Kevin Seyssel1, Michel Suter2,3, François Pattou4, Robert Caiazzo4, Helene Verkindt4, Violeta Raverdy4, Mathieu Jolivet5,6, Emmanuel Disse6,7, Maud Robert5,6, Vittorio Giusti8. 1. Department of Physiology, University of Lausanne (UNIL), Lausanne, Switzerland. 2. Department of Visceral Surgery, University Hospital (CHUV), Lausanne, Switzerland. 3. Department of Surgery, Riviera-Chablais Hospital, Aigle, Monthey, Switzerland. 4. University of Lille, CHU Lille Endocrine and Metabolic Surgery, Inserm UMR 1190 Translational Research for Diabetes, Lille, France. 5. Department of Digestive Surgery, Center of Bariatric Surgery, Hospital Edouard Herriot, Hospices Civils de Lyon, Lyon, France. 6. Fédération Hospitalo-Universitaire DO-IT, Centre Intégré et Spécialisé de L'Obésité de Lyon, Université Lyon 1, CRNH-RA, Hospices Civils de Lyon, Lyon, France. 7. Department of Endocrinology Diabetology and Nutrition, Claude Bernard Lyon 1 University, Lyon Sud University Hospital, Hospices Civils de Lyon, Lyon, France. 8. Metabolic Center, Hôpital du Valais, Sion, Switzerland. giustivik@gmail.com.
Abstract
INTRODUCTION: Different factors, such as age, gender, preoperative weight but also the patient's motivation, are known to impact outcomes after Roux-en-Y gastric bypass (RYGBP). Weight loss prediction is helpful to define realistic expectations and maintain motivation during follow-up, but also to select good candidates for surgery and limit failures. Therefore, developing a realistic predictive tool appears interesting. PATIENTS/ METHODS: A Swiss cohort (n = 444), who underwent RYGBP, was used, with multiple linear regression models, to predict weight loss up to 60 months after surgery considering age, height, gender and weight at baseline. We then applied our model on two French cohorts and compared predicted weight to the one finally reached. Accuracy of our model was controlled using root mean square error (RMSE). RESULTS: Mean weight loss was 43.6 ± 13.0 and 40.8 ± 15.4 kg at 12 and 60 months respectively. The model was reliable to predict weight loss (0.37 < R2 < 0.48) and RMSE between 5.0 and 12.2 kg. High preoperative weight and young age were positively correlated to weight loss, as well as male gender. Correlations between predicted weight and real weight were highly significant in both validation cohorts (R ≥ 0.7 and P < 0.01) and RMSE increased throughout follow-up between 6.2 and 15.4 kg. CONCLUSION: Our statistical model to predict weight loss outcomes after RYGBP seems accurate. It could be a valuable tool to define realistic weight loss expectations and to improve patient selection and outcomes during follow-up. Further research is needed to demonstrate the interest of this model in improving patients' motivation and results and limit the failures.
INTRODUCTION: Different factors, such as age, gender, preoperative weight but also the patient's motivation, are known to impact outcomes after Roux-en-Y gastric bypass (RYGBP). Weight loss prediction is helpful to define realistic expectations and maintain motivation during follow-up, but also to select good candidates for surgery and limit failures. Therefore, developing a realistic predictive tool appears interesting. PATIENTS/ METHODS: A Swiss cohort (n = 444), who underwent RYGBP, was used, with multiple linear regression models, to predict weight loss up to 60 months after surgery considering age, height, gender and weight at baseline. We then applied our model on two French cohorts and compared predicted weight to the one finally reached. Accuracy of our model was controlled using root mean square error (RMSE). RESULTS: Mean weight loss was 43.6 ± 13.0 and 40.8 ± 15.4 kg at 12 and 60 months respectively. The model was reliable to predict weight loss (0.37 < R2 < 0.48) and RMSE between 5.0 and 12.2 kg. High preoperative weight and young age were positively correlated to weight loss, as well as male gender. Correlations between predicted weight and real weight were highly significant in both validation cohorts (R ≥ 0.7 and P < 0.01) and RMSE increased throughout follow-up between 6.2 and 15.4 kg. CONCLUSION: Our statistical model to predict weight loss outcomes after RYGBP seems accurate. It could be a valuable tool to define realistic weight loss expectations and to improve patient selection and outcomes during follow-up. Further research is needed to demonstrate the interest of this model in improving patients' motivation and results and limit the failures.
Authors: Lars Sjöström; Anna-Karin Lindroos; Markku Peltonen; Jarl Torgerson; Claude Bouchard; Björn Carlsson; Sven Dahlgren; Bo Larsson; Kristina Narbro; Carl David Sjöström; Marianne Sullivan; Hans Wedel Journal: N Engl J Med Date: 2004-12-23 Impact factor: 91.245
Authors: Lars Fischer; Felix Nickel; Johannes Sander; Alexander Ernst; Thomas Bruckner; Beate Herbig; Markus W Büchler; Beat P Müller-Stich; Rune Sandbu Journal: Surg Obes Relat Dis Date: 2014-03-12 Impact factor: 4.734
Authors: Benjamin Fox; Ellie Chen; Andrew Suzo; Sally Jolles; Jacob A Greenberg; Guilherme M Campos; Corrine I Voils; Luke M Funk Journal: J Surg Res Date: 2015-04-09 Impact factor: 2.192
Authors: Robert Caiazzo; Guillaume Lassailly; Emmanuelle Leteurtre; Gregory Baud; Hélène Verkindt; Violeta Raverdy; David Buob; Marie Pigeyre; Philippe Mathurin; François Pattou Journal: Ann Surg Date: 2014-11 Impact factor: 12.969
Authors: Izabela A Karpińska; Jan Kulawik; Magdalena Pisarska-Adamczyk; Michał Wysocki; Michał Pędziwiatr; Piotr Major Journal: Obes Surg Date: 2021-03-13 Impact factor: 4.129
Authors: Erik Stenberg; Ingmar Näslund; Carina Persson; Eva Szabo; Magnus Sundbom; Johan Ottosson; Erik Näslund Journal: Int J Obes (Lond) Date: 2020-07-10 Impact factor: 5.095