Total Gastrectomy Risk Model

Abstract

To construct a risk model for total gastrectomy outcomes using a nationwide Internet-based database. Total gastrectomy is a very common procedure in Japan. This procedure is among the most invasive gastrointestinal procedures and is known to carry substantial surgical risks. The National Clinical Database was used to retrieve records on more than 1,200,000 surgical cases from 3500 hospitals in 2011. After data cleanup, 20,011 records from 1623 hospitals were analyzed for procedures performed between January 1, 2011, and December 31, 2011. The average patient age was 68.9 years; 73.7% were male. The overall morbidity was 26.2%, with a 30-day mortality rate of 0.9%, in-hospital mortality rate of 2.2%, and overall operative mortality rate of 2.3%. The odds ratios for 30-day mortality were as follows: ASA (American Society of Anesthesiologists) grade 4 or 5, 9.4; preoperative dialysis requirement, 3.9; and platelet count less than 50,000 per microliter, 3.1. The odds ratios for operative mortality were as follows: ASA grade 4 or 5, 5.2; disseminated cancer, 3.5; and alkaline phosphatase level of more than 600 IU/L, 3.1. The C-index of 30-day mortality and operative mortality was 0.811 (95% confidence interval [CI], 0.744-0.879) and 0.824 (95% CI, 0.781-0.866), respectively. We have performed the first reported risk stratification study for total gastrectomy, using a nationwide Internet-based database. The total gastrectomy outcomes in the nationwide population were satisfactory. The risk models that we have created will help improve the quality of surgical practice.