LAUSR

BACKGROUND Total gastrectomy for gastric cancer is associated with significant 30-day mortality, but this endpoint may underestimate the short-term mortality of the procedure. METHODS Retrospective analysis was performed using the National Cancer Database (2004-2015). Patients who underwent total gastrectomy for stage I to III gastric adenocarcinoma were identified and divided into cohorts based on 90-day mortality. Predictors of mortality were analyzed using multivariable logistic regression, and annual trends in mortality rates were calculated by Joinpoint Regression. RESULTS Of the 5,484 patients who underwent total gastrectomy, 90-day and 30-day mortality rates were 9.1% and 4.7%, respectively. Factors associated with 90-day mortality included increasing age (odds ratio 1.0, P < .001), income below the median (odds ratio 1.2, P = .039), Charlson-Deyo score ≥2 (odds ratio 1.4, P = .039), treatment at low-volume facilities (odds ratio 1.5, P < .001), N1 (odds ratio 2.0, P < .001), N2 (odds ratio 2.0, P < .001), or N3 (odds ratio 2.7, P < .001) stage disease, having <16 lymph nodes harvested (odds ratio 1.5, P < .001), and lack of treatment with chemotherapy (3.7, P < .001). Lack of health insurance (odds ratio 4.1, P = .080), and positive microscopic margins (odds ratio 1.3, P = .080) were correlated, but not significantly associated, with 90-day mortality. The 90-day mortality rate significantly declined from 14.3% in 2004 to 7.9% in 2015 (P = .006), and the 30-day mortality rate significantly declined from 7.7% in 2004 to 4.8% in 2015 (P = .009). CONCLUSION Nearly half of the deaths within 90 days after total gastrectomy for cancer occur beyond 30 days postoperative. Ninety-day mortality has improved over time, but rates remain high, suggesting the need for improved out-of-hospital postoperative care beyond 30 days.