Multidisciplinary Surgical Management of Patients with Clinically Severe

Multidisciplinary surgical management of patients with clinically severe obesity in a publicly funded bariatric surgery service in three public hospitals in Australia

Michelle (Mun Chieng) Tan

A thesis submitted to fulfil requirements for the degree of Doctor of Philosophy

The Boden Collaboration, Central Clinical School, Charles Perkins Centre, Faculty of Medicine and Health, The University of Sydney

2021 Statement of Originality

This is to certify that to the best of my knowledge, the content of this thesis is my own work. This thesis has not been submitted for any degree or other purposes.

I certify that the intellectual content of this thesis is the product of my own work and that all the assistance received in preparing this thesis and sources have been acknowledged.

Signature:

Michelle (Mun Chieng) Tan

i Acknowledgements | Michelle M.C. Tan

ACKNOWLEDGEMENTS

Core Research Centre, Study Hospitals and Fundings The clinical research described in this doctoral thesis was undertaken at the Charles Perkins Centre (CPC)’s Boden Collaboration of the University of Sydney (USyd), Royal Prince Alfred Hospital (RPAH), Concord Repatriation General Hospital (CRGH) and Camden Hospital. Special thanks to several funding bodies that have eased the financial burden during my PhD studies. Firstly, thanks to the Australian Commonwealth Government for awarding the Research Training Program (RTP) Scholarships to support my doctorate studies. Secondly, the CPC Early- and Mid- Career Researcher (EMCR) SEED Funding award for funding my data collection from all the participating hospitals throughout the years. I sincerely appreciate the Postgraduate Research Support Scheme (PRSS) and Boden Travel Support Scheme for providing travel funds enabling me to present my research at international conferences possible during my candidature.

The completion of this multicentre and multidisciplinary research thesis would not have been possible without the support, assistance and contribution of the many individuals I have had the pleasure of working with and learning from over these past three and a half years.

Clinical A/Prof Tania Markovic I would like to thank my main supervisor, Clinical A/Prof Tania Markovic for her clinical expertise and opportunity for leading this multicentre 10-year longitudinal publicly funded bariatric surgery research. I very much appreciated her invaluable guidance especially during the data collection phase in clearing inconsistencies and making collecting the best quality clinical data (both medical and surgical) possible in RPAH. I am also particularly grateful for her kind support, time and advice during my first year of doctoral studies - especially in initiating our cohort study together, providing me my first job in Australia, and her delightful companionship and care when I was starting a new life here. Her input and feedback upon reviewing my thick doctoral thesis that generated many manuscripts is also very much appreciated.

Acknowledgements | Michelle M.C. Tan

Prof Amanda Salis Importantly, I would like to express my deepest gratitude to Prof Amanda Salis, a supervisor, mentor and friend of mine who is no secret one of the best research supervisors that any HDR student could ask for. It has been such a pleasure to work with Amanda over the past few years. Not only is she friendly, kind, caring, cheerful, optimistic and encouraging; but also always supportive and incredibly generous with her time. Thank you for also being my teammate rubbing shoulders that I do not feel alone managing a large longitudinal study; for exchanging ideas and discussing details of the research components with me; moving forward together in confronting any obstacles to the completion; and for putting so much trust in me. I truly appreciate her for always listening to me so attentively and patiently, and for her insightful and helpful feedback on my written work to the highest quality. Her passion for research has also been an inspiration to continue improving myself and thriving in the academia. Thank you so much also for introducing me to her research group in Boden Collaboration who lightened up my working life and treated me like their own research group mate. Not forgetting, Amanda’s kind understanding of international students’ professional and personal difficulties will always be remembered deep in my heart for the rest of my life. Her generous sharing of invaluable experience studying and working abroad has taught me innumerable and invaluable lessons about research and people. Thank you so much Amanda.

Clinical A/Prof Samantha Hocking I would also like to thank my extremely energetic and down-to-earth supervisor, Clinical A/Prof Samantha Hocking for her kind guidance, advice and support over the years. I very much appreciate her help in setting up my new accommodation so that I could focus on the large PhD research in the following years. I really appreciate the time she has spent meeting with me and discussing our patient data, databases, endocrinology and obesity management knowledge, thesis drafts, ideas and solutions together. Thanks also for always moving forward with me in the biggest direction possible, teaching me tactful ways of speaking with local patients to obtain useful information while inviting them to re-engage with clinic, and kind efforts in understanding each other better and better over time. It’s always nice to meet with a mentor with similar high-

iii

Acknowledgements | Michelle M.C. Tan achieving spirit and great passion in clinical research that I hope we could cross paths in the future. I will forever miss our furniture hunting, patient contacting in RPAH together, and those times we prepared and attended conferences together in sharing our meaningful research findings.

Concord Repatriation General Hospital (CRGH), Camden Hospital and Dr Nic Kormas I would also like to thank the amazing teams in the Concord Repatriation General Hospital (which is also the study surgical operation site) and Camden Hospital who selflessly dedicated to my data collection over the years. It has been a pleasure working with all of you, especially Dr Nic Kormas, the Medical Director of both clinics in the two hospitals. He has also been a valuable expert contributor in my doctoral research, and from whom I have learnt so much. Thank you very much for the great help and assistance in facilitating my large data collection of 10-year follow-up and accommodating my many over-time data collection nights in clinics in addition to numerous meetings with him over the years. Thank you very much for also providing me with the invaluable opportunities to meet and interview the study patients face-to-face to obtain the most up-to- date data for rounding-up the research; intruding upon his already hectic daily endeavours. He has always been the most supportive person to me in the publicly funded bariatric surgery program. Thank you for also generously introducing me to the important network in the fields, including the President of World Obesity Federation, National Bariatric Surgery Registry staff and other bariatric surgeons to expand my research collaboration.

Importantly, I also appreciate all the other lovely staff in both clinics who transformed my challenging data collection years into a warm, happy and fun time, especially Prof Markus Seibel, Manager Ms Sue, Zaza, Bev, Maria, Carrie, Dr Veronica Wong, A/Prof Roger Chen, Dr Avinash, Hayley and Joanne. Apart from their warm companionship, I wish to thank them for supporting me with my research, despite their busy clinics, such as promptly tracking patients’ clinic appointments and tracing paper-based hospital records around the hospitals for me to complete data collection. Thanks also for all of your care and warmth when I got burned out and sick, for always dropping me to the train stations, and pampering me with the all-time delicious lunch,

Acknowledgements | Michelle M.C. Tan birthday cakes, chocolates, coffees, and exciting Christmas parties. Our time and photos taken together will be one of the best memories during my doctoral studies.

Royal Prince Alfred Hospital (RPAH) In the RPAH, I am particularly grateful for the staffs in the RPA Medical Records Department, especially Mr Andy You (Manager of Patient Information Services), Mr Vino (Research Administrative Manager), Mr Raisul Sarker (Research Administrative Officer), Ms Anita (Death Review team), Ms Sandra Mitchell (Medico Legal Subpoena), Ms Crystal Matchett (Medical Records team) and Ms Alice (Medical Records team). There are too many of you to name all in the big department, but I do cherish all of you from the bottom of my heart. Thank you so much for providing me with all the essential materials and info, access for the studies and 24-hour technical assistance, including my urgent requests for patient files during the over-time midnights to help through progression. Thanks also for spoiling me with many delicacies during my time in the department. I will always remember all of you and your kind support and help.

Special thanks to the lovely people in RPA Department of Endocrinology, especially Ms Sharon Clibbens and Ms Sylvia Pham who added so much laughter in my course, and organizing so many fantastic Endocrinology seminars throughout the years to enhance my learning.

Sydney Informatics Hub (SIH) of the University of Sydney I would also like to express my utmost appreciation to the people in the Sydney Informatics Hub (SIH), a Core Research Facility of the USyd, who are beyond amazing for the extremely useful technical assistance.

Specifically, Senior Consultant Statistician Mr Jim Matthews for his contagious enthusiasm for advanced biostatistics. Thank you so much for advising me on the advanced modellings in my studies, sharing good statistics books with me, and patiently confirming each of my statistical tests from the development of research protocol all the way to my final statistical analyses and interpretations. His valuable advice has brought clarity to many of my studies.

Acknowledgements | Michelle M.C. Tan

For all his awesome expertise and enthusiasm, I would like to thank the SIH Data Science Group Lead, Dr Gordon McDonald. Thank you very much for all of the time and guidance on my large dataset manipulation, and for going along with my occasional dubious research ventures and efforts in trial and error. The data management for my long-term clinical data combining physicians’ diagnosis, medication use, blood tests and clinical measurements for each comorbidity is not a small task. Because of you, I did it!! Now the impressive changes of statuses of all the comorbidities over long-term follow-up duration were successfully generated and presented in my thesis. Thank you also for his encouragement to step out in trying different statistical software such as STATA and R Studio, apart from the SPSS and GraphPad Prism that I am already familiar with, to achieve the highest standing data management and statistical modellings possible.

Sincere thanks to the eResearch Training Specialist Dr Zhao Jianzhou, Research Data Consultants Dr Taylor Syme and Dr Cameron Fong, and Senior Research Informatics Technical Officer (Visualization) Dr Nathan Butterworth for their tireless support.

I have thoroughly enjoyed working with all of you over the years.

Australian Orthopaedic Association (AOA) and Dr Adrian Low Thanks to the Australian Orthopaedic Association (AOA) for the linked data between my PhD datasets and the Australian Orthopaedic Association National Joint Replacement Registry (AOANJRR).

Not forgotten, Dr Adrian Low (Orthopaedic surgeon and senior clinical lecturer) at the Sydney Adventist Hospital Clinical School and University of New South Wales (UNSW) for his support and discussion of ideas around orthopaedic surgery in making our research a success.

The greatest support and dearest besties in my life

Acknowledgements | Michelle M.C. Tan

Most importantly, utmost appreciations to my most inspiring dearest besties and the greatest supports towards the successful completion of my research, which I would like to express my gratitude to them accordingly. I feel like the luckiest person in the world to have all of them who have been by my side unconditionally throughout my life and during my doctoral studies, I feel really blessed that our years together continue counting - Dr Ng Ooi Chuan, Maw Pin (Prof Dr Tan Maw Pin), Wong sifu (Dr Wong Teck Wee), Wen Lee, KT (Dr Yong Kig Tsuew), mum, eldest brother, granny, Wern Wey, Thomas Lim, Abang Heng (Dr Heng Kiang Soon) and A/Prof Loke Seng Cheong. Thank you so much to all for the endless love, support, guidance, advice, encouragement, motivation, and trust in every step of the way. My life would not be perfect without you.

Second home - The Boden Collaboration, Central Clinical School, Charles Perkins Centre (CPC) I am forever grateful for the family members in the Boden Collaboration, CPC, the second home I have been working at most of the time through the days and nights and weekends, other than my time in the hospitals. It has been a wonderful time with all of the Bodenites. Special thanks to Prof Ian Caterson, my inspiring PhD team advisor, enthusiastic Director of Boden Collaboration as well as the Chair of National Bariatric Surgery Registry Steering Committee. His ambition and vision have set a great example for me.

The amazing Boden administrative staffs, Ms Kristine Maddock, Ms Joyce Calvitto and Ms Melanie Symons - thank you so much for being such a lovely, wonderful and caring family members, really appreciate all of your guidance, advice, support, company, birthday presents and help these years.

To other Boden family members, I really enjoy being with all of you. Especially Ali (Alice Meroni), Gail Del Olmo, Jason Jin, Jia (Reeja Nasir), Alison Coenen, Ang Li, Claudia, Radhika, Melinda, Hamish, Mackenzie Fong, Felipe, Prof Michael Skilton, Kyra Sim, Elif, Shuang, Emma, Prof Stephen Colagiuri, Tara, Tegan, Chelsea, Haley, Talia, Anthony, Andrea, Erica, Michelle Hsu, Jessica Burk, Jovana, Prof Timothy Gill, Eyza and Rachel. Thank you for making my second home such a

vii

Acknowledgements | Michelle M.C. Tan wonderful one. Being with you all have been the highlight of my doctoral studies and I will forever have wonderful memories at the Boden.

Liaison librarians of the University of Sydney (USyd) Not forgotten, sincere appreciation to the USyd’s Liaison librarians Mr Rod Dyson and Ms Elaine Tham for all your guidance and assistance in my systematic reviews. Thanks for teaching me patiently how to perform medical database searches when I was new to systematic reviews, and for constantly helping to thoroughly confirm my multiple database searches.

Charles Perkins Centre (CPC) Royal Prince Alfred (RPA) Clinic Special thanks to the CPC RPA Clinic staffs for their kind support throughout my data collection years and dual-energy X-ray absorptiometry (DEXA) training, including the lovely Manager Ms Cathy (Catherine Yates); my lunch mates Pauline, David and Nadine; and Sally.

Other collaborators in the PhD study I would also like to acknowledge my associate collaborators, including the team Bariatric Surgeons who are always positive and encouraging, and who operated on my study patients ‒ Dr David Martin (RPAH and CRGH), Dr Craig Taylor (CRGH), Dr Philip Le Page (CRGH) and Dr David Joseph (RPAH and CRGH); and the RPAH’s Metabolism and Obesity Services clinic staff ‒ Nurse Manager Ms Elisia Manson, and dietitians Dr Janet Franklin and Ms Gabrielle Maston. Thanks for being both patient and helpful with my research requests and clarification seeking.

Dr Craig Wood, Dr Shih-Chiang Shen and Jon Peck Thank you so much to Dr Craig Wood (Geisinger Health System, Obesity Institute, USA) and Dr Shih-Chiang Shen (Shang-Ho Hospital, New Taipei City, Taiwan) for their kind sharing and guidance on the DiaRem algorithm paper in this thesis. Thank you so much Jon Peck, the book author of “SPSS Statistics for Data Analysis and Visualization” for his kind guidance and advice on the advanced modellings in my research.

viii

Acknowledgements | Michelle M.C. Tan

Other friends and support I am also especially appreciative to other lovely friends I have met here, including Fiona Thien, Geraldine Eggenhuizen and Brenda Nguyen from the Sydney Adventist Hospital Clinical School; my neighbours; SUAMS former President Ying Yi; and all the kind friends from the Global Tzu Chi Compassion Buddhist Foundation (Sydney branch) with whom I built forever-friendships. There are too many of you to list, but you know who you are. It is incredibly warm, happy and inspiring to have all of you. Thank you.

PhD Thesis – Main Abstract | Michelle M.C. Tan

ABSTRACT

Background: Obesity, a National Health Priority Area, has been a major challenge facing the Australian population and is at a crisis level. It has also become a global epidemic with a disproportionate rise in class III obesity (BMI ≥40 Kg/m2), causing substantial burden in obesity and the healthcare systems. There are approximately one million adults in Australia with clinically severe obesity, defined as class III obesity alone or a BMI ≥35 Kg/m2 with at least one major obesity-related comorbidity. Bariatric surgery is well-established as the most effective treatment for severe and complex obesity that has been unmanageable by other modalities. Despite most Australians relying on the public healthcare system, the vast majority of bariatric surgical procedures are performed in private hospitals owing to scarce resources in the public sector. As a result of this reason and sparse research funding, there is a limited understanding of the management, benefits and safety of publicly funded bariatric surgery in the context of clinically severe obesity and its metabolic consequences, especially in long-term (defined by >5 years).

Settings: A multidisciplinary publicly funded bariatric surgery service covering three public hospitals, namely Royal Prince Alfred Hospital, Concord Repatriation General Hospital and Camden Hospital in Sydney, New South Wales (NSW), Australia.

PhD Thesis – Main Abstract | Michelle M.C. Tan

Data linkage with the Australian Orthopaedic Association National Joint Replacement Registry (AOANJRR) was also performed.

Aims and research themes: This thesis aims to address key knowledge deficiencies in a broad range of areas encompassing the full spectrum of bariatric surgery in multidisciplinary clinical obesity services, using both retrospective and prospective study designs. These studies were grouped into four major research themes, with each theme representing one chapter in the thesis. The research themes investigated were: (1) long-term health outcomes after bariatric surgery, (2) adherence to multidisciplinary post-operative follow-up care, (3) whether pre-operative weight loss predicts post-surgical weight loss, and (4) prediction of diabetes remission using an algorithm.

Main findings: The first research theme (CHAPTER 2) is a retrospective cohort study that examined the long- term effectiveness and safety of bariatric surgery in a highly-complex clinically severe obese population. While the overall weight loss and changes of obesity-related comorbidities [especially type 2 diabetes mellitus (T2DM), hypertension and hyperlipidaemia] were significant following bariatric surgery over 6 years; being super obese (BMI ≥50 Kg/m2) had no multiplicative and detrimental effect in these features compared to the morbidly obese patients (<50 Kg/m2). The prevalence of the obesity-related comorbidities studied [i.e. T2DM, hypertension, osteoarthritis (OA) and/or weight-bearing joint pain (WBJP), sleep-disordered breathing and hyperuricaemia] decreased after bariatric surgery and remained lower than baseline as time progressed; whereas hyperlipidaemia and mental illness surpassed baseline level at post-operative 6 years. For nutrient deficiency, vitamin D deficiency was noted in one-third of patients pre-operatively and decreased significantly after bariatric surgery. Iron deficiency anaemia doubled at year 6 post- operation. Low prevalence of vitamin B12 insufficiency was detected before and after bariatric surgery, with no patient developing a deficiency in years 5 and 6 post-operatively. Of the 34.5% with peri- and post-operative complications, none was life-threatening.

The second research theme (CHAPTER 3) is a prospectively conducted study that investigated the reasons for ceasing attendance at clinic reviews after surgery, predictors of adherence to post- bariatric surgery clinic reviews, and the relationship – if any – between adherence to follow-up and weight loss outcomes. The adherence rate to follow-up visits after bariatric surgery (i.e. patients attended follow-up regularly) among the study patients was 63.7% [107 of 168 (63.7%)]; 20 (11.9%) attended irregularly and 41 (24.4%) ceased attending reviews. According to the patients, withdrawal from the publicly funded bariatric surgery service was mainly associated with travel distance. Linear mixed-effects model with random effects revealed no pronounced difference in the mean weight loss between the adherent and nonadherent groups (composite of those who attended irregularly or who ceased attending follow-up) over the years. Logistic regression model shows that older and partnered patients were more likely to adhere to follow- up care after operation. The results could guide care practices for patients needing additional contacts and supports, so they may benefit from additional assistance to experience optimal outcomes.

PhD Thesis – Main Abstract | Michelle M.C. Tan

Subsequent studies (CHAPTER 4) explored the relationship between pre-operative weight loss and weight loss post-bariatric surgery, to determine the necessity of the current requirement for weight loss prior to publicly funded bariatric surgery in the future. The result shows insignificant relationship between these two parameters, suggesting that the pre-operative weight management program (WMP) that is currently-mandated prior to the surgery may not be necessary. However, the WMP might still be important as an opportunity to resolve medical problems; to prevent post-operative psychological issues from emerging; to ensure patients understand the implications of bariatric surgery and its necessary lifestyle changes; and to minimise potential surgical risks. Multiple linear regression analysis demonstrated that age at surgery is a reliable predictor of post-operative weight loss across years 1 to 6 post-operation, suggesting that older patients may achieve better outcomes from bariatric surgery.

The fourth research theme (CHAPTER 5) identified short- and longer-term post-operative diabetes remission prediction following bariatric surgery using the DiaRem scoring system that computed by the following simple variables – age, HbA1c, glucose lowering treatment other than insulin, and insulin treatment. The DiaRem algorithm was shown to perform well in discriminative capacity, predictive ability and calibration in the study cohort with T2DM. The area under the curve (AUC) of the receiver operator characteristic was 0.869 (95% CI=0.800–0.938) for the standard year 1 post-operative follow-up, with the most optimal cut-off score being ≤12, sensitivity of 94.8% and specificity of 64.2%. The discriminative ability is comparatively high for those with 5-year post-operative diabetes remission [AUC=0.835 (95% CI=0.714–0.956), optimal cut-off score ≤12, sensitivity=89.5% and specificity=52.2%]. Logistic regression models demonstrated that DiaRem algorithm reliably predicted diabetes remission 1 year [OR (95% CI)=0.733 (0.655-0.821), p<0.001] and 5 years following surgery [OR (95% CI)=0.753 (0.623- 0.909), p=0.003]. The Hosmer-Lemeshow goodness-of-fit tests indicated good fit of DiaRem prediction models for both 1 year and 5 years post-surgery, thus accurate models to use. This study confirmed that DiaRem is a useful and practical tool to help clinicians with selection and prioritisation of patients with T2DM and seeking bariatric surgery in publicly funded models.

Conclusion: Collectively, this thesis highlights the unique nature of the multidisciplinary surgical management of clinically severe obesity in three specialist obesity services in public hospitals within the carefully-studied cohort. It supported the high-risk and well-characterised clinically severe obese population with long-term effective and safe access to multidisciplinary publicly funded bariatric surgery service and research. These studies significantly contributed to a transparent and improved understanding of the management in the context of complex obesity, including those with super obesity and multiple obesity-related comorbidities. Ultimately, these findings can enhance clinical practice by providing evidence-based knowledge and specific tools for the management of multidisciplinary bariatric surgery in the growing population with clinically severe obesity.

xii

List of Awards | Michelle M.C. Tan

AWARDS

1) Postgraduate Research Support Scheme (PRSS) 2020 Funded by: The University of Sydney 2) Charles Perkins Centre Early- and Mid-Career Researcher (EMCR) SEED Grant Award 2019 Funded by: Charles Perkins Centre, The University of Sydney 3) The Boden Institute Travel Support Scheme 2018 Funded by: The Boden Collaboration, Central Clinical School, Charles Perkins Centre, The University of Sydney 4) Postgraduate Research Support Scheme (PRSS) 2018 Funded by: The University of Sydney 5) Postgraduate Research Support Scheme (PRSS) 2017 Funded by: The University of Sydney 6) Research Training Program (RTP) Fee Offset and Stipend Scholarships 2017-2020 Funded by: Australian Commonwealth Government 7) Sancta Sophia College Entrance Scholarship 2017 Funded by: Sancta Sophia College, The University of Sydney

xiii List of Conference Presentations | Michelle M.C. Tan

CONFERENCE PRESENTATION ABSTRACTS

Conference presentations arising from this thesis are:

Oral presentation

1) Joint Scientific Meeting of Australian and New Zealand Obesity Society (ANZOS), Obesity Surgery Society of Australia and New Zealand (OSSANZ) and Asia Oceania Association of Studies for Obesity (AOASO) 2017 Title: Is bariatric surgery safe and effective for patients with super obesity? Michelle Tan, Samantha Hocking, Amanda Sainsbury, Danny Kim, Caro Badcock, Elisia Manson, Nick Fuller, Nic Kormas, David Martin, Craig Taylor, David Joseph, Philip Le Page, Ian Caterson, and Tania Markovic 4-6 October 2017, Adelaide, Australia.

2) Joint Scientific Meeting of ANZOS, OSSANZ and AOASO 2017 Title: Does attendance at follow-up post bariatric surgery affect outcomes? Michelle Tan, Amanda Sainsbury, Samantha Hocking, Samuel Barclay, Jan Paul de Bruin, Caro Badcock, Elisia Manson, Nick Fuller, Nic Kormas, David Martin, Craig Taylor, David Joseph, Phillip Le Page, Ian Caterson, and Tania Markovic 4-6 October 2017, Adelaide, Australia.

3) 6th Charles Perkins Centre (CPC) Early- and Mid-Career Researcher (EMCR) Symposium 2020 Title: Multidisciplinary surgical management of patients with clinically severe obesity in public hospitals in New South Wales (NSW): Long-term weight trajectories, health outcomes and surgical complications Michelle Tan Supervisors/Core Collaborators: Samantha Hocking, Tania Markovic, Amanda Sainsbury, Nic Kormas, Ian Caterson, Jason Jin, Phillip Le Page, David Martin, Craig Taylor, and David Joseph. 30 October 2020, Sydney, Australia.

xiv

List of Conference Presentations | Michelle M.C. Tan

Poster presentation

1) European Congress on Obesity 2018 Title: Is bariatric surgery safe and effective for patients with super obesity? Michelle Tan, Amanda Sainsbury, Samantha Hocking, Nic Kormas, Janet Franklin, Elisia Manson, Caro-Anne Badcock, David Martin, Craig Taylor, David Joseph, Philip Le Page, Nick Fuller, Danny Kim, Ian Caterson, and Tania Markovic Published in Obesity Facts – The European Journal of Obesity, 11 (Suppl 1), 1-358. 23-26 May 2018, Vienna, Austria.

2) European Congress on Obesity 2018 Title: What happens to patients who do not attend follow-up after bariatric surgery? Michelle Tan, Amanda Sainsbury, Samantha Hocking, Nic Kormas, Janet Franklin, Elisia Manson, Caro Anne Badcock, David Martin, Craig Taylor, David Joseph, Philip Le Page, Jan Paul de Bruin, Samuel Haig Barclay, Nick Fuller, Ian Caterson, and Tania Markovic Published in Obesity Facts – The European Journal of Obesity, 11 (Suppl 1), 1-358. 23-26 May 2018, Vienna, Austria.

3) European Congress on Obesity 2018 Title: Does weight loss before bariatric surgery matter? Michelle Tan, Amanda Sainsbury, Samantha Hocking, Janet Franklin, Elisia Manson, Nic Kormas, Caro-Anne Badcock, David Martin, Craig Taylor, David Joseph, Philip Le Page, Nick Fuller, Ian Caterson, and Tania Markovic Published in Obesity Facts – The European Journal of Obesity, 11 (Suppl 1), 1-358. 23-26 May 2018, Vienna, Austria.

4) Joint Conference of the Australian & New Zealand Gastro and Oesophageal Surgery Association (ANZGOSA)/Australian & New Zealand Metabolic and Obesity Surgery Society (ANZMOSS) 2019 Title: Long-term safety and efficacy of single-stage bariatric operations in patients with

List of Conference Presentations | Michelle M.C. Tan

clinically super-obesity in tertiary public hospitals in New South Wales (NSW) Michelle Tan, Nic Kormas, Samantha Hocking, Amanda Sainsbury, Ian Caterson, Ang Li, David Martin, Craig Taylor, David Joseph, Philip Le Page, and Tania Markovic 2-4 October 2019, Queensland, Australia.

xvi

Table of Contents | Michelle M.C. Tan

TABLE OF CONTENTS

DECLARATION OF ORIGINALITY………………………………………………………………………………………………i ACKNOWLEDGMENTS…………………………………………………………………………………………………………...ii ABSTRACT……………………………………………………………………………………………………….……………………..x AWARDS……………………………………………………………………………………………………………………………..xiii CONFERENCE ABSTRACTS ARISING FROM THIS THESIS………………………………………………………..xiv TABLE OF CONTENTS…………………………………………………………………………………………………..……..xvii LIST OF ABBREVIATIONS…………………………………………………………………………………..………………..xxii LIST OF FIGURES……………………………………………………………………………………………..……….………..xxvi LIST OF TABLES………………………………………………………………………………………………..………………xxviii

CHAPTER 1 INTRODUCTION AND LITERATURE REVIEW…………………………………………….. 1 1.1 THESIS OVERVIEW….……………………………………………………………………………………………....………. 1 1.2 OBESITY………………………………………………………………………………………….……………………………… 3 1.2.1 Definition and diagnostic criteria………………………………….…………………………………….. 4 1.2.2 Obesity-related comorbidities……………..…………………………………………………………….. 8 1.2.3 The epidemic of obesity……………………………………………………………………………………… 11 1.2.4 Aetiology of obesity……………………………………………………………………………………………. 16 1.2.5 Overview of management of obesity………………………………………………………………..… 16 1.3 NON-SURGICAL TREATMENT OF OBESITY……………………………………………………………………………. 17 1.3.1 Changes in diet and exercise………………………………………………………………………….…... 17 1.3.2 Very low energy diet (VLED)……………………………………………………………………………….. 19 1.3.3 Psychological intervention….………………………………………………………………………………. 22 1.3.4 Pharmacotherapy…………………………………………………………………………………………..….. 22 1.4 BARIATRIC SURGERY……………………………………………………………………………………………………..…. 26 1.4.1 Type of bariatric surgical procedures………………………………………………………………….. 26 1.5 OUTCOMES OF BARIATRIC SURGERY………………………………………………………………………………….. 38 1.5.1 Effectiveness of bariatric surgery…………………………………………………………………..……. 38 1.5.2 Can surgical outcomes be predicted?...... 53 1.6 PUBLICLY FUNDED BARIATRIC SURGERY……………………………………………………………………………... 56 1.7 FOLLOW-UP AFTER BARIATRIC SURGERY…………………………………………………………………………..… 57 1.7.1 Definitions of adherence to follow-up……………………………………………………………….. 58 1.7.2 Importance of adherence to follow-up after bariatric surgery…………………………… 59 1.7.3 Predictors of and reasons for loss to follow-up………………………………………………….. 60

xvii

Table of Contents | Michelle M.C. Tan

1.8 PRE-OPERATIVE CARE/WEIGHT LOSS AND ITS IMPLICATIONS ON POST-SURGERY WEIGHT LOSS………………………………………………………………………………………………………………………………. 61 1.9 RATIONALES FOR THE RESEARCH DIRECTIONS AND AIM: BRIDGING THE KNOWLEDGE GAPS, CLINICAL NEEDS AND RESEARCH CONTRIBUTIONS……………..………….. 64 RESEARCH THEME 1: Long-term effectiveness and safety of bariatric surgery in clinically severe obesity and its associated comorbidities………….…. 65 RESEARCH THEME 2: Adherence to post-bariatric surgery care……………...... …………….… 67 RESEARCH THEME 3: Pre-operative weight loss as a requirement for publicly funded bariatric surgery……………………………………………………………………..……. 68 RESEARCH THEME 4: Prediction of T2DM remission in bariatric surgical patients…………. 68

CHAPTER 2 LONG-TERM WEIGHT TRAJECTORIES, HEALTH OUTCOMES AND SURGICAL COMPLICATIONS AFTER BARIATRIC SURGERY OF PUBLICLY FUNDED PATIENTS WITH CLINICALLY SEVERE OBESITY……………………………………….. 72 2.1 ABSTRACT…………………………………………………………………………………………………………………….....72 2.2 BACKGROUND…………………………………………………………………………………………………………….……76 2.3 MATERIALS AND METHODS………………………………………………………………………………………….....94 2.4 RESULTS………………………………………………………………………………………………………………………...110 2.5 DISCUSSION………………………………………………………………….…………………………………………….. 153 2.6 CONCLUSION………………………………………………………………………………………………………………. 178

CHAPTER 3 ADHERENCE TO MULTIDISCIPLINARY FOLLOW-UP CARE AFTER BARIATRIC SURGERY IN AN AUSTRALIAN PUBLICLY FUNDED HEALTHCARE SERVICE: PREDICTORS, REASONS FOR LOSS TO FOLLOW-UP AND OUTCOMES…... 180 3.1 ABSTRACT……………………………………………………………………………………………………………..….… 180 3.2 BACKGROUND…………………………………………………………………………………………………………..… 181 3.3 MATERIALS AND METHODS………………………………………………………………………………………... 183 3.4 RESULTS……………………………………………………………………………………………………………………... 188 3.5 DISCUSSION………………………………………………………………………………………………………………... 198 3.6 CONCLUSION…………………………………………………………………………………………………………….… 202

CHAPTER 4 QUALIFYING FOR FUNDED-BARIATRIC SURGERY IN PUBLIC HOSPITALS: DOES PRE-OPERATIVE WEIGHT LOSS MATTER?...... 204 4.1 ABSTRACT………………………………………………………………………………………………………………...… 204 4.2 INTRODUCTION………………………………………………………………………………………………………….. 205 4.3 METHODS……………………………………………………………………………………………………………….….. 207 4.4 RESULTS…………………………………………………………………………………..……………………….………… 212 4.5 DISCUSSION………………………………………………………………………………………………………..………. 217 4.6 CONCLUSION…………………………………………………………………………………………………………..….. 220

xviii

Table of Contents | Michelle M.C. Tan

CHAPTER 5 PREDICTION OF 1-YEAR AND 5-YEAR DIABETES REMISSION FOLLOWING SLEEVE GASTRECTOMY, MINI GASTRIC BYPASS-ONE ANASTOMOSIS GASTRIC BYPASS AND ADJUSTABLE GASTRIC BANDING USING DIAREM SCORE………………….……..221 5.1 ABSTRACT……………………………………………………………………………………………………………….….. 221 5.2 BACKGROUND………………………………………………………………………………………………………….…. 222 5.3 RESEARCH DESIGNS AND METHODS…………………………………………………………………………... 224 5.4 RESULTS……………………………………………………………………………………………………………………... 230 5.5 DISCUSSION……………………………………………………………………………………………..…………………. 237 5.6 CONCLUSION………………………………………………………………………………………………………….…… 242

CHAPTER 6: CONCLUDING REMARKS…………….…………………………………….………………….. 243 6.1 Main findings and clinical implications……………………………………………………………………….. 244 6.2 Summary, general limitations and future directions………………………………………………..… 250

REFERENCES………………………………………………………………………………………………………………………… 255 APPENDIX Post-bariatric surgery questionnaire……………...……………………………………………….. 284

xix

List of Abbreviations | Michelle M.C. Tan

ABBREVIATIONS

Abbreviation Full form %EWL Percentage of excess weight loss %TWL Percentage of total weight loss ACE inhibitors Angiotensin-converting enzyme inhibitors ACS Acute coronary syndrome ACT Australian Capital Territory ADA American Diabetes Association ADS Australia Diabetes Society AGB Adjustable gastric banding AHI Apnoea-hypopnoea index AIC Akaike’s Information Criterion AIHW Australian Institute of Health and Welfare ANZMOSS Australian & New Zealand Metabolic & Obesity Surgery Society ANZOS Australian & New Zealand Obesity Society AOANJRR Australian Orthopaedic Association National Joint Replacement Registry ASMBS American Society for Metabolic and Bariatric Surgery AUC Area Under the Curve BCa Bias-corrected and accelerated BED Binge eating disorder BiPAP Bilevel positive airway pressure BMD Bone mineral density BMI Body mass index BP Blood pressure BPD/DS Biliopancreatic diversion with duodenal switch CAD Coronary artery disease CBT Cognitive behavioural treatment CKD Chronic kidney disease cm centimetres CI Confidence intervals COSiPH Clinical Obesity Services in Public Hospitals COSM Canberra Obesity Service Management CPAP Continuous positive airway pressure CT Computed tomography CVD Cardiovascular disease DEXA Dual-energy X-ray absorptiometry DiRECT Diabetes Remission Clinical Trial DPP4 inhibitors Dipeptidyl peptidase 4 inhibitors DM Diabetes mellitus DVT Deep vein thrombosis DRS Diabetes Remission Score EASO European Association for the Study of Obesity

xxii

List of Abbreviations | Michelle M.C. Tan eMR Electronic medical record EOSS Edmonton Obesity Staging System FBG Fasting blood glucose FDA Food and Drug Administration FFA Free fatty acids FY Financial year GBD Global Burden of Disease GFR Glomerular filtration rate GLP-1 receptor Glucagon-like peptide-1 receptor agonists agonists GORD Gastro-oesophageal reflux disease HbA1c Glycated haemoglobin HDL-C High-density lipoproteins cholesterol IARC International Agency for Research on Cancer IFSO International Federation for the Surgery of Obesity and Metabolic Disorders IGFs Insulin-like growth factors IHD Ischaemic heart disease IMS Individualized Metabolic Surgery IU International Units JAMA Journal of the American Medical Association Kg Kilograms Km Kilometres LABS study Longitudinal Assessment of Bariatric Surgery study LOHS Length of hospital stay Look AHEAD trial Look Action for Health in Diabetes trial LDL-C Low-density lipoproteins cholesterol LAGB Laparoscopic adjustable gastric banding LSG Laparoscopic sleeve gastrectomy LTFU Loss to follow-up m Metres MA Meta-analysis MCAR Missing data completely at random MGB-OAGB Mini gastric bypass-one anastomosis gastric bypass MI Myocardial infarction mm millimetres mmol/L millimoles per litre MO Morbid obesity MRN Medical record number MRI Magnetic resonance imaging NAFLD Non-alcoholic fatty liver disease NASH Non-alcoholic steatohepatitis NHANES National Health and Nutrition Examination Survey NSW New South Wales

xxiii

List of Abbreviations | Michelle M.C. Tan

OA Osteoarthritis OACs Obesity-associated cancers OHA Oral hypoglycaemic agent OHS Obesity hypoventilation syndrome OR Odds Ratio OSA Obstructive sleep apnoea OXM Oxyntomodulin PCOS Polycystic ovary syndrome PE Pulmonary embolism PORs Pooled odds ratios PTSD Post-traumatic stress disorder PYY Peptide YY QoL Quality of life RAS Renin-angiotensin system RCT Randomized controlled trial RES Respiratory events scoring ROC Receiver Operating Characteristic RR Relative Risk RYGB Roux-en-Y gastric bypass SAT Subcutaneous adipose tissue SG Sleeve gastrectomy SLEEVEPASS trial Sleeve vs Bypass trial SLHD Sydney Local Health District SM-BOSS Swiss Multicentre Bypass or Sleeve Study SO Super obesity SOReg study Scandinavian Obesity Surgery Registry study SOS study Swedish Obesity Subjects study SD Standard deviation SE Standard error SGLT2 inhibitors Sodium-glucose transport protein 2 inhibitors STAMPEDE study Surgical Treatment and Medications Potentially Eradicate Diabetes Efficiently study SWSLHD South Western Sydney Local Health District T2DM Type 2 diabetes mellitus TGA Therapeutic Goods Administration THA Total hip arthroplasty TKA Total knee arthroplasty TJA Total joint arthroplasty TMR Total meal replacements TZD Thiazolidinedione UK United Kingdom US United States USA United States of America

xxiv

List of Abbreviations | Michelle M.C. Tan

VAT Visceral adipose tissue VLED Very low energy diet VLCD Very low calorie diet WBJP Weight-bearing joint pain WC Waist circumference WHO World Health Organization WMP Weight management program

Authors' initials and the corresponding full names: Initials Full name AL Adrian Low AS Amanda Sainsbury/Amanda Salis CT Craig Taylor DJ David Joseph DM David Martin MT Michelle Tan NK Nic Kormas PL Philip Le Page SH Samantha Hocking TM Tania Markovic

xxv

List of Figures | Michelle M.C. Tan

LIST OF FIGURES

Figure number Subtitle Figure 1.1 Top-10 global causes of death in the year 2016 Figure 1.2 Comorbidities associated with obesity Figure 1.3(a) Age-standardised prevalence of (i) overweight and (ii) obesity in adults over 20 years old by geographical region and year (1980–2015) Figure 1.3(b) Age-standardised prevalence of obesity worldwide (174 countries) in 2015 among (A) men and (B) women aged over 20 years old Figure 1.4 Age-standardised global prevalence of (a) overweight and (b) obesity in men and women over 20 years old by year, from 1980 to 2015 Figure 1.5 Distribution of body mass index (BMI) in adults aged 18 and over in Australia, 1995 and 2014-2015, demonstrating the growing obesity trend in Australia Figure 1.6 Bariatric procedures performed in Australia and New Zealand (ANZ) in years 2014/15–2018/19 according to the ANZ Bariatric Surgery Registry Figure 1.7 Pictorial representation of sleeve gastrectomy (SG) Figure 1.8 Mechanism of action of sleeve gastrectomy (SG) Figure 1.9 Pictorial representation of mini gastric bypass-one anastomosis gastric bypass (MGB-OAGB) Figure 1.10 Pictorial representation of laparoscopic adjustable gastric banding (LAGB) Figure 1.11 Summary of the mechanism of action of adjustable gastric banding (AGB) Figure 1.12 Pictorial representation of the Roux-en-Y gastric bypass (RYGB) Figure 1.13 Summary of mechanism of action of Roux-en-Y gastric bypass (RYGB) Figure 1.14 Mechanisms underlying cancer promotion and development induced by obesity Figure 1.15 Conceptual framework and research themes for the associated studies in this thesis Figure 2.1 Distribution of the study patients with clinically severe obesity in each BMI- obesity category at pre-operative baseline (n=168) Figure 2.2 Edmonton Obesity Staging System (EOSS) Staging tool Figure 2.3 Observed mean weight change from initial clinic visit (lifestyle modifications) to pre-operative baseline through 8-year post-operative follow-up Figure 2.4 Observed mean and median BMI change from initial clinic visit (lifestyle modifications) to pre-operation to Year 8 of bariatric surgery Figure 2.5 Overall marginal estimated mean (a) weight (Kg) and (b) BMI loss (Kg/m2) at each visit over time modelled from the mixed-effects model taking into account the repeated measures, adjusted for age at time of surgery, sex and race Figure 2.6 Comparisons of modelled weight change (Kg) between the SO and MO groups at pre-operative baseline and at years 1, 2, 3, 4, 5 and 6 following bariatric surgery

xxvi

List of Figures | Michelle M.C. Tan

Figure 2.7 Comparison of estimated marginal mean differences of post-surgical percent total weight loss (%TWL) of the pre-surgery weight between SO and MO groups by clinic follow-up visits adjusted for age at surgery, sex and race Figure 2.8 Distributions of Edmonton Obesity Staging System (EOSS) stages among the publicly funded bariatric surgery service patients (n=168) Figure 2.9 Prevalence of obesity-related comorbidities pre-operatively and at 1, 2, 3, 4, 5 and 6 years following bariatric surgery (%) Figure 2.10 Yearly remission, improved, persisting and worsened rates of T2DM following bariatric surgery Figure 2.11 Yearly remission, improved, unchanged and worsened rates of hypertension following bariatric surgery Figure 2.12 Yearly remission, improved, persisting and worsened rates of hyperlipidaemia following bariatric surgery Figure 2.13 Comparison of SO and MO groups on the yearly changes of remission, (a)–(c) improvement, persisting and worsening rates for the selected obesity-related comorbidities post-operatively Figure 2.14 Prevalence of sleep-disordered breathing (OSA/OHS) and CPAP/BiPAP prescription among the study population undergoing bariatric surgery Figure 2.15 Prevalence of depression and/or severe anxiety alongside the antidepressants and/or antianxiety agents use among the study population undergoing bariatric surgery Figure 2.16 Prevalence prescribed opioid use and total joint arthroplasty (TJA) among the patients with OA and/or weight-bearing joint pain (WBJP) Figure 3.1 Definition of ‘Adherence’ or regularly attending follow-up care post- operatively Figure 3.2 Flow diagram of patient tracking Figure 3.3 Comparison of estimated marginal mean differences of post-surgical weight from pre-operative baseline (Kg) between Adherent and Nonadherent groups by clinic visits adjusted for age at surgery, sex and race Figure 4.1 Observed mean weight change (Kg) from initial clinic visit to pre-operative baseline through 8-year post-operative follow-up Figure 5.1 Flow chart describing the patient selection strategy for the study cohort Figure 5.2 Distribution of the DiaRem scoring system Figure 5.3(a) Area under the Receiver Operating Characteristic (ROC) curve (AUC) for diagnostic value of DiaRem score as a predictor of T2DM remission at 1-year post bariatric surgery (n=111) Figure 5.3(b) Area under the Receiver Operating Characteristic (ROC) curve (AUC) for DiaRem score in predicting T2DM remission at 5-year post bariatric surgery (n=42)

xxvii

List of Tables | Michelle M.C. Tan

LIST OF TABLES

Table number Subtitle Table 1.1 World Health Organization (WHO) definitions of body weight disorders in adults by body mass index (BMI) Table 1.2 Common weight loss drugs Table 1.3 DiaRem scores Table 1.4 ABCD score Table 1.5 Distributions of bariatric procedures stratified by public and private hospitals Table 1.6 Study outline for 10 years of pre- and post-operative follow-up Table 2.1 Studies reporting on bariatric surgery outcomes Table 2.2 Edmonton Obesity Staging System (EOSS) Staging tool Table 2.3 Baseline socio-demographics of study cohort by primary procedures Table 2.4 Clinical profiles of all the study patients pre- and post-bariatric surgery (n=168) Table 2.5 Annual remission, improvement, persistence and aggravation rates of obesity-related comorbidities after bariatric surgery between the SO and MO groups over 6 years of follow-up Table 2.6 Prevalence of hyperuricaemia Table 2.7 Peri-operative and late post-operative complications by bariatric procedures (n=168) Table 2.8 Prevalence of nutrient deficiencies pre- and post-bariatric surgery Table 2.9 Mean values of nutritional parameters pre- and post-bariatric surgery Table 2.10 Characteristics of study patients who died in-hospital (n=3) Table 3.1 Baseline characteristics of patients and the relationship with follow-up adherence after bariatric surgery (n=168) Table 3.2 Baseline obesity-related comorbidities and mental illness stratified by adherence to follow-up appointments Table 3.3 Final multivariate logistic regression modelling for predictors of adherence to postoperative follow-up Table 3.4 Reasons for withdrawal from the publicly funded bariatric surgery service or loss to follow-up to scheduled post-operative aftercare appointments (n=41) Table 4.1 Potential advantages and disadvantages of pre-operative weight management program (WMP) Table 4.2 Baseline characteristics of the study cohort (n=168) Table 4.3 Initial and pre-operative baseline obesity-related comorbidities of all patients Table 4.4 Summary of the final prediction models of the percentage total weight loss (%TWL) at 1 to 7 years after bariatric surgery. Table 5.1 DiaRem scoring system Table 5.2 Socio-demographic and clinical characteristics of publicly funded patients with type 2 diabetes mellitus (T2DM) who underwent bariatric surgery (n=114) Table 5.3 Proportions of patients with T2DM remission according to their DiaRem score categories Table 5.4 Binary logistic regression models of T2DM remission at 1 and 5 years after bariatric surgery with DiaRem score as the only independent variable

xxviii

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan

CHAPTER 1

INTRODUCTION AND LITERATURE REVIEW

1.1 Thesis Overview

This PhD thesis centres around a multidisciplinary publicly funded bariatric surgery management based at three specialist obesity clinics at three public hospitals in Sydney, Australia, and aimed to address the clinically severe obesity epidemic and its associated burden from a broad range of perspectives in a well‐defined adult bariatric surgical population. It is also aimed at improving the surgical management and existing care in the public hospitals and Australian healthcare system. In view of the highly complex nature of clinically severe obesity and its impact on different aspects of patient’s life, a multidisciplinary team is well‐placed to manage the multifactorial interplay of the obesity, associated comorbidities, and their chronic relapsing and remitting course. The physician‐led multidisciplinary approach in our publicly funded bariatric surgery service across two local health districts in the state of New South Wales (NSW) include medical directors, consultant endocrinologists, bariatric surgeons, clinical nurse consultants, dietitians, exercise physiologists, physiotherapists and clinical psychologists. This has enabled substantial collaboration between our team from multiple health disciplines with diverse skills, to treat all aspects of bariatric surgical patients’ obesity and comorbidities that will be systematically covered in the thesis. This thesis is divided into six chapters over four major research themes, as follows:

CHAPTER 1: Introduction and Literature Review This chapter provides an overview of the background and evidence for bariatric surgical treatment, obesity, obesity‐related metabolic and systemic consequences, the burden of these diseases, feasible tools in practice, and clinical management. This is followed by descriptions of the rationales for research directions, and the development of the research themes and objectives. These contextualise the key area of knowledge gaps identified within the review of

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan literature and clinical observations. The clinical needs and knowledge gaps were then framed against the research opportunities provided by bariatric surgery in the setting of clinically severe obesity; individual aims of each thesis chapter were specified; the unique research themes were highlighted and further connected with feasible methodologies.

CHAPTERS 2 to 5: Thesis Studies The studies in the thesis that were framed into the following four main research themes are presented as individual chapters in the thesis: 1. Long‐term effectiveness, health impacts and inherent risks of bariatric surgery in the setting of clinically severe obesity 2. Adherence to post‐bariatric surgery care 3. The role of pre‐operative weight loss 4. Prediction of type 2 diabetes mellitus (T2DM) remission using an algorithm, and its feasibility as a diagnostic tool in everyday practice in the bariatric surgical population

These studies were conceived and carried out during the PhD tenure of 3.5 years. The roles of the PhD candidate/researcher (MT) included conceptualisation of the study, research protocol development, systematic literature review, project planning, preparation of patient questionnaires and postage‐paid envelopes, ethics amendment application, research funding application, coordination and implementation of all big and small milestones, database creation, collation of full list of patients and their identifiers, data collection in three participating hospitals, patient contacting and interviews with team consultant physicians, sending questionnaires and letters to patients, data entry, data quality monitoring and audit, data linkage with a national surgical registry, data cleaning, data analysis, result interpretation, writing, reporting, and dissemination of research outputs. Throughout the study, the PhD candidate/researcher was also responsible for collaborating with a wide range of stakeholders in managing the research deliverables. They included PhD supervisors, multidisciplinary team clinicians, bariatric surgeons, orthopaedic surgeon, hospital administrative staff, Medical Record Department staff, internal and external academics, Ethics and Governance, National Bariatric Surgery Registry, Australian

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan

Orthopaedic Association National Joint Replacement Registry (AOANJRR), death review team, Medico Legal Subpoena, Department of Forensic Medicine of NSW, and NSW State Coroner.

The specific retrospective and prospective study designs, methods and resources adopted in this thesis will be detailed within each chapter representing each study. Generally, this longitudinal study involved collection of repeated measures of clinical, biochemical, nutritional, surgical, lifestyle behaviour and patient management data before and after bariatric surgery over a 10‐ year follow‐up period.

CHAPTER 6: Concluding Remarks Finally, the conclusions are presented, with a critical discussion of direct translation of the findings into optimised clinical practice and future research directions to further enhance understanding of clinically severe obesity, its associated diseases, bariatric surgical treatment, orthopaedic surgery, lifestyle modifications, publicly funded services and the management of surgical complications.

1.2 Obesity

Obesity is one of the National Health Priority Areas of Australia (1) and has existed as a major challenge facing the national and global population at a crisis level for many years. Obesity is a chronic and complex condition characterized by excessive fat accumulation, which alters anatomy and physiology, thus resulting in unfavourable metabolic, biomechanical and psychosocial health consequences (2). Obesity results from the multifactorial interplay of genetic, epigenetic, hormonal, biological, behavioural, sociocultural, and environmental dynamics that interrupt the balance between energy intake and energy expenditure (3, 4). Clinically, the body mass index (BMI) is the most practical diagnostic tool for obesity, albeit it is a crude index not distinguishing specific distribution of fat and lean body mass.

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan

Despite its relatively simplistic definition, obesity is an officially recognized condition that is difficult to treat and prevent. Lifestyle interventions remain the foundation for obesity management and can be effective in the short‐term; however, maintenance of weight loss can be challenging in the long‐term. Pharmacotherapies for weight loss and its maintenance are expensive, and long‐term studies longer than 2–3 years in duration are lacking. Given effective conservative treatment options (lifestyle modification and pharmacotherapy) appear to be modest for individuals with BMI ≥35.0 Kg/m2, bariatric surgery is considered the most effective and beneficial treatment for promoting long‐term weight loss while reducing and controlling obesity‐related comorbidities (5, 6).

This chapter covers currently available modalities of obesity management. It starts by providing general background to obesity, including its definition, obesity‐related comorbidities and the epidemic. Subsequently, a range of non‐surgical and surgical treatment for obesity will be walked through, from diet modification, exercise, very low energy diet (VLED), psychological intervention, pharmacotherapy to bariatric surgery. The available evidence regarding the most popular bariatric surgical procedures to date will also be highlighted, alongside their mechanisms of action and outcomes including the effectiveness on weight loss, various obesity‐related comorbidities and mortality, as well as the prediction of diabetes remission post‐surgery. An overview of publicly funded bariatric surgery, the importance of follow‐up after bariatric surgery, and the implications of pre‐operative weight loss on post‐surgery weight loss will also be provided. This chapter will be concluded with a brief introduction to existing knowledge gaps and the rationales and aims of this research.

1.2.1 Definition and diagnostic criteria

Body mass index (BMI)

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan

The body mass index (BMI) is an internationally recognised standard for measuring and classifying obesity in adults (7), calculated by dividing an individual’s weight in kilograms by the height in metres squared: Weight in Kg BMI = (Equation 1.1) (Height in m)2

Despite the fact that BMI does not necessarily reflect body fat distribution or describe the same degree of fatness in different individuals (8), at a population level in clinical settings, the BMI is generally a practical, easy and useful measure for identifying overweight and obesity.

The classifications established and published by the World Health Organization (WHO) provide the values depicted in Table 1.1 below, with obesity further classified into three classes according to severity: Table 1.1 World Health Organization (WHO) definitions of body weight disorders in adults by BMI (7)

BMI (Kg/m2) Classifications

Below 18.5 Underweight

18.5–24.9 Healthy weight

25.0–29.9 Pre‐obesity/overweight

30.0–34.9 Class I obesity

35.0–39.9 Class II obesity

Above 40 Class III obesity Abbreviations: BMI=Body mass index; Kg=Kilograms; m=metres

Certain racial/ethnic groups, such as Asian populations, have different definitions of obesity that use lower cut‐offs, i.e. ≥27.5 Kg/m2, compared to the standard BMI classification (9). All BMI thresholds should be reconsidered depending on the ancestry of individuals.

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan

In addition to these WHO classifications, some organizations and studies further classify obesity into the following terms:

(i) Severe obesity, clinically severe obesity or morbid obesity (MO) Severe and morbid obesity are often used interchangeably to describe the combination of class II and class III obesity: Severe obesity: BMI ≥35 Kg/m2 Generally, the World Obesity Federation, Australian Bureau of Statistics, and Australian Institute of Health and Welfare (AIHW) defined class II and/or class III obesity as severe obesity, regardless of the presence of obesity‐related comorbidities (8, 10‐12). Likewise, a number of landmark or large studies in the field of bariatric surgery research, including the Longitudinal Assessment of Bariatric Surgery (LABS) Consortium study, defined patients with ≥BMI 35 Kg/m2 as severely obese (range: 33.0–94.3) (13, 14). A large, retrospective, matched‐cohort study reported by Fisher et al. (15) which extracted data of 20,235 patients with type 2 diabetes mellitus (T2DM) (5,301 surgical and 14,934 matched non‐surgical patients) from electronic medical records (eMRs), insurance claims and other data systems in the USA, also defined patients with BMI ≥35.0 Kg/m2 as severely obese.

Clinically severe obesity or morbid obesity (MO): BMI ≥40.0 Kg/m2 or ≥35.0 Kg/m2 with the presence of at least one obesity‐related comorbidity The representatives from the Australian Clinical Obesity Services in Public Hospitals (COSiPH), Australian and New Zealand Obesity Society (ANZOS), and the Australian and New Zealand Metabolic & Obesity Surgery Society (ANZMOSS) recently published an expert consensus that defined a BMI greater than 35.0 Kg/m2 with at least one obesity‐related health condition, or BMI ≥40.0 Kg/m2 alone, as clinically severe obesity (16). Whereas the international landmark bariatric surgery studies, including the Sleeve vs Bypass (SLEEVEPASS) (17, 18) and Sleeve Multicentre Bypass or Sleeve Study (SM‐BOSS) randomized clinical trials (RCTs) (19) classify the same criteria as morbid obesity.

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan

(ii) Super obesity (SO) It is noteworthy that within class III obesity, there is a further sub‐division for those with BMI greater than 50 Kg/m2, namely super obesity (SO), which is widely used in research on this population (20‐25).

In the present studies, these definitions are applied to our highly complex bariatric surgical cohort with an initial BMI of ≥40 Kg/m2 and at least one obesity‐related comorbidity.

Limitation of BMI There is debate about the use of BMI for accurately defining obesity and its associated health risk. This is because BMI does not distinguish body fat and lean mass, and it is likely that only increased fat mass induces health risks (26‐28). Conversely, fat mass depletion due to weight loss is associated with improvement in obesity‐related comorbidities and overall health (29), reflecting the importance of accurate fat mass measurement.

Body fat is compartmentalized into two major types: subcutaneous adipose tissue (SAT) and visceral adipose tissue (VAT). The metabolic risks associated with obesity have been attributed to augmentation in VAT (30), which requires expensive, labour‐intensive and inaccessible gold‐ standard imaging techniques for accurate measurement, such as computed tomography (CT) or magnetic resonance imaging (MRI). Dual‐energy X‐ray absorptiometry (DEXA) is an alternative method for estimating VAT, which has been proven to correlate strongly with VAT measurements by CT and MRI. DEXA delivers a minimal radiation dose and is less costly and more time‐saving than both the former methods. However, DEXA equipment is expensive and this machine requires trained technicians (30).

Waist circumference (WC) is a simple measure of obesity, specifically, abdominal adiposity, and reasonably indicates the amount of VAT. A WC of ≥102 centimetres (cm) for Caucasian men and ≥88 cm for Caucasian women indicates a substantially increased risk of metabolic complications (11, 12). There are three different WC measurement sites that are mainly and routinely used,

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan with no consensus regarding the optimal one (30). The WHO recommends measuring WC at the midpoint between the bony landmarks of the lowest rib and the top of the iliac crest (WCmid) (31). Whereas the Anthropometric Standardization Reference Manual recommends measuring WC at the narrowest point of the torso (WCnarrow) (32). On the other hand, some research studies have measured WC at the level of the umbilicus (WCumbilicus) (33, 34).

1.2.2 Obesity‐related comorbidities

Three out of the ten leading causes of death worldwide, including ischaemic heart disease (IHD), cerebrovascular events/stroke and T2DM (35), are to a certain extent clearly linked to obesity as shown in Figure 1.1:

Figure 1.1 Top‐10 global causes of death in the year 2016 (35) (from the World Health Organization (WHO). Global Health Estimates 2016: Deaths by cause, age, sex, by country and by region, 2000‐2016. Geneva, Switzerland: World Health Organization (WHO); 2018)

Of the 56.9 million deaths that occurred worldwide in 2016, over half (54%, n=15.2 million deaths) were owing to the listed top‐10 causes. Of these, IHD and stroke were the biggest killers

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan and have remained the leading causes of death globally for the past 15 years. Diabetes mellitus (DM) accounted for 1.6 million deaths in 2016, up from less than one million in 2000.

Indeed, obesity greatly impacts health status through its associations with the following array of significant medical complications (36‐50), and the illustration in Figure 1.2. Greater details will be explained in Section 1.7: ▪ T2DM ▪ Hyperlipidaemia ▪ Hypertension ▪ Metabolic syndrome (defined as a cluster of the abovementioned conditions that occur together, including obesity, T2DM, hyperlipidaemia and hypertension) ▪ Cardiovascular disease [e.g. coronary artery disease (CAD) and ischaemic stroke] ▪ Certain types of malignancies (e.g. oesophageal, pancreatic, renal and endometrial adenocarcinomas, gastric, colorectal, postmenopausal breast, ovarian, gallbladder and thyroid cancers, hepatocellular carcinoma, and multiple myeloma) ▪ Sleep disturbances [e.g. obstructive sleep apnoea (OSA) and obesity hypoventilation syndrome (OHS)] ▪ Osteoarthritis (OA) ▪ Mental illness [e.g. depression, anxiety disorders and binge eating disorder (BED)] ▪ Gastro‐oesophageal reflux disease (GORD) ▪ Non‐alcoholic fatty liver disease (NAFLD), progressing in four stages from (1) steatosis (simple build‐up of fat in the liver cells) to (2) non‐alcoholic steatohepatitis (NASH) (a more serious form of NAFLD where the liver becomes inflamed), (3) fibrosis (persistent inflammation causing scar tissue around the liver and nearby blood vessels) and, ultimately, (4) permanent liver damage, cirrhosis, if not detected and managed. ▪ Hyperuricemia and gout ▪ Polycystic ovary syndrome (PCOS)

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan

Figure 1.2 Comorbidities associated with obesity (50) (adapted from Hanipah ZN, Schauer PR. Bariatric surgery as a long‐term treatment for type 2 diabetes/metabolic syndrome. Annual Review of Medicine. 2020;71:1‐15)

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan

In the USA, it has been estimated that the health costs incurred by a single individual with obesity were US$1901 per annum in 2014, projected to US$149.4 billion at the national level (51). In Australia, approximately one‐quarter of children and two‐thirds of adults are overweight or obese, and rates are rising largely due to the increasing epidemic of obesity, which cost the economy $8.6 billion in 2011–2012 (11). It is also interesting to note that obesity accounts for increased rates of comorbidities that affect quality of life (QoL) and complicates obesity treatments (52, 53), costing in excess of $58.2 billion per year in terms of workplace productivity, government assistance, and costs to the Australian health sector (54).

1.2.3 The epidemic of obesity

The WHO has officialised obesity as an epidemic. In parallel to this, Chooi and co‐authors (55) recently summarised the established knowledge on the epidemiology of obesity. They provided current trends and demonstrated that the prevalence of overweight and obesity worldwide has doubled since 1980 to an extent that nearly one‐third of the world’s population is nowadays overweight or obese, rendering obesity an epidemic and major public health threat (55). Figure 1.3(a) sets out the prevalence of overweight and obesity worldwide over time, divided into six regions according to the WHO, namely, the American, European, East Mediterranean, African, South East Asian and Western Pacific regions, which represent approximately 78% of the world’s population (55). The subsequent Figure 1.3(b) presents the age‐standardized prevalence of obesity worldwide identified in 174 countries in 2015 using the Global Burden of Disease (GBD) study’s data (56). The prevalence of overweight and obesity remained greater in women than in men throughout this period and increased with age (Figure 1.4).

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan

(i)

(ii)

Figure 1.3(a) Age‐standardised prevalence of (i) overweight and (ii) obesity in adults over 20 years old by geographical region and year (1980–2015) (55) (from Chooi YC, Ding C, Magkos F. The epidemiology of obesity. Metabolism. 2019;92:6‐10) Note: Americas includes USA, Brazil, Argentina, Mexico and Colombia; the European region covers the United Kingdom (UK), France, Germany, Russia and Turkey; the East Mediterranean region includes Pakistan, Egypt, Iran, Iraq and Afghanistan; the African region includes South Africa, Nigeria, Ethiopia, Congo and Tanzania; the South East Asian region includes India, Indonesia, Bangladesh, Thailand and Myanmar; and the Western Pacific region includes China, Japan, South Korea, Philippines and Vietnam

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan

Figure 1.3(b) Age‐standardised prevalence of obesity worldwide (174 countries) in 2015 among (A) men and (B) women aged over 20 years old (56) (from GBD Obesity Collaborators. Health effects of overweight and obesity in 195 countries over 25 years. New England Journal of Medicine. 2017;377(1):13‐27) Abbreviations: ATG=Antigua and Barbuda, BRB=Barbados, COM=Comoros, DMA=Dominica, E. Med.=Eastern Mediterranean, FJI=Fiji, FSM=Federated States of Micronesia, GRD=Grenada, KIR=Kiribati, LCA=Saint Lucia, MDV=Maldives, MHL=Marshall Islands, MLT=Malta, MUS=Mauritius, SGP=Singapore, SLB=Solomon Islands, SYC=Seychelles, TLS=Timor‐Leste, TON=Tonga, TTO=Trinidad and Tobago, VCT=Saint Vincent and the Grenadines, VUT=Vanuatu, W. Africa=Western Africa, and WSM=Samoa

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan

(a)

(b)

Figure 1.4 Age‐standardised global prevalence of (a) overweight and (b) obesity in men and women over 20 years old by year, from 1980 to 2015 (55) (from Chooi YC, Ding C, Magkos F. The epidemiology of obesity. Metabolism. 2019;92:6‐10)

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan

Obesity can start in childhood and adolescence. Recent data show that one in four children and adolescents (25%) aged 2–17 years in Australia were overweight or obese in 2017–2018 (57). This is a major concern because their obesity may continue into adulthood if left untreated, and lead to obesity‐related chronic diseases such as CVD and T2DM (58). Approximately 2 in 3 Australian adults (67.0%) were overweight or obese in 2017‐2018, among which, 31.3% were obese (8, 10). This data indicates a nearly two‐fold increment in the prevalence of obesity as compared to the prevalence of 19% in 1995.

There has also been a shift in the population distribution of BMI from 1995 to 2014–2015 (59), towards increased prevalence of obesity and higher BMI values as illustrated in Figure 1.5.

Figure 1.5 Distribution of body mass index (BMI) in adults aged 18 and over in Australia, 1995 and 2014‐ 2015, demonstrating the growing obesity trend in Australia (59) (from Australian Institute of Health and Welfare. Australia’s health 2018. Canberra, Australia: Australian Institute of Health and Welfare (AIHW); 2018)

It can be seen from the trend lines in the above graph (Figure 1.5) that since 1995, the proportions of Australian adults with a BMI in the more severe obesity ranges, namely, classes II

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan and III obesity, have increased dramatically. It almost doubled over the 20 years from a prevalence of 4.9% to 9.4% among Australian adults over 18 years old.

1.2.4 Aetiology of obesity

Obesity is of multi‐factorial aetiology that varies from patient to patient. Generally, a patient’s weight is tightly regulated, with the brain appearing central in weight regulation. There are a number of key drivers that can interfere with body weight regulation and contribute to an individual’s obesity. The factors include genetics, mental health factors (e.g. depression, BED, emotional eating, substance abuse and sexual abuse), medical conditions (e.g. hypothyroidism, Cushing’s syndrome and hypothalamic damage), and iatrogenic factors (i.e. prescribed medications). These factors combined with the modern ‘Western lifestyle’, appear to be driving the current obesity epidemic (49, 60). The Western lifestyle factors that influence the current obesity epidemic include high‐energy intake (food that contains a lot of fat, meat, and sugar); less consumption of vegetables and fruits; and decreased daily physical activity. This is not surprising given that unhealthy but highly palatable calorie‐rich foods and drinks, such as fast foods and sugar‐sweetened beverages, are now abundant and cheap. The modernization of transportation and changes in work and leisure activities have also reduced daily physical activity.

1.2.5 Overview of management of obesity

As the worldwide trends in obesity continue rising, treating obesity is at the top of the medical agenda. The management of obesity is complex, partly owing to the combination of contributing factors to obesity, as well as the presence of wider treatment objectives than weight loss alone. Management of obesity‐related comorbidities and improvement of patients’ QoL and well‐being are often included in the treatment aims. Treatments for obesity range from lifestyle modifications, behavioural therapy, pharmacotherapy to bariatric surgery. In general, a step‐up approach is often adopted in managing patients with more severe obesity and its related comorbidities, due to increasing risk with increasing invasiveness. In this context, multi‐modal

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan treatment is often required for successful management. Each of these obesity treatments will be elaborated in Sections 1.3‐1.4, divided into non‐surgical (lifestyle modification, psychological intervention and pharmacological management) and surgical treatments.

1.3 Non‐surgical treatments of obesity

1.3.1 Changes in diet and exercise

A plethora of studies have demonstrated that obesity (including ≥class I obesity or BMI ≥27 Kg/m2 with comorbidities) can be managed effectively through some combination of diet (e.g. changes in dietary choices and restriction of dietary energy intake) and/or exercise in conjunction with psychological/behavioural therapy as explained in the sub‐sections below. As a non‐invasive cornerstone of obesity management, such comprehensive lifestyle and behavioural interventions could usually result in a modest initial weight loss of 5−10% of inial weight loss. This has minimal risks while achieving significant clinical benefits (61), including improvement in cardiovascular risk factors and reduction in mortality (62‐66). Unfortunately, long‐term adherence to lifelong behavioural change has been a significant issue and, hence, many patients return to their pre‐ treatment baseline weight within five years of lifestyle modification (67). a) Dietary modification

There is no one superior dietary treatment for obesity management, rather a series of evidence‐ based diets that can be matched to patient preferences, past experiences, and the presence of comorbidities. In a recent review of dietary modifications for weight loss and its maintenance by Yannakoulia and team (68), it is apparent from the evidence that mainly an energy deficit is the more efficacious way to achieve weight loss. The authors also confirmed that no specific dietary scheme, be it nutrient‐, food group‐ or dietary pattern‐ based, is favoured (68). This was corroborated in the European Guidelines for Obesity Management in Adults, which states that a dietary intervention should target an energy deficit of ~500–1000 Kilocalories to achieve a steady

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan rate of weight loss of 0.5–1 Kg/week, although weight plateau generally occurs by six months (69‐71). b) Exercise

Exercise is an important component of a weight loss program in conjunction with caloric reduction. Several studies have reported additive benefits of combining exercise with caloric restriction on promotion of weight loss, prevention of weight regain, and body fat reduction, compared to dietary modification alone (72). Independent of weight loss, exercise has been related to a 6.1% decrease in VAT in patients with obesity, as quantified by radiographic imaging, whilst dietary modification showed virtually no change of 1.1% (73). Previous reports have shown that weight loss from exercise also exerts a number of beneficial actions, including improved serum leptin and adiponectin abnormalities in overweight and obese individuals, suggesting its therapeutic implications (74). Along the same lines, recent evidence further supports that supervised exercise programs can yield significant improvements in components of metabolic syndrome, particularly in reducing WC, as demonstrated by a systematic review conducted on eight eligible studies representing 1,218 patients (75). Longer supervised exercise programs and frequent interval sessions appear to have the greatest benefit and, thus, may help reduce cardiovascular risk and DM (75). Aerobic and resistance exercises play vital roles in bringing about benefits for patients with obesity and related comorbidities. For this reason, the European Guidelines for Obesity Management in Adults 2019 recommends that, initially, at least 150 min/week of moderate aerobic physical activity (e.g. brisk walking) should be combined with 2‐3 sessions/week of resistance exercise at least twice a week to increase muscle strength (76).

As highlighted earlier, a major health concern is the markedly rising prevalence of severe obesity. There is evidence that lifestyle interventions can also be effective for inducing weight loss in adults with severe obesity for a short‐term, and that exercise added to dietary modification further enhanced the magnitude of weight loss achieved. For instance, in a single‐blinded RCT of 130 adults with severe obesity (77), participants were randomised to a 1‐year lifestyle

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan intervention consisting of diet and exercise; delivered with a combination of group, individual and telephone contacts. One group was randomised to diet and exercise for 12 months (initial‐ activity) (n=67, mean BMI=43.5±4.8 Kg/m2), and the other group was to an identical dietary intervention but with exercise delayed for 6 months (delayed‐activity) (n=63, mean BMI=43.7±5.9 Kg/m2). The initial‐activity group lost significantly more weight in the first 6 months compared with the counterpart [10.9 Kg (95% CI=9.1‐12.7) vs 8.2 Kg (95% CI=6.4‐9.9), p=0.02]. Weight loss at 12 months was similar in the two intervention groups [12.1 Kg (95% CI=10.0‐14.2) vs 9.9 Kg (95% CI=8.0‐11.7), p=0.25]. The addition of exercise also resulted in greater reductions in WC and hepatic fat content.

1.3.2 Very low energy diet (VLED)

While lifestyle modifications remain the cornerstone of obesity treatment, adjunct therapies are an important part of the treatment armamentarium. To date, a very low energy diet (VLED), also called a very low calorie diet (VLCD), is the most effective and affordable non‐surgical, non‐ pharmacological treatment for inducing rapid weight loss over the short‐term (78, 79). A VLED is defined as restricting the daily energy intake to 2,500–3,350 Kilojoules (600−800 Kilocalories), and is the most intensive form of dietary intervention for management of obesity (71). It is most commonly achieved by total meal replacements (TMR), which involve replacing all meals and snacks with pre‐packaged, nutritionally complete shakes or soups to which water or milk are added, or bars or other pre‐packaged meal replacement products. To help induce mild ketosis and consequently appetite control, selected low‐carbohydrate and low‐energy additional items are sometimes allowed in clinical practice, such as tea, coffee, low‐joule jelly and low‐starch vegetables. The energy content of a TMR is typically less than 3,350 Kilojoules/day (800 Kilocalories/day) and, hence, TMRs are often referred to as a VLED which, by definition, provide less than 3,350 Kilojoules/day (80). TMRs are safe in the short‐term and have been reported to be used for up to five months with no adverse effect (81). A systematic review of 14 studies (eight RCTs, two cohort and six pre‐post studies without controls) among 6,579 adults with ≥class I obesity who used a VLED strategy for weight loss found that a

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan

VLED is effective for short‐term weight loss, but not all initial weight loss is maintained long‐term. In this investigation, it was determined that a VLED induced 8.4 to 22.0 Kg of initial weight loss over 12−16 weeks, 5.5 to 13.1 Kg one year aer commencement of VLED treatment, and 5.5 to 9.1 Kg after 2 years (82).

In spite of this, VLED has been determined as often underutilized by healthcare providers (83, 84). A survey of dietitians in Australia revealed that only 1.5% of local dietitians recommended a VLED to their clients for weight loss (83), possibly due to concerns regarding adverse psychological outcomes (e.g. inducing or exacerbating BED), although these concerns may be unfounded (85). Another safety concern is that rapid weight loss may possibly have adverse effects on body composition [e.g. bone mineral density (BMD)], in the sense that the greater the degree of energy restriction used to achieve rapid weight loss, the stronger the potential adaptive responses that may be induced compared with those seen during slow weight loss with a conventional food‐based diet (30). Other adverse effects that have also been frequently reported in studies included constipation, headaches, dizziness, fatigue, cholelithiasis and hair loss (86). Its underutilization by healthcare professionals may also be because of the lack of training and resources available for pre‐treatment evaluation and monitoring during these diets (87).

Besides substantial weight loss, VLED results in successful T2DM remission. This was shown in the Diabetes Remission Clinical Trial (DiRECT) study (88), an open‐label, cluster‐randomised, controlled trial performed at primary care practices in the UK. This trial was aimed to examine the efficacy of aggressive lifestyle intervention and weight reduction on T2DM. The lifestyle intervention consisted of withdrawal of antidiabetic and antihypertensive medications, VLED (825–853 Kilocalories per day formula diet for 12–20 weeks), stepped food reintroduction (2–8 weeks), followed by a structured support for weight‐loss maintenance. At 2 years, T2DM of 53 (36%) of 149 intervention participants were in remission, and 17 (11%) had achieved at least 15 Kg of weight loss (88).

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan

Published data to date suggested VLED, administered mostly via liquid meal replacement products, with wide variations of macronutrient composition, effectively produce clinically relevant weight loss (i.e. ≥10% of initial body weight) when used for ≥6 weeks even in individuals with class III obesity. For example, a recent systematic review and meta‐analysis (MA) (89) (11 eligible studies included, moderate risk of bias score, 10 studies were suitable for MA) found a pooled average weight loss of 9.8 Kg (95% CI=8.8‐10.8) after using VLED, representing a loss of approximately 4.1‐8.6% of initial body weight. VLED lasting for ≥6 weeks achieved 25.8 Kg (95% CI=22.2‐29.4) pooled average weight loss, representing 10.2‐28.0% weight loss.

A VLED should be used under an allied health professional’s supervision even in the absence of obesity‐related comorbidities, and regular medical review is required to manage the effect of VLED on obesity‐related comorbidities (70). As an example of this, antihypertensive agents may need to be reduced, while significant dose‐reductions of glucose‐lowering agents (such as sulphonylureas or insulin) are required to avoid hypoglycaemia in the setting of the rapid weight loss and carbohydrate restriction provided by a VLED (70).

VLED before bariatric surgery

In 2018, Holderbaum and colleagues addressed a similar topic from another scientific angle: a systematic review of quantitative studies on the effects of VLEDs on liver size and weight loss in the pre‐operative period of bariatric surgery (5). VLED treatment led to significant weight loss (90‐92) and liver size reduction in bariatric surgical patients during the pre‐operative period, thereby improving surgical access (93‐97). Pre‐operative weight loss is also desirable to optimise the safety of bariatric surgery in the sense of reducing the surgical risk of conversion from laparoscopic to open procedure (92). Besides this, VLEDs with various nutritional compositions and durations have been adopted during the pre‐surgery period among patients undergoing bariatric surgery for reduction of visceral fat (98‐100), blood loss in the trans‐operative period (101), and operating time (92, 101), which may affect the outcomes and costs of bariatric surgical procedures.

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan

1.3.3 Psychological intervention

Obesity is often implicated in psychosocial difficulties that can aggravate mental health. Problematic eating behaviours are common in individuals with obesity. Uncontrolled eating, variably operationalized as binge eating disorder (BED), binge eating, or loss‐of‐control eating, has emerged as a behaviour of particular interest (102). The term “BED” will be referred to across the thesis context for uniform reporting. BED is one of the most common eating disorders in the general population. It is characterized by the consumption of an objectively large amount of food in a discrete period of time, e.g. two hours, with an accompanying sense of loss of control over eating and subsequent distress (103). Current evidence supports that as BMI increases, the prevalence of BED increases (104, 105). Several other psychosocial conditions are also identified as associated with BED, including depression and anxiety.

On the basis of the abovementioned factors, the impact of psychological/behavioural therapies in obesity management has blossomed. Psychosocial interventions are progressively being recommended for weight loss and for bariatric surgery candidates before and after their operation. A number of systematic reviews have focused on weight loss outcomes and collectively suggest that behavioural interventions are beneficial. These include Liu and co‐ workers’ (106) review of eight studies assessing the impact of pre‐operative behavioural interventions on weight loss; Rudolph’s systematic review (of 15 studies) and meta‐analysis (MA) (of five RCTs) investigating the impact of post‐operative behavioural management interventions on weight loss (107); and Stewart and Anevell’s systematic review (of 11 studies) and MA (of five RCTs) of pre‐ and/or post‐operative behavioural lifestyle interventions focused on changing dietary behaviour or physical activity (108).

1.3.4 Pharmacotherapy Despite being the cornerstones of obesity management; lifestyle and behavioural modifications often fail to achieve or sustain predefined long‐term weight loss targets. Pharmacotherapy added to lifestyle modifications results in additional, albeit limited, weight reduction (109). In general,

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan pharmacotherapy achieves weight reduction intermediate to that of lifestyle intervention and bariatric surgery (70). Besides weight maintenance, pharmacological management has been proven to promote the amelioration of obesity‐related comorbidities and improve physical functions (particularly among patients with ≥class I obesity or BMI ≥27 Kg/m2 with comorbidities).

Anti‐obesity pharmacotherapy has experienced a recent revival following the US Food and Drug Administration’s (FDA) approval of several new agents for obesity management since 2012. An extensive review by Pilitsi et al. (109) that focused on pharmacological therapies for adult obesity is summarised in Table 1.2. From the evidence, it seems that the current FDA‐approved anti‐ obesity medications, including orlistat, liraglutide, and the combinations of phentermine/topiramate, and naltrexone/bupropion lead to an additional weight reduction of 5% when combined with lifestyle modifications. This may translate to multiple valuable metabolic and cardiovascular benefits (109). Orlistat is the only FDA‐approved medication for long‐term management of obesity that can be prescribed to adolescents >12 years of age, while phentermine can be used for short‐term treatment in individuals >16 years old (110).

Among the FDA‐approved drugs for obesity, naltrexone/bupropion and higher‐dose liraglutide have been approved by the Australian Therapeutic Goods Administration (TGA), thereby joining orlistat and phentermine.

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan

Table 1.2 Common weight loss drugs The common and FDA‐approved anti‐obesity medications, along with the eligible studies that established their safety and efficacy, mechanism of action, main health benefits, frequent side effects and contraindications (109)

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan

Note: All values shown are placebo‐subtracted. ^Analysis performed at 36 months. Abbreviations: BID=Twice daily; DA=Dopamine; DBP=Diastolic blood pressure; GABA=Gamma‐aminobutyric acid; GI=Gastrointestinal; GLP‐1=Glucagon‐like peptide 1; HbA1c=Glycosylated haemoglobin; HDL‐C=High‐density lipoprotein cholesterol; HT‐2C R=Hydroxytryptamine 2C receptor; HTN=Hypertension; LDL‐ C=Low‐density lipoprotein cholesterol; MAOI=Monoamine oxidase inhibitors; MEN=Multiple endocrine neoplasia; NA=Not available; NE=Norepinephrine; NS=Non‐significant; QD=Once per day; SBP=Systolic blood pressure; TC=Total cholesterol; TG=Triglycerides; TID=Three times daily; XR=Extended release; y=Year .

(from Pilitsi E, Farr OM, Polyzos SA, Perakakis N, Nolen‐Doerr E, Papathanasiou A‐E, et al. Pharmacotherapy of obesity: Available medications and drugs under investigation. Metabolism Clinical and Experimental. 2019;92:170‐92)

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan

1.4 Bariatric surgery

The Greek term ‘bariatric’ was coined from the combination of “bar‐” (meaning weight), “‐iatr‐” (which is treatment), and “‐ic” (i.e. pertaining to); hence, it translates to “pertaining to the treatment of weight”. Bariatric surgery has proven to be the most effective therapy for clinically severe obesity, leading to remission of obesity‐related comorbidities and long‐term sustained weight loss. The fundamental goals of bariatric surgical procedures remain the same – to meaningfully maximise weight loss and weight maintenance, and ameliorate obesity‐related comorbidities while maintaining nutritional health.

Traditionally, bariatric procedures have been classified as restrictive or malabsorptive, or a combination of both (111, 112). This classification is currently out‐of‐date, and the newer recognised mechanisms of action of bariatric surgery include alterations in gut hormones that favour improved insulin dynamics, increased satiety, reduced hunger, and augmented energy expenditure (113, 114). The main hormones playing a role consist of glucagon‐like‐peptide 1 (GLP‐1) (an incretin), oxyntomodulin (OXM), and peptide YY (PYY), all of which originate from the endocrine L‐cells in the gut. Other hormones, such as ghrelin, have also been identified as involved in post‐operative changes in appetite and insulin secretion (114).

This section discusses in great detail: 1) the current and most‐commonly performed bariatric surgeries, 2) recent evidence demonstrating its efficacy and safety in treating obesity and associated chronic conditions, 3) adherence to post‐operative follow‐up, and 4) prediction of DM remission among patients who have had bariatric surgery.

1.4.1 Type of bariatric surgical procedures

Bariatric surgery encompasses a range of procedures intended to achieve promising weight loss and resolve comorbidities. Several bariatric operations have fallen out of favour or have gradually become obsolete due to unacceptable post‐operative complication profiles in the long‐term,

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan such as the vertical banded gastroplasty (for examples, long‐term digestive difficulties and psychological problems) and biliopancreatic diversion with duodenal switch (BPD/DS) (for instance, technical complexity, long‐term significant nutrient deficiencies, GORD, diarrhoea, internal hernias and duodenal dissection) (115, 116). In contrast, two newer procedures − sleeve gastrectomy (SG) and mini gastric bypass‐one anastomosis gastric bypass (MGB‐OAGB) − are becoming much more popular. The four most widely accepted and performed bariatric procedures with proven efficacy and safety profiles that are presently in the mainstream bariatric surgery options are SG, MGB‐OAGB, adjustable gastric banding (AGB) and Roux‐en‐Y gastric bypass (RYGB) (117, 118). Each procedure comprises its unique effectiveness, durability and risk profile. The indication for each of these operations varies according to the patient’s pre‐existing comorbidities, demographics, risks and goals for weight loss.

Today, most bariatric surgical procedures are performed using minimally invasive technique, the laparoscopic procedure (keyhole surgery), since its advent in the 1990s (119). Since then, the operation became less invasive, and safety profile and acceptability of bariatric surgery improved. The improved documentation of effectiveness in terms of operation time, operative blood loss, length of intensive care stay, post‐operative pain, length of hospital stay (LOHS), recovery time, reversibility, and comorbidity incidence (119).

(a) Sleeve gastrectomy (SG) Today, sleeve gastrectomy (SG) is the most commonly employed primary bariatric procedure worldwide. Australia is no exception as depicted in the latest National Bariatric Surgery Registry 2018/19 Annual Report (Figure 1.6):

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan

LSG

LAGB

RYGB

MGB‐OAGB

Band reversals

Other procedures

Figure 1.6 Bariatric procedures performed in Australia and New Zealand (ANZ) in years 2014/15–2018/19 according to the ANZ Bariatric Surgery Registry (6) Abbreviations: FY=Financial year; LSG=Laparoscopic sleeve gastrectomy; LAGB=Laparoscopic adjustable gastric banding; RYGB=Roux‐en‐Y gastric bypass; MGB‐OAGB=Mini gastric bypass‐one anastomosis gastric bypass. (from The Australian and New Zealand Bariatric Surgery Registry Steering Committee. Bariatric Surgery Registry 2018/19 Report. Melbourne, Australia: Monash University; 2019)

It is a vertical gastrectomy involves excising a significant portion of the stomach, which results in the formation of a tubular remnant (120), as illustrated in Figure 1.7. Its effects as a bariatric surgery are mediated through both restriction and hormonal changes (114), as denoted in Figure 1.8. The simple SG procedure demonstrates exceptional advantages over other procedures in maintaining a functioning pylorus, avoiding an anastomosis and insertion of a foreign body, and preserving normal intestinal absorption (121).

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan

Figure 1.7 Pictorial representation of sleeve gastrectomy (SG) (122) (from American Society for Metabolic and Bariatric Surgery (ASMBS). Bariatric surgery procedures 2019 [Available from: https://asmbs.org/patients/bariatric‐surgery‐procedures])

Figure 1.8 Mechanism of action of sleeve gastrectomy (SG) (114) Sleeve gastrectomy (SG) reduces stomach volume with no alteration of nutrient flow. The newer mechanism of action acknowledges that the physiological changes, such as gut hormone secretion, are more important than the stomach volume reduction. Sleeve gastrectomy (SG) also

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan increases intestinal transit time, potentially due to the generation of high intraluminal gastric remnant pressures, the excision of the gastric pacemaker at the bariatric surgery, and neural signalling (from Neff KJ, le Roux CW. Mechanisms of action of bariatric surgical procedures. In: Agrawal S, editor. Obesity, bariatric and metabolic surgery: A practical guide. Cham: Springer International Publishing; 2016. p. 519‐27)

From a historical perspective, SG was developed in the late 1990s as a bridge to a definitive procedure in SO patients (123). It was soon found that short‐term, one‐year weight loss was comparable to that after AGB, and subsequently became a stand‐alone operation that is less technically demanding than others (118).

Despite its simplicity, SG is not without complications which, sometimes, can be life‐threatening. Long‐term post‐operative complications relate to stenosis/stricture, which can occur proximally at the gastro‐oesophageal junction or distally at the incisura angularis, and manifest as dysphagia and vomiting. Investigation is done with an upper‐GI contrast study or endoscopy and it can be successfully treated with endoscopic dilation (124). Additionally, SG may also lead to an increased risk of developing Barrett's oesophagus – a potentially pre‐cancerous condition – with an incidence of up to 18.8% after 5 years (125) and 17% after 10 years (125‐127).

Nutritional deficiencies are common after bariatric surgery and require lifelong screening during post‐operative follow‐up care. Some of these deficiencies can result in severe consequences. Patients who have had SG are at risk of vitamin D deficiency and metabolic bone disease (128, 129). The management will be outlined in greater detail in Section 1.5.1 (viii). Anaemia and iron‐, folate‐, vitamin B12‐, and vitamin D‐deficiencies are seen post‐SG (130) and require adherence to daily recommended multivitamins, vitamin B12 and calcium supplementation (131, 132). The details of nutritional profiles will be discussed in CHAPTER 2.

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan

(b) Mini gastric bypass‐one anastomosis gastric bypass (MGB‐OAGB)

Mini gastric bypass‐one anastomosis gastric bypass (MGB‐OAGB) is an emerging procedure rapidly gaining acceptance worldwide and in Australia for treating more severe obesity (133). MGB‐OAGB is a modification of Mason’s loop gastric bypass, and was first described by Rutledge in 1997 and officially reported in 2001 (134).

In 2002, Carbajo and Caballero from Spain proposed a technical variation to prevent gastroesophageal bile reflux. They renamed their technique “one anastomosis gastric bypass” (OAGB). According to this technique, OAGB involves a latero‐lateral anastomosis between the loop of the jejunum and the pouch, and the distance to the Treitz’ ligament averages 250–350 cm (135) (see Figure 1.9).

Figure 1.9 Pictorial representation of mini gastric bypass‐one anastomosis gastric bypass (MGB‐OAGB) (130) (from Mahawar KK, Borg C‐M, Kular KS, Courtney MJ, Sillah K, Carr WR, et al. Understanding objections to one anastomosis (mini) gastric bypass: A survey of 417 surgeons not performing this procedure. Obesity Surgery. 2017;27(9):2222‐8)

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan

Since then, other names, such as “single anastomosis gastric bypass” or “omega loop gastric bypass” have been proposed for the same technique (137, 138). In 2013, the confusion created by the various names led a group of surgeons using the name “mini gastric bypass‐one anastomosis gastric bypass” (MGB‐OAGB) to define this surgery (139).

The procedure is now recognised by the International Federation for the Surgery of Obesity and Metabolic Disorders (IFSO) as a mainstream bariatric procedure (117). In Australia, this procedure has been increasingly favoured in recent years. As revealed in the ANZ Bariatric Surgery Registry 2018/19 Annual Report (Figure 1.6), there was a 3.5‐fold increase in the use of this procedure compared to four years ago.

MGB‐OAGB has the advantages of technical simplicity, theoretically fewer sites for internal hernias (a worrisome complication often associated with a gastric bypass procedure), shorter operative time, weight loss durability, and confirmed safety and efficacy (140). Despite an increase in its utilisation, there remains concern that MGB‐OAGB could create bilio‐enteric reflux, which may increase the risk of oesophageal and gastric cancers (136).

Laparoscopic adjustable gastric banding (LAGB) was a popular bariatric surgery option between its FDA‐approval in 2001, until around 2008 when SG emerged and joined RYGB as the most popular primary procedures. LAGB is a restrictive procedure involving the placement of an adjustable silicone ring below the gastro‐oesophageal junction to create a pouch (120) (Figure 1.10). It is further connected to a tube that traverses the peritoneum and subcutaneous tissues and attaches to a reservoir port, which is placed within the anterior abdominal wall to create a stoma (141). The inflatable band can be inflated or deflated to adjust the degree of restriction. LAGB is reversible and there is no anastomosis.

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan

Figure 1.10 Pictorial representation of laparoscopic adjustable gastric banding (LAGB) (122) (from American Society for Metabolic and Bariatric Surgery (ASMBS). Bariatric surgery procedures 2019 [Available from: https://asmbs.org/patients/bariatric‐surgery‐procedures])

Figure 1.11 Summary of the mechanism of action of AGB (114) The LAGB procedure limits the rate of food emptying from the oesophagus to the stomach by applying pressure to the stomach inlet. This pressure can be adjusted by either inflation or deflation of the gastric band and modulate neural signalling, producing effects on satiety (from Neff KJ, le Roux CW. Mechanisms of action of bariatric surgical procedures. In: Agrawal S, editor. Obesity, bariatric and metabolic surgery: A practical guide. Cham: Springer International Publishing; 2016. p. 519‐27)

Initially, AGB was born out of a search for caloric restriction without disturbing the continuity of the gastrointestinal tract (142). The adjustable band could provide patients with a varying sized

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan stoma that can be changeable according to symptoms. It has been found to provide better weight loss outcomes compared with non‐adjustable bands.

Nonetheless, late post‐operative complications do arise. Slippage of the band, pouch dilation, band erosion, oesophageal motility and port site complications are the most common problems associated with LAGB (120, 132). In Australia, as high as 34.9% of reoperation rate (surgical reversals of bands) has been reported (6). In the USA, as another instance, based on a large retrospective database analysis, approximately 20% of patients underwent an average of nearly four reoperations each (143). Of the 25,042 patients who underwent AGB with an average of 4.5‐ years of follow‐up, gastric band‐related reoperations included device removal, device replacement, or revision to a different bariatric procedure such as SG or RYGB (143). In France, it was reported that the gastric band removal rate was up to 4% per year and, disappointingly, as high as 50% of removals at 15 years (144). Besides surgical complications, failure to sustain weight loss have also led to removal of the bands (112). It is estimated that up to 30% of patients undergoing LAGB might experience inadequate weight loss; i.e., ≤25 percentage of excess weight loss (%EWL) (145).

(d) Roux‐en‐Y gastric bypass (RYGB)

Roux‐en‐Y gastric bypass (RYGB) involves the formation of a gastric pouch of approximately 30 ml volume from the proximal stomach and division of the small bowel at a point 50–100 cm from the duodenojejunal flexure. The distal small bowel (Roux limb) is then anastomosed to the newly‐ formed pouch to create a gastro‐jejunostomy. The bypassed segment of the stomach and proximal small bowel (biliopancreatic limb) is then anastomosed to the small bowel 100–150 cm from the newly formed gastro‐jejunostomy (see Figure 1.12). Figure 1.13 summarizes the mechanism of action of RYGB.

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan

Figure 1.12 Pictorial representation of the Roux‐en‐Y gastric bypass (RYGB) (122) (from the American Society for Metabolic and Bariatric Surgery (ASMBS). Bariatric surgery procedures 2019 [Available from: https://asmbs.org/patients/bariatric‐surgery‐procedures])

Figure 1.13 Summary of mechanism of action of RYGB (114)

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan

The key features of RYGB consist of exclusion of the duodenum, increased delivery of nutrients to the ileum and, possibly, transection of vagal fibres during gastrectomy (from Neff KJ, le Roux CW. Mechanisms of action of bariatric surgical procedures. In: Agrawal S, editor. Obesity, bariatric and metabolic surgery: A practical guide. Cham: Springer International Publishing; 2016. p. 519‐27)

On the basis of the nature of RYGB being a much more complicated procedure and more technically challenging, surgical complications can occur due to physiological changes and alteration of anatomy. These include obstruction (gastric remnant distension and anastomotic stenosis), haemorrhage and cholelithiasis. Other complications include short bowel syndrome (a malabsorption disorder caused by a lack of functional small intestine), dumping syndrome (a condition that results from rapid passage of chyme into the small intestine, with uncomfortable symptoms including diarrhoea, nausea and feeling light‐headed or tired after a meal), and nutritional deficiencies (120).

Patients undergoing RYGB need to be monitored for adherence to multivitamin consumption and be scheduled for periodic surveillance of their important micro‐ and macronutrient levels (146). There can be exacerbation and onset of various nutritional complications, including deficiencies in vitamin B12, folate, zinc, copper and selenium; and iron deficiency anaemia (reviewed in references(147, 148); and hypovitaminosis D (128).

Operative choice between SG and RYGB

Nowadays, SG and RYGB are the two most commonly performed bariatric surgeries, including as revisional surgeries for failed LAGB (149). Generally, SG and RYGB are comparable in terms of effectiveness and safety in terms of the following (132). There are two recent landmark RCTs comparing RYGB and SG: the Sleeve vs Bypass (SLEEVEPASS) (18) and the Swiss Multicenter Bypass or Sleeve Study (SM‐BOSS) (19) trials. Both demonstrated similar excess weight reduction over 5 years of follow‐up, which is also in line with the weight loss trajectory that Vitiello and team (150) described in a 5‐year retrospective, matched, case‐control study. Both the SLEEVEPASS and SM‐BOSS RCTs also found similar outcomes with respect to the resolution rates

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan

of obesity‐related comorbidities (T2DM, hypertension and dyslipidaemia), operative complications, complications requiring operative re‐intervention or mortality, and QoL at study conclusion (18, 19). It should be cautioned that the SM‐BOSS trial found patients undergoing SG had a significantly higher rate of worsened (32% vs 6%) and de novo GORD (32% versus 11%), while the RYGB group had a significantly higher remission rate of GORD (60% versus 25%).

Of note, another landmark trial for DM in obesity, the Surgical Treatment and Medications Potentially Eradicate Diabetes Efficiently (STAMPEDE) study (151), suggests that RYGB may be preferable to SG as a DM intervention, because over 5 years of follow‐up, RYGB patients were more likely to be off glucose‐lowering medications than patients who had had SG (45% versus 25%). A descriptive cohort study performed in the Netherlands found that the rate of early dumping symptoms after SG was similar to that following RYGB, but patients who had had SG appeared to experience less late dumping syndrome (152). RYGB was also shown to have higher rates of myocardial infarction (MI) and pulmonary embolism (PE) than SG or LAGB in immediate post‐operative period (<30 days following bariatric surgery) (153). On the contrary, Chang and co‐authors (153) found that SG [1.21%, 95% Confidence Interval (CI)=0.23–2.19%] resulted in a higher anastomotic leak rate than RYGB (1.14%, 95% CI=0.84–1.43%). Healthcare professionals should tailor the operative choice of SG or RYGB to individual patients' goals and weigh the benefits and risk tolerances through a process of shared decision‐making of SG or RYGB.

In super obesity, a two‐stage approach consisted of SG followed by RYGB is a popular model of care in the event of de novo GORD after the first procedure, with or without migration of the sleeved stomach into the chest (154).

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan

1.5 Outcomes of bariatric surgery

The level of adiposity is only part of the problem of obesity; the much more serious problem is the comorbidities related to obesity. Bariatric surgery is endorsed as the most effective treatment for obesity and its associated complications with respect to long‐term weight reduction, and improvement or remission of obesity‐associated comorbidities.

1.5.1 Effectiveness of bariatric surgery i) Weight loss

Weight loss has tremendous benefits in obesity. The fact that bariatric surgery induces significant weight loss in patients with clinically severe obesity is well‐established. The first and largest prospective case‐controlled study examining surgical weight loss was, and remains, the Swedish Obesity Subjects (SOS) study (155, 156). Despite study sample being over‐represented with outdated vertical banded gastroplasty, the SOS study was pioneering. It demonstrated that there was superior weight loss with bariatric surgery compared with non‐surgical alternatives (157). More recently, numerous RCTs have confirmed that bariatric surgery leads to significantly great weight loss. These RCTs include the Look Action for Health in Diabetes (AHEAD), SM‐BOSS and SLEEVEPASS trials. At trial’s end (i.e. approximately 10 years), the Look AHEAD RCT reported 19.3% weight loss from baseline among 196 participants who had bariatric surgery (158). The SM‐BOSS, a 2‐group RCT, exhibited 61.1% excess BMI loss (SG) versus 68.3% excess BMI loss (RYGB) at 5 years (19). Likewise, at 5 years after bariatric surgery, the SLEEVEPASS trial revealed an estimated mean %EWL of 49% (95% CI=45%‐52%) after SG and 57% (95% CI=53%‐61%) after RYGB among the 240 patients enrolled (18).

In view of the fact that the effectiveness and durability of weight loss are key attributes of the broad acceptance of bariatric surgery, a systematic review and MA performed by O’Brien et al. (159) reported very long‐term (defined by 10 or more years of follow‐up) weight loss outcomes

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan after bariatric surgery. A total of 57 datasets were identified, of which 33 were eligible for MA. Weighted mean %EWL values were calculated for all the articles included in the systematic review. Among them, eighteen reports of RYGB showed a weighted mean of 56.7 %EWL, 17 reports of LAGB demonstrated 45.9 %EWL, nine reports of BPD/DS showed 74.1 %EWL, and two reports of the newer procedure SG presented 58.3 %EWL (159). As expected, all the bariatric procedures were shown to be associated with substantial and durable weight loss. However, it was pointed out that more long‐term data are needed for the newer procedures, particularly SG and MGB‐OAGB, which have gained popularity among bariatric surgeons in recent years. ii) Type 2 diabetes mellitus (T2DM)

There is substantial evidence suggesting that weight loss can reverse and improve a number of consequences of obesity. The link between T2DM remission and bariatric surgery is not a new concept. Weight loss has been shown to improve insulin sensitivity and β‐cell function, and induce remission of T2DM (160). Numerous observational studies and RCTs have shown marked improvements in the majority of patients with T2DM immediately after bariatric surgery (118, 132).

First of all, the landmark STAMPEDE trial randomized 150 overweight or obese adults with DM into treatment groups that received intensive medical therapy alone or in combination with RYGB or SG. After 5 years, among the remaining 134 patients, both bariatric surgical procedures were found to be superior to medical therapy alone, i.e. 29% and 23% of patients in the RYGB and SG arms, respectively, achieved greater reductions in glycaemic control measures, as compared to 5% of the patients in the intensive medical therapy arm (151). Another large, long‐term, multicentre, observational prospective LABS cohort study showed improved blood glucose control and remission of DM among patients who underwent bariatric surgery up to 7 years post‐ operation (14). More specifically, for RYGB, the proportion of DM in remission at 1, 3, 5, and 7 years were 71.2% (95% CI=67.0‐75.4), 69.4% (95% CI=65.0‐73.8), 64.6% (95% CI=60.0‐69.2), and 60.2% (95% CI=54.7‐65.6), respectively. Whereas for the LAGB group, 30.7% (95% CI=22.8‐38.7),

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan

29.3% (95% CI=21.6‐37.1), 29.2% (95% CI=21.0‐37.4), and 20.3% (95% CI=9.7‐30.9) of patients were observed to have DM remission (14).

Bariatric surgery has also proven to have considerable efficacy in DM prevention. In a study that prospectively followed patients with obesity who sought and underwent RYGB, the incidence of DM in the RYGB patients was 3%, compared to 26% in non‐surgical patients after 12 years (161). iii) Cardiovascular risk and cardiovascular disease (CVD)

Hypertension An emerging body of literature confirms that bariatric surgery has a significant capacity to improve cardiovascular disease (CVD) and its risk (132), such as hypertension. Hypertension is undoubtedly one of the commonest obesity‐related comorbidities in patients undergoing bariatric surgery. It has been suggested to be related to the neuroendocrine mechanisms (e.g. the renin‐angiotensin‐aldosterone system) (162). Multiple RCTs and systematic reviews have shown improvements in hypertension after bariatric surgery (reviewed in reference(118). Among the studies included in that review, hypertension resolved in 50–62% cases and was improved in 63–86% of patients who had bariatric surgery. In another systematic review, BPD/DS was shown to exhibit the greatest percentage of remission of hypertension, followed by RYGB, SG and AGB (163).

A more recent RCT comparing patients with hypertension treated with either RYGB or medical therapy (164) also confirmed that bariatric surgery leads to reduced cardiovascular risk, as measured by the Framingham Risk Score (165, 166). This RCT demonstrated that the primary endpoint (i.e. blood pressure control of ≥30% reduction in the number of antihypertensive medications) was achieved by as high as 83.7% of patients receiving surgery compared to 12.8% receiving medical therapy (164).

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan

Hyperlipidaemia/Dyslipidaemia Hyperlipidaemia or dyslipidaemia follows similar trends to that of hypertension, with some studies reported improvement and/or resolution of dyslipidaemia up to 70–80% of patients (118). The way this has been determined varies among studies, possibly included reductions in statin medication, and/or changes in cholesterol, triglycerides, low‐density lipoprotein cholesterol (LDL‐C), and/or high‐density lipoprotein cholesterol (HDL‐C) levels (163). The obesity‐related comorbidities including hyperlipidaemia are discussed in greater detail in CHAPTER 2.

Cardiovascular disease (CVD) Patients with obesity often sustain from various cardiovascular diseases (CVD). In the landmark SOS study of over 2,000 patients who had bariatric surgery and a prospectively‐matched, non‐ surgical control group that was followed for 10 years, the surgical patients appeared to have a lower risk of cardiovascular mortality and total first‐time acute coronary syndrome (ACS) or stroke (167). Bariatric surgery has also been proven to decrease coronary arterial calcification as well as improve both diastolic and systolic function in obese patients with heart failure (168, 169). In terms of being a preventive strategy, bariatric surgery has also been reported by, for example, these three large, observational cohort studies to have favourable effect on the risk of heart failure (170‐172). When comes to neurovascular risk, bariatric surgery has been observed to reduce carotid atherosclerosis and atrial fibrillation (173, 174).

On the flip side, it has been reported that bariatric surgery may be detrimental to the control of certain types of CVD, such as dysrhythmia, in the early post‐operative period (175). iv) Obstructive sleep apnoea (OSA)

Another strong correlation exists between obesity and obstructive sleep apnoea (OSA), a sleep‐ related breathing disorder characterized by repeated episodes of apnoea and hypopnoea during sleep (176). A recent systematic review and MA (177) was conducted to examine the relationship between surgical weight loss and OSA severity (i.e., apnoea‐hypopnoea index [AHI]), and how

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan this relationship is altered by the various respiratory events scoring (RES) criteria used to derive the AHI. It was determined that surgical weight loss resulted in reductions in BMI and AHI; however, OSA persisted at follow‐up in the majority of the study patients. There was high between‐study heterogeneity, which was largely attributable to baseline AHI and the duration of follow‐up when analysed using meta‐regression. In spite of this, there was insufficient data for evaluating the impact of different RES criteria on OSA severity.

In the subsequent year, a research team from the UK (176) performed an updated systematic review and MA comparing LSG and RYGB in patients with obesity. The studies included in this review were RCTs, comparative prospective, and matched‐cohort studies. It was found that there was a trend towards increased resolution of OSA after LSG at 1 year follow‐up (OR=0.47, 95% CI=0.20‐1.06, p=0.07) and at 3 years follow‐up (OR=0.52, 95% CI=0.16‐1.71, p=0.28), when compared to RYGB. It was suggested that the improved OSA resolution post‐LSG could be due to differences in vagal nerve damage during the procedures (176). Another study manifested that there is a significant difference between microglia activation in vagal structures following RYGB and LSG, which can lead to differences in the remodelling of gut‐brain communication. This could then influence the collapsibility of pharyngeal airways and, therefore, OSA resolution rates (178). v) Osteoarthritis (OA) and total joint arthroplasty (TJA)

Osteoarthritis (OA) is a progressive degenerative joint disease that leads to joint damage, chronic pain and joint stiffness, muscle atrophy, impaired mobility, poor balance and, eventually, overall physical disability. There is ample data to suggest that obesity accelerates the development of OA of the knee and hip by exerting deleterious effects on joints through both biomechanical and systemic inflammatory changes (118, 179). As an ultimate treatment, lower‐extremity total joint arthroplasty (TJA), comprising total knee arthroplasty (TKA) and total hip arthroplasty (THA), remains the gold standard management for highly severe end‐stage OA of the knee and hip. As rates of more severe obesity increase, rates of TKA and THA performed have also increased, outpacing the non‐obese population.

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan

Having said that, because some orthopaedic surgeons may not operate on patients with extreme BMIs, bariatric surgery has been recommended to assist patients in losing weight before TJA is performed. Bariatric surgery has been shown to not only maintain weight loss, but also reduce pain and stiffness, and improve knee function in patients with knee OA. For example, a prospective observational study in the USA by Hacken and colleagues (180) evaluated 13 patients who underwent bariatric surgery and had symptoms and radiographic evidence of knee OA. They determined that at the fifth year of follow‐up, significant improvement in pain (p=0.0005) and daily living function/activities (p=0.0088) were achieved, as measured by Knee Osteoarthritis Outcome Score.

The majority of studies also speculated improvements in gait biomechanics, joint pain, joint function and range of motion amongst patients undergoing both bariatric surgery and TJA (reviewed in references(179, 181), there always seems to have some debate remains over whether performing bariatric surgery prior to TJA improves complication and revision rates in patients with more severe obesity.

A recent systematic review (182) comprehensively studied 13 studies that included moderate to high quality of evidence on a total of 11,770 patients who had undergone bariatric surgery prior to TJA. It found no consensus in the literature in terms of complication and revision rates. For example, a Propensity Score‐Matched Analysis of a New York state‐wide database analysis (183) that matched patients undergoing bariatric surgery with the obese controls not receiving surgery (TKA: 2,636 bariatric surgical group versus 2,636 controls; and TJA: 792 bariatric surgical group versus equal number of controls), found that bariatric surgery did not reduce the risk of revisional surgery for TKA nor THA.

In the same year, Li and team (184) conducted a further MA on a total of nine studies with 38,728 patients. Their findings from the body of eligible evidence revealed that bariatric surgery prior to TJA was associated with partial improvement of short‐term outcomes after TJA, including medical complications, LOHS and operative time. Nonetheless, bariatric surgery did not improve the risk

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan for long‐term outcomes, including dislocation, periprosthetic infection, periprosthetic fracture and revision (184).

Fundamental clinical questions remain regarding the optimal management of more severe obesity and lower‐extremity OA. This should be the focus of future research interdisciplinary collaborations into providing care to patients with both conditions to provide some surgical tips to improve weight loss and TJA outcomes in obesity. vi) Mental health illness

Clinical and psychosocial outcomes have been studied after bariatric surgery, in long‐term and short‐term as reviewed by Castaneda et al. (185). Bariatric surgery candidates, an already‐ vulnerable population even before surgery, have been reported to have a higher proportion of pre‐surgical psychiatric illnesses and psychotropic medication use compared to the general population (53, 186‐188). Around 21‐61% of bariatric surgery candidates suffer from a psychiatric disease (185).

Mood disorders (Depression and anxiety) Over half of patients with obesity who present for medical or surgical management of their obesity meet the criteria for a psychiatric illness, most often a mood disorder (189). The prevalence of depression and anxiety, the most common pre‐operative mood disorders, in this population, is significant (187, 189). Patients with obesity have approximately double the rate of depression compared to the general population. Roberts et al. found that 15.5% of patients with obesity met the criteria for major depression, compared to 7.4% of people with normal weight (190). It is, therefore, unsurprising that a significant proportion of patients who present to specialty obesity clinics for surgical intervention struggle with depressive symptomology (191). Depressive symptoms may also be exacerbated by the presence of medical comorbidities and the occurrence of post‐surgical complications and persistence/recurrence of comorbidities after surgery (192). In addition to bariatric surgery, mood disorders are prevalent in patients before

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan any major surgery, due to associated pain, morbidity, and decreased well‐being (193). If not addressed promptly, they may predict increased morbidity and mortality after the procedures. These issues highlighted the potential impact of bariatric surgery on mental health and outcomes in a population that is considered already at risk at baseline.

A recent systematic review (2019) was carried out to assess the effects of bariatric surgery on long‐term changes (≥2 years) in anxiety and depressive symptom severity in patients with a BMI ≥35 Kg/m2 (194). A total of 2,058 articles were reviewed for eligibility and 14 prospective studies were included. Thirteen studies (93%) reported significant reductions in depressive symptom severity 2–3 years after bariatric surgery. However, the authors highlighted that all 14 studies recorded statistically significant reductions in depressive symptoms at the conclusion of the studies. Similarly, there were reductions in overall anxiety symptom severity at ≥2 years follow‐ up (k=8 studies, n=1,590 pooled). Pre‐operative anxiety or depression scores, however, did not predict outcomes of post‐operative BMI. Likewise, post‐surgery weight loss did not predict changes in anxiety symptoms. This new systematic review supports some existing literature that shows metabolic treatments may be viable therapeutic interventions for mood disorders.

Despite a range of long‐term health improvements following bariatric surgery, as well as a positive short‐term impact on mental health and psychosocial functioning, the literature has raised concerns regarding the risk of adverse psychiatric events that may emerge following surgery, particularly substance abuse, self‐harm, suicidality and binge eating (195‐200), which will be detailed in the subsequent paragraphs.

Self‐harm/suicide attempts While bariatric surgery is generally beneficial in terms of depressive symptoms, there exists a cohort of patients undergoing bariatric surgery who might worsen following surgery and be at increased risk of self‐harm and/or suicide. This is not unexpected, given the likely multifactorial consequences of weight regain, unrealistic expectations or other life stresses, which patients often realise these issues in retrospect (201).

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan

This is underpinned by an updated systematic review published by Castaneda and colleagues (185), which reviewed the published literature to evaluate the association of bariatric surgery with suicide events and suicide/self‐harm attempts in patients who had undergone bariatric surgery. In this review of 32 studies of 148,643 subjects, there were predominantly females (76.9%), and RYGB being the most commonly performed procedure (58.9%) (185). The post‐ bariatric suicide event rate was 2.7/1000 patients (95% CI=0.0019–0.0038), while the suicide/self‐harm attempt event rate was 17/1000 patients (95% CI=0.01–0.03). The findings indicate that the self‐harm/suicide attempt risk was higher after bariatric surgery within the same population [Odds Ratio (OR)=1.9, 95% CI=1.23–2.95] compared to age‐, sex‐, and BMI‐matched control subjects (OR=3.8, 95% CI=2.19–6.59). Castaneda and team concluded that post‐bariatric surgery patients had higher self‐harm/suicide attempt risk compared to the matched controls (185).

The findings of this systematic review further supported by a new study from Australia published in JAMA recently (i.e. year 2020) (202) that was not included in the systematic review. This state‐ wide, mirror‐image, longitudinal cohort study was conducted to investigate the association of bariatric surgery with the incidence and predictors of mental health presentations, deliberate self‐harm, and suicide for determinants of mental health service use. Data from the Western Australian Department of Health Data Linkage Branch was used. It comprised records from 24,766 patients who underwent their first bariatric surgery. Over a mean [±standard deviation (SD)] follow‐up period of 5 (±2.9) years, 1 in 10 bariatric patients (10%) who had bariatric surgery were found to have used at least one episode of mental health care post‐surgery, and about 7% of the patients required some form of psychiatric services after bariatric surgery. Morgan and co‐ workers also confirmed that existing psychiatric conditions were a major risk factor for requiring mental health services after bariatric surgery, and suicides accounted for 9.6% of all deaths. These findings suggest that caution should be used when considering the hypothesis that weight reduction due to bariatric surgery will improve mental health, particularly suicidal thoughts and self‐harm in patients with obesity (202).

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan

Eating disorder Studies have exposed that eating disorder is not uncommon in patients prior to bariatric surgery, and may persist or emerge after operation. Up to date, binge eating disorder (BED) is the most common eating disorder prior to bariatric surgery, with prevalence rates ranging from 4% to 49% (21, 203). The prevalence of BED often decreases after surgery (204). The focus has been particularly on BED and other maladaptive eating behaviours (e.g. grazing and night eating syndrome), given they have been consistently linked to attenuated weight loss after bariatric surgery (205). An article by Kalarchian and Marcus (195) reviewed the current status of research on psychosocial concerns post‐bariatric surgery. It found from the available studies that psychosocial interventions have a positive impact on post‐surgery outcomes, particularly maladaptive eating.

On the flip side, sub‐studies of longitudinal LABS reported that 12.1% of RYGB and 14.6% of LAGB patients fulfilled the criteria for BED before bariatric surgery (206), although those prevalence rates remained lower than initial pre‐surgical estimates for both surgical groups (196, 206). The rates of BED slowly increased over a 7‐year period following bariatric surgery, culminating in 3.3% of RYGB and 6.6% of AGB patients meeting the criteria for BED (206). These findings suggested that binge eating is a clinically relevant behaviour that may impede weight loss, and thus, they highlighted the importance of conducting ongoing assessment of maladaptive eating following surgery (206).

Psychiatric illnesses A systematic review and MA of RCTs that compared surgical and non‐surgical treatments and assessed mental health QoL, to better understand the impact of bariatric surgery on the risk of adverse mental health outcomes, underlined that the final mental health QoL scores were similar for the two treatment groups (standardized mean difference=0.37, 95% CI=−0.07‐0.81) (199). They explained these results, in conjunction with the fact that individuals who choose bariatric surgery tend to have high‐risk baseline characteristics, suggesting that intensive mental health follow‐up after bariatric surgery should be routinely considered (199).

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan

Psychiatric evaluations before and after bariatric surgery are strongly recommended by the International Federation for the Surgery of Obesity (IFSO) (207), the American Society for Metabolic and Bariatric Surgery (ASMBS) (208), and The Obesity Society (TOS) (131) to identify patients with high risk of self‐harm or suicide. Current guidelines recommend a face‐to‐face interview by a qualified evaluator to determine psychosocial, developmental, cognitive, personality, lifestyle, motivational and social support aspects pre‐operatively. Post‐operative follow‐up is also highly recommended to assess the need for psychotherapy or pharmacological interventions (131, 207, 208). Taken together, these results and guidelines propose the use of a multidisciplinary team that include psychiatrists, psychologists and other mental health professionals to optimize patient care in bariatric surgery candidates. vii) Cancer

The International Agency for Research on Cancer (IARC) has identified 13 malignancies as obesity‐ associated cancers (OACs) (209). Oesophageal, pancreatic, colorectal, post‐menopausal breast, endometrial, kidney, liver, thyroid, gastric cardia, meningioma, ovarian, oesophageal adenocarcinoma, gallbladder and myeloma cancers have been implicated (118, 209‐211). It has been suggested that obesity promotes and accelerates cancer development and growth via multiple pathways (211). On the basis that a comprehensive elaboration of all plausible obesity and cancer mechanisms is beyond the scope of this thesis, a summary is illustrated in Figure 1.14. These mechanisms create great challenges in combating the obesity‐cancer linkage with a single or focused targeted approach.

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan

Figure 1.14 Mechanisms underlying cancer promotion and development induced by obesity. In summary, the mechanisms that have received the most study are insulin and insulin‐like growth factors (IGFs), sex hormones, and adipokines (212) (from Bruno DS, Berger NA. Impact of bariatric surgery on cancer risk reduction. Annals of Translational Medicine. 2019:1‐15)

Unlike the seemingly more direct relationship between obesity and T2DM, the relationship between cancer and obesity remains intriguing, and varies among different type of cancers (132). A recent systematic review and MA (205) of eight population‐based studies (total 635,642 patients) assessed the effect of bariatric surgery on cancer incidence at the population level. Pooled odds ratios (PORs) were calculated for the incidence of cancer after bariatric surgery compared to controls. It was determined that bariatric surgery was associated with a significant reduction in the incidence of obesity‐related cancer (POR=0.55, 95% CI=0.31‐0.96; p=0.04). Bariatric surgery was also protective for breast cancer development (POR=0.50, 95% CI=0.25‐ 0.99; p=0.045) at the population level (213).

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan

Interestingly, one of the studies included in the abovementioned systematic review and MA, a national population‐based cohort study in England that conducted using data of 716,960 patients with obesity from the Hospital Episode Statistics database (214) indicated that bariatric surgery was associated with a 77% lower risk of developing hormonal cancers (breast, endometrial and prostate) (OR=0.23, 95% CI=0.18–0.30). However, on the flip side, RYGB resulted in an increased risk of colorectal cancer (OR=2.63, 95% CI=1.17–5.95). Among the 8,794 patients who underwent RYGB, LAGB and SG who were matched with 8,794 obese patients who did not undergo surgery, patients who had bariatric surgery demonstrated a consistently lower risk of breast (OR=0.25, 95% CI=0.19–0.33), endometrium (OR=0.21, 95% CI=0.13–0.35) and prostate (OR=0.37, 95% CI=0.17–0.76) cancers compared to non‐surgical patients. Likewise, in the SOS study (157, 215), men with a BMI ≥34 Kg/m2 and women ≥38 Kg/m2 were followed‐up for a mean of 10.9 years (range: 0–18.1 years). Women who underwent surgery showed a reduced incidence of new cancers compared to controls, with significant differences for malignant melanoma and hematologic malignancies by 42% (HR=0.58, 0.44–0.77). In men, there was no cancer risk reduction (HR=0.97, 0.62–1.52). It should be noted that the mortality benefit of bariatric surgery observed in, for instance, the SOS study (156), was primarily driven by decreased cancer‐related death rather than major cardiovascular outcomes.

The links between bariatric surgery and cancer are strong, and bariatric surgery appears to be associated with reduced cancer incidence at a population‐based level (132). viii) Bone health

While bariatric surgery can effectively elucidate the problem of obesity and its related complications, accumulating evidence indicates that it, unfortunately, has negative skeletal consequences, causing unfavourable bone loss and higher fracture risk (216, 217). As a result, the field of research into bone health following bariatric surgery is rapidly growing. For example, a 12‐year nationwide cohort study in Taiwan comparing 2,064 patients who had a history of RYGB

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan with 5,027 obese patients who did not perform any bariatric surgery appeared having a 1.2‐fold increased fracture risk in the surgical group compared to the controls (95% CI=1.02–1.43) (217).

Skeletal deterioration after bariatric surgery may act via multiple mechanisms, mainly mechanical unloading, nutritional deficiencies and hormonal changes (recently reviewed (year 2020) in reference(216). The severity of bone outcomes was deemed to be likely due to the degree of malabsorption depending upon the specific surgical procedures (216). Lu and colleagues also proposed the view that more malabsorptive procedures are significantly correlated with higher fracture risk 1–2 years following bariatric surgery (adjusted‐HR=1.21, 95% CI=1.02–1.43) (217). Consistent with Lu’s findings, Elaine and research team confirmed that, compared with restrictive AGB, RYGB (a combination of restrictive and malabsorptive procedures) is associated with a 43% increase in non‐vertebral fracture (218). The significant differences in non‐vertebral fracture risk between RYGB and AGB groups were particularly significant at year‐5 post‐operation, with the estimated HR increased to 3.91 (95% CI=1.58–9.64).

Current post‐operative clinical managements focus on ameliorating nutritional deficiencies and reducing bone loss, in order to attenuate bone deterioration after bariatric surgery. For instance, the Clinical Guidelines of the Obesity Management Task Force of the European Association for the Study of Obesity (EASO) (219) recommend that patients undergoing bariatric surgery should receive follow‐up care in a multidisciplinary setting, which will be explained further in Section 1.7 of the body of this chapter. Pre‐bariatric evaluation and post‐surgery follow‐up are stipulated to be in place to optimize patient outcomes. This could be a structured screening for risk factors of bone loss, bone metabolic or turnover markers, or bone mineral density screened by dual‐energy X‐ray absorptiometry (DEXA). It is also highly recommended that a screening is to be done during the first two years following bariatric surgery (219). Routine baseline and annual post‐operative evaluation and monitoring for vitamin D deficiency after RYGB and SG are also encouraged (115). In patients who have experienced, for example, RYGB, treatment with oral vitamin D and calcium citrate is indicated to prevent or minimise secondary hyperparathyroidism. An initial vitamin D supplement of 3000 IU/day post‐operatively has been recommended for most patients

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan regardless of the type of bariatric surgery. Patients with severe vitamin D malabsorption are recommended to be given initial oral doses of vitamin D2 or D3 of 50,000 IU 1‐3 times/weekly. Recalcitrant cases may require concurrent calcitriol or calcium injection, such as for people with extremely low serum calcium.

Considering the pros and cons of bariatric surgery in terms of bone health, further comprehensive investigation of its effects of these bariatric procedures on bone metabolism is needed. Further research may be valuable in exploring the association between skeletal changes and bariatric surgery, such as SG that is increasingly becoming a more prevalent bariatric surgical procedure. ix) Mortality

Although some older retrospective epidemiological studies suggested that weight loss may be associated with increased mortality, contradictory findings have been reported; for example, a better overall survival rates after bariatric surgery have been observed in the newer studies (reviewed in reference(220). Sheng and team discovered, in their systematic review and MA, that in the long‐term (defined by >5 years), bariatric surgery reduced mortality among T2DM patients. A total of 10 eligible articles (one RCT and nine cohort studies) were included in this review, with pooled estimates revealed that bariatric surgery was associated with significantly lower mortality (RR=0.21; 95% CI=0.20–0.21) compared with non‐surgical treatment (220).

Recently, the American National Health and Nutrition Examination Survey (NHANES) linked to the US Mortality Registry updated to 2011, was used to estimate all‐cause and specific causes of death in adults with BMI ≥40 Kg/m2 who underwent bariatric surgery supplemented with behavioural intervention (221). The authors also found that bariatric surgery supplemented with behavioural intervention could reduce mortality. Multivariate‐adjusted proportional hazard Cox models showed two‐ to seven‐fold increases in risk of all‐cause and specific‐cause mortality in this patient group compared with patients with BMIs of 30–32 Kg/m2 and BMIs of 25–28 Kg/m2.

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan

1.5.2 Can surgical outcomes be predicted?

Type 2 diabetes mellitus (T2DM) Obesity‐related type 2 diabetes mellitus (T2DM) as an indication for bariatric surgery is a clear paradigm shift in diabetes management in recent years (222, 223). Since each bariatric procedure is also associated with some risks and surgical complications, it is important to apply tools that can better predict the long‐term resolution of T2DM to improve clinical decisions and help set realistic outcome expectations for patients and healthcare systems. Predictive tools for diabetes remission after bariatric surgery have been suggested and are widely‐applied, such as the DiaRem score (224), ABCD score (225), Diabetes Remission Score (DRS) (226) and Individualized Metabolic Surgery (IMS) score (227).

(a) DiaRem Score The DiaRem scoring system was initially a validated tool for evaluating the likelihood of remission of T2DM after RYGB (224), and has been used for other bariatric surgical procedures such as AGB and SG lately (228‐231).

The DiaRem components include age, pre‐operative HbA1c level, and types of pre‐operative anti‐ diabetes medications to calculate a score ranging between 0–22 across four variables (Table 1.3) (224).

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan

Table 1.3 DiaRem scores (224) Variable Score Age (years) <40 0 40‐49 1 50‐59 2 ≥60 3 HbA1c (%) <6.5 0 6.5‐6.9 2 7.0‐8.9 4 ≥9.0 6 Glucose‐lowering drugs No 0 Yes 3 Insulin therapy No 0 Yes 10 Overall score (sum of the four components) 0‐22 Probability of remission in each DiaRem score subgroup 0‐2 83‐90 3‐7 61‐70 8‐12 24‐40 13‐17 12‐21 18‐22 0‐9 Abbreviation: HbA1c=Glycated haemoglobin (from Still CD, Wood GC, Benotti P, Petrick AT, Gabrielsen J, Strodel WE, et al. Preoperative prediction of type 2 diabetes remission after Roux‐en‐Y gastric bypass surgery: a retrospective cohort study. The Lancet Diabetes & Endocrinology. 2014;2(1):38‐45)

As shown in Table 1.3 above, the 4‐point score ranging from 0 to 3 is used for the simple factors of age and HbA1c level. With respect to medications, there is a 3‐point score for classical oral hypoglycaemic agents (OHAs) and a 10‐point score for insulin treatment. The total scores for each variable are summed. Patients with lower scores are predicted to have a higher probability of T2DM remission after bariatric surgery.

The easy‐to‐use DiaRem score from the USA (224, 232, 233) has been shown to provide acceptable predictive power for diabetes remission in various populations 1‐year post‐RYGB,

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan

with a predictive performance superior to other scores (232). The DiaRem has also been shown to predict diabetes remission following other types of bariatric surgery in the longer term (>1 year) (229, 231). The score is easily implemented in clinical practice since it relies on simple pre‐ operative patient characteristics and basic clinical parameters, rather than less‐conventional or non‐standard biomarkers, like C‐peptide, which is included in the ABCD score system (229) described next.

(b) ABCD/Diabetes Surgery Score Lee and co‐authors (2013) adopted the parameters of age, BMI, C‐peptide level, and duration of diabetes for scoring (0–10) in the ABCD score to calculate the remission of T2DM in patients undergoing bariatric surgery. Details of the ABCD score are presented in Table 1.4:

Table 1.4 ABCD score (225) Variable Point on ABCD score 0 1 2 3 Age (years) ≥40 <40 BMI (Kg/m2) <27.0 27.0‐34.9 35.0‐41.9 ≥42.0 C‐Peptide concentration (mg/L) <2.0 2.0‐2.9 3.0‐4.9 ≥5.0 Duration of diabetes (years) >8 4‐8 1‐3.9 <1 Abbreviation: BMI=Body mass index (from Lee W‐J, Hur KY, Lakadawala M, Kasama K, Wong SK, Chen S‐C, et al. Predicting success of metabolic surgery: Age, body mass index, C‐peptide, and duration score. Surgery for Obesity and Related Diseases. 2013;9(3):379‐84)

For age, 1 point is given to <40 years old. Each of the remaining three variables has four categories to which 0–3 points are assigned. The points for each variable are summed to give a total ABCD score ranging from 0 to 10 points. Patients with higher ABCD scores are predicted to have a higher probability of T2DM remission after bariatric surgery.

(c) Diabetes Remission Score (DRS) Ugale et al. (226) created the Diabetes Remission Score (DRS) for predicting diabetes remission in patients with T2DM undergoing ileal interposition with SG or ileal interposition with diverted SG. Seven components are included in the score: age, BMI, duration of diabetes, macrovascular

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan complications, microvascular complications, pre‐operative insulin use, and stimulated C‐peptide. Each of these seven variables is scored by a two‐point scoring system ranging from 7 to 12. Lower scores indicate a higher probability of achieving diabetes remission after bariatric surgery.

1.6 Publicly funded bariatric surgery

Publicly funded multidisciplinary specialist obesity services play an important role in the management of socioeconomically‐disadvantaged patients with clinically severe obesity. Demand for these services is increasing, driven by a large number of potentially eligible patients. In an effort to address this dilemma, several Australian state and territory health departments, including the state of New South Wales (NSW), have introduced publicly funded specialist obesity services (16, 234) that provide surgical treatments.

Patients referred for specialist obesity services have often been unable to sustain their weight loss and have complex care needs requiring the evaluation and treatment of multiple medical conditions (16) for which these patients may require access to bariatric surgery. Approximately one million Australian adults have obesity with complications and are potentially eligible for bariatric surgery, making them candidates for referral to specialist obesity services (235). Apart from Australia (6), other countries with publicly funded healthcare systems that are offering bariatric surgery treatment are the UK (236, 237), USA (238), Canada (239‐241), New Zealand (242), France (243, 244), Brazil (245, 246), and Norway (41).

Despite most Australians relying on the public health system, the vast majority of all bariatric procedures (93.9%) are performed in private hospitals, as tabulated in Table 1.5 (6).

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan

Table 1.5 Distributions of bariatric procedures stratified by public and private hospitals (6) Primary Procedures Revision Procedures Total Procedures n (%) n (%) n (%) Public Private All Public Private All Public Private All Number of 918 15,915 16,833 370 3,840 4,210 1,288 19,755 21,043 procedures (5.5%) (94.5%) (80.0% of (8.8%) (91.2%) (20.0% of (6.1%) (93.9%) (100.0%) total total procedures) procedures)

(from The Australian and New Zealand Bariatric Surgery Registry Steering Committee. Bariatric Surgery Registry 2018/19 Report. Melbourne, Australia: Monash University; 2019)

The majority of patients with clinically severe obesity and a range of complex health issues are restricted from the publicly funded bariatric surgery services due to limited surgeries each year, prolonged wait times, geographical inaccessibility to the services, lack of long‐term follow‐up systems and out‐of‐pocket costs (16, 70).

1.7 Follow‐up after bariatric surgery

While it is clear that patients who have had bariatric surgery are likely to have improvement in multiple health outcomes, ongoing lifestyle intervention remains crucial for long‐term weight maintenance and monitoring of obesity‐related comorbidities. Other important considerations are nutritional and bone issues, psychological well‐being and post‐operative complications (e.g. dysphagia, GORD and oesophageal stricture). Hence, long‐term involvement of a multidisciplinary team and medical follow‐up are essential for all bariatric surgery patients. This has enabled the team with diverse skills from multiple health disciplines to effectively treat all aspects of the patients’ obesity and its associated comorbidities. Theoretically, patients undergoing bariatric surgery are advised of many behavioural recommendations to follow after their operations, including attending regular follow‐up appointments and closely adhering to dietary recommendations (e.g. consuming and prioritizing adequate protein), regular exercise, and daily vitamin supplementation. Multidisciplinary team also plays important roles in communicating with each other regularly regarding ongoing patient care delivery, while facilitating screening and referrals for any obesity‐related comorbidities.

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan

1.7.1 Definitions of adherence to follow‐up

Besides adherence to follow‐up appointments, non‐adherence to dietary supplementation recommendations is widely recognized as a critical causal factor in nutritional deficiency after bariatric surgery (247, 248). In this context, lack of adherence to post‐operative recommendations has been linked with poorer outcomes after bariatric surgery and appears to be a key component of less successful outcomes (249). On top of that, the lack of consistent measurement of adherence to post‐operative behavioural recommendations may also contribute to the apparent variability in adherence rates.

One challenge in interpreting the results of studies on follow‐up appointment adherence is the conflicting definitions of adherence. Some studies determined appointment attendance based on the percentage of scheduled visits that were attended, while others used attendance at a specific time‐point as to yearly post‐operative clinic visit. Studies also differed in the types of visits considered, for example, visits with surgeons versus other care providers in a multidisciplinary bariatric surgery team, which may include a medical provider (physician expertized in pharmacotherapy, surgeon with expertise in surgery or nurse), a dietitian, an exercise physiologist and a psychologist. In addition, bariatric surgery programs may have different practices in terms of supporting or enhancing patient attendance, such as providing frequent reminder calls or incentivising attendance (248). Noting general patterns in adherence may be the most beneficial way to interpret these studies.

A recent comprehensive systematic review was conducted to assess the measurement of appointment attendance adherence in bariatric surgery, which included 44 eligible articles (249). There seems no standardized methodology used to define adherence to follow‐up appointments. The review showed that the most‐commonly used definition of adherence to follow‐up appointments was whether patients attended a percentage of recommended visits/minimum number of visits (31% of the included studies, n=14) (249). On the other hand, 24% of studies reported appointment adherence as a frequency or percentage of appointments attended

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan without direct adherence (249). Even in studies that used a dichotomous definition of adherence or non‐adherence, definitions that were used are again varied significantly, including “attendance at ≥1 appointment during a specified period”, “attendance at 50‐75% of scheduled visits”, and “attendance at all recommended visits”. One‐third of the reviewed studies did not clearly define which provider the patient was scheduled to see during their visit to a multidisciplinary team, whether it was a physician, nurse, dietitian or surgeon. The wide variability in the definition of follow‐up adherence contributes to the broad range of reported adherence rates.

1.7.2 Importance of adherence to follow‐up after bariatric surgery

Long‐term success in weight loss and improvement in comorbid conditions after bariatric surgery are thought to largely depend on the patient’s ability to adhere to a complex set of behaviours, including regular attendance at follow‐up appointments and following stringent dietary, exercise and vitamin recommendations. The wide range of both pre‐existing and post‐surgery issues that can be experienced by patients also highlights the necessity for good‐quality, long‐term follow‐ up care (250).

The American Association of Clinical Endocrinologists/The Obesity Society/American Association of Metabolic and Bariatric Surgery guidelines recommend routine metabolic and nutritional monitoring for all patients who have had bariatric surgery as well as a review of weight, eating behaviours and physical activity (131, 146). Additionally, the European Association for the Study of Obesity (EASO) has also published detailed guidelines for non‐specialists, highlighting the need for longer‐term follow‐up for patients following bariatric surgery (219). Within this framework, the EASO recommended a strategy that longer‐term follow‐up needs be transferred to primary‐ care physicians and other non‐surgical healthcare professionals when clinically appropriate (219).

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan

Unfortunately, adherence to required clinic visits or follow‐up care tends to be low in the bariatric surgery population with attrition rates as high as 89% depending on the procedure and follow‐ up program. Retention of patients becomes increasingly difficult as the time since bariatric surgery increases. A study found that 68–78% of patients attended 3‐, 6‐, and 12‐month follow up appointments, while only 41% attended a 24‐month follow up (240). Another study reported that only 33% of patients who had RYGB attended a follow‐up at 2 years post‐surgery and only 7% attended at 10 years, despite multiple contact attempts (251). Long‐term attrition is problematic regardless of surgery type, with a MA of four studies (n=365) revealed a 6.4% greater %EWL in patients who had RYGB and who adhered to follow‐ups, compared to those who did not at 12‐months post‐surgery (252). In addition, a database review of over 51,000 patients found that attendance was associated with greater weight loss outcomes at 12 months in patients undergoing RYGB and SG (253).

1.7.3 Predictors of and reasons for loss to follow‐up

Given the growing demand for bariatric surgery as a treatment modality for obesity, examining the process of the pre‐surgical assessment phase may identify factors associated with surgical outcomes and attrition. Numerous studies have examined demographic variables that are associated with or predict appointment attrition. These include pre‐operative weight, sex, age, employment status, presence of comorbid conditions, distance from clinic, and insurance coverage (248). Notwithstanding, results have been mixed (248).

It is also imperative to understand the post‐operative barriers to care by quantitatively examining patient and operational factors contributing to attrition. Improvement in the operational efficiency of bariatric surgery programs would contribute to wider access to care for patients deemed suitable for surgical intervention (239). A better understanding of this gap is needed.

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan

1.8 Pre‐operative care/weight loss and its implications on post‐surgery weight loss

Pre‐operative management and assessment of candidates for bariatric surgery have long been considered a critical part of post‐operative success (131). In the USA, this tradition has been tied to third‐party payers who have required a pre‐operative medically supervised weight‐loss program and counselling before surgical treatment. In some states of Australia, especially NSW, a normally one‐year pre‐operative weight‐loss program is often mandated for publicly funded patients. This helps to understand patients’ lifestyle changes needed for bariatric surgery to be successful, while also aims to properly assess and manage all medical problems. These have led to recommendations that pre‐operative weight loss be a routine component of the pre‐operative preparation process for bariatric surgery, and that amount of pre‐surgery weight loss has been suggested to be a potential positive predictive factor for long‐term bariatric surgery success. However, the evidence in the current literature on the effects of pre‐bariatric surgery care as well as weight loss prior to bariatric surgery are mixed (254‐260), with the majority of studies, especially recent ones, suggested that long‐term mandated pre‐operative weight loss periods should be abandoned (254‐256, 261, 262).

For example, a study in the USA retrospectively examined the relationship between pre‐ operative care, and post‐operative weight loss and follow‐up in the first two years post‐ operatively in 1,303 patients who underwent an index bariatric surgical procedure between 2009 and 2014 (mean BMI=48±8.6 Kg/m2) (255). Pre‐operative follow‐up was defined as documented contact with a clinical member of the team before surgery and was coded as individual visits, support group encounters, telephone encounters, electronic messaging, or educational sessions. In this study, the percentage of total weight loss (%TWL) at 6, 12, and 24 months post‐operation was the primary outcome variable. Secondary outcomes assessed were completion of visits with the bariatric surgery team at recommended timepoints (6, 12, and 24 months), which was referred to as post‐operative follow‐up. It was found that the intensity and length of pre‐ operative follow‐up were not associated with post‐operative weight loss or follow‐up. A greater

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan number of individual visits with the bariatric surgery team and additional psychologist visits before surgery were associated with less post‐operative weight loss (p<0.05).

Another retrospective chart review of 218 patients by Krimpuri and co‐authors (254) from Ohio investigated if pre‐operative weight loss is the best predictor of post‐operative weight loss. The findings demonstrated that weight loss preceding bariatric surgery was a significant predictor of 1‐year change in post‐operative BMI (as a single independent variable) (p=0.006); and became non‐significant when age, race and sex were added to the regression equation (p=0.543). The authors indicated that pre‐operative weight loss should not be considered in isolation when clearance for bariatric surgery is being evaluated.

Some other older systematics reviews have been conducted. For example, a review by Livhits and team (258) studied the effect of pre‐operative weight loss on post‐operative outcomes in bariatric surgical patients. It covered multiple pre‐operative predictors of post‐operative weight loss and found mixed results, with seven of eligible 14 studies showing positive association between mandatory pre‐operative weight loss in the weeks immediately preceding surgery and post‐operative weight loss; six showed no association and one showed negative association (258). With the presence of large amount of variation in pre‐operative weight loss methods (e.g. informal/unclear prescription, structured program, meal replacement products), this review did not assess the best way to achieve this weight loss upfront (258). The components of a successful pre‐operative care program, such as exercise plan and nutritional education, alongside the most effective dietary plan to maximise weight loss and minimise surgical complications among the bariatric surgical candidates were also unreported (258). Consistent with some of these previous findings, Cassie and co‐authors (263) reviewed 27 studies also revealed that 62.5% of them (n=15) found no beneficial effect of pre‐operative weight loss on post‐operative weight loss. These results are similar to those of another systematic review of studies that evaluated the relationship between pre‐operative weight loss and post‐operative outcomes by Ochner and co‐ workers, focusing on the content and effectiveness of pre‐operative diets and post‐operative weight loss results of 2 RCTs, 5 prospective studies and 6 retrospective studies, also concluded

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan that these existing literature does not support mandatory or necessary weight loss before bariatric surgery (257). In fact, the authors expressed reservations that a pre‐operative weight loss requirement may lead to denial of surgery and subsequent health benefits to patients who are unable to achieve a pre‐specified amount of weight loss (257). An updated literature review by Gerber and colleagues (259) also determined that much of the research on the impact of pre‐ operative weight loss, regardless of length of dietary interventions, on post‐operative outcome is inconsistent. Kim et al. (264) reviewed studies that focused on mandated pre‐operative weight loss, indicating that there was no data from any RCT, large prospective study or MA to support the practice of insurance‐mandated pre‐operative weight loss in the USA (264).

In a nutshell, there is little strong evidence to support or refute a pre‐operative weight loss requirement for bariatric surgery candidates. Contrarily, the current literature suggested that pre‐operative requirements may not be necessary or safe. The conclusion is, however, defensible. The lack of details on pre‐operative education and direction calling into question the validity and overall clinical applicability of the outcomes. Both recommendation and level of clinical intervention during the pre‐operative weight loss programs were either inconsistent or not discussed, making any discernible effective pre‐operative interventions may have on post‐ operative outcomes difficult to evaluate. Benefits of pre‐operative care in optimizing patients’ medical and psychological conditions, and potential in decreasing surgical risks were also not appropriately studied. Nevertheless, the question of which pre‐operative programs are most effective is important only if the actual pre‐operative weight loss is related to post‐operative outcomes, including weight loss, medical and psychological conditions, and surgical complication rates. It is also crucial to consider if a pre‐operative weight loss mandate would carry a risk of denying bariatric surgery to candidates who were unable to meet this requirement but would nonetheless benefit from the surgery (265).

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan

1.9 Rationales for the research directions and aims: Bridging the knowledge gaps, clinical needs and research contributions

The literature review highlighted the various clinical, epidemiological, biochemical, physical, nutritional and physiological aspects of obesity and their intricate links with metabolic diseases. It importantly demonstrates significant increases in the more severe form of obesity over the last few decades, along with predispositions to consequential comorbidities that result in increased all‐cause morbidity and mortality. To date, bariatric surgery is the most powerful treatment for more severe obesity and its related comorbidities. As such, obesity, with its implications for comorbidities and mortality are now emerging as increasingly recognised serious public health concerns. Bariatric surgery also provides an excellent platform for research, as the bariatric surgical cohort is a high‐risk population with both the greatest degrees of obesity and higher rates of intimately associated comorbidities.

Yet, there are numerous key knowledge gaps in a range of areas in the setting of clinically severe obesity and surgical treatment identified. There is limited understanding of the long‐term outcomes of bariatric surgery across all domains, ranging from epidemiology, effectiveness, durability, diagnostic modalities, patient adherence and multidisciplinary management, which are particularly pervasive in public healthcare systems. This poses considerable difficulty in developing clinical guidelines and subsequent management of the bariatric surgical population in the publicly funded services. The hallmark of undergoing bariatric surgery in the publicly funded bariatric surgery service in Australia, especially the state of NSW, is having clinically severe obesity (defined as class III obesity alone or a BMI ≥35 Kg/m2 with at least one major obesity‐related comorbidity).

The premise for the series of research studies contained in this thesis arose from the key areas of knowledge gaps summarised from the literature review and key observations from our clinical bariatric surgical practice that restricted optimal obesity management in the Australian publicly funded bariatric surgery services. This was identified as associated with the lack of access to publicly funded bariatric surgery services, limited long‐term evidence, and uncertainties in

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan patients with extreme obesity and at least one related comorbidity. By basing the research of highly complex obesity and its interconnecting elements in a publicly funded bariatric surgical cohort, it was possible to capitalise on a highly‐valuable interface of in‐depth knowledge deficiency, real‐world clinical needs and research contributions. In this thesis, a broad range of important questions regarding the multidisciplinary management of clinically severe obesity in the bariatric surgical cohort in three public hospitals was closely examined and systematically addressed.

The body of work presented in this thesis was written around the following four main research themes, each comprising specific objectives and studies that addressed key gaps in the understanding of multidisciplinary management of bariatric surgery in the context of clinically severe obesity. Each study also involved unique and advanced modellings, permitting the robust and reliable exploration of the following outcome measures: 1. Long‐term effectiveness, health impacts and pitfalls of bariatric surgery; 2. Challenges and prediction of adherence to post‐bariatric surgery care; 3. Role of pre‐operative weight loss; and 4. Prediction of T2DM remission using the DiaRem algorithm, and its applicability as a diagnostic tool in everyday practice in the bariatric surgical population.

RESEARCH THEME 1: Long‐term effectiveness and safety of bariatric surgery in clinically severe obesity and its associated comorbidities Knowledge gaps and clinical needs Bariatric surgery reliably attains substantial weight loss, with proven resolution and improvement of metabolic diseases. However, there is a paucity of understanding of the long‐term (>5 years) management, health benefits, and safety of publicly funded bariatric surgery in the setting of clinically severe obesity. The data on outcomes for highly‐complex populations that have undergone newer bariatric procedures such as SG was limited. This include weight loss, weight sustainability, patterns of change in obesity‐related comorbidity, nutritional status, surgical risks and mortality following bariatric surgery between 6‐9 years period, all of which have not been

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan well explored. Concerns regarding the balance of benefits and harms of surgery for people with super obesity (SO) was unresolved.

Objectives As a prelude to more specific studies, this research theme studied the Edmonton Obesity Staging System (EOSS) stages as well as the BMI levels of the study bariatric surgical cohort to screen the obesity severity of the patients. Subsequently, the effectiveness and durability of bariatric surgery on the long‐term weight change, a full spectrum of obesity‐related comorbidities and surgical consequences were studied at a multisite publicly funded bariatric surgery service.

Therefore, the specific aims of this longitudinal study were: 1) To develop an electronic database by retrospectively capturing unified data of various aspects. Besides the research purposes, this data meaningfully provides an opportunity for great utility in the clinical obesity services for future ongoing patient monitoring; 2) To measure the EOSS stages of the bariatric surgical cohort; 3) To determine the BMI levels of the patients; 4) To detail the effect of multidisciplinary bariatric surgery management on the weight loss over 6 to 9 years post‐surgery; 5) To assess the significant impact of bariatric surgery on changes in comorbidity status, particularly T2DM, hypertension, hyperlipidaemia, osteoarthritis (OA), weight‐bearing joint pain (WBJP), sleep‐disordered breathing, depression, severe anxiety, and hyperuricaemia. To gain an even finer understanding of the powerful effect of bariatric surgery, clear changes in the status of obesity‐related comorbidities were further calculated based on laboratory findings, medications, physicians’ diagnoses, orthopaedic surgery registry linkages, and physical measurements to generate the prevalence, resolution, improvement, persistence, worsening and incidence of the metabolic diseases; 6) To evaluate peri‐ and post‐surgical complications; 7) To address pre‐ and post‐operative nutrient outcomes; 8) To characterise mortality following bariatric surgery; and

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan

9) To investigate the impact of bariatric surgery on weight loss, length of hospital stay (LOHS), surgical outcomes, and obesity‐related comorbidity changes in individuals with SO, who are often labelled as a “challenging” population.

RESEARCH THEME 2: Adherence to post‐bariatric surgery care Knowledge gaps and clinical needs A key contributing factor to the challenges in managing and studying clinically severe obesity is the considerable proportion of patients who do not attend routine appointments post‐surgery. This is potentially risky for the bariatric surgical cohort due to its disruption of the long‐term management of weight loss, diseases, nutritional deficiencies and potential surgical complications. The inter‐correlation between adherence to post‐surgical routine clinic appointments and health outcomes is not fully understood. Little is known of the underlying barriers that contributed to withdrawal from the healthcare services. The pre‐operative predictors of attrition from attending clinic reviews after bariatric surgery have not been well described. These are particularly important, as the outcomes may guide future patient selection or clinical care practices for patients needing additional supports. The findings also make essential to contributions to the establishment of guidelines for the multidisciplinary management of bariatric surgery in public services.

Objectives Using both prospective and retrospective study designs, and various means including questionnaires, patient contacts and interviews, this study aimed: 1) To determine the rate of adherence to clinic reviews after bariatric surgery; 2) To understand the reasons for ceasing attendance at clinic reviews; 3) To determine the predictors of adherence to post‐surgery clinic reviews; and 4) To examine whether there is any relationship between adherence to post‐operative clinic review attendance and weight outcomes over long‐term.

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan

RESEARCH THEME 3: Pre‐operative weight loss as a requirement for publicly funded bariatric surgery Knowledge gaps and clinical needs There has been conflicting literature on the utility of pre‐operative weight loss for patients undergoing bariatric surgery in public hospitals. This currently‐mandated requirement is often a one‐year integrated pre‐operative behavioural weight management program (WMP). There is little evidence that it improves weight loss outcomes after bariatric surgery.

Objectives This study aimed to explore the relationship between pre‐ and post‐bariatric surgery weight loss in order to determine the need for lifestyle weight‐loss interventions by specialist obesity services prior to publicly funded bariatric surgery.

RESEARCH THEME 4: Prediction of T2DM remission in bariatric surgical patients Knowledge gaps and clinical needs Review of the literature revealed realistic limitations that exist in predictive and diagnostic tests for T2DM remission in patients with obesity. Complex scoring systems hindered their widespread use due to feasibility issues. The current T2DM scores were also poorly validated in exclusively bariatric surgical cohorts. There is also an increasing clinical demand for bariatric surgery and the need to establish useful and clinically‐applicable tools to predict T2DM outcomes that are highly interconnected with obesity. All these highlighted the pressing need for further validation and development in this area, in order to examine the suitability and generalizability of DiaRem use in an Australian bariatric surgical population. The application of DiaRem algorithm – a simple, non‐invasive and structured tool – in the Australian bariatric surgical cohort with both clinically severe obesity and T2DM has not been assessed.

Objectives We aimed to identify and validate the short‐term and longer‐term (1‐ and 5‐year) post‐operative T2DM remission prediction using the DiaRem tool in the bariatric surgical cohort with T2DM.

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan

The conceptual framework and study outline in the research entirety and each individual study in this thesis are illustrated in Figure 1.15 and Table 1.6, respectively.

Collectively, through this series of studies, this thesis aimed to provide a better understanding of the long‐term public healthcare system‐focused knowledge, therapeutic strategies, patient adherence and management of bariatric surgery in a well‐characterised population in the setting of clinically severe obesity. The outcomes of this longitudinal study will potentially and directly impact upon patient care, expansion of current literature, addition to clinical knowledge, utility of predictive and diagnostic tools, development of guidelines, and management strategies for the treatment of the highest grade of obesity.

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan

Figure 1.15 Conceptual framework and research themes for the associated studies in this thesis

RESEARCH THEME 1 – Health outcomes ‐ Super obesity versus morbid obesity 6‐Year modelled weight trajectories, health outcomes and surgical complications . Modelled weight changes after bariatric surgery . Changes of statuses of comorbidities . Socio‐demographics (remission, improvement, persisting and . Observed and modelled weight changes (weight, BMI, %TWL) worsened) for T2DM, hypertension and . Edmonton Obesity Staging System (EOSS) hyperlipidaemia . Changes of statuses of comorbidities (remission, improvement, persisting and worsened) . Surgical complications for T2DM, hypertension and hyperlipidaemia . Prevalence of T2DM, hypertension, hyperlipidaemia, OA and/or WBJP, OSA/OHS, depression and/or severe anxiety, and hyperuricemia Australian Orthopaedic Association . Total joint arthroplasty Data Linkage National Joint Replacement Registry . Medications (antidiabetic, antihypertensive, lipid‐lowering agents, opiates, CPAP/BiPAP, antidepressants and antianxiety agents) . Blood tests [HbA1c, FBG, lipid profile (total cholesterol, triglycerides, HDL‐C and LDL‐C),

urate, transferrin saturation, ferritin, Vitamin D, Vitamin B ] 12 Publicly Funded Bariatric Surgery . Blood pressure . SG | MGB‐OAGB | AGB | RYGB Nutritional status (iron deficiency anaemia, vitamin D and vitamin B12) . Surgical complications . Mortality RESEARCH THEME 2 – Follow‐up study Adherence to multidisciplinary follow‐up after bariatric surgery RESEARCH THEME 4 ‐ Predictors of loss to follow‐up Prediction of T2DM remission following . Socio‐demographics RESEARCH THEME 3 bariatric surgery using DiaRem algorithm . Distance to hospitals Does weight loss before bariatric surgery . Pre‐operative lifestyle behaviours . Socio‐demographics predict post‐operative weight loss? . Baseline comorbidities & medications . DiaRem scoring system . Socio demographics . LOHS . Weight change ‐ . Initial weight . Weight changes . Area under the curve (AUC) of the Receiver . Pre and post operative weight changes . Surgical complications Operating Characteristic (ROC) ‐ ‐ . Initial and baseline comorbidities . Predictive ability ‐ Reasons for loss to follow‐up . Multiple linear regression models . Calibration ‐ Outcomes . Modelled weight changes 70

CHAPTER 1 – Introduction and Literature Review | Michelle M.C. Tan

Table 1.6 Study outline for 10 years of pre‐ and post‐operative follow‐up Data collection Initial Pre‐op Bariatric 3‐Month 6‐Month 1‐Year 18‐ 2‐Year 3‐Year 4‐Year 5‐Year 6‐Year 7‐Year 8‐Year 9‐Year timepoint visit to baseline surgery post‐op post‐op post‐op Month post‐op post‐op post‐op post‐op post‐op post‐op post‐op post‐op clinic post‐op Variable Informed consent √ Referral and follow‐up √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ clinic(s) Date of birth √ Sex √ Race √ Marital status √ Smoking status √ Alcohol‐drinking √ Postcode √ √ √ √ √ √ √ √ √ √ √ √ √ √ Employment status √ Date of surgeries √ Bariatric surgery (Primary) √ Bariatric surgery √ √ √ √ √ √ √ √ √ √ √ √ (Secondary) Length of hospital stay √ √ √ √ √ √ √ √ √ √ √ √ √ Height √ Weight √ √ √ √ √ √ √ √ √ √ √ √ √ √ Clinical diagnosis of √ √ √ √ √ √ √ √ comorbidities Blood pressure √ √ √ √ √ √ √ Medications √ √ √ √ √ √ √ Blood tests √ √ √ √ √ √ √ Peri‐op complications √ √ Post‐op complications √ √ √ √ √ √ √ √ √ √ √ Mortality √ √ √ √ √ √ √ √ √ √ √ √ √ Abdominoplasty √ √ √ √ √ √ √ √ √ √ Follow‐up status √ √ √ √ √ √ √ √ √ √ √ √ Reasons for LTFU √ √ √ √ √ √ √ √ √ √ √ √ Total joint arthroplasty √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ Abbreviations: op=Operation; m=Metres; Kg=Kilograms; LTFU=Loss to follow‐up

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

CHAPTER 2

RESEARCH THEME 1

Long‐term weight trajectories, health outcomes and surgical complications after bariatric surgery of publicly funded patients with clinically severe obesity

2.1 ABSTRACT

Introduction: Obesity has emerged as a global epidemic with a disproportionate rise in class III obesity [body mass index (BMI) ≥40 Kg/m2]. Long‐term results (defined by >5 years) of the durability of weight loss, health improvements and safety outcomes after bariatric surgery are needed. There is a limited understanding of the management, benefits and safety of publicly funded bariatric surgery in the context of clinically severe obesity [defined as class III obesity alone or a BMI ≥35 Kg/m2 with at least one major obesity‐related comorbidity] and its associated metabolic comorbidities, especially in the long‐term. This multicentre longitudinal study was conducted to assess the effectiveness and safety of the current range of bariatric surgical procedures in publicly funded patients with clinically severe obesity in three public hospitals in geographically diverse districts in Sydney, Australia, to help guide clinical decision‐making.

Objectives: The overarching aims of this chapter were to closely examine long‐term weight change, health status and surgical complications following laparoscopic sleeve gastrectomy (SG), mini gastric bypass‐one anastomosis gastric bypass (MGB‐OAGB), adjustable gastric banding (AGB) and roux‐en‐Y gastric bypass (RYGB).

Settings: Royal Prince Alfred Hospital, Concord Repatriation General Hospital, and Camden Hospital, Sydney, New South Wales (NSW), Australia.

Methods: As a part of the multidisciplinary referral‐based publicly funded bariatric practice, all the 168 patients with clinically severe obesity and who were deemed suitable for bariatric surgery were enrolled into a minimum of one‐year integrated physician‐led weight management program (WMP). The WMP incorporated dietary intervention, behavioural management and exercise prescription. Following this, bariatric surgery was undertaken among the study cohort between 2009 and 2017, with at least one year of follow‐up through 9 years post‐surgery until 31st December 2019. Patient socio‐demographics, comorbidities, medication prescriptions, clinical measurements, blood tests, nutritional status and surgical complications were thoroughly recorded and assessed retrospectively in this cohort study. Linear mixed‐effects models taken into account repeated measures and a generalised mixed‐effects regression, both of which accounted for random effects at the individual level were applied to model weight changes of the patients over a 6‐year period.

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

Main Outcome and Measures: The primary outcomes were weight loss, expressed as BMI change, weight change in kilograms (Kg) and percentage of total weight loss (%TWL) from pre‐ operative baseline. The secondary endpoints were peri‐ and post‐operative surgical complications; mortality; nutrient deficiencies; and obesity‐related comorbidities including type 2 diabetes mellitus (T2DM), hypertension, hyperlipidaemia, osteoarthritis (OA), weight‐bearing joint pain (WBJP), sleep‐disordered breathing [obstructive sleep apnoea (OSA) and obesity hypoventilation syndrome (OHS)], depression, severe anxiety, and hyperuricaemia determined by clinical diagnoses, physical measurements, laboratory testings, medication use, medical records and surgical reports. Post‐operative course of T2DM, hypertension and hyperlipidaemia were calculated from baseline and classified into four main statuses, namely remission, improvement, persistence and worsening post‐operatively, in addition to prevalence and incidence rates. Full total joint arthroplasty (TJA) data alongside the OA diagnoses were obtained via data linkage from the nationwide Australian Orthopaedic Association National Joint Replacement Registry (AOANJRR). All the patient outcomes were assessed at baseline, 1, 2, 3, 4, 5, and 6 years post‐operative follow‐up. A subanalysis for super obesity (SO) context was conducted with all the patients further stratified into two groups based on their pre‐operative baseline BMI ‒ morbid obesity (MO) (BMI <50 Kg/m2) versus SO (BMI ≥50 Kg/m2); and their magnitude of weight loss, length of hospital stay (LOHS), surgical complications and health conditions were compared.

Results: Socio‐demographics, severity of study population and bariatric surgical information Patients were aged 21‐72 years, 66.1% were female, 71.4% were Caucasian, 57.1% were current/ex‐smoker, and 43.5% were unemployed on government support payments, with initial weight of 141.9 Kg (range: 83.5–272.8), a pre‐operative baseline mean weight 133.2 Kg (range: 80.3–239.9 Kg) corresponding to a mean BMI 48.0 Kg/m2 (range: 33.6‐78.8 Kg/m2). All the study patients had Edmonton Obesity Staging System (EOSS) of at least Stage 2 obesity at baseline (61.3%, 27.4%, and 11.3% Stages 2, 3 and 4 respectively), with a median of six obesity‐related complications prior to surgery (range: 1–12). The most performed primary procedure was SG (83.9%). All procedures were performed laparoscopically in a single bariatric centre with no conversion to open. The average LOHS was 3 days (range: 1–8). Six patients underwent a secondary bariatric procedure. Of the 34.5% (58 patients) with peri‐ and post‐operative surgical complications, none was life‐threatening. There were no 30‐day mortality, and the three deaths occurred were deemed unrelated to bariatric surgical operations. Weight loss Statistical modelling demonstrated substantial weight change was achieved over time (p<0.001 versus baseline); with a significant estimated marginal mean %TWL of baseline weight (±SE) of 24.1±1.1% at 1 year, 24.1±1.1% at 2 years, 22.1±1.1% at 3 years, 20.4±1.2% at 4 years, 19.0±1.3% at 5 years, and 18.8±1.5% at 6 years. OA, WBJP, opioid use and TJA OA and/or WBJP is the commonest obesity‐related comorbidities in the study cohort, and a steady trend was identified indicating bariatric surgery most effectively improved OA and/or WBJP symptoms following marked weight loss secondary to bariatric surgery at all the 6 annual follow‐up timepoints. After bariatric surgery, prevalence of prescribed opioid analgesic use or

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan dependency mostly increased to surpass baseline prevalence (11.3%), with patients either initiated their opioid or continued with their opioids during the follow‐up assessment, suggesting the need for alternative methods of pain management in the patients with extreme obesity. Nonetheless, the prevalence of patients undergoing TJA declined over time with variability in patterns following surgical weight loss. Sleep‐disordered breathing (OSA/OHS) There appeared to be a quadratic trend to sleep‐disordered breathing (i.e. OSA/OHS) throughout 6 years of follow‐up period. The prevalence of CPAP or BiPAP device required by the study patients at baseline was as high as 51.8%. Of the 32 patients with OSA and were prescribed CPAP at baseline, 36.8% no longer required CPAP at their last observations following bariatric surgery. T2DM, hypertension and hyperlipaemia The prevalence of other obesity‐related comorbidities studied, including T2DM and hypertension decreased after bariatric surgery and remained lower than baseline as time progressed, but not the case with hyperlipidaemia. Among those with T2DM and hypertension at baseline, drastic composite cumulative rates of remission and improvement seen in majority of the patients, being 77.1% and 70.0%, respectively. With the T2DM, the remission rates decrease over time but remained high in long‐term, i.e. 52.3%, 50.0%, 48.3%, 50.0%, 46.5%, and 35.5% at post‐surgical 1, 2, 3, 4, 5, and 6 years, respectively. There appeared to be a similar trend in hypertension remission rates, the proportions in remission were 39.6%, 32.1%, 37.3%, 31.5%, 30.2%, and 34.4% from years 1 to 6 post‐surgery. Whereas hyperlipidaemia showed no such trend. Hyperlipidaemia based on our strict definition of assessing all the four subfractions of lipid profile, in addition to medication use, produced a result of approximately half of the patients remained persisting hyperlipidaemia across all timepoints. Despite the stringent assessment criteria, over one‐third of our patients achieved remission or improvement in their hyperlipidaemia after surgery. Depression and/or severe anxiety Compared with pre‐surgery (47.0%), the prevalence of patients having depression and/or severe anxiety was lower between 1 and 4 years post‐surgery, but surpassed baseline level at post‐ operative 5 years (50.0%) and 6 years (48.8%), respectively. Given the high rates of mental illness in the patients, pharmacotherapy is commonly seen, i.e. 31.1% patients at year 1 post‐surgery and hit a new high record through year 6 (45.0%). Hyperuricaemia Bariatric surgery is associated with decreased serum uric acid levels in the study cohort, evident from pre‐operative baseline prevalence of hyperuricaemia of 30.4% improved to 12.5% at year 6 post‐surgery. The mean pre‐operative serum uric acid level was 0.36±0.10 mmol/L and dropped from first year of surgery onwards through 6 years of follow‐up. SO versus MO Statistical modellings show that substantial weight loss was achieved by both the SO and MO groups at all time‐points following bariatric surgery (p<0.001 versus baseline), but no significant difference between the two groups in terms of magnitude of %TWL, probability of successful ≥20 %TWL, post‐operative obesity‐related comorbidity course, LOHS and surgical complications over 6 years time. Therefore, we concluded that being at the highest level of obesity, SO, had negligible impact on outcome measures compared to the lighter MO group.

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

Nutrient deficiencies Although the mean levels of serum vitamins and trace minerals were at adequate levels before and after bariatric surgery, we found a high prevalence of nutrient deficiencies in the study cohort, mainly of vitamin D and iron. Pre‐operatively, vitamin D deficiency was noted in 31.9% of patients, and decreased significantly after bariatric surgery. Whereas iron deficiency anaemia doubled at year 6 post‐operation (14.3%) than that of pre‐surgery (7.1%). As for vitamin B12 insufficiency, low prevalence was detected before and after bariatric surgery, with no patient developing a deficiency in years 5 and 6 post‐operatively (0%).

Conclusion: This study shows bariatric surgery in public hospitals using a multidisciplinary approach is durable, safe and effective. Bariatric surgery appears to be equally safe and effective in the super obese and morbidly obese groups.

Keywords: Bariatric surgery, Long‐term outcomes, Publicly funded, Multidisciplinary management, Clinically severe obesity, Public hospitals.

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

2.2 BACKGROUND

Obesity, a National Health Priority Area, is currently a major challenge facing the Australian population at a crisis level. It has also become a global epidemic with a disproportionate rise in class III obesity [body mass index (BMI) of ≥40 Kg/m2], causing substantial burden in obesity and the healthcare systems. There are approximately one million adults in Australia with severe levels of obesity, namely clinically severe obesity, defined as class III obesity alone or a ≥35 Kg/m2 with at least one major obesity‐related comorbidity. In an effort to address this epidemic, several Health Departments of Australian state and territories have launched multidisciplinary specialist obesity services (i.e. specialist hospital‐based multidisciplinary healthcare services for the management of obesity and its associated medical conditions), with a few successfully establishing publicly funded bariatric surgery services to manage highly complex patients with clinically severe obesity (16, 234).

In the past two decades, bariatric surgery has universally become an increasingly popular treatment for clinically severe obesity (13, 14, 18, 19, 266‐269). It has been proven the most effective treatment with superior results in weight reduction, and resolution or improvement of obesity‐related comorbidities, and contribute to enhanced quality of life and increased life expectancy (14, 151). Although generally accepted as the most effective means for inducing dramatic weight loss, typically losing 10‐15 BMI points, most evaluations of bariatric surgical outcomes have been hampered by inadequate and incomplete long‐term follow‐up, or significantly contrast procedures no longer performed today, such as the nonadjustable gastric band and vertical banded gastroplasty reported in the well‐respected Swedish Obese Subjects (SOS) study (192, 270). The Longitudinal Assessment of Bariatric Surgery (LABS) Consortium from the United States have published 3‐ and 7‐year outcomes from two bariatric surgical procedures, adjustable gastric banding (AGB) and roux‐en‐Y gastric bypass (RYGB) (13, 14). However, these reports do not include the widely‐accepted, stand‐alone bariatric operations gaining popularity and acceptance among the bariatric surgeons − sleeve gastrectomy (SG) and mini gastric bypass‐ one anastomosis gastric bypass (MGB‐OAGB). To date, there are continued gaps in evidence

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan about the long‐term effectiveness and durability (i.e. >5 years) of these newer operations that have supplanted the other procedures (24, 25). There is also significant variability in weight loss and durability outcomes within different study populations, which remains largely unexplained. Therefore, the primary aim of this study was to examine the effect of multidisciplinary bariatric surgery management on the weight loss in the study cohort over a long‐term follow‐up duration.

In the new decade, beyond effective and durable weight loss, bariatric surgery has also proven to offer metabolic benefits, resulting in a shift from mere weight management as the primary focus of bariatric surgery to the improvement or resolution of obesity‐related comorbidities. Yet, more information and understanding are needed about the longer‐term sustainability of weight loss, control for comorbidities associated with obesity, and surgical complications after bariatric procedures. In addition to questions about the long‐term sustainability of weight loss, it is important to understand whether remission and improvement in obesity‐related comorbidities are durable over time, and if there is worsening or emergence of incident conditions after surgical treatments. Many previous studies have not measured the progressive and combined measures (both medication use and blood tests) of obesity‐related comorbidities over a long‐term period (23, 271). If we were able to observe these changes at annual time points over a long period, this would greatly improve the understanding of the impact of bariatric surgery on metabolic diseases. Consequently, this would aid in defining evidence‐based targets and inform decisions for publicly funded bariatric surgery services. In line with such rationales, the effectiveness and durability of bariatric surgery on type 2 diabetes mellitus (T2DM), hypertension, hyperlipidaemia, osteoarthritis (OA), weight‐bearing joint pain (WBJP), sleep‐disordered breathing [including obstructive sleep apnoea (OSA) and obesity hypoventilation syndrome (OHS)], depression, severe anxiety and hyperuricaemia are critically examined and discussed in this study.

‘Diabesity’ is a term that was coined to clearly emphasize the close pathophysiologic interconnection between the metabolic diseases of T2DM and obesity (272). In recent years, the Australia Diabetes Society (ADS) and multiple international diabetes organizations (222, 223, 273) have endorsed and issued recommendation for bariatric surgery as a part of the standard

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan treatment algorithm for T2DM in adults with clinically severe obesity whose weight and associated comorbidities including hyperglycaemia is inadequately controlled by lifestyle interventions and medical therapy. A plethora of evidence has proven that bariatric surgical procedures induce remission of T2DM in 12% to 73% of surgically treated individuals, but most of these data demonstrated short‐ or mid‐term (≤5 years) efficacy of bariatric surgery (14, 18, 19, 41, 151, 274‐279). Traditionally viewed as an intractable chronic disease, complete remission of T2DM is now feasible following bariatric surgery. The durability of its effectiveness on T2DM remains a pressing clinical question. The wide range of T2DM remission rates reported in these studies, which also examined both medications and glycaemic control, is likely to be multifactorial. The heterogeneity can be attributed to the varied definitions of remission used, different populations studied, different levels of comorbidity severity included, and a diversity of bariatric surgical procedures performed. Furthermore, studies varied with respect to reporting styles, often either cumulative remission counted as patient who ever achieved remission in T2DM or to the last observation, or prevalent remission (i.e. only patients who were in T2DM remission at the time of measurement compared to pre‐operative baseline). As a further evidence of the impact of bariatric surgery on T2DM over long‐term follow up duration in a highly complex study cohort, we aimed to assess all the: changes in T2DM status (remission, improvement, persistence and worsening); cumulative remission and improvement to the end of observation period; as well as prevalence of T2DM.

Based on the growing body of evidence, patients struggling with obesity is also tightly related to increased cardiovascular risk, including hypertension (280, 281). Proposed mechanisms of actions include upregulation of the renin‐angiotensin system (RAS); rise in free fatty acids (FFA) and angiotensinogen; as well as the pro‐inflammatory effects of adipocytes, thromboxane A‐2 and a declined insulin sensitivity, thereby causing endothelial dysfunction, augmented arterial stiffness and eventually hypertension (281). These effects are ameliorated following weight loss, resulting in remission or improvement in obesity‐related hypertension. Like other outcome variables associated with obesity such as T2DM and hyperlipidaemia/dyslipidaemia in the bariatric surgical literature, definitions for hypertension and its resolution are also significantly

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan non‐homogenous across studies, as some reported as antihypertensive therapy use alone (18, 41); some diagnostic criteria were made regardless of pharmacologic management (282); many studies adopted a blood pressure of >140/90 mmHg or medication use (283); and others greater than 135/85 mmHg (284, 285). A number of studies did not specify definitions for remission at all (286‐289). As a result, improvement of hypertension following bariatric surgery remains incompletely understood, due to nonuniformity in its definition. In contrast to T2DM, there is a paucity of literature on the hypertension outcomes after bariatric surgery. We further delineate the effects of bariatric surgery on long‐term outcomes of hypertension in the complex patients with clinically severe obesity, using clear definitions of both antihypertensive therapy and blood pressure measurements for detailed changes in statuses (resolution, improvement, persistence and worsening).

Hyperlipidaemia or dyslipidaemia affects up to 80% of patients with obesity (290, 291), and is a key driver for cardiovascular disease (CVD) via the promotion of atherosclerosis and vascular thrombosis (292). Bariatric surgery is an established strategy to improve hyperlipidaemia. However, a recurring issue in bariatric surgery studies have been the definition of hyperlipidaemia/dyslipidaemia. Several studies have either assessed medications alone (41, 293, 294); lipid parameters but with no medication assessed (14, 295); lipid‐lowering agents and lipid profiles separately with no combined outcomes as a true whole representation of hyperlipidaemia status (294); definition not specified/patient self‐reported (296, 297); or separated the subfractions of lipid (e.g. hypercholesterolemia or hypertriglyceridemia) without evaluation of the whole lipid profile (298, 299). Some studies had short follow‐up periods post‐ bariatric surgery, e.g. 1 year (298), which limits the ability to study rare outcomes with a long latency period. Measuring the magnitude of resolution and improvement in hyperlipidaemia by monitoring both the changes of medications and lipid parameters [all the four measurements including total cholesterol, triglycerides, high‐density lipoprotein cholesterol (HDL‐C) and low‐ density lipoprotein cholesterol (LDL‐C)] following 6 years of bariatric surgery would more appropriately help in understanding the role of weight loss in management of hyperlipidaemia and cardiovascular risk in the publicly funded bariatric surgery care. Importantly, it could serve

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan as evidence‐based targets to inform clinical decisions to continue or trial cessation of lipid‐ lowering agents in those who manage lipid abnormalities well or who achieve optimal body weight loss and at most fitting post‐surgery year.

Another common obesity‐related comorbidity is osteoarthritis (OA). It is a progressive disease that leads to joint pain and significant disability. Obesity has consistently been identified as an accelerator of lower extremity OA by exerting the deleterious effects on joints through biochemical (weight‐loaded transmission on the joints) and adipose‐induced inflammation (179). Severe obesity also augments risk of chronic pain through mechanical factors, chemical mediators and depression (300). There is ample evidence in the published literature that weight loss is effective in reducing the symptoms of OA (179, 180, 301‐303). Yet, fundamental clinical questions still remain around the optimal management of clinically severe obesity and lower extremity OA. This was because many patients report that it is difficult for them to lose weight due to limitations in their physical activity level related to the joint pain, leaving the patients trapped in a vicious cycle. Looking in the mirror of obesity pandemic, orthopaedic surgeons are now seeing heavier and younger patients in need of total joint arthroplasty (TJA). With the combined presence of extremes of obesity and joint diseases, some patients are completely disabled and potentially house‐bound (179). Therefore, it will be interesting and helpful to address both synergistic disciplines across clinically severe obesity and OA/weight‐bearing joint pain (WBJP) that may be beneficial in long‐term. Accumulating evidence points towards the notion that bariatric surgery could improve gait biomechanics, and in the patients with both severe obesity and OA, it improves joint and pain functions (179). We hypothesised that bariatric surgery would not only a reliable method of reducing and sustaining weight loss, but also for the improvements in OA and/or WBJP in the patient cohort.

Obstructive sleep apnoea (OSA) and obesity hypoventilation syndrome (OHS) are the most common sleep disorders that involve cessation or a significant reduction in airflow in breathing effort, characterised by recurrent episodes of partial or complete upper airway obstruction during sleep due to the collapse of the pharyngeal muscles (304). Sleep‐disordered breathing is

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan typically associated with obesity, and it affects up to 80.5% of obese population who underwent bariatric surgery (305). Conversely, majority individuals with OSA (i.e. two‐third) have obesity (306), and even higher prevalence (85.7%) in clinically severe obese patients (307). Studies have proven that bariatric surgery is an effective treatment for OSA and OHS in short term (177, 306). Long‐term efficacy of bariatric surgery in sleep‐disordered breathing in the patients with clinically severe obesity is less robust, due to lack of publications, differing study designs, low numbers of well‐powered studies, and insufficient concordance regarding screening methods/objective measurements of outcomes. In drawing the evidence together, we also investigate the effects of bariatric surgery on sleep‐disordered breathing in an attempt to expand the understanding among the patients with both this disorder and clinically severe obesity in long‐term. We hypothesized bariatric surgery would significantly control the sleep‐disordered breathing of the study cohort.

Bariatric surgery has emerged as the most helpful treatment for obesity, however, the effects on mental health conditions are unclear. Recent studies have shown that the rates of mental health conditions are relatively common in patients with extreme obesity ‐ as high as 30‐40% are demonstrated in bariatric surgery candidates (196, 308). Considerable evidence shows post‐ operative improvement in depressive and anxiety symptoms, but conversely an increased suicide rate after bariatric surgery (194, 309), suggesting that mental health assessment is now more important than in the past. Long‐term data are again limited on mental illness after bariatric surgery, therefore, we explore the relationship between the two parameters in our population through 6 years.

The relevant articles reporting on the effectiveness or efficacy of bariatric surgery is summarised in Table 2.1.

Table 2.1 Studies reporting on bariatric surgery outcomes Authors, publication Study details Outcomes in bariatric surgical group year (Reference) Weight loss Sjöström et al., 2004 Prospective matched cohort study at 25 ‐23% (2 years), ‐17% (10 years), ‐16% (270) surgical departments and 480 primary (15 years), and ‐18% (20 years). Sjöström et al., 2014 health care centres in Sweden. Bariatric (192) surgical patients (n=2,010) with matched [Swedish Obese controls (n=2,037) recruited between Subjects (SOS) study] between 1987 and 2001. Multiple sub‐studies performed Adjustable or nonadjustable banding, vertical banded gastroplasty, or RYGB in the surgery group.

Courcoulas et al., 2015 Multicentre observational cohort study RYGB ‐41 Kg (31‐52 Kg) or 31.5 %TWL (275) at 10 US hospitals in 6 geographically (26.4‐38.4 %TWL); AGB ‐20 Kg (10‐29 Courcoulas et al., 2018 diverse clinical centres. Patients with Kg) or 15.9 %TWL (7.9‐23.0 %TWL). (14) severe obesity (median BMI 47 Kg/m2, [The Longitudinal range 34‐81 Kg/m2) undergoing bariatric Assessment of surgical procedures between 2006 and Bariatric Surgery 2009, followed up until January 31, (LABS) Study] 2015. Research assessments: Pre‐ Multiple sub‐studies surgery, 6‐month and annual up to 7 performed years.

Ece et al., 2018 (25) Retrospective analysis of 186 The mean %TWL and %EWL at 1, 2, 3, consecutive patients (mean BMI and 3.4 years post‐operatively were 52.6 Kg/m2) undergoing SG: MO vs SO vs similar in the groups, except the SSO. Mean follow‐up 3.4 years (2.8±4.5 %EWL for the SSO group was years). significantly lower (48.3%) at the end of the follow‐up period. 50.9% MO (n=83) vs 31.9% SO (n=52) vs 17.2% SSO (n=28).

Database source: Department of Surgery, Faculty of Medicine, Selcuk University, Turkey (Hospital unspecified)

Plamper et al., 2017 1‐Year retrospective analysis of 287 SO MGB‐OAGB significantly achieved (24) patients who operated between 2007 superior weight loss at 1 year in and 2015 at a single bariatric centre. comparison with SG. MGB‐OAGB (n=169, mean baseline

BMI=54.1 Kg/m2) vs SG (n=118, mean BMI: MGB‐OAGB (34.9±4.8 Kg/m2) vs baseline BMI=54.6 Kg/m2). SG (38.5±8.6 Kg/m2) (p=0.001)

%EWL: MGB‐OAGB (66.2±13.9%) vs SG (57.3±19.0%) (p<0.0001)

‘Diabesity’ Arterburn et al., 2013 Retrospective cohort study of 4,434 68.2% (95% CI=66‐70) experienced an (277) adults with uncontrolled or medication‐ initial complete T2DM remission controlled T2DM who underwent RYGB within 5 years after surgery. from 1995 to 2008 in three integrated health care delivery systems in the US. About one‐third [35.1% (95% CI=32‐ 38)] redeveloped T2DM within 5 Definition for main outcome: ADA years. definition ‐ Complete remission HbA1c level <6.0% and FBG <5.6 mmol/L for at least 1 year in the absence of active pharmacological therapy; Partial remission by HbA1c level <6.5% and FBG concentration 5.6–6.9 mmol/L, for at least 1 year in the absence of active pharmacological therapy.

Remission and relapse events by diabetes medication use and clinical laboratory measures of glycaemic control.

Courcoulas et al., Multicentre observational cohort study % remission at 1, 3, 5, and 7 years: 2015 (275) at 10 US hospitals in 6 geographically RYGB: 71.2% (95% CI=67.0‐75.4), Courcoulas et al., diverse clinical centres. Patients with 69.4% (95% CI=65.0‐73.8), 64.6% 2018 (14) severe obesity (median BMI 47 Kg/m2, (95% CI=60.0‐69.2), and 60.2% (95% [The Longitudinal range 34‐81 Kg/m2) undergoing bariatric CI=54.7‐65.6), respectively. Assessment of surgical procedures between 2006 and Bariatric Surgery 2009, followed up until January 31, LAGB: 30.7% (95% CI=22.8‐38.7), (LABS) Study] 2015. 29.3% (95% CI=21.6‐37.1), 29.2% Multiple sub‐studies Research assessments: Pre‐surgery, 6‐ (95% CI=21.0‐37.4), and 20.3% (95% performed month and annual up to 7 years. CI=9.7‐30.9), respectively.

Definition for main outcome: Diabetes diagnosis was defined as currently taking antidiabetic medication,

or HbA1c level ≥6.5%, or FBG ≥126 mg/dL. Remission = % of patients who did not have DM at follow‐up among those who had DM at baseline.

Jakobsen et al., 2018 Registry‐based cohort study of 1,888 At baseline, prevalence of use of (41) consecutive treatment seeking patients antidiabetics in the surgical with clinically severe obesity in a publicly group=25.5% (236/932). funded tertiary care outpatient centre in Norway, who underwent either bariatric DM remission [Absolute risk, % (95% surgery or specialized medical CI)]=57.5% (53.8‐61.2). treatment. Baseline data of exposures from 2005 to 2010 and follow‐up data New‐onset DM [Absolute risk, % (95% from 2006 until death/2015, with a CI)]=0.3% (0‐0.7). median 6.5 years of follow‐up (range 0.2‐10.1 years). Bariatric surgery (92% RYGB, 7% SG; n=932) vs medical treatment (individual/group‐based lifestyle intervention programs, n=956).

Definition for main outcome: Drug dispense alone (Norwegian Prescription Database).

Lager et al., 2018 Retrospective cohort study Follow‐up rates from 714 patients (279) (multidisciplinary University of Michigan initially were 657 (92%), 556 (78%), Adult Bariatric Surgery Program) of 507 (71%), and 498 (70%) at 1–4 years consecutive 714 clinically severe obese post‐operatively. patients (mean baseline BMI 48.4 Kg/m2) ≥18 years undergoing RYGB (n=380) or Among patients with DM, HbA1c LSG (n=334) between 2008 and 2013. improved more in RYGB than 1‐4 years post‐operative electronic laparoscopic SG: medical review and manual chart review HbA1c improvements at 1 year were prior to augmentation to automatic data 1.5±0.2% in RYGB and abstraction. 0.9±0.1% in laparoscopic SG (p=0.01), whereas improvements at 4 years Definition for main outcomes: were 1.3±0.2% (RYGB) vs 0.5±0.2% Remission of DM was defined as HbA1c (SG) (p=0.002). <6.5% without antidiabetic medications. A decrease in medications was defined There were greater rates of DM as a dose reduction and/or fewer remission, discontinuation of medications for the specified condition. antidiabetic medications, and decreased medications in patients

who underwent RYGB than laparoscopic SG at 3 and 4 years post‐ surgery.

Madsen et al., 2019 Population‐based data linkage At 1 year of follow‐up, 74% of the (278) observational cohort study of 1,111 bariatric surgical cohort experienced patients with T2DM and severe obesity T2DM remission, 27% had relapsed treated by RYGB in Northern Denmark after 5 years. (2006–2015), and 1,074 matched non‐ operated patients with T2DM. Data was linked from the Danish Civil Registration System (CRS), the Danish National Patients Registry (DNPR), the Danish National Health Service Prescription Database (DNHPD), and the LABKA database (with clinical laboratory information from both primary and secondary care).

Definition for main outcome: T2DM remission: Off glucose‐lowering drug with HbA1c <6.5%, or metformin monotherapy with HbA1c <6.0%.

Peterli et al., 2018 A 2‐group randomized trial, conducted 205 (94.5%) patients completed the (19) from 2007‐2011, with 5‐year follow‐up trial. [The Swiss period (last follow‐up in March 2017). Multicenter Bypass 217 Swiss patients with clinically severe At 5 years following surgery, or Sleeve Study obesity (mean BMI 43.9) were randomly proportions achieving T2DM (SM‐BOSS)] assigned to laparoscopic SG (n=107) vs remission (SG: 61.5% vs RYGB: 67.9%), laparoscopic RYGB (n=110) groups. improvement (SG: 15.4% vs RYGB: 7.1%), unchanged (SG: 11.5% vs Definition for main outcome: ADA RYGB: 10.7%) and worsening (SG: definition ‐ Complete remission HbA1c 11.5% vs RYGB: 14.3%). level <6.0% and FBG <5.6 mmol/L for at least 1 year in the absence of active Marked amelioration of glycaemic pharmacological therapy; Partial control after 5 years compared with remission by HbA1c level <6.5% and FBG baseline, with no significant concentration 5.6–6.9 mmol/L, for at differences between the treatment least 1 year in the absence of active groups in FBG (SG: 114.1 mg/dL vs pharmacological therapy. RYGB: 101.1 mg/dL; absolute difference 13.0 mg/dL; 95% CI=−7.5 to 33.5 mg/dL; p=0.21) or HbA1c (SG: 6.2% vs RYGB: 5.9%; absolute

difference 0.3%; 95% CI=−0.1% to 0.8%; p=0.09), uncorrected for multiple comparisons.

Salminen et al., 2018 Multicentre, open‐label, randomized At baseline, 42.1% had T2DM. (18) clinical equivalence trial of 240 clinically Complete or partial remission of severe obese patients aged 18‐60 years T2DM: 37% after SG (n=15/41) and from 2008 to 2010 in Finland, with a 5‐ 45% after RYGB (n=18/40) (p>0.99). year follow‐up period (last follow‐up, October 14, 2015); SG (n=121) vs RYGB (n=119).

Schauer et al., 2014 RCT of 134 overweight or obese patients After 5 years, glycaemic control of (274) (BMI 27‐43 Kg/m2) with T2DM. Intensive both bariatric surgical procedures was Schauer et al., 2017 medical therapy alone vs medical plus found to be superior to medical (151) RYGB vs medical plus SG. therapy alone, i.e. 14/49 (29%) and [Surgical Treatment 11/47 (23%) patients in RYGB and SG and Medications Definition for main outcome: arms, respectively, as compared to Potentially Eradicate T2DM remission ‐ HbA1c level <6.0%, or 5% of intensive medical therapy arm. Diabetes Efficiently off antidiabetic medication. (STAMPEDE) trial] Multiple sub‐studies performed

Hypertension Jakobsen et al., 2018 Registry‐based cohort study of 1,888 At baseline, prevalence of use of (41) consecutive treatment seeking patients antihypertensive therapy in the with clinically severe obesity in a publicly surgical group=47.8% (430/932). funded tertiary care outpatient centre in Norway, who underwent either bariatric Hypertension remission [Absolute surgery or specialized medical risk, % (95% CI)]=31.9% (28.4‐35.4). treatment. Baseline data of exposures from 2005 to 2010 and follow‐up data New‐onset hypertension [Absolute from 2006 until death/2015, with a risk, % (95% CI)]=3.5% (2.2‐4.9). median 6.5 years of follow‐up (range 0.2‐10.1 years). Bariatric surgery (92% RYGB, 7% SG; n=932) vs medical treatment (individual/group‐based lifestyle intervention programs, n=956).

Definition for main outcome:

Drug dispense alone (Norwegian Prescription Database).

Salminen et al., 2018 Multicentre, open‐label, randomized At baseline, 70.8% hypertension. (18) clinical equivalence trial of 240 clinically Remission for hypertension: 29% after [The Sleeve vs Bypass severe obese patients aged 18‐60 years SG (n=20/68) and 51% after RYGB (SLEEVEPASS) RCT] from 2008 to 2010 in Finland, with a 5‐ (n=37/73) (p=0.02). year follow‐up period (last follow‐up, October 14, 2015); SG (n=121) vs RYGB (n=119).

Definition for main outcome: Antihypertensive therapy use alone.

Casella et al., 2016 Retrospective analysis of prospectively 148 patients (81.4%) completed 6‐ (282) collected data from 182 patients (mean year follow‐up. 37 patients (25%) initial BMI was 45.9±7.3 Kg/m2) seeking reached a follow‐up of 7 years. SG between 2006 and 2008 in a university hospital in Italy. Pre‐operative hypertension was diagnosed in 45.2% of the bariatric Long‐term outcomes at 6 and 7 years surgical patients. were analysed. Remission and improvement of Definition for main outcome: hypertension occurred in 59.7% and Hypertension diagnosis: 38.8% of these patients, respectively. BP alone (SBP/DBP >140/90 mmHg) Note: Improvement and remission Improvement of hypertension: year (whether it was at post‐operative SBP or DBP values reduced or year 6 or 7) unspecified. Details of normalised with lower medication blood pressure readings and dosage. medications unreported.

Remission of hypertension: SBP or DBP values normalised without use of antihypertensive therapy.

Gadiot et al., 2017 Analysis of prospectively maintained At pre‐operative baseline, 33.6% had (286) database for 277 patients who hypertension (n=93). underwent SG between 2007 and 2010. Resolution of hypertension was Definition for main outcome: achieved in 53% of patients (n=37). Unspecified Reduction of antihypertensive therapy in 47% patients (n=33).

Kehagias et al., 2013 Retrospective analysis of prospectively At pre‐operative baseline, only 14 (285) collected data of 208 patients with patients had hypertension (6.7%). morbid obesity (up to BMI 50 Kg/m2) who underwent SG at University At post‐operative year 1, Hospital of Patras, Greece from January hypertension occurred in 8 of the 2005 to December 2010. 14 patients (4.4%) Resolution=42.9% Definition for main outcome: SBP/DBP ≥130/85 mmHg At post‐operative year 2, hypertension occurred in 7 of 13 patients (5.1%) Resolution=46.1%

At post‐operative year 3, hypertension occurred in 6 of 11 patients (6.6%) Resolution=45.5%

At post‐operative year 4, hypertension occurred in 4 of 7 patients (7.2%) Resolution=42.8%

At post‐operative year 5, hypertension occurred in 0 of 2 patients (0%) Resolution=100.0%

Note: Calculation of remission of comorbidity unexplained. Details of blood pressure readings and medications unreported. Only briefly reported 95.8% of the patients no longer required antihypertensive agents after bariatric surgery, without specifying post‐operative year and types of antihypertensive therapy.

Neagoe et al., 2017 Review of prospectively maintained At pre‐operative baseline, (288) database of the 101 consecutive patients hypertension was found in 60.4% of [mean BMI 44.5 Kg/m2 (range 31.5– patients (n=61). 1 year after surgery, 73.0)] who underwent laparoscopic SG in remission of hypertension was seen in 73.3% patients (n=33).

a Balkan emerging bariatric centre between 2010 and 2016. Note: Details of blood pressure readings and medications unreported. Definition for main outcome: Unspecified

Nocca et al., 2017 Retrospective review of a 5‐year At pre‐operative baseline, 309 (289) prospectively collected database of patients (33.1%) had hypertension. 1,050 severely obese patients (mean BMI=44.6 Kg/m2) who underwent The improvement and remission of laparoscopic SG from 2005 to 2013 at a hypertension was found in 76.9% and hospital in France. 19.2% patients, respectively.

Definition for main outcome: Note: Details of blood pressure Unspecified readings and medications unreported.

Hyperlipidaemia/ Dyslipidaemia Jakobsen et al., 2018 Registry‐based cohort study of 1,888 At baseline, prevalence of use of lipid‐ (41) consecutive treatment seeking patients modifying agents in the surgical with clinically severe obesity in a publicly group=21.2% (198/932). funded tertiary care outpatient centre in Norway, who underwent either bariatric Dyslipidaemia remission [Absolute surgery or specialized medical risk, % (95% CI)]=43.0% (39.3‐46.7). treatment. Baseline data of exposures from 2005 to 2010 and follow‐up data New‐onset diabetes [Absolute risk, % from 2006 until death/2015, with a (95% CI)]=1.1% (0.3‐1.9). median 6.5 years of follow‐up (range 0.2‐10.1 years). Bariatric surgery (92% RYGB, 7% SG; n=932) vs medical treatment (individual/group‐based lifestyle intervention programs, n=956).

Definition for main outcome: Drug dispense alone (Norwegian Prescription Database).

Pajecki et al., 2020 Retrospective, single‐centre 2‐year chart Prevalence of medication use of (293) review of 247 clinically severe obese dyslipidaemia dropped from 15.4% patients (mean BMI 46.7±6.7 Kg/m2) (pre‐operation, n=38) to 7.7% (post‐ who underwent RYGB in a university surgical year 1, n=19) and 6.9% (year hospital that performs publicly funded 2, n=17). bariatric surgery in Brazil.

Definition for main outcome: Medications alone

Patkar et al., 2017 Retrospective matched cohort study of Patients who underwent LAGB (294) obese patients who underwent (n=4,208), RYGB (n=4,308) or SG laparoscopic bariatric surgeries between (n=545) vs control cohorts (n=9,061). 2006 and 2013 using a US claim database. Surgical patients (LAGB, RYGB Compared with control subjects, and SG) were matched to medically patients who had bariatric surgery managed patients (controls). had significantly lower medication Comorbidity management was assessed usage for dyslipidaemia at 6 months every 6 months up to 5 years after the and continued for up to 5 years of surgery or an assigned index date for follow‐up. Subanalyses of changes in control subjects. selected laboratory test values over follow‐up corroborated the primary Definition for main outcome: analyses. Lipid‐lowering agents and lipid profiles separately

Courcoulas et al., Multicentre observational cohort study No significant trend in year 3 to 7 in 2015 (275) at 10 US hospitals in 6 geographically prevalence or remission of high LDL‐C Courcoulas et al., diverse clinical centres. Patient with levels. There was a significant 2018 (14) severe obesity (median BMI 47 Kg/m2, increase in incidence [2.4% (95% CI= [The Longitudinal range 34‐81 Kg/m2) undergoing bariatric 1.1‐3.8) vs 4.4% (95% CI=2.1‐6.7); Assessment of surgical procedures between 2006 and p=0.04]. There was not a significant Bariatric Surgery 2009, followed up until January 31, trend in prevalence, remission, or (LABS) Study] 2015. incidence of low HDL‐C or high Multiple sub‐studies Research assessments: Pre‐surgery, 6‐ triglyceride levels from years 3 to 7, performed month and annual up to 7 years. with 1 exception. The prevalence of high triglycerides followed a quadric Definition for main outcome: trend. However, values were similar Lipid parameters, no medication at post‐surgical year 3 [4.6% (95% assessed CI=3.3‐6.0); p=0.02] and year 7 [4.9% (95% CI=3.3‐6.6); p=0.02].

Spivak et al., 2017 Retrospective Israel national bariatric RYGB achieved the greatest mean (295) surgery registry of 4,526 patients reduction in plasma lipids (TC and with dyslipidaemia (at least 1 abnormal LDL‐C). HDL‐C levels were most plasma lipid level) and improved post‐SG. underwent bariatric surgery. 1‐Year follow‐up, enrolled from 2013 to 2014. 3742 underwent SG, 495 underwent RYGB, and 289 underwent LAGB.

Definition for main outcome: Lipid parameters, no medication assessed

Miras et al., 2018 The United Kingdom National Bariatric Unspecified remission rate ‐ Reported (297) Surgery Registry with 50,782 entries pictorially that remission of the study (mean BMI 48.0±8.0 Kg/m2), operations comorbidities significantly evident at between 2000 and 2015. Follow‐up: 1 to 1 year post‐operatively and reached a 5 years. plateau 2‐5 years after bariatric surgery. Definition for main outcome: Dependent on the judgement of the clinicians’ submitting data to the registry.

Osteoarthritis (OA) Hacken et al., 2019 5‐Year prospective observational study Significant improvements at 5‐year (180) in severely obese patients with follow‐up from baseline in all symptoms and radiographic evidence of subscales—pain, stiffness, and knee OA who were undergoing bariatric physical function (p<0.01). surgery (n=13).

Outcome measure: The Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) and Knee Osteoarthritis Outcome Score Surveys (KOOS) were administered at baseline, 6 months, 12 months, 2 years, and 5 years.

King et al., 2016 (302) Longitudinal observational study of A large percentage experienced [Longitudinal 2,458 patients with severe obesity who improvement compared with Assessment of underwent RYGB or LAGB between 2006 baseline, in pain, physical function, Bariatric Surgery‐2 and 2009, at 1 of 10 hospitals at 6 US and walk time over 3 years of follow‐ (LABS‐2) study] clinical centres. 3‐Year follow‐up through up. But the percentage with October 2012 was reported. improvement in pain and physical function deteriorated between year 1 Outcome measures: and year 3 after bariatric surgery. 36‐Item Short‐Form Health Survey (SF‐ 36). 400‐meter walk time, and the Western Ontario and McMaster

Universities Osteoarthritis Index (WOMAC).

OSA and OHS Peromaa‐Haavisto et 1‐Year prospective multicentre study of Prevalence of OSA decreased from al., 2017 (306) 132 patients who had OSA before 71% at baseline to 44% at 1 year after bariatric surgery in Finland. bariatric surgery (p<0.001).

45% OSA resolved, 78% resolved or improved, 8% de novo OSA. Moderate Definition for main outcome measures: or severe OSA persisted in 20% Changes in the prevalence of OSA and patients after the operation. apnoea–hypopnea index (AHI). Sleep symptom questionnaire was Total AHI decreased from 27.8 administered at baseline and at 1 year. events/hour to 9.9 events/hour (p<0.001).

Mental illness Hawkins et al., 2020 1‐year prospective cohort study of 190 32.1% patients were taking (308) patients (unknown BMI) who underwent psychiatric medications before bariatric surgery based on psychotropic surgery; of those, 82% (50/61) medication use. continued to take psychiatric medications 1‐year after surgery.

Kalarchian et al., 2019 Sub‐study of the LABS Consortium study Compared with pre‐surgery (34.7%), (195) at 3 US academic medical centres. prevalence of having any mental [The Longitudinal disorder was significantly lower 4 Assessment of 173 Participants (BMI range 36.1‐76.0 years (21.3%; p<0.01) and 5 years Bariatric Surgery Kg/m2) who completed the structured (19.2%; p=0.01), but not 7 years (LABS) Study] clinical interview for Diagnostic and (29.1%; p=0.27) after RYGB. Statistical Manual of Mental Disorders, 4th Edition prior to RYGB (n=104) or LAGB (n=69), as well as ≥1 follow‐up through 7 years post‐surgery.

Abbreviations: CI=Confidence intervals; %EWL=Percentage of excess weight loss; %TWL=Percentage total weight loss; BMI=Body mass index; Kg=Kilograms; m=metres; ADA=American Diabetes Association; BP=Blood pressure; SBP=Systolic blood pressure; DBP=Diastolic blood pressure; HbA1c=Glycated haemoglobin; OA=Osteoarthritis; OSA=Obstructive sleep apnoea; OHS=Obesity hypoventilation syndrome; MO=Morbid obesity; SO=Super obesity; SSO=Super‐super obesity; SG=Sleeve gastrectomy; LAGB=Laparoscopic adjustable gastric banding; RYGB=Roux‐en‐Y gastric bypass; MGB‐OAGB=Mini gastric bypass‐one anastomosis gastric bypass; RCT=Randomized Controlled Trial; DM=Diabetes mellitus; T2DM=Type 2 diabetes mellitus; vs=versus; FBG=Fasting blood glucose; TC=Total cholesterol; HDL‐C=High‐density lipoprotein cholesterol; LDL‐C=Low‐density lipoprotein cholesterol.

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

Notwithstanding being the most effective treatment for clinically severe obesity that produces long‐term weight loss and improvement in obesity‐related comorbidities, bariatric surgery carries potential for significant complications including nutritional deficiencies (146, 310‐313). Contrarily, individuals with obesity are commonly deficient in at least one micronutrient prior to seeking treatment and these deficiencies can worsen after bariatric surgery. The deficiencies are believed to arise from decreased gastric acid and intrinsic factor secretion, anatomic changes causing nutrient malabsorption, food intolerance, reduced food intake and poor food choices (130, 146, 310). The outcomes from a highly complex population over a long‐term follow‐up remain uncertain, we hope to help answer this key question with our study cohort.

The health outcomes after bariatric surgery in individuals with super obesity (SO) (BMI ≥50 Kg/m2) is a hidden component of the current obesity epidemic. The results in the literature are mixed, despite SO having been an issue of debate for more than a decade. A more complete set of key outcomes with longer follow‐up is less clear in this population with the highest‐degree of obesity, assumed greater technical difficulties, potentially higher risk and thus high priority (20). Some authors have observed that patients with SO are at an increased risk of surgical complications, a prolonged operating time, a longer length of hospitalization, increasing rates of 30‐day hospital re‐admission and rising healthcare‐related costs (314‐317). The common technical difficulties in the patients with SO include the ones related to the body size of the patient with SO by itself, which the surgical navigation appears more complex (316). Thicker layers of abdominal wall and intra‐abdominal fat, longer distance between the xiphoid and the oesophagus, and hepatomegaly resulting in reduced workspace are some of the surgical hindrances known to be associated with this challenging patient population (315, 318). Most of the time, specialized instrumentation such as longer instruments and lenses are needed, requiring greater force to manipulate them, consequently increasing operating time and surgeon fatigue (319). Notwithstanding, these claims are in part due to limited literature on the safety and long‐term outcomes of various bariatric procedures in the population with SO. As minimally‐invasive techniques and multidisciplinary care evolve, there is growing evidence that SO patients are benefiting from bariatric surgery with acceptable safety. There is recent evidence demonstrating

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan equivalent results in the SO as in other bariatric patients (25, 320, 321), suggesting that traditional perceptions of the peri‐operative risks, safety and resource challenges in only SO patient population are antiquated and understudied. There are limited studies directly compared the outcome measures between SO patients and those with morbid obesity (MO) (BMI <50 Kg/m2), which is in pressing need of further investigations. As such, the goal of this part of study is to expand on the current literature in this area by comparing the weight loss, length of hospital stay (LOHS), peri‐ and post‐surgical outcomes, and the rate of obesity‐related comorbidity changes (remission, improvement, persistence and worsening) between SO and MO groups, to formulate recommendations regarding the surgical care of the population with SO. We focused on the most common obesity‐related comorbidities − T2DM, hypertension, and hyperlipidaemia.

Collectively, in this multicentre cohort study with standardized data collection, we aimed to closely evaluate the effectiveness and durability weight loss of a range of up‐to‐date bariatric surgical procedures, long‐term effectiveness on the obesity‐related comorbidities and the safety outcomes of bariatric surgery in a well‐defined and carefully‐studied publicly funded cohort population attending a public hospital service. We also identify pre‐operative and 6 years of post‐ operative nutrient deficiencies in the patients undergoing bariatric surgery. The detailed 9‐year peri‐ and post‐operative surgical complications and mortality are also reported in this chapter.

2.3 MATERIALS AND METHODS

Study design This study involved a retrospective data collection and analysis of real‐world evidence of all eligible patients who underwent any type of publicly funded bariatric surgery in Sydney, Australia between June 2009 and December 2017 (n=168), and were followed up for at least 1 year to 9 years post‐operatively. The total sample size was underpowered but sufficient to highlight the study design effect. The patients are aimed to be followed up indefinitely in the event any

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan adverse symptoms can be monitored, investigated and managed in a timely manner. During the follow‐up, patients are encouraged of healthy lifestyle behaviours and any of their queries are addressed during the medical consultations.

Upon ethics approval from the institutional review board at each site (Protocol No. X15‐0339 & LNR/15/RPAH/463) and written informed consents obtained from all the patients, we carried out the present multicentre, multidisciplinary, multistage and multisurgeon bariatric surgery research comprised three public hospitals in the Greater Sydney, Australia providing publicly funded care (i.e. Royal Prince Alfred, Concord Repatriation General and Camden Hospitals). This publicly funded bariatric surgery service represents Australian state of New South Wales (NSW)’s first, longest established, and the largest publicly funded bariatric surgery research program to improve the health and wellbeing of patients with clinically severe obesity (defined as class III obesity alone or BMI ≥35 Kg/m2 with at least one significant obesity‐related comorbid condition). All the patients were operated at our own local bariatric surgical facility, the Concord Repatriation General Hospital that also operated Australia’s heaviest man with a maximum weight of 468 Kg prior to seeking for treatments.

Research assessments were conducted by thorough data extraction from the electronic and paper‐based medical records, interviews, questionnaires, telephone and mail. All data were assessed and inputted in a standardized database by the same study‐qualified research investigator (MT), and verifications on the inconsistencies and ambiguous cases were sought from the experienced study consultant physicians (TM, NK and SH). The research protocol and databases were created by the study investigator (MT) in discussion with the aforementioned study consultant physicians (TM, SH and NK), a professor of obesity research (AS) and senior biostatisticians to ensure quality data collection of clinically important variables across all timepoints to generate meaningful analyses, in which the details of the thesis overview and study outline are as described in CHAPTER 1. Data collection for the analyses of this study was accomplished in compliance with the Declaration of Helsinki.

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

The blood test results for biochemistry and nutritional parameters available were extracted from both the electronic and paper‐based hospital records, whichever that were available. In the absence of laboratory results in both the hospital records, phone call attempts were made to request for a copy of the patient’s blood test findings from the centralized pathology laboratories to be emailed and faxed to the clinics. The laboratory results were also accessed and obtained from the online NSW Health Pathology portal, if any readings were available in the online portal but not other sources. Generally, the biochemistry and nutritional parameters for pre‐operative baseline and post‐operative years 1, 2, 3, 4, 5 and 6 were obtained, which included glycaemic control, lipid profile, iron studies, vitamin D, and vitamin B12 that were detailed in each subsection in this chapter.

Patient inclusion and exclusion Briefly, the inclusion criteria for the publicly funded bariatric surgery were: (1) Aged between 18–70, (2) Initial BMI ≥40 Kg/m2 with at least one obesity‐related comorbidity which will improve with pre‐operative weight loss regimens (e.g. T2DM, OSA/OHS, hypertension, hyperlipidaemia, OA and/or WBJP and non‐alcoholic fatty liver disease), and (3) Pregnancy not anticipated in first two years post‐surgery.

Exclusion criteria were: (1) Irreversible endocrine or other disorders that can cause obesity, (2) Current drug or alcohol abuse, (3) Uncontrolled, severe psychiatric illness, (4) Lack of comprehension of risks, benefits, expected outcomes, alternatives, and lifestyle changes required with bariatric surgery, and (5) Inability to attend post‐surgical follow‐up appointments.

All the patients underwent a formal comprehensive multidisciplinary assessment as fit for bariatric surgery by an endocrinologist, bariatric surgeon, nurse manager, exercise physiologist/physiotherapist, dietitian and a co‐located psychologist. A psychiatric evaluation was obtained if considered necessary.

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

Surgical procedures All the treating bariatric surgeons who operated the patients were part of the study team (PL, DM, CT and DJ). The advantages and disadvantages of the bariatric procedures in the treatment of clinically severe obesity were extensively discussed with all patients by the treating surgeons. Subsequently, a general recommendation was made based on the severity of obesity, comorbidities, surgical risks and patient’s preference. Patients made the final decisions proceeding to bariatric procedures performed. The standardized surgical techniques for the four bariatric procedures adopted are as described as follows:

Sleeve gastrectomy (SG) The greater curvature was resected using a 60‐mm linear stapler, along a 32‐French (Fr) calibration bougie that was inserted to the pylorus along the lesser curve. The volume of the remaining stomach was approximately 100 cm3.

Mini gastric bypass‐one anastomosis gastric bypass (MGB‐OAGB) Pneumoperitoneum was established using optical insertion of 12‐mm visiport. Two further 15‐ mm ports and one 5‐mm port were used as working ports. A minimum 10 cm length and 36‐Fr gastric pouch was created using TriStapler® 45‐ and 60‐mm purple and black cartridges. A loop of the small bowel 150‐200 cm from the DJ flexure was then brought up to the gastric pouch in an ante‐colic, ante‐gastric fashion. Next, either a linear stapled or hand‐sewn anastomosis was performed according to our operating surgeon’s preference. A leak test was done by using a dilute methylene blue solution.

Adjustable gastric banding (AGB) Optical entry was obtained followed by 15‐mmHg pneumoperitoneum. A total of four ports were used (15 mm and 3 x 5 mm). Next, the patient was placed in reverse Trendelenburg position. The angle of His was first dissected to expose the left crus. Subsequently, the gastrohepatic ligament was incised to expose the right crus, allowing dissection of a retrogastric tunnel on the right crus using a blunt‐tipped grasper. This grasper was passed from the base of the right crus to the apex

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan of the left crus. Then, the band was introduced into the peritoneal cavity and passed around the stomach of approximately 1‐2 cm below the cardia. The buckle was secured, creating a small gastric pouch above the band. To minimize the risk of subsequent band slippage, a series of gastropexy sutures including gastrophrenic (Birmingham), gastrogastric, and anterior gastropexy sutures were used as per our normal practice using 2.0 Novafil. Band tubing then was exteriorized via a 5‐mm port and shortened appropriately. The port was connected and secured to the abdominal wall.

Roux‐en‐Y gastric bypass (RYGB) The standardized surgical technique for the RYGB entailed creating a small stomach pouch of approximately 30 cm3 using a 60‐mm stapler. The alimentary limb was set to 100‐150 cm, the biliopancreatic limb was measured to 50‐80 cm, and a side‐to‐side jejunojejunostomy was created using a 60‐mm linear stapler. The entry hole was closed by hand‐sewn sutures or a stapler. Hand‐sewn end‐to‐side 36‐Fr anastomosis (with antecolic and antegastric positioning) was performed for gastrojejunostomy.

Subsequent revisional bariatric procedures were identified by medical record reviews or surgical reports.

Pre‐operative body mass index (BMI) classification

Body mass index (BMI) was categorised according to the World Health Organization (WHO) classification (7), with all the study patients having class III obesity (BMI of ≥40 Kg/m2) upon initial consultation. Subsequent to the integrated lifestyle weight management program (WMP) at the clinics and immediately before undergoing bariatric surgery, patients were re‐assessed and classified based on their pre‐operative BMI, with one‐fifth of the patients improved to class II obesity (BMI <40 Kg/m2) (Figure 2.1).

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

In the context of the highest grade of obesity, all the patients were further stratified into the following groups for subanalysis for comparison of outcome measures: (i) Morbid obesity (MO) (BMI <50 Kg/m2) (ii) Super obesity (SO) (BMI ≥50 Kg/m2)

Figure 2.1 Distribution of the study patients with clinically severe obesity in each BMI‐ obesity category at pre‐operative baseline (n=168)

Class II Obesity Class III Obesity (BMI ≥40 Kg/m2) (n=133, 79.2%) 50 (n=35) 45.2% 45

% 30

25 20.8% 22.0%

Patient, 20

15 11.9% 10

0 BMIBMI <40< 40 Kg/m2Kg/m2 BMI BMI 4040 ‐ ‐49.949 Kg/m Kg/m22 BMI BMI ≥ 5050‐59 Kg/m2 Kg/m2 BMI BMI ≥ ≥60 Kg/m2Kg/m2

Morbid Obesity (MO) Super Obesity (SO) n=111 (66.1%) n=57 (33.9%)

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

The Edmonton Obesity Staging System (EOSS) stages

While all our patients in the complexity‐oriented publicly funded specialist obesity clinics also had one or more comorbidity related to obesity that a BMI classification alone does not distinguish patients with higher‐disease severity or overall health status, we recognised the importance of understanding their disease severity to prioritize those who would benefit the most from early access to bariatric surgery. Therefore, we further adopted a disease‐related and functional staging scoring system that has been proposed as a simple clinical staging tool in assessing the indication for bariatric surgery (322) ‐ the Edmonton Obesity Staging System (EOSS) (323). The EOSS was applied to classify the study patients into five stages, namely EOSS stages 0, 1, 2, 3 and 4 based on the impact of obesity, taking into account the presence, symptoms and degree of their obesity‐related illness severity, as well as psychopathologic manifestations. Patients were assigned individually to the highest‐stage they have presented in any of the EOSS domains displayed in Table 2.2 and Figure 2.2.

Stage 0 of EOSS indicates the obese phenotype with no comorbidity, risk factor or symptom. Stage 1 represents obesity‐related subclinical disease (e.g. borderline hypertension or impaired fasting glucose) whereby medical management is favoured. Whereas patients in stages 2 and 3 of EOSS are candidates for medical and surgical treatments for obesity, with established or significant obesity‐related comorbidities. Stage 4 represents severe to end‐stage disease associated with obesity to the extent that surgical intervention is less likely to improve long‐term prognosis and may be harmful (324).

100

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

Table 2.2 Edmonton Obesity Staging System (EOSS) Staging tool (325) Stage Conceptual definition Study operational definitions 0 No apparent obesity‐related risk factors, None reported physical symptoms, psychopathology, functional limitations, and/or impairments of well‐being 1 Presence of obesity‐related subclinical risk Any of the following: factors, mild physical symptoms, mild 1. FBG: 5.6–6.9 mmol/L psychopathology, mild functional limitations, 2. Total cholesterol: 5.2–6.1 mmol/L and/or impairment of well‐being 3. Triglycerides: 1.7–2.3 mmol/L 4. HDL‐C: 1.0–1.6 mmol/L 5. LDL‐C: 3.4–4.0 mmol/L 6. GFR: 60.0–89.9 mL/min/1.73 m2 2 Presence of established obesity‐related Any of the following: comorbidities, moderate limitations in 1. FBG: ≥7.0 mmol/L activities of daily living, and/or well‐being 2. Diagnosed T2DM or on glucose‐ lowering medication 3. Total cholesterol: ≥6.2 mmol/L 4. Diagnosed hypercholesterolemia 5. Triglycerides: >2.3 mmol/L 6. HDL‐C: <1.0 mmol/L 7. LDL‐C: ≥4.1 mmol/L 8. Diagnosed hyperlipidaemia or on hyperlipidaemia medication 9. Diagnosed arterial hypertension or on hypertension medication 10. Osteoarthropathy 11. Fatty liver 12. GFR: 30.0–59.9 mL/min/1.73 m2 3 Established end‐organ damage, significant Any of the following: psychopathology, significant functional 1. Coronary artery disease limitations, and/or impairment of well‐being 2. Myocardial infarction 3. Congestive heart failure 4. Ischaemic stroke 5. GFR: <30.0 mL/min/1.73 m2 4 Severe (potentially end‐stage) disabilities Not applicable from obesity‐related chronic diseases, disabling psychopathology, functional limitations, and/or impairment of well‐being Abbreviations: EOSS=Edmonton Obesity Staging System; HDL‐C=High‐density lipoprotein cholesterol; LDL‐C=Low‐ density lipoprotein cholesterol; FBG=Fasting blood glucose; GFR=Glomerular filtration rate; T2DM=Type 2 diabetes mellitus. (from Ogassavara NC, Dias JGM, Pajecki D, de Oliveira Siqueira J, Santo MA, Tess BH. The Edmonton Obesity Staging System: Assessing a potential tool to improve the management of obesity surgery in the Brazilian public health services. Surgery for Obesity and Related Diseases. 2020;16(1):40‐7) 101

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

Figure 2.2 Edmonton Obesity Staging System (EOSS) Staging tool (323)

(Reprinted from Sharma AM, Kushner RF. A proposed clinical staging system for obesity. International Journal of Obesity. 2009;33(3):289‐95)

102

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

Assessments and outcome measures

Assessment of weight change

Weight change, the primary endpoint of the study was evaluated based on weight loss in kilogram (Kg), BMI loss or percent of total weight loss (%TWL), from pre‐operative baseline, in which baseline weight was measured closest to the time of bariatric surgery. The weight loss indices were calculated according to the following standard equations (326):

Change in BMI (ΔBMI) (Kg/m2) = Pre‐operative baseline BMI – Post‐operative BMI (Equation 2.1)

Weight loss (Kg) = Post‐operative weight – Pre‐operative baseline weight (Equation 2.2)

Post‐operative weight – Pre‐operative baseline weight %TWL = × 100% (Equation 2.3) Pre‐operative baseline weight

Definition and post‐operative course of obesity‐related comorbidities

The predefined secondary endpoints ‒ the changes in status of comorbidities that are strongly associated with obesity were assessed at each follow‐up visit over 6 years based on symptoms, laboratory findings, physical measures and medication use. In addition to prevalence rates, the post‐operative course of comorbidities was defined as: in remission (asymptomatic and medications no longer needed), improved (reduction in number of active medications and/or fewer symptoms), persisting (same symptoms and equivalent medications as before bariatric surgery) or worsened (increase in therapy or a change from non‐insulin treatment to insulin use in the case of T2DM) for T2DM, hypertension and hyperlipidaemia. The incidence of a comorbidity was defined as patients without the comorbidity at pre‐operative baseline who newly developed the comorbidity after bariatric surgery. As this study was not intended to measure medication adherence but reflective of comorbidity complexity, medication dosage was not factored into the calculation.

103

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

Type 2 diabetes mellitus (T2DM) status Pre‐operative T2DM status was determined based on physician’s diagnosis, prescription of any diabetes medication [insulin, oral hypoglycaemic agents (OHAs) and/or injectable diabetes medication], or a plasma glycosylated haemoglobin (HbA1c) measure of ≥6.5%. If HbA1c level was not available, fasting blood glucose (FBG) of ≥7.0 mmol/L was used. T2DM remission status after bariatric surgery was modified from the American Diabetes Association (ADA) criteria (273).

Remission of T2DM was determined as a HbA1c <6.5% and a FBG of <7.0 mmol/L in the absence of antidiabetic medications. Improvement in T2DM was defined as improved parameters in terms of reduced number of medications or a change from insulin use to non‐insulin treatment (i.e. OHAs). Persisting T2DM status was defined by unchanged antidiabetic medications calculated from pre‐operative baseline. Worsening T2DM was assessed as newly prescribed antidiabetic medications after the surgery, increased number of OHAs, and/or changing from OHAs to insulin use. Patients reporting as having polycystic ovarian syndrome (PCOS) who did not meet the laboratory criteria for T2DM and were not on an anti‐diabetic medication other than metformin were not considered to have T2DM. Incident T2DM was any new onset T2DM not present at pre‐operative baseline that developed throughout the 6 years post‐operatively.

Hypertension status Hypertension status was determined based on an abnormal increase in blood pressure with a systolic/diastolic blood pressure of greater than or equal to 140/90 mmHg, or treatment with antihypertensive drugs. Hypertension remission post‐surgery was defined as normotensive (<140/90 mmHg) without any antihypertensive therapy, according to the American Society for Metabolic and Bariatric Surgery criteria (326). Improvement in hypertension was defined as fewer antihypertensive therapies across timepoints, whereas persisting status indicated by the same number of medications as before bariatric surgery. Worsened hypertension indicated an increased number of therapies in reference to pre‐surgical status.

104

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

Hyperlipidaemia status Hyperlipidaemia status was determined based on a total cholesterol level of ≥5.2 mmol/L (200 mg/dL), low‐density lipoprotein cholesterol (LDL‐C) ≥3.5 mmol/L (130 mg/dL), high‐density lipoprotein cholesterol (HDL‐C) ≤1.0 mmol/L (40 mg/dL), triglycerides ≥1.7 mmol/L (150 mg/dL); and/or treatment with lipid‐lowering agents (i.e. statin, fibrate and ezetimibe). At the annual follow‐ups, hyperlipidaemia remission required a complete return to normal lipid panel (of all the four serum lipid subfractions) with cessation of all lipid‐lowering drugs (326). Improvement in hyperlipidaemia was defined as reduction of number of prescribed lipid‐lowering agents. Similar to hypertension, persisting hyperlipidaemia is defined as unchanged number of lipid‐ lowering medications, and worsened status was indicated by a higher number of lipid‐lowering agents.

Sleep‐disordered breathing status The status of pre‐operative sleep‐disordered breathing, which consists of obstructive sleep apnoea (OSA) and obesity hypoventilation syndrome (OHS), were determined based on a Respiratory and Sleep Physician’s diagnosis according to previous diagnostic overnight inpatient polysomnography, and review of continuous positive airway pressure (CPAP) or bilevel positive airway pressure (BiPAP) device requirements. In view that the objective measurement ‐ polysomnography could not be performed for all patients, remission was not defined for sleep‐ disordered breathing. Instead, an improvement was considered based on a discontinuation of CPAP or BiPAP use, in addition to physicians’ diagnoses of decreased symptoms or normalized sleep pattern. This status adheres to both the objective and subjective improvement guidelines of the American Society for Metabolic and Bariatric Surgery (326). Besides that, adherence to CPAP or BiPAP device use among the study patients was retrieved to accurately eliminate self‐ discontinuation of the devices.

Osteoarthritis (OA) and/or weight‐bearing joint pain (WBJP) status The osteoarthritis (OA) and/or weight‐bearing joint pain (WBJP) status were evaluated based on physicians’ diagnoses of OA, and patients’ complaints of the WBJP symptoms (including back pain,

105

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

knee pain, hip pain and other joint pain) and were on medications (opioids and/or non‐opioid analgesic anti‐inflammatory and pain‐relief agents). In particular, opioid use was defined as prescription of regular or daily opioid analgesics, typically codeine, buprenorphine, fentanyl, oxycodone, tapentadol and tramadol.

Additionally, via data linkage, the full data of total joint arthroplasty (TJA) of our patients were retrieved from the nationwide Australian Orthopaedic Association National Joint Replacement Registry (AOANJRR), assisted by an orthopaedic surgeon (AL) in our team. The data linkage was performed according to the AOANJRR guidelines and legal requirements. Specifically, upon ethics approval, the matching process between our database and the AOANJRR was initiated by providing the patient identifiers such as the medical record number (MRN) to the registry. The registry team then matched the study patients with the AOANJRR databases, and exported the data requested onto a Microsoft Excel spreadsheet with encrypted coding before emailing the re‐identifiable dataset to the study investigator (MT). These included diagnoses for confirmation (whether it was OA, rheumatoid arthritis, loosening, fracture or pain), procedure types, specific joint undergoing the respective orthopaedic surgery, type of revisional surgery (if applicable), and procedure dates.

Depression and/or severe anxiety status Depression and severe anxiety were classified based on psychiatrists’ diagnoses, patient’s history/symptoms and/or psychologists’ assessments, and use of pharmacotherapy (antidepressants, antianxiety agents, or antipsychotic agents and mood stabilisers) or ongoing non‐pharmacological treatment for mental health illness such as cognitive behavioural treatment (CBT). As depression and severe anxiety are often interrelated with each other, we merged the two conditions as one variable, namely depression and/or severe anxiety in the analysis. In this study, the specific psychopharmacotherapy for depressions included amitriptyline, citalopram, desvenlafaxine, duloxetine, escitalopram, fluoxetine, fluvoxamine, mirtazapine, moclobemide, paroxetine, sertraline and venlafaxine; and the antianxiety agents included alprazolam, diazepam, lorazepam and oxazepam.

106

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

Hyperuricaemia Hyperuricaemia was defined as a serum urate level of 0.40 mmol/L and above, given that this cut‐off is the saturation point for uric acid in physiological conditions (327).

Nutritional deficiencies Patients were instructed by dietitians to consume multivitamin and mineral supplements daily for lifelong after bariatric surgery. The following micronutrients were monitored on an annual basis following surgery: serum iron studies (ferritin and transferrin saturation), 25‐hydroxy vitamin D (25‐OH vitamin D), and Vitamin B12 (cobalamin). These blood assays were collected before and up to 6 years after bariatric surgery.

The reference ranges of pathological cut‐offs for nutrient deficiencies adopted in this study were established by the NSW’s Health Pathology (Sydney, Australia). Nutritional deficiency or abnormal levels were defined as serum concentrations below the normal reference ranges, in which:

a) Iron deficiency anaemia Iron studies were considered low at levels of serum ferritin (<20 μg/L) and transferrin saturation (<15%). Patients who fulfilled both low concentrations of ferritin and transferrin saturation were diagnosed as having iron deficiency anaemia. b) Vitamin D deficiency Serum vitamin D level was considered deficient at <50 nmol/L. c) Vitamin B12 insufficiency

A serum vitamin B12 level <150 pmol/L indicated vitamin B12 insufficiency.

Surgical complications Peri‐ and post‐operative complications were documented thoroughly and classified as early/late and minor/major according to the established standardized reporting recommendations for bariatric surgery (326).

107

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

Generally, early complications include any occurrence of surgical complication within 30 days of bariatric surgical procedures, whereas late complications constitute any occurrence >30 days post‐operation. Major complications include any death or a complication that result in a prolonged hospital stay exceeding 7 days, administration of an anticoagulant (e.g. warfarin), reintervention, reoperation or a need for blood transfusion of four or more units. Minor complications consist of all other post‐operative adverse events that is not included under major.

Mortality and cause of death Deaths and the causes were adjudicated according to all information available including hospital records (e.g. palliative care medical notes), post‐mortem reports and death certificates. These were verified with the team physicians, death review team, Medico Legal Subpoena, Department of Forensic Medicine of NSW state, and State Coroner.

STATISTICAL ANALYSES

Descriptive statistics summarize baseline characteristics for overall sample, each bariatric surgical procedure and by follow‐up year. An overview of 9‐year descriptive weight change from initial clinic visit to pre‐surgery through 8 years following bariatric surgery are demonstrated. Due to small number of patients at years 7 and 8 post‐surgery; the primary and secondary outcomes up to 6 years were analysed and reported, and results beyond year 6 were eliminated, except the mortality and the national TJA data from the AOANJRR. Frequencies and percentages (%) are reported for categorical variables. Means and standard deviation (SD), and median (IQR) are reported for continuous data. Normality was assessed visually and using the Shapiro‐Wilks test. The bootstrap 95% confidence intervals (CI) for the comorbidity statuses and prevalence of nutrient deficiencies in this study were computed by the bias‐corrected and accelerated (BCa) method.

The marginal estimated means of the weight change (Kg), BMI (Kg/m2) and % total weight loss (%TWL) modelled from linear mixed‐effects models (also known as linear mixed models or

108

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

mixed‐effects models) for indicator time after bariatric surgery are reported (1) as a whole, as well as (2) separately for the two BMI groups (MO versus SO) at each follow‐up year. We required that the smallest trajectory group include at least 5% of our total sample size at any follow‐up timepoint, therefore, the modellings were plotted up to 6 years post‐surgery, and the later years including years 7 and 8 with fewer observations were omitted from statistical modellings. On the same justification of small sample that leading to inadequate statistical power to detect the level of difference, comparison between each bariatric procedures for weight change was not included in the analysis, given that RYGB and some of the follow‐up years for MGB‐OAGB did not fulfil the a minimum of 5% of total sample for statistical modelling. The longitudinal analyses constructed from the linear mixed‐effects models were based on restricted maximum likelihood (REML) method, with a person‐level random intercept (i.e. subject ID) included, and models were adjusted for age at surgery, sex and race. The modelling included fixed effect terms for age at time of surgery, sex, race and annual clinic visits. A repeated measures term was included for the visits within each subject to take into account multiple observations over time, expressed as continuous data. The modelled weights have accounted for weight that are missing completely at random (MCAR), thus the analyses should not bias the results. The best fitting models were determined using the Akaike’s Information Criterion (AIC). The modelled trajectories of overall or each group are plotted with bars indicating the 95% CI of the modelled weight (Kg), BMI (Kg/m2) and %TWL.

Likewise, weight loss between the SO and MO groups over 6 years of time were modelled and compared using linear mixed‐effects model. A generalised mixed‐effects regression accounted for random effects at the individual level was applied to estimate probability of weight loss in the two BMI groups, defined by achievement of ≥20% TWL as per the annual Australia and New Zealand Bariatric Surgery Registry report guideline. Confidence intervals (CI) were set at 95%, and a two‐sided p value of <0.05 was considered statistically significant. The proportions of changes of comorbidity status (remission, improvement, persisting and worsening) of T2DM, hypertension and hyperlipidaemia were compared using Pearson’s chi‐square (χ2) test or Fisher’s exact test as appropriate.

109

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

Data analyses were implemented using IBM SPSS Statistics (Version 26, Armonk, New York, USA), STATA (Version 15, Stata Corp LP, College Station, TX, USA) and GraphPad Prism (Version 8.4.3, GraphPad Software LLC, San Diego, CA, USA). All reported p values are 2‐sided. A p value less than 0.05 are considered statistically significant.

2.4 RESULTS

Study Cohort Overall, a total of 168 patients suffering from clinically severe obesity were offered bariatric surgery between 2009‐2017, with SG (83.9%, n=141) as the most commonly performed primary (index) bariatric procedure, followed by MGB‐OAGB (n=15), AGB (n=11), and RYGB (n=1). All the bariatric surgical operations were performed laparoscopically with no conversion to open. The average length of hospital stay (LOHS) was 3 days (ranged 1‐8). Mean follow‐up time after bariatric surgery was four years.

At initial medical review, the mean (±SD) weight and BMI of the patients were 141.9±35.9 Kg and 51.2±10.8 Kg/m2, respectively. Mean age at surgery was 52.0±11.4 years old (ranged 21–72), with a mean pre‐operative baseline weight 133.2 Kg (ranged 80.3–239.9) and pre‐operative BMI was 48.0 Kg/m2 (ranged 33.6‐78.8), respectively. Most patients were of obesity class III, with 33.9% being considered super obese (BMI ≥50 Kg/m2). At baseline, two‐thirds of patients were female; majority were Caucasian; over half were current or ex‐smoker; 40.6% were either separated, divorced or widowed; 43.5% were unemployed on Australian government support payments; and 7.7% were heavy alcohol consumers (i.e. ≥4 standard drinks most days of the week).

Table 2.3 details the baseline characteristics of the study patients by their primary (index) procedures. There was no significant difference in the characteristics between the surgery groups.

110

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

Table 2.3 Baseline socio‐demographics of study cohort by primary procedures Variables Overall SG MGB‐OAGB AGB RYGB (n=168) (n=141) (n=15) (n=11) (n=1)

Age, mean±SD (years) 52.0±11.4 52.0±11.7 49.5±8.2 57.1±9.7 33.0 (Range) (21–72) (21–72) (30–59) (39–68)

Sex, n (%) Female 111 (66.1) 92 (65.2) 10 (66.7) 8 (72.7) 1 (100.0) Male 57 (33.9) 49 (34.8) 5 (33.3) 3 (27.3) 0 (0)

Race, n (%) Caucasian 120 (71.4) 98 (69.5) 10 (66.7) 11 (100) 1 (100.0) Middle Eastern 24 (14.3) 21 (14.9) 3 (20.0) 0 (0) 0 (0) Other§ 24 (14.3) 22 (15.6) 2 (13.3) 0 (0) 0 (0)

Relationship status, n (%) Separated/Divorced/Widowed 67 (40.6) 51 (36.7) 9 (64.3) 6 (54.5) 1 (100.0) Married/Domestic partner 82 (49.7) 74 (53.2) 4 (28.6) 4 (36.4) 0 (0) Single 16 (9.7) 14 (10.1) 1 (6.3) 1 (9.1) 0 (0)

Smoking status, n (%) Current/Ex‐smoker 96 (57.1) 82 (58.2) 8 (53.3) 5 (45.5) 1 (100.0) Never smoker 72 (42.9) 59 (41.8) 7 (46.7) 6 (54.5) 0 (0)

Heavy alcohol consumption¥, n (%) 13 (7.7) 11 (84.6) 2 (15.4) 0 (0) 0 (0) (Current/Ex)

Employment status, n (%) Employed 57 (33.9) 46 (32.6) 4 (26.7) 7 (63.6) 0 (0) Unemployed# 17 (10.1) 15 (10.6) 2 (13.3) 0 (0) 0 (0)

On government support payment, 73 (43.5) 61 (36.3) 8 (4.8) 4 (2.4) 0 (0) n (%)ψ Disability support pension 44 (26.2) 35 (24.8) 7 (46.7) 2 (18.2) 0 (0) Age pension 15 (8.9) 13 (9.2) 0 (0) 2 (18.2) 0 (0) Carers pension 6 (3.6) 6 (4.3) 0 (0) 0 (0) 0 (0) NewStart Allowance 5 (3.0) 5 (3.5) 0 (0) 0 (0) 0 (0) 7 (4.2) 6 (4.3) 1 (6.7) 0 (0) 0 (0) Others† #Unemployment including those who were not working or retired, and not on any government support payment †Veteran pension, Naonal Disability Insurance Scheme, Workers compensaon, Department of Housing, or Unemployment pension §Indigenous Australian, Pacific Islander, Americas, Black African, Mauritian, Filipino and Pakistanis ¥Heavy drinker is defined as ≥4 standard drinks most days of the week ψFour patients were on two government support payments each

111

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

Primary Outcomes:

Weight changes Weight outcomes were charted up to 8 years post‐operation for the actual observation in accordance with follow‐up data available, and 6 years for modelled weight loss accounted for pre‐operative baseline factors and nature of the data.

Observed overview body mass index (BMI) and weight changes The observed weight change and BMI change by follow‐up timepoints from initial clinic visit (i.e. 12 months preceding bariatric surgery on lifestyle modifications) to pre‐operative baseline through 8 years of follow‐up after bariatric surgery are depicted in Figures 2.3 and 2.4. The weight loss with diet and exercise interventions 12 months before bariatric surgery was a mean of 8.7 Kg (corresponding to 3.2 Kg/m2), which then decreased to a mean weight at the time of surgery of 133.2 Kg, corresponding to a mean BMI of 48.0 Kg/m2.

112

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

Figure 2.3 Observed mean weight change from initial clinic visit (lifestyle modifications) to pre‐operative baseline through 8‐year post‐operative follow‐up 160.0

140.0

120.0

100.0 (Kg)

80.0 weight

60.0 Observed 40.0

20.0

0.0 Initial visit Pre‐op m3 m6 m12 m18 m24 m36 m48 m60 m72 m84 m96 Follow‐up, months

No. of patients 168 168 165 164 157 136 129 98 79 63 42 25 11 Weight (Kg) Mean (SD) 141.9 133.2 113.0 105.3 99.8 97.4 99.1 101.7 103.2 105.9 111.7 101.5 110.6 (35.9) (32.4) (27.7) (26.0) (24.6) (24.5) (24.0) (24.5) (25.5) (27.1) (28.2) (23.2) (30.3) Minimum 83.5 80.3 67.3 63.0 58.2 58.6 58.0 57.5 60.0 59.1 63.4 70.5 75.8 Maximum 272.8 239.9 205.0 184.2 174.1 171.9 175.00 176.80 182.2 192.3 189.90 163.7 163.7 Abbreviations: BMI=Body mass index; m=month; SD=Standard deviations

113

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

Figure 2.4 Observed mean and median BMI change from initial clinic visit (lifestyle modifications) to pre‐operation to Year 8 of bariatric surgery

100

)

2 60

(Kg/m

BMI 50

40 Observed

0 Initial visit 0 3 6 12 18 24 36 48 60 72 84 96 Follow‐up, months

No. of patients 168 168 165 164 157 136 129 98 79 63 42 25 11

BMI (Kg/m2)

Median 48.2 45.6 39.2 36.2 34.3 33.7 34.4 34.7 34.8 36.0 38.2 37.2 37.2 Minimum 34.0 33.6 28.1 25.2 22.2 22.2 22.0 23.5 24.4 23.9 24.2 26.5 31.1 Maximum 83.8 78.8 68.3 62.3 60.1 57.8 55.5 56.4 57.4 59.5 59.1 63.2 63.2 Mean (SD) 51.2 48.0 40.8 37.9 36.1 35.2 35.7 36.3 36.8 37.4 39.1 38.0 41.5 (10.8) (9.5) (8.4) (7.9) (7.6) (7.7) (7.6) (7.5) (7.9) (8.0) (7.9) (7.9) (10.2) Abbreviations: BMI=Body mass index; Kg=Kilograms; m=meters BMI is calculated as weight in kilograms divided by height in meters squared

Boxplots represent the observed median body mass index (BMI) change from initial visit (i.e. 12 months preceding surgery) (on lifestyle modifications) to pre‐operative baseline to 8 years after bariatric surgery. Lower and upper ends of error bars indicate minimum and maximum values (i.e. ranges), respectively; lower and upper borders of boxes represent interquartile range (IQR) (25th and 75th percentiles, respectively); horizontal lines in boxes indicate median values; plus signs

114

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan indicate mean values. BMI at time 0 represents pre‐operative BMI between initial visit and day of surgery.

As can be seen in Figures 2.3 and 2.4, there were significant weight changes from pre‐operative baseline to each post‐surgical visit over the observation period. The greatest actual weight loss occurred within the first 18 to 24 post‐operative months (97.4 Kg to 99.1 Kg), regained slightly and remained constant through. Weight recidivism from the weight loss nadir was seen between years 2 and 9 post‐surgery. In comparison to the year 6 post‐surgery, the weight loss was generally more pronounced at 7 and 8 years of follow‐up after bariatric surgery. Howbeit further modellings are limited by small number of patients in these last two years of observations.

Modelled weight and body mass index (BMI) change As the key focus of this study is the effectiveness of bariatric surgery, the linear mixed‐effects modellings were performed on weight loss from the pre‐surgical baseline taken into account repeated measure and controlled for baseline confounding factors. Figures 2.5(a) and 2.5(b) represent the core results of our study, i.e. the overall estimated marginal mean weight change of the study patients by post‐surgical follow‐up years. This consists of measures of the mean weight loss, %TWL and BMI loss from baseline to 6 years after operation.

The linear mixed‐effects model revealed that substantial and steady weight loss was achieved over time (p<0.001 versus pre‐operative baseline); with a significant %TWL (±SE) of 24.1±1.1% at 1 year, 24.1±1.1% at 2 years, 22.1±1.1% at 3 years, 20.4±1.2% at 4 years, 19.0±1.3% at 5 years, and 18.8±1.5% at 6 years. This was corresponding to a change in BMI (±SE) of 11.6±0.7 Kg/m2 at year 1, 11.7±0.6 Kg/m2 at year 2, 10.7±0.6 Kg/m2 at year 3, 10.0±0.7 Kg/m2 at year 4, 9.3±0.7 Kg/m2 at year 5, and 9.1±0.9 Kg/m2 at year 6 post‐operatively.

As outlined in the respective panel (a) and panel (b) that complement each other’s data, at the start of the surgical treatment, patients experienced most of their total weight loss in the first 2 years after bariatric surgery and stabilization occurred through the second year of follow‐up. From 2‐year level, modelled weight change began to demonstrate some modest weight regain

115

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan afterward and remained relatively stable again at years 5 to 6. Analysis of these data also shows that the maximal amount of %TWL was achieved between year 1 and year 2 (literally at year 1.5 post‐surgery as discussed earlier), which was rarely been observed in other previous studies which usually only observed and reported annual interval measurement.

Figure 2.5 Overall marginal estimated mean (a) weight (Kg) and (b) BMI loss (Kg/m2) at each visit over time modelled from the mixed‐effects model taking into account the repeated measures, adjusted for age at time of surgery, sex and race

(a) Modelled weight loss (Kg) year 1 to 6 post‐bariatric surgery

(Kg)

change

weight

mean * * * * * * marginal

Estimated

Years since bariatric surgery

Likewise, there was a significant and dramatic BMI loss by 1 year after bariatric surgery, followed by a slightly steady and stabilized weight regain towards their last follow‐up observations. Taken altogether, the study cohort significantly achieved and maintained a successful marginal estimated mean weight loss over each timepoints to their last follow‐up (p<0.05).

116

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

(b) Modelled BMI loss (Kg/m2) year 1 to 6 post‐bariatric surgery

) 2 (Kg/m

change

BMI

* * mean

* * * * marginal

Estimated

Years since bariatric surgery

No. of patients 168 157 129 98 79 63 42 Weight loss (Kg) 0 33.5 33.6 30.7 28.6 26.4 25.7 (95%CI) (29.9–37.0) (30.1–37.1) (27.2–34.2) (24.9–32.3) (22.5–30.3) (20.8–30.7) %TWL 0 24.1 24.1 22.1 20.4 19.0 18.8 (95%CI) (21.9–26.3) (21.9–26.3) (19.8–24.3) (18.0–22.8) (16.5–21.6) (15.7–21.8) BMI loss (Kg/m2) 0 11.6 11.7 10.7 10.0 9.3 9.1 (95%CI) (10.3–12.9) (10.4–13.0) (9.5–12.0) (8.6–11.3) (7.9–10.7) (7.3–10.8) * p<0.001 versus baseline value Abbreviations: Kg=Kilograms; %TWL=Percentage total weight loss; BMI=Body mass index; CI=Confidence intervals

The overall marginal estimated mean weight loss at each visit over time modelled from the mixed‐ effects model with random effects, taking into account the repeated measures nature of the data. Lines indicate modelled weight change from baseline based on mixed models, adjusted for baseline factors (age at time of surgery, sex and race). A negative value represents weight loss of pre‐surgery weight. Data markers (estimated marginal mean values) indicate weight change data taken into account missing data completely at random (MCAR). Error bars represent the 95% CI.

117

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

Super obesity (SO) (BMI ≥50 Kg/m2) versus morbidly obesity (MO) (BMI <50 Kg/m2)

A total of 33.9% (n=57) patients were identified as having super obesity (SO) (BMI ≥50 Kg/m2) and 66.1% (n=111) with morbid obesity (MO) (BMI <50 Kg/m2). The subanalyses comparing the modelled weight outcomes between the SO versus MO groups are shown in Figures 2.6 and 2.7, irrespective of bariatric surgical procedures undergone.

Comparisons of estimated marginal mean weight change (Kg) were modelled using linear mixed‐ effects model that adjusted for pre‐operative baseline factors, taking into account repeated measures. Overall, the modelled weights from pre‐operative baseline within each BMI category are shown to be significantly different at each visit as illustrated in Figure 2.6. Following a similar trend, both the SO and MO groups experienced maximum modelled mean weight loss within the first 2 years following bariatric surgery [33.6 Kg (95% CI=30.1–37.1) and 24.1% TWL (95% CI=21.9– 26.3), p<0.001]. After reaching the weight nadir, between years 2 and 6, there was a slight weight regain that followed a quadratic trend (p<0.001) with a weight change ranging from ‐25.7 Kg to ‐ 33.6 Kg, corresponding to a %TWL of baseline weight of ‐18.8% to ‐24.1%, such that the weight regain decreased during this time period.

118

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

Figure 2.6 Comparisons of modelled weight change (Kg) between the SO and MO groups at pre‐operative baseline and at years 1, 2, 3, 4, 5 and 6 following bariatric surgery

(Kg)

weight

mean

marginal

Estimated

Years since bariatric surgery

No. of patients 168 157 129 98 79 63 42 Morbid obesity 111 105 88 68 55 43 30 Super obesity 57 52 41 30 24 20 12

Lines indicate modelled weight change from baseline based on mixed‐effects model for

repeated measures, adjusted for baseline factors (age at surgery, sex and race). Data markers (estimated marginal mean values) and bars (SE) indicate weight change data taken into account missing data completely at random (MCAR).

Both BMI groups significantly reduced weight with perfect parallel graphic trajectories across the follow‐up period. Unquestionably, compared with MO, the weights were significantly higher for patients with SO over the observation period (p<0.001) (Figure 2.6).

However, these kilogram loss are deceitful, because when further comparison between the groups was implemented using the %TWL, there were no statistically significant %TWL difference between the SO and MO groups over 6 years of time as shown in Figure 2.7.

119

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

Figure 2.7 Comparison of estimated marginal mean differences of post‐surgical percent total weight loss (%TWL) of the pre‐surgery weight between SO and MO groups by clinic follow‐up visits adjusted for age at surgery, sex and race

%TWL

mean

marginal

Estimated

Years since bariatric surgery

No. of patients 168 157 129 98 79 63 42 Morbid obesity 111 105 88 68 55 43 30 Super obesity 57 52 41 30 24 20 12

In other words, regardless of baseline BMI whether is smaller or greater than 50 Kg/m2, there were similar percent weight losses between the two groups over 6 years of time post‐surgery, respectively, at 1 year (estimated marginal mean difference=‐1.3%, 95% CI=‐2.4% to ‐4.9%), at 2 years (estimated marginal mean difference=‐1.7%, 95% CI=‐1.9% to ‐5.3%), at 3 years (estimated marginal mean difference=‐1.3%, 95% CI=‐2.3% to ‐4.9%), at 4 years (estimated marginal mean difference=‐1.0%, 95% CI=‐3.0% to ‐5.0%), at 5 years (estimated marginal mean difference=‐0.9%, 95% CI=‐3.6% to ‐5.5%), and at 6 years (estimated marginal mean difference=‐0.8%, 95% CI=‐5.1% to ‐6.8%).

To further delineate the variability or equality (or equivalence) of substantial weight change in the two groups across follow‐up, the mean probability of %TWL ≥20% from baseline and 95% CI by visit in MO and SO groups adjusted for sex, age at surgery, race and surgery type were

120

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan examined. The generalised mixed‐effects regression accounted for random effects at the individual level revealed no significant difference between the groups in terms of probability of successful ≥20 %TWL.

Secondary Outcomes:

Obesity‐related comorbidities

Edmonton Obesity Staging System (EOSS) stages In the present study, the most common Edmonton Obesity Staging System (EOSS) stage among the patients were stage 2 (61.3% of patients), followed by stages 3 (27.4%) and 4 (11.3%) (Figure 2.8). We observed no patients in stages 0 and 1 of EOSS. These results clearly demonstrated the high complexity of their diseases that required referral to our publicly funded bariatric surgery service that are well‐equipped to handle severe clinical complications.

Figure 2.8 Distributions of Edmonton Obesity Staging System (EOSS) stages among the publicly funded bariatric surgery service patients (n=168)

Stage 0 = 0%

StageStage 4 4 Stage 1 = 0% 11%

11.3%

27.4% StageStage 3 3 61.3% 27.4%28% StageStage 2 2 27.4%61%

121

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

Prevalence of type 2 diabetes mellitus (T2DM), hypertension, hyperlipidaemia, osteoarthritis (OA)/weight‐bearing joint pain (WBJP), obstructive sleep apnoea (OSA)/obesity hypoventilation syndrome (OHS), and depression and/or severe anxiety

Figure 2.9 below shows the prevalence of obesity‐related comorbidities by timepoints. At pre‐ operative baseline, there were high rates of all the obesity‐related comorbidities of interest. These include OA and/or WBJP, T2DM, hypertension, hyperlipidaemia, OSA/OHS, and depression and/or severe anxiety.

122

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

Figure 2.9 Prevalence of obesity‐related comorbidities pre‐operatively and at 1, 2, 3, 4, 5 and 6 years following bariatric surgery (%) 80 All patients

50 Patient,

20 Baseline Year 1Year 2Year 3Year 4Year 5Year 6

Number of years since bariatric surgery, Year

OA and/or WBJP Hypertension Hyperlipidaemia T2DM

OSA/OHS Depression and/or severe anxiety

Year Baseline 1 2 3 4 5 6

OA and/or WBJP, 120/168 100/158 76/122 56/90 44/76 32/56 23/41 n (%) (71.4%) (63.3%) (62.3%) (62.2%) (57.9%) (57.1%) (56.1%)

T2DM, n (%) 114/168 53/163 44/131 32/98 29/84 24/58 23/44 (67.9%) (32.5%) (33.6%) (32.7%) (34.5%) (41.4%) (52.3%)

Hypertension, n (%) 110/168 63/163 55/134 38/94 40/81 32/58 23/43 (65.5%) (38.7%) (41.0%) (40.4%) (49.4%) (55.2%) (53.5%)

OSA/OHS, n (%) 107/168 90/161 73/125 54/95 46/77 32/59 20/38 (63.7%) (55.9%) (58.4%) (56.8%) (59.7%) (54.2%) (52.6%)

Hyperlipidaemia, n 99/168 102/163 80/130 63/99 52/83 39/58 34/43 (%) (58.9%) (62.6%) (61.5%) (63.6%) (62.7%) (67.2%) (79.1%)

Depression and/or 79/168 55/167 43/132 35/100 31/81 31/62 20/41 severe anxiety, n (%) (47.0%) (32.9%) (32.6%) (35.0%) (38.3%) (50.0%) (48.8%) Abbreviations: T2DM=Type 2 diabetes mellitus; OSA=Obstructive sleep apnoea; OHS=Obesity hypoventilation syndrome; OA=Osteoarthritis; WBJP=Weight‐bearing joint pain

123

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

The prevalence of all the obesity‐related comorbidities followed a quadratic trend, which declined dramatically from baseline to the first year post‐surgery, followed by a slight increase in the comorbidity prevalence during years 2 through 6 (Figure 2.9), with the exception of depression and/or severe anxiety and hyperlipidaemia, which exceed baseline levels as time progressed.

Prediabetes was seen in 24 patients (14.3%) at pre‐operative baseline (not shown in Figure 2.9). Before surgery, over half of the study patients (53.0%, n=89) experienced some degree of gastro‐ oesophageal reflux disease (GORD). We did not report GORD under this section but detailed it under the subsection for peri‐ and post‐surgical complications for those who had worsening of reflux symptoms, or patients who had no GORD at baseline but reported de novo severe reflux symptoms on medications or required additional surgical interventions after bariatric surgery, which we think may be more meaningful.

The clinical profiles presented in Table 2.4 below for the entire study population regardless of baseline metabolic diseases, demonstrated that the post‐operative blood pressure measurements, glycaemic control and triglycerides levels were lower than the baseline for the whole study population. Whereas only total cholesterol, HDL‐C and LDL‐C levels were higher than that of pre‐operation (Table 2.4).

124

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

Table 2.4 Clinical profiles of all the study patients pre‐ and post‐bariatric surgery (n=168) Follow‐up, year Baseline 1 2 3 4 5 6 Clinical measurement Systolic BP Mean (±SD) (mmHg) 131.9±15.0 121.8±14.6 124.8±14.6 124.2±15.5 126.6±17.1 128.7±13.7 126.5±17.4 Diastolic BP Mean (±SD) (mmHg) 77.0±8.9 72.3±8.4 72.0±9.0 73.6±8.6 73.0±8.5 73.0±8.8 74.8±9.9 Laboratory parameters HbA1c (%) Mean (±SD) 6.8±1.5 5.9±1.2 6.0±1.4 6.2±1.3 6.4±1.3 6.5±1.2 6.7±1.4 Range (4.8‐11.1) (4.3‐11.4) (4.2‐10.7) (4.5‐10.0) (4.7‐9.9) (4.7‐9.6) (4.8‐9.7) FBG (mmol/L) Mean (±SD) 6.9±2.6 5.7±1.9 6.1±2.1 6.0±2.2 6.2±1.9 6.4±2.1 7.0±2.5 Range (3.4‐18.4) (3.1‐14.8) (2.8‐14.7) (3.1‐17.3) (4.0‐11.5) (2.9‐14.6) (3.8‐15.9) Total cholesterol, mmol/L 4.5±1.0 4.7±0.9 4.6±1.0 4.7±1.0 4.7±0.9 4.8±0.9 4.8±1.2 Triglycerides, mmol/L 1.7±0.9 1.3±0.8 1.3±0.9 1.3±0.6 1.3±0.7 1.4±0.8 1.6±0.7 HDL‐C, mmol/L 1.2±0.4 1.4±0.4 1.6±0.5 1.5±0.4 1.5±0.5 1.5±0.5 1.3±0.3 LDL‐C, mmol/L 2.5±1.0 2.6±0.8 2.5±0.9 2.6±0.9 2.6±0.9 2.5±0.8 2.7±1.0 Use of medications Antidiabetics N=168 N=159 N=122 N=88 N=75 N=57 N=39 Yes, n (%) 104 (61.9) 47 (29.6%) 40 (32.8%) 29 (33.0%) 22 (29.3%) 21 (36.8%) 19 (48.7%) Insulin treatment, n (%) 39 (23.2%) 16 (10.1%) 15 (12.3%) 11 (12.5%) 6 (8.0%) 5 (8.8%) 7 (18.0%) Glucose‐lowering agents, n (%) 0 66 (39.3%) 119 (74.8%) 88 (72.1%) 64 (72.7%) 56 (74.7%) 38 (66.7%) 21 (53.8%) 1 56 (33.3%) 27 (17.0%) 22 (18.0%) 15 (17.0%) 12 (16.0%) 11 (19.3%) 10 (25.6%) 2 41 (24.4%) 12 (7.5%) 10 (8.2%) 7 (8.0%) 5 (6.7%) 5 (8.8%) 6 (15.4%) 3 4 (2.4%) 1 (0.6%) 2 (1.6%) 2 (2.3%) 2 (2.7%) 3 (5.3%) 2 (5.1%) 4 1 (0.6%) 0 (0.0%) 0 (0.0%) 0 (0.0%) 0 (0.0%) 0 (0.0%) 0 (0.0%) Mean number of drug (±SD) 0.9±0.9 0.3±0.6 0.4±0.7 0.4±0.7 0.4±0.7 0.5±0.9 0.7±0.9 Range 0–4 0–3 0–3 0–3 0–3 0–3 0–3 Antihypertensive therapy N=168 N=153 N =118 N=86 N=73 N=55 N=39 Yes, n (%) 103 (61.3) 57 (37.3%) 44 (37.3%) 33 (38.4%) 32 (43.8%) 28 (50.9%) 20 (51.3%) Mean number of drug (±SD) 1.2±1.3 0.6±1.0 0.7±1.1 0.7±1.1 0.8±1.1 0.9±1.1 1.0±1.1 Range 0–5 0–4 0–5 0–5 0–4 0–4 0–4 Lipid‐lowering drugs, n (%) N=168 N=159 N=123 N=91 N=75 N=56 N=39 Yes, n (%) 80 (47.6%) 55 (34.6%) 49 (39.8%) 35 (38.5%) 31 (41.3%) 26 (46.6%) 19 (48.7%) Mean number of drug (±SD) 0.6±0.7 0.4±0.6 0.5±0.6 0.4±0.6 0.5±0.6 0.6±0.7 0.6±0.7 Range 0–3 0–2 0–2 0–2 0–2 0–2 0–2

Abbreviations: BP=Blood pressure; SD=Standard deviation; HbA1c=Glycated haemoglobin; FBG=Fasting blood glucose; HDL‐ C=High‐density lipoprotein cholesterol; LDL‐C=Low‐density lipoprotein cholesterol; N=Total number of patients with data of the variables available at that follow‐up year.

125

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

Changes in comorbidity statuses

The detailed changes of statuses in comorbidities (T2DM, hypertension and hyperlipidaemia) were also generated based on clinical measures, biochemical tests, medication use and physicians’ examinations. The illustrations, proportions and 95% CI of the post‐operative statuses alongside with the detailed blood tests, medication profiles and clinical measurements are reported accordingly in Figures 2.10–2.12, respectively.

Changes in type 2 diabetes mellitus (T2DM) status

Figure 2.10 displays the annual changes of comorbidity status after bariatric surgery at each follow‐up period.

Figure 2.10 Yearly remission, improved, persisting and worsened rates of T2DM following bariatric surgery

40 %

30 Patient, 20

0 0123456 Follow‐up after bariatric surgery, Year

Remission Improved Persisting Worsened

Follow‐up, year 0 1 2 3 4 5 6 Remission, n (%) 56 (52.3%) 42 (50.0%) 29 (48.3%) 25 (50.0%) 20 (46.5%) 11 (35.5%) (95 CI%) (43.9–59.8) (39.3–59.5) (36.7–58.3) (36.3–62.0) (34.9–58.1) (22.6–48.4) Improved, n (%) 31 (29.0%) 24 (28.6%) 17 (28.3%) 15 (30.0%) 14 (32.6%) 9 (29.0%) (95 CI%) (21.5–38.2) (21.4–36.9) (20.0–38.3) (20.0–40.0) (20.9–44.2) (16.1–41.9)

126

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

Persisting, n (%) 18 (16.8%) 16 (19.0%) 10 (16.7%) 8 (16.0%) 7 (16.3%) 6 (19.4%) (95 CI%) (11.2–22.4) (11.9–26.2) (8.3–25.0) (8.0–24.0) (7.0–25.6) (9.7–29.0) Worsened, n (%) 2 (1.9%) 2 (2.4%) 4 (6.7%) 2 (4.0%) 2 (4.7%) 5 (16.1%) (95 CI%) (0.0–4.7) (0.0–6.0) (3.3–11.7) (0.0–10.0) (0.0–11.6) (6.5–25.8) Total patientsa 114 (67.9%) 107 (100%) 84 (100%) 60 (100%) 50 (100%) 43 (100%) 31 (100%)

HbA1c (%) Mean (±SD) 7.2±1.5 6.2±1.2 6.4±1.4 6.4±1.3 6.6±1.3 6.7±1.2 6.9±1.4 Range (4.90–11.1) (4.3–11.4) (4.2–10.7) (4.5–10.0) (4.7–9.9) (4.9–9.6) (4.8–9.7) FBG (mmol/L) Mean (±SD) 7.6±2.8 6.1±2.1 6.7±2.3 6.6±2.5 6.7±2.0 6.8±2.3 7.3±2.6 Range (4.1–18.4) (3.1–14.8) (4.0–14.7) (3.1–17.3) (4.0–11.5) (2.9–14.6) (3.8–15.9) Antidiabetics Yes, n (%) 104 (91.2%) 47 (43.9%) 40 (47.6%) 29 (48.3%) 22 (44.9%) 20 (47.6%) 18 (58.1%) Insulin treatment, n (%) 39 (34.2%) 16 (15.0%) 15 (17.9%) 11 (18.3%) 6 (12.2%) 5 (4.4%) 7 (6.1%) Glucose‐lowering agents†, n (%) 114 107 84 60 49 42 31 0 12 (10.5%) 67 (62.6%) 50 (59.5%) 36 (60.0%) 30 (61.2%) 24 (57.1%) 14 (45.2%) 1 56 (49.1%) 27 (25.2%) 22 (26.2%) 15 (25.0%) 12 (24.5%) 10 (23.8%) 9 (29.0%) 2 41 (36.0%) 12 (11.2%) 10 (11.9%) 7 (11.7%) 5 (10.2%) 5 (11.9%) 6 (19.4%) 3 4 (3.5%) 1 (0.9%) 2 (2.4%) 2 (3.3%) 2 (4.1%) 3 (7.1%) 2 (6.5%) 4 1 (0.9%) 0 (0.0%) 0 (0.0%) 0 (0.0%) 0 (0.0%) 0 (0.0%) 0 (0.0%) Mean number of drug (±SD) 1.4±0.8 0.5±0.7 0.6±0.8 0.6±0.8 0.6±0.8 0.7±0.9 0.9±1.0 Range 0–4 0–3 0–3 0–3 0–3 0–3 0–3 aPatients with T2DM present at pre‐operative baseline Bootstrap 95% confidence interval (CI) was computed by the bias‐corrected and accelerated (BCa) method †Biguanides, sulphonylureas, dipeptidyl peptidase 4 (DPP4) inhibitors, glucagon‐like peptide‐1 (GLP‐1) receptor agonists, sodium‐glucose transport protein 2 (SGLT2) inhibitors, α‐glucosidase inhibitor and thiazolidinedione (TZD). Abbreviations: SD=Standard deviation; HbA1c=Glycated haemoglobin; FBG=Fasting blood glucose

At baseline, 114 (67.9%) of 168 patients were diagnosed with T2DM; 39 of the patients with T2DM (34.2%) were receiving insulin treatment while 102 (89.5%) were on glucose‐lowering

agents with a mean elevated HbA1c of 7.2%. In all, the resultant post‐surgical weight loss following bariatric surgery substantially and rapidly resolved and improved T2DM, with 35.5%‐52.3% of patients experienced remission and 29.0%‐32.6% improvement of their T2DM in between the 6 years follow‐up period (Figure 2.10). The remission and improvement rates decrease over time but stabilized until year 5 and remained high in long‐term at year 6, which explain the trend of increasing prevalence of T2DM and hypertension at the later years after surgery, as shown in Figure 2.9.

Besides that, the cumulative remission and improvement rates were also calculated. It was found that 48.2% achieved a cumulative remission rate of T2DM following bariatric surgery to their last

127

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan observations (55 of the 114 patients with T2DM at baseline). While 28.9% (i.e. 33 patients of the 114 patients with T2DM at baseline) reached a cumulative improvement rate of T2DM following bariatric surgery to their last observations. As a whole, the composite cumulative remission and improvement rates of T2DM was as high as 77.1% (n=88), equivalent to more than three‐quarters of those who reported T2DM at baseline. In other words, although there appeared some declination in T2DM remission over time, most of the patients with T2DM undergoing bariatric surgery are still in remission and improvement to their last observations, while the worsening and incidence rates were extremely low. Worthy of note, there was only one negligible new onset (incident) of T2DM requiring antidiabetic medications during the long‐term follow‐up period.

As for the glycaemic control and medication breakdowns, HbA1c level fell from an average of 7.2% to below 7.0% at all timepoints after bariatric surgery. Mean FBG responded well between years 1 and 5 post‐surgery but did not reduce substantially at year 6, compared to the pre‐operative baseline. Antidiabetic medication use (insulin and/or oral glucose‐lowering agents) significantly reduced post‐surgery, with only a‐fifth of patients required insulin at year‐6 follow‐up, and a significant fall of mean number of oral drugs from 1.4 to 0.9 was observed.

Changes in hypertension status

The results of changes in hypertension statuses are presented in detail in Figure 2.11. Two patients who have had a diagnosis of heart failure and one patient with atrial fibrillation who were treated with beta‐blockers for their cardiac problems rather than for hypertension were excluded from analysis in the present study (n=3).

128

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

Figure 2.11 Yearly remission, improved, unchanged and worsened rates of hypertension following bariatric surgery

30 % 25

20 Patient, 15

0 0123456 Follow‐up after bariatric surgery, Year

Remission Improved Persisting Worsened

Follow‐up, year 0 1 2 3 4 5 6 Remission, n (%) 40 (39.6%) 26 (32.1%) 22 (37.3%) 17 (31.5%) 13 (30.2%) 11 (34.4%) (95 CI%) (31.7–47.5) (23.5–40.7) (25.4–49.2) (22.2–40.7) (18.6–41.9) (21.9–46.9) Improved, n (%) 28 (27.7%) 28 (34.6%) 15 (25.4%) 18 (33.3%) 13 (30.2%) 11 (34.4%) (95 CI%) (20.8–35.6) (24.7–44.4) (16.9–35.6) (24.1–44.4) (18.6–41.9) (21.9–46.9) Unchanged, n (%) 29 (28.7%) 22 (27.2%) 18 (30.5%) 16 (29.6%) 13 (30.2%) 7 (21.9%) (95 CI%) (21.8–35.6) (18.5–35.8) (20.3–39.0) (18.5–40.7) (20.9–41.9) (10.2–34.4) Worsened, n (%) 4 (4.0%) 5 (6.2%) 4 (6.8%) 3 (5.6%) 4 (9.3%) 3 (9.4%) (95 CI%) (1.0–7.9) (2.5–9.9) (3.4–11.9) (0.0–13.0) (4.7–16.3) (0.0–18.8) Total patientsa 110 (65.5%) 101 (100%) 81 (100%) 59 (100%) 54 (100%) 43 (100%) 32 (100%)

Systolic BP (mmHg) Mean (±SD) 133.7±14.6 124.9±15.0 128.7±14.8 127.8±15.4 129.2±17.5 130.2±13.0 128.5±17.4 Diastolic BP (mmHg) Mean (±SD) 77.0±9.2 73.4±8.3 72.9±8.9 73.3±9.0 74.4±8.4 73.0±8.7 75.1±9.5 Antihypertensive therapy† Yes, n (%) 103 (93.6%) 57 (56.4%) 44 (53.7%) 33 (55.0%) 32 (59.3%) 28 (65.1%) 20 (62.5%) 0 7 (6.4%) 42 (42.4%) 37 (45.7%) 26 (44.1%) 22 (40.7%) 15 (34.9%) 12 (37.5%) 1 39 (35.5%) 33 (33.3%) 23 (28.4%) 14 (23.7%) 15 (27.8%) 14 (32.6%) 7 (21.9%) 2 40 (36.4%) 15 (15.2%) 14 (17.3%) 13 (22.0%) 11 (20.4%) 9 (20.9%) 10 (31.3%) 3 13 (11.8%) 6 (6.1%) 2 (2.5%) 2 (3.4%) 2 (3.7%) 3 (7.0%) 2 (6.3%) 4 9 (8.2%) 3 (3.0%) 4 (4.9%) 3 (5.1%) 4 (7.4%) 2 (4.7%) 1 (3.1%) 5 2 (1.8%) 0 (0.0%) 1 (1.2%) 1 (1.7%) 0 (0.0%) 0 (0.0%) 0 (0.0%) Mean number of drug (±SD) 1.9±1.1 0.9±1.0 1.0±1.2 1.1±1.2 1.1±1.2 1.1±1.1 1.2±1.1 Range 0–5 0–4 0–5 0–5 0–4 0–4 0–4

aPatients with hypertension present at pre‐operative baseline

129

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

Bootstrap 95% confidence interval (CI) was computed by the bias‐corrected and accelerated (BCa) method †Beta‐blockers, angiotensin‐converting enzyme (ACE) inhibitors, angiotensin receptor antagonists, calcium channel blockers and thiazide diuretics. Abbreviations: BP=Blood pressure; SD=Standard deviation

Among two‐thirds of the patients who underwent bariatric surgery and had hypertension at pre‐ surgical baseline, complete remission of hypertension was achieved in one‐thirds of them over 6 years with no indication of hypertension at all following bariatric surgery, and improved in another similar proportions of patients across follow‐up (Figure 2.11). There appeared to be a similar trend to T2DM in hypertension remission and improvement rates, despite small and variable fluctuations after 1 year of surgery, yet most patients maintained much of their remission and improvement in hypertension across the long‐term follow‐up period. Compared to the pre‐operative baseline, antihypertensive medications were discontinued in significant proportions of the patients (Figure 2.11), with a mean pre‐operative number of antihypertensive drugs of 1.9 reduced to between 0.9‐1.2 during the 6 years post‐operative follow‐up.

Of interest, the cumulative remission rate of hypertension following bariatric surgery to their last observations was 40 patients (i.e. 36.4% of the 110 patients with hypertension at baseline). Additionally, another 37 patients achieved a cumulative improvement of hypertension after bariatric surgery to their last follow‐up observations (33.6% of the 110 patients with hypertension at baseline). Over long‐term follow‐up, the composite cumulative remission and improvement rate of hypertension was as high, i.e. 70.0% (n=77). This monitoring of clear statuses of hypertension in long‐term follow‐up is crucial, on the basis that hypertension is a complex measurement that often the hypertensive patients experience a period of remission while not receiving antihypertensive therapy, but many eventually relapse (328).

Changes in hyperlipidaemia status

Health improvements in hyperlipidaemia showed variability in response as contrast to T2DM and hypertension (Figure 2.12). At pre‐operative baseline, 99 patients (58.9%) had hyperlipidaemia. Owing to the strict criteria of needing to fulfill discontinuation of lipid‐lowering medications and

130

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

normalization of the whole lipid profile comprising four subfractions (namely total cholesterol, triglycerides, HDL‐C and LDL‐C), hyperlipidaemia remission and improvement rates over the post‐ operative years in the current study were unsurprisingly relatively low.

Figure 2.12 Yearly remission, improved, persisting and worsened rates of hyperlipidaemia following bariatric surgery

40 %

30 Patient,

0 0123456

Follow‐up after bariatric surgery, Year

Remission Improved Persisting Worsened

Follow‐up, year 0 1 2 3 4 5 6 Remission, n (%) 16 (19.3%) 9 (12.9%) 9 (17.0%) 6 (14.3%) 6 (17.6%) 4 (15.4%) (95 CI%) (12.0–26.5) (7.1–18.6) (9.4–24.5) (7.1–23.8) (8.8–26.5) (7.7–26.9) Improved, n (%) 15 (18.1%) 16 (22.9%) 11 (20.8%) 7 (16.7%) 5 (14.7%) 6 (23.1%) (95 CI%) (10.8–25.3) (15.7–30.0) (13.2–28.3) (7.1–26.2) (5.9–23.5) (11.5–34.6) Unchanged, n (%) 49 (59.0%) 41 (58.6%) 28 (52.8%) 25 (59.5%) 19 (55.9%) 13 (50.0%) (95 CI%) (49.4–69.8) (48.6–68.6) (41.5–66.0) (47.6–71.4) (41.2–70.6) (34.6–65.4) Worsened, n (%) 3 (3.6%) 4 (5.7%) 5 (9.4%) 4 (9.5%) 4 (11.8%) 3 (11.5%) (95 CI%) (0.0–8.4) (1.4–10.0) (3.8–15.1) (4.8–14.3) (5.9–17.6) (0.0–23.1) Total patientsa 99 (58.9%) 83 (100%) 70 (100%) 53 (100%) 42 (100%) 34 (100%) 26 (100%)

Total cholesterol, 4.4±1.2 4.7±1.0 4.6±1.1 4.6±1.1 4.6±0.9 4.6±0.8 4.6±1.2 mmol/L Triglycerides, mmol/L 1.9±0.9 1.4±0.9 1.5±1.1 1.4±0.7 1.4±0.7 1.5±0.8 1.6±0.8 HDL‐C, mmol/L 1.2±0.3 1.4±0.3 1.5±0.5 1.5±0.4 1.5±0.5 1.5±0.4 1.3±0.3 LDL‐C, mmol/L 2.5±1.2 2.7±0.9 2.4±0.9 2.5±1.0 2.5±0.8 2.4±0.8 2.6±1.0

131

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

Lipid‐lowering drugs† Yes, n (%) 80 (80.8%) 55 (59.1%) 49 (64.5%) 35 (60.3%) 31 (64.6%) 25 (64.1%) 18 (64.3%) Mean number of drug (±SD) 1.0±0.6 0.7±0.6 0.7±0.6 0.7±0.6 0.7±0.6 0.8±0.7 0.8±0.7 Range 0–3 0–2 0–2 0–2 0–2 0–2 0–2 aPatients with hyperlipidaemia present at pre‐operative baseline Bootstrap 95% confidence interval (CI) was computed by the bias‐corrected and accelerated (BCa) method †Statin, fibrate and/or ezetimibe Abbreviations: SD=Standard deviation; HDL‐C=High‐density lipoprotein cholesterol; LDL‐C=Low‐density lipoprotein cholesterol

Hyperlipidaemia was in remission for over 10% of study patients at year 1 to nearly 20% of at year 6 post‐operation. Slightly higher proportions of patients experienced improvement across the follow‐up timepoints, as compared to remission (Figure 2.12). The proportion of patients with improvement of hyperlipidaemia increased from 18.1% (95% CI=10.8–25.3) at year 1 to 23.1% (95% CI=11.5–34.6) at year 6 post‐surgery. As a result of the same strict criteria, approximately half of the patients who continued with their current daily lipid‐lowering agent treatment were considered having persisting hyperlipidaemia, as portrayed in Figure 2.12. Whereas there was an increasing linear trend of patients who had worsening hyperlipidaemia from 3.6% to 11.5%. Despite so, the number of patients remained at 3 individuals at year 1 and year 6 post‐surgery. Similar to T2DM, only one incidence requiring lipid‐lowering therapy emerged de novo.

Subsequently, all the lipid values and details of lipid‐lowering drugs at all timepoints are reported in Figure 2.12. With the control of lipid profile, compared to baseline, amelioration was seen across all years after surgery more so in triglyceride and HDL‐C values, but not the case in total cholesterol and LDL‐C levels. The changes in both the latter lipid levels were less marked as if there is no improvement during follow‐up. However, it might be worth mentioning that these lipid parameters, specifically total cholesterol and LDL‐C levels have reached normalization since baseline and sustained across post‐operative follow‐up. Of note, they were likely plateaued by the lipid‐lowering agents thereafter. In spite of the fact that the total number of patients who were prescribed lipid‐lowering agents steeply decreased over time (with approximately 20% patients needed less medications each year), over half of the patients remained on the therapy post‐operatively. Despite fallen in the abnormal range at pre‐operative baseline, triglycerides marker achieved progressive improvement well after bariatric surgery throughout the 6 years of

132

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan follow‐up period. Irrespective of the normal HDL‐C level at pre‐operative baseline, the mean value increased progressively over follow‐up time points until 5 years, with the year 6 timepoint one exception but still remained slightly higher than the baseline average reading.

Super obesity (SO) versus morbid obesity (MO)

The yearly remission, improvement, persisting and worsening rates of T2DM, hypertension and hyperlipidaemia between the SO and MO groups following bariatric surgery over 6 years of time are illustrated in Figures 2.13 (a) to (c). The exact number of observations, couple with its percentages, are summarized in the following Table 2.5. Fisher’s exact test was used to compare the rates of comorbidity statuses over Pearson’s chi‐square (χ2) test due to a low number of observations in certain comorbidity status in each year, particularly the worsening status that required Fisher’s exact test.

133

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

Figure 2.13 (a)–(c) Comparison of SO and MO groups on the yearly changes of remission, improvement, persisting and worsening rates for the selected obesity‐related comorbidities post‐operatively£, ¥

(a) Changes in T2DM 100%

90%

80%

2.4 0.9 4.7 70% 12.9 4.0 9.3 6.7 11.2 14.3 60% 6.0 12.9 8.3

% 50% 27.9 23.4 26.0 22.6 21.7 40% 25.8 Patients, 0.0 0.0 0.9 0.0 30% 8.3 0.0 5.6 10.0 4.8 3.2 5.6 6.7 7.0 20% 6.0 4.0 33.6 4.7 32.1 30.0 32.0 32.6 6.5 25.8 3.2 10% 18.7 17.9 18.3 18.0 14.0 9.7 0% MO SO SO MO MO SO SO MO SO SO MO SO SO MO SO SO MOMO SO Year 1 Year 3 Year 4 Year 5 Year 6 Year 1Year Year 2 2Year 3Year 4Year 5Year 6 Follow‐up after bariatric surgery, Year Remission Improvement Persistence Worsened Series5

134

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

(b) Changes in hypertension

100%

90%

80% 3.1 70% 3.0 3.7 7.0 3.7 3.4 9.4 60%

% 15.8

14.8 18.6 14.8 50% 18.6 28.1 3.4 Patients, 40% 2.5 1.9 18.8 1.0 13.6 24.1 20.9 2.3 30% 27.2 12.9 12.3 11.9 14.8 6.3 11.6 20% 7.4 8.9 11.9 31.3 12.5 27.7 27.1 9.3 10% 22.2 23.3 9.3 17.3 14.8 11.9 10.2 6.3 9.3 7.0 0% 3.1 MO SO SO MO SO SO MO SO MOMO SO SO MO SO SO MO SO SO YearYear 1 1Year Year 2 2Year Year 3 3Year Year 4 4Year Year 5 5Year Year 6 6 Follow‐up after bariatric surgery, Year Remission Improvement Persistence Worsened Series5

135

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

100%

90%

2.9 1.2 3.8 80% 11.5 4.8 8.8 70%

60% %

47.0 48.6 41.5 42.3 50% 47.6 47.1 Patients, 40%

30% 2.9 13.3 2.4 17.0 2.9 4.8 5.7 19.2 20% 17.1 0.0 14.3 8.8 11.8 12.0 10.0 11.9 10% 11.3 2.9 7.7 16.9 15.1 11.4 2.4 11.5 3.8 4.8 5.7 3.8 9.5 8.8 8.8 4.8 3.8 0% 2.4 1.4 1.9 MOMO SO SO MO SO SO MO SO SO MOMO SO SO MOMO SO SO MOMO SO SO Year 3 Year 4 Year 6 YearYear 1 1Year Year 2 2Year 3Year 4Year Year 5 5Year 6 Follow‐up after bariatric surgery, Year Remission Improvement Persistence Worsened Series5

Abbreviations: SO=Super obesity; MO=Morbid obesity £Fisher’s exact test demonstrated p values <0.05 for comparison between MO and SO subgroups of each year ¥The percentage of patients achieving remission, improvement, persistence and worsening statuses in each year of each metabolic comorbidity for MO and SO groups adds up to 100%.

136

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

Table 2.5 Annual remission, improvement, persistence and aggravation rates of obesity‐related comorbidities after bariatric surgery between the SO and MO groups over 6 years of follow‐up£ Patient, n (%) Year 1 Year 2 Year 3 Year 4 Year 5 Year 6 MO SO MO SO MO SO MO SO MO SO MO SO T2DM Remission 36 20 27 15 18 11 16 9 14 6 8 3 (33.6%) (18.7%) (32.1%) (17.9%) (30.0%) (18.3%) (32.0%) (18.0%) (32.6%) (14.0%) (25.8%) (9.7%) Improvement 25 6 19 5 13 4 13 2 12 2 8 1 (23.4%) (5.6%) (22.6%) (6.0%) (21.7%) (6.7%) (26.0%) (4.0%) (27.9%) (4.7%) (25.8%) (3.2%) Persistence 12 6 12 4 5 5 3 5 4 3 4 2 (11.2%) (5.6%) (14.3%) (4.8%) (8.3%) (8.3%) (6.0%) (10.0%) (9.3%) (7.0%) (12.9%) (6.5%) Worsening 1 1 2 0 4 0 2 0 2 0 4 1 (0.9%) (0.9%) (2.4%) (0.0%) (6.7%) (0.0%) (4.0%) (0.0%) (4.7%) (0.0%) (12.9%) (3.2%)

Hypertension Remission 28 12 14 12 16 6 12 5 10 3 10 1 (27.7%) (11.9%) (17.3%) (14.8%) (27.1%) (10.2%) (22.2%) (9.3%) (23.3%) (7.0%) (31.3%) (3.1%) Improvement 19 9 22 6 8 7 13 5 9 4 9 2 (18.8%) (8.9%) (27.2%) (7.4%) (13.6%) (11.9%) (24.1%) (9.3%) (20.9%) (9.3%) (28.1%) (6.3%) Persistence 16 13 12 10 11 7 8 8 8 5 3 4 (15.8%) (12.9%) (14.8%) (12.3%) (18.6%) (11.9%) (14.8%) (14.8%) (18.6%) (11.6%) (9.4%) (12.5%) Worsening 3 1 3 2 2 2 2 1 3 1 1 2 (3.0%) (1.0%) (3.7%) (2.5%) (3.4%) (3.4%) (3.7%) (1.9%) (7.0%) (2.3%) (3.1%) (6.3%)

Hyperlipidaemia Remission 14 2 8 1 8 1 4 2 3 3 3 1 (16.9%) (2.4%) (11.4%) (1.4%) (15.1%) (1.9%) (9.5%) (4.8%) (8.8%) (8.8%) (11.5%) (3.8%) Improvement 11 4 12 4 9 2 6 1 4 1 5 1 (13.3%) (4.8%) (17.1%) (5.7%) (17.0%) (3.8%) (14.3%) (2.4%) (11.8%) (2.9%) (19.2%) (3.8%) Persistence 39 10 34 7 22 6 20 5 16 3 11 2 (47.0%) (12.0%) (48.6%) (10.0%) (41.5%) (11.3%) (47.6%) (11.9%) (47.1%) (8.8%) (42.3%) (7.7%) Worsening 1 2 2 2 2 3 2 2 3 1 3 0 (1.2%) (2.4%) (2.9%) (2.9%) (3.8%) (5.7%) (4.8%) (4.8%) (8.8%) (2.9%) (11.5%) (0.0%)

£Fisher’s exact test demonstrated a p value of <0.05 for all years; Abbreviations: SO=Super obesity; MO=Morbid obesity; T2DM=Type 2 diabetes mellitus

137

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

The annual changes of comorbidity statuses of T2DM, hypertension and hyperlipidaemia were demonstrated proportionally to their respective pre‐operative baseline conditions. The subanalysis comparing outcomes of the above selected comorbidities between the SO and MO groups revealed no significant difference (p>0.05) between the SO and MO groups according to the rates of the remission, improvement, persistence and worsening of the comorbidities over the entire follow‐up period.

The subgroup analysis for changes in surgical complications and LOHS over time showed no significant difference between these two groups (p>0.05). Our study concluded that both the SO and MO groups benefit equally from bariatric surgery.

138

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

Sleep‐Disordered Breathing [Obstructive Sleep Apnoea (OSA) or Obesity Hypoventilation Syndrome (OHS)] and continuous positive airway pressure (CPAP)/bilevel positive airway pressure (BiPAP) device use

The impact of bariatric surgery on the sleep‐disordered breathing [obstructive sleep apnoea (OSA) or obesity hypoventilation syndrome (OHS)] as well as the use of continuous positive airway pressure (CPAP)/bilevel positive airway pressure (BiPAP) devices are noted in Figure 2.14.

Figure 2.14 Prevalence of sleep‐disordered breathing (OSA/OHS) and CPAP/BiPAP prescription among the study population undergoing bariatric surgery

TITLE 40

AXIS 30 Patient,

0 0123456 Follow‐up afterAXIS bariatric TITLE surgery, Year

OSA/OHS diagnosis CPAP/BiPAP Prescription

Follow‐up, year 0 1 2 3 4 5 6 Sleep‐disordered breathing (OSA/OHS) Number of patients (n) 107/168 90/161 73/125 54/95 46/77 32/59 20/38 Percentage (%) 63.7% 55.9% 58.4% 56.8% 59.7% 54.2% 52.6% CPAP/BiPAP prescription Number of patients (n) 87/168ᵠ 59/161 48/125 36/95 28/77 20/59 12/38 Percentage (%) 51.8% 36.6% 38.4% 37.9% 36.4% 33.9% 31.6% Abbreviations: OSA=Obstructive sleep apnoea; OHS=Obesity hypoventilation syndrome; CPAP=Continuous positive airway pressure; BiPAP=Bilevel positive airway pressure ᵠ8 patients had OHS on BiPAP at pre‐operative baseline, the remaining were OSA requiring CPAP

139

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

Prior to bariatric surgery, as high as 63.7% of the patients (107 of 168 patients) who underwent bariatric surgery were diagnosed as having either OSA or OHS. Ameliorations of OSA/OHS were seen to improve over timepoints, reduced to a prevalence of 52.6% at year 6 post‐surgery. The prevalence of CPAP or BiPAP device required by the study patients was also observed to decrease tremendously from 51.8% at bariatric surgery and on a slight upward trend across year 2, then continue to downtrend again year 3 onwards throughout the follow‐up period to 31.6%. Aside from this, a total of 32 patients with OSA and were prescribed CPAP no longer required CPAP to their last observations following bariatric surgery, equivalent to a cumulative improvement rate of 36.8%. There was only one new onset of diagnosed mild OSA during the post‐operative follow‐ up period, with no CPAP required.

The CPAP/BiPAP adherence rate at pre‐operative baseline was 64.4%, i.e. 47 out of the 73 patients requiring CPAP/BiPAP. In other terms, 35.6% (26 patients) were not adherent to CPAP/BiPAP prescriptions before surgery. Whereas 18 chose to stop CPAP/BiPAP use post‐ surgery, i.e. self‐discontinuation. It is worthwhile to note that the top two patient self‐reported factors found to influence non‐progression to CPAP/BiPAP device were intolerance to the machines and financial issue.

140

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

Mental illness and antidepressants and/or antianxiety agents use

Figure 2.15 provides the prevalence of mental illness, specifically depression and/or severe anxiety, as well as the use of antidepressants and/or antianxiety agents.

Figure 2.15 Prevalence of depression and/or severe anxiety alongside the antidepressants and/or antianxiety agents use among the study population undergoing bariatric surgery

TITLE 30 AXIS

Patient, 20

0 0123456 Follow‐up after bariatricAXIS TITLE surgery, Year

Depression/Severe anxiety Antidepressants/ Antianxiety agents

Follow‐up, year 0 1 2 3 4 5 6 Depression and/or severe anxiety Number of patients (n) 79/168 55/167 43/132 35/100 31/81 31/62 20/41 Percentage (%) 47.0% 32.9% 32.6% 35.0% 38.3% 50.0% 48.8% Antidepressants† and/or antianxiety agents‡ Number of patients (n) 52/167 42/165 31/129 26/99 24/80 25/61 18/40 Percentage (%) 31.1% 25.5% 24.0% 26.3% 30.0% 41.0% 45.0% †Amitriptyline, citalopram, desvenlafaxine, duloxetine, escitalopram, fluoxetine, fluvoxamine, mirtazapine, moclobemide, paroxetine, sertraline, venlafaxine ‡Alprazolam, diazepam, lorazepam, oxazepam

Almost half of the patients (47.0%) were diagnosed as having depression and/or severe anxiety before bariatric surgery. As can be seen from Figure 2.15, change in the prevalence of these mental illnesses over time did not differ greatly between the baseline and year 6 of follow‐up.

141

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

Specifically, compared to pre‐surgery (47.0%), although the prevalence dropped at years 1 to 4 post‐surgery, it then increased and became slightly higher versus baseline by the fifth (50.0%) and sixth (48.8%) years.

Of the patients, 31.1% were found to be on regular antidepressants and/or antianxiety agents pre‐surgically. In agreement with its prevalence of depression and severe anxiety, the prevalence of use of the medications followed a quadratic trend ‐ first decreased from baseline to year 2 post‐surgery (24.0%), then gradually increasing over time to a prevalence higher than baseline at years 5 (41.0%), and continued to hit a new record high through year 6 (45.0%) (Figure 2.15).

Our additional evaluation of mental health profiles of the study cohort shows that pre‐existing post‐traumatic stress disorder (PTSD) was co‐existed with 4 patients with depression and/severe anxiety despite after bariatric surgery. According to the psychological or psychiatric assessment reports, 3 of these 4 patients had complex traumas from sexual abuse and parent’s sudden death (n=1), witness parent being killed in a fatal car accident (n=1) and domestic violence (n=1). Besides PTSD and depression and/or severe anxiety, these traumas had also cited as contributing to their ongoing emotional eating and/or binge eating disorders. On the other hand, it is interesting to note that the other 2 patients with depression and/or severe anxiety and PTSD reported reversed mental issues following multidisciplinary surgical treatment.

Opioid use and total joint arthroplasty (TJA)

Further to the continual improvements in osteoarthritis (OA) and the associated joint pain and symptoms from baseline through 6 years post‐surgery as demonstrated in Figure 2.9, Figure 2.16 below shows the prevalence of prescribed opioid use and total joint arthroplasty (TJA) in the clinically severe obese patients having concurrent OA and/or weight‐bearing joint pain (WBJP).

142

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

Figure 2.16 Prevalence prescribed opioid use and total joint arthroplasty (TJA) among the patients with OA and/or weight‐bearing joint pain (WBJP) 25

15 %

TITLE

AXIS 10 Patient,

0 0123456 Follow‐up afterAXIS bariatric TITLE surgery, Year

Opioid use TJA (inclusive of revision)

Follow‐up, year 0 1 2 3 4 5 6 Opioid use† Number of patients (n) 19/168 18/158 15/122 11/88 8/75 11/56 7/39 Percentage (%) 11.3% 11.4% 12.3% 12.5% 10.7% 19.6% 17.9% Total joint arthroplasty (TJA) (including revisional surgery)ᵠ Number of patients (n) 13/168 5/158 8/122 6/90 2/76 2/56 2/41 Percentage (%) 7.7% 3.2% 6.6% 6.7% 2.6% 3.6% 4.9% †Codeine, buprenorphine, fentanyl, oxycodone, tapentadol, tramadol ᵠDerived from data linkage with the nationwide Australian Orthopaedic Association National Joint Replacement Registry (AOANJRR)

Among those with OA and/or WBJP, 11.3% of them were prescribed opioids before bariatric surgery, the cornerstone of pain management in surgeries (329). Out of these 11.3% patients, 11 of the 19 patients had ongoing opioid prescriptions through their last observations. The opioid use or dependency increased annually post‐surgery across all timepoints from baseline to particularly year 5 (19.6%) and year 6 (17.9%), with year 4 the exception (Figure 2.16), despite steady improvement in OA, WBJP and the localized pain. Specifically, among the 11 patients with OA and/or WBJP who were prescribed opioids at year 5, they were suffering from severe degree of OA (n=3), both OA and gout (n=2), both OA and depression (n=3), both OA and degenerative

143

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

disc disease (n=1), and pain after TJA (n=2) that caused significant pain that required opioid pain relievers. Whereas 6 years after bariatric surgery, the 7 patients with OA and/or WBJP who were prescribed opioids were persistent users for severe OA (n=3), and 4 patients were found suffering from both OA and depression. However, at any given timepoint, the prevalence of opioid use was less than 20% across follow‐ups.

Among the 19 patients with continued baseline chronic use of opioids, 8 patients off opioid medications or changed to non‐opioid anti‐inflammatory and pain‐relief agents to their last observations (42.1%). Whereas 14 patients were post‐surgery initiated opioid user (i.e. post‐ surgery use of opioids without pre‐surgery use). In contrast, TJA rate decreased from 7.7% at baseline to 4.9% at year 6 post‐bariatric surgery (i.e. 2 of 41 patients at year 6). Additional results from the national registry were also extended to years 7 and 8 of follow‐up (not shown in Figure 2.16), with only 2 and 1 patients underwent TJA, respectively.

Hyperuricaemia

The mean pre‐operative serum uric acid level was 0.36 mmol/L, with 31 cases (30.4%) classified as having hyperuricemia at the time of bariatric surgery (Table 2.6).

Table 2.6 Prevalence of hyperuricaemia§ Follow‐up, year 0 1 2 3 4 5 6 Number of patients (n) 31/102 11/100 10/55 8/46 10/35 9/31 3/24 Percentage (%) 30.4% 11.0% 18.2% 17.4% 28.6% 29.0% 12.5% Urate level, mean (±SD) 0.36±0.10 0.31±0.08 0.34±0.11 0.33±0.08 0.34±0.09 0.34±0.09 0.32±0.07 (mmol/L) §Hyperuricaemia is defined as a uric acid level of >0.40 mmolL−1

Hyperuricaemia improved greatly after bariatric surgery to as low as 3 events (12.3%) at year 6 post‐operatively, although with variable fluctuations in between the follow‐up term. Mean uric acid level dropped as well, despite the mean values have been within the normal range across all timepoints.

144

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

Safety of bariatric surgery: Surgical complications and subsequent surgery (revisions and reversals)

The details of a variety of major/minor and early/late post‐operative complications is summarized in Table 2.7. Of patients subdivided into four groups according to the type of bariatric surgery, 58 (34.5%) suffered from early‐ and late‐ post‐operative complications. Six patients underwent revisional surgeries due to surgical complications after their initial bariatric procedures, of which two patients underwent unplanned conversion to RYGB for severe GORD.

Table 2.7 Peri‐operative and late post‐operative complications by bariatric procedures (n=168) Complication category and type SG MGB‐OAGB AGB RYGB (n=141) (n=15) (n=11) (n=1)

Peri‐operative outcomes (<30 days), No. (%)

Major

Haemorrhage 1 (0.7%) 0 (0) 0 (0) 0 (0)

Severe constipation Severe constipation requiring admission 1 (0.7%) 0 (0) 0 (0) 0 (0)

Severe GORD requiring admission and endoscopy 0 (0) 0 (0) 1 (9.1%) 0 (0)

Pneumonia 1 (0.7%) 0 (0) 0 (0) 0 (0)

Intestinal obstruction 0 (0) 1 (6.7%) 0 (0) 0 (0)

Bariatric surgical‐related hernia repair (Umbilical) 0 (0) 1 (6.7%) 0 (0) 0 (0)

Minor

Refractory hypertension 1 (0.7%) 0 (0) 0 (0) 0 (0)

Bile leak 1 (0.7%) 0 (0) 0 (0) 0 (0)

Wound dehiscence 0 (0) 0 (0) 1 (9.1%) 0 (0)

Haematoma Haematoma 0 (0) 0 (0) 1 (9.1%) 0 (0) Haematoma on anticoagulant 0 (0) 0 (0) 1 (9.1%) 0 (0)

Severe constipation 1 (0.7%) 0 (0) 0 (0) 0 (0)

145

Complication category and type SG MGB‐OAGB AGB RYGB (n=141) (n=15) (n=11) (n=1)

Persistent tachycardia and/or tachypnoea 1 (0.7%) 0 (0) 0 (0) 0 (0)

Iron deficiency anaemia 1 (0.7%) 0 (0) 0 (0) 0 (0)

Acute gout 1 (0.7%) 0 (0) 0 (0) 0 (0)

Dumping syndrome 0 (0) 1 (6.7%) 0 (0) 0 (0)

Late post‐operative outcomes (>30 days), No. (%)

Major

Bariatric surgical‐related hernia repair Umbilical 0 (0) 1 (6.7%) 1 (9.1%) 0 (0) Internal 0 (0) 1 (6.7%) 0 (0) 0 (0) Paraumbilical 1 (0.7%) 0 (0) 0 (0) 0 (0) Incisional 1 (0.7%) 0 (0) 0 (0) 0 (0) Ventral 1 (0.7%) 0 (0) 0 (0) 0 (0)

Severe GORD Severe GORD requiring surgical correction (hiatus hernia repair) 3 (2.1%) 0 (0) 0 (0) 0 (0) Severe GORD requiring surgical correction (conversion to RYGB) 2 (1.4%) 0 (0) 0 (0) 0 (0)

Gastric band erosion 0 (0) 0 (0) 1 (9.1%) 0 (0)

Incisura obstruction 1 (0.7%) 0 (0) 0 (0) 0 (0)

Intestinal obstruction 1 (0.7%) 0 (0) 0 (0) 0 (0)

Obstruction of gastric band 0 (0) 0 (0) 1 (9.1%) 0 (0)

Gastric perforation 1 (0.7%) 0 (0) 0 (0) 0 (0)

Cholelithiasis requiring surgery 0 (0) 2 (13.3%) 0 (0) 0 (0)

Wound dehiscence 0 (0) 1 (6.7%) 0 (0) 0 (0)

Minor

Haematoma 1 (0.7%) 0 (0) 0 (0) 0 (0)

Severe GORD Significant GORD 12 (8.5%) 1 (6.7%) 2 (13.3%) 0 (0) Severe GORD requiring endoscopy 3 (2.1%) 1 (6.7%) 0 (0) 0 (0) GORD requiring endoscopy and dilation 4 (2.8%) 0 (0) 0 (0) 0 (0)

146

Complication category and type SG MGB‐OAGB AGB RYGB (n=141) (n=15) (n=11) (n=1)

Barrett's oesophagus 1 (0.7%) 0 (0) 0 (0) 0 (0)

Stricture 3 (2.1%) 1 (6.7%) 0 (0) 2 (200%)

Adhesion 0 (0) 1 (6.7%) 0 (0) 0 (0)

Dumping syndrome 0 (0) 2 (13.3%) 1 (9.1%) 0 (0)

Severe constipation 5 (3.5%) 0 (0) 0 (0) 0 (0)

Iron deficiency anaemia 1 (0.7%) 0 (0) 0 (0) 0 (0)

Dilated oesophagus due to gastric band 0 (0) 0 (0) 1 (9.1%) 0 (0)

Oesophageal dysmotility 1 (0.7%) 0 (0) 1 (9.1%) 0 (0)

Cholelithiasis not requiring surgery 1 (0.7%) 0 (0) 0 (0) 0 (0) Abbreviations: BMI=Body mass index, GORD=Gastro‐oesophageal reflux disease

Of the early/peri‐operative surgical complications, there were 3 major adverse events, namely haemorrhage, pneumonia and severe constipation requiring admission. The late major post‐ operative complications occurred more common following LSG. This includes severe GORD requiring surgical corrections, incisura obstruction, intestinal obstruction, gastric perforation, and bariatric surgical‐related hernia repair. MGB‐OAGB followed after SG closely, with 1 wound dehiscence and 2 cholelithiasis (requiring additional surgery) developed after bariatric surgery, while 2 of the MGB‐OAGB patients were treated with hernia repairs. Surgical complications directly related to the AGB were mainly severe GORD, wound dehiscence, haematoma, band erosion, obstruction of gastric band, dilated oesophagus and oesophagus dysmotility. Of the only patient underwent RYGB as index surgery, there were occurrence of two strictures as shown in Table 2.7.

A substantial number of patients (n=19) underwent LSG experienced severe GORD. In this series, 3 of them required endoscopy, and 4 required endoscopy and dilation. This is not uncommon,

147

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan given one potential drawback of SG is always the high incidence of exacerbation or new onset of GORD, and Barrett’s oesophagus, which is in line with our result tabulated in Table 2.7 that the LSG patient has also presented with a Barrett’s oesophagus.

A total of 11.5% of patients who were treated with the gastric bypass procedures (MGB‐OAGB and AGB) experienced dumping syndrome (n=3), this is perhaps not surprising given its known association with gastric bypass that more commonly seen than other type of procedures such as SG.

In total, there were 6 patients who required major surgical‐related hernia repair, of which one was at the early stage (<30 days after bariatric surgery), and 5 beyond post‐operative day 30. This included 2 umbilical, 1 internal, 1 paraumbilical, 1 incisional and 1 ventral hernia repair, respectively, most commonly seen among our patients who were performed LSG and MGB‐OAGB (Table 2.7).

Most frequently occurred minor complications among the 55 incidences, regardless of early or late, being severe GORD (n=23), strictures (n=6) and severe constipations (n=5).

Amongst the patients with late post‐operative complications, 9 patients had 2 complications, 1 patient suffered from 3 complications. One patient who had MGB‐OAGB reported 4 post‐surgical complications, including stricture, bariatric surgical‐related repairs (umbilical and internal, respectively), and adhesiolysis (i.e. division of adhesions) with closure of petersen defect.

148

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

Nutrient deficiencies

Table 2.8 summarizes the prevalence of nutrient abnormalities before and after bariatric surgery. A double elevation of prevalence of iron deficiency anaemia was noted between pre‐ operative baseline and 6th‐year post‐operative mark; started with as low as 7.1% patients have had iron deficiency anaemia before surgery, as reflected by the ferritin and transferrin saturation levels. Over time, the prevalence of iron deficiency anaemia exacerbated post‐ surgery in approximately two‐fold (14.3%) at year 6 of follow‐up.

Table 2.8 Prevalence of nutrient deficiencies pre‐ and post‐bariatric surgery Nutritional parameter Iron deficiency Vitamin D Vitamin B12 anaemia† deficiency‡ insufficiency₤

Pre‐operative baseline Number of patients 10/140 43/135 6/120 Percentage (95% CI) 7.1% (3.6–10.7) 31.9% (24.4–39.3) 5.0% (2.5–8.3) Year 1 post‐surgery Number of patients 4/136 25/138 10/133 Percentage (95% CI) 2.9% (0.7–5.9) 18.1% (12.3–24.6) 7.5% (4.5–10.5) Year 2 post‐surgery Number of patients 6/101 23/106 4/99 Percentage (95% CI) 5.9% (3.0–8.9) 21.7% (15.1–29.2) 4.0% (1.0–8.1) Year 3 post‐surgery Number of patients 7/76 13/70 3/71 Percentage (95% CI) 9.2% (5.3–13.2) 18.6% (11.4–25.7) 4.2% (0.0–8.5) Year 4 post‐surgery Number of patients 6/56 8/57 1/53 Percentage (95% CI) 10.7% (5.4–16.1) 14.0% (8.8–19.3) 1.9% (0.0–7.5) Year 5 post‐surgery Number of patients 2/42 3/41 0/40 Percentage (95% CI) 4.8% (0.0–11.9) 7.3% (0.0–14.6) 0.0% Year 6 post‐surgery Number of patients 5/35 4/31 0/32 Percentage (95% CI) 14.3% (5.7–22.9) 12.9% (6.5–22.6) 0.0% Bootstrap 95% confidence interval (CI) was computed by the bias‐corrected and accelerated (BCa) method †Iron deficiency anaemia is defined by a ferritin level <20 μg/L and a transferrin saturation level <15% ‡Vitamin D deficiency is defined by vitamin D level <50 nmol/L ₤ Vitamin B12 insufficiency is defined by vitamin B12 level <150 pmol/L

149

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

By contrast, the prevalence of vitamin D deficiency was considerably high before bariatric surgery, as observed in nearly one‐third (31.9%) of the study cohort. It then followed a secular trend over the follow‐up years and rapidly decreased after surgery to 12.9% at the year 6 evaluation (Table 2.8). Of the 43 patients presented with vitamin D deficiency pre‐surgically, 27 with vitamin D levels normalized to their last observations after surgery.

In the assessment of vitamin B12, or cobalamin deficiency, low prevalence of insufficiency in

vitamin B12 was detected before bariatric surgery (5.0%). After bariatric surgery, the deficiency occurred in 7.5%, 4.0%, 4.2% and 1.9% patients at years 1, 2, 3 and 4, respectively, with no patient developing a deficiency in years 5 and 6 post‐operatively (0%). Nonetheless, all the average levels of serum vitamins and a trace mineral before and after bariatric surgery remained well within normal ranges as presented in the following Table 2.9.

Table 2.9 Mean values of nutritional parameters pre‐ and post‐bariatric surgerya Nutritional parameter Iron studies 25‐OH Vitamin D Vitamin B12 (nmol/L) (pmol/L) Ferritin Transferrin (µg/L) saturation (%) Pre‐operative baseline 121.5±116.4 19.5±8.3 60.5±21.8 354.5±246.6 Year 1 post‐surgery 134.1±126.5 27.3±9.6 70.1±26.4 351.1±185.6 Year 2 post‐surgery 109.1±105.5 25.9±10.5 66.0±20.6 382.4±169.2 Year 3 post‐surgery 79.5±81.6 25.2±11.0 66.8±20.2 393.2±187.8 Year 4 post‐surgery 86.3±69.5 24.6±10.8 67.0±23.2 394.0±206.6 Year 5 post‐surgery 66.7±51.3 22.9±10.4 77.1±18.4 398.3±175.2 Year 6 post‐surgery 73.1±61.4 22.3±9.6 69.5±18.2 363.8±162.1 aValues are expressed as mean±SD. Number of patients is specified for each measurement according to available data. Abbreviation: SD=Standard deviation

Meaningful comparison of pre‐ and post‐surgery mean levels of serum ferritin, transferrin

saturation, 25‐OH vitamin D and vitamin B12 is limited by no significant trend across time. However, as shown in Table 2.9, the lowest average of vitamin D level in the study cohort was observed at pre‐surgery timepoint, rather than the rest of post‐surgery clinic follow‐up years. This implies a possibility of high prevalence of vitamin D deficiency even prior to surgery, and this

postulation was further confirmed by Table 2.8.

150

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

Mortality For this part of the study, the mortality was evaluated from the time at bariatric surgery until the 9th year of follow‐up. As listed in Table 2.10, three deaths were reported during years 6, 8, and 9 of follow‐up period (i.e. 1.8% mortality rate of the study cohort), with none occurring in the immediate 30‐day post‐surgical period (0%). All causes of death were deemed not related to the bariatric surgery, of which two patients died from cardiac events and one patient died from an acute kidney injury related to vasculitis. Among the 11 patients (18.2%) who underwent LAGB, two out of them were reported death, which was two‐thirds of the deaths reported in the present study. All the 3 deceased patients had the shortest LOHS of only one day during their primary surgical procedure.

151

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

Table 2.10 Characteristics of study patients who died in‐hospital (n=3) Patient Operation Sex Race Age at Weight BMI LOHS Cause of death No. primary (Kg) (Kg/m2) (days) surgery At At At At (Year) bariatric palliative bariatric palliative surgery care surgery care ward ward 1 LAGB Female Caucasian 63 121.3 88.3 49.2 35.8 1 Acute kidney injury secondary to vasculitis, unrelated to bariatric procedure

2 LAGB Female Caucasian 59 167.4 162.8 52.8 51.4 1 Cardiac event, unrelated to bariatric procedure

3 LSG, Male Middle 49 174 117.5 53.1 35.9 Primary Cardiac event, unrelated to followed Eastern surgery=1, bariatric procedure by conversion Revisional to RYGB surgery=6

Abbreviations: Kg=Kilograms; LOHS=Length of hospital stay, LAGB=Laparoscopic adjustable gastric banding, LSG=Laparoscopic sleeve gastrectomy, RYGB=Roux‐ en‐Y gastric bypass.

152

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

2.5 DISCUSSION

The burden of the obesity epidemic in modern society has prompted the development of novel and a broad range of weight management options to support the efforts of those needing to lose weight and alleviate obesity‐related comorbidities, with a consensus that bariatric surgery is at the top of the obesity treatment pyramid. Bariatric surgery is well‐established as the most effective treatment for severe and complex obesity that has proven unmanageable by pharmacotherapy and, dietary and exercise modifications. This treatment also provides an ideal platform for research of the highest‐grade obesity and its impacts in view of the fact that the bariatric surgical cohort is a high‐risk population, with detailed clinical data in our publicly funded bariatric surgery service. This state‐wide bariatric surgery service adopts a multidisciplinary approach employing endocrinology, bariatric surgical, nursing, nutrition and dietetics, exercise physiological and clinical psychological domains in managing the complex patients. Moreover, this population facilitates the opportunity to observe the long‐term effects of bariatric surgery on weight loss, improvement of comorbidities, and safety, which to date no Australian report has been able to provide an in‐depth evaluation of comprehensive datasets from three multidisciplinary settings, at least in Sydney, in addition to linkage with the interrelated orthopaedic surgery data, the Australian Orthopaedic Association National Joint Replacement Registry (AOANJRR).

In light of the overwhelming demand for surgical treatment for obesity via publicly funded specialist obesity services, it is important to systematically capture standardized clinical data for the ongoing monitoring and evaluation of clinical care. It is also important for promoting collaboration between health services such as information‐sharing for quality assurance as well as data linkages for future research purposes (16). As a result of limited bariatric surgical procedures in the public sector and sparse research funding, there is still a limited understanding of the management, benefits and risks of publicly funded bariatric surgery in the settings of clinically severe obesity and its metabolic comorbidities. We conducted this longitudinal study, developed an electronic database and captured standardized data of a full spectrum of aspects

153

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan concerning bariatric surgery, medicine, diet, exercise, psychology, biochemistry and anthropometry from three public hospitals before and after bariatric surgical procedures; in addition to mortality and pre‐operative socio‐demographics. This study presents the first ever in‐ depth longitudinal report of a publicly funded bariatric surgery service and research comprised of three specialist obesity clinics in the Australian state of NSW. Not mentioning the well‐ established publicly funded bariatric surgery services in the Australian state of Victoria, which mostly perform LAGB procedures (330, 331), there are only a few public hospitals with the staff resources and physical infrastructure required for bariatric surgery (16). Hence, there are limited published reports on publicly funded bariatric services with which could facilitate comparisons with our results. For instance, Canberra Obesity Service Management (COSM) – the only publicly funded obesity service in Canberra, Australian Capital Territory (ACT) – had no publicly funded bariatric surgery patients upon their publication (234). This was due to the fact that the Canberra team had just started a publicly funded bariatric surgery service like other similar services at the end of 2017 and there were only 15 COSM patients who had recently undergone publicly funded bariatric surgery as of February 2019 (234). As one of the first comprehensive reports encompassing a full spectrum of aspects regarding bariatric surgery within multidisciplinary clinical obesity services across a wide range of up‐to‐date bariatric surgical procedures, we hope it can serve as a foundation for effective and safe publicly funded bariatric surgery; monitoring of long‐term patient outcomes; standardized clinical data recording; and future comparisons of multiple publicly funded bariatric surgery services.

In this chapter, we address several knowledge deficiencies regarding long‐term weight change, health outcomes and surgical complications following currently‐performed bariatric surgeries with in‐depth data from a multisite publicly funded bariatric surgery service. There are ongoing concerns remain regarding the balance of the benefits and harms of bariatric surgery in patients with the highest level of obesity, namely, super obesity (SO). Thus, we also expand on the current literature on this topic by comparing weight loss, peri‐ and post‐surgical outcomes, and rates of obesity‐related comorbidity remission between SO patients and those with BMI <50 Kg/m2

154

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

(morbid obesity, MO), to formulate recommendations regarding the surgical care of SO patients.

Weight loss outcomes

In this longitudinal cohort study observing 168 publicly funded bariatric surgery service patients, significant weight loss was found over the long‐term follow‐up period at all timepoints. Besides weight loss, our results also show durability of weight loss in the majority of patients across all timepoints. Weight was maintained at at‐least 15 to 20 %TWL for 6 years following bariatric surgery, although there was an acceptable degree of weight regain in the long‐term. Similar weight loss patterns were also observed in a 10‐year study among Taiwanese patients who presented for LSG (332), a 5‐year Finnish Sleeve vs Bypass (SLEEVEPASS) randomized controlled trial (RCT) (18), the 3‐ and 7‐year LABS studies (13, 14), and the 6‐year Utah Obesity study (161). These patterns involved an overall maximum weight loss at 1 to 2 years after bariatric surgery followed by a slight weight regain thereafter. The findings are also comparable to the 1‐ and 5‐ year post‐bariatric surgery weight loss outcomes reported in a retrospective observational cohort study of 65,093 patients with severe obesity, the PCORnet (333). This study reported an estimated mean 25.2 %TWL at 1 year (95% CI=25.1–25.4%) and 18.8 %TWL at 5 years (95% CI=18.0–19.6%) following sleeve gastrectomy (SG).

Several mechanisms of post‐operative weight relapse have been proposed. They were associated with mental health (depression, binge eating disorder, emotional eating, grazing, sweet cravings), maladaptive lifestyle behaviours (adherence to exercise and/or dietary recommendation), inadequate follow‐up support, hormonal/metabolic imbalance (increased ghrelin levels), and anatomical/surgical factors (initial sleeve size, sleeve dilatation, amount of gastrointestinal track bypassed) (334‐336). Despite ongoing annual clinic follow‐up visits, modest weight regains were observed at post‐surgical year 2 onwards in our study cohort. This suggests a more complex interaction between weight regain and follow‐up requiring further investigation, and highlights the need for clinical series to adequately report the follow‐up support provided. The details of

155

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan our follow‐up care are extensively discussed in CHAPTER 3 (Adherence to follow‐up research theme).

Sleeve gastrectomy (SG) is a popular first‐stage procedure that is often and increasingly used as a stand‐alone procedure due to its simple technique, rapid performance and potentially greater safety compared to other procedures (337, 338), and beneficial effects in inducing substantial weight loss and remission of comorbidities (22, 318, 339, 340). Therefore, our surgical team introduced SG as the primary procedure for the majority of the study patients after the commencement of the publicly funded bariatric surgery service. At an earlier time, the percentage excess weight loss (%EWL) was a commonly reported outcome of bariatric surgery. However, since the time this research was designed in recent years, standard outcome reporting was adopted using the percentage weight loss (%TWL) relative to baseline body weight, being the preference nowadays (326). Furthermore, this measure is also the simplest to understand and has broader acceptability than %EWL, making it a practical assessment metric that is easily translatable into clinical practice. Thus, besides BMI loss and absolute weight change, the %TWL metric was chosen and presented in our study to facilitate the comparison of the findings with those from other studies. It is also important to note that given that most of our study patients had class III obesity and even super obesity upon entry to the clinics, even after significant weight loss following bariatric surgery, they may never reach the unrealistic goal of an “ideal” or “normal” body weight (traditionally set at 25 Kg/m2). Thus, the %EWL metric used to define success and failure according to the difference of pre‐operative and ideal weights after bariatric surgery was not considered appropriate and was omitted from this report. In reality, the “ideal” BMI benchmark is rarely achieved with any weight loss method (266), including the irrefutably powerful bariatric surgery. On top of that, all our patients were coupled with at least one major obesity‐related condition, with most having multiple illnesses. Hence, an improvement in classification of the EOSS stages may be a more meaningful achievement to the patients than normal body weight in the sense of the long‐term metabolic benefits, functionality and well‐ being already ascertained at this stage.

156

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

Obesity‐related comorbidities

The Edmonton Obesity Staging System (EOSS) In light of the nature of our tertiary care hospitals also being clinical obesity services specialising in the management of highly complex patients, it is not surprising that all the study patients with clinically severe obesity were at ≥ EOSS stage 2; i.e., a more severe pattern of obesity and metabolic diseases. It does support the fact that clinically severe obese patients requiring bariatric surgery in the public healthcare system are seriously affected by their obesity and its complications, as can also be seen in our extremely strict intake eligibility for bariatric surgery described earlier. Retrospective application of the EOSS staging tool to the American National Health and Nutrition Examination Survey’s (NHANES) data proved that patients in stages 2–4 of EOSS have higher all‐cause mortality compared to the lower stages of 0 or 1 (341) if not treated appropriately. Clearly, determining access to clinical obesity services or treatment intensity based on weight or BMI alone will likely neglect a considerable proportion of patients who would otherwise benefit from specialized clinical obesity care, such as publicly funded bariatric surgery services. Besides that, it could also lead to overtreatment in low‐risk patients (EOSS stages 0‐2, with no or few obesity complications) with high BMIs or undertreatment of high‐risk patients (EOSS stages 3‐4 with more health impairments) with lower weights or BMIs. Correspondingly, preventing patients from progressing through the higher EOSS stages in addition to BMIs has been our key clinical objective, rather than weight loss alone. We have been considering wider use of EOSS in clinical practice and health services planning to potentially better target early access to effective medical and surgical treatments of obesity to those patients who stand to benefit the most. Particularly, there is an ongoing discussion about the limited access to bariatric surgery in many public healthcare systems worldwide, including the current multidisciplinary publicly funded bariatric surgery service in Sydney, Australia. As a key efficient planning resource, we postulated that obesity classification systems such as the EOSS could be useful in assessing patients’ health status according to their disease severity. It provides an alternative to simple reliance on BMI‐defined criteria to support waiting‐list prioritization of publicly funded surgical treatment.

157

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

In the present study, we observed significant decreases in the prevalence of all studied comorbidities, except hyperlipidaemia and depression and/or severe anxiety. Our patients were found to have a median of 6 obesity‐related comorbidities. We analysed and discussed changes in the statuses of the major metabolic comorbidities in this study. The remission and improvement of co‐existing obesity‐related comorbidities have always been desirable outcomes of obesity management and, in most cases, may be the main reasons for treating clinically severe obese patients. The beneficial effects of bariatric surgery are well‐known and the present study adds knowledge about the totality of long‐term surgical and metabolic outcomes in a sample of highly complex patients who underwent publicly funded specialised obesity treatment.

Type 2 diabetes mellitus (T2DM)/Diabesity Diabesity poses individual and global health challenges on an unprecedented scale. While bariatric surgery is recognised as a potent therapeutic option in patients with diabesity, different gradients of efficiency have been reported among different studies (14, 18, 19, 41, 151, 192, 274‐ 279). In the current study, bariatric surgery appeared more successful in managing T2DM and hypertension than hyperlipidaemia, as measured by a combination of medications, changes in blood test results, changes in clinical measurements and physicians’ diagnoses. Despite a greater severity of body weight and metabolic comorbidities, tremendous T2DM remission and improvements rates were observed in the study patients. The incidence of T2DM cases following bariatric procedures was negligible low at all follow‐up assessment timepoints, with only one new onset observed among our study population. The underlying mechanisms of T2DM remission and improvement after bariatric surgery involve the combined effects of incretin hormone secretion, bile acid metabolism, intestinal physiology, lipid regulation, neuronal signalling, microbiome changes, and weight loss (342‐345).

Despite our cohort study being focused on high‐severity patients (EOSS stages ≥ 2) with heavy weights, multiple metabolic diseases and an average baseline HbA1c ≥7%, the strong association between bariatric surgery and T2DM remission are promising and comparable to the results of other longitudinal studies. This includes the 6.5‐year Norwegian Prescription Database cohort

158

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan study of clinically severe obese patients (57.5% absolute risk for remission in bariatric surgical group, based on drugs dispensed), and LABS study of severe obese patients (BMI ranged 34‐81 Kg/m2, with or without a comorbidity) (remission rates: 67.5% RYGB and 28.6% LAGB at 3 years). The remission rate of T2DM in the current study was also comparable to that of the nationwide Scandinavian Obesity Surgery Registry (SOReg) study (346) that also observed both glucose lowering medication and glycaemic control endpoints, although was of shorter duration than our study. This registry‐based study from Sweden covered 8,546 adults with diabesity who underwent bariatric surgery. Some – 58.2% and 46.6% achieved T2DM remission at 2 and 5 years after bariatric surgery, respectively (346). Our findings are also similar to those reported by another registry, the United Kingdom National Bariatric Surgery Registry with 50,782 entries for operations between 2000 and 2015 (297). This reported particular effectiveness on T2DM (around 50% remission over 5 years); however, no standardized clinical definitions or laboratory ranges were used to define remission. Instead, remission was dependent on the judgement of the clinicians’ submitting data to the registry, thus the evidence was considered less convincing.

Amongst other landmark studies, the prospective matched SOS cohort study (192) revealed that surgically‐treated patients had excellent T2DM remission rate at 2 years post‐bariatric surgery (72.3%, 219 of 303 patients); however, the rate relapsed to 30.4% at 15 years post‐surgery (35 of 115 patients). A similar outcome in terms of changes of T2DM statuses was also reported in the prospective Swiss Multicentre Bypass or Sleeve Study (SM‐BOSS) RCT (217 Swiss patients with clinically severe obesity, mean BMI=43.9 Kg/m2) (19), which reported proportions achieving T2DM remission (SG: 61.5% vs RYGB: 67.9%), improvement (SG: 15.4% vs RYGB: 7.1%), unchanged (SG: 11.5% vs RYGB: 10.7%) and worsening (SG: 11.5% vs RYGB: 14.3%) at 5 years post‐operation. Our analyses of changes in T2DM statuses over 6 years of follow‐up also corroborated the findings of a retrospective study of 74 Indiana patients who underwent LSG, which also reported a significant cumulative remission and improvement rate of 77% over the entire follow‐up period (340). The proportion achieving T2DM remission following bariatric surgery in the current study was in line with the Longitudinal Assessment of Bariatric Surgery (LABS) study, which documented an observed T2DM remission rates at 4, 5 and 7 years of 63.7%,

159

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

61.4% and 58.9%, respectively, for RYGB patients and 33.0%, 26.3% and 24.0%, respectively, for LAGB patients (14). Similar to other reports including this LABS study (14), there was little incidence of T2DM after bariatric surgery (<1.5% for RYGB and LAGB unreported), and our incidence rate of T2DM was only 0.6% by 6 years of timepoints.

Regarding the cumulative T2DM remission rate, our study findings were also almost identical to those of a retrospective observational study of 110 obese patients (BMI ≥35 Kg/m2) with T2DM in Portugal (347) that reported a cumulative T2DM remission rate of a maximum of 57.3% at year‐5 post‐operation, except that the authors considered patients who had ever attained a remission as cumulative remission. Whereas in our studies, we only allowed patients with cumulative remission and improvements of T2DM following bariatric surgery towards their last observations to be accounted in the accumulative rates.

Regardless of the high severity of our patients’ obesity, the present findings demonstrated the significant ability of bariatric surgery in improving glucose homeostasis and induce T2DM remission. It should also be underlined that the study patients exhibited remarkably low worsening and incidence rates of T2DM despite the long follow‐up period. Consequently, we support the already established position that bariatric surgery can serve as both the most effective treatment tool and an illuminating scientific model with which to address the diabesity crisis in the publicly funded healthcare system in long‐term.

Hypertension Obesity is a major risk factor for cardiovascular disease (CVD) and mortality, and contributes to the pathogenesis of cardiovascular risk factors including hypertension (348). In this study, bariatric surgery conferred not only a marked weight loss up to a modelled mean of nearly 25 %TWL and significant concurrent remission and improvement of T2DM, it also conferred substantial remission and improvement in hypertension. The observed decrease in medication use and blood pressure over the 6 years of follow‐up period reflects the substantial impact of bariatric surgery. The findings are consistent with numerous previous studies among patients

160

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan who have undergone bariatric surgery. In the same line with a long‐term registry‐based cohort study (median: 6.5 years, range: 0.2‐10.1 years) among patients with clinically severe obesity that based on drug dispensation in the Norwegian Prescription Database published by Jakobsen and co‐workers (41), remission of hypertension was observed in 31.9% of the surgically‐treated group (n=932). However, it is important to point out that using dispensed drugs as proxy outcomes such as this study might have underestimated comorbidity rates, therefore should be interpreted with caution. In this respect, our results are also in line with a systematic review of 29 studies that included 7,971 patients with more than 2 years of post‐surgery follow‐up (283), which found a hypertension remission rate of 38.2% after RYGB using the same definitions as ours. Although when compared to another systematic review (14 studies included, 3,550 subjects, mean follow‐ up 5.35 years) that examined the efficacy of LSG on hypertension among patients with obesity (mean pre‐operative BMI=47.7±8.8 kg/m2) (280), the remission rate is much higher (62.2%, range 0–100%) and the improvement rate is similar to our findings (35.7%, range 13.3–76.9%). The authors also highlighted significant heterogeneity in the definitions of hypertension, resolution and improvement rates of hypertension, lack of RCTs, and majority of the included studies being retrospective cohort studies that made meta‐analysis unfeasible. Therefore, our comparison with other studies of similar groups of patients and methodologies, could provide in‐depth insight of the effectiveness of bariatric surgery on hypertension in patients with clinically severe obesity. With the lack of RCTs in bariatric surgery field, this is not uncommon, given that RCTs of surgical research are not always feasible nor considered ethically appropriate. Even when funding is not an impediment, patients’ preference, rarity of some medical conditions, and lack of equipoise among surgeons were proved to be principal precluding problems in performing RCTs of bariatric surgery (349, 350). Observational studies such as cohort studies have been the most appropriate alternative study design when a RCT is not feasible.

Regarding the high cumulative remission and improvement hypertension rate following bariatric surgery, similar results were also seen in an Indiana study by Eid and co‐workers. They reported a cumulative remission and improvement rates of 74% in hypertensive patients who underwent

161

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

LSG (340). This is in agreement with our findings that 70% of patients experienced major improvement or remission of hypertension over a 6‐year follow‐up period.

As was found in other only landmark studies with the full changes in statuses like our studies, the SM‐BOSS RCT (19) showed that their study patients possessed higher rates of remitted, unchanged and worsened hypertension at 5 years post‐operation than our study patients, but not the improvement rate. The differences were likely due to the initial disease severity of the study cohorts, whereby all our study cohort had complex clinically severe obesity. Comparison with another landmark study, the SLEEVEPASS trial (18) was unfortunately tempered by their reliance on antihypertensive medication use only as an indicator of hypertension, with no blood pressure measurements taken; in addition to not having the complete detailed changes in hypertension statuses. Hence, this does not conceivably facilitate comparison with our findings.

Hyperlipidaemia In this part of the study, we closely detailed progressive changes in hyperlipidaemia status (remission, improvement, persistence and worsening), by combining lipid‐lowering medication use with all the four important clinical serum lipid measures (total cholesterol, triglycerides, HDL‐ C and LDL‐C), in order to address meaningful knowledge deficiencies in the effectiveness of bariatric surgery on lipid profiles. In other terms, it was four times stricter to call a hyperlipidaemia remission in the present study compared to studies that only assessed lipid subfractions separately (298, 299). This has also, thus, generated superiority of persistent hyperlipidaemia, i.e., among over half of the study cohort, whereas one‐third of the study patients attained either remission or improvement of hyperlipidaemia over the follow‐up duration.

Adipose tissue, or body fat, is a lipid storage, as well as an endocrinologically and immunologically active organ that can be involved in the pathogenic mechanisms underlying hyperlipidaemia. Adipose tissue dysfunction is caused by obesity‐induced stress (351), whilst bariatric surgery improves its function via a significant reduction of fat mass. In spite of this, the relationship

162

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan between weight loss and the lipid profile remains inconclusive. It has been discussed that the improvement in the levels of HDL‐C and triglycerides may be associated with weight loss (352), while LDL‐C and total cholesterol might not. This theory exposes the mechanism of malabsorptive procedures that might tend to favour the drops in total cholesterol and LDL‐C levels due to decreases of their absorption, but not by SG or other restrictive procedures that seen in the majority of our population, i.e., a total of 83.9% of our sample, which explains the plausibility of our findings regarding increased HDL‐C and lower triglycerides over the timepoints, but not the others. Our data support other previous studies that considered lipid variables but lacked lipid‐ lowering regimen data and only reported short‐term results. One indicated their observation at 1‐year post‐surgery that the mean HDL‐C levels improved post‐SG procedure, but not in their patients presenting for RYGB (295). Similarly, the greatest normalization of HDL‐C was seen in patients after SG compared with those post‐RYGB. Contrarily, patients who underwent RYGB were able to reduce their mean total cholesterol and LDL‐C levels to below the abnormal threshold, and a higher proportion of RYGB patients in the normalization of total cholesterol and LDL‐C (76% and 74.2%, respectively), which exceeds that achieved after SG (43.5% total cholesterol and 46.9% LDL‐C) (295). The small change in measured total cholesterol and LDL‐C levels observed in the present study was also likely plateaued by their use of lipid‐lowering agents. Low rates of cessation of treatment in the current cohort over the study period, despite good responses in lipid parameters were likely due to guideline recommendations for the patients’ use in cardiovascular risk reduction, particularly in individuals with other cardiometabolic risk factors (353).

As explained earlier, it is not possible to directly compare the results of the present study with the literature for several reasons. Apart from the different cut‐offs used and follow‐up durations, the definitions used for hyperlipidaemia or dyslipidaemia vary widely. Some studies primarily evaluated medications alone (41, 293, 294); others assessed some of the lipid parameters with no medication appraised at all (14, 295); some investigated lipid‐lowering agents and lipid profiles separately without combining the outcomes as a true whole representation of hyperlipidaemia status (294); some did not specified definition at all/patient self‐reported (296,

163

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

297); and a couple of studies assessed medications but separated the lipid subfractions (e.g. hypercholesterolemia or hypertriglyceridemia) without evaluation of the whole lipid profile (298, 299). Only a few studies have reported hyperlipidaemia statuses, including remission, based on both lipid‐lowering medications and lipid profiles, even rarer, for all four fundamental lipid components to have been studied. The only longitudinal study that followed the same line as our detailed study is the 5‐year SM‐BOSS trial of 217 patients with clinically severe obesity (SG: 107 patients vs RYGB: 110 patients) (19), with the exception that total cholesterol and lipid‐lowering agents other than statins were not taken into consideration for the assessment of post‐operative course. This difference potentially explains the contrasts between the findings of our population and this publication that reported higher remission and improvement of dyslipidaemia after SG procedures (remission: 42.6%, improvement: 41.2%, unchanged: 16.2% and worsened: 0%) and RYGB procedures (remission: 62.3%, improvement: 30.2%, unchanged: 7.5% and worsened: 0%), respectively. Therefore, it may not necessarily identify more events of hyperlipidaemia remission and improvement if the other studies have incorporated the strict definitions that we used. Notably, the SLEEVEPASS randomized trial that had a protocol very similar to the SM‐BOSS study however only reported changes in lipid levels separately. True remission of dyslipidaemia was defined based only on LDL‐C levels and medications which, again does not facilitate precise comparisons with our findings.

OA and/or WBJP, TJA and prescribed opioids Obesity is a staggering epidemic and a modifiable risk factor for osteoarthritis (OA) and weight‐ bearing joint pain (WBJP) (301). The clinical magnitude of the improvements in OA and/or WBJP in our study patients, from a pre‐operative 71.4% to 56.1% following surgical weight loss over 6 follow‐up years has provided important confirmatory data to the literature. A 5‐year observational study of 13 Pennsylvania patients with symptoms and radiographic evidence of knee OA who underwent bariatric surgery found that patients maintained improvements in knee symptoms, pain and daily living activities for 5 years following surgical weight loss; with only one patient having knee replacement surgery (7.7%) (180). The Western Ontario and McMaster Universities Index of Osteoarthritis and Knee Osteoarthritis Outcome Score Surveys were

164

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan administered in this study. Not mentioning the distinctive sample size of this underpowered study (n=13), their results were similar to our findings at year‐5 post‐surgery that also noted maintenance of impressive improvements in OA and WBJP symptoms, as well as mean percentage of TWL of 26.4%, in addition to very small number of patients performed total joint arthroplasty (TJA) (2 of 168 patients, 4.9%).

There is a dearth of research studying opioid use beyond the first year after bariatric surgery. The phenomenon observed in our study is in the same direction as that of the LABS‐2 study, which also determined changes in the short‐ and long‐term use of prescribed opioid analgesics following bariatric surgery. Like our study, the 7‐year LABS‐2 observational cohort study (354) (median BMI=46 Kg/m2), which also studied changes in prescribed analgesics post‐surgery, reported an increased use of opioid medication following bariatric surgery, from 14.7% at pre‐ operative baseline increased to 20.3%, i.e., above baseline levels as time progressed at year 7. However, contrary to their proposed tandem that additional surgeries (such as back, knee, hip or ankle surgery, or a subsequent bariatric surgical procedure) were related to increased risk of initiating opioid use following bariatric surgery; we found a declined TJA performed throughout the 6 follow‐up years including revisional surgeries in the current investigation. As a matter of fact, this study sheds further light on the beneficial effects of bariatric surgery on TJA other than symptomatic relief of OA and WBJP in this carefully‐studied cohort. In contrast to our accurate medication data captured from hospital records and our TJA data obtained from the national AOANJRR registry, the LABS‐2 study has noted their reliance on self‐reported opioid analgesic medication use, and that the accuracy in this sample was unmeasured. Despite this limitation, this study (354) and our findings are distinct from other studies that have reported relatively short timeframes to examine the use of prescribed opioid medications after bariatric surgery (300, 355‐357).

There is an increased trend in the opioid use of our bariatric surgical cohort post‐surgery across nearly all timepoints, particularly year 5 (19.6%) and year 6 (17.9%). Majority of the new opioid users did not suffer a surgical complication. Furthermore, most of the post‐operative

165

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan complications following bariatric surgery do not cause significant pain [e.g. gastroesophageal reflux disease (GORD), bleeding]. Therefore, post‐operative complications are unlikely to account for the development of new opioid use in our study cohort. Instead, our detailed evaluation of indication for opioid prescriptions showed that the opioid consumptions were associated with patients’ significant pains caused by various deteriorated physiological and psychological outcomes. It is understandable that these pains should not be disregarded nor undertreated. On the other hand, bariatric surgical population may be more vulnerable to misuse and dependence than the general population, given the risk of “addiction‐transfer” (i.e. exchange of one compulsive behaviour for another) (358) that may confer a similar misuse of opioids after surgery (329, 359). The full clinical significance of this aspect and its direct causal effect in bariatric surgical population has not been explored in the current work, and may warrant future studies. However, based on patients’ positive motivations for pursuing bariatric surgery and risks associated with opioid abuse after surgery, we suggest healthcare providers to develop surveillance programs to monitor patients’ post‐operative opioid use, to maximise benefits from bariatric surgery. This could be helpful to detect patients at serious risk for misuse of opioids, promptly counsel patients on the risks, and identify those at higher risk of poor surgical outcomes. Surgeons may also minimise patients’ exposure to the risks by following the current opioid prescribing guidelines, while actively reduce excess prescription of opioids.

With much research demonstrating the benefits of weight loss to OA, the problem still arises in daily surgical practice as to how best combat this issue. Besides hospital policies that prevent orthopaedic surgeons from performing TJA on some patients with severe obesity, the latter did technically limit the ability to perform TJA (360). We believe that more research in these areas should also be conducted to determine optimal interventions for patients with both clinically severe obesity and OA symptoms, translating into practice to help those in need. All said and done, consideration should be given to referring patients to multidisciplinary services that comprise physicians, orthopaedic surgeons and bariatric surgeons, among other healthcare providers, to obtain the most comprehensive care possible.

166

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

Sleep‐disordered breathing: Obstructive sleep apnoea (OSA) and obesity hypoventilation syndrome (OHS) Our results suggest that there were long‐term decreases in the prevalence of obstructive sleep apnoea (OSA) and obesity hypoventilation syndrome (OHS) after bariatric surgery. The prevalence of use of continuous positive airway pressure (CPAP) or bilevel positive airway pressure (BiPAP) devices required by the study patients decreased tremendously from 51.8% following bariatric surgery and on a slight upward trend across year 2, then continued to downtrend again year 3 onwards throughout the follow‐up period to 31.6%. Aside from this, a total of 32 patients with OSA that were prescribed CPAP no longer required CPAP to their last observations following bariatric surgery, which is equivalent to a cumulative improvement rate of 36.8%. There was only one new onset of diagnosed mild OSA during the post‐operative follow‐ up period, with no CPAP required. Several other studies have also demonstrated the effectiveness of bariatric surgery in reducing OSA severity (306, 361). In a study of 132 clinically severe obese patients with OSA who underwent RYGB, the prevalence of OSA decreased from 71% at baseline to 44% at 1 year after bariatric surgery (p<0.001) (306). The authors also determined that OSA resolved in 45% of patients and improved in another 33% of the study patients (306). These different findings could potentially be explained by the different definitions of OSA and its remission/improvement that were used, or by different types of surgery, severities of patients, as well as other differences in the nature of follow‐up. Regardless of the controversy, bariatric surgery clearly demonstrates its additive effect on improving OSA and OHS.

Depression and/or severe anxiety It is noteworthy that our study cohort displayed lower prevalence of depression and/or severe anxiety as well as the use of antidepressants and/or antianxiety agents in the first few years from pre‐surgery, but not years 5 and 6 post‐surgery in which the rate climbed marginally back to similar prevalence as baseline. This findings verifies that of the LABS Consortium substudy (196), which reported a similar trend post‐RYGB with respect to change in the prevalence of mental disorders over a 7 year follow‐up (n=173). Consistent with our findings, a prospective study of 190 patients with clinically severe obesity within the Toronto Western Hospital Bariatric Surgery Program, which examined psychiatric medication use 1 year after bariatric surgery, found 32.1%

167

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

(61/190) of patients took these medications before surgery and 26.3% (50/190) after surgery (308). However, further comparison was halted when this study had only 1 year of available follow‐up data to date and psychiatric diagnoses were lacking. Short‐term significant and the largest reductions in depression and anxiety within the first two years of bariatric surgery were also seen in the studies included in a recent systematic review (14 prospective studies of participants with severe obesity, i.e. BMI ≥35 Kg/m2) (194). While this improvement may hold true, depression severity did increase after 2‐3 year mark post‐surgery as shown in a meta‐ analysis of 58 studies (189). Our longitudinal design helped confirm the answer that bariatric surgery shows a short‐term but not long‐term improvements in depression and anxiety.

We recommend all bariatric surgical candidates to receive pre‐operative education and psychiatric evaluation as to the possibility of mental illness, such as depression and severe anxiety. This high‐risk population should also be offered education on the symptoms of depression, and concrete steps to follow to get help and support if patients notice they are experiencing any of the depressive symptoms. Further studies should seek to validate these findings and confirm the long‐term impacts of bariatric surgery on depression and/or severe anxiety in the clinically severe obese population. Bariatric surgery is a life‐saving and ‐changing treatment. Pre‐operative fantasy expectations about future weight reduction, environment or social acceptability, and obesity treatments that create storm for serotonin depletion that leads to anxiety and depressive symptoms, anger, fatigue, and irritability are some of the important areas that need further long‐ term investigation in the setting of clinically severe obesity. As our additional findings also revealed that there were other issues impacting on mental health of this group of patients in the post‐surgical period, such as post‐traumatic stress disorder (PTSD), emotional eating and binge eating disorder; more in‐depth evaluation in these areas should be performed in the bariatric surgical population in the future.

Hyperuricaemia Hyperuricemia, or increased serum uric acid levels, raises the risk of developing the debilitating chronic condition − gout (362), although majority of individuals with hyperuricaemia do not suffer

168

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan from gout. Gout is both an inflammatory and metabolic disease, characterized by high serum uric acid levels, resulting in the deposition of monosodium urate crystals in the joints, which then triggers an inflammatory response with tissue damage and associated consequential functional impairment (363‐367). With obesity being the most common risk factor for gout development (368), the American College of Rheumatology (369) and the European League Against Rheumatism’s guidelines (370) recommend weight loss as part of gout management in patients with obesity. Concurring with our study, reductions in hyperuricaemia following bariatric surgery were also witnessed in another few studies (327, 371, 372), which reported a general fall in the incidence of hyperuricemia over time, ranging from 2% to 9.5% post‐operatively. For instance, Golomb and co‐authors showed a steady decline from 6% to 5% to 3% at the 12th, 24th and 36th post‐operative marks, respectively (372). The study findings prove the impact of bariatric surgery on serum uric acid levels. The impact of post‐bariatric surgical serum uric acid levels on gout and acute gouty arthritis in clinically severe obese population warrants future attention.

Peri‐ and post‐operative complications This study has, however, revealed a high incidence of various surgical and gastro‐intestinal complications across the surgery types, likely founded by our patients already being very severe at entry to clinic and at pre‐operative baseline. Most occurring complications in this study with majority of the population underwent LSG were gastroesophageal reflux disease (GORD) or gastric reflux. The worsening or de novo GORD are not completely a new issue for LSG in accordance with previous studies (19, 318, 340). As an illustration of this point, the SM‐BOSS study, a 2‐group randomized trial by Peterli and team (19) also found that GORD worsened (defined by more symptoms or increase in therapy) after SG more frequently than after RYGB (31.8% vs 6.3%). Many patients with clinically severe obesity experienced intermittent GORD prior to bariatric surgery, which can be exacerbated after LSG. Likewise, GORD‐naïve patients at baseline more often report de novo GORD symptoms after SG than after RYGB, as was also observed in the 5‐year SM‐BOSS study (19). In most cases, GORD symptoms can be treated conservatively with proton pump inhibitors (PPIs). In the major complicated cases where

169

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan pharmaceutical treatment was insufficient, their severe GORD had to be appropriately managed by introduction of the 2‐stage operative approach with RYGB (which, in our study, 1.4% of patients, n=2), or hiatal hernias been repaired (2.1%, n=3).

Just as with any other weight loss modalities, a hundred percent of efficacy with negligible adverse events is impossible. Instead of labelling the operative complications as “surgery failures”, the risk‐benefit balance for bariatric surgery is clearly still in favour of performing the surgical procedures in most cases, on the beyond significant lasting weight loss witnessed among the patients, which bariatric surgery also effectively solved multiple obesity‐related comorbidities. A more practical goal to help both the bariatric surgeons and patients may be to weigh the precise short‐ and long‐term pitfalls, expected benefits and surgical techniques when deciding whether bariatric surgery is the best option, and, if so, which procedure is the most adequate and preferable to the individual. Based on our results of 9‐year post‐operative study observation period, the high surgical and gastro‐intestinal complication rates could also be a sign of a strong element of our study, as special emphasis was given to the thorough recording of all complications. These findings provide a true reflection of the adverse effects of bariatric surgery, which do exist in many patients and many bariatric centres. They may also serve as a source of knowledge for future bariatric surgeons, healthcare providers, researchers, stakeholders and the public.

Nutrient deficiencies Despite bariatric surgery’s health benefits in weight loss and the remission and improvement of comorbidities, it has some potential nutritional complications. Bariatric surgery can exacerbate pre‐existing micronutrient deficiencies in patients with obesity. Considering these scenarios, we assessed the nutritional status of the study patients pre‐ and post‐operatively over a long duration. On the whole, the mean levels of serum vitamins and a trace mineral fell within the normal ranges before and after bariatric surgery, however, a high prevalence of vitamin D and iron deficiencies were evident, even prior to surgery. The occurrence of micronutrient deficiency before bariatric surgery is well‐established, with a couple of mechanisms contributing to

170

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan malnutrition in the obese patients. Firstly, obesity itself predisposes patients to expanded adipose tissue, greater anatomical distribution of fat deposition, and chronic low‐grade systemic inflammation. These alterations lead to accelerated oxidative stress, altered nutrient transporters and increased antioxidant nutrient utilization (313). Secondly, poor diet quality is often more noticeable in obese patients, which leads to a higher intake of processed foods that are low in micronutrients (313, 373).

Vitamin D deficiency Deficiency in vitamin D (calciferol, a fat‐soluble vitamin) before bariatric surgery is the most consistent pre‐operative micronutrient deficiency (313), particularly in the patients with severe and complex obesity. The findings of our study are comparable to ample existing evidence. Some examples are data from a cross‐sectional study assessing the pre‐operative baseline nutritional status of Kuwaiti bariatric surgery candidates. It found that vitamin D deficiency was most prevalent among their patients before surgery (374). It was determined that as many as 76% of bariatric surgery candidates (of 610 patients with vitamin D data available) were vitamin‐D deficient at baseline. The results of the current study are also concordant with another short‐ term retrospective study also conducted in Sydney and was published over 10 years ago (311). Likewise, the authors identified a total pre‐operative prevalence of vitamin D deficiency of 57% (24 of 42 patients), which improved slightly 1 year post‐operatively [RYGB: 30% (9 of 30 patients), SG: 43% (3 of 7 patients), total prevalence unreported, and no post‐operative AGB data]. However, it is important to point out that the patients in this study had class I obesity with unknown obesity‐related comorbidity, hence findings should be compared with caution. In our study, the total prevalence of vitamin D deficiency of the clinically severe obese population steadily declined each year following bariatric surgery, from a pre‐operative rate of 31.9% (43 of 135 patients) to post‐operatively rates of 18.1% (at year 1), 21.7% (at year 2), 18.6% (at year 3), 14.0% (at year 4), 7.3% (at year 5) and 12.9% (at year 6). An additional possible explanation specifically linked with the deficiency of the “sunshine” vitamin, vitamin D pre‐surgery in our study cohort is sunlight under‐exposure and hepatic hydroxylation (313) due to limited outdoor activities as a result of heavy weights, given that all our patients had class II or III obesity.

171

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

Alarmingly, pre‐operative vitamin D deficiency often carried forward as post‐operative deficiency. These are especially where the bariatric surgical procedure is likely to result in clinically significant vitamin D deficiency, osteoporosis, and fractures. Therefore, assessment and treatment of vitamin D insufficiency or deficiency must be done pre‐operatively, as recommended by evidence‐based clinical practice guidelines (375, 376). In this study, the trend towards improvement in vitamin D status post‐surgery may be related to the consumption of prescribed vitamin D supplement regimens.

Iron deficiency On the other hand, iron deficiency, which commonly progresses to anaemia (377), has been reported to be the most common specific nutrient deficiency occurring after bariatric surgery (310, 313, 378), as it was in our study. In the present study, not only did we measure the serum ferritin level (indicator of depleted iron stores), we also measured it in combination with the percent trasferrin saturation marker to accurately diagnose iron deficiency anaemia. The latter may be especially helpful where inflammatory conditions preclude the adoption of serum ferritin alone (379). On top of that, a retrospective series, such as the assessment of an administrative database of 18,783 Ontario adults who underwent bariatric surgery in the multicentre publicly funded Ontario Bariatric Network (310) also demonstrated that the most common severe nutritional complication requiring hospital admission was iron deficiency, occuring in 190 patients (1%). Several factors are known to contribute to iron deficiency after bariatric surgery: (i) Patients experience intolerance or aversion to meat, limiting heme intake (115, 313). (ii) SG (a restrictive procedure), or the restrictive component of RYGB (a mixed restrictive and malabsorptive procedure) reduces the gastric acid production needed for iron conversion, while the malabsorptive component reduces iron absorption (313, 380). (iii) The divalent cations of calcium and copper compete with iron for absorption which, in the event of their supplementation, may reduce iron uptake (115, 313).

172

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

Vitamin B12 deficiency

Cobalamin, or vitamin B12 deficiency has been reported to mostly affect patients undergoing RYGB (313, 373, 381, 382). This explains the findings in our study cohort, where only one patient underwent RYGB but the majority had SG as the primary surgical procedure. The mechanisms contributing to malnutrition following RYGB may be the reduction in gastric acid secretion, release of vitamin B12 from R‐binders, and availability of intrinsic factor (313, 381, 383), the details of which are beyond the scope of this discussion. The absorption of vitamin B12 can also be affected by SG, but hydrochloric acid production still occurs in the gastric pouch, as does intrinsic factor secretion by the parietal cells of the stomach upon stimulation by foods (382).

Unlike the glucose and lipid profiles, the nutritional parameters were unfortunately not routine blood tests in our hospitals until recent years. Missing data were also existed for some of the patients who were loss to follow‐up or attended our clinic appointments irregularly, which we address in the next chapter (CHAPTER 3). Besides utmost effort being made to contact these patients, we also attempted to order blood tests comprised nutritional parameters to be measured on their blood request forms. This was in addition to the tracing of blood test results from not only hospital paper and electronic medical records, but also missing blood investigations (not available in the hospital records) from the laboratories via telephone calls, faxes, emails and online pathology service portal to collect blood data as complete as possible from all the participating study settings over the entire study observation period. In summary, the findings of this part of the study advocate routine evaluation, tailoring of nutritional supplementation (especially vitamin D and iron), and adherence monitoring before and after bariatric surgery. For dietitians, the acquisition of specific knowledge of this area, such as subsequent repletion regimens, is required to deliver appropriate and effective care to post‐bariatric surgery patients.

Mortality The mortality rate observed in our study is consistent with that of a prior retrospective cohort study (384) that reported 1.3% all‐cause mortality among 8,385 patients in Israel (43% SG, 40% AGB and 17% RYGB). This demonstrates that there are extremely low death rates among bariatric

173

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan surgical populations (<2%). Of note, all the causes of death in the study patients were deemed unrelated to the surgery and only occurred several years post‐operatively (i.e. 6, 8 and 9 years after bariatric surgery, respectively).

Comparison of super obesity (SO) and morbid obesity (MO) Many studies have discussed the weight outcomes between SO and MO groups; nevertheless, thus far, there are only a few studies that have been able to compare the comorbidity outcomes of the two BMI groups. To the best of our knowledge, this is the first study that has exclusively presented not only weight and safety outcomes, but also detailed changes in comorbidity status after bariatric surgery on adults in the MO and SO groups across a wide variety of bariatric surgery types in the long‐term. Given the scant attention that has been devoted to real‐world head‐to‐ head comparisons of effectiveness and safety between these two BMI groups, whether SO or non‐SO counterparts in a case‐control fashion in the same study, our findings could lead to better knowledge and patient care.

This study did not detect a statistically significant difference in weight loss measured as %TWL from pre‐operative baseline between the SO and MO groups. The post‐operative comorbidity outcomes of the two BMI groups were also remarkably similar, with pre‐operative BMI having no bearing on the SO and changes in post‐operative comorbidity courses. Our results are comparable to those of a study from New York of 930 patients with SO and who underwent either SG (n=732) or RYGB (n=198) in the sense of %TWL and the resolution rate of comorbidities (T2DM, hypertension, hyperlipidaemia and OSA), although this study only reached 1‐year of follow‐ up(321). This phenomenon was also observed in another recent comparative study of 186 Turkish patients with MO (50.9%, n=83), SO (31.9%, n=52) and super‐super obesity (SSO) (17.2%, n=28) who underwent LSG. The authors found similar weight losses (%TWL and %EWL) and rates of improvement or remission of comorbidities in all comparison groups in short‐term (up to 41.2 months), except the significantly lower %EWL for the SSO group (48.3%) at the end of follow‐up period (25). There are, however, a few shortcomings in this study, including their retrospective

174

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan analysis that based on a database from a single institution (hospital unknown) with a limited number of patients in each group.

We postulated that the equal outcomes observed between the two BMI groups were partly due to the fact that the two groups demonstrated similar %TWLs, which had been translated into remission or improvement of all of these obesity‐related comorbidities (321). With respect to safety, despite the SO group being fraught with more complex technical challenges, both BMI groups also encountered similar rates of surgical complications during and after bariatric surgery. Indeed, risk aversion is important. However, high‐risk patients, including those with extreme BMIs like the SO population, may also attain the greatest potential benefit from bariatric surgery. While both SO and MO groups demonstrated significant remission and amelioration of obesity‐ related comorbidities as well as equal safety profiles, very few studies have investigated these metrics closely, so there is little support for our claims. Together with the results reported here, these findings point to a need for future research on other populations regardless of publicly funded healthcare system or less severe study cohort to identify and characterize the outcomes between SO versus MO patients in a head‐to‐head fashion. But not zooming into one BMI group alone such as SO (22‐24, 271, 340, 385‐388) or SSO only with no other BMI group comparator (318, 389, 390), to define and justify the true benefits and risks involved with bariatric surgery in the often‐labelled “challenging” population.

Strengths This longitudinal study has several strengths. It is a multicentre study with observations of long‐ term follow‐up, hence maximising the capacity to provide conclusive results. Another advantage of this study is its investigation of real‐world long‐term evidence, despite retrospectively, we do not believe results from a prospectively conducted study will differ vastly, as has been reconfirmed by findings in the literature. Conversely, our full spectrum of findings contributes knowledge and values to the literature, as well as provides an example of a successful physician‐ led multidisciplinary approach. This report is also one of the few to focus on government‐funded bariatric surgery services that help disadvantaged patients with limited access to publicly funded

175

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan bariatric surgery, which is a known reality worldwide. The virtually high follow‐up rates of surgical complications, physicians’ diagnoses, dispensed medications for obesity‐related comorbidities, clinical measurements and mortality, as well as precisely defined definitions and calculations for each comorbidity, provide highly meaningful and accurate data for generalisation to bariatric surgery services in public healthcare systems. The inclusion of bariatric surgery as the top of hierarchy of the publicly funded obesity treatments represents an advancement in obesity management. There is still an ongoing debate regarding the best surgical procedures for maintaining weight loss and resolving associated comorbidities. However, this was not the main objective of our study and hence minimal discussion. We included patients who underwent four different bariatric surgical procedures, especially SG. This is an important advantage given that SG is gaining popularity nowadays; however, most previous longitudinal studies only reported on RYGB or other outdated procedures or on the short‐term effectiveness of SG. Importantly, as the data were collected directly from hospital records and surgical reports and entered into a standardized database by a single study investigator, it can be ruled out the underreporting bias regarding surgical complications or over‐reporting bias regarding the positive outcomes of bariatric surgery. Furthermore, the differences in baseline parameters such as age, sex and BMI among patients potentially resulting in bias were minimised by adjustments of these factors in the statistical modellings. Additionally, the study protocol did not include an upper limit for BMI. While we had a number of patients with a BMI ≥50 Kg/m2, we compared SO and MO patients in a case‐control manner. Hence, we were able to address the question of whether patients with extremely high BMIs get greater or lower benefits and risks from bariatric surgery, as compared to the less heavy MO patients.

Limitations The study is not without some limitations. Firstly, this study may be underpowered by its relatively small sample size related to the scarce resources of the public healthcare system and, hence, stringent selection of the population, which may affect the ability to draw robust conclusions. This is particularly the case as we have necessarily divided the study cohort into the BMI subgroups, SO and MO. Nonetheless, we were able to collect data from multiple resources

176

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan and means with minimal missing data, as well as substantial in‐depth variables in great detail. Additionally, we conducted modellings with repeated measures and inference tests up to post‐ surgical 6th‐year observations consisting of statistically adequate sample size and data, regardless of overall or subgroup analyses. The results show that we were able to find significant and meaningful patterns in long‐term changes in weight loss and obesity‐related comorbidities in the study cohort. Secondly, this study did not address some other outcomes such as dietary supplements, quality of life, cost‐effectiveness and measures of changes in lifestyle (e.g. exercise and dietary alterations) post‐surgery. Particularly the changes in lifestyle that were not quantified in this cohort, given that this is a multidisciplinary service that lifestyle changes often accompany bariatric surgery. We considered that this was not the primary focus of this study in addition to patients attended mostly yearly clinic reviews; we decided not to measure the effect of modification of lifestyle and behaviours after bariatric surgery due to impracticalities in our study designs and settings, and instead focused on bariatric surgery. Numerous studies have shown beneficial effects of lifestyle interventions in terms of cardiometabolic markers. Therefore, we consider that incorporation of the measures of diet and exercise modification following bariatric surgery may warrant future research and may be helpful in identifying factors that independently or synergistically result in benefits. Thirdly, only a minority of patients underwent MGB‐OAGB (n=15), AGB (n=11) and RYGB (n=1) as index procedures in this study. Therefore, it does not allow for comparison and firm conclusions between the surgical procedures. Even if we compared the four bariatric procedures considered in our study, the generalizability of the results might be compromised, based on the fact that our study was dominated by patients who underwent SG. Fourthly, this study included mostly Caucasian patients with clinically severe obesity, which limits generalizability of the results to populations of other races, or with less severe obesity, or individuals not seeking specialist obesity treatments. Fifthly, it is worth noting that, studies have also shown that pre‐operative duration of T2DM is related to remission of diabesity following bariatric surgery and its durability. Our exploration of this relationship was hindered by unavailability of the duration of T2DM data in the hospital records for all our patients. Addition of duration of T2DM or initial T2DM incidence as a routine medical history retrieval in all the

177

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan hospitals, whether in electronic medical records or clinic notes, would be beneficial for clinical care and research.

2.6 CONCLUSION

This study confirms that bariatric surgery performed in public hospitals using a multidisciplinary approach is effective, durable and safe for long‐term weight loss and the management of obesity‐ related comorbidities. Having said that, although there were not many surgical complications necessitating surgical or endoscopic revision within the 9 years of post‐operative follow‐up and benefit‐to‐risk ratio of bariatric surgery was decidedly in favour of the benefits, the risks of complications should be taken into account in decision‐making processes.

Osteoarthritis and/or weight‐bearing joint pain most heavily representing the obesity‐related comorbidities in the study cohort, and bariatric surgery most effectively improved the symptoms following marked weight loss at all the 6 annual follow‐up timepoints. Patients undergoing total joint arthroplasty declined over time with variability in patterns following bariatric surgical weight loss. There appeared to be a quadratic trend to sleep‐disordered breathing throughout 6 years of follow‐up period. One‐third of patients no longer required CPAP devices at their last observations following bariatric surgery. Among those with type 2 diabetes mellitus and hypertension at pre‐operative baseline, drastic composite cumulative rates of remission and improvement seen in majority of them. Hyperlipidaemia based on our strict definition of assessing all the four subfractions of lipid profile and medication use, generated a result of approximately half of the patients remained persisting hyperlipidaemia across all timepoints. The prevalence of patients having depression and/or severe anxiety was lower between 1 and 4 years post‐surgery, but surpassed baseline level at 5 and 6 years of follow‐up. Bariatric surgery is associated with decreased serum uric acid levels in the study cohort over 6 years of period.

178

CHAPTER 2 ‐ Weight trajectories, comorbidities, surgical complications and nutrients | Michelle M.C. Tan

Despite mean serum vitamins and trace minerals were at adequate levels before and after bariatric surgery, high prevalence of nutrient deficiencies was noted in the study cohort, mainly of vitamin D. Pre‐operatively, vitamin D deficiency was noted in one‐third of patients, and decreased significantly post operation. Iron deficiency anaemia doubled at year 6 post‐operation than that of pre‐surgery. As for vitamin B12 insufficiency, low prevalence was detected before and after bariatric surgery, with no patient developing a deficiency in years 5 and 6 post‐ operatively. Bariatric surgery appears to be equally safe and effective in the super obese and morbidly obese groups in achieving substantial weight loss, obesity‐related comorbidities, length of hospital stay and surgical complications.

179

CHAPTER 3 – Adherence to clinic follow‐up | Michelle M.C. Tan

CHAPTER 3

RESEARCH THEME 2

Adherence to multidisciplinary follow‐up care after bariatric surgery in an Australian publicly funded healthcare service: Predictors, reasons for loss to follow‐up and outcomes

3.1 ABSTRACT

Introduction Nonadherence to routine follow‐up after bariatric surgery is common despite recommendations for such. Review after bariatric surgery is necessary for management of obesity‐related comorbidities, treatment of post‐operative complications and monitoring of nutritional status, and may assist with long‐term weight maintenance. The aims of this study were to examine predictors of adherence to post‐bariatric surgery follow‐up, identify the reasons for ceasing attendance at clinical follow‐up reviews after surgery, and to assess the relationship between adherence to follow‐up and weight loss outcomes.

Methods A prospective longitudinal follow‐up study was conducted on 168 patients with clinically severe obesity who underwent bariatric surgery from 2009 to 2017 at a publicly funded service in New South Wales (NSW) in Australia. This service mandates multidisciplinary post‐operative follow‐ up care at 3, 6, 12, 18 and 24 months, and annually thereafter. Patients’ socioeconomic, psychosocial, anthropometric, medical and surgical data were collected from medical records, clinic reviews and patient interviews (phone or in person). Patients who attended for irregular review were contacted by telephone or mail, and a clinic visit was scheduled if the patient was willing to attend. If a patient was uncontactable or if responders were willing to complete our mixed‐method research questionnaire upon request, a questionnaire was sent. Patients who stopped coming to clinics, not reached, refused to return to clinics, declined to take part in the questionnaire, or died were defined as loss to follow‐up (LTFU). The patients who attended for regular review (defined as ‘Adherent’ group) were compared to those who attended for irregular review and LTFU, namely the ‘Nonadherent’ group.

Results Patients were aged 21‐73 (average 52) years with a mean (±SD) body mass index (BMI) of 48.2±9.5 Kg/m2. Among the obesity‐related complications, osteoarthritis (OA) and/or weight‐ bearing joint pain being the most prevalent one, endorsed by as high as nearly three quarters of our patients. 107 (63.7%) patients attended follow‐up regularly, 20 (11.9%) attended irregularly, and 41 (24.4%) were no longer attending reviews or considered LTFU. Baseline BMI was not statistically different between those who were Adherent and Nonadherent to follow‐up care

180

CHAPTER 3 – Adherence to clinic follow‐up | Michelle M.C. Tan

(47.1±9.3 Kg/m2 versus 49.6±9.8 Kg/m2, respectively). Patients who attended reviews regularly had no greater number of comorbidities, mental illness and length of hospital stay (LOHS) compared to the Nonadherent group. Patients experienced low rates of serious adverse events after surgery, with no discrepancies observed between the two groups. Likewise, mixed model with random effects revealed no pronounced difference in the mean weight loss between the two groups over time. Of the 14 reasons for withdrawal from post‐surgical follow‐up visits captured, the most common reasons according to patients were travel distance (n=11), other significant illness (n=7), family‐related issues (n=5), and work‐related problems (n=4). In multivariate logistic regression analyses, pre‐operative predictors of adherence to follow‐up were determined to be increasing age and relationship status. Distance from the clinic was found not one of the predictors of adherence to follow up.

Conclusion The adherence rate to follow‐up visits after bariatric surgery among our patients was 63.7%. Although the adherence rate in our service was high while nonadherence did not impair weight loss results, there was still a significant minority of patients who did not adhere with follow‐up recommendations that putting them at risk of inadequate medical care. Withdrawal from publicly funded bariatric surgery service was mainly associated with travel distance. Older and partnered patients were more likely to adhere to follow‐up care after operation. The findings of this research may guide future patient selection or care practices for patients needing additional contacts and supports. With this, the patients may instead benefit from additional assistance to experience optimal outcomes.

Keywords Bariatric surgery; Adherence; Follow‐up; Predictors of adherence; Reasons for attrition

3.2 BACKGROUND

Bariatric surgery is known as a highly effective treatment option for extreme obesity in terms of both weight loss, and remission or improvement of obesity‐related comorbidities. Nevertheless, long‐term weight loss and maintenance is variable and not always achieved. Up to 50% of patients experience weight regain and return of the comorbidities within 5–10 years of their surgical treatment (161, 240, 391, 392). Some studies have found that patients with higher rates of clinical follow‐up attendance are more likely to have better long‐term weight maintenance and remission of comorbidities (393‐397). Besides that, by attending clinical follow‐up reviews following surgery, complications such as nutritional deficiency and Barrett’s oesophagus (a

181

CHAPTER 3 – Adherence to clinic follow‐up | Michelle M.C. Tan potentially precancerous condition) could be identified and treated at an early stage (127, 398). Thus, pre‐ and post‐operative multidisciplinary evaluation and follow‐up and long‐term adherence are highly recommended for patients undergoing bariatric surgical treatment, leading to mounting availability of multidisciplinary services in specialist obesity clinics.

Despite the recommendations, a considerable proportion of patients do not attend adequate routine post‐bariatric surgery review visits. A number of definitions have been adopted for adherence to follow‐up, such as attended at least one clinic appointment in 3 or 12 months (393, 399, 400); half or more of pre‐scheduled post‐operative reviews (396, 401); or the majority of scheduled follow‐up visits (i.e. 3 or 4 out of 4) (240). These wide variance in the definition of ‘adherence’ contributes to the broad range of reported nonadherence rates, vary markedly between different bariatric centres and populations, ranged between 3% to 63% (240, 401‐408).

In accordance with the World Health Organization (WHO), the term ‘adherence’ is preferable to ‘compliance’ owing to the fact that the latter implies patient submission to healthcare professionals’ commands without mutual negotiation. Whereas adherence requires the patient’s agreement to treatment recommendations, based on the suggestion that patients should be active partners with healthcare professionals in their own care, and that good communication between patients and healthcare professionals is a must for an effective clinical practice (409).

Although it was speculated that follow‐up attrition contributes to defective long‐term post‐ surgical outcomes, little is known of the pre‐operative patient characteristics that predict nonadherence to post‐surgical follow‐up visits. Only a handful of studies have focused on this matter but different results have been reported (240, 400, 401, 410, 411). Moreover, a substantial number of patients undergoing bariatric surgery have been reported as loss to follow‐ up (LTFU) for no apparent reasons, which in turn may cause an overestimation of the benefits of operation, negligence of potential surgical complications, loss of a support network and lack of reinforcement to follow medical regimen. Understanding the barriers and reasons for which patients are unable to adhere to post‐operative clinic reviews could provide knowledge

182

CHAPTER 3 – Adherence to clinic follow‐up | Michelle M.C. Tan associated with the real outcomes after surgery. Furthermore, it could contribute to monitoring strategies to solve the nonadherence issue and facilitate patient adherence to post‐surgical follow‐up and multidisciplinary weight management programs (WMPs), which could be essential to ensure weight loss maintenance, to manage obesity‐related comorbidities and to minimize risk of complications. Importantly, central to the perpetual scarcity of funds of public healthcare system, more information is needed to optimise the healthcare system to maximise the attendance of follow‐up care.

Taken together, the primary aim of the present study was to assess baseline socio‐demographic and medical predictors of adherence to post‐operative follow‐up in patients who underwent bariatric surgery. These included sleeve gastrectomy (SG), mini gastric bypass‐one anastomosis gastric bypass (MGB‐OAGB), adjustable gastric banding (AGB) and roux‐en‐Y gastric bypass (RYGB). Additionally, we also aimed to evaluate characteristics of patients who adhere to follow‐ up visits and those who do not, and to investigate the relationship of post‐operative attendance to pre‐scheduled clinic appointments with weight control. We also sought to identify the reasons for failure to attend clinic reviews after bariatric surgery in a publicly funded healthcare service.

3.3 MATERIALS AND METHODS

Study population This was a cohort study of retrospectively collected data and prospectively following up on patients who underwent bariatric surgery from 2009 to 2017 through the first and oldest state of New South Wales (NSW)’s Publicly funded Bariatric Surgery Collaborative. We are a partnership of The University of Sydney, Royal Prince Alfred (RPA) Hospital, Concord Repatriation General Hospital, and Camden Hospital located within Sydney Local Health District (SLHD) and South Western Sydney Local Health District (SWSLHD), Sydney, Australia. This service consists of an inter‐professional team of consultant endocrinologists, bariatric surgeons, nurse practitioners, dietitians, and psychologists, who assess patients for bariatric surgery and provide bariatric

183

CHAPTER 3 – Adherence to clinic follow‐up | Michelle M.C. Tan aftercare in the abovementioned participating public hospitals. Patients underwent a multidisciplinary lifestyle intervention at one of the three hospitals before undergoing bariatric surgery at one single operating centre, and then returned to respective clinics for scheduled follow‐up visits following operations.

Eligible patients were 18 to 70 years of age, with an initial body mass index (BMI) ≥40 Kg/m2 with at least one comorbidity of obesity which improves with weight loss. The study was approved by the institutional review board of SLHD, SWSLHD, and the three hospitals (Ethics Approval Number: X15‐0339 & LNR/15/RPAH/463). Written informed consent was obtained from all patients.

Bariatric surgery assessment process Before undergoing bariatric surgery, baseline comorbidities were assessed by operating bariatric surgeons, consultant endocrinologists and nurse practitioners. A psychiatrist or psychologist evaluated all patients for any history of psychiatric illness along with any ongoing psychiatric comorbidities. All bariatric procedures were attempted laparoscopically by the same surgical team, and the details of surgical techniques have been previously described (CHAPTER 2).

Follow‐up and questionnaire All patients were informed details about the clinic follow‐up regimen, including the need and importance of post‐operative follow‐up reviews, and were asked and agreed to commit to lifelong follow‐up in the clinics. The follow‐up regimen includes multidisciplinary post‐operative follow‐up at 3 months, 6 months, 1 year, 18 months, 2 years and annually thereafter up to 9 years. At each time interval, patients were assessed and counselled by a consultant endocrinologist, nurse practitioner or dietitian, and/or exercise physiologist who were also involved in assessing patients pre‐operatively. At each visit following bariatric surgery, patients were monitored with respect to their diet, nutritional supplementation, exercise, medical conditions, mental health, wellbeing, medications, weight, blood pressure and blood profiles. There is also an option for patients who need to follow up through telemedicine.

184

CHAPTER 3 – Adherence to clinic follow‐up | Michelle M.C. Tan

After the identification of all patients who missed appointments, they were called by the study investigators to re‐schedule their review appointments. Three telephone calls were attempted for each patient, and if a patient was uncontactable by phone; a letter, mixed‐methods research questionnaire and blood test request form accompanied by a postage‐paid return envelope were sent to the patient’s home address requesting them to contact the clinic to make an appointment while filling in the questionnaire. The detailed questionnaire developed covers collection of the current quantitative and qualitative data on socio‐demographics, lifestyle behaviours, medical status, blood parameters, and open‐ended questions devoted to identifying reasons for not attending clinic follow‐up appointments.

Data collection Patients’ demographic characteristics, socioeconomic status, health status, medications and blood tests were gathered. Patients’ demographic variables comprised age, sex, race, employment, relationship status, and distance to clinic. Employment status was categorised into two groups on the basis of occupational time commitment: working versus non‐working. The working group included individuals with full‐time, part‐time and casual occupation status, and the non‐working group included those who were retired, on the government support schemes, unemployed or being a homemaker. Travel distance (Km) to clinics were calculated (http://www.australiapostcodes.com/distance‐between‐postcodes) and categorised into two categories: 0–49 and ≥50 to reflect geographic areas and proximity to patients’ respective clinics. Variables describing patients’ anthropometric characteristics included height, weight, and body mass index (BMI). Patients’ comorbidities and medications at entry to clinics and pre‐operation were collected, including diabetes mellitus, hypertension, hyperlipidaemia, osteoarthritis (OA) and/or weight‐bearing joint pain (WBJP), obstructive sleep apnoea (OSA)/obesity hypoventilation syndrome (OHS), gastro‐oesophageal reflux disease (GORD), asthma, cardiovascular disease, gout, hyperuricaemia and fatty liver disease. Patients’ mental illness included depressive disorder and/or severe anxiety. The psychiatric diagnoses were either lifetime diagnoses or active conditions that were well‐managed at the time of pre‐surgical evaluation with no contraindication to surgery.

185

CHAPTER 3 – Adherence to clinic follow‐up | Michelle M.C. Tan

Outcomes and definitions of adherence to follow‐up care The primary outcome was adherence to follow‐up to specialist obesity clinic reviews by endocrinologists, surgeons, dietitians, exercise physiologists and psychologists. ‘Adherence’ or regular review was defined a priori as attended a minimum of two visits within the first year and second year of post‐operation, as well as annually since third year of surgery of scheduled post‐ operative multidisciplinary review appointments as portrayed in Figure 3.1.

Bariatric Surgery Month

0 3 6 12 18 24 36 48 60 72 84 96 108 Adherent if attended at least: 2 visits 2 visits 1 visit 1 visit 1 visit 1 visit 1 visit 1 visit 1 visit

Figure 3.1 Definition of ‘Adherence’ or regularly attending follow‐up care post‐operatively

In other terms, the remainder were thereafter classified as ‘Nonadherent’, which consisted of patients who were irregularly reviewed and ceased attending clinic appointments/LTFU. Particularly, irregular review was defined as not attended a minimum of two appointments within the first and second year post bariatric surgery, or missed any of the annual follow‐up from third year post‐surgery onwards. If a patient returns to clinic for follow‐up after study investigators’ phone call, mail and/or questionnaire follow‐up (See APPENDIX for the questionnaire), the patient’s status was counted as irregularly attended clinic reviews. Whereas loss to follow‐up (LTFU) was defined as patients who ceased coming to clinic review, who were not reached, refused to return to clinics, refused to take part in the questionnaire, or died. For patients who were LTFU and did not return their questionnaire, the last clinical data in hospital records available were used for analysis.

STATISTICAL ANALYSIS A statistical analysis was carried out using the SPSS statistical software version 26 (SPSS Inc., Armonk, NY, USA). Continuous variables were reported as means and standard deviations (SD) and categorical variables as frequencies and proportions (%). Adherent/Regularly review group

186

CHAPTER 3 – Adherence to clinic follow‐up | Michelle M.C. Tan and Nonadherent (irregularly reviewed + no longer attending clinic reviews/LTFU) group were compared using independent t‐test for continuous variables, whereas categorical variables were compared using Chi‐square (χ2) test. Mann‐Whitney U test was performed for skewed data.

A mixed‐effects model with random effects was fitted to model the changes in weight over time, expressed as continuous data, with adjustment for age at time of surgery, sex and race. Comparisons of the modelled means included the change from baseline to each visit between follow‐up groups and within each group. Specifically, the mixed‐effects model included fixed effect terms for age at time of surgery, sex, race and visits. A repeated measures term was included for visits within each subject to take into account multiple observations over time. The random intercept for subject ID was included to allow for different baseline weights for each individual (within the groups). To compare groups of subjects, fixed effects terms were included in the model identifying group membership (follow‐up type) and allowing the mean response at each visit to differ between the groups (interaction term). With the small number of patients and the subsequent even smaller numbers of observations at each clinic visit year in the irregular group, comparisons using this approach will be weighted towards the irregular group. Therefore, the model was fitted with only two follow‐up groups (i.e. the irregular, ceased attending follow‐ up or LTFU categories were combined into Nonadherent group) for the corresponding comparisons of interest.

A multivariable logistic regression model using the Forward Likelihood Ratio (LR) method with entry criteria of α=0.05 and removal criteria=0.10 was performed to identify predictors of adherence to follow‐up appointments, adjusted for possible confounders (age, sex and race). The model was also assessed for multicollinearity, and none was found. All statistical analyses were two‐sided with a significance of p<0.05.

187

CHAPTER 3 – Adherence to clinic follow‐up | Michelle M.C. Tan

3.4 RESULTS

A total of one hundred and sixty‐eight patients successfully completed the pre‐operative work up process and underwent bariatric surgery between 2009 and 2017, and were monitored for at least a year. Of study patients, 66.1% were females, 71.4% Caucasian and 83.9% had a sleeve gastrectomy (SG) performed as their primary surgical procedure. The mean follow‐up duration after bariatric surgery was 4 years, with the maximal length of post‐surgical follow‐up being 9 years. Of these, 107 (63.7%) patients attended post‐surgical follow‐up in our hospitals regularly, 20 (11.9%) attended irregularly, and 41 (24.4%) no longer attending follow‐up/LTFU throughout the entire follow‐up period. In aggregate, the composited Nonadherence group was 36.3% of the total number of patients (n=61). A flow diagram detailing the patient disposition is as laid out in Figure 3.2:

188

CHAPTER 3 – Adherence to clinic follow‐up | Michelle M.C. Tan

168 Bariatric surgeries were publicly funded and performed from 2009‐2017

Post‐operative follow‐up care

Missed clinic review appointments

Contacts attempted

Phone calls answered

Willing to re‐engage (n=45) and have returned to clinic appointments (n=9) Returned a questionnaire (n=14) In prison (n=3)

Met with patients outside Deceased (n=3) clinics, invited but patients refused to attend clinic Unreachable review appointments (n=4) (n=2)

Irregularly attended No longer attending clinic reviews or loss to follow up (LTFU) clinic reviews ‐ (n=20, 11.9%) (n=41, 24.4%)

Adherent group Nonadherent group (n=107, 63.7%) (n=61, 36.3%)

Figure 3.2 Flow diagram of patient tracking

189

CHAPTER 3 – Adherence to clinic follow‐up | Michelle M.C. Tan

The baseline sociodemographic, anthropometric and clinical characteristics for the study patients are outlined in Table 3.1.

Table 3.1 Baseline characteristics of patients and the relationship with follow‐up adherence after bariatric surgery (n=168) Overall Adherent group Nonadherent group p value† (n=168) (n=107) (n=61)

Socio‐demographics

Age at time of surgery, mean 52.0±11.4 53.5±10.6 49.2±12.2 0.018* (±SD) (years)

Sex, n (%) 0.565

Female 111 (66.1%) 69 (64.5%) 42 (68.9%)

Male 57 (33.9%) 38 (35.5%) 19 (31.1%)

Relationship status, n (%) 0.010*

Married/In a relationship 82 (49.7%) 59 (35.8%) 23 (13.9%)

Divorced/Separated/ 67 (40.6%) 33 (31.7%) 34 (55.7%) Widowed

Single 16 (9.7%) 12 (11.5%) 4 (6.6%)

Race, n (%) 0.550

Caucasian 120 (71.4%) 79 (73.8%) 41 (67.2%)

Middle Eastern 24 (14.3%) 15 (14.0%) 9 (14.8%)

Othersᵠ 24 (14.3%) 13 (12.1%) 11 (18.0%)

Employment status

Working (Employed) 57 (33.9%) 39 (36.4%) 18 (29.5%) n/a

Distance to clinic (Km)

Mean (±SD) 38.7±68.9 40.6±71.5 35.2±64.5 0.490

Median 23.0 23.0 23.0 0.870‡

190

Overall Adherent group Nonadherent group p value† (n=168) (n=107) (n=61)

Range 0−452 0−452 1−442

n (%) 0.431

0–49 147 (87.5%) 92 (86.0%) 55 (90.2%)

≥50 21 (12.5%) 15 (14.0%) 6 (9.8%)

Anthropometric measurements

Weight at surgery, 0.099 mean (±SD) (Kg) 133.2±32.4 130.0±32.4 138.6±32.1

BMI at surgery, 0.104 mean (±SD), (Kg/m2) 48.0±9.5 47.1±9.3 49.6±9.8

Clinical characteristics

Length of hospital stay, days 2.9±1.1 2.8±1.0 3.0±1.2 0.187 Range (days) 1–8 1–8 1–8

Peri‐ and post‐surgery 58 (34.5%) 39 (36.4%) 19 (31.1%) 0.487 adverse events, n (%) Abbreviations: BMI=Body mass index; SD=Standard deviation; Km=Kilometres †Independent t‐test or Chi‐square (χ2) test used unless otherwise indicated *p<0.05 ᵠIndigenous Australian, Pacific Islander, Americas, Black African, Mauritian and Pakistanis ‡Mann‐Whitney U test

At the time of bariatric surgery, the mean age of all patients who received bariatric surgery was 52.0±11.4 years. Their BMI being 48.0±9.5 Kg/m2, females were two‐folds as many of males and approximately one‐third was working. Furthermore, majority of the patients were Caucasian (71.4%), and were more often either married, had a partner, divorced, separated or became widow. The mean travel distance to attending post‐surgical follow‐up care was 50.8±117.6 Km. Distributions of sociodemographic, anthropometric and clinical characteristics were similar for those who regularly attended clinic visits following bariatric surgery and those who did not, as demonstrated in Table 3.1. Similarly, there is no significant difference in the length of hospital stay and surgical complication rates between the two cohorts. The only exception was age and

191

CHAPTER 3 – Adherence to clinic follow‐up | Michelle M.C. Tan

relationship status where the Adherent group appears significantly older as well as having higher proportion of partnered patients than the Nonadherent group (p<0.05).

Table 3.2 Baseline obesity‐related comorbidities and mental illness stratified by adherence to follow‐up appointments

Overall Adherent Nonadherent p value (n=168) (n=107) (n=61) Total number of comorbidities, mean (±SD) 5.8±2.2 5.9±2.1 5.8±2.2 0.782 Selected obesity‐associated comorbidities, n (%)

Osteoarthritis (OA) and/or weight‐bearing joint pain¥ 120 (71.4%) 78 (72.9%) 42 (68.9%) 0.577 Type 2 diabetes mellitus 114 (67.9%) 69 (64.5%) 45 (73.8%) 0.211 Hypertension 110 (65.5%) 70 (63.6%) 40 (65.6%) 0.984 Obstructive sleep apnoea (OSA) or 107 (63.7%) 69 (64.5%) 38 (62.3%) 0.776 Obesity hypoventilation syndrome (OHS) Hyperlipidaemia 99 (58.9%) 64 (59.8%) 35 (57.4%) 0.758 Asthma 46 (27.4%) 29 (27.1%) 17 (27.9%) 0.915 Fatty liver disease 43 (25.6%) 30 (28.0%) 13 (21.3%) 0.332 Mental illness, n (%)

Depression and/or severe anxiety, n (%) 79 (47.0%) 49 (45.8%) 30 (49.2%) 0.672 Depression or mood disorders, n (%) 74 (44.0%) 46 (43.0%) 28 (45.9%) 0.715 Severe anxiety, n (%) 44 (26.2%) 28 (26.2%) 16 (26.2%) 0.993 ¥Weight‐bearing joint pain defined as knee pain, back pain and/or hip pain

In the presence of detailed data, the top seven most common comorbidities associated with obesity at baseline are described in Table 3.2, with osteoarthritis (OA) and/or weight‐bearing joint pain (WBJP) being the most prevalent pre‐existing comorbidity, being present in 71.4% patients, i.e. almost three quarters of our total sample. Further clinical information such as depression and severe anxiety are also displayed in Table 3.2. Of the 168 patients in our program, there was a median of six obesity‐related complications prior to surgery (ranged 1–12). When comparing the Adherent and Nonadherent groups, there was no significant difference between

192

CHAPTER 3 – Adherence to clinic follow‐up | Michelle M.C. Tan their presence and number of obesity‐related comorbidities at baseline. Likewise, the occurrence of baseline mental illness between the comparison groups were not different.

Figure 3.3 Comparison of estimated marginal mean differences of post‐surgical weight from pre‐operative baseline (Kg) between Adherent and Nonadherent groups by clinic visits adjusted for age at surgery, sex and race

(Kg)

weight

mean

marginal

Estimated

Years since bariatric surgery

No. of patients 168 157 129 98 79 63 42 Adherent 107 106 86 71 55 43 24 Nonadherent 61 51 43 27 24 20 18

The weight trajectories over time were modelled from fitting of patient observational data stratified into follow‐up type with two categories, the Adherent and Nonadherent groups. Lines and data markers indicate modelled weight from baseline based on mixed‐effects models, adjusted for baseline factors (age, sex and race) independently related to missing follow‐up appointments. Bars correspond to standard errors.

The graphical view of the modelled mean weights between the Adherent and Nonadherent groups from baseline by timepoint are presented in Figure 3.3, estimated from the mixed‐effects model with random effects that takes into account the repeated measures nature of the data. The body weight measurements of the patients who ceased attending our clinic reviews were obtained from the medical records or questionnaires, whichever that were available. Despite

193

CHAPTER 3 – Adherence to clinic follow‐up | Michelle M.C. Tan achieved significant weight loss since initial and pre‐operative medical reviews (p<0.001), we did not find a significant difference in change in weight between the two follow‐up groups over time. There is also no statistical significance observed in terms of the interaction between time and follow‐up groups. The most considerable mean weight reduction was observed within the first‐ year post‐operative window (Figure 3.3). Between years 2 and 6, there was a mean weight regain that followed a quadratic trend for both follow‐up type groups. The modelled mean weight difference between the two follow‐up groups after bariatric surgery from baseline to years 6 was: 7.2 Kg (95% CI=1.8‐16.1) (pre‐operative baseline), 6.6 Kg (95% CI=0.7‐13.8) (post‐operative year 1), 6.2 Kg (95% CI=1.0‐13.3) (year 2), 5.4 Kg (95% CI=1.7‐12.6) (year 3), 3.7 Kg (95% CI=3.8‐11.2) (year 4), 1.8 Kg (95% CI=6.4‐10.0) (year 5), and 5.3 Kg (95% CI=5.1‐15.7) (year 6), respectively.

Predictors of adherence to follow‐up appointments

Table 3.3 depicts the multivariate logistic regression that modelled the relationship between adherence to clinical follow‐up regimens and pre‐operative baseline predictors. The list of independent variables included in the binary logistic regression model included variables described in the aforementioned results from bivariate analyses (Tables 3.1 and 3.2), which comprised a broad range of potential baseline predictors.

194

CHAPTER 3 – Adherence to clinic follow‐up | Michelle M.C. Tan

Table 3.3 Final multivariate logistic regression modelling for predictors of adherence to post‐ operative follow‐up Parameter B Adjusted‐ p 95% Confidence interval (CI) Odds Ratio value (OR) Lower bound Upper bound

Age at surgery 0.029 1.029 0.048 1.000 1.059

Relationship status

Unpartnered ‐0.704 0.495 0.036 0.256 0.954 (Single/Separated/Divorced/ Widowed)

Partnered 0 1 (Married/In a relationship) (Reference) All independent variables included in the binary logistic regression models were: age at surgery (years) (continuous data), distance between follow‐up hospitals and homes (Km) (continuous data), sex (Female vs Male) (categorical data), race (Caucasian vs Middle‐Eastern vs Others) (categorical data), relationship status (Unpartnered vs Partnered) (categorical data), working (Yes vs No) (categorical data), pre‐operative body mass index (Kg/m2) (continuous data), number of baseline comorbidities (continuous data), baseline depression (Yes vs No) (categorical data), and baseline severe anxiety (Yes vs No) (categorical data).

In view of the small number of single relationship patients, while to improve the accuracy and stability of estimates for this parameter in the logistic regression model, the relationship status was collapsed into two categories − partnered (married/in a relationship) and not partnered (single/separated/divorced/widowed).

The logistic regression model (Table 3.3) shows that age and relationship status are significant predictors of adherence to clinic follow‐up after surgery. Increasing age was a significant factor in engagement in follow‐up care (Adjusted‐OR=1.029, 95% CI=1.000‐1.059), with the parameter estimate being marginally significant. Compared to the unpartnered group, the patients who were married or in a relationship at baseline were twice as likely to be adherent to follow‐up after bariatric surgery (Adjusted‐OR=2.022, 95% CI=1.048‐3.899).

195

CHAPTER 3 – Adherence to clinic follow‐up | Michelle M.C. Tan

In summary, the findings in the present study revealed that older age and partnered relationship at baseline are significant positive factors likely leading to higher adherence to follow‐up among the publicly funded patients who underwent bariatric surgery.

Pathways to post‐operative follow‐up care withdrawal

Since we also aimed to understand the particularities of the withdrawal group for further improvement of multidisciplinary publicly funded bariatric surgery services, we retrieved the patients’ reasons for no longer attending clinic appointments. Table 3.4 lists the reasons to justify ceasing attendance at clinic appointments that we managed to capture during our phone interviews and the open‐ended questionnaires patients had returned to us. Reasons such as imprisonment and death were retrieved from the hospital records.

Some patients who missed earlier annual visits returned and participated in data collection during a later visit upon ringing and posting questionnaires to them. These patients expressed willingness to re‐engage with our service, thus adherence status changed from LTFU to irregular (n=9). Reasons for withdrawal from the service were thus retrieved and documented for the remaining 41 patients who were no longer attending follow‐up care.

Table 3.4 Reasons for withdrawal from the publicly funded bariatric surgery service or loss to follow‐up to scheduled post‐operative aftercare appointments (n=41) Reasons Frequency of responses

Distance to clinic 11§

Other significant illness 7§

Family‐related issues 5§

Work‐related issues 4§

Unwilling to return to return for clinic review appointments 3§

In prison 3

196

Reasons Frequency of responses

Death 3

Patient considered follow‐up unnecessary 2§

Reported not being told by clinic that follow‐up is necessary 2§

Weight regain/Dissatisfaction with results 1§

Intellectual impairment 1

Prefer to receive medical care locally 1§

Social commitment 1

“Lazy” 1§

Uncontactable 4 §8 Patients with multiple reasons

As shown in the binary logistic regression modelling for predictors of adherence in Table 3.3, the objective distance (Km) between hospitals and residing homes was found not a significant predictive factor leading to attrition. Interestingly, it was however the most common reason for ceasing attendance at clinic review appointments after bariatric surgery, according to the patients (n=11). Amongst the patients who withdrew from the current program and expressed distance as a barrier to attend further clinic visits, over half of them reported having moved outside of Sydney or Australia (n=7). This has further been verified with their new home addresses provided by the patients or medical records. Other common reasons recorded included significant illnesses (e.g. leukaemia, cancer and major depression) (n=7), family‐ related issues (n=5), work‐related circumstances (n=4), unwillingness to return to clinic follow‐ up appointments (n=3) and imprisonment (n=3). Three patients had died during the follow‐up period due to non‐surgical‐related causes as previously elaborated in CHAPTER 2. Only one patient cited weight regain as the reason for ceasing attendance to clinic visits. Interestingly, two patients decided themselves that they did not need to attend follow up, and another two patients believed that clinics have discharged them. This was despite emphasis placed in pre‐surgical

197

CHAPTER 3 – Adherence to clinic follow‐up | Michelle M.C. Tan intervention and screening of bariatric surgery candidates regarding the significance of the post‐ operative follow‐up and adherence to it.

Some patients (n=8) had more than one reason for withdrawing from the service. For example, a patient indicated failure to meet weight loss target as well as far travel distance to clinic, and hence left the service. In another instance, a patient blamed himself for being lazy to engage to follow‐up care in addition to distance, despite living only 5 Km from the hospital. Whereas another patient raised both work‐ and family‐related issues as barriers to return to clinic visits. For others, there was a patient who expressed her preference to receive medical care locally was also one of those who reported have had moved to another state. Only four patients were unreachable by any attempts and methods of contact.

3.5 DISCUSSION

This chapter shows that the overall rate for adherence to follow‐up appointments in the current study was 63.7%, in other words, 36.3% was considered nonadherent. This adherence rate is comparable to another similar publicly funded bariatric surgery program in Canada which reported 62.1% of adherence rate at 2‐years endpoints (240). The authors defined adherence to follow‐up care as having attended at least 3 out of 4 of clinic visits. A study in Spain among 263 patients also of public National Health Care Service recorded a lower nonadherence rate of 17.5% (404), defined by not attending any post‐surgical appointments for more than six months, primarily as a result of work‐related circumstances (n=12), family‐related problems (n=6) and living outside the city or country (n=5). Whereas another study of 178 Israeli patients reported a far higher nonadherence rate of 53.4%, defined by attending ≥ 6 of a total of 9 meetings in the first post‐operative year, with work‐related issues being the main self‐reported reason for missing follow‐up appointments with the surgeon (n=77) and dietitian (n=37), followed by mobility difficulties/distance to the clinic (n=72 and n=30, respectively). Of note, the different

198

CHAPTER 3 – Adherence to clinic follow‐up | Michelle M.C. Tan rates may result from the variability from study to study, in the sense of the definition of and content of “a clinical follow‐up visit” and the follow‐up time period.

Consistent with some literature (401), the adherence to routine post‐operative follow‐up was unassociated with weight loss outcomes. Post‐operative clinic attendance is probably not an adequate indicator of a patient’s adherence to the behavioural and nutritional changes required after bariatric surgery. Significant variation does exist in clinic review content and healthcare provider’s discipline, philosophic approach, knowledge, and patient relationship or communication skills (412).

To our best knowledge, only a handful of studies have been able to uncover predictors and/or specific reasons of nonadherence or loss to follow‐up (LTFU) (240, 400, 401, 404, 410, 411). Recognizing the importance of both aspects, our research included examinations of both predictors of adherence and reasons for withdrawal from the service from patients’ perspectives. In the part of our study where we explored the baseline predictors of follow‐up adherence after bariatric surgery, age and relationship status were found to be positive predictors of adherence to follow‐up care. The tendency of older patients in adhering post‐operative review appointments have been reported in several previous studies (400, 401, 410, 411). This result is not particularly surprising, as older age has been associated with declining health (413) and more positive help‐seeking attitudes (414). Hence, our multidisciplinary service entailing medical consultation, dietetic advice, exercise program, and psychosocial support could be directed at older patients, with close follow‐up in their management of weight loss and obesity‐related comorbidities. On top of that, younger patients tend to divide their attention to family obligations especially if they have young children in the household, work demands and social activities more than older patients (401, 410, 411). Special efforts should be implemented to encourage the younger patients to attend follow‐up care.

Furthermore, we also found a significant positive trend between people who had a partner at the time of surgery and adherence to follow‐up. With the other literature has shown that men with

199

CHAPTER 3 – Adherence to clinic follow‐up | Michelle M.C. Tan

HIV are more likely to attend clinic follow up if they were living with a partner, perhaps because their partners served as treatment supporters (415). In a situation of chronic disease such as clinically severe obesity, patients may not prioritize follow‐up visits if they are not highly‐ motivated or do not appreciate the potential positive impact on their health. Indeed, encouraging and motivating patients to attend long‐term post‐operative follow‐up care is an important issue. Strategies to motivate the unpartnered patients could improve higher adherence, hence ameliorate the public healthcare systems and patient outcomes over time. For instance, with over half of our study population being in separation, divorced, widowed or single at baseline, the integration of peer support program or expansion of patients’ current support network may be beneficial to serve as powerful treatment supporters or companionships. In fact, there is abundant consistent evidence demonstrating that peer support, regardless formal or informal, received by patients from other members of their community, can encourage patients to stay connected to services and improve adherence to treatment (409, 416‐419). Concurrently, this could reduce the amount of time devoted by the healthcare professionals to the care of chronic conditions such as obesity (409).

In another part of our study, we also managed to capture the subjective reasons for withdrawal from the publicly funded bariatric surgery service. Travel distance has been emerged as a barrier to clinic review (420, 421). In our service, distance from clinic review prior to surgery was generally short, as the catchment area of majority of our patients was within the city of Sydney, although we are tertiary referral service which is also accepting interstate referrals. This explains the ambiguity that although travel distance was the principal barrier to post‐surgical follow‐up adherence in this study, it did not show a significant impact as a pre‐operative predictor in the logistic regression model. Some of the patients indeed had moved to living outside of Sydney after their surgery over the follow‐up years and, hence, a valid justification for withdrawal from the service. Undoubtedly, this information provides a vast insight into patients’ perception regarding the usefulness of the post‐surgical follow‐up service and barriers that hindered their adherence to follow‐up care. To resolve this barrier, we have started to implement a telemedicine option in our clinics to help the patients in need. This could not only enhance access

200

CHAPTER 3 – Adherence to clinic follow‐up | Michelle M.C. Tan to health information for patients and clinicians, but also could facilitate remote patient monitoring, and deliver timely healthcare recommendations to patients. Having said that, the subjective reasons reported were only among patients who no longer attending clinic reviews which in our opinion is more clinically meaningful, but not among the irregularly reviewed patients. Some clinical studies have suggested strategies to increase clinic attendance rates, which might be applicable to our cohort and patients of other specialist obesity services. Some examples that might be effective and plausible in clinical practice include flexible schedules; reminder via letter or telephone from clinic personnel (who are seen as caring and engaged and who provide reinforcement for follow up); specific call protocols; virtual therapy groups (e.g. group WhatsApp/Telegram/video conferencing); and frequent contact information updates (248, 411, 412, 422). The follow‐up strategy shall also be tailored to the individual requirements and prerequisites of each patient, such as give thought to a patient’s travel distance to clinic, by flexibly scheduling their next appointment.

The present findings might be best interpreted taking into account some other considerations. A publicly funded bariatric surgery service such as ours improves access to patients who are highly severe and complex, as in heavier in body weights and living with more coexistent obesity‐related comorbidities, but cannot afford private health insurance. In this scenario, our patients have no financial limitations for the surgical treatment and follow‐up as some other authors have described, because all patients are part of a publicly funded program, and there is no additional medical cost to patients for clinical follow‐up visits. Despite removal of this known barrier, post‐ operative non‐attendance remained an issue within our bariatric surgical cohort. It is noteworthy that understandably, patients were incurred travel expenses to get to our service particularly if they live far away, and for those who are working, it costs them “time off from work”.

To date, there is also no one universal definition that is agreed upon by all healthcare providers and researchers. Consequently, different groups apply different definitions of adherence to follow‐up. One imperative step that should be taken is collectively deriving a unified definition for adherence to follow‐up care to standardize all the future research studies. Concerning

201

CHAPTER 3 – Adherence to clinic follow‐up | Michelle M.C. Tan patients’ marital status as a positive predictor of adherence to follow‐up, this warrants future studies. Given that bariatric surgery is a life‐changing treatment for patients with clinically severe obesity, the changes in relationship status after bariatric surgery are along the same lines. Changes of relationship status after bariatric surgery have been frequently observed in our clinical practice and other reports. Of particular, the recent study of two large Swedish cohorts who had bariatric surgery with long‐term follow‐up periods (median: 10 years, range: 0.5‐20 years) – the landmark Swedish Obese Subjects (SOS) study (n=3,870) and the Scandinavian Obesity Surgery Registry (SOReg) cohort study (n=29,234) (423). The authors found that besides its associations with obesity‐related comorbidities, bariatric surgery‐induced weight loss was also significantly associated with changes in relationship status, including increased incidence of divorce, remarriage, or marriage/having new partner, suggesting it could be the subject of further investigation in the future.

Although strengths of this study include determination of pre‐operative predictors of adherence to follow‐up in specialist obesity services, and clear elucidation of subjective reasons for withdrawal from the publicly funded bariatric surgery service, which have only been discussed scarcely until now, limitation must be acknowledged too. Some of our patients who ceased attending clinic review appointments were loss to follow‐up in clinics as well as the hospitals. Because of this, their weight trajectories are unknown, pointing to our question how similar or different are they to the Adherent group or patients who were irregularly reviewed.

3.6 CONCLUSION

The adherence rate to post‐surgery clinic review visits in our publicly funded bariatric surgery service was high. There is no difference in the weight loss outcomes identified between the Adherent and Nonadherent groups following bariatric surgery. However, the latter group of patients who did not adhere with follow‐up recommendations might be at risk of inadequate medical care. Therefore, it is crucial to develop a policy for specialised services or local general

202

CHAPTER 3 – Adherence to clinic follow‐up | Michelle M.C. Tan practitioners (GPs) to manage the patients with clinically severe obesity who underwent publicly funded bariatric surgery in the primary care setting in long‐term. This will be most effective in reinforcing medical, diet and exercise recommendations at each consultation, while monitoring for possible nutritional deficiencies and surgical complications. The specialist services or GPs will also be helpful in ensuring adequate psychological support for the patients and manage their obesity‐related comorbidities.

The findings of this longitudinal study indicate that certain patient characteristics, namely older age and relationship status may be useful to identify patients likely to attend follow‐up visits regularly. The younger and unpartnered groups should receive more attention and support to boost their motivation to attend post‐surgical follow‐up visits to improve long‐term outcomes in the provision of care. Incorporation of the strategies for the addressable reasons that might influence adherence, such as travel distance to hospitals, could also optimize patients’ adherence to follow‐up appointments, health and outcomes.

203

CHAPTER 4 – Pre‐operative predicts post‐operative weight loss | Michelle M.C. Tan

CHAPTER 4

RESEARCH THEME 3

Qualifying for funded‐bariatric surgery in public hospitals: Does pre‐operative weight loss matter?

4.1 ABSTRACT

Background Despite the efficacy and durability of bariatric surgery, it is only available at a limited number of public hospitals and is only offered to individuals who have chronic complications of their obesity and when all other weight loss treatments have failed. Many publicly funded bariatric surgery services have mandated that bariatric surgery candidates undergo an integrated pre‐operative behavioural weight management program (WMP) as a pre‐requisite for the surgical treatment. The goal of pre‐operative physician‐led WMP is to establish positive pre‐surgical lifestyle changes, to assess and manage all medical issues, and to reinforce concomitant weight loss that is sustainable post‐operatively. However, there is little evidence to support that pre‐operative weight loss results in better outcomes after bariatric surgery. The identification of any predictive association would improve patient selection and help develop interventions targeting those most likely to benefit from bariatric surgery in the competitive publicly funded settings. It would also be useful in developing strategies to improve post‐operative weight loss. Therefore, this study investigated the relationship between pre‐operative weight loss during the WMP and the long‐ term weight loss post‐surgery, up to 7 years of post‐operative follow‐up, among the complex and severe bariatric surgical cohort. Pre‐operative contributors to post‐operative weight loss achievement were also extensively studied.

Methods 168 patients with clinically severe obesity (i.e. BMI ≥35 Kg/m2 with at least one obesity‐ associated complication) were enrolled into a one‐year publicly funded and referral‐based WMP incorporating dietary intervention, behavioural management and exercise prescription; followed by bariatric surgery between 2009 and 2017. After surgery, mandatory multidisciplinary follow‐ up visits were scheduled for 2 weeks, 4 weeks, 8 weeks, 3 months, 6 months, 1 year, 1.5 years, 2 years and annually thereafter. The median post‐operative follow‐up was 4 years. Pre‐operative and post‐operative weight loss in percentage of total weight loss (%TWL) and absolute kilogram (Kg) were reported in this study. Bariatric surgical outcome predictor variables evaluated included pre‐operative baseline age, race, sex, marital status, pre‐operative weight, employment status, government support payment recipient status, smoking habit and pre‐operative medical conditions. Pearson 2‐tailed correlation was calculated to determine if initial pre‐operative weight loss predicted the degree of weight loss after bariatric surgery. A multiple linear regression model was fitted to estimate contributions of independent variables to weight loss achievement after bariatric surgery.

204

CHAPTER 4 – Pre‐operative predicts post‐operative weight loss | Michelle M.C. Tan

Results All patients had a mean %TWL(±SD) of 9.3±10.3%, 24.4±8.7%, 23.9±11.1%, 21.5±11.7%, 21.1±12.6%, 20.6±12.8%, 17.8±12.3% and 19.7±9.5% at baseline (i.e. weight loss during the initial pre‐surgery WMP), 1, 2, 3, 4, 5, 6 and 7 years post‐operatively. Pearson 2‐tailed correlation analysis revealed a complete absence of correlation between initial %TWL and post‐operative weight loss across 1 to 7 years after bariatric surgery. Multiple linear regression models indicated that age at surgery, is a reliable predictor of post‐operative weight loss, statistically significant across years 1 to 6 post operation, but not at year 7, likely due to smaller number of observations.

Conclusion Weight loss achieved by a pre‐bariatric surgery WMP did not predict the amount of weight loss achieved by bariatric surgery, or successful weight loss maintenance after bariatric surgery in the study cohort. The present findings suggest that WMP‐mandated pre‐operative weight loss may not be essential when clearance for publicly funded bariatric surgery is evaluated. Increased age at surgery is a significant independent predictor of long‐term weight loss after bariatric surgery, implying that older patients may achieve better outcomes from bariatric surgery. Further evaluation is required on the impact of bariatric surgery on frailty, sarcopenia and cortical bone loss in this older population.

Keywords Bariatric surgery; Predictors; Pre‐operative weight loss; Post‐operative weight loss; Outcomes

4.2 INTRODUCTION

As with any surgical procedure, patient selection based on adequate risk versus benefit evaluation by an expert multidisciplinary team is critical for minimizing risks and optimising outcomes following bariatric surgery (269). In Australia, a one‐year pre‐operative weight loss program is often mandated as an eligibility criterion for publicly funded patients to educate patients regarding the lifestyle changes needed for bariatric surgery to be successful. This also allows the multidisciplinary team an opportunity to properly assess and manage all the medical issues of the complex patients, many of whom have class II or class III obesity with multiple obesity‐related comorbidities. Pre‐operative weight loss was also proposed to help assess patient adherence and assist with patient selection. This might identify patients who will comply better with the dietary restrictions and healthy exercise habits required after bariatric surgery. Weight loss preceding bariatric surgery may also reduce operative risks. On the basis that a pre‐operative

205

CHAPTER 4 – Pre‐operative predicts post‐operative weight loss | Michelle M.C. Tan weight loss requirement preceding bariatric surgery might also potentially exclude some patients who do not or refuse to lose weight, it is critical to appraise the evidence regarding whether pre‐ operative weight loss leads to improved weight outcomes after the surgery.

A few guidelines with the goal of establishing best clinical practices for bariatric surgical care have been produced by professional societies (131, 424). Much of the focus of these guidelines is the systematic identification and evaluation of an appropriate patient from medical and surgical perspectives. These guidelines also emphasize the role of the multidisciplinary, integrated health team such as physicians, dietitians, nurses, exercise physiologists and mental health professionals in the pre‐ and post‐operative care. An important element of this care is a pre‐operative weight management program (WMP), in which patients are asked to engage in a program designed to produce weight loss prior to bariatric surgical treatment and, theoretically, reduce the risk of surgical complications and maximise post‐operative weight loss (425). The relationship between pre‐operative WMP, and both surgical complications and post‐operative weight loss are as elaborated in the following table (Table 4.1), with several potential advantages and potential disadvantages outlined.

Table 4.1 Potential advantages and disadvantages of pre‐operative weight management program (WMP) (425) Potential advantages Potential disadvantages

. Opportunity to lose weight through . Inconsistent definitions, treatments, and less‐invasive means measurements of WMP . Greater post‐operative weight losses . Unnecessary dieting as bariatric surgery . Improved technical aspects of patients are already considered dieting bariatric surgery veterans . Shorter operating time . Discouraging for patients . Shorter length of hospital stay (LOHS) . Possible unnecessary delay of necessary . Opportunity to practise post‐ treatment (particularly bariatric surgery) operative behaviour changes (Adapted from Tewksbury C, Williams NN, Dumon KR, Sarwer DB. Preoperative medical weight management in bariatric surgery: A review and reconsideration. Obesity Surgery. 2017;27(1):208‐14)

206

CHAPTER 4 – Pre‐operative predicts post‐operative weight loss | Michelle M.C. Tan

While these mandates may be consistent with the popular psychology maxim that “the best predictor of future behaviour is past behaviour” (254); they may not be consistent with obesity research literature that reported mixed findings. Some studies support the concept that pre‐ operative WMP leads to greater weight loss after bariatric surgery, but many other studies did not found a clear relationship (101, 254, 256‐259, 261, 263, 264, 426‐428). In the setting of clinical severe obesity at publicly funded bariatric surgery services in Australia, the effect of one‐ year MWP on post‐operative weight loss is unknown. The data of highly complex population in long‐term is even sparse.

Therefore, the primary aim of the present study was to determine if pre‐operative weight loss predicts weight loss after bariatric surgery as an independent variable in a publicly funded bariatric surgical cohort. In the second tier, we sought to explore predictors of pre‐operative weight loss achievement. This was done via an evaluation of multiple potential pre‐operative factors to determine if pre‐operative weight loss is the best predictor of post‐operative weight loss, or another factor instead. These will help to develop a treatment algorithm for guiding clinical decision‐making.

4.3 METHODS

Patients with clinically severe obesity participated in a minimum one‐year of physician‐led WMP incorporating dietary intervention, exercise prescription and psychological support, as described in detail in the following Sections (i) to (iii). This was a comprehensive and intensive lifestyle program that aimed to thoroughly prepare patients with clinically severe obesity, multiple highly complex health issues and were deemed suitable for bariatric surgery.

At the initial visit to the clinics, all patients underwent an evaluation by a multidisciplinary team consisting of consultant endocrinologists, nurses and dietitians. Standard investigations were carried out during the pre‐operative care, including endocrinologic, nutritional and psychological

207

CHAPTER 4 – Pre‐operative predicts post‐operative weight loss | Michelle M.C. Tan evaluations. Psychosocial, functional status and support of family or friends were also carefully assessed. These indications for bariatric surgery were validated during a monthly multidisciplinary staff meeting before referral to bariatric surgeons. A specific weight loss target has not been a prerequisite for referral to the publicly funded bariatric surgery.

(i) Dietary intervention All patients’ dietary patterns were thoroughly assessed by dietitians or therapists at the first face‐ to‐face clinic visits. Under the supervision of registered dietitians at the Royal Prince Alfred (RPA) Hospital, patients participated in a structured dietary intervention during the pre‐operative period. This entailed three stages. In the first stage, patients were encouraged to be on an intensive very low energy diet (VLED) program for at least 12 weeks, particularly the super obese (SO) patients and those from interstate. Those who were on the VLED had a detailed one‐hour group education session regarding the diet, and were encouraged to attend fortnightly group follow‐up. In some cases, the duration of this intensive VLED can be extended if patient desired. For others who were not SO at initial assessment, the initial dietary advice was refined based on the patient’s current dietary habits, medical conditions, behavioural factors, social factors and so on. Recommendations were tailored to address any particular issues identified and to optimise their dietary habits.

In the second stage of the dietary intervention, patients were given group‐based nutrition advice to modify their eating patterns, encouraging a goal of 5‐10 percentage total weight loss (%TWL) prior to bariatric surgery. During this period, all patients have also received dietary counselling on a regular basis to prepare them for their subsequent post‐operative dietary phases and restrictions, such as transition from liquids to puree, and to reintroduction of textured/solid foods. The patients were closely monitored by the multidisciplinary team at least trimonthly and advice was individualised, keeping them in active treatment.

During the third stage, all patients were placed on a mandatory VLED at least two weeks prior to bariatric surgery. Same applied to another two sites, the Concord Repatriation General and

208

CHAPTER 4 – Pre‐operative predicts post‐operative weight loss | Michelle M.C. Tan

Camden Hospitals at this final stage of dietary intervention. The two‐week VLED was importantly aimed to reduce abdominal adiposity and liver volume of the study cohort to prevent large liver size from complicating the subsequent surgical access.

(ii) Supervised exercise prescription The supervised and individualised exercise program at the Concord Repatriation General and Camden Hospitals involved an exercise physiologist or physiotherapist, available five days per week in well‐equipped clinics. The total dose of individualised exercise prescribed to patients was 330 minutes of at least moderate‐intensity physical activity per week (i.e. 180 minutes of supervised exercise classes at the Concord Repatriation General Hospital or Camden Hospital’s gymnasium. In addition to a further target of 150 minutes of unsupervised home‐based exercise on off‐clinic days). Exercise intensity was established on a target heart rate response of 60‐80% of predicted maximal heart rate, i.e. approximate range of 110‐140 beats/minute. The prescribed exercises usually included 20‐30 minutes, each of aerobic exercises and resistance training activities (429).

(iii) Psychological support/Behavioural management At the Concord Repatriation General and Camden Hospitals, an onsite psychologist was available for an initial consultation and four group sessions conducted throughout the WMP. On the other hand, a co‐located psychologist conducted the similar consultation and sessions in the RPA Hospital. The initial assessment focused on screening for any major psychological and social barriers to weight loss and eating disorders. Additionally, pre‐operative review appointments provided patients with psychological treatments to manage barriers and psychological problems using a patient‐centred approach.

After the WMP and consultations with the multidisciplinary team, patients chose one of four surgical procedures in accordance to surgeon’s suggestion and decided on an individualized basis: sleeve gastrectomy (SG), mini gastric bypass‐one anastomosis gastric bypass (MGB‐OAGB), adjustable gastric banding (AGB) or roux‐en‐Y gastric bypass (RYGB). The details of surgical

209

CHAPTER 4 – Pre‐operative predicts post‐operative weight loss | Michelle M.C. Tan techniques have been extensively described in CHAPTER 2. After bariatric surgery, mandatory multidisciplinary post‐operative follow‐ups occurred at 2 weeks, 4 weeks, 8 weeks (transitioning diet phase), 3 months (diet advancement), 6 months, 1 year, 1.5 years, 2 years and annually thereafter.

Study outcome measures In our study of exclusively clinically severe obese patients, the outcome measures were collected and assessed retrospectively as a health service evaluation. This included the absolute weights of the study patients throughout the entire observation period, measured in kilogram (Kg). Initial/Pre‐operative weight loss (defined as the 12 months preceding bariatric surgery) and post‐ operative weight loss in %TWL were calculated. This calculation was for determining if initial pre‐ operative weight loss predicts post‐operative weight loss. Last recorded weight prior to bariatric surgery, specifically 1 to 14 days before the surgery was used as pre‐operative weight, if any patient’s weight was not measured at surgery. If measurement in clinics was not possible or if documented weight from our follow‐up clinics was unavailable for post‐operative windows, weight of the patient was obtained from hospital records.

For the second tier of this study, potential predictors were selected based on theoretical considerations, variable availability and previous studies (254, 265, 430, 431). Bariatric surgical outcome predictor variables assessed in this study included age at surgery, race, sex, marital status, employment status, government support payment, smoking status, pre‐operative weight, type of surgery, and baseline number of obesity‐related medical conditions. To account for multiple pre‐operative baseline comorbidities among the bariatric surgical cohort, the number of medical conditions was defined as a count of the following diagnosed obesity‐related comorbidities: prediabetes, type 2 diabetes mellitus (T2DM), hypertension, hyperlipidaemia, sleep‐disordered breathing [i.e. either obstructive sleep apnoea (OSA) or obesity hypoventilation syndrome (OHS)], osteoarthritis (OA) and/or weight‐bearing joint pain (WBJP), depression and/or severe anxiety, gastro‐oesophageal reflux disease (GORD), coronary artery disease (CAD), stroke, lymphoedema or chronic venous insufficiency, other cardiac disease, diabetic

210

CHAPTER 4 – Pre‐operative predicts post‐operative weight loss | Michelle M.C. Tan microvascular disease (including nephropathy, neuropathy and retinopathy), asthma, gout, polycystic ovary syndrome (PCOS), fatty liver disease, chronic kidney disease (CKD), and deep vein thrombosis (DVT)/pulmonary embolism (PE).

STATISTICAL ANALYSIS Statistical analyses were performed using IBM SPSS Statistics (Version 26, Armonk, NY, USA). P values <0.05 were considered statistically significant. Pearson 2‐tailed correlation test was used to determine if pre‐operative weight loss reliably predicted the degree of weight loss after bariatric surgery to 7 post‐operative years. Post‐operative weight at year 8 mark was not included in the Pearson 2‐tailed correlation analysis owing to very small number of observations.

A multiple linear regression model using stepwise selection and enter methods were constructed to study the power of the independent variables to predict post‐operative change in %TWL. The model included the following independent variables: (a) continuous variables − pre‐operative weight (Kg), and age at surgery (years); (b) categorical variables − sex (Male versus Female), race (Caucasian versus Middle Eastern versus Others), marital status (Married/Domestic partner versus Separated/Divorced/Widowed versus Single), smoking status (Current/Former smoker versus Never smoker), working status (Yes versus No), whether on government support payment (Yes versus No), type of surgery (SG versus AGB versus MGB‐OAGB versus RYGB), and total number of baseline comorbidities. The criteria probability was fitted at inclusion 0.05 and removal of 0.10, respectively.

211

CHAPTER 4 – Pre‐operative predicts post‐operative weight loss | Michelle M.C. Tan

4.4 RESULTS

A summary of baseline demographic and clinical characteristics of the study patients is provided in Table 4.2.

Table 4.2 Baseline characteristics of the study cohort (n=168) Characteristics Mean±SD or frequency (%) Age (year) Mean±SD 52.0±11.4 Range 21‐72 Age groups <40 30 (17.9%) 40–60 95 (56.5%) >60 43 (25.6%) Sex Female 111 (66.1%) Male 57 (33.9%) Race Caucasian 120 (71.4%) Middle Eastern 24 (14.3%) Other§ 24 (14.3%) Married/Cohabitation, n (%) 82 (49.7%) Smoking status Current/Former smoker 96 (57.1%) Never smoker 72 (42.9%) Employment status Employed 57 (33.9%) Unemployed 17 (10.1%) Student 8 (4.8%) Housewife 20 (11.8%) On government support paymentβ 73 (43.5%)

212

CHAPTER 4 – Pre‐operative predicts post‐operative weight loss | Michelle M.C. Tan

Characteristics Mean±SD or frequency (%) Weight±SD (range) (Kg) Initial 141.9±35.9 (83.5–272.8) Pre‐operative weight after WMP and at bariatric surgery 133.2±32.4 (80.3–239.9) BMI (Kg/m2) Initial 51.2±10.8 (34.0–83.8) Pre‐operative BMI after WMP and at surgery 48.0±9.5 (33.6–78.8) Index bariatric surgical procedure Sleeve gastrectomy (SG) 141 (83.9%) Mini gastric bypass‐one anastomosis gastric bypass 15 (8.9%) (MGB‐OAGB) Adjustable gastric banding (AGB) 11 (6.5%) Roux‐en‐Y gastric bypass (RYGB) 1 (0.6) Number of obesity‐related comorbidities at baseline Median (range) 6 (1–12) Abbreviations: SD=Standard deviation, Kg=Kilograms, BMI=Body mass index; WMP=Weight Management Program §Indigenous Australian, Pacific Islander, Americas, Black African, Mauritian, Filipino and Pakistanis βDisability support pension, Age pension, Carers pension, NewStart Allowance, Veteran pension, National Disability Insurance Scheme, Workers compensation

Also shown in Table 4.2, the largest ethnic group that receiving publicly funded bariatric surgery were Caucasians (n=120, 71.4%), followed by Middle Eastern (n=24, 14.3%), and the remaining were of other races (Indigenous Australian, Pacific Islander, Americas, Black African, Mauritian, Filipino and Pakistanis) (n=24, 14.3%). The main primary procedure that was performed during this period was laparoscopic sleeve gastrectomy (83.9%). Of the major comorbidities of interest among the study cohort, the median number of diagnoses per patient was six (ranged 1–12).

Figure 4.1 below displays the observed weight variation (Kg) of the study cohort from initial (lifestyle intervention) to pre‐operation to 8 years following bariatric surgery.

213

CHAPTER 4 – Pre‐operative predicts post‐operative weight loss | Michelle M.C. Tan

Figure 4.1 Observed mean weight change (Kg) from initial clinic visit to pre‐operative baseline through 8‐year post‐operative follow‐up 180.0 Predict

160.0

140.0 141.9

(Kg) 133.2

120.0 weight

111.7 110.6 105.9 101.7 103.2 101.5 100.0 99.8 99.1 Observed

80.0

60.0

40.0 Initial visit Pre‐op m12 m24 m36 m48 m60 m72 m84 m96

Follow‐up, months

No. of 168 168 157 129 98 79 63 42 25 11 patients Weight (Kg) Minimum 83.5 80.3 58.2 58.0 57.50 60.0 59.1 63.4 70.5 75.8 Maximum 272.8 239.9 174.10 175.0 176.80 182.2 192.3 189.9 163.7 163.7 Whiskers represent the standard deviation (SD) Abbreviation: Kg=Kilograms, m=month

Between initial and pre‐operative baseline, 168 patients attended and recorded a weight. At 1 year post‐operatively, 157 (93.5%) patients attended clinics and recorded a weight. At years 2, 3, 4, 5, 6, 7 and 8 post‐surgery, 129 (76.8%), 98 (58.3%), 79 (47. 0%), 63 (37.5%), 42 (25.0%), 25 (14.9%) and 11 (6.5%) patients attended hospitals and documented a weight. Initially, mean weight at the entrance to study clinics was 141.9 Kg, equating to a BMI of 51.2±10.8 Kg/m2. After the WMP, i.e. after the pre‐surgical weight loss requirement, patients lost a mean of %TWL(±SD) of 9.3±10.3%. After bariatric surgery, all patients had an observed overall mean %TWL(±SD) of 24.4±8.7%, 23.9±11.1%, 21.5±11.7%, 21.1±12.6%, 20.6±12.8%, 17.8±12.3% and 19.7±9.5% at 1, 2, 3, 4, 5, 6 and 7 years post‐operatively.

214

CHAPTER 4 – Pre‐operative predicts post‐operative weight loss | Michelle M.C. Tan

Table 4.3 tabulates the initial and pre‐operative baseline obesity‐related comorbidities of the study population. It can be clearly seen that at initial stage, OA and/or WBJP were the most common comorbidity among the patients, followed by hypertension, sleep‐disordered breathing and T2DM. Three patients were found died of causes unrelated to bariatric procedures during the period, of which the details of surgical complications and mortality have been reported in CHAPTER 2. Likewise, the pre‐operative baseline comorbidities of the study cohort have been extensively described in CHAPTER 2.

Table 4.3 Initial and pre‐operative baseline obesity‐related comorbidities of all patients Initial Pre‐operative baseline n (%) n (%) Type 2 diabetes mellitus (T2DM) 105 (62.5) 114 (67.9) Prediabetes† 22 (13.1) 24 (14.3) Hypertension 111 (66.1) 110 (65.5) Hyperlipidaemia 94 (56.0) 99 (58.9) Sleep‐disordered breathing (OSA/OHS) 109 (64.9) 107 (63.7) Requiring CPAP/BiPAP device 67 (39.9) 87 (51.8) OA and/or WBJP 113 (67.3) 120 (71.4) Depression and/or severe anxiety§ 86 (51.2) 79 (47.0) Gastro‐oesophageal reflux disease (GORD) 82 (48.8) 89 (53.0) †Prediabetes includes diagnosis of impaired glucose tolerance (IGT) or impaired fasting glucose (IFG) Abbreviations: OSA=Obstructive sleep apnoea; OHS=Obesity hypoventilation syndrome; CPAP=Continuous positive airway pressure; BiPAP=Bilevel positive airway pressure; OA=Osteoarthritis; WBJP=Weight‐bearing joint pain. §Due to collinearity, patient who was suffering from either ‘depression' or ‘severe anxiety’ or both were combined into one comorbidity/variable – ‘depression and/or severe anxiety’. This mental health condition was considered as present only if the patient’s general physician has confirmed the diagnosis, and if the patient was on a documented treatment.

Pearson correlation coefficient value (r) is a measure of the strength of the relationship between two studied variables, ranging from ‐1 to 1. An r of ‐1 represents a perfect negative linear relationship between the variables, whereas an r of 0 indicates no linear relationship in between, and an r of 1 means a perfect positive linear relationship between variables. While coefficient of determination (R square) is the proportion of the variance in the dependent variable that is

215

CHAPTER 4 – Pre‐operative predicts post‐operative weight loss | Michelle M.C. Tan

predictable from the independent variables. The present findings indicated complete absence of correlations between initial %TWL and the weight loss after bariatric surgery across years 1 to 7 (data not shown) (i.e. r <0.2 and R square <0.04 across all post‐operative years were observed).

Multiple linear regression models demonstrated that %TWL increased as age at surgery increased (Table 4.4). The other predictors that were added into the regression model stayed non‐ significant. The insignificant predictors comprised the following variables: pre‐operative weight, demographic variables (age at surgery, race, sex, marital status, smoking status, employment status and government support payment), type of surgery, and baseline total number of obesity‐ related comorbidities.

Table 4.4 Summary of the final prediction models of the percentage total weight loss (%TWL) at 1 to 7 years after bariatric surgery. The multiple linear regression models were based on stepwise selection method for years 1 to 6 post operation, and variable entry for year 7 after bariatric surgery Regression model Dependent R square Independent β (95% CI) p value selection method variable variable

Stepwise %TWL after 1 year Age at surgery 0.226 0.081 <0.001 (0.104‐0.348) Stepwise %TWL after 2 years Age at surgery 0.255 0.065 0.004 (0.081‐0.418) Stepwise %TWL after 3 years Age at surgery 0.237 0.056 0.021 (0.039‐0.457) Stepwise %TWL after 4 years Age at surgery 0.355 0.094 0.007 (0.101‐0.608) Stepwise %TWL after 5 years Age at surgery 0.392 0.104 0.011 (0.092‐0.692) Stepwise %TWL after 6 years Age at surgery 0.328 0.096 0.049 (0.001‐0.654) Enter %TWL after 7 years Age at surgery 0.220 0.048 0.314 (‐0.223‐0.664) All variables included in the multiple linear regression analyses were: Age at surgery (years), pre‐operative weight (Kg), sex (Male versus Female), race (Caucasian versus Middle‐Eastern vs Others§), marital status (Married/Domestic partner versus Separated/Divorced/Widowed versus Single), smoking status (Current/Former smoker versus Never

216

CHAPTER 4 – Pre‐operative predicts post‐operative weight loss | Michelle M.C. Tan

smoker), working status (Yes versus No), welfare (Yes versus No), type of surgery (SG versus AGB versus MGB‐OAGB versus RYGB), and total number of baseline comorbidities. §Others: Indigenous Australian, Pacific Islander, Americas, Black African, Mauritian, Filipino and Pakistanis. Abbreviations: %TWL=Percentage of total weight loss; CI=Confidence interval; Kg=Kilograms; SG=Sleeve gastrectomy; AGB=Adjustable gastric banding; MGB‐OAGB=Mini gastric bypass‐one anastomosis gastric bypass; and RYGB=Roux‐ en‐Y gastric bypass.

Overall, the results (Table 4.4) indicate that for each additional year of age, there is an additional of 0.226, 0.255, 0.237, 0.355, 0.392, 0.328 and 0.220 %TWL, respectively over 7 years of time. Age at surgery is a significant predictor of post‐operative %TWL at 1 to 6 years following bariatric surgery, but not at 7 years likely due to small sample size.

4.5 DISCUSSION

The present study found that pre‐operative weight loss was not associated with long‐term post‐ surgery weight loss. This finding adds to the growing body of literature that calling into question regarding the need for a minimum one‐year of mandatory WMP prior to publicly funded bariatric surgery, although there are previous other studies that found contradictory results (258, 432). For example, a systematic review and meta‐analysis of 15 articles (included a total of 3,403 patients) found a mandatory pre‐operative weight loss prior to bariatric surgery appears to be associated with a greater significant increase in post‐operative weight loss (7 of 14 studies with positive association) (258). However, there was considerable heterogeneity among the studies, including the amount of post‐operative weight loss and post‐operative follow‐up period (3 to 48 months). Further complicating the issue, the pre‐operative weight loss examined in their work was immediate, i.e. weeks preceding to surgery. Whereas Cassie and co‐authors (263) who reviewed a larger group of studies (n=27) revealed that 62.5% of them (n=15) found no beneficial effect of pre‐operative weight loss on post‐operative weight loss. These results are similar to those of another systematic review of studies that evaluated the relationship between pre‐ operative weight loss and post‐operative outcomes by Ochner and colleagues, focusing on the content and effectiveness of pre‐operative programs and post‐operative weight loss results, also concluded that these existing literature reported mixed findings and do not strongly support

217

CHAPTER 4 – Pre‐operative predicts post‐operative weight loss | Michelle M.C. Tan mandatory weight loss before bariatric surgery (257). The pre‐operative requirements or recommendations in the studies included (2 randomized controlled trials, 5 prospective studies and 6 retrospective studies) were either ranged from several weeks to 12 months or unspecified; or programs varied (257). An updated review by Gerber and colleagues (259) that included 23 studies as well as the aforementioned Levhits and Oncher’s reviews, also determined that much of the research on the impact of pre‐operative weight loss, regardless of length of dietary interventions, on post‐operative weight loss outcome is inconsistent.

Our results raised questions regarding the utility and timing of pre‐operative lifestyle interventions for bariatric surgery patients. In fact, our findings suggest that a mandatory pre‐ operative weight loss program may not be required for entry into publicly funded bariatric surgery. The present findings was further supported by another recent (published in 2020) but short‐term study from the UK (428), which found no relationship between weight loss during the pre‐operative WMP (6 to 12 months subjected to patients’ BMIs) and post‐surgery (either 12 or 24 months) in 208 patients who underwent LAGB (n=128) and RYGB (n=80). In the absence of a funding requirement in Australia and clear evidence of clinical benefits of mandated MWP weight loss, these findings do not support practice of mandating participation in a pre‐operative weight loss regimen, or attainment of an arbitrary weight loss target as a criterion required for multidisciplinary bariatric surgical candidacy.

Despite this lack of correlation, based on other researchers (431) and clinical experience of our team experts, pre‐operative preparation within a weight loss educational program may still be important in several ways if not mandated for bariatric surgical patients. Firstly, it is important that during the pre‐operative WMP care, patients are appropriately educated the management of weight, obesity‐related comorbidities, surgical complications and psychological conditions. Secondly, bariatric surgery itself may lead to drastic changes in the volume of food consumed but does not resolve difficulties faced by the patients in terms of choice, texture, and timing of food intake. In this regard, multidisciplinary healthcare team plays a vital role in this by providing counselling to patients. Furthermore, bariatric surgeons usually recommend weight loss before

218

CHAPTER 4 – Pre‐operative predicts post‐operative weight loss | Michelle M.C. Tan bariatric surgery in the publicly funded services, because of the following potential surgical benefits post operation: reduction in the liver mass, improvement of obesity‐related chronic conditions, facilitation of laparoscopic completion, minimisation of surgical risks, reduction of surgical time, and improvement of recovery (433, 434). These can be an opportunity for future studies in the setting of highly‐complex clinically severe obesity and multidisciplinary bariatric surgery services. With clinically severe obesity disproportionately affecting disadvantaged populations, it is also imperative to develop and test the most robust pre‐operative treatment strategies that could positively impact long‐term post‐operative outcomes in the bariatric surgical populations. Shifting the focus of outcome from the current measure of pre‐operative weight loss alone, to the lifestyle modification and psychoeducation as comprehensive preparation of surgery may likely be realistic and sensible approaches (257, 425). These are vital to accurately determine patients’ pre‐operative clinical needs for the development of standards of practice.

Among the multiple predictors of post‐operative weight loss, age has previously been presented as a relevant factor for weight loss after bariatric surgery (254, 430, 431, 435). Our findings support this factor as being predictive of post‐operative outcome. However, as contrast to these previous studies, the present results tend to show a trend that is in favour of increased age for greater weight loss instead. Our data suggests that entering into a publicly funded bariatric surgery service with an advancing age could be more beneficial in achieving better outcomes. With the degree of obesity and high prevalence of its associated comorbidities, in addition to older patients have shown better outcomes from bariatric surgery, increasing the age of patients for publicly funded bariatric surgery can be considered. However, these promising results in favour of advancing age should also be considered carefully owing to the principal limitation of this study that has not taken the physiologic function of elderly candidates into account. The senior citizens are more susceptible to weight loss in response to illness and the aging process itself, such as nutritional frailty, sarcopenia (characterized by progressive and generalized loss of muscle mass and strength) and cortical bone loss (436). In this case, the benefits of weight loss must be weighed against the unintentional adverse consequences of weight loss following

219

CHAPTER 4 – Pre‐operative predicts post‐operative weight loss | Michelle M.C. Tan bariatric surgery in the elderly. In the absence of available data, we suggest evaluation of the impact of bariatric surgery on frailty, sarcopenia and cortical bone loss in older population in the future. Involvement of geriatric specialists in pre‐operative, post‐operative, medical and psychosocial management of the older bariatric surgical patients may be of great importance for patient safety and optimal outcomes.

4.6 CONCLUSION

In summary, the findings of the present longitudinal study demonstrate a limited relationship between pre‐operative lifestyle interventions (i.e. the structured WMP) and the post‐operative weight loss, suggesting that the one‐year pre‐operative WMP that is presently‐mandated prior to bariatric surgery may not be necessary. These findings add to the literature, unsupportive of the obligatory cumbersome, lengthy and burdensome pre‐operative weight loss for highly‐ complex patients seeking publicly funded bariatric surgery. Pre‐operative optimizing patients for fitness of surgery and general anaesthesia may however confer general benefits on post‐ operative complications and medical conditions, thus require further validation as a necessary process. Future research should also study more pre‐operative behavioural‐ and education‐ based outcome measures rather than magnitude of weight loss alone. The current results also drew an inference that increased age is a consistent and reliable predictor of post‐operative weight loss across years 1 to 6 following bariatric surgery. Further careful assessment is required on the impact of bariatric surgery on frailty, sarcopenia and cortical bone loss in this older adult population in a longer term.

220

CHAPTER 5 – DiaRem algorithm | Michelle M.C. Tan

CHAPTER 5

RESEARCH THEME 4

Prediction of 1‐year and 5‐year diabetes remission following sleeve gastrectomy, mini gastric bypass‐one anastomosis gastric bypass and adjustable gastric banding using DiaRem score

5.1 ABSTRACT

Introduction

DiaRem is a clinical scoring system designed to predict diabetes remission after Roux‐en‐Y gastric bypass (RYGB). To date, there is no score used for predicting diabetes remission following any bariatric surgical procedure in Australia. Additionally, there is little known about the performance of DiaRem in predicting diabetes remission following sleeve gastrectomy (SG), mini gastric bypass‐one anastomosis gastric bypass (MGB‐OAGB) and adjustable gastric banding (AGB) in longer‐term. We examined 1‐year and 5‐year post‐operative diabetes remission predictions by DiaRem score following SG, MGB‐OAGB and AGB in our study cohort at three public hospitals.

Methods We retrospectively collected and validated data from 114 patients with type 2 diabetes mellitus (T2DM) and class III obesity who underwent publicly funded bariatric surgical procedures, and determined diabetes remission status 1‐ and 5‐year post‐operatively according to pre‐operative DiaRem profiles. Age, T2DM treatment [oral hypoglycaemic agents (OHAs)/glucose‐lowering drugs and insulin therapy], and plasma haemoglobin A1c (HbA1c) were included to compute the DiaRem scores. Area Under the Receiver Operating Characteristic (ROC) Curve (AUC) was fitted to assess the discriminative ability of the DiaRem to distinguish between patients who did and did not remit their T2DM. Logistic regression analysis was performed to determine the predictive association between DiaRem scores and T2DM remission, while calibration was evaluated by Hosmer‐Lemeshow goodness‐of‐fit test.

Results At 1 year and 5 years after surgery, 52.3% and 45.2% T2DM remission was observed, respectively. DiaRem scores performed well in discriminative capacity, predictive ability and calibration in the study population. For year‐1 follow‐up post‐operatively, AUC was 0.869 (95% CI=0.800–0.938), with the most optimal cut‐off score being ≤12, sensitivity of 94.8% and specificity of 64.2%. The discriminative ability is comparatively high for those with 5‐year post‐operative diabetes remission [AUC=0.835 (95% CI=0.714–0.956), optimal cut‐off score ≤12, sensitivity=89.5% and specificity=52.2%]. Logistic regression analyses yielded models that demonstrate DiaRem score

221

CHAPTER 5 – DiaRem algorithm | Michelle M.C. Tan reliably predicted diabetes remission 1 year [OR (95% CI)=0.733 (0.655‐0.821), p<0.001] and 5 years following bariatric surgery [OR (95% CI)=0.753 (0.623‐0.909), p=0.003]. Low DiaRem scores in the study cohort pointed to highly expected post‐operative T2DM remissions. The Hosmer‐ Lemeshow goodness‐of‐fit test results indicated good fits of our DiaRem prediction models for both 1 and 5 years post‐surgery, thus accurate models to use.

Conclusions This is the first comprehensive external validation of prediction of T2DM remission of DiaRem scoring system in Australia and publicly funded bariatric care models. This study provides evidence that DiaRem is a useful and practical tool to help clinicians with selection and prioritisation of patients with T2DM and seeking for bariatric surgery in the publicly funded services.

Keywords Diabetes remission, Bariatric surgery, Prediction, Validation, DiaRem, Short‐term, Long‐term

5.2 BACKGROUND

Bariatric surgery has emerged as the most effective treatment for major and sustained weight loss with amelioration of obesity‐associated diseases, such as improving and even reversing type 2 diabetes mellitus (T2DM). Over the past decade, the surgical treatment options for patients with T2DM have grown. The ability to distinguish between patients for whom surgery will and will not induce remission of T2DM can help to facilitate a more tailored approach to treatment decisions and inform guidelines determining patient eligibility for bariatric surgery. Therefore, there is a growing need for tools that could better predict long‐term outcomes, particularly the resolution of major obesity‐related comorbidities, such as T2DM. These tools could improve clinical decisions and help set realistic outcome expectations, which are important for patients and healthcare systems.

Predictive tools for diabetes remission after bariatric surgery — predominantly following Roux‐ en‐Y gastric bypass (RYGB) have been suggested (224, 225, 227). Among which, DiaRem scoring system is one of the easiest to implement and the most reported. DiaRem score was derived on a retrospective American Caucasian cohort of 690 patients with T2DM with an average body mass

222

CHAPTER 5 – DiaRem algorithm | Michelle M.C. Tan index (BMI) of 49.4 Kg/m2 and who underwent RYGB. It has shown to exhibit an acceptable predictive power for diabetes remission 1‐year post‐surgery (224). Two years later, the authors further established an electronic health record review of a subset of the original validation study patients (n=407, mean BMI of 48.5 Kg/m2) and demonstrated that the DiaRem score also predicts diabetes remission in the cohort lasting at least 5 years after surgery, despite no validation was performed (437).

The DiaRem algorithm (224) has also shown to be capable of predictive performance superior to other scores (232), as well as predicting diabetes remission following other types of bariatric surgeries (228, 231). The score is established as an easily‐implemented tool in clinical practice, since it relies on simple pre‐operative patient characteristics and basic clinical parameters [age,

BMI, HbA1c, and the use of insulin therapy and oral hypoglycaemic agents (OHAs)]. Rather than on less‐conventional or non‐standard biomarkers, like C‐peptide that is included in the ABCD score (225) (which has been detailed in CHAPTER 1 ‐ Section 1.5.2). Importantly, the ABCD score that includes duration of diabetes was developed in Asian populations with a lower mean BMI (36.5 Kg/m2), whereas our study population was predominantly Caucasians with clinically severe obesity, indicates insufficient evidence and suitability to utilise this score in our bariatric surgical study cohort. Nevertheless, the following issues may have influenced the universal utility of the DiaRem score: (i) DiaRem was mainly demonstrated to be predictive of diabetes remission 1‐year after RYGB. Its predictive capacity for longer‐term diabetes remission is limited and controversial (229, 438, 439). (ii) Although DiaRem performs well at the extreme score values (low values strongly predict diabetes remission and vice versa), its performance in the middle score range was reported as suboptimal (440). (iii) Other than RYGB, limited information is available regarding ability of DiaRem to predict diabetes remission after other bariatric procedures, such as the more widely‐performed procedures nowadays − SG, MGB‐OAGB and AGB. These limitations have raised questions about the applicability of DiaRem to other ethnic cohorts, longer‐term use and other bariatric surgery types.

223

CHAPTER 5 – DiaRem algorithm | Michelle M.C. Tan

Together, the anticipated increasing number of bariatric surgery procedures and the uncertainty in predicting clinical outcomes, both short‐ and long‐term, emphasise the need to identify a useful and clinically applicable tool to predict bariatric surgery outcomes, especially diabetes remission. To date, there is no method for predicting diabetes remission outcomes following any bariatric procedures commonly performed in Australia, including DiaRem score. While more research is needed in this field to reliably predicting diabetes remission, we designed this study to identify and validate the short‐term and longer‐term (1‐ and 5‐year) post‐operative diabetes remission prediction by DiaRem score following laparoscopic SG, MGB‐OAGB and AGB with our patient cohort.

5.3 RESEARCH DESIGNS AND METHODS

Study design and settings A longitudinal retrospective study was undertaken among the publicly funded bariatric patients at the Royal Prince Alfred Hospital, Concord Repatriation General Hospital and Camden Hospital at the Sydney Metropolitan Area of New South Wales (NSW) state, Australia. Informed consent was obtained from all eligible patients. Ethics approval was obtained from local Research Ethics Committee at each of the participating sites.

Patient selection and data collection We included consenting adult patients (18 years of age or older) who had a BMI greater than 35 Kg/m2 at baseline, a diagnosis of T2DM according to the definition set out by the American Diabetes Association (ADA), in addition to at least one other obesity‐related comorbidity, who underwent bariatric surgery between 2009 and 2017. A complete pre‐operative work up and post‐operative follow‐up investigations were done. Only patients who had post‐bariatric surgery follow‐up of at least one year were included. Demographic and clinical variables were collected from hospitals’ medical records.

224

CHAPTER 5 – DiaRem algorithm | Michelle M.C. Tan

Definition and remission of type 2 diabetes mellitus (T2DM) The definitions of T2DM and its remission were modified from the ADA guidelines (441). T2DM was defined by a HbA1c ≥6.5% or a fasting blood glucose (FBG) ≥7 mmol/L. All pre‐operative medication including sulphonylureas, insulin sensitizing agents (biguanides and thiazolidinediones), dipeptidyl peptidase 4 (DPP4) inhibitors, glucagon‐like peptide‐1 (GLP‐1) receptor agonists, sodium‐glucose transport protein 2 (SGLT2) inhibitors, α‐glucosidase inhibitors, insulin, or combinations of these before bariatric surgery were taken into account for defining

T2DM. T2DM remission was based on HbA1c level of <6.5%, FBG concentration of less than 7 mmol/L, and when the patient has been off antidiabetic medication for least 12 months.

225

CHAPTER 5 – DiaRem algorithm | Michelle M.C. Tan

Procedures

Patients who underwent bariatric surgery 2009‐2017 (n=168)

Non‐diabetic (including pre‐diabetic) (n=53) and type 1 diabetic (n=1) preoperatively were excluded

Type 2 diabetic patients (n=114) (SG=92; MGB‐OAGB=13; AGB=9)

Baseline DiaRem data available (n=114)

Baseline DiaRem and 1‐year post‐ Baseline DiaRem and 5‐year post‐

operative diabetes status measures operative diabetes status measures

available (n=111) available (n=42)

Outcome: Diabetes Remission Outcome: Diabetes Remission

1‐year post‐operatively (n=58) 5 year post operatively (n=19) ‐ ‐ (52.3%) (45.2%)

Figure 5.1 Flow chart describing the patient selection strategy for the study cohort

226

CHAPTER 5 – DiaRem algorithm | Michelle M.C. Tan

The DiaRem scoring system The DiaRem score was calculated for each eligible patient using four baseline variables as described by Still et al. (224), as summarised in Table 5.1 and Figure 5.2.

Table 5.1 DiaRem scoring system (224) Variable Score Age (Year) <40 0 40‐49 1 50‐59 2 ≥60 3 HbA1c (%) <6.5 0 6.5‐6.9 2 7.0‐8.9 4 ≥9.0 6 Anti‐diabetes drugs No 0 Yes 3 Insulin therapy No 0 Yes 10 Overall score (sum of the four components) 0‐22 Probability of T2DM remission in each DiaRem score subgroup 0‐2 Higher probability of T2DM remission 88‐99% 3‐7 64‐88% 8‐12 23‐49% 13‐17 11‐33% 18‐22 Lower probability of T2DM remission 2‐16% Abbreviations: T2DM=Type 2 diabetes mellitus; HbA1c=Glycated haemoglobin (adapted from Still CD, Wood GC, Benotti P, Petrick AT, Gabrielsen J, Strodel WE, et al. Preoperative prediction of type 2 diabetes remission after Roux‐en‐Y gastric bypass surgery: a retrospective cohort study. The Lancet Diabetes & Endocrinology. 2014;2(1):38‐45)

227

CHAPTER 5 – DiaRem algorithm | Michelle M.C. Tan

The DiaRem score spans from 0 to 22 was further divided into five subgroups corresponding to probability ranges for T2DM remission (2% to 99%) (Table 5.1 and Figure 5.2). A low DiaRem score predicts a high chance of diabetes remission, and a high DiaRem score predicts non‐ diabetes remission.

Figure 5.2 Distribution of the DiaRem scoring system The total score is calculated by adding the points for each of the DiaRem components.

DATA ANALYSIS

Demographic and clinical variables were reported at baseline, 1‐year and 5‐year follow‐ups after bariatric surgery using descriptive statistics. Fisher's exact test was adopted to determine the effect of an increment in the DiaRem score groups (categorical data) on post‐operative T2DM remission following 1 year and 5 years of bariatric surgery, respectively.

228

CHAPTER 5 – DiaRem algorithm | Michelle M.C. Tan

The Area Under the Curve (AUC) from the Receiver Operating Characteristic (ROC) curve, also referred to as the c‐statistic, along with 95% confidence intervals (CIs), were derived to determine the discriminative value of the DiaRem score between remitted and non‐remitted T2DM cases at year‐1 and year‐5 post‐operation. A score has a low discriminative ability when it generates an AUC value between 0.5 and 0.7, moderate or good discriminative ability when its value is between 0.7 and 0.9, and outstanding discriminative ability when its value is greater than or equal to 0.9 (442). The sensitivity and specificity for each potential threshold of the DiaRem scores was evaluated to determine the optimal DiaRem cut‐off scores in predicting T2DM remissions. The sensitivity describes the ability of DiaRem scoring system to detect a true positive, reflecting its capacity to correctly identify all patients with T2DM in remission. The specificity is defined as the capability of the DiaRem to detect a true negative, correctly identifying patients who do not have their T2DM remitted.

Next, to better understand the predictive ability of DiaRem score towards T2DM remission, we further performed multivariate logistic regression analyses side‐by‐side with the ROC analysis, and DiaRem score was entered independently into the logistic regression models as a single independent variable.

Calibration measures the model accuracy, i.e. agreement between predicted and observed outcomes. In this study, calibration was statistically assessed based on the most widely used method, the Hosmer‐Lemeshow goodness‐of‐fit test (442). A p value greater than 0.05 is needed to conclude that there are no significant differences between the predicted and observed outcomes and therefore the model has good overall calibration.

For all inferential statistics, a p value of less than 0.05 was considered significant. All data analyses were performed using IBM SPSS Statistics Version 26.0 (IBM Corp, Armonk, NY, USA).

229

CHAPTER 5 – DiaRem algorithm | Michelle M.C. Tan

5.4 RESULTS

Patient characteristics Overall, there were 168 patients with clinical severe obesity who underwent bariatric surgical procedures between 2009 and 2017. Among them, 114 patients with T2DM and had data for the DiaRem score were included in the present study for analysis. Within this population, 111 and 42 patients with T2DM met the criteria for DiaRem modelling in the year‐1 and year‐5 post‐surgical timepoints, respectively (Figure 5.1).

The pre‐operative baseline socio‐demographics, 1‐year and 5‐year post‐surgical clinical characteristics of the study population included in the analyses are displayed in Table 5.2. At surgery, the mean age of eligible patients was 53.0±10.6 years (range 21.0–72.0 years). There was a higher prevalence of female patients than males in this study cohort (59.6%). The most common race in this cohort being Caucasian (68.4%), followed by Middle Eastern (14.9%).

Table 5.2 Socio‐demographic and clinical characteristics of publicly funded patients with type 2 diabetes mellitus (T2DM) who underwent bariatric surgery (n=114)

Pre‐surgery 1 year 5 years (n=114) post‐surgery post‐surgery (n=111) (n=42)

Age, mean (±SD) 53.0±10.6 (range) (Year) (21.0‐72.0) Sex, n (%) Male 46 (40.4%) Female 68 (59.6%) Race, n (%) Caucasian 78 (68.4%) Middle Eastern 17 (14.9%) Other (Indigenous Australian, Pacific 19 (16.7%) Islander, Americas, Black African, Mauritian and Pakistanis) Marital status, n (%)

230

Married/Domestic partner 53 (47.7%) Separated/Divorced/Widowed 48 (43.2%) Single 10 (9.0%) Weight (Kg) Mean (±SD) 131.9±32.3 (Range) (80.3–239.9) BMI (Kg/m2) Mean (±SD) 47.1±9.4 (Range) (33.9–78.8)

HbA1c (%) Mean (±SD) 7.2±1.5 6.2±1.2 6.7±1.2 (Range) (4.9–11.1) (4.3–11.4) (4.9–9.6) FBG (mmol/L) Mean (±SD) 7.6±2.8 6.1±2.1 6.8±2.3 (Range) (4.1–18.4) (3.1–14.8) (2.9–14.6) Antidiabetic medication Yes, n (%) 104 (91.2%) 47 (43.9%) 20 (47.6%) Insulin therapy Yes, n (%) 39 (34.2%) 16 (15.0%) 5 (11.9%) Glucose‐lowering drugs Yes, n (%) 102 (89.5%) 40 (35.1%) 18 (15.8%) Mean number of drug (±SD) 1.4±0.8 0.5±0.7 0.7±0.9 Range (0–4) (0–3) (0–3) Number of patients who were prescribed the following number of glucose‐lowering drugs: 0 12 (10.5%) 67 (62.6%) 24 (57.1%) 1 56 (49.1%) 27 (25.2%) 10 (23.8%) ≥2 46 (40.4%) 13 (12.1%) 8 (19.0%) Number of prescriptions of the classes of glucose‐lowering drugs: Biguanides 100 30 15 Sulphonylureas 17 3 4 DPP4 inhibitors 15 13 5 GLP‐1 receptor agonists 12 6 1 SGLT2 inhibitors 7 2 3 Other (α‐Glucosidase inhibitor and 3 ‐ 1 TZD)

231

CHAPTER 5 – DiaRem algorithm | Michelle M.C. Tan

Abbreviations: BMI=Body mass index; SD=Standard deviation; HbA1c=Glycated haemoglobin; FBG=Fasting blood glucose; DPP4=Dipeptidyl peptidase 4; GLP‐1=Glucagon‐like peptide‐1; SGLT2=Sodium‐glucose transport protein 2; TZD=Thiazolidinedione. Note: Any missing data are missing completely at random (MCAR)

The mean weight was 131.9±32.3 Kg and congruent with the patients studied by Still et al. (224) (mean BMI=49.4±8.3 Kg/m2), our patients had a mean BMI of 47.1±9.4 Kg/m2 at baseline. With

reference to glycaemic control, the average baseline HbA1c level was 7.2±1.5% (range 4.9–11.1%). Three patients were on insulin therapy exclusively, while 36 (31.6% of all eligible patients) were on both oral antidiabetic medication and insulin. A total of 102 patients (89.5%) were on glucose‐ lowering drugs at baseline, accounting for a mean number of 1.4±0.8 medications (range: 0–4 medications). At 1‐year and 5‐year follow‐up, the use of treatment with insulin and glucose‐ lowering drugs decreased significantly (p<0.05).

DiaRem scoring and diabetes remission Table 5.3 represents the rates of T2DM remission for the study cohort at 1‐ and 5‐year post‐ operation, stratified by the DiaRem score groups. The detailed breakdown of each DiaRem components of the patients are also presented in the same table. The results highlight that majority of the patients fell within the 3‐7 and 18‐22 DiaRem score ranges.

Table 5.3 Proportions of patients with T2DM remission according to their DiaRem score categories Baseline patients Remission rate of Remission rate of with T2DM T2DM at 1 year T2DM at 5 years (n=114) (%) post‐surgery (of % post‐surgery (of % baseline patients) baseline patients) (n=111) (n=42) DiaRem total score Mean of total score (±SD) 10.4±6.7 Classification by total scores, n (%) 0–2 4 (3.5%) 4 (100%) ‐ 3–7 53 (46.5%) 41 (77.4%) 15 (75.0%) 8–12 18 (15.8%) 10 (58.8%) 2 (25.0%)

232

13–17 11 (9.6%) 3 (27.3%) 2 (40.0%) 18–22 28 (24.6%) 0 (0.0%) 0 (0.0%) Total number of patients with 0 58 19 T2DM in remission DiaRem components Age (Year) <40 16 (14.0%) 40‐49 21 (18.4%) 50‐59 41 (36.0%) ≥60 36 (31.6%)

HbA1c (%) <6.5 44 (38.6%) 6.5‐6.9 16 (14.0%) 7.0‐8.9 40 (35.1%) ≥9.0 14 (12.3%) Glucose‐lowering drugs Yes 102 (89.5%) No 12 (10.5%) Insulin therapy Yes 39 (34.2%) No 75 (65.8%)

Abbreviations: T2DM=Type 2 diabetes mellitus; HbA1c=Glycated haemoglobin.

1‐Year outcomes after bariatric surgery At the 1‐year follow‐up, 58 out of 111 patients (52.3%) achieved remission of T2DM. In line with the literature (224) listed in Table 5.1 and Figure 5.2, our cohort with lower DiaRem scores had a higher success rate of T2DM remission. Precisely, 100% of those with DiaRem score 0‐2 had remission of T2DM, 77.4% of those with score 3‐7, 58.8% of those with score 8‐12, 27.3% with those with score 13‐17 and 0% in those with score 18‐22 as depicted in Table 5.3. The proportions of patients with T2DM remission predictably and consistently declined in the groups of patients with higher DiaRem score categories as demonstrated by the fisher’s exact test (p<0.001).

233

CHAPTER 5 – DiaRem algorithm | Michelle M.C. Tan

5‐Years outcomes after bariatric surgery At 5 years after bariatric surgery, remission of T2DM was seen in 19 patients (45.2%). As denoted in Table 5.3, 75.0% with DiaRem score 3‐7 had remission of T2DM at 5 years after bariatric surgery, followed by 25.0% and 40.0% in those with scores 8‐12 and 13‐17, respectively. Likewise, patients with the highest DiaRem score of 18‐22 were not likely to remit their T2DM. The fisher’s exact test further confirmed the negative association between the DiaRem score groups and T2DM remission (p<0.001).

DiaRem performance as a predictor of T2DM remission

Area Under the Curve (AUC) from the Receiver Operating Characteristic (ROC) analysis Discriminative ability was assessed based on the Area Under the Curve (AUC), sensitivity and specificity. On applying the visual Receiver Operating Characteristic (ROC) analysis, the AUC generated provides us with the sensitivity and specificity of the DiaRem scoring system for predicting the remission of T2DM. Figures 5.3(a) and (b) present the 1‐year and 5‐year findings of AUC and its predictive ability of T2DM remission following bariatric surgery.

The large AUC of 0.869 (95% CI=0.800–0.938) indicates that the DiaRem score has a good discriminative ability in predicting T2DM remission 1 year following bariatric surgery. It was also noted that a DiaRem cut‐off score of ≤12 had the most optimal sensitivity and specificity 94.8% and 64.2%, respectively to predict the remission of T2DM at year 1. By way of further explanation, a DiaRem cut‐off score of 12 would correctly predict the true positive of T2DM remission by 94.8% at year‐1 following bariatric surgery. While 64.2% of the true negative outcome (non‐remitted T2DM cases) will correctly be specified as non‐remitted.

234

CHAPTER 5 – DiaRem algorithm | Michelle M.C. Tan

AUC = 0.869

Figure 5.3(a) Area under the Receiver Operating Characteristic (ROC) curve (AUC) for diagnostic value of DiaRem score as a predictor of T2DM remission at 1‐year post bariatric surgery (n=111). The nonparametric ROC plot was applied. This diagnostic test is used to classify the accuracy of the prediction model and the closer the AUC to a value of 1, the more accurate is the model. Sensitivity is shown on the y‐axis and specificity on the x‐axis.

In our study, the optimal DiaRem cut‐off point was similar for 5th year post‐surgery. DiaRem score of ≤12 best predicting year‐5 remission of T2DM, with 89.5% sensitivity and 52.2% specificity, with an AUC of 0.835 (95% CI=0.714–0.956). This indicates a moderate to high discriminant ability between those whose T2DM went into remission and those with persisting T2DM.

235

CHAPTER 5 – DiaRem algorithm | Michelle M.C. Tan

AUC = 0.835

Figure 5.3(b) Area under the Receiver Operating Characteristic (ROC) curve (AUC) for DiaRem score in predicting T2DM remission at 5‐year post bariatric surgery (n=42). The nonparametric ROC plot was applied. This diagnostic test is used to classify the accuracy of the prediction model and the closer the AUC to a value of 1, the more accurate is the model. Sensitivity is shown on the y‐axis and specificity on the x‐axis.

Collectively, the diagnostic values of DiaRem scores for T2DM remission between year 1 and year 5 after surgery using the values of AUC, sensitivity and specificity were similar.

Multivariate binary logistic regression models The logistic regression analyses with DiaRem score (continuous data) as a single independent variable further proven the strong predicting ability of DiaRem score in both year 1 and year 5 follow‐up in the bariatric surgical cohort (Table 5.4):

236

CHAPTER 5 – DiaRem algorithm | Michelle M.C. Tan

Table 5.4 Binary logistic regression models of T2DM remission at 1 and 5 years after bariatric surgery with DiaRem score as the only independent variable Dependent variable Independent B Odds Ratio (95% CI) p value variable

Year‐1 T2DM DiaRem score ‐0.310 0.733 (0.655‐0.821) <0.001*** remission following bariatric surgery Year‐5 T2DM DiaRem score ‐0.284 0.753 (0.623‐0.909) 0.003** remission following bariatric surgery *p<0.05, **p<0.01, ***p<0.001

Calibration The calibration of the DiaRem scoring system that was manifested using the Hosmer‐Lemeshow goodness‐of‐fit test, indicated that both post‐operative 1‐year (p=0.748) and 5‐year models (p=0.486) in our study demonstrated good fits statistically (p>0.05), thus excellent model accuracy.

5.5 DISCUSSION

DiaRem provides acceptable predictability and has been internally (228) and externally validated in several countries, including Canada (443), United Kingdom (440), Taiwan (444), China (445), Israel (229), India (446) and Brazil (447). Nevertheless, evidence on the usefulness and accuracy of DiaRem score to predict diabetes remission above 1 year post‐operatively is limited, and only sparse data on other bariatric procedures have been published other than RYGB. To the best of our knowledge, no study has validated the DiaRem risk prediction tool of diabetes remission after bariatric surgery in Australia. The present study is the first to address the gap and validated the performance of DiaRem scoring system in Australia to predict diabetes remission as well as publicly funded bariatric surgery patients with T2DM. Furthermore, we are also the first few who accurately predicting not only short‐term (1‐year) diabetes remission, but also extended the

237

CHAPTER 5 – DiaRem algorithm | Michelle M.C. Tan application to longer‐term (5‐years) important outcomes of SG, MGB‐OAGB and AGB in predominantly Caucasian cohort.

Performance of a clinical prediction model like the DiaRem tool requires precise statistical evaluation, often through discrimination power, predictive ability and calibration. Discriminative power refers to the score’s ability to distinguish between the positive and negative cases, whereas the calibration refers to how similar is the predicted and actual observed outcomes. However, a systematic review of clinical prediction models for diabetes remission after bariatric surgery determined that amongst the studies underwent external validation, most did not look into the essential discrimination and calibration (448). Our comprehensive discriminative analysis found that DiaRem score has a reasonable discriminative power for T2DM remission following both 1 year (AUC value=0.869) and 5 years (AUC value=0.835) of bariatric surgery. Validation of the DiaRem score for use in the long‐term follow‐up assessments was carried out by a study that did not perform calibration but reported fair discriminatory power (229). Our 5‐year result is comparable and slightly better than this 5‐year post‐surgery study which demonstrated an AUC value of 0.78, 0.82 and 0.73 for RYGB, SG and AGB, respectively (229). As evidenced by larger AUC, our 1‐year ROC analysis also showed slightly better discriminatory properties than a Canadian cohort study of 1‐year post‐RYGB (443) and a recent detailed validation study conducted among Chinese patients who were follow‐up for 1‐ and 3‐years after RYGB (445), which both reported AUC values below 0.800 and greatly differ optimal cut‐off DiaRem scores. Particularly, the study done by Honarmand et al. (443) had an AUC of 0.776 (optimal DiaRem cut‐ off of <5, a sensitivity 71.8% and specificity of 71.3%). Whereas Kam et al. reported a 1‐year AUC of 0.700 (optimal cut‐off value ≤9, sensitivity=54.0% and specificity=76.2%) and 3‐year AUC of 0.790 (optimal cut‐off value ≤10, sensitivity=74.6% and specificity=73.6%), respectively.

Furthermore, not many previous studies have explored the optimal DiaRem cut‐off scores for maximal patient benefit in terms of diabetes remission. To optimize the use of the DiaRem score for patient selection and prioritization among those in the queue for publicly funded bariatric surgery, determining the best cut‐off score most predictive of DM remission is of high clinical

238

CHAPTER 5 – DiaRem algorithm | Michelle M.C. Tan value. In our study, we discovered that a DiaRem score of ≤12 had the most optimal sensitivity and specificity for both diabetes remission of 1‐ and 5‐years post‐surgical follow‐up, respectively. Our finding is clinically relevant and may aid healthcare practitioners to ensure optimal outcomes of bariatric surgery.

In some situations, even if the model demonstrates effective discriminative ability, its predictive ability may not. In light with this notion, we confirmed its predictive performance using the logistic regression models for year 1 and year 5 diabetes remission, with DiaRem score as the only independent variable in the modellings. Intriguingly, our findings proven that the DiaRem score has excellent predictive ability towards diabetes remission for both years post‐operatively. Notably the trends were observed towards slightly lower predictive performance with time from surgery, and this was largely consistent with another published reports which go beyond short‐ term observations (229, 438). Other studies which have conducted similar analysis at 1 to 2 years of follow‐up are also in agreement with our results that, lower DiaRem scores were associated with a greater likelihood of diabetes remission and that remission rates of patients in the lower DiaRem score categories at year‐1 follow‐up (439, 443). Our reports were also similar to those reported by the original study, Still and colleagues (224), that there were differences in diabetes remission rates among those in the higher categories also after 1 year post‐operatively.

A limited number of studies have examined the calibration of the DiaRem model. To the best of our knowledge, no study has validated the accuracy of DiaRem prediction model in Caucasian population specifically. To shade a better understanding of calibration of DiaRem score among the Australian population, we further applied the Hosmer‐Lemeshow goodness‐of‐fit test and filled this gap. Consistent with our findings, the only two other studies which also have calculated the calibration, were those among Taiwanese (444) and Chinese cohorts (445), supported that DiaRem is an accurate prediction model to use.

This study evinced a number of advantages. The DiaRem algorithm was originally created from RYGB patients only. Our study validated the prediction scoring model among patients pursuing

239

CHAPTER 5 – DiaRem algorithm | Michelle M.C. Tan

SG, MGB‐OAGB and AGB, thereby offer improved and broader range of evidence for current clinical utility. In contrast to our findings covering both short‐ and longer‐term, the major limitation of nearly all the published studies only followed‐up for 1 year or less than 3 years after bariatric surgery. This is insufficient to reflect the prediction of long‐term diabetes remission after surgery. Our analyses assessed up to 5 years of diabetes remission post‐operatively support the use of DiaRem to predict patients’ chances of experiencing short‐term and prolonged T2DM remission after surgery. The DiaRem scoring system notably also does not include BMI, one of the primary criteria used to determine bariatric surgery eligibility. BMI was excluded in the original study (224) based on the analysis ran for the development of the DiaRem that found BMI was not predictive of diabetes remission. Our verification of DiaRem adds to the mounting proof that there are other factors in addition to BMI, such as T2DM remission, should be considered when determining eligibility for bariatric surgery to ensure those likely to benefit from surgery are not denied this potentially life‐saving treatment option (228). This information addresses the ongoing debate with respect to the eligibility requirements for the competitive funded bariatric surgery in public hospitals. Besides clinic visits, the simple DiaRem score can also be used by patients at home in real‐time, if their HbA1c values are also available. Several studies have reported difficulties in applying the DiaRem score in the sense of prescription of antidiabetic medications. These include drug licencing of SGLT2 inhibitors launched in USA during the development of the DiaRem score (224), historical practice, varied practicing patterns in different geographic regions, as well as trainings of prescribing physicians (440, 446). We managed to take into account all insulin and glucose‐lowering drugs in the present study to reflect the reality of our patients and physicians in Australia.

Whilst this study has contributed meaningful validation and support to the DiaRem score, as well as patients and clinicians, there are a few limitations that should be noted. Due to the restricted resources of publicly funded bariatric surgery in Australia where only 1,000 procedures in public hospitals per year that could be subsidized, and only less than 5% subsidy for the state of NSW, our sample size appears relatively small despite covered all the patients operated from year 2009 to 2017 in NSW. Future validation with a larger sample size may be necessary to generalize the

240

CHAPTER 5 – DiaRem algorithm | Michelle M.C. Tan present findings to the entire population with T2DM in Australia. Additionally, the existing DiaRem scoring system is not adequately validated in various ethnicities and BMI ranges. Albeit similar race and BMI with the original study, our study cohorts contain no Asian, therefore, the DiaRem scoring system is not generalizable to Asian Australian as well as bariatric surgical patients with a BMI lower than 35 Kg/m2. Further studies including greater diversity in race and BMI ranges could be helpful to parse the potential ethnic and obesity‐level variations. In our study, we included complete T2DM remission according to the criteria modified from ADA as our only outcome variable, while the DiaRem score was developed based on complete + partial remission of T2DM. One argument in favour of this approach is that partial remission of T2DM may be considered merely an “improvement” in the medical condition. Complete remission of T2DM is arguably a more clinically significant outcome.

As a side note, there is no one perfect diabetes remission score. Each scoring system has its own strengths and weaknesses in predicting diabetes remission, and the same is true of the DiaRem scoring system. DiaRem algorithm does not take into account duration of diabetes, one of the strong risk factors for T2DM remission suggested by several studies (446, 449, 450), likely to be a surrogate for pancreatic function (449). On the flip side, the absence of a systematic screening process in most countries and vague symptoms associated with T2DM made a determination of the duration of diabetes imprecise (440). The DiaRem algorithm also demonstrates a major advantage of simplicity, requiring no additional clinical tests that are less frequently available in standard clinic visits, such as C‐reactive protein. The Diabetes Remission Score (DRS) (226) does not specify the percentage of T2DM remission in each of the three grades (mild, moderate and severe). This made the score difficult to compare the outcome; in opposition to the DiaRem scoring system where a low score clearly predicts a high chance of diabetes remission, and a high score predicts non‐diabetes remission, as listed in Table 5.1 and Figure 5.2. It is also important to note that both the DRS and ABCD score (225) do not include HbA1c, a vital marker of glycaemic control or chronicity of DM that DiaRem score has included. HbA1c level is a standard criterion for the diagnosis of T2DM, hence it should be easily accessible for scoring, especially during the pre‐surgery assessment. Like our study cohort, the DiaRem score was developed based on

241

CHAPTER 5 – DiaRem algorithm | Michelle M.C. Tan

American Caucasian population with a mean BMI >45 Kg/m2. Whereas the ABCD score was developed in Asian populations with a mean BMI of approximately 35 Kg/m2. This difference must be taken into consideration for future analysis of different racial groups and BMI ranges, since the Asian population tend to develop T2DM at a younger age and at a lower BMI than the Caucasian population (230, 451).

5.6 CONCLUSION

The results of this study add to the growing evidence that diabetes remission can be accurately predicted prior to bariatric surgery. Our findings suggest that the easy‐to‐use DiaRem score is a convenient, practical and accurate tool to use in daily clinical practice to help select and consult the patients with T2DM who are undergoing bariatric surgery in Australia. Not only could the DiaRem scoring method accurately predict 1‐year diabetes remission outcome following bariatric surgery, but also longer‐term of 5 years post‐operatively. With this clinical application, patients identified as having higher tendency of diabetes remission could be prioritised for the competitive publicly funded bariatric surgery.

242

CHAPTER 6 – Concluding Remarks | Michelle M.C. Tan

CHAPTER 6

CONCLUDING REMARKS

This chapter draws together the significance of the major findings of all studies, each representing its unique purpose as discussed in detail in chapters 2 to 5, respectively. The central goal of this thesis is to address the key knowledge deficiencies of long‐term (i.e. 6 to 9 years) effectiveness and safety outcomes of bariatric surgery, along with its intersection with metabolic and systemic outcomes, as well as treatment and patient management in the setting of clinically severe obesity in a multidisciplinary publicly funded bariatric surgery service covering three public hospitals. These series of longitudinal studies, framed into four main research themes were conducted using various methodologies, study designs, means and resources to optimise clinical management, treatment pathway and patient engagement, as well as validate a useful diagnostic tool in mainstream care for clinically severe obesity.

The basis for these studies stemmed from clinical observation and literature concerning the increasing prevalence and severity of obesity, particularly the clinically severe obese population requiring bariatric surgical treatment. However, only a few existing clinical practice guidelines aid physicians, bariatric surgeons, orthopaedic surgeons, nurses, dietitians, exercise physiologists and psychologists, among others in the multidisciplinary model of care for managing bariatric surgery in publicly funded healthcare systems. The pathways for managing the highest degree of obesity and its accompanying medical consequences in the highly complex cohort who often have longstanding obesity with multiple obesity‐related conditions are unclear. This is despite bariatric surgery being the most powerful means of tackling obesity and its associated diseases. Notably, evidence of the effectiveness and durability of long‐term health and safety outcomes of bariatric surgery is lacking, particularly for the newer procedures like sleeve gastrectomy (SG). This is especially alarming given the phenomena of weight regain and recurrence of metabolic diseases often seen over time, and their interconnections with age. Most evaluations of bariatric surgical outcomes have been hampered by inadequate and incomplete follow‐up. There is no

243

CHAPTER 6 – Concluding Remarks | Michelle M.C. Tan feasible tool for predicting diabetes remission following bariatric surgical procedures in Australia to prioritize patients in the competitive publicly funded bariatric surgery services. These significantly reflect a suboptimal evidence base in these domains that urge the pressing need to improve public healthcare system‐focused knowledge, understanding, diagnosis, effectiveness, durability, therapeutic strategies, patient adherence and management of bariatric surgery.

The specific key results and direct constructive implications that have been found and discussed in this thesis were categorised into four broad themes, as demonstrated below:

6.1 Main findings and clinical implications

Research theme 1 Long‐term effectiveness and durability of health impacts of bariatric surgery in patients with clinically severe obesity

The first research theme, presented in CHAPTER 2, is a retrospective cohort study that examined the long‐term effectiveness, health outcomes and safety of bariatric surgery in a highly complex clinically severe obese population. This research theme was framed from 15 various significant topics. It was conducted to help close the continued gaps by understanding the effectiveness and durability of bariatric surgery across a full‐range of health aspects in the context of clinically severe obesity at our three hospital‐based and complexity‐oriented specialist obesity clinics. We collected and assessed annual repeated measures of a complete set of study parameters for up to 10 years. By doing so, we could determine the long‐term and significant weight loss, weight maintenance, meaningful disease statuses, nutritional status and any potential surgical risks following bariatric surgery in the setting of clinically severe obesity over a minimum of 6 years. To gain an even more in‐depth understanding of the powerful impact of bariatric surgery, clear changes in statuses of obesity‐related comorbidities were further calculated. This was done by combining the laboratory results, medication use, physician’s diagnosis and physical measurements, and generated the prevalence, resolution, improvement, persistence, worsening and incidence of the diseases.

244

CHAPTER 6 – Concluding Remarks | Michelle M.C. Tan

Overall, the study bariatric surgical cohort had a mean pre‐operative BMI of 48.0 Kg/m2. Each of the patients had Edmonton Obesity Staging System (EOSS) of ≥ Stage 2 obesity at baseline, with a median of six obesity‐related comorbidities before bariatric surgery, demonstrating the high complexity of their diseases that required referral to the publicly funded bariatric surgery service in managing severe clinical complications. Adopting EOSS has readily contributed to the accurate assessment of patients’ health status according to their disease severity rather than solely relying on BMI‐defined criteria. It also lends a better support to the prioritization of the waiting list for publicly funded surgical treatment.

The most performed primary procedure was SG (83.9%), with negligibly few patients requiring a revisional bariatric procedure. One‐third of the patients experienced surgical complications, yet none were life‐threatening. Gastroesophageal reflux disease (GORD) was the commonest surgical complication in this cohort. These findings were concordant with other bariatric surgery studies where GORD worsened more frequently after SG procedure than RYGB. The three patient deaths occurred at 6, 8 and 9 years post‐operatively and were deemed unrelated to bariatric surgical operations. A significant weight loss, typically losing around 10 BMI points was achieved at 1‐year post‐surgery and maintained over 6 years of follow‐up. Notwithstanding a greater severity of body weight and obesity‐associated comorbidities, vital remission and improvements of metabolic and systemic comorbidities were seen in the study cohort. Osteoarthritis (OA) and/or weight‐bearing joint pain (WBJP) were identified as the most common obesity‐related comorbidities in this population, and the symptoms effectively improved following marked weight loss secondary to bariatric surgery at all 6 follow‐up timepoints. With the sleep‐ disordered breathing, over one‐third of study patients no longer required a CPAP device to their last observations after bariatric surgery. The prevalence of other obesity‐related comorbidities including type 2 diabetes mellitus (T2DM) and hypertension declined after bariatric surgery and remained lower than baseline as time progressed. However, this was not the case with hyperlipidaemia. Substantial cumulative rates of remission and improvement were observed among more than two‐third of study population with T2DM and hypertension at baseline. Meanwhile, hyperlipidaemia that was determined based on our strict definition generated a

245

CHAPTER 6 – Concluding Remarks | Michelle M.C. Tan result of approximately half of the patients having persisting hyperlipidaemia across all follow‐ up years. Despite the stringent assessment criteria, over one‐third of the bariatric surgical cohort achieved remission or improvement in hyperlipidaemia following surgery. Consistent with the literature, the prevalence of depression and/or severe anxiety was lower in short‐term post‐ surgery, but surpassed baseline level in later years. As for hyperuricaemia, a marked reduction in serum urate level was observed in the study cohort throughout follow‐up. Altogether, this research study has assuredly proven the long‐term benefits of bariatric surgery for weight loss and obesity‐related comorbidities, whilst also has highlighted the inherent risks involved in the vulnerable patients with clinically severe obesity in a publicly funded specialist obesity service. We also identified the meaningful changes in multiple comorbidities secondary to bariatric surgery in the long‐term by combining detailed, real‐world evidence.

This longitudinal study has created an electronic database and captured standardized data of a full spectrum of aspects concerning clinically severe obesity and its treatments based at a publicly funded bariatric surgery service covering three specialist obesity clinics. The database could meaningfully serve as a foundation for the effectiveness and safety of the multidisciplinary management of bariatric surgery; monitoring of long‐term outcomes; uniform clinical data collection and reporting; and future comparison across multiple publicly funded bariatric surgery services.

Our data challenges the traditional belief that bariatric surgery is technically difficult, resource challenging, greater surgical risk, requiring longer length of hospital stay (LOHS) and inducing more unfavourable outcome measures for patients with the highest level of obesity [i.e. super obesity (SO) (BMI ≥50 Kg/m2)], than those with a lower degree of obesity. Our evidence is based on a head‐to‐head comparison in a case‐control fashion. Stratifying patients into the two BMI categories shows equivalent benefits from bariatric surgery in the SO and morbidly obese (MO) groups. The similar gains consist of the magnitude of percent total weight loss (%TWL), probability of successful ≥20 %TWL, post‐operative obesity‐related comorbidity course, LOHS and surgical complications over 6 years of follow‐up appointments. As very few studies have

246

CHAPTER 6 – Concluding Remarks | Michelle M.C. Tan investigated at these metrics closely, our findings could expand the current literature and lead to better knowledge and prompt non‐bias recommendations for patient care. These findings also warrant future research of other populations within the publicly funded healthcare system or less severe study cohort to better characterize the SO population.

Nutritional deficiency has been a key consideration in the ever‐growing bariatric surgical research field. This prompted our investigation into the patients’ vitamin D, iron and vitamin B12 statuses. The average serum trace mineral and vitamins were at adequate levels before and after bariatric surgery. Still, a high prevalence of nutrient deficiencies prior to bariatric surgery could be seen in the study cohort, especially in vitamin D and iron. Before the surgery, vitamin D deficiency was noted in nearly one‐third of patients and decreased significantly after surgery. Contrarily, iron deficiency anaemia doubled at post‐operative year 6 mark compared to pre‐ surgery. With vitamin B12 insufficiency, low prevalence was detected before and after bariatric surgery, with no patient developing a deficiency year 5 onwards. These findings support the notion that evaluating nutritional parameters should be routine in clinical obesity services. It also suggests that patients need long‐term follow‐up after bariatric surgery to monitor for nutritional deficiencies. This could be at a specialised service or with the general practitioner (GP). If a GP or dietitian is to care for the patient, education and clinical guidelines are needed to deliver appropriate and effective care.

In summary, this study conceivably shows that bariatric surgery in public hospitals using a physician‐led multidisciplinary approach, is durable, safe and effective in managing a combination of clinically severe obesity, metabolic abnormality and other obesity‐related consequences. This study has provided reassuring, in‐depth and practical information regarding the multidisciplinary surgical management of clinically severe obesity within the cohort seeking publicly funded bariatric surgical treatment in a great duration of follow‐up. These findings that based on real‐world evidence have addressed the key clinical issues across a broad range of essential areas concerning clinically severe obesity and bariatric surgical treatment. Meanwhile, the results also add knowledge about the totality of its accompanying diseases, the burden of these diseases, other modalities for obesity (e.g. behavioural modifications), and the

247

CHAPTER 6 – Concluding Remarks | Michelle M.C. Tan multidisciplinary model of care in the Australian public healthcare system. The current results significantly contributed to a transparent and improved understanding of the management regarding complex obesity, extended to those with super obesity, multiple difficult obesity‐ related complications, and even mortality.

Research theme 2 Adherence to post‐surgical follow‐up

The second theme (CHAPTER 3) is a prospective follow‐up study that mapped out the rate of adherence to clinical follow‐up after bariatric surgery and the underlying reasons for ceasing attendance at clinic reviews. It also analysed the predictors of adherence to post‐surgery clinic reviews, and the relationship between adherence to post‐operative follow‐up and weight outcomes over 6 years. The adherence rate to follow‐up visits after bariatric surgery was high at 63.7%. During the interviews, patients pointed out that travel distance was their main reason for withdrawing from the publicly funded bariatric surgery service. Advanced statistical modellings show no significant difference in the mean weight loss between the adherent and nonadherent groups over the years. The predictive logistic regression model reveals that older and partnered patients were more likely to adhere to follow‐up care.

These findings could meaningfully guide clinical care practices for patients needing additional contacts and supports. This would encourage and likely maximise their benefits from additional assistance to potentially experience optimal outcomes. Nonetheless, our data challenge the importance of routine adherence to clinic reviews as nonadherence did not impair the weight loss results among our study patients. It is believed that nonadherence to routine clinic reviews might pose a risk of inadequate medical care, including potential surgical complications and nutrient deficiencies. There is also no single universal definition of adherence to follow‐ups that is agreed upon by researchers, healthcare providers and stakeholders. A unified definition for future research studies would be helpful.

248

CHAPTER 6 – Concluding Remarks | Michelle M.C. Tan

Research theme 3 Does pre‐operative weight loss matter?

The third theme (CHAPTER 4) explored the relationship between pre‐operative and post‐bariatric surgery weight losses. The study sought to determine the necessity of the current requirement for participation in a year‐long weight loss program before receiving publicly funded bariatric surgery in the nearer future. The result in our study cohort unveils little appreciable relations between these two parameters (pre‐operative and post‐operative weight losses), suggesting that the pre‐operative lifestyle weight management program (WMP) presently mandated before surgery may not be necessary. Nevertheless, the WMP might still be important as an opportunity to resolve medical problems and prevent post‐operative psychological issues from emerging. The WMP is also crucial to ensure patients understand the implications of bariatric surgery and its necessary lifestyle changes, and minimise potential surgical risks. Practically, weight loss is not mandatory during the WMP. Still, there also needs to have been some weight loss over this pre‐ operative time to assess patients’ ability to change and maintain behaviour for weight loss. These are some areas for future investigation. It would also be helpful to examine the most robust pre‐ operative treatment strategies that could positively impact long‐term post‐operative outcomes in the bariatric surgical patients with clinically severe obesity. Shifting the focus of outcome from the current measure of pre‐operative weight loss alone, to the lifestyle modification and psychoeducation as comprehensive preparation of surgery may likely be realistic and safe approaches.

In the subsequent tier, a multiple linear regression analysis demonstrated that advancing age is a reliable predictor of post‐operative weight loss across years 1 to 6 post‐operation. This suggests that older patients may achieve better outcomes from bariatric surgery. However, the benefits of weight loss must be weighed against the adverse consequences of weight loss following bariatric surgery in the older adult population. In the absence of available data, further careful assessment is required on the impact of bariatric surgery on frailty, sarcopenia and cortical bone loss in the elderly in the future. Involvement of geriatricians in the clinical care of the senior citizens may be essential for patient safety and optimal outcomes.

249

CHAPTER 6 – Concluding Remarks | Michelle M.C. Tan

Research theme 4 DiaRem algorithm use in bariatric surgical cohort

The last research theme (CHAPTER 5) outlines a simple algorithm that could be a practical and state‐of‐the‐art tool in predicting diabetes remission for those with clinically severe obesity and T2DM. Based on the work presented in this thesis, it was evidenced that the DiaRem algorithm performs excellently in discriminative capacity, predictive ability and calibration in the bariatric surgical cohort with T2DM. Therefore, this easy‐to‐use and affordable DiaRem score can be appreciated as a useful, reliable and accurate tool to help clinicians select and prioritise growing patients with T2DM seeking bariatric surgery in the everyday practice.

As this is the first validation of the DiaRem scoring system in Australia’s publicly funded bariatric care models, it is noteworthy that integration and utilization of this new knowledge and predictive algorithm in a population vastly different from the setting of clinically severe obesity may worth validation in other bariatric surgical population such as the less severe patients.

6.2 Summary, general limitations and future directions

Summary Collectively, the series of studies performed in this thesis support the high‐risk and well‐ characterised clinically severe obese population with long‐term effective and safe access to multidisciplinary publicly funded bariatric surgery service and highly valuable research. The studies have significantly contributed to the important knowledge, practice, health system, policy, and understanding and outcomes of the clinical applicability of bariatric surgery for clinically severe obesity. They also offer insight into future research directions and research collaboration endeavours. Specifically, the research findings in this work will better inform other researchers and clinicians who specialised in managing severe forms of obesity. Indeed, the findings provide a more robust and accurate context and real‐world population‐based evidence. This thesis offers one of the first comprehensive reports encompassing a full spectrum of aspects

250

CHAPTER 6 – Concluding Remarks | Michelle M.C. Tan around bariatric surgery in multidisciplinary clinical obesity services across a range of up‐to‐date bariatric surgical procedures. Thus, we hope the novel data asset could further serve as a foundation for standardized clinical data capturing, monitoring long‐term patient outcomes, and future comparison across multiple publicly funded bariatric surgery services.

General limitations Limitations in this work can be divided into general limitations to the overall research, as well as specific constraints to individual research themes. Specific limitations relating to findings in the studies have been systematically discussed in chapters 2 to 5. A key general limitation was practical considerations around the duration of 3.5‐year PhD tenure and time investment in data collection for the longitudinal study at three publicly hospitals across NSW. Whilst this prevented very long‐term (i.e. ≥10 years) data collection and analysis and more in‐depth exploration of other research questions, these studies were able to address the key clinical challenges, covering a broad range of areas around the intersection of health outcomes, bariatric surgery and publicly funded multidisciplinary obesity management. As such, this thesis has addressed all its primary and secondary aims, and enabled a broad range of meaningful research that drove new care strategies that will improve patient wellbeing and ease the burden on the existing care and Australian healthcare system, particularly the public sector. This thesis has also created the framework for future exploration of the clinical data, as discussed in a greater detail in section Future directions below.

Although highly applicable to other Australian states, the differences in health care systems and standards of care in our studies may have lower applicability to those of other countries. This includes type and level of multidisciplinary care, pre‐surgery preparation, post‐surgery support and accreditations. The heterogeneity of outcome definitions in bariatric surgery field have also made combination and comparison across studies difficult; work toward standardization of outcome definitions would be beneficial.

251

CHAPTER 6 – Concluding Remarks | Michelle M.C. Tan

Future directions Future endeavours in the area of obesity and its associated diseases, bariatric surgery and other adjunct treatment modalities could be aided by the following undertakings to address other current gaps: 1) Cost‐effectiveness analysis in multidisciplinary publicly funded bariatric surgery services; 2) Measuring changes in health‐related quality of life and lifestyle pre‐ and post‐surgery; 3) Development of guidelines for GPs to manage patients after bariatric surgery in the primary care setting; 4) Future use of smart innovations, and advancement in medicine and healthcare in the multidisciplinary bariatric surgery services. These include a virtual hospital to remotely monitor patients from home, augmentation to automated data entering and abstraction technology, and virtual group therapy (e.g. using mobile apps). In the age of Covid‐19 and the era of digital revolution, the digital health would be extremely practical to enhance the efficiency of healthcare delivery, reduce patient loss to follow‐up, and make patient care more precise and personalised; 5) Testing the most robust pre‐operative treatment strategies that could positively impact long‐ term post‐operative outcomes in the bariatric surgical populations. It should be skewed towards pre‐operative behavioural‐ and education‐based outcome measures rather than magnitude of weight loss alone; 6) Further careful assessment on the impact of bariatric surgery on frailty, sarcopenia and cortical bone loss in the older adult population in a longer term; 7) Expansion of the number of publicly funded bariatric surgeries in the state of NSW.

Over and above that, this thesis has created a meaningful framework for future exploration and monitoring of the detailed clinical, patient background, and laboratory data already collected for 10 years since the patients’ entrance to clinics to the finish line of the PhD study. Additionally, work within this PhD study interfacing clinical obesity services, clinical predictive tool and large data linkage have established the foundation for current and future collaborative research endeavours to achieve success. Importantly, it has also established vital scientific and clinical

252

CHAPTER 6 – Concluding Remarks | Michelle M.C. Tan collaborations that can advance research efforts. This involves researchers, physicians, bariatric surgeons, orthopaedic surgeons, biostatisticians, data scientists and other bariatric clinicians.

From this series of PhD research, several manuscripts covering a broad range of areas in addressing the key knowledge deficiencies have been conceived: 1) Long‐term effects of bariatric surgery on weight trajectories, health and safety outcomes (T2DM, hypertension, hyperlipidaemia, sleep‐disordered breathing, depression, severe anxiety, hyperuricaemia, surgical complications, and mortality); 2) Benefits of bariatric surgery before total joint arthroplasty (TJA) in patients with OA and/or WBJP; 3) Utilization of EOSS in patients with clinically severe obesity in an Australian publicly funded bariatric surgery service; 4) Nutrient deficiencies before and after bariatric surgery; 5) Is bariatric surgery safe and effective for patients with super obesity? 6) Adherence to follow‐up after bariatric surgery: predictors, reasons for loss to follow‐up and outcomes; 7) Qualifying for funded‐bariatric surgery in public hospitals: does weight loss before bariatric surgery matter? 8) DiaRem score for predicting T2DM remission following bariatric surgery procedures in Australian diabetic patients.

The collaborative approach to studying clinically severe obesity and its management in the patients undergoing bariatric surgery has opened multiple research contributions of need. At this stage, the following collaborative research efforts are being planned or underway. They link the key clinical questions, feasible methodologies and scientific breakthroughs: 1) A systematic review of the publicly funded bariatric surgery services in different countries; 2) Characterisation of metabolic syndrome in the bariatric surgical cohort; 3) Comparison of the usefulness of the DiaRem algorithm between Australian and American populations;

253

CHAPTER 6 – Concluding Remarks | Michelle M.C. Tan

4) In‐depth investigation of surgical outcomes at a large scale (i.e. national level); 5) Evaluation of risk factors for obesity‐related disorders in the setting of clinically severe obesity undergoing bariatric surgery.

Ultimately, all these findings serve as a foundation to enhance clinical practice by contributing firm evidence base and specific tools. This will help develop evidence‐based guidelines and policies for the optimal management of multidisciplinary bariatric surgery in the growing clinically severe obesity population in the long‐term.

254

REFERENCES | Michelle M.C. Tan

REFERENCES

1. Australian Bureau of Statistics. National Health Priority Areas 2010 [Available from: http://www.abs.gov.au/ausstats/[email protected]/0/5317BAD6B8EEE19ACA25757C001EED30?o pendocument. 2. Broskey NT, Obanda DN, Burton JH, Cefalu WT, Ravussin E. Skeletal muscle ceramides and daily fat oxidation in obesity and diabetes. Metabolism. 2018;82:118‐23. 3. Bray GA, Frühbeck G, Ryan DH, Wilding JP. Management of Obesity. The Lancet. 2016;387(10031):1947‐56. 4. Heymsfield SB, Wadden TA. Mechanisms, pathophysiology, and management of obesity. New England Journal of Medicine. 2017;376(3):254‐66. 5. Holderbaum M, Casagrande DS, Sussenbach S, Buss C. Effects of very low calorie diets on liver size and weight loss in the preoperative period of bariatric surgery: A systematic review. Surgery for Obesity and Related Diseases. 2018;14(2):237‐44. 6. The Australian and New Zealand Bariatric Surgery Registry Steering Committee. Bariatric Surgery Registry 2018/19 Report. Melbourne, Australia: Monash University; 2019. 7. World Health Organization (WHO). Body mass index ‐ BMI Geneva: World Health Organization (WHO); 2019 [Available from: http://www.euro.who.int/en/health‐ topics/disease‐prevention/nutrition/a‐healthy‐lifestyle/body‐mass‐index‐bmi. 8. Australian Institute of Health and Welfare. Overweight and obesity: An interactive insight. Cat. no. PHE 251. Canberra, Australia: Australian Institute of Health and Welfare (AIHW); 2019. 9. WHO Expert Consultation. Appropriate body‐mass index for Asian populations and its implications for policy and intervention strategies. The Lancet. 2004;363(9403):157‐63. 10. Australian Bureau of Statistics. National Health Survey: First results 2017‐18 (cat. no. 4364.0.55.001). Australia: Commonwealth of Australia; 2018. 11. Australian Institute of Health and Welfare (AIHW). A picture of overweight and obesity in Australia 2017. Cat. no. PHE 216. Canberra, Australia: Australian Institute of Health and Welfare (AIHW); 2017. 12. World Obesity Federation. Obesity Classification 2019 [cited 2019. Available from: https://www.worldobesity.org/about/about‐obesity/obesity‐classification. 13. Courcoulas AP, Christian NJ, Belle SH, Berk PD, Flum DR, Garcia L, et al. Weight change and health outcomes at 3 years after bariatric surgery among individuals with severe obesity. Jama. 2013;310(22):2416‐25. 14. Courcoulas AP, King WC, Belle SH, Berk P, Flum DR, Garcia L, et al. Seven‐year weight trajectories and health outcomes in the Longitudinal Assessment of Bariatric Surgery (LABS) study. JAMA surgery. 2018;153(5):427‐34. 15. Fisher DP, Johnson E, Haneuse S, Arterburn D, Coleman KJ, O’Connor PJ, et al. Association between bariatric surgery and macrovascular disease outcomes in patients with type 2 diabetes and severe obesity. Jama. 2018;320(15):1570‐82. 16. Atlantis E, Kormas N, Samaras K, Fahey P, Sumithran P, Glastras S, et al. Clinical Obesity Services in Public Hospitals in Australia: a position statement based on expert consensus. Clinical Obesity. 2018;8(3):203‐10.

255

17. Helmiö M, Victorzon M, Ovaska J, Leivonen M, Juuti A, Jaser N, et al. SLEEVEPASS: a randomized prospective multicenter study comparing laparoscopic sleeve gastrectomy and gastric bypass in the treatment of morbid obesity: preliminary results. Surgical endoscopy. 2012;26(9):2521‐6. 18. Salminen P, Helmiö M, Ovaska J, Juuti A, Leivonen M, Peromaa‐Haavisto P, et al. Effect of laparoscopic sleeve gastrectomy vs laparoscopic Roux‐en‐Y gastric bypass on weight loss at 5 years among patients with morbid obesity: The SLEEVEPASS randomized clinical trial. JAMA. 2018;319(3):241‐54. 19. Peterli R, Wölnerhanssen BK, Peters T, Vetter D, Kröll D, Borbély Y, et al. Effect of laparoscopic sleeve gastrectomy vs laparoscopic Roux‐en‐Y gastric bypass on weight loss in patients with morbid obesity: The SM‐BOSS randomized clinical trial. JAMA. 2018;319(3):255‐65. 20. Peterson K, Anderson J, Boundy E, Ferguson L, Erickson K. Rapid evidence review of bariatric surgery in super obesity (BMI≥ 50 kg/m2). Journal of General Internal Medicine. 2017;32(1):56‐64. 21. Williams‐Kerver GA, Steffen KJ, Mitchell JE. Eating pathology after bariatric surgery: An updated review of the recent literature. Current Psychiatry Reports. 2019;21(9):86. 22. Hong J, Park S, Menzo EL, Rosenthal R. Midterm outcomes of laparoscopic sleeve gastrectomy as a stand‐alone procedure in super‐obese patients. Surgery for Obesity and Related Diseases. 2018;14(3):297‐303. 23. Singla V, Aggarwal S, Singh B, Tharun G, Katiyar V, Bhambri A. Outcomes in super obese patients undergoing one anastomosis gastric bypass or laparoscopic sleeve gastrectomy. Obesity Surgery. 2019;29(4):1242‐7. 24. Plamper A, Lingohr P, Nadal J, Rheinwalt KP. Comparison of mini‐gastric bypass with sleeve gastrectomy in a mainly super‐obese patient group: First results. Surgical Endoscopy. 2017;31(3):1156‐62. 25. Ece I, Yilmaz H, Alptekin H, Yormaz S, Colak B, Yilmaz F, et al. Comparative effectiveness of laparoscopic sleeve gastrectomy on morbidly obese, super‐obese, and super‐super obese patients for the treatment of morbid obesity. Obesity Surgery. 2018;28(6):1484‐91. 26. Romero‐Corral A, Somers VK, Sierra‐Johnson J, Jensen MD, Thomas RJ, Squires RW, et al. Diagnostic performance of body mass index to detect obesity in patients with coronary artery disease. European Heart Journal. 2007;28(17):2087‐93. 27. Kennedy AP, Shea JL, Sun G. Comparison of the classification of obesity by BMI vs Dual‐ energy X‐ray Absorptiometry in the Newfoundland population. Obesity. 2009;17(11):2094‐ 9. 28. Romero‐Corral A, Somers VK, Sierra‐Johnson J, Thomas RJ, Collazo‐Clavell M, Korinek Jec, et al. Accuracy of body mass index in diagnosing obesity in the adult general population. International Journal of Obesity. 2008;32(6):959‐66. 29. Beavers KM, Beavers DP, Nesbit BA, Ambrosius WT, Marsh AP, Nicklas BJ, et al. Effect of an 18‐month physical activity and weight loss intervention on body composition in overweight and obese older adults. Obesity. 2014;22(2):325‐31. 30. Seimon RV, Gibson AA, Harper C, Keating SE, Johnson NA, Da Luz FQ, et al. Rationale and protocol for a randomized controlled trial comparing fast versus slow weight loss in postmenopausal women with obesity ‐ The TEMPO Diet Trial. Healthcare. 2018;6(3):85.

256

31. World Health Organization (WHO). Waist circumference and waist‐hip ratio: Report of a WHO expert consultation. Geneva, Switzerland: World Health Organization (WHO); 2011 8‐ 11 December 2008. Report No.: 9241501499. 32. Lohman TG, Roche AF, Martorell R. Anthropometric standardization reference manual: Human kinetics books Champaign; 1988. 33. Ross R, Berentzen T, Bradshaw AJ, Janssen I, Kahn HS, Katzmarzyk PT, et al. Does the relationship between waist circumference, morbidity and mortality depend on measurement protocol for waist circumference? Obesity Reviews. 2008;9(4):312‐25. 34. Rexrode KM, Carey VJ, Hennekens CH, Walters EE, Colditz GA, Stampfer MJ, et al. Abdominal adiposity and coronary heart disease in women. JAMA. 1998;280(21):1843‐8. 35. World Health Organization (WHO). Global Health Estimates 2016: Deaths by cause, age, sex, by country and by region, 2000‐2016. Geneva, Switzerland: World Health Organization (WHO); 2018. 36. Alford S, Patel D, Perakakis N, Mantzoros C. Obesity as a risk factor for Alzheimer's disease: Weighing the evidence. Obesity Reviews. 2018;19(2):269‐80. 37. Avgerinos KI, Spyrou N, Mantzoros CS, Dalamaga M. Obesity and cancer risk: Emerging biological mechanisms and perspectives. Metabolism. 2018;92:121‐35. 38. Borrell LN, Samuel L. Body mass index categories and mortality risk in US adults: The effect of overweight and obesity on advancing death. American Journal of Public Health. 2014;104(3):512‐9. 39. Glueck CJ, Goldenberg N. Characteristics of obesity in polycystic ovary syndrome: Etiology, treatment, and genetics. Metabolism. 2018;92:108‐20. 40. Huh JH, Kim KJ, Kim SU, Han SH, Han K‐H, Cha B‐S, et al. Obesity is more closely related with hepatic steatosis and fibrosis measured by transient elastography than metabolic health status. Metabolism. 2017;66:23‐31. 41. Jakobsen GS, Småstuen MC, Sandbu R, Nordstrand N, Hofsø D, Lindberg M, et al. Association of bariatric surgery vs medical obesity treatment with long‐term medical complications and obesity‐related comorbidities. JAMA. 2018;319(3):291‐301. 42. Klatsky AL, Zhang J, Udaltsova N, Li Y, Tran HN. Body mass index and mortality in a very large cohort: Is it really healthier to be overweight? The Permanente Journal. 2017;21. 43. Koliaki C, Liatis S, Kokkinos A. Obesity and cardiovascular disease: Revisiting an old relationship. Metabolism. 2018;92:98‐107. 44. Koren D, Taveras EM. Association of sleep disturbances with obesity, insulin resistance and the metabolic syndrome. Metabolism. 2018;84:67‐75. 45. Polyzos SA, Kountouras J, Mantzoros CS. Obesity and nonalcoholic fatty liver disease: From pathophysiology to therapeutics. Metabolism. 2018;92:82‐97. 46. Polyzos SA, Mantzoros CS. Obesity: Seize the day, fight the fat. Metabolism Clinical and Experimental. 2019;92:1‐5. 47. Suárez AL. Burden of cancer attributable to obesity, type 2 diabetes and associated risk factors. Metabolism. 2018;92:136‐46. 48. Vekic J, Zeljkovic A, Stefanovic A, Jelic‐Ivanovic Z, Spasojevic‐Kalimanovska V. Obesity and dyslipidemia. Metabolism. 2018;91:71‐81. 49. Park A. Pathophysiology and aetiology and medical consequences of obesity. Medicine. 2019;47(3):169‐74.

257

50. Hanipah ZN, Schauer PR. Bariatric surgery as a long‐term treatment for type 2 diabetes/metabolic syndrome. Annual Review of Medicine. 2020;71(1):1‐15. 51. Kim DD, Basu A. Estimating the medical care costs of obesity in the United States: Systematic review, meta‐analysis, and empirical analysis. Value in Health. 2016;19(5):602‐ 13. 52. Bischoff S, Damms‐Machado A, Betz C, Herpertz S, Legenbauer T, Löw T, et al. Multicenter evaluation of an interdisciplinary 52‐week weight loss program for obesity with regard to body weight, comorbidities and quality of life—a prospective study. International journal of obesity. 2012;36(4):614‐24. 53. Buchwald H, Avidor Y, Braunwald E, Jensen MD, Pories W, Fahrbach K, et al. Bariatric surgery: A systematic review and meta‐analysis. JAMA. 2004;292(14):1724‐37. 54. Norton L, Harrison J, Pointer S, Lathlean T. Obesity and injury in Australia: A review of the literature. Injury Research and Statistics. 2012(60). 55. Chooi YC, Ding C, Magkos F. The epidemiology of obesity. Metabolism. 2019;92:6‐10. 56. GBD Obesity Collaborators. Health effects of overweight and obesity in 195 countries over 25 years. New England Journal of Medicine. 2017;377(1):13‐27. 57. Australian Institute of Health Welfare (AIHW). Overweight and obesity: An interactive insight. Canberra, Australia: Australian Institute of Health and Welfare (AIHW); 2020. 58. O’Connor EA, Evans CV, Burda BU, Walsh ES, Eder M, Lozano P. Screening for obesity and intervention for weight management in children and adolescents: Evidence report and systematic review for the US Preventive Services Task Force. JAMA. 2017;317(23):2427‐44. 59. Australian Institute of Health and Welfare. Australia’s health 2018. Canberra, Australia: Australian Institute of Health and Welfare (AIHW); 2018. 60. Baqai N, Wilding JPH. Pathophysiology and aetiology of obesity. Medicine. 2015;43(2):73‐6. 61. Washburn RA, Szabo AN, Lambourne K, Willis EA, Ptomey LT, Honas JJ, et al. Does the method of weight loss effect long‐term changes in weight, body composition or chronic disease risk factors in overweight or obese adults? A systematic review. PloS One. 2014;9(10). 62. Gallagher R, Kirkness A, Zelestis E, Hollams D, Kneale C, Armari E, et al. A randomised trial of a weight loss intervention for overweight and obese people diagnosed with coronary heart disease and/or type 2 diabetes. Annals of Behavioral Medicine. 2012;44(1):119‐28. 63. Lantz H, Peltonen M, Ågren L, Torgerson JS. A dietary and behavioural programme for the treatment of obesity. A 4‐year clinical trial and a long‐term post‐treatment follow‐up. Journal of Internal Medicine. 2003;254(3):272‐9. 64. Morrison DS, Boyle S, Morrison C, Allardice G, Greenlaw N, Forde L. Evaluation of the first phase of a specialist weight management programme in the UK National Health Service: Prospective cohort study. Public Health Nutrition. 2012;15(1):28‐38. 65. Ryan DH, Johnson WD, Myers VH, Prather TL, McGlone MM, Rood J, et al. Nonsurgical weight loss for extreme obesity in primary care settings: Results of the Louisiana Obese Subjects Study. Archives of Internal Medicine. 2010;170(2):146‐54. 66. Wadden TA, Volger S, Sarwer DB, Vetter ML, Tsai AG, Berkowitz RI, et al. A two‐year randomized trial of obesity treatment in primary care practice. New England Journal of Medicine. 2011;365(21):1969‐79.

258

67. World Health Organization (WHO). Obesity: preventing and managing the global epidemic. Geneva: World Health Organization; 2000. Report No.: 9241208945. 68. Yannakoulia M, Poulimeneas D, Mamalaki E, Anastasiou CA. Dietary modifications for weight loss and weight loss maintenance. Metabolism Clinical and Experimental. 2019;92:153‐62. 69. Olson K, Bond D, Wing R. Behavioral approaches to the treatment of obesity. Rhode Island Medical Journal. 2017;100(3):21‐4. 70. Sweeting AN, Caterson ID. Approaches to obesity management. Internal Medicine Journal. 2017;47(7):734‐9. 71. Yumuk V, Tsigos C, Fried M, Schindler K, Busetto L, Micic D, et al. European guidelines for obesity management in adults. Obesity Facts. 2015;8(6):402‐24. 72. Jakicic JM, Rogers RJ, Davis KK, Collins KA. Role of physical activity and exercise in treating patients with overweight and obesity. Clinical Chemistry. 2018;64(1):99‐107. 73. Verheggen R, Maessen M, Green DJ, Hermus A, Hopman M, Thijssen D. A systematic review and meta‐analysis on the effects of exercise training versus hypocaloric diet: Distinct effects on body weight and visceral adipose tissue. Obesity Reviews. 2016;17(8):664‐90. 74. Yu N, Ruan Y, Gao X, Sun J. Systematic review and meta‐analysis of randomized, controlled trials on the effect of exercise on serum leptin and adiponectin in overweight and obese individuals. Hormone and Metabolic Research. 2017;49(03):164‐73. 75. Joseph MS, Tincopa MA, Walden P, Jackson E, Conte ML, Rubenfire M. The impact of structured exercise programs on metabolic syndrome and its components: A systematic review. Diabetes, Metabolic Syndrome and Obesity: Targets and Therapy. 2019;12:2395. 76. Durrer Schutz D, Busetto L, Dicker D, Farpour‐Lambert N, Pryke R, Toplak H, et al. European Practical and Patient‐Centred Guidelines for Adult Obesity Management in Primary Care. Obesity Facts. 2019;12(1):40‐66. 77. Goodpaster BH, DeLany JP, Otto AD, Kuller L, Vockley J, South‐Paul JE, et al. Effects of diet and physical activity interventions on weight loss and cardiometabolic risk factors in severely obese adults: A randomized trial. JAMA. 2010;304(16):1795‐802. 78. Gibson A, Franklin J, Pattinson A, Cheng Z, Samman S, Markovic T, et al. Comparison of very low energy diet products available in australia and how to tailor them to optimise protein content for younger and older adult men and women. Healthcare. 2016;4(3):71. 79. Harper C, Maher J, Grunseit A, Seimon R, Sainsbury A. Experiences of using very low energy diets for weight loss by people with overweight or obesity: A review of qualitative research. Obesity Reviews. 2018;19(10):1412‐23. 80. Delbridge E, Proietto J. State of the science: VLED (Very Low Energy Diet) for obesity. Asia Pac J Clin Nutr. 2006;15(suppl):49‐54. 81. Sumithran P, Proietto J. Safe year‐long use of a very‐low‐calorie diet for the treatment of severe obesity. The Medical Journal of Australia. 2008;188(6):366‐8. 82. Asher RC, Burrows TL, Collins CE. Very low‐energy diets for weight loss in adults: A review. Nutrition & Dietetics. 2013;70(2):101‐12. 83. Purcell K, Sumithran P, Prendergast LA, Bouniu CJ, Delbridge E, Proietto J. The rate of weight loss does not influence long term weight maintenance: A randomised controlled trial. Obesity Research & Clinical Practice. 2014;8:80.

259

84. Collins C. Survey of dietetic management of overweight and obesity and comparison with best practice criteria. Nutrition and Dietetics. 2003;60(3):177‐84. 85. Da Luz F, Hay P, Gibson AA, Touyz SW, Swinbourne JM, Roekenes JA, et al. Does severe dietary energy restriction increase binge eating in overweight or obese individuals? A systematic review. Obesity Reviews. 2015;16(8):652‐65. 86. Mustajoki P, Pekkarinen T. Very low energy diets in the treatment of obesity. Obesity Reviews. 2001;2(1):61‐72. 87. Seimon RV, Wild‐Taylor AL, Keating SE, McClintock S, Harper C, Gibson AA, et al. Effect of weight loss via severe vs moderate energy restriction on lean mass and body composition among postmenopausal women with obesity: The TEMPO Diet randomized clinical trial. JAMA Network Open. 2019;2(10):e1913733‐e. 88. Lean ME, Leslie WS, Barnes AC, Brosnahan N, Thom G, McCombie L, et al. Durability of a primary care‐led weight‐management intervention for remission of type 2 diabetes: 2‐year results of the DiRECT open‐label, cluster‐randomised trial. The Lancet Diabetes & Endocrinology. 2019;7(5):344‐55. 89. Maston G, Gibson AA, Kahlaee HR, Franklin J, Manson E, Sainsbury A, et al. Effectiveness and characterization of severely energy‐restricted diets in people with class III obesity: Systematic review and meta‐analysis. Behavioral Sciences. 2019;9(12):144. 90. Carbajo M, Castro MJ, Kleinfinger S, Gómez‐Arenas S, Ortiz‐Solórzano J, Wellman R, et al. Effects of a balanced energy and high protein formula diet (Vegestart complet®) vs. low‐ calorie regular diet in morbid obese patients prior to bariatric surgery (laparoscopic single anastomosis gastric bypass): A prospective, double‐blind randomized study. Nutricion Hospitalaria. 2010;25(6):939‐48. 91. Leonetti F, Campanile FC, Coccia F, Capoccia D, Alessandroni L, Puzziello A, et al. Very low‐ carbohydrate ketogenic diet before bariatric surgery: Prospective evaluation of a sequential diet. Obesity Surgery. 2015;25(1):64‐71. 92. Schouten R, van der Kaaden I, van’t Hof G, Feskens PG. Comparison of preoperative diets before bariatric surgery: A randomized, single‐blinded, non‐inferiority trial. Obesity Surgery. 2016;26(8):1743‐9. 93. Benjaminov O, Beglaibter N, Gindy L, Spivak H, Singer P, Wienberg M, et al. The effect of a low‐carbohydrate diet on the nonalcoholic fatty liver in morbidly obese patients before bariatric surgery. Surgical Endoscopy. 2007;21(8):1423‐7. 94. Brody F, Vaziri K, Garey C, Shah R, LeBrun C, Takurukura F, et al. Preoperative liver reduction utilizing a novel nutritional supplement. Journal of Laparoendoscopic & Advanced Surgical Techniques. 2011;21(6):491‐5. 95. Edholm D, Kullberg J, Haenni A, Karlsson FA, Ahlström A, Hedberg J, et al. Preoperative 4‐ week low‐calorie diet reduces liver volume and intrahepatic fat, and facilitates laparoscopic gastric bypass in morbidly obese. Obesity Surgery. 2011;21(3):345‐50. 96. González‐Pérez J, Sánchez‐Leenheer S, Delgado AR, González‐Vargas L, Díaz‐Zamudio M, Montejo G, et al. Clinical impact of a 6‐week preoperative very low calorie diet on body weight and liver size in morbidly obese patients. Obesity Surgery. 2013;23(10):1624‐31. 97. Leyba JL, Aulestia SN, Llopis SN. Laparoscopic Roux‐en‐Y gastric bypass versus laparoscopic sleeve gastrectomy for the treatment of morbid obesity. A prospective study of 117 patients. Obesity Surgery. 2011;21(2):212‐6.

260

98. Colles SL, Dixon JB, Marks P, Strauss BJ, O’Brien PE. Preoperative weight loss with a very‐ low‐energy diet: Quantitation of changes in liver and abdominal fat by serial imaging–. The American Journal of Clinical Nutrition. 2006;84(2):304‐11. 99. Collins J, McCloskey C, Titchner R, Goodpaster B, Hoffman M, Hauser D, et al. Preoperative weight loss in high‐risk superobese bariatric patients: A computed tomography‐based analysis. Surgery for Obesity and Related Diseases. 2011;7(4):480‐5. 100. Faria SL, Faria OP, de Almeida Cardeal M, Ito MK. Effects of a very low calorie diet in the preoperative stage of bariatric surgery: A randomized trial. Surgery for Obesity and Related Diseases. 2015;11(1):230‐7. 101. Van Nieuwenhove Y, Dambrauskas Z, Campillo‐Soto A, Van Dielen F, Wiezer R, Janssen I, et al. Preoperative very low‐calorie diet and operative outcome after laparoscopic gastric bypass: A randomized multicenter study. Archives of Surgery. 2011;146(11):1300‐5. 102. Mitchell JE, King WC, Courcoulas A, Dakin G, Elder K, Engel S, et al. Eating behavior and eating disorders in adults before bariatric surgery. International Journal of Eating Disorders. 2015;48(2):215‐22. 103. American Psychiatric Association. Diagnostic and statistical manual of mental disorders (DSM‐5®): American Psychiatric Pub; 2013. 104. Brown KL, LaRose JG, Mezuk B. The relationship between body mass index, binge eating disorder and suicidality. BMC Psychiatry. 2018;18(1):196. 105. Mustelin L, Bulik CM, Kaprio J, Keski‐Rahkonen A. Prevalence and correlates of binge eating disorder related features in the community. Appetite. 2017;109:165‐71. 106. Liu RH. Do behavioral interventions delivered before bariatric surgery impact weight loss in adults? A systematic scoping review. Bariatric Surgical Practice and Patient Care. 2016;11(2):39‐48. 107. Rudolph A, Hilbert A. Post‐operative behavioural management in bariatric surgery: A systematic review and meta‐analysis of randomized controlled trials. Obesity Reviews. 2013;14(4):292‐302. 108. Stewart F, Avenell A. Behavioural interventions for severe obesity before and/or after bariatric surgery: A systematic review and meta‐analysis. Obesity Surgery. 2016;26(6):1203‐ 14. 109. Pilitsi E, Farr OM, Polyzos SA, Perakakis N, Nolen‐Doerr E, Papathanasiou A‐E, et al. Pharmacotherapy of obesity: Available medications and drugs under investigation. Metabolism Clinical and Experimental. 2019;92:170‐92. 110. Chao AM, Wadden TA, Berkowitz RI. The safety of pharmacologic treatment for pediatric obesity. Expert Opinion on Drug Safety. 2018;17(4):379‐85. 111. Forse RA, Apovian CM. Bariatric surgery: What every provider needs to know. NJ, USA: SLACK Incorporated; 2016. 112. Yorke E. Bariatric surgical options and future directions. British Columbia Medical Journal. 2018;60(3):148‐50. 113. Sandoval D. Bariatric surgeries: beyond restriction and malabsorption. International Journal of Obesity. 2011;35(3):S45‐S9. 114. Neff KJ, le Roux CW. Mechanisms of action of bariatric surgical procedures. In: Agrawal S, editor. Obesity, bariatric and metabolic surgery: A practical guide. Cham: Springer International Publishing; 2016. p. 519‐27.

261

115. Mechanick JI, Apovian C, Brethauer S, Garvey WT, Joffe AM, Kim J, et al. Clinical Practice Guidelines for the perioperative nutrition, metabolic, and nonsurgical support of patients undergoing bariatric procedures – 2019 Update. Surgery for Obesity and Related Diseases. 2019;In Press. 116. Shanti H, Patel AG. Surgery for obesity. Medicine. 2019;47(3):184‐7. 117. De Luca M, Tie T, Ooi G, Higa K, Himpens J, Carbajo M‐A, et al. Mini gastric bypass‐one anastomosis gastric bypass (MGB‐OAGB)‐IFSO position statement. Obesity Surgery. 2018;28(5):1188‐206. 118. Radvinsky D, Iskandar M, Ferzli G. Bariatric surgery today: The good, the bad, and the ugly. Annals of Laparoscopic and Endoscopic Surgery. 2017;2(4). 119. Fried M, Peskova M. New approach in surgical treatment of morbid obesity: Laparoscopic gastric banding. Obesity Surgery. 1995;5(1):74‐6. 120. Healy P, Clarke C, Reynolds I, Arumugasamy M, McNamara D. Complications of bariatric surgery: What the general surgeon needs to know. The Surgeon. 2016;14(2):91‐8. 121. Shi X, Karmali S, Sharma AM, Birch DW. A review of laparoscopic sleeve gastrectomy for morbid obesity. Obesity Surgery. 2010;20(8):1171‐7. 122. American Society for Metabolic and Bariatric Surgery (ASMBS). Bariatric surgery procedures 2019 [Available from: https://asmbs.org/patients/bariatric‐surgery‐procedures. 123. Regan J, Inabnet W, Gagner M, Pomp A. Early experience with two‐stage laparoscopic Roux‐ en‐Y gastric bypass as an alternative in the super‐super obese patient. Obesity Surgery. 2003;13(6):861‐4. 124. Dapri G, Cadière GB, Himpens J. Laparoscopic seromyotomy for long stenosis after sleeve gastrectomy with or without duodenal switch. Obesity Surgery. 2009;19(4):495‐9. 125. Sebastianelli L, Benois M, Vanbiervliet G, Bailly L, Robert M, Turrin N, et al. Systematic endoscopy 5 years after sleeve gastrectomy results in a high rate of Barrett’s esophagus: Results of a multicenter study. Obesity Surgery. 2019;29(5):1462‐9. 126. Felsenreich DM, Kefurt R, Schermann M, Beckerhinn P, Kristo I, Krebs M, et al. Reflux, sleeve dilation, and Barrett’s esophagus after laparoscopic sleeve gastrectomy: Long‐term follow‐ up. Obesity Surgery. 2017;27(12):3092‐101. 127. Genco A, Soricelli E, Casella G, Maselli R, Castagneto‐Gissey L, Di Lorenzo N, et al. Gastroesophageal reflux disease and Barrett’s esophagus after laparoscopic sleeve gastrectomy: A possible, underestimated long‐term complication. Surgery for Obesity and Related Diseases. 2017;13(4):568‐74. 128. Maghrabi AH, Wolski K, Abood B, Licata A, Pothier C, Bhatt DL, et al. Two‐year outcomes on bone density and fracture incidence in patients with T2DM randomized to bariatric surgery versus intensive medical therapy. Obesity. 2015;23(12):2344‐8. 129. Chakhtoura MT, Nakhoul NN, Shawwa K, Mantzoros C, Fuleihan GAEH. Hypovitaminosis D in bariatric surgery: A systematic review of observational studies. Metabolism. 2016;65(4):574‐85. 130. Ben‐Porat T, Elazary R, Goldenshluger A, Dagan SS, Mintz Y, Weiss R. Nutritional deficiencies four years after laparoscopic sleeve gastrectomy: Are supplements required for a lifetime? Surgery for Obesity and Related Diseases. 2017;13(7):1138‐44. 131. Mechanick JI, Youdim A, Jones DB, Timothy Garvey W, Hurley DL, Molly McMahon M. Clinical practice guidelines for the perioperative nutritional, metabolic, and nonsurgical

262

support of the bariatric surgery patient—2013 update: Cosponsored by American Association of Clinical Endocrinologists, the Obesity Society, and American Society for Metabolic & Bariatric Surgery. Surgery for Obesity and Related Diseases. 2013;21(Suppl 1):S1‐S27. 132. Nudel J, Sanchez VM. Surgical management of obesity. Metabolism ‐ Clinical and Experimental. 2019;92:206‐16. 133. Parmar CD, Mahawar KK. One anastomosis (mini) gastric bypass is now an established bariatric procedure: A systematic review of 12,807 patients. Obesity Surgery. 2018;28(9):2956‐67. 134. Rutledge R, Walsh TR. Continued excellent results with the mini‐gastric bypass: Six‐year study in 2,410 patients. Obesity Surgery. 2005;15(9):1304‐8. 135. Carbajo M, García‐Caballero M, Toledano M, Osorio D, García‐Lanza C, Carmona JA. One‐ anastomosis gastric bypass by laparoscopy: Results of the first 209 patients. Obesity Surgery. 2005;15(3):398‐404. 136. Mahawar KK, Borg C‐M, Kular KS, Courtney MJ, Sillah K, Carr WR, et al. Understanding objections to one anastomosis (mini) gastric bypass: A survey of 417 surgeons not performing this procedure. Obesity Surgery. 2017;27(9):2222‐8. 137. Chevallier JM, Arman GA, Guenzi M, Rau C, Bruzzi M, Beaupel N, et al. One thousand single anastomosis (omega loop) gastric bypasses to treat morbid obesity in a 7‐year period: Outcomes show few complications and good efficacy. Obesity Surgery. 2015;25(6):951‐8. 138. Lee W‐J, Lin Y‐H. Single‐anastomosis gastric bypass (SAGB): Appraisal of clinical evidence. Obesity Surgery. 2014;24(10):1749‐56. 139. Musella M, Milone M. Still “controversies” about the mini gastric bypass? Obesity Surgery. 2014;24(4):643‐4. 140. British Obesity & Metabolic Surgery Society (BOMSS). BOMSS Position Statement on Mini Gastric Bypass (Loop or Single Anastomosis Gastric Bypass). UK: British Obesity & Metabolic Surgery Society (BOMSS); 2014. 141. Sonavane SK, Menias CO, Kantawala KP, Shanbhogue AK, Prasad SR, Eagon JC, et al. Laparoscopic adjustable gastric banding: What radiologists need to know. Radiographics. 2012;32(4):1161‐78. 142. Forsell P, Hallberg D, Hellers G. Gastric banding for morbid obesity: Initial experience with a new adjustable band. Obesity Surgery. 1993;3(4):369‐74. 143. Ibrahim AM, Thumma JR, Dimick JB. Reoperation and medicare expenditures after laparoscopic gastric band surgery. JAMA Surgery. 2017;152(9):835‐42. 144. Carandina S, Tabbara M, Galiay L, Polliand C, Azoulay D, Barrat C, et al. Long‐term outcomes of the laparoscopic adjustable gastric banding: Weight loss and removal rate. A single center experience on 301 patients with a minimum follow‐up of 10 years. Obesity Surgery. 2017;27(4):889‐95. 145. Loubnan Z., Parikh M., Kurian M. Reoperative options after gastric banding. In: Nguyen N., Brethauer S., Morton J., Ponce J., Rosenthal R., editors. The ASMBS Textbook of Bariatric Surgery. Cham: Springer; 2020. 146. Parrott J, Frank L, Rabena R, Craggs‐Dino L, Isom KA, Greiman L. American Society for Metabolic and Bariatric Surgery integrated health nutritional guidelines for the surgical

263

weight loss patient 2016 update: Micronutrients. Surgery for Obesity and Related Diseases. 2017;13(5):727‐41. 147. Bal BS, Finelli FC, Shope TR, Koch TR. Nutritional deficiencies after bariatric surgery. Nature Reviews Endocrinology. 2012;8(9):544. 148. Weng T‐C, Chang C‐H, Dong Y‐H, Chang Y‐C, Chuang L‐M. Anaemia and related nutrient deficiencies after Roux‐en‐Y gastric bypass surgery: A systematic review and meta‐analysis. BMJ Open. 2015;5(7):e006964. 149. Wu C, Wang F‐g, Yan W‐M, Yan M, Song M‐m. Clinical outcomes of sleeve gastrectomy versus Roux‐En‐Y gastric bypass after failed adjustable gastric banding. Obesity Surgery. 2019;29(10):3252‐63. 150. Vitiello A, Pilone V, Ferraro L, Forestieri P. Is the sleeve gastrectomy always a better procedure? Five‐year results from a retrospective matched case‐control study. Obesity Surgery. 2018;28(8):2333‐8. 151. Schauer PR, Bhatt DL, Kirwan JP, Wolski K, Aminian A, Brethauer SA, et al. Bariatric surgery versus intensive medical therapy for diabetes: 5‐Year outcomes. New England Journal of Medicine. 2017;376(7):641‐51. 152. Emous M, Wolffenbuttel BH, van Dijk G, Totté E, van Beek AP. Long‐term self‐reported symptom prevalence of early and late dumping in a patient population after sleeve gastrectomy, primary, and revisional gastric bypass surgery. Surgery for Obesity and Related Diseases. 2018;14(8):1173‐81. 153. Chang SH, Freeman NL, Lee JA, Stoll CR, Calhoun AJ, Eagon JC, et al. Early major complications after bariatric surgery in the USA, 2003–2014: A systematic review and meta‐ analysis. Obesity Reviews. 2018;19(4):529‐37. 154. Angrisani L, Vitiello A, Ferraro L. Comment on: Two‐stage approach is still the gold standard for super‐super obese patients (SSO) undergoing bariatric surgery. Surgery for Obesity and Related Diseases. 2019;15(1):33‐5. 155. Karlsson J, Taft C, Ryden A, Sjöström L, Sullivan M. Ten‐year trends in health‐related quality of life after surgical and conventional treatment for severe obesity: The SOS intervention study. International Journal of Obesity. 2007;31(8):1248. 156. Sjöström L, Narbro K, Sjöström CD, Karason K, Larsson B, Wedel H, et al. Effects of bariatric surgery on mortality in Swedish Obese Subjects. New England Journal of Medicine. 2007;357(8):741‐52. 157. Sjöström L. Review of the key results from the Swedish Obese Subjects (SOS) trial: A prospective controlled intervention study of bariatric surgery. Journal of Internal Medicine. 2013;273(3):219‐34. 158. Wadden TA, Chao AM, Bahnson JL, Bantle JP, Clark JM, Gaussoin SA, et al. End‐of‐trial health outcomes in Look AHEAD participants who elected to have bariatric surgery. Obesity. 2019;27(4):581‐90. 159. O’Brien PE, Hindle A, Brennan L, Skinner S, Burton P, Smith A, et al. Long‐term outcomes after bariatric surgery: A systematic review and meta‐analysis of weight loss at 10 or more years for all bariatric procedures and a single‐centre review of 20‐year outcomes after adjustable gastric banding. Obesity Surgery. 2019;29(1):3‐14. 160. Hofsø D, Fatima F, Borgeraas H, Birkeland KI, Gulseth HL, Hertel JK, et al. Gastric bypass versus sleeve gastrectomy in patients with type 2 diabetes (Oseberg): A single‐centre,

264

triple‐blind, randomised controlled trial. The Lancet Diabetes & Endocrinology. 2019;7(12):912‐24. 161. Adams TD, Davidson LE, Litwin SE, Kolotkin RL, LaMonte MJ, Pendleton RC, et al. Health benefits of gastric bypass surgery after 6 years. JAMA. 2012;308(11):1122‐31. 162. Sarkhosh K, Birch DW, Shi X, Gill RS, Karmali S. The impact of sleeve gastrectomy on hypertension: A systematic review. Obesity Surgery. 2012;22(5):832‐7. 163. Li J, Lai D, Wu D. Laparoscopic Roux‐en‐Y gastric bypass versus laparoscopic sleeve gastrectomy to treat morbid obesity‐related comorbidities: A systematic review and meta‐ analysis. Obesity Surgery. 2016;26(2):429‐42. 164. Schiavon CA, Bersch‐Ferreira AC, Santucci EV, Oliveira JD, Torreglosa CR, Bueno PT, et al. Effects of bariatric surgery in obese patients with hypertension: The GATEWAY randomized trial (Gastric Bypass to Treat Obese Patients With Steady Hypertension). Circulation. 2018;137(11):1132‐42. 165. Aftab H, Risstad H, Søvik TT, Bernklev PDT, Hewitt S, Kristinsson JA, et al. Five‐year outcome after gastric bypass for morbid obesity in a Norwegian cohort. Surgery for Obesity and Related Diseases. 2014;10(1):71‐8. 166. Heneghan HM, Huang H, Kashyap SR, Gornik HL, McCullough AJ, Schauer PR, et al. Reduced cardiovascular risk after bariatric surgery is linked to plasma ceramides, apolipoprotein‐ B100, and ApoB100/A1 ratio. Surgery for Obesity and Related Diseases. 2013;9(1):100‐7. 167. Sjöström L, Peltonen M, Jacobson P, Sjöström CD, Karason K, Wedel H, et al. Bariatric surgery and long‐term cardiovascular events. JAMA. 2012;307(1):56‐65. 168. Priester T, Ault TG, Davidson L, Gress R, Adams TD, Hunt SC, et al. Coronary calcium scores 6 years after bariatric surgery. Obesity Surgery. 2015;25(1):90‐6. 169. Aggarwal R, Harling L, Efthimiou E, Darzi A, Athanasiou T, Ashrafian H. The effects of bariatric surgery on cardiac structure and function: A systematic review of cardiac imaging outcomes. Obesity Surgery. 2016;26(5):1030‐40. 170. Benotti PN, Wood GC, Carey DJ, Mehra VC, Mirshahi T, Lent MR, et al. Gastric bypass surgery produces a durable reduction in cardiovascular disease risk factors and reduces the long‐ term risks of congestive heart failure. Journal of the American Heart Association. 2017;6(5):e005126. 171. Persson CE, Björck L, Lagergren J, Lappas G, Giang KW, Rosengren A. Risk of heart failure in obese patients with and without bariatric surgery in Sweden: A registry‐based study. Journal of Cardiac Failure. 2017;23(7):530‐7. 172. Sundström J, Bruze G, Ottosson J, Marcus C, Näslund I, Neovius M. Weight loss and heart failure: A nationwide study of gastric bypass surgery versus intensive lifestyle treatment. Circulation. 2017;135(17):1577‐85. 173. Marchesi F, Giacosa R, Reggiani V, De Sario G, Tartamella F, Melani E, et al. Morphological changes in the carotid artery intima after gastric bypass for morbid obesity. Obesity Surgery. 2017;27(2):357‐63. 174. Jamaly S, Carlsson L, Peltonen M, Jacobson P, Sjöström L, Karason K. Bariatric surgery and the risk of new‐onset atrial fibrillation in Swedish obese subjects. Journal of the American College of Cardiology. 2016;68(23):2497‐504.

265

175. Kuno T, Tanimoto E, Morita S, Shimada YJ. Effects of bariatric surgery on cardiovascular disease: A concise update of recent advances. Frontiers in Cardiovascular Medicine. 2019;6:94. 176. Al‐Rubaye H, McGlone ER, Farzaneh B, Mustafa L, Johnson M, Kayal A, et al. Roux‐en‐Y gastric bypass or sleeve gastrectomy for obstructive sleep apnea: A systematic review and meta‐analysis. Laparoscopic, Endoscopic and Robotic Surgery. 2019. 177. Wong A‐M, Barnes HN, Joosten SA, Landry SA, Dabscheck E, Mansfield DR, et al. The effect of surgical weight loss on obstructive sleep apnoea: A systematic review and meta‐analysis. Sleep Medicine Reviews. 2018;42:85‐99. 178. Ballsmider L, Vaughn A, David M, Hajnal A, Di Lorenzo P, Czaja K. Sleeve gastrectomy and Roux‐en‐Y gastric bypass alter the gut‐brain communication. Neural Plasticity. 2015;2015. 179. Springer BD, Carter JT, McLawhorn AS, Scharf K, Roslin M, Kallies KJ, et al. Obesity and the role of bariatric surgery in the surgical management of osteoarthritis of the hip and knee: A review of the literature. Surgery for Obesity and Related Diseases. 2017;13(1):111‐8. 180. Hacken B, Rogers A, Chinchilli V, Silvis M, Mosher T, Black K. Improvement in knee osteoarthritis pain and function following bariatric surgery: 5‐Year follow‐up. Surgery for Obesity and Related Diseases. 2019;15(6):979‐84. 181. Hooper JM, Deshmukh AJ, Schwarzkopf R. The role of bariatric surgery in the obese total joint arthroplasty patient. Orthopedic Clinics. 2018;49(3):297‐306. 182. Gu A, Cohen JS, Malahias M‐A, Lee D, Sculco PK, McLawhorn AS. The effect of bariatric surgery prior to lower‐extremity total joint arthroplasty: A systematic review. HSS Journal. 2019;15(2):190‐200. 183. McLawhorn AS, Levack AE, Lee Y‐y, Ge Y, Do H, Dodwell ER. Bariatric surgery improves outcomes after lower extremity arthroplasty in the morbidly obese: A Propensity Score‐ Matched analysis of a New York statewide database. The Journal of Arthroplasty. 2018;33(7):2062‐9.e4. 184. Li S, Luo X, Sun H, Wang K, Zhang K, Sun X. Does prior bariatric surgery improve outcomes following total joint arthroplasty in the morbidly obese? A meta‐analysis. The Journal of Arthroplasty. 2019;34(3):577‐85. 185. Castaneda D, Popov VB, Wander P, Thompson CC. Risk of suicide and self‐harm is increased after bariatric surgery: A systematic review and meta‐analysis. Obesity Surgery. 2019;29(1):322‐33. 186. Duarte‐Guerra LS, Gorenstein C, Paiva‐Medeiros PF, Santo MA, Neto FL, Wang Y‐P. Clinical utility of the Montgomery‐Åsberg Depression Rating Scale for the detection of depression among bariatric surgery candidates. BMC Psychiatry. 2016;16(1):119. 187. Mitchell JE, King WC, Chen JY, Devlin MJ, Flum D, Garcia L, et al. Course of depressive symptoms and treatment in the longitudinal assessment of bariatric surgery (LABS‐2) study. Obesity. 2014;22(8):1799‐806. 188. Osterhues A, von Lengerke T, Mall JW, de Zwaan M, Müller A. Health‐related quality of life, anxiety, and depression in bariatric surgery candidates compared to patients from a psychosomatic inpatient hospital. Obesity Surgery. 2017;27(9):2378‐87. 189. Dawes AJ, Maggard‐Gibbons M, Maher AR, Booth MJ, Miake‐Lye I, Beroes JM, et al. Mental health conditions among patients seeking and undergoing bariatric urgery: A meta‐analysis. JAMA. 2016;315(2):150‐63.

266

190. Roberts RE, Kaplan GA, Shema SJ, Strawbridge WJ. Are the obese at greater risk for depression? American Journal of Epidemiology. 2000;152(2):163‐70. 191. Green DD, Engel SG, Mitchell JE. Psychological aspects of bariatric surgery. Current Opinion in Psychiatry. 2014;27(6):448‐52. 192. Sjöström L, Peltonen M, Jacobson P, Ahlin S, Andersson‐Assarsson J, Anveden Å, et al. Association of bariatric surgery with long‐term remission of type 2 diabetes and with microvascular and macrovascular complications. JAMA. 2014;311(22):2297‐304. 193. Liberzon I, Abelson JL, Amdur RL, King AP, Cardneau JD, Henke P, et al. Increased psychiatric morbidity after abdominal aortic surgery: Risk factors for stress‐related disorders. Journal of Vascular Surgery. 2006;43(5):929‐34. 194. Gill H, Kang S, Lee Y, Rosenblat JD, Brietzke E, Zuckerman H, et al. The long‐term effect of bariatric surgery on depression and anxiety. Journal of Affective Disorders. 2019;246:886‐ 94. 195. Kalarchian MA, Marcus MD. Psychosocial concerns following bariatric surgery: Current status. Current Obesity Reports. 2019;8(1):1‐9. 196. Kalarchian MA, King WC, Devlin MJ, Hinerman A, Marcus MD, Yanovski SZ, et al. Mental disorders and weight change in a prospective study of bariatric surgery patients: 7 Years of follow‐up. Surgery for Obesity and Related Diseases. 2019;15(5):739‐48. 197. Mitchell JE, Crosby R, de Zwaan M, Engel S, Roerig J, Steffen K, et al. Possible risk factors for increased suicide following bariatric surgery. Obesity. 2013;21(4):665‐72. 198. Peterhänsel C, Petroff D, Klinitzke G, Kersting A, Wagner B. Risk of completed suicide after bariatric surgery: A systematic review. Obesity Reviews. 2013;14(5):369‐82. 199. Szmulewicz A, Wanis KN, Gripper A, Angriman F, Hawel J, Elnahas A, et al. Mental health quality of life after bariatric surgery: A systematic review and meta‐analysis of randomized clinical trials. Clinical Obesity. 2019;9(1):e12290. 200. Tindle HA, Omalu B, Courcoulas A, Marcus M, Hammers J, Kuller LH. Risk of suicide after long‐term follow‐up from bariatric surgery. The American Journal of Medicine. 2010;123(11):1036‐42. 201. Switzer NJ, Debru E, Church N, Mitchell P, Gill R. The impact of bariatric surgery on depression: A review. Current Cardiovascular Risk Reports. 2016;10(3):12. 202. Morgan DJ, Ho KM, Platell C. Incidence and determinants of mental health service use after bariatric surgery. JAMA Psychiatry. 2020. 203. Opozda M, Chur‐Hansen A, Wittert G. Changes in problematic and disordered eating after gastric bypass, adjustable gastric banding and vertical sleeve gastrectomy: A systematic review of pre‐post studies. Obesity Reviews. 2016;17(8):770‐92. 204. Meany G, Conceição E, Mitchell JE. Binge eating, binge eating disorder and loss of control eating: Effects on weight outcomes after bariatric surgery. European Eating Disorders Review. 2014;22(2):87‐91. 205. Conceição E, Orcutt M, Mitchell J, Engel S, LaHaise K, Jorgensen M, et al. Eating disorders after bariatric surgery: A case series. International Journal of Eating Disorders. 2013;46(3):274‐9. 206. Smith KE, Orcutt M, Steffen KJ, Crosby RD, Cao L, Garcia L, et al. Loss of control eating and binge eating in the 7 years following bariatric surgery. Obesity Surgery. 2019;29(6):1773‐80.

267

207. Fried M, Yumuk V, Oppert JM, Scopinaro N, Torres A, Weiner R, et al. Interdisciplinary European guidelines on metabolic and bariatric surgery. Obesity Surgery. 2014;24(1):42‐55. 208. Sogg S, Lauretti J, West‐Smith L. Recommendations for the presurgical psychosocial evaluation of bariatric surgery patients. Surgery for Obesity and Related Diseases. 2016;12(4):731‐49. 209. Lauby‐Secretan B, Scoccianti C, Loomis D, Grosse Y, Bianchini F, Straif K. Body fatness and cancer—viewpoint of the IARC Working Group. New England Journal of Medicine. 2016;375(8):794‐8. 210. Basen‐Engquist K, Chang M. Obesity and cancer risk: Recent review and evidence. Current Oncology Reports. 2011;13(1):71‐6. 211. Bruno DS, Berger NA. Impact of bariatric surgery on cancer risk reduction. Annals of Translational Medicine. 2019:1‐15. 212. Bruno DS, Berger NA. Impact of bariatric surgery on cancer risk reduction. Annals of Translational Medicine. 2019:1‐15. 213. Wiggins T, Antonowicz SS, Markar SR. Cancer risk following bariatric surgery: Systematic review and meta‐analysis of national population‐based cohort studies. Obesity Surgery. 2019;29(3):1031‐9. 214. Mackenzie H, Markar SR, Askari A, Faiz O, Hull M, Purkayastha S, et al. Obesity surgery and risk of cancer. British Journal of Surgery. 2018;105(12):1650‐7. 215. Sjöström L, Gummesson A, Sjöström CD, Narbro K, Peltonen M, Wedel H, et al. Effects of bariatric surgery on cancer incidence in obese patients in Sweden (Swedish Obese Subjects Study): A prospective, controlled intervention trial. The Lancet Oncology. 2009;10(7):653‐ 62. 216. Ye Y, Yang D, Han J. The effect of bariatric surgery on bone health: From mechanism to management. Obesity Medicine. 2020:100199. 217. Lu C‐W, Chang Y‐K, Chang H‐H, Kuo C‐S, Huang C‐T, Hsu C‐C, et al. Fracture risk after bariatric surgery: A 12‐year nationwide cohort study. Medicine. 2015;94(48). 218. Elaine WY, Lee MP, Landon JE, Lindeman KG, Kim SC. Fracture risk after bariatric surgery: Roux‐en‐Y gastric bypass versus adjustable gastric banding. Journal of Bone and Mineral Research. 2017;32(6):1229‐36. 219. Busetto L, Dicker D, Azran C, Batterham RL, Farpour‐Lambert N, Fried M, et al. Practical recommendations of the obesity management task force of the European Association for the Study of Obesity for the post‐bariatric surgery medical management. Obesity Facts. 2017;10(6):597‐632. 220. Sheng B, Truong K, Spitler H, Zhang L, Tong X, Chen L. The long‐term effects of bariatric surgery on type 2 diabetes remission, microvascular and macrovascular complications, and mortality: A systematic review and meta‐analysis. Obesity Surgery. 2017;27(10):2724‐32. 221. Gaeta M, Rausa E, Malavazos AE, Bonavina L, Smuts CM, Ricci C. Bariatric surgery to reduce mortality in US adults. A public health perspective from the analysis of the American National Health and Nutrition Examination Survey linked to the US mortality register. Obesity Surgery. 2018;28(4):900‐6. 222. Rubino F, Nathan D, Eckel R, Schauer P, Alberti K, Zimmet P, et al. Delegates of the 2nd Diabetes Surgery Summit. Metabolic surgery in the treatment algorithm for type 2 diabetes:

268

a joint statement by international diabetes organizations. Diabetes Care. 2016;39(6):861‐ 77. 223. Rubino F, Nathan DM, Eckel RH, Schauer PR, Alberti KGM, Zimmet PZ, et al. Metabolic surgery in the treatment algorithm for type 2 diabetes: A joint statement by international diabetes organizations. Obesity Surgery. 2017;27(1):2‐21. 224. Still CD, Wood GC, Benotti P, Petrick AT, Gabrielsen J, Strodel WE, et al. Preoperative prediction of type 2 diabetes remission after Roux‐en‐Y gastric bypass surgery: a retrospective cohort study. The Lancet Diabetes & Endocrinology. 2014;2(1):38‐45. 225. Lee W‐J, Hur KY, Lakadawala M, Kasama K, Wong SK, Chen S‐C, et al. Predicting success of metabolic surgery: Age, body mass index, C‐peptide, and duration score. Surgery for Obesity and Related Diseases. 2013;9(3):379‐84. 226. Ugale S, Gupta N, Modi KD, Kota SK, Satwalekar V, Naik V, et al. Prediction of remission after metabolic surgery using a novel scoring system in type 2 diabetes–a retrospective cohort study. Journal of Diabetes & Metabolic Disorders. 2014;13(1):89. 227. Aminian A, Brethauer SA, Andalib A, Nowacki AS, Jimenez A, Corcelles R, et al. Individualized metabolic surgery score: procedure selection based on diabetes severity. Annals of surgery. 2017;266(4):650‐7. 228. Wood GC, Horwitz D, Still CD, Mirshahi T, Benotti P, Parikh M, et al. Performance of the DiaRem score for predicting diabetes remission in two health systems following bariatric surgery procedures in Hispanic and non‐Hispanic white patients. Obesity surgery. 2018;28(1):61‐8. 229. Dicker D, Golan R, Aron‐Wisnewsky J, Zucker J‐D, Sokolowska N, Comaneshter DS, et al. Prediction of Long‐Term Diabetes Remission After RYGB, Sleeve Gastrectomy, and Adjustable Gastric Banding Using DiaRem and Advanced‐DiaRem Scores. Obesity surgery. 2019;29(3):796‐804. 230. Ha J, Kwon Y, Kim NH, Park S, Menzo EL, Rosenthal RJ. Discordance in prediction for prognosis of type 2 diabetes after metabolic surgery: Comparison of the ABCD, DiaRem, and individualized metabolic surgery models. Annals of Surgical Treatment and Research. 2019;97(6):309‐18. 231. Pucci A, Tymoszuk U, Cheung W, Makaronidis J, Scholes S, Tharakan G, et al. Type 2 diabetes remission 2 years post Roux‐en‐Y gastric bypass and sleeve gastrectomy: the role of the weight loss and comparison of DiaRem and DiaBetter scores. Diabetic Medicine. 2018;35(3):360‐7. 232. Cotillard A, Poitou C, Duchâteau‐Nguyen G, Aron‐Wisnewsky J, Bouillot J‐L, Schindler T, et al. Type 2 diabetes remission after gastric bypass: what is the best prediction tool for clinicians? Obesity surgery. 2015;25(7):1128‐32. 233. Still CD, Benotti P, Mirshahi T, Cook A, Wood GC. DiaRem2: Incorporating duration of diabetes to improve prediction of diabetes remission after metabolic surgery. Surgery for Obesity and Related Diseases. 2019;15(5):717‐24. 234. Brightman L, Huang HCC, Dugdale P. Determining patient attendance, access to interventions and clinical outcomes in a publicly funded obesity programme: Results from the Canberra Obesity Management Service. Clinical Obesity. 2019;9:e12325. 235. Sharman M. J., Breslin M. C., Kuzminov A., Palmer A. J., Blizzard L., Hensher M., et al. Population estimates and characteristics of Australians potentially eligible for bariatric

269

surgery: Findings from the 2011–13 Australian Health Survey. Australian Health Review. 2018;42(4):429‐37. 236. Allen P, Keen J, Wright J, Dempster P, Townsend J, Hutchings A, et al. Investigating the governance of autonomous public hospitals in England: Multi‐site case study of NHS foundation trusts. Journal of Health Services Research & Policy. 2012;17(2):94‐100. 237. Hopkins J, Welbourn R. The importance of national registries/databases in metabolic surgery: The UK experience. Surgery for Obesity and Related Diseases. 2016;12(6):1178‐85. 238. Hennings DL, Baimas‐George M, Al‐Quarayshi Z, Moore R, Kandil E, DuCoin CG. The inequity of bariatric surgery: Publicly insured patients undergo lower rates of bariatric surgery with worse outcomes. Obesity Surgery. 2018;28(1):44‐51. 239. Diamant A, Milner J, Cleghorn M, Sockalingam S, Okrainec A, Jackson TD, et al. Analysis of patient attrition in a publicly funded bariatric surgery program. Journal of the American College of Surgeons. 2014;219(5):1047‐55. 240. Larjani S, Spivak I, Guo MH, Aliarzadeh B, Wang W, Robinson S, et al. Preoperative predictors of adherence to multidisciplinary follow‐up care postbariatric surgery. Surgery for Obesity and Related Diseases. 2016;12(2):350‐6. 241. Padwal RS, Chang H‐J, Klarenbach S, Sharma AM, Majumdar SR. Characteristics of the population eligible for and receiving publicly funded bariatric surgery in Canada. International Journal for Equity in Health. 2012;11(1):54. 242. Taylor T, Wang Y, Rogerson W, Bavin L, Sharon C, Beban G, et al. Attrition after acceptance onto a publicly funded bariatric surgery program. Obesity Surgery. 2018;28(8):2500‐7. 243. Debs T, Petrucciani N, Kassir R, Iannelli A, Amor IB, Gugenheim J. Trends of bariatric surgery in France during the last 10 years: Analysis of 267,466 procedures from 2005–2014. Surgery for Obesity and Related Diseases. 2016;12(8):1602‐9. 244. Thereaux J, Lesuffleur T, Païta M, Czernichow S, Basdevant A, Msika S, et al. Long‐term follow‐up after bariatric surgery in a national cohort. British Journal of Surgery. 2017;104(10):1362‐71. 245. Rasera I, Luque A, Junqueira SM, Brasil NC, Andrade PC. Effectiveness and safety of bariatric surgery in the Public Healthcare System in Brazil: Real‐world evidence from a high‐volume obesity surgery center. Obesity Surgery. 2017;27(2):536‐40. 246. Zubiaurre PR, Bahia LR, da Rosa MQM, Assumpção RP, Padoin AV, Sussembach SP, et al. Estimated costs of clinical and surgical treatment of severe obesity in the Brazilian Public Health System. Obesity Surgery. 2017;27(12):3273‐80. 247. Dagan SS, Keidar A, Raziel A, Sakran N, Goitein D, Shibolet O, et al. Do bariatric patients follow dietary and lifestyle recommendations during the first postoperative year? Obesity Surgery. 2017;27(9):2258‐71. 248. Hood MM, Corsica J, Bradley L, Wilson R, Chirinos DA, Vivo A. Managing severe obesity: Understanding and improving treatment adherence in bariatric surgery. Journal of Behavioral Medicine. 2016;39(6):1092‐103. 249. Hood MM, Kelly MC, Feig EH, Webb V, Bradley LE, Corsica J. Measurement of adherence in bariatric surgery: A systematic review. Surgery for Obesity and Related Diseases. 2018;14(8):1192‐201. 250. Parretti HM, Hughes CA, Jones LL. ‘The rollercoaster of follow‐up care’ after bariatric surgery: A rapid review and qualitative synthesis. Obesity Reviews. 2019;20(1):88‐107.

270

251. Higa K, Ho T, Tercero F, Yunus T, Boone KB. Laparoscopic Roux‐en‐Y gastric bypass: 10‐Year follow‐up. Surgery for Obesity and Related Diseases. 2011;7(4):516‐25. 252. Kim HJ, Madan A, Fenton‐Lee D. Does patient compliance with follow‐up influence weight loss after gastric bypass surgery? A systematic review and meta‐analysis. Obesity surgery. 2014;24(4):647‐51. 253. Spaniolas K, Kasten KR, Celio A, Burruss MB, Pories WJ. Postoperative follow‐up after bariatric surgery: effect on weight loss. Obesity surgery. 2016;26(4):900‐3. 254. Krimpuri RD, Yokley JM, Seeholzer EL, Horwath EL, Thomas CL, Bardaro SJ. Qualifying for bariatric surgery: is preoperative weight loss a reliable predictor of postoperative weight loss? Surgery for Obesity and Related Diseases. 2018;14(1):60‐4. 255. Tewksbury C, Wu J, Allison K, Gardiner H, Dumon K, Williams N, et al. Prebariatric surgery care and postoperative outcomes: Increased number of visits associated with smaller weight losses over first 2 postoperative years. Surgery for Obesity and Related Diseases. 2019;15:1548‐53. 256. Kalarchian MA, Marcus MD, Courcoulas AP, Cheng Y, Levine MD. Preoperative lifestyle intervention in bariatric surgery: A randomized clinical trial. Surgery for Obesity and Related Diseases. 2016;12(1):180‐7. 257. Ochner C, Dambkowski C, Yeomans B, Teixeira J, Pi‐Sunyer FX. Pre‐bariatric surgery weight loss requirements and the effect of preoperative weight loss on postoperative outcome. International Journal of Obesity. 2012;36(11):1380. 258. Livhits M, Mercado C, Yermilov I, Parikh JA, Dutson E, Mehran A, et al. Preoperative predictors of weight loss following bariatric surgery: Systematic review. Obesity Surgery. 2012;22(1):70‐89. 259. Gerber P, Anderin C, Thorell A. Weight loss prior to bariatric surgery: an updated review of the literature. Scandinavian Journal of Surgery. 2015;104(1):33‐9. 260. Pratt KJ, Jalilvand A, Needleman B, Urse K, Ferriby M, Noria S. Postoperative outcomes based on patient participation in a presurgery education and weight management program. Surgery for Obesity and Related Diseases. 2018;14(11):1714‐23. 261. Parikh M, Dasari M, McMacken M, Ren C, Fielding G, Ogedegbe G. Does a preoperative medically supervised weight loss program improve bariatric surgery outcomes? A pilot randomized study. Surgical endoscopy. 2012;26(3):853‐61. 262. Sherman WE, Lane AE, Mangieri CW, Choi YU, Faler BJ. Does preoperative weight change predict postoperative weight loss after laparoscopic sleeve gastrectomy? Bariatric Surgical Practice and Patient Care. 2015;10(3):126‐9. 263. Cassie S, Menezes C, Birch DW, Shi X, Karmali S. Effect of preoperative weight loss in bariatric surgical patients: a systematic review. Surgery for Obesity and Related Diseases. 2011;7(6):760‐7. 264. Kim JJ, Rogers AM, Ballem N, Schirmer B. ASMBS updated position statement on insurance mandated preoperative weight loss requirements. Surgery for Obesity and Related Diseases. 2016;12(5):955‐9. 265. Ochner CN, Jochner MC, Caruso EA, Teixeira J, Pi‐Sunyer FX. Effect of preoperative body mass index on weight loss after obesity surgery. Surgery for Obesity and Related Diseases. 2013;9(3):423‐7.

271

266. O'Brien PE, MacDonald L, Anderson M, Brennan L, Brown WA. Long‐term outcomes after bariatric surgery: fifteen‐year follow‐up of adjustable gastric banding and a systematic review of the bariatric surgical literature. Annals of Surgery. 2013;257(1):87‐94. 267. Risstad H, Søvik TT, Engström M, Aasheim ET, Fagerland MW, Olsén MF, et al. Five‐year outcomes after laparoscopic gastric bypass and laparoscopic duodenal switch in patients with body mass index of 50 to 60: a randomized clinical trial. JAMA surgery. 2015;150(4):352‐61. 268. Gagner M, Deitel M, Erickson AL, Crosby RD. Survey on laparoscopic sleeve gastrectomy (LSG) at the Fourth International Consensus Summit on Sleeve Gastrectomy. Obesity surgery. 2013;23(12). 269. Lainas P, Dammaro C, Gaillard M, Donatelli G, Tranchart H, Dagher I. Safety and short‐term outcomes of laparoscopic sleeve gastrectomy for patients over 65 years old with severe obesity. Surgery for Obesity and Related Diseases. 2018;14(7):952‐9. 270. Sjöström L, Lindroos A‐K, Peltonen M, Torgerson J, Bouchard C, Carlsson B, et al. Lifestyle, diabetes, and cardiovascular risk factors 10 years after bariatric surgery. New England Journal of Medicine. 2004;351(26):2683‐93. 271. Prachand VN, Ward M, Alverdy JC. Duodenal switch provides superior resolution of metabolic comorbidities independent of weight Loss in the super‐obese (BMI≥ 50 kg/m 2) compared with gastric bypass. Journal of Gastrointestinal Surgery. 2010;14(2):211‐20. 272. Sims EA, Danforth Jr E, Horton ES, Bray GA, Glennon J, Salans L, editors. Endocrine and metabolic effects of experimental obesity in man. Proceedings of the 1972 Laurentian Hormone Conference; 1973: Elsevier. 273. American Diabetes Association (ADA). Standards of Medical Care in Diabetes—2020. Diabetes Care. 2020;43(Suppl. 1):S1‐S212. 274. Schauer PR, Bhatt DL, Kirwan JP, Wolski K, Brethauer SA, Navaneethan SD, et al. Bariatric surgery versus intensive medical therapy for diabetes—3‐year outcomes. New England Journal of Medicine. 2014;370(21):2002‐13. 275. Courcoulas AP, Belle SH, Neiberg RH, Pierson SK, Eagleton JK, Kalarchian MA, et al. Three‐ year outcomes of bariatric surgery vs lifestyle intervention for type 2 diabetes mellitus treatment: a randomized clinical trial. JAMA surgery. 2015;150(10):931‐40. 276. Hanipah ZN, Schauer PR. Bariatric surgery as a long‐term treatment for type 2 diabetes/metabolic syndrome. Annual Review of Medicine. 2020;71:1‐15. 277. Arterburn DE, Bogart A, Sherwood NE, Sidney S, Coleman KJ, Haneuse S, et al. A multisite study of long‐term remission and relapse of type 2 diabetes mellitus following gastric bypass. Obesity Surgery. 2013;23(1):93‐102. 278. Madsen LR, Baggesen LM, Richelsen B, Thomsen RW. Effect of Roux‐en‐Y gastric bypass surgery on diabetes remission and complications in individuals with type 2 diabetes: A Danish population‐based matched cohort study. Diabetologia. 2019;62(4):611‐20. 279. Lager CJ, Esfandiari NH, Luo Y, Subauste AR, Kraftson AT, Brown MB, et al. Metabolic parameters, weight loss, and comorbidities 4 years after Roux‐en‐Y gastric bypass and sleeve gastrectomy. Obesity Surgery. 2018;28(11):3415‐23. 280. Graham C, Switzer N, Reso A, Armstrong C, Church N, Mitchell P, et al. Sleeve gastrectomy and hypertension: A systematic review of long‐term outcomes. Surgical Endoscopy. 2019;33(9):3001‐7.

272

281. Kotsis V, Stabouli S, Papakatsika S, Rizos Z, Parati G. Mechanisms of obesity‐induced hypertension. Hypertension Research. 2010;33(5):386‐93. 282. Casella G, Soricelli E, Giannotti D, Collalti M, Maselli R, Genco A, et al. Long‐term results after laparoscopic sleeve gastrectomy in a large monocentric series. Surgery for Obesity and Related Diseases. 2016;12(4):757‐62. 283. Puzziferri N, Roshek TB, Mayo HG, Gallagher R, Belle SH, Livingston EH. Long‐term follow‐ up after bariatric surgery: A systematic review. JAMA. 2014;312(9):934‐42. 284. Ruiz‐Tovar J, Martínez R, Bonete JM, Rico JM, Zubiaga L, Diez M, et al. Long‐term weight and metabolic effects of laparoscopic sleeve gastrectomy calibrated with a 50‐Fr Bougie. Obesity Surgery. 2016;26(1):32‐7. 285. Kehagias I, Spyropoulos C, Karamanakos S, Kalfarentzos F. Efficacy of sleeve gastrectomy as sole procedure in patients with clinically severe obesity (BMI ≤50 Kg/m2). Surgery for Obesity and Related Diseases. 2013;9(3):363‐9. 286. Gadiot RP, Biter LU, van Mil S, Zengerink HF, Apers J, Mannaerts GH. Long‐term results of laparoscopic sleeve gastrectomy for morbid obesity: 5 to 8‐year results. Obesity Surgery. 2017;27(1):59‐63. 287. Jammu GS, Sharma R. A 7‐year clinical audit of 1107 cases comparing sleeve gastrectomy, Roux‐En‐Y gastric bypass, and mini‐gastric bypass, to determine an effective and safe bariatric and metabolic procedure. Obesity Surgery. 2016;26(5):926‐32. 288. Neagoe RM, Mureșan M, Bancu Ș, Balmos I, Băișan V, Voidăzan S, et al. Results of laparoscopic sleeve gastrectomy—5‐Year follow‐up study in an eastern european emerging bariatric center. Obesity Surgery. 2017;27(4):983‐9. 289. Nocca D, Loureiro M, Skalli EM, Nedelcu M, Jaussent A, Deloze M, et al. Five‐year results of laparoscopic sleeve gastrectomy for the treatment of severe obesity. Surgical Endoscopy. 2017;31(8):3251‐7. 290. Ricci C, Gaeta M, Rausa E, Macchitella Y, Bonavina L. Early impact of bariatric surgery on type II diabetes, hypertension, and hyperlipidemia: A systematic review, meta‐analysis and meta‐regression on 6,587 patients. Obesity Surgery. 2014;24(4):522‐8. 291. Catapano AL, Graham I, De Backer G, Wiklund O, Chapman MJ, Drexel H, et al. 2016 ESC/EAS guidelines for the management of dyslipidaemias. European Heart Journal. 2016;37(39):2999‐3058. 292. Bastien M, Poirier P, Lemieux I, Després J‐P. Overview of epidemiology and contribution of obesity to cardiovascular disease. Progress in Cardiovascular Diseases. 2014;56(4):369‐81. 293. Pajecki D, Kawamoto F, Dantas ACB, Andrade PC, Brasil NC, Junqueira SM, et al. Real‐world evidence of health outcomes and medication use 24 months after bariatric surgery in the public healthcare system in Brazil: A retrospective, single‐center study. Clinics. 2020;75. 294. Patkar A, Fegelman E, R. Kashyap S, Brethauer S, Bour E, Yoo A, et al. Assessing the real‐ world effect of laparoscopic bariatric surgery on the management of obesity‐related comorbidities: A retrospective matched cohort study using a US C laims Database. Diabetes, Obesity and Metabolism. 2017;19(2):181‐8. 295. Spivak H, Sakran N, Dicker D, Rubin M, Raz I, Shohat T, et al. Different effects of bariatric surgical procedures on dyslipidemia: A registry‐based analysis. Surgery for Obesity and Related Diseases. 2017;13(7):1189‐94.

273

296. Wazir N, Arshad MF, Finney J, Kirk K, Dewan S. Two years remission of type 2 diabetes mellitus after bariatric surgery. Journal of the College of Physicians and Surgeons Pakistan. 2019;29(10):967‐71. 297. Miras AD, Kamocka A, Patel D, Dexter S, Finlay I, Hopkins JC, et al. Obesity surgery makes patients healthier and more functional: Real world results from the United Kingdom National Bariatric Surgery Registry. Surgery for Obesity and Related Diseases. 2018;14(7):1033‐40. 298. Vigilante A, Signorini F, Marani M, Paganini V, Viscido G, Navarro L, et al. Impact on dyslipidemia after laparoscopic sleeve gastrectomy. Obesity Surgery. 2018;28(10):3111‐5. 299. Climent E, Benaiges D, Flores‐Le Roux JA, Ramon JM, Pedro‐Botet J, Goday A. Changes in the lipid profile 5 years after bariatric surgery: Laparoscopic Roux‐en‐Y gastric bypass versus laparoscopic sleeve gastrectomy. Surgery for Obesity and Related Diseases. 2018;14(8):1099‐105. 300. Okifuji A, Hare BD. The association between chronic pain and obesity. Journal of Pain Research. 2015;8:399. 301. Gill R, Al‐Adra D, Shi X, Sharma A, Birch D, Karmali S. The benefits of bariatric surgery in obese patients with hip and knee osteoarthritis: A systematic review. Obesity Reviews. 2011;12(12):1083‐9. 302. King WC, Chen J‐Y, Belle SH, Courcoulas AP, Dakin GF, Elder KA, et al. Change in pain and physical function following bariatric surgery for severe obesity. JAMA. 2016;315(13):1362‐ 71. 303. Speck RM, Bond DS, Sarwer DB, Farrar JT. A systematic review of musculoskeletal pain among bariatric surgery patients: Implications for physical activity and exercise. Surgery for Obesity and Related Diseases. 2014;10(1):161‐70. 304. Punjabi NM. The epidemiology of adult obstructive sleep apnea. Proceedings of the American Thoracic Society. 2008;5(2):136‐43. 305. Loo GH, Rajan R, Mohd Tamil A, Ritza Kosai N. Prevalence of obstructive sleep apnea in an Asian bariatric population: An underdiagnosed dilemma. Surgery for Obesity and Related Diseases. 2020;16(6):778‐83. 306. Peromaa‐Haavisto P, Tuomilehto H, Kössi J, Virtanen J, Luostarinen M, Pihlajamäki J, et al. Obstructive sleep apnea: the effect of bariatric surgery after 12 months. A prospective multicenter trial. Sleep Medicine. 2017;35:85‐90. 307. Kositanurit W, Muntham D, Udomsawaengsup S, Chirakalwasan N. Prevalence and associated factors of obstructive sleep apnea in morbidly obese patients undergoing bariatric surgery. Sleep and Breathing. 2018;22(1):251‐6. 308. Hawkins M, Leung SE, Lee A, Wnuk S, Cassin S, Hawa R, et al. Psychiatric medication use and weight outcomes one year after bariatric surgery. Psychosomatics. 2020;61(1):56‐63. 309. Müller A, Hase C, Pommnitz M, de Zwaan M. Depression and suicide after bariatric surgery. Current Psychiatry Reports. 2019;21(9):84. 310. Bielawska B, Ouellette‐Kuntz H, Patel SV, Anvari M, Zevin B. Severe nutritional complications after bariatric surgery in Ontario adults: A population based descriptive study. Surgery for Obesity and Related Diseases. 2020. 311. Toh SY, Zarshenas N, Jorgensen J. Prevalence of nutrient deficiencies in bariatric patients. Nutrition. 2009;25(11‐12):1150‐6.

274

312. Souza NM, Santos ACO, Santa‐Cruz F, Guimarães H, Silva LM, de‐Lima DSC, et al. Nutritional impact of bariatric surgery: A comparative study of Roux‐en‐Y Gastric Bypass and Sleeve gastrectomy between patients from the public and private health systems. Revista do Colégio Brasileiro de Cirurgiões. 2020;47. 313. Patel JJ, Mundi MS, Hurt RT, Wolfe B, Martindale RG. Micronutrient deficiencies after bariatric surgery: An emphasis on vitamins and trace minerals. Nutrition in Clinical Practice. 2017;32(4):471‐80. 314. Mehaffey JH, LaPar DJ, Turrentine FE, Miller MS, Hallowell PT, Schirmer BD. Outcomes of laparoscopic Roux‐en‐Y gastric bypass in super‐super‐obese patients. Surgery for Obesity and Related Diseases. 2015;11(4):814‐9. 315. Zerrweck C, Sepúlveda EM, Maydón HG, Campos F, Spaventa AG, Pratti V, et al. Laparoscopic gastric bypass vs. sleeve gastrectomy in the super obese patient: early outcomes of an observational study. Obesity surgery. 2014;24(5):712‐7. 316. Schwartz A, Etchechoury L, Collet D. Outcome after laparoscopic gastric bypass for super‐ super obese patients. Journal of visceral surgery. 2013;150(2):145‐9. 317. Villamere J, Gebhart A, Vu S, Nguyen NT. Body mass index is predictive of higher in‐hospital mortality in patients undergoing laparoscopic gastric bypass but not laparoscopic sleeve gastrectomy or gastric banding. The American Surgeon. 2014;80(10):1039‐43. 318. Arapis K, Macrina N, Kadouch D, Parenti LR, Marmuse JP, Hansel B. Outcomes of Roux‐en‐ Y gastric bypass versus sleeve gastrectomy in super‐super‐obese patients (BMI≥ 60 kg/m2): 6‐year follow‐up at a single university. Surgery for Obesity and Related Diseases. 2019;15(1):23‐33. 319. Zak Y, Petrusa E, Gee DW. Laparoscopic Roux‐en‐Y gastric bypass patients have an increased lifetime risk of repeat operations when compared to laparoscopic sleeve gastrectomy patients. Surgical Endoscopy. 2016;30(5):1833‐8. 320. McGlone ER, Bond A, Reddy M, Khan OA, Wan AC. Super‐obesity in the elderly: is bariatric surgery justified? Obesity surgery. 2015;25(9):1750‐5. 321. Hariri K, Guevara D, Dong M, Kini SU, Herron DM, Fernandez‐Ranvier G. Is bariatric surgery effective for co‐morbidity resolution in the super‐obese patients? Surgery for Obesity and Related Diseases. 2018;14(9):1261‐8. 322. Gill RS, Karmali S, Sharma AM. The potential role of the Edmonton obesity staging system in determining indications for bariatric surgery. Obesity Surgery. 2011;21(12):1947‐9. 323. Sharma AM, Kushner RF. A proposed clinical staging system for obesity. International Journal of Obesity. 2009;33(3):289‐95. 324. Whyte M, Velusamy S, Aylwin S. Disease severity and staging of obesity: A rational approach to patient selection. Current Atherosclerosis Reports. 2014;16(11):456. 325. Ogassavara NC, Dias JGM, Pajecki D, de Oliveira Siqueira J, Santo MA, Tess BH. The Edmonton Obesity Staging System: Assessing a potential tool to improve the management of obesity surgery in the Brazilian public health services. Surgery for Obesity and Related Diseases. 2020;16(1):40‐7. 326. Brethauer SA, Kim J, El Chaar M, Papasavas P, Eisenberg D, Rogers A, et al. Standardized outcomes reporting in metabolic and bariatric surgery. Obesity Surgery. 2015;25(4):587‐ 606.

275

327. Maglio C, Peltonen M, Neovius M, Jacobson P, Jacobsson L, Rudin A, et al. Effects of bariatric surgery on gout incidence in the Swedish Obese Subjects study: A non‐randomised, prospective, controlled intervention trial. Annals of the Rheumatic Diseases. 2017;76(4):688‐93. 328. Dannenberg AL, Kannel WB. Remission of hypertension: The'natural'history of blood pressure treatment in the Framingham Study. JAMA. 1987;257(11):1477‐83. 329. Smith ME, Lee JS, Bonham A, Varban OA, Finks JF, Carlin AM, et al. Effect of new persistent opioid use on physiologic and psychologic outcomes following bariatric surgery. Surgical Endoscopy. 2019;33(8):2649‐56. 330. Burton P, Brown W, Chen R, Shaw K, Packiyanathan A, Bringmann I, et al. Outcomes of high‐ volume bariatric surgery in the public system. ANZ Journal of Surgery. 2016;86(7‐8):572‐7. 331. Clough A, Hamill D, Jackson S, Remilton M, Eyre R, Callahan R. Outcome of three common bariatric procedures in the public sector. ANZ Journal of Surgery. 2017;87(11):930‐4. 332. Chang D‐M, Lee W‐J, Chen J‐C, Ser K‐H, Tsai P‐L, Lee Y‐C. Thirteen‐year experience of laparoscopic sleeve gastrectomy: Surgical risk, weight loss, and revision procedures. Obesity Surgery. 2018;28(10):2991‐7. 333. Arterburn D, Wellman R, Emiliano A, Smith SR, Odegaard AO, Murali S, et al. Comparative effectiveness and safety of bariatric procedures for weight loss: A PCORnet cohort study. Annals of Internal Medicine. 2018;169(11):741–50. 334. Lauti M, Kularatna M, Hill AG, MacCormick AD. Weight regain following sleeve gastrectomy: A systematic review. Obesity Surgery. 2016;26(6):1326‐34. 335. Karmali S, Brar B, Shi X, Sharma AM, de Gara C, Birch DW. Weight recidivism post‐bariatric surgery: A systematic review. Obesity Surgery. 2013;23(11):1922‐33. 336. Yu Y, Klem ML, Kalarchian MA, Ji M, Burke LE. Predictors of weight regain after sleeve gastrectomy: An integrative review. Surgery for Obesity and Related Diseases. 2019;15(6):995‐1005. 337. Angrisani L, Santonicola A, Iovino P, Formisano G, Buchwald H, Scopinaro N. Bariatric surgery worldwide 2013. Obesity Surgery. 2015;25(10):1822‐32. 338. Ali M, El Chaar M, Ghiassi S, Rogers AM. American Society for Metabolic and Bariatric Surgery updated position statement on sleeve gastrectomy as a bariatric procedure. Surgery for Obesity and Related Diseases. 2017;13(10):1652‐7. 339. Rosenthal RJ, Panel ISGE. International Sleeve Gastrectomy Expert Panel Consensus Statement: Best practice guidelines based on experience of> 12,000 cases. Surgery for Obesity and Related Diseases. 2012;8(1):8‐19. 340. Eid GM, Brethauer S, Mattar SG, Titchner RL, Gourash W, Schauer PR. Laparoscopic sleeve gastrectomy for super obese patients: Forty‐eight percent excess weight loss after 6 to 8 years with 93% follow‐up. Annals of Surgery. 2012;256(2):262‐5. 341. Kuk JL, Ardern CI, Church TS, Sharma AM, Padwal R, Sui X, et al. Edmonton Obesity Staging System: Association with weight history and mortality risk. Applied Physiology, Nutrition, and Metabolism. 2011;36(4):570‐6. 342. Affinati AH, Esfandiari NH, Oral EA, Kraftson AT. Bariatric surgery in the treatment of type 2 diabetes. Current Diabetes Reports. 2019;19(12):156. 343. Batterham RL, Cummings DE. Mechanisms of diabetes improvement following bariatric/metabolic surgery. Diabetes Care. 2016;39(6):893‐901.

276

344. Cho YM. A gut feeling to cure diabetes: Potential mechanisms of diabetes remission after bariatric surgery. Diabetes & Metabolism Journal. 2014;38(6):406‐15. 345. Penney N, Kinross J, Newton R, Purkayastha S. The role of bile acids in reducing the metabolic complications of obesity after bariatric surgery: A systematic review. International Journal of Obesity. 2015;39(11):1565‐74. 346. Jans A, Näslund I, Ottosson J, Szabo E, Näslund E, Stenberg E. Duration of type 2 diabetes and remission rates after bariatric surgery in Sweden 2007–2015: A registry‐based cohort study. PLoS Medicine. 2019;16(11):e1002985. 347. Souteiro P, Belo S, Magalhães D, Pedro J, Neves JS, Oliveira SC, et al. Long‐term diabetes outcomes after bariatric surgery—Managing medication withdrawal. International Journal of Obesity. 2019;43(11):2217‐24. 348. Singh P, Subramanian A, Adderley N, Gokhale K, Singhal R, Bellary S, et al. Impact of bariatric surgery on cardiovascular outcomes and mortality: A population‐based cohort study. British Journal of Surgery. 2020;107(4):432‐42. 349. Solomon MJ, McLeod RS. Should we be performing more randomized controlled trials evaluating surgical operations? Surgery. 1995;118(3):459‐67. 350. Fung EK, Loré JM, Jr. Randomized controlled trials for evaluating surgical questions. Arch Otolaryngol Head Neck Surg. 2002;128(6):631‐4. 351. Bays HE, Toth PP, Kris‐Etherton PM, Abate N, Aronne LJ, Brown WV, et al. Obesity, adiposity, and dyslipidemia: A consensus statement from the National Lipid Association. Journal of Clinical Lipidology. 2013;7(4):304‐83. 352. Cunha FM, Oliveira J, Preto J, Saavedra A, Costa MM, Magalhães D, et al. The effect of bariatric surgery type on lipid profile: An age, sex, body mass index and excess weight loss matched study. Obesity Surgery. 2016;26(5):1041‐7. 353. Bibbins‐Domingo K, Grossman DC, Curry SJ, Davidson KW, Epling JW, García FA, et al. Statin use for the primary prevention of cardiovascular disease in adults: US Preventive Services Task Force recommendation statement. JAMA. 2016;316(19):1997‐2007. 354. King WC, Chen J‐Y, Belle SH, Courcoulas AP, Dakin GF, Flum DR, et al. Use of prescribed opioids before and after bariatric surgery: Prospective evidence from a US multicenter cohort study. Surgery for Obesity and Related Diseases. 2017;13(8):1337‐46. 355. Raebel MA, Newcomer SR, Reifler LM, Boudreau D, Elliott TE, DeBar L, et al. Chronic use of opioid medications before and after bariatric surgery. JAMA. 2013;310(13):1369‐76. 356. Raebel MA, Newcomer SR, Bayliss EA, Boudreau D, DeBar L, Elliott TE, et al. Chronic opioid use emerging after bariatric surgery. Pharmacoepidemiology and Drug Safety. 2014;23(12):1247‐57. 357. Weingarten TN, Sprung J, Flores A, Baena AMO, Schroeder DR, Warner DO. Opioid requirements after laparoscopic bariatric surgery. Obesity Surgery. 2011;21(9):1407‐12. 358. Brummett CM, Waljee JF, Goesling J, Moser S, Lin P, Englesbe MJ, et al. New persistent opioid use after minor and major surgical procedures in US adults. JAMA Surgery. 2017;152(6):e170504‐e. 359. Bak M, Seibold‐Simpson SM, Darling R. The potential for cross‐addiction in post‐bariatric surgery patients: Considerations for primary care nurse practitioners. Journal of the American Association of Nurse Practitioners. 2016;28(12):675‐82.

277

360. Hill DS, Freudmann M, Sergeant JC, Board T. Management of symptomatic knee osteoarthritis in obesity: A survey of orthopaedic surgeons’ opinions and practice. European Journal of Orthopaedic Surgery & Traumatology. 2018;28(5):967‐74. 361. Cohen JB, Cohen DL. Cardiovascular and renal effects of weight reduction in obesity and the metabolic syndrome. Current Hypertension Reports. 2015;17(5):34. 362. Shiozawa A, Szabo SM, Bolzani A, Cheung A, Choi HK. Serum uric acid and the risk of incident and recurrent gout: A systematic review. The Journal of Rheumatology. 2017;44(3):388‐96. 363. Richette P, Bardin T. Gout. Lancet (London, England). 2010;375(9711):318‐28. 364. Yeo C, Kaushal S, Lim B, Syn N, Oo AM, Rao J, et al. Impact of bariatric surgery on serum uric acid levels and the incidence of gout: A meta‐analysis. Obesity Reviews. 2019;20(12):1759‐ 70. 365. Joosten LA, Netea MG, Mylona E, Koenders MI, Malireddi RS, Oosting M, et al. Engagement of fatty acids with toll‐like receptor 2 drives interleukin‐1β production via the ASC/caspase 1 pathway in monosodium urate monohydrate crystal–induced gouty arthritis. Arthritis & Rheumatism. 2010;62(11):3237‐48. 366. Cleophas MC, Crişan TO, Joosten LA. Factors modulating the inflammatory response in acute gouty arthritis. Current Opinion in Rheumatology. 2017;29(2):163‐70. 367. McGill NW. The epidemiology and treatment of gout. Open access rheumatology: Research and reviews. 2011;3:73. 368. Roddy E, Choi HK. Epidemiology of gout. Rheumatic Disease Clinics. 2014;40(2):155‐75. 369. Khanna D, Fitzgerald JD, Khanna PP, Bae S, Singh MK, Neogi T, et al. 2012 American College of Rheumatology guidelines for management of gout. Part 1: Systematic nonpharmacologic and pharmacologic therapeutic approaches to hyperuricemia. Arthritis Care & Research. 2012;64(10):1431‐46. 370. Richette P, Doherty M, Pascual E, Barskova V, Becce F, Castaneda‐Sanabria J, et al. 2016 updated EULAR evidence‐based recommendations for the management of gout. Annals of the Rheumatic Diseases. 2017;76(1):29‐42. 371. Ahluwalia JS, Chang P‐C, Tai C‐M, Tsai C‐C, Sun P‐L, Huang C‐K. Comparative study between laparoscopic adjustable gastric banded plication and sleeve gastrectomy in moderate obesity ‐ 2 year results. Obesity Surgery. 2016;26(3):552‐7. 372. Golomb I, David MB, Glass A, Kolitz T, Keidar A. Long‐term metabolic effects of laparoscopic sleeve gastrectomy. JAMA Surgery. 2015;150(11):1051‐7. 373. Strohmayer E, Via MA, Yanagisawa R. Metabolic management following bariatric surgery. Mount Sinai Journal of Medicine: A Journal of Translational and Personalized Medicine. 2010;77(5):431‐45. 374. Al‐Mutawa A, Anderson AK, Alsabah S, Al‐Mutawa M. Nutritional status of bariatric surgery candidates. Nutrients. 2018;10(1):67. 375. Mechanick JI, Apovian C, Brethauer S, Garvey WT, Joffe AM, Kim J, et al. Clinical Practice Guidelines for the perioperative nutrition, metabolic, and nonsurgical support of patients undergoing bariatric procedures – 2019 Update: cosponsored by American Association of Clinical Endocrinologists/American College of Endocrinology, The Obesity Society, American Society for Metabolic & Bariatric Surgery, Obesity Medicine Association, and American Society of Anesthesiologists. Surgery for Obesity and Related Diseases. 2019;In Press.

278

376. O'Kane M, Parretti HM, Pinkney J, Welbourn R, Hughes CA, Mok J, et al. British Obesity and Metabolic Surgery Society Guidelines on perioperative and postoperative biochemical monitoring and micronutrient replacement for patients undergoing bariatric surgery—2020 update. Obesity Reviews. 2020;21(11):e13087. 377. de Cleva R, Cardia L, Riccioppo D, Kawamoto M, Kanashiro N, Santo MA. Anemia before and after Roux‐en‐Y gastric bypass: Prevalence and evolution on long‐term follow‐up. Obesity Surgery. 2019;29(9):2790‐4. 378. Rolim F, Cruz F, Campos J, Ferraz Á. Long‐term repercussions of Roux‐en‐Y gastric bypass in a low‐income population: Assessment ten years after surgery. Revista do Colegio Brasileiro de Cirurgioes. 2018;45(4):e1916. 379. Camaschella C. Iron‐deficiency anemia. New England Journal of Medicine. 2015;372(19):1832‐43. 380. Ruz M, Carrasco F, Rojas P, Codoceo J, Inostroza J, Rebolledo A, et al. Iron absorption and iron status are reduced after roux‐en‐Y gastric bypass. The American Journal of Clinical Nutrition. 2009;90(3):527‐32. 381. Saltzman E, Philip Karl J. Nutrient deficiencies after gastric bypass surgery. Annual Review of Nutrition. 2013;33:183‐203. 382. Kornerup L, Hvas C, Abild C, Richelsen B, Nexo E. Early changes in vitamin B12 uptake and biomarker status following Roux‐en‐Y gastric bypass and sleeve gastrectomy. Clinical Nutrition. 2019;38(2):906‐11. 383. Stabler SP. Vitamin B12 deficiency. New England Journal of Medicine. 2013;368(2):149‐60. 384. Reges O, Greenland P, Dicker D, Leibowitz M, Hoshen M, Gofer I, et al. Association of bariatric surgery using laparoscopic banding, Roux‐en‐Y gastric bypass, or laparoscopic sleeve gastrectomy vs usual care obesity management with all‐cause mortality. JAMA. 2018;319(3):279‐90. 385. Bettencourt‐Silva R, Neves JS, Pedro J, Guerreiro V, Ferreira MJ, Salazar D, et al. Comparative effectiveness of different bariatric procedures in super morbid obesity. Obesity Surgery. 2019;29(1):281‐91. 386. Celio AC, Wu Q, Kasten KR, Manwaring ML, Pories WJ, Spaniolas K. Comparative effectiveness of Roux‐en‐Y gastric bypass and sleeve gastrectomy in super obese patients. Surgical Endoscopy. 2017;31(1):317‐23. 387. Zhang H, Han X, Yu H, Di J, Zhang P, Jia W. Effect of Roux‐en‐Y gastric bypass on remission of T2D: Medium‐term follow‐up in Chinese patients with different BMI obesity class. Obesity Surgery. 2017;27(1):134‐42. 388. Thereaux J, Corigliano N, Poitou C, Oppert J‐M, Czernichow S, Bouillot J‐L. Comparison of results after one year between sleeve gastrectomy and gastric bypass in patients with BMI ≥ 50 kg/m². Surgery for Obesity and Related Diseases. 2015;11(4):785‐90. 389. Serrano OK, Tannebaum JE, Cumella L, Choi J, Vemulapalli P, Melvin WS, et al. Weight loss outcomes and complications from bariatric surgery in the super super obese. Surgical Endoscopy. 2016;30(6):2505‐11. 390. Gonzalez‐Heredia R, Sanchez‐Johnsen L, Valbuena VS, Masrur M, Murphey M, Elli E. Surgical management of super–super obese patients: Roux‐en‐Y gastric bypass versus sleeve gastrectomy. Surgical endoscopy. 2016;30(5):2097‐102.

279

391. Magro DO, Geloneze B, Delfini R, Pareja BC, Callejas F, Pareja JC. Long‐term weight regain after gastric bypass: a 5‐year prospective study. Obesity surgery. 2008;18(6):648‐51. 392. Christou NV, Look D, MacLean LD. Weight gain after short‐and long‐limb gastric bypass in patients followed for longer than 10 years. Annals of surgery. 2006;244(5):734. 393. Compher CW, Hanlon A, Kang Y, Elkin L, Williams NN. Attendance at clinical visits predicts weight loss after gastric bypass surgery. Obesity surgery. 2012;22(6):927‐34. 394. Gould JC, Beverstein G, Reinhardt S, Garren MJ. Impact of routine and long‐term follow‐up on weight loss after laparoscopic gastric bypass. Surgery for Obesity and Related Diseases. 2007;3(6):627‐30. 395. Odom J, Zalesin KC, Washington TL, Miller WW, Hakmeh B, Zaremba DL, et al. Behavioral predictors of weight regain after bariatric surgery. Obesity surgery. 2010;20(3):349‐56. 396. Shen R, Dugay G, Rajaram K, Cabrera I, Siegel N, Ren CJ. Impact of patient follow‐up on weight loss after bariatric surgery. Obesity surgery. 2004;14(4):514‐9. 397. Pontiroli AE, Fossati A, Vedani P, Fiorilli M, Folli F, Paganelli M, et al. Post‐surgery adherence to scheduled visits and compliance, more than personality disorders, predict outcome of bariatric restrictive surgery in morbidly obese patients. Obesity surgery. 2007;17(11):1492‐ 7. 398. Dogan K, Homan J, Aarts EO, de Boer H, van Laarhoven CJ, Berends FJ. Long‐term nutritional status in patients following Roux‐en‐Y gastric bypass surgery. Clinical Nutrition. 2018;37(2):612‐7. 399. Poole NA, Atar AA, Kuhanendran D, Bidlake L, Fiennes A, McCluskey S, et al. Compliance with surgical after‐care following bariatric surgery for morbid obesity: a retrospective study. Obesity surgery. 2005;15(2):261‐5. 400. Wheeler E, Prettyman A, Lenhard MJ, Tran K. Adherence to outpatient program postoperative appointments after bariatric surgery. Surgery for obesity and related diseases. 2008;4(4):515‐20. 401. Goldenshluger A, Elazary R, Cohen M, Goldenshluger M, Ben‐Porat T, Nowotni J, et al. Predictors for adherence to multidisciplinary follow‐up care after sleeve gastrectomy. Obesity surgery. 2018;28(10):3054‐61. 402. Moroshko I, Brennan L, O’Brien P. Predictors of attrition in bariatric aftercare: a systematic review of the literature. Obesity surgery. 2012;22(10):1640‐7. 403. Schwoerer A, Kasten K, Celio A, Pories W, Spaniolas K. The effect of close postoperative follow‐up on co‐morbidity improvement after bariatric surgery. Surgery for Obesity and Related Diseases. 2017;13(8):1347‐52. 404. Vidal P, Ramón JM, Goday A, Parri A, Crous X, Trillo L, et al. Lack of adherence to follow‐up visits after bariatric surgery: reasons and outcome. Obesity surgery. 2014;24(2):179‐83. 405. Lara MD, Baker MT, Larson CJ, Mathiason MA, Lambert PJ, Kothari SN. Travel distance, age, and sex as factors in follow‐up visit compliance in the post‐gastric bypass population. Surgery for Obesity and Related Diseases. 2005;1(1):17‐21. 406. Sarwer DB, Wadden TA, Moore RH, Baker AW, Gibbons LM, Raper SE, et al. Preoperative eating behavior, postoperative dietary adherence, and weight loss after gastric bypass surgery. Surgery for Obesity and Related Diseases. 2008;4(5):640‐6. 407. Ma Y, Pagoto SL, Olendzki BC, Hafner AR, Perugini RA, Mason R, et al. Predictors of weight status following laparoscopic gastric bypass. Obesity surgery. 2006;16(9):1227‐31.

280

408. Moroshko I, Brennan L, O'Brien P. Predictors of dropout in weight loss interventions: a systematic review of the literature. Obesity reviews. 2011;12(11):912‐34. 409. World Health Organization (WHO). Adherence to long‐term therapies: Evidence for action. Geneva, Switzerland: World Health Organization; 2003. 410. Khorgami Z, Zhang C, Messiah SE, de la Cruz‐Muñoz N. Predictors of postoperative aftercare attrition among gastric bypass patients. Bariatric Surgical Practice and Patient Care. 2015;10(2):79‐83. 411. Kedestig J, Stenberg E. Loss to follow‐up after laparoscopic gastric bypass surgery–a post hoc analysis of a randomized clinical trial. Surgery for Obesity and Related Diseases. 2019;15(6):880‐6. 412. Gourash WF, Lockhart JS, Kalarchian MA, Courcoulas AP, Nolfi D. Retention and attrition in bariatric surgery research: An integrative review of the literature. Surgery for Obesity and Related Diseases. 2016;12(1):199‐209. 413. Gill RS, Karmali S, Hadi G, Al‐Adra DP, Shi X, Birch DW. Predictors of attrition in a multidisciplinary adult weight management clinic. Canadian Journal of Surgery. 2012;55(4):239. 414. Mackenzie CS, Gekoski WL, Knox V. Age, gender, and the underutilization of mental health services: The influence of help‐seeking attitudes. Aging and Mental Health. 2006;10(6):574‐ 82. 415. Sileo K, Wanyenze R, Kizito W, Reed E, Brodine S, Chemusto H, et al. Multi‐level determinants of clinic attendance and antiretroviral treatment adherence among fishermen living with HIV/AIDS in communities on Lake Victoria, Uganda. AIDS and Behavior. 2019;23(2):406‐17. 416. Broadhead RS, Heckathorn DD, Altice FL, Van Hulst Y, Carbone M, Friedland GH, et al. Increasing drug users’ adherence to HIV treatment: Results of a peer‐driven intervention feasibility study. Social Science & Medicine. 2002;55(2):235‐46. 417. Magura S, Laudet AB, Mahmood D, Rosenblum A, Knight E. Adherence to medication regimens and participation in dual‐focus self‐help groups. Psychiatric Services. 2002;53(3):310‐6. 418. Hoch C, Göbel U, Janssen G. Psychosocial support of patients with homozygous beta‐ thalassaemia. Klinische Padiatrie. 2000;212(4):216. 419. Getahun H, Maher D. Contribution of ‘TB clubs’ to tuberculosis control in a rural district in Ethiopia [Planning and Practice]. The International Journal of Tuberculosis and Lung Disease. 2000;4(2):174‐8. 420. Mehaffey JH, Michaels AD, Mullen MG, Meneveau MO, Pender JR, Hallowell PT. Patient travel for bariatric surgery: Does distance matter? Surgery for Obesity and Related Diseases. 2017;13(12):2027‐31. 421. Monfared S, Martin A, Selzer D, Butler A. Travel distance reduces follow‐up compliance but has no effect on long‐term weight loss success in bariatric patients. Surgical Endoscopy. 2020:1‐5. 422. Gourash WF, Ebel F, Lancaster K, Adeniji A, Iacono LK, Eagleton JK, et al. Longitudinal Assessment of Bariatric Surgery (LABS): Retention strategy and results at 24 months. Surgery for Obesity and Related Diseases. 2013;9(4):514‐9.

281

423. Bruze G, Holmin TE, Peltonen M, Ottosson J, Sjöholm K, Näslund I, et al. Associations of bariatric surgery with changes in interpersonal relationship status: Results from 2 Swedish cohort studies. JAMA Surgery. 2018;153(7):654‐61. 424. Jensen MD, Ryan DH, Apovian CM, Ard JD, Comuzzie AG, Donato KA, et al. 2013 AHA/ACC/TOS guideline for the management of overweight and obesity in adults: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines and The Obesity Society. Journal of the American college of cardiology. 2014;63(25 Part B):2985‐3023. 425. Tewksbury C, Williams NN, Dumon KR, Sarwer DB. Preoperative medical weight management in bariatric surgery: a review and reconsideration. Obesity surgery. 2017;27(1):208‐14. 426. Livhits M, Mercado C, Yermilov I, Parikh JA, Dutson E, Mehran A, et al. Does weight loss immediately before bariatric surgery improve outcomes: A systematic review. Surgery for Obesity and Related Diseases. 2009;5(6):713‐21. 427. Solomon H, Liu GY, Alami R, Morton J, Curet MJ. Benefits to patients choosing preoperative weight loss in gastric bypass surgery: new results of a randomized trial. Journal of the American College of Surgeons. 2009;208(2):241‐5. 428. Abbott S, Lawson J, Singhal R, Parretti H, Tahrani A. Weight loss during medical weight management does not predict weight loss after bariatric surgery: A retrospective cohort study. Surgery for Obesity and Related Diseases. 2020;In Press, Corrected Proof. 429. Atlantis E, Langford K, Piya M, Ho V, Skelsey K, Rickards L, et al. Physical capacity outcomes in patients with severe obesity after 12 months of physician‐led multidisciplinary team care: A case series from a public hospital clinical obesity service. Clinical Obesity. 2019;9(6):e12337. 430. Andersen JR, Aadland E, Nilsen RM, Våge V. Predictors of weight loss are different in men and women after sleeve gastrectomy. Obesity surgery. 2014;24(4):594‐8. 431. Parri A, Benaiges D, Schröder H, Izquierdo‐Pulido M, Ramón J, Villatoro M, et al. Preoperative predictors of weight loss at 4 years following bariatric surgery. Nutrition in Clinical Practice. 2015;30(3):420‐4. 432. Alvarado R, Alami R, Hsu G, Safadi B, Sanchez B, Morton J, et al. The impact of preoperative weight loss in patients undergoing laparoscopic Roux‐en‐Y gastric bypass. Obesity Surgery. 2005;15(9):1282‐6. 433. Gilbertson NM, Paisley AS, Kranz S, Weltman A, Kirby JL, Hallowell PT, et al. Bariatric surgery resistance: Using preoperative lifestyle medicine and/or pharmacology for metabolic responsiveness. Obesity Surgery. 2017;27(12):3281‐91. 434. Kim JJ. Evidence base for optimal preoperative preparation for bariatric surgery: Does mandatory weight loss make a difference? Current Obesity Reports. 2017;6(3):238‐45. 435. Contreras JE, Santander C, Bravo J. Correlation between age and weight loss after bariatric surgery. Obesity Surgery. 2013;23(8):1286‐9. 436. Miller SL, Wolfe RR. The danger of weight loss in the elderly. The Journal of Nutrition, Health & Aging. 2008;12(7):487‐91. 437. Wood GC, Mirshahi T, Still CD, Hirsch AG. Association of DiaRem score with cure of type 2 diabetes following bariatric surgery. JAMA Surgery. 2016;151(8):779‐81.

282

438. Mehaffey JH, Mullen MG, Mehaffey RL, Turrentine FE, Malin SK, Kirby JL, et al. Type 2 diabetes remission following gastric bypass: does diarem stand the test of time? Surgical endoscopy. 2017;31(2):538‐42. 439. Aminian A, Brethauer SA, Kashyap SR, Kirwan JP, Schauer PR. DiaRem score: external validation. The Lancet Diabetes & Endocrinology. 2014;2(1):12‐3. 440. Tharakan G, Scott R, Szepietowski O, Miras AD, Blakemore AI, Purkayastha S, et al. Limitations of the DiaRem score in predicting remission of diabetes following Roux‐En‐Y gastric bypass (RYGB) in an ethnically diverse population from a single institution in the UK. Obesity surgery. 2017;27(3):782‐6. 441. American Diabetes Association. Standards of Medical Care in Diabetes—2019. Diabetes Care. 2019;40(Supplement 1):S1‐S2. 442. Hosmer Jr DW, Lemeshow S, Sturdivant RX. Applied logistic regression: John Wiley & Sons; 2013. 443. Honarmand K, Chetty K, Vanniyasingam T, Anvari M, Chetty V. Type 2 diabetes remission rates 1‐year post‐Roux‐en‐Y gastric bypass and validation of the DiaRem score: the Ontario Bariatric Network experience. Clinical obesity. 2017;7(3):176‐82. 444. Shen S‐C, Wang W, Tam K‐W, Chen H‐A, Lin Y‐K, Wang S‐Y, et al. Validating risk prediction models of diabetes remission after sleeve gastrectomy. Obesity Surgery. 2019;29(1):221‐9. 445. Kam H, Tu Y, Pan J, Han J, Zhang P, Bao Y, et al. Comparison of four risk prediction models for diabetes remission after roux‐en‐Y gastric bypass surgery in obese Chinese patients with type 2 diabetes mellitus. Obesity Surgery. 2020:1‐11. 446. Ahuja A, Tantia O, Chaudhuri T, Khanna S, Seetharamaiah S, Majumdar K, et al. Predicting remission of diabetes post metabolic surgery: A comparison of ABCD, diarem, and DRS scores. Obesity Surgery. 2018;28(7):2025‐31. 447. Sampaio‐Neto J, Nassif LS, Branco‐Filho AJ, Bolfarini LA, Loro LS, SOUZA MPd, et al. External validation of the diarem score as remission predictor of diabetes mellitus type 2 in obese patients undergoing roux‐en‐y gastric bypass. ABCD Arquivos Brasileiros de Cirurgia Digestiva (São Paulo). 2015;28(Suppl 1):19‐22. 448. Alba AC, Agoritsas T, Walsh M, Hanna S, Iorio A, Devereaux P, et al. Discrimination and calibration of clinical prediction models: Users’ guides to the medical literature. JAMA. 2017;318(14):1377‐84. 449. Panunzi S, Carlsson L, De Gaetano A, Peltonen M, Rice T, Sjöström L, et al. Determinants of diabetes remission and glycemic control after bariatric surgery. Diabetes Care. 2016;39(1):166‐74. 450. Lee MH, Lee W‐J, Chong K, Chen J‐C, Ser K‐H, Lee Y‐C, et al. Predictors of long‐term diabetes remission after metabolic surgery. Journal of Gastrointestinal Surgery. 2015;19(6):1015‐21. 451. Chan JC, Malik V, Jia W, Kadowaki T, Yajnik CS, Yoon K‐H, et al. Diabetes in Asia: Epidemiology, risk factors, and pathophysiology. JAMA. 2009;301(20):2129‐40.

283

APPENDIX Post Bariatric Surgery Questionnaire Patient Label

Dear ______,

We note that it is over 12 months since you last attended the bariatric surgery follow-up clinic. It is important for your well-being that we review you at least once a year following surgery, to ensure that the surgery has not resulted in any adverse effects or caused any new problems. Furthermore, we need to monitor outcomes in order to receive ongoing government funding for our bariatric surgery service to continue.

If you are able to attend, please call us to make an appointment on (02) 95154220. If you are not able to attend, it is important that you undergo a health check with your General Practitioner (GP). This is to make sure that you do not have any nutritional deficiencies and to check on your metabolic health, even though you may feel fine. We are also interested to know the reason(s) why it might have been difficult for you to attend the follow-up review appointment, so that we can improve our service.

We would be grateful if you could complete this questionnaire and have the attached blood tests. Could you please also ask your GP to measure your weight, waist circumference and blood pressure, and send these back to us with a copy of your current medical conditions and medications in the enclosed self-addressed postage-paid envelope.

There are 20 questions that will only take approximately 10-15 minutes of your time to complete. Be assured that all answers you provide will be kept confidential. Responses will be compiled together and analysed as a group.

If you have any questions or concerns, please do not hesitate to contact:

Royal Prince Alfred Hospital: Ms. Elisia Manson Clinical Nurse Consultant at Metabolism & Obesity Services Tel: (02) 9515 4220 Fax: (02) 9515 5820 Email: [email protected]

Camden/Concord Hospital: Dr. Nic Kormas Senior Staff Specialist Endocrinologist Tel: (02) 9767 6747 Email: [email protected]

Date: ______

Name: Address:

Date of birth: Phone no:

Sex: Email:

284 Protocol X15-0339 & LNR/15/RPAH/463 Page 1 of 7

Instructions

Please answer all questions as accurately as possible. All answers are confidential and complete anonymity is assured. Your participation is voluntary and will help us and other patients greatly. We would like to thank you for your time.

Part I: Basic information

1) What is your current weight (in Kilograms)?

______Kg Estimated/Weighed (please circle)

2) What is your current waist circumference (in centimetres)? i.e. the narrowest horizontal measurement between the bottom of the ribs and the top of your hips.

______cm Estimated/Weighed (please circle)

3) What is your ethnicity? a. Caucasian b. Indigenous Australian c. Pacific Islander d. Middle Eastern e. Asian f. Other, namely ______

4) Do you use tobacco products (e.g. cigarettes, e-cigarettes, hookah)? If yes, how many on average per day? a. No, I have never smoked a cigarette in my life b. No, I have stopped smoking c. Yes, on average ______per day Please circle the type(s) of tobacco products used: cigarettes/e-cigarettes/hookah

5) Do you drink alcohol? i.e. beer, wine, spirits, ciders, other? Yes/No (please circle)

If yes, what type of alcohol and how many glasses (on average) do you drink per week? None 1 to 5 glasses 6 to 10 glasses 11 to 15 glasses 16 to 20 glasses More than 20 glasses Beer Wine Spirits Ciders Other

Protocol X15-0339 & LNR/15/RPAH/463 285 Page 2 of 7

6) What was your marital status when you had bariatric surgery? a. Single b. Married c. Divorced/Separated d. Widowed e. Other, namely ______

7) What is your marital status currently? a. Single b. Married c. Divorced/Separated d. Widowed e. Other, namely ______

8) Who do you live with? ______

9) What was your highest level of education at time of the surgery? a. Primary school b. High school not completed c. High school completed d. More than high school completed e. Tertiary education, please specify: ______f. Other, namely ______g. I don’t know

10) What is your highest level of education now (i.e. since the surgery)? a. Primary school b. High school not completed c. High school completed d. More than high school completed, please specify: ______e. Tertiary education, please specify: ______f. Other, namely ______g. I don’t know

Protocol X15-0339 & LNR/15/RPAH/463 286 Page 3 of 7

11) What was your status of employment before bariatric surgery? a. Housewife/Homemaker b. Unemployed c. Employed d. Self-employed e. Pensioned f. Other, namely ______

12) What is your current role? a. Housewife/Homemaker b. Unemployed c. Employed d. Self-employed e. Pensioned f. Other, namely ______

13) What type of work (paid or unpaid) do you do?

Part II: Medical information

14) Current medical conditions (Circle relevant answers) a. Type 2 diabetes mellitus Yes/No b. High blood pressure Yes/No c. Asthma/Chronic obstructive pulmonary disease/Emphysema Yes / No d. High cholesterol/Dyslipidaemia/Hyperlipidaemia Yes/No Yes/No e. Osteoarthritis and/or joint pain Yes/No f. Gastric reflux Yes/No g. Heart disease Yes/No h. Cancer Yes/No i. Depression Yes/No j. Anxiety Yes/No k. Gout l. Others, namely ______

Protocol X15-0339 & LNR/15/RPAH/463 287 Page 4 of 7

15) Have you had, or do you currently still have obstructive sleep apnoea (OSA)? a. No, I have never had obstructive sleep apnoea b. My sleep doctor has told me I no longer have obstructive sleep apnoea c. Yes, I still have obstructive sleep apnoea

16) Did you require CPAP to treat your sleep apnoea? a. I have never been prescribed CPAP b. I was prescribed CPAP but have not been using it, because ______c. I used CPAP, but do not need it anymore d. I still use CPAP

17) Have you had any medical problems or procedures since the operation? Yes/No If yes, please specify what and when: ______

18) What medication(s), vitamins and natural supplement(s) do you currently take? What is the dosage and when do you take it? e.g. Aspirin 100mg 1 tablet per day

a. ______

b. ______

c. ______

d. ______

e. ______

f. ______

g. ______

h. ______

i. ______

j. ______

Protocol X15-0339 & LNR/15/RPAH/463 288 Page 5 of 7

Part IV: Follow-up details

19) What was the reason(s) for not coming to follow-up appointments after your surgery? a. The distance (e.g. too far away/living outside the state or country) b. Work-related (e.g. could not combine with work) c. Family-related (e.g. family emergency) d. I regained weight e. I forgot about the appointment f. I was too ill to come for follow-up g. The follow-up was unnecessary in my opinion h. Someone at the hospital told me I did not have to come back for follow-up i. I did not like the experience of attending follow-up at the hospital clinic j. Other, namely ______

20) What do you think of the following aspects with respect to your experience with the process of bariatric surgery? Please tick the relevant box for each line: Very bad Bad Acceptable Good Very good

The overall satisfaction with surgery

Education and information provided prior to surgery

The usefulness of follow-up

The follow-up addressed individual needs

The support you got from the hospital staff

The expertise of the hospital staff

The administration and organization of the follow-up

The support you got from family and friends

Protocol X15-0339 & LNR/15/RPAH/463 289 Page 6 of 7

If you have any other remarks or suggestions for improving the follow-up of bariatric surgery, please feel free to leave your comments below:

______

The doctor/clinician to complete Values Date Weight (Kg) Waist circumference (cm) Blood pressure (mmHg)

Protocol X15-0339 & LNR/15/RPAH/463 290 Page 7 of 7