Internal and External Validation of a multivariable Model to Define Hospital-Acquired Pneumonia After Esophagectomy

Abstract

Background: Pneumonia is an important complication following esophagectomy; however, a wide range of pneumonia incidence is reported. The lack of one generally accepted definition prevents valid inter-study comparisons. We aimed to simplify and validate an existing scoring model to define pneumonia following esophagectomy. Patients and methods: The Utrecht Pneumonia Score, comprising of pulmonary radiography findings, leucocyte count, and temperature, was simplified and internally validated using bootstrapping in the dataset (n = 185) in which it was developed. Subsequently, the intercept and (shrunk) coefficients of the developed multivariable logistic regression model were applied to an external dataset (n = 201) Results: In the revised Uniform Pneumonia Score, points are assigned based on the temperature, the leucocyte, and the findings of pulmonary radiography. The model discrimination was excellent in the internal validation set and in the external validation set (C-statistics 0.93 and 0.91, respectively); furthermore, the model calibrated well in both cohorts. Conclusion: The revised Uniform Pneumonia Score (rUPS) can serve as a means to define post-esophagectomy pneumonia. Utilization of a uniform definition for pneumonia will improve inter-study comparability and improve the evaluations of new therapeutic strategies to reduce the pneumonia incidence.