Risk model for right hemicolectomy based on 19,070 Japanese patients in the National Clinical Database

Abstract

Right hemicolectomy is a very common procedure throughout the world, although this procedure is known to carry substantial surgical risks. The present study aimed to develop a risk model for right hemicolectomy outcomes based on a nationwide internet-based database. The National Clinical Database (NCD) collected records on over 1,200,000 surgical cases from 3,500 Japanese hospitals in 2011. After data cleanup, we analyzed 19,070 records regarding right hemicolectomy performed between January 2011 and December 2011. The 30-day and operative mortality rates were 1.1 and 2.3 %, respectively. The 30-day mortality rates of patients after elective and emergency surgery were 0.7 and 6.0 %, respectively (P < 0.001). The odds ratios of preoperative risk factors for 30-day mortality were: platelet <50,000/μl, 5.6; ASA grade 4 or 5, 4.0; acute renal failure, 3.2; total bilirubin over 3 mg/dl, 3.1; and AST over 35 U/l, 3.1. The odds ratios for operative mortality were: previous peripheral vascular disease, 3.1; cancer with multiple metastases, 3.1; and ASA grade 4 or 5, 2.9. Sixteen and 26 factors were selected for risk models of 30-day and operative mortality, respectively. The c-index of both models was 0.903 [95 % confidence interval (CI) 0.877-0.928; P < 0.001] and 0.891 (95 % CI 0.873-0.908; P < 0.001), respectively. We performed the first reported risk stratification study for right hemicolectomy based on a nationwide internet-based database. The outcomes of right hemicolectomy in the nationwide population were satisfactory. The risk models developed in this study will help to improve the quality of surgical practice.