Scholarlink

Introduction <p>Hypoalbuminemia is a potential postoperative complication of one‑anastomosis gastric bypass (OAGB). This study aims to identify predictors of hypoalbuminemia and develop a machine learning based clinical scoring system to stratify risk in OAGB patients.</p> Methods <p>This retrospective study analyzed data from patients who underwent primary OAGB at a single academic metabolic and bariatric surgery (MBS) center over the past six years. Key demographic, clinical, and laboratory variables were compared between groups. Machine learning based model was developed to predict hypoalbuminemia after OAGB surgery, then, the top-performing model was selected, and feature importance analysis was conducted using an explainable AI (XAI) technique to identify the most influential variables. Subsequently, a scoring system was generated using AutoScore to predict hypoalbuminemia.</p> Results <p>From 2018 to 2023, 3,242 patients who underwent OAGB at our academic center were assessed for this study, and the patients were classified into the Patients Without Hypoalbuminemia group and the Patients with Hypoalbuminemia group. Potential risk factors for hypoalbuminemia included age ≥ 41, and biliopancreatic limb length adjusted for height ≥ 101, lower height, and lower ferritin and hemoglobin level. The clinical prediction score showed an area under the receiver-operating-characteristic curve (AUROC) of 0.80 with a sensitivity of 83.3% and specificity of 67% at the optimal cutoff. This analysis revealed significant differences in patients in the risk of hypoalbuminemia between low-risk (score &lt; 55) and high-risk (score ≥ 55) groups.</p> Conclusion <p>Older age, lower height, severe obesity, and higher biliopancreatic limb length adjusted for height ratio independently predict hypoalbuminemia after OAGB. The score system stratifies risk and may guide intensified nutritional monitoring during the vulnerable early postoperative period.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Unveiling Predictors of Hypoalbuminemia Following One Anastomosis Gastric Bypass (OAGB): A Retrospective Analysis

  • 期刊论文

  • 发表日期：

    2025年11月7日

查看全文
  • Arshad Ali,
  • Shahab Shahabi,
  • Delaram Moosavi,
  • Seyed Mohammad Mahdi Khadem,
  • Farid Ahmad Qaderi,
  • Fahimeh Yarigholi,
  • Abdolreza Pazouki,
  • Mohammad Kermansaravi

摘要

Introduction

Hypoalbuminemia is a potential postoperative complication of one‑anastomosis gastric bypass (OAGB). This study aims to identify predictors of hypoalbuminemia and develop a machine learning based clinical scoring system to stratify risk in OAGB patients.

Methods

This retrospective study analyzed data from patients who underwent primary OAGB at a single academic metabolic and bariatric surgery (MBS) center over the past six years. Key demographic, clinical, and laboratory variables were compared between groups. Machine learning based model was developed to predict hypoalbuminemia after OAGB surgery, then, the top-performing model was selected, and feature importance analysis was conducted using an explainable AI (XAI) technique to identify the most influential variables. Subsequently, a scoring system was generated using AutoScore to predict hypoalbuminemia.

Results

From 2018 to 2023, 3,242 patients who underwent OAGB at our academic center were assessed for this study, and the patients were classified into the Patients Without Hypoalbuminemia group and the Patients with Hypoalbuminemia group. Potential risk factors for hypoalbuminemia included age ≥ 41, and biliopancreatic limb length adjusted for height ≥ 101, lower height, and lower ferritin and hemoglobin level. The clinical prediction score showed an area under the receiver-operating-characteristic curve (AUROC) of 0.80 with a sensitivity of 83.3% and specificity of 67% at the optimal cutoff. This analysis revealed significant differences in patients in the risk of hypoalbuminemia between low-risk (score < 55) and high-risk (score ≥ 55) groups.

Conclusion

Older age, lower height, severe obesity, and higher biliopancreatic limb length adjusted for height ratio independently predict hypoalbuminemia after OAGB. The score system stratifies risk and may guide intensified nutritional monitoring during the vulnerable early postoperative period.