Preoperative risk factors for postoperative intra-abdominal infectious complication after gastrectomy for gastric cancer using a Japanese web-based nationwide database

• Published in: Gastric cancer : official journal of the International Gastric Cancer Association and the Japanese Gastric Cancer Association 2021-01, Vol.24 (1), p.205-213
• Main Authors: Fujiya, Keiichi, Kumamaru, Hiraku, Fujiwara, Yoshiyuki, Miyata, Hiroaki, Tsuburaya, Akira, Kodera, Yasuhiro, Kitagawa, Yuko, Konno, Hiroyuki, Terashima, Masanori
• Format: Article
• Language: English
• Subjects: Abdomen; Abdominal Surgery; Activities of daily living; Anastomotic leak; Body mass index; Cancer Research; Gastrectomy; Gastric cancer; Gastroenterology; Medicine; Medicine & Public Health; Oncology; Original Article; Pancreas; Patients; Risk assessment; Risk factors; Risk groups; Surgical Oncology; Vascular diseases
• ISSN: 1436-3291; 1436-3305
• DOI: 10.1007/s10120-020-01083-3

Background Postoperative intra-abdominal infectious complication (PIIC) after gastrectomy for gastric cancer worsens in-hospital death or long-term survival. However, the methodology for PIIC preoperative risk assessment remains unestablished. We aimed to develop a preoperative risk model for postgastrectomy PIIC. Methods We collected 183,936 patients’ data on distal or total gastrectomy performed in 2013–2016 for gastric cancer from the Japanese National Clinical Database and divided into development (2013–2015; n  = 140,558) and validation (2016; n  = 43,378) cohort. The primary outcome was the incidence of PIIC. The risk model for PIIC was developed using 18 preoperative factors: age, sex, body mass index, activities of daily living, 12 comorbidity types, gastric cancer stage, and surgical procedure in the development cohort. Secondarily, we developed another model based on the new scoring system for clinical use using selected factors. Results The overall incidence of PIIC was 4.7%, including 2.6%, 1.7%, and 1.3% in anastomotic leakage, pancreatic fistula, and intra-abdominal abscess, respectively. Among the 18 preoperative factors, male [odds ratio, (OR) 1.92], obesity (OR, 1.52–1.96), peripheral vascular disease (OR, 1.55), steroid use (OR, 1.83), and total gastrectomy (OR, 1.89) strongly correlated with PIIC incidence. The entire model using the 18 factors had good discrimination and calibration in the validation cohort. We selected eight relevant factors to create a simple scoring system, using which we categorized the patients into three risk groups, which showed good calibration. Conclusion Using nationwide clinical practice data, we created a preoperative risk model for postgastrectomy PIIC for gastric cancer.