Morbidity and mortality of gastrectomy for cancer in Department of Veterans Affairs Medical Centers

BACKGROUND: The purpose of this study was to define risk factors that predict 30-day morbidity and mortality after gastrectomy for cancer in Veterans Affairs (VA) Medical Centers. METHODS: The VA National Surgical Quality Improvement Program prospectively collected data on 708 patients undergoing gastrectomy for cancer in 123 participating VA medical centers from 1991 to 1998. Independent variables included 68 preoperative patient characteristics and 12 intraoperative variables; the dependent variables were 21 defined adverse outcomes and death. Predictive models for 30-day morbidity and mortality were constructed by using stepwise logistic regression analysis. RESULTS: The 30-day morbidity rate was 33.3% (236 of 708). The overall 30-day mortality rate was 7.6% (54 of 708). Significant positive predictors of morbidity (P <.05) included current pneumonia, American Society of Anesthesiologists class IV (threat to life), partially dependent functional status, dyspnea on minimal exertion, preoperative transfusion, extended operative time, and increasing age. Significant positive predictors of mortality (P <.05) included do not resuscitate status, prior stroke, intraoperative transfusion, preoperative weight loss, preoperative transfusion, and elevated preoperative alkaline phosphatase level. CONCLUSIONS: Risk factors predicting morbidity and mortality rates at VA hospitals after gastrectomy for gastric cancer are reported by using a prospectively collected, multi-institutional database. Assigning relative weights to factors associated with adverse outcomes may help improve patient care.