Possible models for risk stratification of overall survival free of other indicated treatment depending on TACE refractoriness cri- liver decompensation were then developed based on the above teria. analyses and performance was measured by likelihood ratio chi- Adverse events were accessed retrospectively in patient records square. A ROC curve analysis was developed to determine the best and were classified according to Common Terminology Criteria of cut-off value for the model. Kaplan-Meier analysis compared over- Adverse Events (CTCAE) 5.0 [16]. Only CTCAE grade ˰3 adverse all survival free of liver decompensation between the two groups events were recorded. created. Liver Decompensation in Intermediate- GE Port J Gastroenterol 2023;30:29–37 31 Stage HCC DOI: 10.1159/000520530

Table 1. Baseline characteristics Differences were considered statistically significant when cor- responding p values were less than 0.05. All statistical analyses Parameter n (%) were performed using SPSS Version 23 (IBM Corporation, Chi- cago, IL, USA). Sex 20 (20.2%) Female 79 (79.8%) Results Male 24 (24.2%) Patients’ Baseline Characteristics Age, years 21 (21.2%) From a total of 161 patients with intermediate-stage <60 54 (54.5%) HCC and Child-Pugh A cirrhosis treated with TACE, 99 60–70 patients were included (Fig.  1). Mean age was 73 (IQR >70 42 (42.4%) 59–86) years, 78 (79%) were male. Median follow-up time 34 (34.3%) was 40 (IQR 11–62) months. The etiology of hepatic dis- Etiology 7 (7.1%) ease was alcohol in 42%, chronic hepatitis C in 34%, Alcohol 5 (5.1%) chronic hepatitis B (HBV) in 7%, non-alcoholic steato- HCV 11 (11.1%) hepatitis in 5%, and others in 11% (Table 1). Median tu- HBV mor size was 37 mm (10–180 mm). Median tumor num- NASH 86 (86.8%) ber of nodules was 1.6 (IQR 1–2). Eighty-one patients Others 7 (7.1%) (82%) were within up-to-7 criteria and 18 patients (18%) 6 (6.1%) were beyond up-to-7 criteria. Median albumin was 35.5 Tumor size, mm g/L (IQR 34–38), median bilirubin was 1 mg/dL (IQR <50 58 (58.6%) 0.7–1.5), median alpha-fetoprotein was 10.9 ng/mL (IQR 50–70 25 (25.3%) 4–70), and median international normalized ratio was 1.2 >70 8 (8.1%) (IQR 1.1–1.3). No patient presented ascites or hepatic en- 8 (8.1%) cephalopathy. Number of nodules 1 81 (81.8%) Overall and 6-Month Survival Free of Liver 2 18 (18.2%) Decompensation before TACE Refractoriness 3 Liver function deteriorated to Child-Pugh B/C be- >3 43 (43.4%) fore TACE refractoriness in 51 (51.5%) patients and 56 (56.6%) median time to liver decompensation was 14 (IQR Up-to-7 8–20) months, after first TACE. In univariable analysis, Within 78 (78.8%) overall survival free of liver decompensation was sig- Beyond 21 (21.2%) nificantly longer in patients within versus beyond up- to-7 criteria (median, 40.9 vs. 20.3, p = 0.041), albumin Albumin, g/L 43 (43.4%) >35 versus ˯35 mg/dL (median, 43.7 vs. 24.5 months, ≤35 56 (56.6%) p < 0.001), bilirubin <2 versus ˰2 mg/dL (median, 43.9 >35 vs. 18.9 months, p = 0.029), and presence versus absence 61 (61.6%) of ORR (median, 56.7 vs. 27.6 months, p = 0.002) (Table Bilirubin, mg/dL 38 (38.4%) 2). In multivariable analysis, beyond up-to-7 criteria <2 (HR 2.4, p = 0.031), albumin <35 mg/dL (HR 3.5, p < ≥2 41 (41.4%) 0.001), and absence of ORR (HR 2.4, p = 0.020) present- 58 (58.6%) ed a negative association with overall survival free of Platelets, ×109 liver decompensation (Table 3). Moreover, 6-month ≥150 32 (32.3%) survival free of liver decompensation presented a nega- <150 34 (34.3%) tive association, in multivariable analysis, with beyond 17 (17.2%) up-to-7 criteria (HR 3.7, p = 0.012), albumin <35 mg/dL AFP, ng/mL 16 (16.2%) (HR 4.4, p = 0.006), and absence of ORR (HR 2.6, p = <20 0.025) (Table 2). ≥20 22 (22.2%) 28 (28.3%) ORR 49 (49.5%) Present Absent TACE procedures 1 2 3 ≥4 Doxorubicin, mg <35 35–70 ≥70 HCV, hepatitis C virus; HBV, hepatitis B virus; NASH, non-alcoholic steatohepatitis; BCLC, Barcelona Clinic Liver Cancer; AFP, alpha- fetoprotein; ORR, objective radiological response at 1 month after first transarterial chemoembolization; TACE, transarterial chemoembolization. 32 GE Port J Gastroenterol 2023;30:29–37 Ferreira-Silva et al. DOI: 10.1159/000520530

Table 2. Univariable analysis for overall survival free of liver decompensation, 6-month survival free of liver decompensation, survival free of TACE refractoriness, and overall survival Parameter OS free of LD p value 6-month survival p value Survival free of p value OS (median, p value (median, 95% CI), free of LD (median, TACE refractoriness 95% CI), months months 95% CI), months (median, 95% CI), months Sex 30.3 (12.2–48.4) 0.764 5.6 (5.1–6.2) 0.0661 33.0 (12.0–53.9) 0.857 35.4 (18.1–42.8) 0.626 Female 27.2 (8.4–46.3) 0.095 5.4 (5.2–5.7) 0.234 29.0 (10.6–47.4) 0.270 44.4 (27.2–51.5) 0.017 Male 0.011 0.223 0.070 0.006 43.6 (25.7–61.2) 5.2 (4.6–5.7) 21.0 (12.2–25.7) 46.5 (21.8–61.2) Age, years 20.3 (11.4–28.2) 0.079 5.4 (4.7–5.9) 0.483 18.0 (11.2–24.7) 0.008 29.3 (13.0–35.6) 0.427 <60 35.4 (29.2–42.5) 0.032 5.5 (4.9–6.1) 0.194 39.0 (12.8–49.2) 0.010 43.5 (27.2–49.9) 0.122 60–70 >70 25.4 (6.4–43.5) 0.041 5.7 (5.4–5.9) 0.007 25 (9.6–40.3) 0.001 37.7 (21.4–43.9) 0.005 58.2 (21.7–94.3) <0.001 5.4 (4.9–5.9) 0.006 49 (32.1–65.9) 0.711 52.9 (31.0–64.8) 0.019 Etiology 9.6 (3.4–14.6) 0.029 4.8 (4.1–5.6) 0.655 19 (10.2–39.0) 0.477 32.7 (5.0–30.4) 0.667 Alcohol 39.3 (18.3–59.7) 0.067 5.2 (4.6–5.7) 0.027 39 (9.3–68.6) 0.144 41.0 (19.7–52.3) 0.660 HCV 19.4 (8.1–38.5) 0.355 5.4 (4.7–5.9) 0.892 38 (9.3–40.2) 0.358 35.0 (8.7–51.2) 0.401 HBV 0.002 0.014 0.002 0.080 NASH 38.2 (28.9–47.4) 0.074 5.7 (5.4–5.9) 0.651 38.0 (23.0–52.2) 0.131 43.7 (26.9–50.5) 0.743 Others 30.9 (13.9–47.9) 5.4 (4.9–5.9) 18 (8.2–27.7) 38.4 (19.5–47.4) 5.3 (3.1–7.4) 0.457 4.6 (3.3–5.8) 0.934 8 (3.8–12.1) 0.333 37.9 (10.3–55.4) 0.663 Tumor size, mm <50 45.8 (34.6–56.9) 5.7 (5.5–5.9) 41.0 (19.3–62.6) 50.1 (30.9–59.2) 50–70 21.1 (10.9–31.3) 4.9 (4.4–5.6) 12.0 (2.7–21.6) 33.6 (16.3–40.9) >70 21.1 (7.8–34.3) 4.9 (3.5–6.2) 29.0 (17.0–62.4) 30.5 (11.0–39.9) 16 (2.5–25.4) 3.6 (3.4–6.2) 20.0 (13.7–46.9) 34.9 (12.3–47.6) Number of nodules 1 40.9 (31.5–50.3) 5.6 (5.4–5.8) 39 (27.0–50.9) 46.9 (29.6–54.2) 2 20.3 (9.9–30.7) 4.8 (4.1–5.4) 6 (3.4–8.5) 27.9 (10.3–35.6) 3 >3 20.4 (10.1–30.7) 5.6 (5.4–5.8) 29.0 (16.7–41.3) 37.5 (19.4–45.7) 40.9 (34.5–47.2) 4.8 (4.1–5.4) 39.0 (17.3–60.7) 45.9 (28.9–52.9) Up-to-7 Within 43.9 (33.8–54.1) 5.4 (5.0–5.8) 31.0 (17.2–44.7) 45.6 (25.9–55.3) Beyond 18.9 (7.4–30.3) 5.5 (5.2–5.9) 29.0 (12.8–54.2) 39.9 (27.3–50.6) Albumin, g/L 39.0 (30.9–47.1) 5.7 (5.2–6.4) 20.0 (15.2–24.8) 21.0 (16.7–25.3) ≤35 15.0 (5.1–24.8) 5.2 (4.1–5.7) 39.9 (31.1–46.9) 28.0 (15.6–40.4) >35 43.6 (31.4–55.7) 5.3 (4.6–6.1) 40.0 (35.9–44.0) 43.3 (24.6–51.9) Bilirubin, mg/dL 27.3 (17.3–37.4) 5.6 (5.1–6.1) 19.0 (15.3–22.7) 41.3 (22.1–50.4) <2 ≥2 56.8 (43.5–70.0) 5.6 (5.2–5.8) 57.3 (43.5–70.0) 39.6 (22.4–46.8) 27.7 (19.3–36.0) 4.8 (4.1–5.7) 31.7 (19.3–38.0) 49.6 (28.8–60.4) Platelets, ×109 ≥150 21.3 (8.6–34.5) 5.3 (4.5–6.2) 17.0 (0.5–33.5) 43.4 (24.6–51.9) <150 39.7 (29.8–54.3) 5.5 (5.2–6.4) 39.0 (22.7–55.3) 42.1 (22.1–50.4) 23.6 (14.6–35.3) 5.3 (4.7–6.1) 20.0 (17.2–22.7) 39.3 (22.4–46.8) AFP, ng/mL 41 (31.2–56.2) 5.4 (5.3–6.3) 41.0 (35.6–46.4) 46.5 (28.8–60.4) <20 ≥20 38.4 (19.3–57.5) 5.5 (5.0–6.1) 38.0 (23.1–52.9) 48.0 (21.7–64.3) 25.9 (15.7–36.1) 5.4 (4.8–6.0) 17.0 (11.8–22.1) 36.4 (16.8–45.9) ORR 35.9 (29.9–48.2) 5.5 (5.1–5.7) 39.0 (21.4–56.6) 41.0 (24.6–47.5) Present Absent TACE procedures 1 2 3 ≥4 Doxorubicin, mg <35 35–70 ≥70 OS, overall survival; LD, liver decompensation; CI, confidence interval; HCV, hepatitis C virus; HBV, hepatitis B virus; NASH, non-alcoholic steatohepatitis; BCLC, Barcelona Clinic Liver Cancer; AFP, alpha-fetoprotein; ORR, objective radiological response at 1 month after first transarterial chemoembolization; TACE, transarterial chemoembolization. Liver Decompensation in Intermediate- GE Port J Gastroenterol 2023;30:29–37 33 Stage HCC DOI: 10.1159/000520530

Table 3. Multivariable analysis for overall survival free of liver decompensation, 6-month survival free of liver function deterioration, sur- vival free of TACE refractoriness, and overall survival Parameter OS free of LD 6-month survival free Survival free of TACE OS of LD refractoriness hazard ratio hazard ratio (95% CI) p value hazard ratio p value hazard ratio p value (95% CI) p value (95% CI) (95% CI) Age, years 1 – 1 – 1 – 1 – ≤70 2.6 (0.7–8.6) 0.246 1.7 (0.6–7.1) 0.275 1.9 (0.5–6.5) 0.623 2.4 (1.3–8.1) 0.012 >70 1 – 1 – 1 – 1 – Etiology 4.0 (0.8–11.1) 0.123 1.7 (0.4–8.1) 0.483 1.2 (0.2–9.5) 0.844 1.4 (0.8–5.1) 0.237 Others HBV 1 – 1 – 1 – 1 – 2.4 (1.1–5.5) 0.031 3.7 (1.3–10.1) 0.012 3.9 (1.8–8.3) 0.001 2.4 (1.2–4.7) 0.011 Up-to-7 Within 1 – 1 – 1 – 1 – Beyond 3.5 (1.9–6.7) <0.001 4.4 (1.5–12.8) 0.006 1.3 (0.6–2.5) 0.503 1.7 (1.0–2.9) 0.042 Albumin, g/L 1 – 1 – 1 – 1 – >35 1.0 (0.9–1.2) 0.843 0.9 (0.5–1.4) 0.537 1.1 (0.4–2.5) 0.983 0.9 (0.8–1.1) 0.387 ≤35 1 – 1 – 1 – 1 – Bilirubin, mg/dL 2.4 (1.1–5.5) 0.343 3.1 (0.8–9.1) 0.127 1.7 (0.6–7.1) 0.275 1.8 (0.7–2.8) 0.387 <2 ≥2 1 – 1 – 1 – 1 – 2.4 (1.1–5.2) 0.020 2.8 (1.2–8.1) 0.025 2.7 (1.1–3.5) 0.025 1.6 (0.8–2.5) 0.074 Platelets, ×109 ≥150 <150 ORR Present Absent OS, overall survival; LD, liver decompensation; CI, confidence interval; HBV, hepatitis B virus; ORR, objective radiological response at 1 month after first transarterial chemoembolization. Table 4. Six-month survival free of liver decompensation for pa- We used the cut-off value of –1.49 to classify patients tients stratified as low- or high-risk in accordance with out model as low-risk (<–1.49) and high-risk (˰–1.49) in accor- dance with the model developed. Patients classified as Model classification 6-month SFLD p value high-risk (n = 52) presented a higher prevalence of liver (median, 95% CI), months decompensation (35%) when compared with patients classified as low-risk (10%). This difference was statisti- Low-risk 26.1 0.002 cally significant in chi-square test (p = 0.022). Further- High-risk 50.9 more, in Kaplan-Meyer analysis, high-risk patients pre- sented smaller overall survival free of liver decompensa- SFLD, survival free of liver decompensation. tion when compared to low-risk patients (median, 26.1 vs. 50.9 months, p = 0.002) (Fig. 2; Table 4). Using up-to-7-criteria, albumin, and ORR, we created Transarterial Chemoembolization Treatment a model predictive of 6-month survival free liver decom- Outcome pensation. The model formula is: 1.601 (if beyond up- The median TACE procedures per patient was 2 (IQR to-7 criteria) – 0.428 (if albumin >35 mg/dL) – 0.464 (if 1–3) and 41 patients (41%) presented ORR after first presence of ORR) – 1.285. TACE. Fifty (51%) patients developed TACE refractori- ness during follow-up and the median time to TACE re- ROC curve presented an AUROC 0.701 (p = 0.006) for fractoriness was 33 (IQR 11.7–54.3) months. Twelve the prediction of liver decompensation at 6 months. Us- (12.1%) patients presented post chemoembolization syn- ing the cut-off value of –1.49, the model presented a sen- drome. No other major (CTCAE grade ˰3) complica- sitivity of 81% and a specificity of 72% for liver decom- tions related to TACE were registered. pensation at 6 months. 34 GE Port J Gastroenterol 2023;30:29–37 Ferreira-Silva et al. DOI: 10.1159/000520530

100 ++ High-risk Discussion + Low-risk + + ++ Based on our findings, we suggest that baseline albu- min <35 g/L, beyond up-to-7 criteria, and absence of re- 80 + sponse to initial TACE are predictors of liver decompen- sation to Child-Pugh B/C before TACE refractoriness, in Child-Pugh class A, % ++ ++ ++ patients with intermediate-stage HCC submitted to TACE. Moreover, early (6 months) liver decompensation + ++ + was associated with baseline albumin <35 g/L, beyond 60 up-to-7 criteria, and absence of response to initial TACE. Indeed, our model created using those variables was able +++ + to predict liver decompensation after initial TACE, and ++++ patients classified as high risk presented shorter survival 40 + ++ + free of liver decompensation. + Previous studies associated different markers of liver function and tumor burden as predictors of liver decom- 20 + + + pensation. However, those studies did not attempt to fo- cus on predictors of liver decompensation before TACE + refractoriness [18–20]. Furthermore, previous studies fail to analyze a specific population of patients with interme- 0 20 40 60 80 diate-stage HCC or a population exclusively treated with 0 Time after initial TACE, months DEM-TACE. Identification of predictors of liver decom- 4 pensation previous to TACE refractoriness is specifically Patients at risk 22 11 4 15 useful in patients with intermediate-stage HCC, since the High-risk 52 concept of therapeutic stage migration could be antici- Low-risk 37 25 21 15 pated in such patients before TACE refractoriness. Fur- thermore, DEM-TACE, as opposed to conventional Fig. 2. Kaplan-Meier curves for the time to Child-Pugh class dete- TACE or transarterial embolization is currently the most rioration in patients with intermediate-stage hepatocellular carci- commonly used locoregional procedure in intermediate- noma who underwent transarterial chemoembolization (TACE) stage HCC, with a better security profile and, in some as first-line treatment. Comparing patients classified as high-risk studies, better efficacy [21–23]. Facciorusso et al. report a and low-risk in accordance with our model using the cut-off value meta-analysis that evaluated the efficacy and safety of of –1.49. TACE versus bland embolization [23]. In this study no differences in survival were identified. However, despite In univariable analysis, up-to-7 criteria, number of the apparently worst security profile of conventional nodules, total tumor size, and absence of ORR were iden- TACE versus bland embolization, DEM-TACE in par- tified as predictive factors associated with survival free of ticular appears to present lower rates of adverse events. TACE refractoriness (Table 2). In multivariable analysis, Indeed, in our population, only twelve patients presented beyond up-to-7 criteria and absence of ORR were identi- postembolization syndrome and no other complications fied as predictive factors of survival free of TACE refrac- were registered. Also, being a single-center study, TACE toriness (Table 3). procedures were standardized and the choice of antican- cer agents was uniformized. Overall Survival Presence of liver decompensation at 6 months after The ORR is proposed as a prognostic tool and is useful initial TACE was negatively associated with overall sur- in the selection of retreatment strategies [1]. Assessment vival. Median survival was 38.7 (IQR 31.2–46.1) months for retreatment with TACE (ART) score is a scoring sys- for patients with Child-Pugh A and 12 (IQR 8.6–15.3) tem using radiological response to initial TACE as a tool months for patients with Child-Pugh B/C at 6-month useful in the selection of patients that do not benefit from evaluation post first TACE procedure. This difference additional TACE [24, 25]. Our findings suggest ORR was statistically significant (p < 0.001). Moreover, albu- might also be useful to predict early liver dysfunction be- min <35 and beyond up-to-7 criteria were also negatively fore TACE refractoriness. associated with overall survival in univariable analysis (Table 2). In multivariable analysis, beyond up-to-7 cri- teria (HR 2.4, p = 0.011) and albumin <35 mg/dL (HR 1.7, p < 0.042) were negatively associated with overall surviv- al (Table 3). Liver Decompensation in Intermediate- GE Port J Gastroenterol 2023;30:29–37 35 Stage HCC DOI: 10.1159/000520530

Cumulative TACE procedure has been previously In conclusion, our findings suggest that in patients identified as a risk factor for liver decompensation [13, with intermediate-stage HCC and Child-Pugh class A, 26]. However, the number of TACE procedures and cu- beyond up-to-7 criteria, albumin <35 mg/dL, and ab- mulative doxorubicin dose did not present an association sence of ORR to initial TACE are predictors of early liver with overall survival free of liver decompensation in our decompensation before TACE refractoriness. We created population. This could be explained by the type of locore- a model using those variables that is able to predict liver gional therapy used. In our population, only DEM-TACE decompensation. Patients classified as high risk using this was used, contrary to previous studies. DEM-TACE pres- model have a smaller window of opportunity for alterna- ents less toxicity to liver function as opposed to conven- tive therapies and might benefit from inclusion in TACE tional TACE [21, 22]. plus ST combination therapy or ST monotherapy clinical trials. The concept of TACE refractoriness is useful to con- sider the opportunity to switch from TACE retreatment Statement of Ethics to ST [8, 17]. Previous studies demonstrated an associa- tion between liver decompensation and TACE refractori- This study was carried out in compliance with the ethical prin- ness [18]. Indeed, in our population, beyond up-to-7 cri- ciples outlined in the Declaration of Helsinki and was approved by teria and absence of ORR were independently associated the Ethical Committee of our center. The requirement for written with both liver decompensation and TACE refractori- informed consent was waived. ness. As such, patients beyond up-to-7 criteria, with ab- sence of ORR, or with albumin <35 g/L not only have a Conflict of Interest Statement smaller window of opportunity to initiate ST but have a The authors have no conflicts of interest to declare. greater probability of TACE refractoriness. Funding Sources Several randomized controlled trials evaluating effica- The authors have no funding to declare. cy and safety of ST plus TACE or TACE alone in patients with intermediate-stage HCC have been done and others Author Contributions are currently in progress [11, 27]. Such studies have, so far, suggested a benefit in progression-free survival with J.F.-S. conceived the study. J.F.-S. and H.C. designed the study. this combination. However, the sub-population of pa- J.F.-S. and P.C.-M. performed data collection. J.F.-S., H.C., R.L., tients who benefit the most from this combination is not and P.P. undertook the bibliographic research and drafted the yet described. We hypothesize that patients classified as manuscript. G.M. critically revised the draft and approved the final high-risk of liver decompensation, using our model, paper. might benefit from inclusion in such studies. References 1 Galle PR, Tovoli F, Foerster F, Wörns MA, Furthermore, overall survival in patients with HCC Cucchetti A, Bolondi L. The treatment of in- treated with locoregional therapy is influenced by both termediate stage tumours beyond TACE: recurrence and successive treatments. Previous studies from surgery to systemic therapy. J Hepatol. demonstrate that a significant portion of patients experi- 2017 Jul;67(1):173–83. ence HCC recurrence after locoregional therapy [28]. In patients with recurrence with advance-stage HCC, ST is 2 Raoul JL, Sangro B, Forner A, Mazzaferro V, the only palliate treatment available. Such patients are de- Piscaglia F, Bolondi L, et al. Evolving strate- prived of this option in the presence of liver decompensa- gies for the management of intermediate- tion. Patients classified as high-risk of liver decompensa- stage hepatocellular carcinoma: available evi- tion, using our model, present an increased risk of liver dence and expert opinion on the use of trans- decompensation and in consequence, might have an in- arterial chemoembolization. Cancer Treat creased risk of being deprived of adequate treatment in Rev. 2011 May;37(3):212–20. case of HCC recurrence. 3 de Lope CR, Tremosini S, Forner A, Reig M, As a retrospective study, there are some limitations Bruix J. Management of HCC. J Hepatol. from the presence of possible bias. Moreover, sample size 2012;56 Suppl 1:S75–87. limited the number of variables included in multivariable analysis, and only significant variables in univariable analysis were included. Further prospective large studies are necessary to confirm our results. 36 GE Port J Gastroenterol 2023;30:29–37 Ferreira-Silva et al. DOI: 10.1159/000520530

4 Marrero JA, Kulik LM, Sirlin CB, Zhu AX, 12 Abou-Alfa GK, Amadori D, Santoro A, Figer 20 Eso Y, Takai A, Takahashi K, Ueda Y, Taura Finn RS, Abecassis MM, et al. Diagnosis, stag- A, De Greve J, Lathia C, et al. Safety and effi- K, Marusawa H, et al. Combination of Mac-2 ing, and management of hepatocellular carci- cacy of sorafenib in patients with hepatocel- binding protein glycosylation isomer and up- noma: 2018 practice guidance by the Ameri- lular carcinoma (HCC) and Child-Pugh A to-seven criteria as a useful predictor for can Association for the Study of Liver Diseas- versus B cirrhosis. Gastrointest Cancer Res. Child-Pugh grade deterioration after transar- es. Hepatology. 2018 Aug;68(2):723–50. 2011 Mar;4(2):40–4. terial chemoembolization for hepatocellular carcinoma. Cancers (Basel). 2019 Mar;11(3): 5 Burrel M, Reig M, Forner A, Barrufet M, de 13 Hiraoka A, Kumada T, Kudo M, Hirooka M, 405. Lope CR, Tremosini S, et al. Survival of pa- Koizumi Y, Hiasa Y, et al.; Real-life Practice tients with hepatocellular carcinoma treated Experts for HCC (RELPEC) Study Group and 21 Ferrer Puchol MD, la Parra C, Esteban E, by transarterial chemoembolisation (TACE) HCC 48 Group (hepatocellular carcinoma ex- Vaño M, Forment M, Vera A, et al. Quimio- using Drug Eluting Beads. Implications for perts from 48 clinics). Hepatic function dur- embolización del hepatocarcinoma: partícu- clinical practice and trial design. J Hepatol. ing repeated TACE procedures and prognosis las cargadas frente a quimioembolización 2012 Jun;56(6):1330–5. after introducing sorafenib in patients with convencional. Radiologia (Madr). 2011;53(3): unresectable hepatocellular carcinoma: mul- 246–53. 6 Vitale A, Trevisani F, Farinati F, Cillo U. ticenter analysis. Dig Dis. 2017;35(6):602–10. Treatment of Hepatocellular Carcinoma in 22 Kloeckner R, Weinmann A, Prinz F, Pinto dos the Precision Medicine Era: From Treatment 14 Mayhew D, Mendonca V, Murthy BV. A re- Santos D, Ruckes C, Dueber C, et al. Conven- Stage Migration to Therapeutic Hierarchy. view of ASA physical status - historical per- tional transarterial chemoembolization ver- Hepatology. 2020 Dec;72(6):2206–18. spectives and modern developments. Anaes- sus drug-eluting bead transarterial chemoem- thesia. 2019 Mar;74(3):373–9. bolization for the treatment of hepatocellular 7 Bruix J, Raoul JL, Sherman M, Mazzaferro V, carcinoma. BMC Cancer. 2015 Jun;15(1):465. Bolondi L, Craxi A, et al. Efficacy and safety 15 Lencioni R, Llovet JM. Modified RECIST of sorafenib in patients with advanced hepa- (mRECIST) assessment for hepatocellular 23 Facciorusso A, Bellanti F, Villani R, Salvatore tocellular carcinoma: subanalyses of a phase carcinoma. in Seminars in liver disease. 2010. V, Muscatiello N, Piscaglia F, et al. Transarte- III trial. J Hepatol. 2012 Oct;57(4):821–9. ©Thieme Medical Publishers. rial chemoembolization vs bland emboliza- tion in hepatocellular carcinoma: A meta- 8 Ogasawara S, Chiba T, Ooka Y, Kanogawa N, 16 Trotti A, Colevas AD, Setser A, Rusch V, analysis of randomized trials. United Euro- Motoyama T, Suzuki E, et al. Efficacy of Jaques D, Budach V, et al. CTCAE v3.0: devel- pean Gastroenterol J. 2017 Jun;5(4):511–8. sorafenib in intermediate-stage hepatocellu- opment of a comprehensive grading system lar carcinoma patients refractory to transarte- for the adverse effects of cancer treatment. Se- 24 Yin W, Ye Q, Wang F, Liang J, Xu B, Zhang rial chemoembolization. Oncology. 2014; min Radiat Oncol. 2003 Jul;13(3):176–81. X, et al. ART score and hepatocellular carci- 87(6):330–41. noma: an appraisal of its applicability. Clin 17 Kudo M, Matsui O, Izumi N, Iijima H, Kadoya Res Hepatol Gastroenterol. 2016 Dec;40(6): 9 Arizumi T, Ueshima K, Minami T, Kono M, M, Imai Y, et al.; Liver Cancer Study Group of 705–14. Chishina H, Takita M, et al. Effectiveness of Japan. JSH Consensus-Based Clinical Practice sorafenib in patients with transcatheter arte- Guidelines for the Management of Hepato- 25 Tseng CL, Lai WJ, Huang CJ, Huang YH, Su rial chemoembolization (TACE) refractory cellular Carcinoma: 2014 Update by the Liver CW, Lee IC, et al. The Effectiveness of ART and intermediate-stage hepatocellular carci- Cancer Study Group of Japan. Liver Cancer. Score in Selecting Patients for Transarterial noma. Liver Cancer. 2015 Dec;4(4):253–62. 2014 Oct;3(3-4):458–68. Chemoembolization Retreatment: A Cohort Study in Taiwan. Medicine (Baltimore). 2015 10 Peck-Radosavljevic M, et al. Outcomes of pa- 18 Kohla MA, Abu Zeid MI, Al-Warraky M, Nov;94(47):e1659–1659. tients (pts) with hepatocellular carcinoma Taha H, Gish RG. Predictors of hepatic de- (HCC) treated with transarterial chemoem- compensation after TACE for hepatocellular 26 Miksad RA, Ogasawara S, Xia F, Fellous M, bolization (TACE): Global OPTIMIS final carcinoma. BMJ Open Gastroenterol. 2015 Piscaglia F. Liver function changes after analysis. American Society of Clinical Oncol- Jun;2(1):e000032. transarterial chemoembolization in US hepa- ogy; 2018. tocellular carcinoma patients: the LiverT 19 Yasui Y, Tsuchiya K, Kurosaki M, Takeguchi study. BMC Cancer. 2019 Aug;19(1):795. 11 Kudo M, Ueshima K, Ikeda M, Torimura T, T, Takeguchi Y, Okada M, et al. Up-to-seven Tanabe N, Aikata H, et al.; TACTICS study criteria as a useful predictor for tumor down- 27 Hindson J. Combined TACE and sorafenib group. Randomised, multicentre prospective staging to within Milan criteria and Child- for HCC treatment. Nat Rev Gastroenterol trial of transarterial chemoembolisation Pugh grade deterioration after initial conven- Hepatol. 2020 Feb;17(2):66–66. (TACE) plus sorafenib as compared with tional transarterial chemoembolization. Hep- TACE alone in patients with hepatocellular atol Res. 2018 May;48(6):442–50. 28 Facciorusso A, Del Prete V, Antonino M, carcinoma: TACTICS trial. Gut. 2020 Aug; Crucinio N, Neve V, Di Leo A, et al. Post-re- 69(8):1492–501. currence survival in hepatocellular carcinoma after percutaneous radiofrequency ablation. Dig Liver Dis. 2014 Nov;46(11):1014–9. Liver Decompensation in Intermediate- GE Port J Gastroenterol 2023;30:29–37 37 Stage HCC DOI: 10.1159/000520530

Research Article Received: August 4, 2021 Accepted: October 14, 2021 GE Port J Gastroenterol 2023;30:38–48 Published online: December 14, 2021 DOI: 10.1159/000520562 Anastomotic Leaks following Esophagectomy for Esophageal and Gastroesophageal Junction Cancer: The Key Is the Multidisciplinary Management Raquel Ortigãoa Brigitte Pereirab Rui Silvaa Pedro Pimentel-Nunesa, c Pedro Bastosa Joaquim Abreu de Sousad Filomena Fariab Mário Dinis-Ribeiroa, c Diogo Libânioa, c aGastroenterology Department, Portuguese Oncology Institute of Porto, Porto, Portugal; bIntensive Care Unit, Portuguese Oncology Institute of Porto, Porto, Portugal; cMEDCIDS, Faculty of Medicine, University of Porto, Porto, Portugal; dSurgery Department, Portuguese Oncology Institute of Porto, Porto, Portugal Keywords between these patients, but sepsis and large leakages were Esophageal cancer · Esophagectomy · Anastomotic leakage · tendentially managed by surgery. At follow-up, 3 patients in Endoscopic treatment the surgery group (23%) and 9 in the endoscopic group (90%) were discharged under an oral diet (p = 0.001). The in- Abstract hospital mortality rate was 38% in the surgical group, 33% in Introduction: Anastomotic leakage after esophagectomy is the conservative group, and 10% in endoscopic group (p = associated with high mortality and impaired quality of life. 0.132). In patients with EAL, the presence of septic shock at Aim: The objective of this work was to determine the effec- leak diagnosis was the only predictor of mortality (p = 0.004). tiveness of management of esophageal anastomotic leak- ICU length-of-stay was non-significantly lower in the endo- age (EAL) after esophagectomy for esophageal and gastro- scopic therapy group (median 4 days, vs. 16 days in the sur- esophageal junction (GEJ) cancer. Methods: Patients sub- gical group, p = 0.212). Conclusion: Risk factors for EAL may mitted to esophagectomy for esophageal and GEJ cancer at help change pre-procedural optimization. The results of this a tertiary oncology hospital between 2014 and 2019 (n = study suggest including an endoscopic approach for EAL. 119) were retrospectively reviewed and EAL risk factors and its management outcomes determined. Results: Older age © 2021 Sociedade Portuguesa de Gastrenterologia. and nodal disease were identified as independent risk fac- Published by S. Karger AG, Basel tors for anastomotic leak (adjusted OR 1.06, 95% CI 1.00– 1.13, and adjusted OR 4.89, 95% CI 1.09–21.8). Patients with Deiscências anastomóticas após esofagectomia por EAL spent more days in the intensive care unit (ICU; median neoplasia esofágica / juncional: a importância da 14 vs. 4 days) and had higher 30-day mortality (15 vs. 2%) and abordagem multidisciplinar higher in-hospital mortality (35 vs. 4%). The first treatment option was surgical in 13 patients, endoscopic in 10, and Palavras Chave conservative in 3. No significant differences were noticeable Cancro esofágico · Esofagectomia · Deiscência anastomótica · Tratamento endoscópico [email protected] © 2021 Sociedade Portuguesa de Gastrenterologia. Correspondence to: www.karger.com/pjg Published by S. Karger AG, Basel Raquel Ortigão, [email protected] is is an Open Access article licensed under the Creative Commons Attribution-NonCommercial-4.0 International License (CC BY-NC) (http://www.karger.com/Services/OpenAccessLicense), applicable to the online version of the article only. Usage and distribution for com- mercial purposes requires written permission.

Resumo esophagectomy carries a high risk of death (3.6–4.5%) Introdução: A deiscência anastomótica após esofagecto- compared with most surgically treated cancers [2–4]. mia está associada a uma elevada taxa de mortalidade e Many efforts have been made to improve the esophagec- qualidade de vida comprometida. Objetivo: Avaliar a tomy technique and to reduce postoperative complica- eficácia da abordagem da deiscência de anastomose es- tions, but esophageal anastomotic leakage (EAL) remains ofágica após esofagectomia por neoplasia do esófago e a frequent and feared postoperative complication, associ- da junção esofagogastrica (JEG). Métodos: Foram revistos ated with high mortality and impaired quality of life. retrospetivamente todos os doentes submetidos a esofa- However, improvement of surgical techniques and man- gectomia por neoplasia do esófago e da JEG num hospital agement of complications has led to a steady decrease in terciário entre 2014 e 2019 (n = 119) e analisados os fa- postoperative mortality over the years [5–8]. tores de risco e as diferentes abordagens na deiscência anastomótica. Resultados: A idade avançada e a presença Some factors have been associated with EAL develop- de metastização ganglionar foram identificados como fa- ment, such as patients’ nutritional status and comorbidi- tores de risco independentes para deiscência anastomóti- ties, cancer stage, surgical procedure, and neoadjuvant ca (OR 1.06, 95% IC 1.00–1.13 e 4.89, IC 1.09–21.8). Os therapy, but there are some controversies in the literature doentes com deiscência anastomótica estiveram mais about the significant risk factors for this adverse event [9, dias internados na unidade de cuidados intensivos (UCI) 10]. The identification of patients at risk for EAL can thus (mediana 14 vs. 4 dias) e tiveram uma mortalidade aos 30 help in postsurgical management. dias e intra-hospitalar mais elevada (15% vs. 2% e 35% vs. 4%, respectivamente). A primeira abordagem terapêutica EAL treatment success relies on early diagnosis, but foi cirúrgica em 13 doentes, endoscópica em 10 e conser- optimal treatment remains controversial. The treatment vadora em 3. Não foram encontradas diferenças estatisti- decision is dependent on the characteristics of the leak camente significativas entre estes doentes, com uma and the severity of the patient’s condition. In the past, tendência para a presença de sépsis e de deiscências de surgical revision with re-anastomosis or esophageal de- maior dimensão nos doentes abordados cirurgicamente. viation was the treatment of choice. Since the emergence Durante o seguimento, 3 doentes do grupo cirúrgico of endoscopic techniques, several potential endoscopic (23%) e 9 do grupo endoscópico (90%) tiveram alta hos- interventions have been used, such as clipping, self-ex- pitalar sob dieta oral (p = 0.001). A taxa de mortalidade pandable metal stents (SEMS), endoscopic vacuum ther- intra-hospitalar foi de 38% no grupo cirúrgico, 33% no apy (EVT), and endoscopic suturing devices [11–13]. grupo conservador e 10% no grupo endoscópico (p = 0.132). Nos doentes com deiscência anastomótica, a pre- While there are several studies evaluating the success sença de choque sético ao diagnóstico foi o único predi- of endoscopic and surgical treatment, there are few com- tor de mortalidade (p = 0.004). O tempo de internamento parative studies evaluating treatment outcomes of differ- na UCI não foi significativamente menor no grupo sub- ent management strategies specifically following onco- metido a tratamento endoscópico (mediana de 4 dias vs. logical esophagectomy. The aim of this study was to assess 16 dias no grupo cirúrgico, p = 0.212). Conclusão: A iden- anastomotic leakage rates after esophagectomy for esoph- tificação de fatores de risco para deiscência anastomótica ageal and gastroesophageal junction cancer (GEJ), to após esofagectomia pode ajudar a alterar a optimização identify possible risk factors for EAL, and compare the pré-procedimento. Os resultados deste estudo sugerem outcomes of patients with EAL according to management incluir uma abordagem endoscópica nos doentes com strategy. deiscência anastomótica. Materials and Methods © 2021 Sociedade Portuguesa de Gastrenterologia. Publicado por S. Karger AG, Basel Patients and Methods This was a retrospective cohort study including consecutive pa- Introduction tients submitted to esophagectomy for esophageal or GEJ cancer between January 2014 and December 2019 in Instituto Português Esophageal cancer is the sixth leading cause of cancer- de Oncologia do Porto – Francisco Gentil. All the patients submit- related death overall, mainly due to diagnosis in advanced ted to esophagectomy are initially managed in ICU where an elec- stages [1]. Even when esophageal cancer is resectable, tronic registry of all admissions ensure consecutive sampling. Data collection was performed through analysis of electronic medical records and patient charts. Patient demographic characteristics were collected along with the following clinical, surgical, and pathological characteristics: Esophageal Anastomotic Leakage GE Port J Gastroenterol 2023;30:38–48 39 DOI: 10.1159/000520562