Recent studies have shown that body composition is an important factor that affects surgical site infection (SSI). However, each study has utilized different body composition criteria. Therefore, in this study, we aim to determine the most predictable body composition criteria for the prediction of SSI after gastrectomy. The visceral fat area (VFA), subcutaneous fat area (SFA), and muscle area were assessed by a preoperative-stage computed tomographic (CT) scan. To compare the predictive performance of body composition for SSI, logistic regression models were used, and the models were compared using the receiver operation characteristic (ROC) curve and the area under the curve (AUC) value. Of the 1038 eligible patients, 58 patients (5.6%) developed SSI. The VFA-to-SFA ratio showed the best predictive performance (mean AUC 75.11). The cutoff value for the SSI of the VFA-to-SFA ratio was 0.94, and the sensitivity and specificity were 67.86% and 77.65%, respectively. A multivariate logistic analysis indicated that a total gastrectomy (OR, 2.13; p = 0.017), stage III or IV cancer (OR, 2.66; p = 0.003), and a high VFA-to-SFA ratio (OR, 8.09; p < 0.001) were independent risk factors for SSI after gastrectomy. The VFA-to-SFA ratio is the most predictable body composition model for use in predicting the incidence of SSI after gastrectomy.
Keywords: body composition; gastrectomy; gastric cancer; outcomes; risk factors; surgical site infection; visceral-to-subcutaneous fat ratio.