|
|
Back to 2015 Annual Meeting Program
Readmission Following Gastric Cancer Resection Predicts Survival
Alexandra W. Acher*1, Malcolm H. Squires2, Ryan Fields3, George a. Poultsides4, Carl Schmidt5, Konstantinos I. Votanopoulos6, Timothy M. Pawlik7, Linda X. Jin3, Aslam Ejaz7, David Kooby2, Mark Bloomston5, David Worhunsky4, Edward a. Levine6, Neil D. Saunders5, Emily Winslow1, Clifford S. Cho1, Glen Leverson1, Shishir K. Maithel2, Sharon M. Weber1
1University of Wisconsin School of Medicine and Public Health, Madison, WI; 2General Surgery, Emory University, Atlanta, GA; 3General Surgery, Washington University School of Medicine, St. Louis, MO; 4General Surgery, Stanford University, Stanford, CA; 5General Surgery, Ohio State University, Columbus, OH; 6General Surgery, Wake Forest University, Winston-Salem, NC; 7General Surgery, The Johns Hopkins University, Baltimore, MD
BACKGROUND: Readmission following gastric cancer resection has not been studied. We evaluated the predictors of readmission following gastric cancer resection utilizing a multi-institutional database.
METHODS: Patients who underwent curative resection of gastric adenocarcinoma from 2000-2012 from 7 academic institutions of the U.S. Gastric Cancer Collaborative were analyzed. In-hospital deaths and palliative surgeries were excluded and readmission was defined as within 30 days of discharge. Univariate and multivariable logistic regression analyses were employed and survival analysis was conducted using Kaplan and Meier methods. Using parameter estimates, a readmission prediction tool was created to optimize negative predictive value (NPV).
RESULTS: Of the 855 patients, 121 patients (14.2%) were readmitted. Overall, 60% of readmitted patients had post-operative complications versus 39% of non-readmitted patients (p<0.0001). Univariate analysis identified age, cardiac disease, hypertension, ASA score > 3, smoking, BUN, pancreatectomy, increasing operation time, increasing blood loss, postoperative complication, type of nodal dissection, and total gastrectomy as risk factors for readmission. Multivariable analysis identified cardiac disease, postoperative complication, and pancreatectomy as independent risk factors for readmission (Table 1). The NPV of the readmission-prediction tool was 94%. Readmitted patients had a decreased median survival compared to non-readmitted patients (Table 2).
CONCLUSIONS: Using a prediction tool to identify high risk patients may allow focused readmission-prevention interventions in patients undergoing surgery for gastric cancer. The etiology of decreased median overall survival in readmitted patients should be the focus of future investigations.
Table 1: Multivariable Analysis of Readmission
| Variable | N (% of readmitted patients) | OR | 95% CI | p-value | | Cardiac Disease | 49 (26%) | 3.05 | 1.88-4.91 | <0.0001 | | Post-operative Complication | 73 (60%) | 2.78 | 1.78-4.40 | <0.0001 | | Pancreatectomy | 14 (12%) | 2.36 | 1.04-5.00 | 0.0294 |
Table 2: Median Overall Survival | Readmitted | Not Readmitted | | Median Overall Survival (months) | 28 | 41 |
Back to 2015 Annual Meeting Program
|
