Analysis of the Prognostic Value of the Ratio of Blood White Cells to Albumin in Critically Ill Patients with Malignant Gastrointestinal Tumors: A Retrospective Study Based on the MIMIC-IV Database

Background This study aimed to predict the prognosis of critical gastrointestinal malignancies using the white blood cell to serum albumin ratio in venous blood (WAR). Methods Utilizing the MIMIC IV 2.0 database, we selected patients diagnosed with gastrointestinal malignancies, specifically esophageal cancer, gastric cancer, and colorectal cancer. The primary outcomes assessed were in-hospital mortality and 28-day survival post-discharge. Logistic regression identified factors associated with in-hospital mortality, while controlling for confounding variables. The predictive ability of the model was evaluated using receiver operating characteristic (ROC) curves. Additionally, Cox regression analysis helped identify variables associated with 28-day mortality. A clinical model was developed, and a nomogram was created, with the C-index used to gauge predictive accuracy. Results The original dataset was randomly divided into a training cohort of 411 cases and a validation cohort of 172 cases. WAR was categorized into high and low groups based on an optimal threshold of 6.1. Kaplan-Meier survival analysis and logistic regression revealed a significant association between WAR and in-hospital mortality (fit index R²: 0.081, 95% CI: 1.151–1.303; P < 0.001). The area under the ROC curve was 0.7002 for the original dataset, 0.7184 for the training set, and 0.7158 for the validation set. The high WAR group displayed a significantly higher risk of 28-day mortality (hazard ratio HR: 1.65; 95% CI: 1.037–2.026; P < 0.05). The C-index for the clinical model nomogram was 0.787 for the original dataset, 0.797 for the training set, and 0.819 for the validation set, reflecting excellent predictive performance for short-term mortality. Conclusion WAR is a reliable and stable predictor of short-term prognosis in patients with severe gastrointestinal malignancies. The clinical model nomogram incorporating WAR exhibits robust predictive capabilities, demonstrating its utility in clinical practice.