Statistical games to predict mortality and length of hospital stay after cardiac surgery

Abstract

Abstract Purpose Neutrophil-to-lymphocyte ratio (NLR), and platelet-to-lymphocyte ratio (PLR) are widely accepted indices positively correlated with the disease severity, progression, and mortality. Since in a previous study we had correlated nutritional parameters with morbidity and mortality in cardiac surgery patients, we decided to perform a post-hoc analysis to test whether these parameters are also correlated with the NLR and PLR indices. Methods NLR and PLR were calculated on days 0, 3, 5, and 7 postoperatively. ROC curve was generated to assess their prognostic value and multivariate logistic analysis to identify independent risk factors for 90day mortality. Results Analysis was performed on 179 patients-data, 11 of which (6.15%) died within 90 days. The discriminatory performance for predicting 90day mortality was better for NLR7 (AUC = 0.925, 95%CI: 0.865–0.984) than for NLR5 (AUC = 0.810, 95%CI: 0.678–0.942), the respective cut-off points being 7.10 and 6.60. PLR3 exhibited a significantly strong discriminatory performance (AUC = 0.714, 95%CI: 0.581–0.847), with a cut-off point of 126.34. Similarly, a significant discriminative performance was prominent for PLR3, NLR5, and NLR7 with respect to the length of hospital stay. Moreover, NLR7 (OR: 2.143, 95% CI: 1.076–4.267, p = 0.030) and ICU length of stay (OR:1.361, 95% CI: 1.045–1.774, p = 0.022) were significant independent risk factors. Conclusion NLR and PLR are efficient predictive factors for 90day mortality and hospital length of stay in cardiac surgery patients. Owing to the simplicity of determining NLR and PLR, their postoperative monitoring may offer a reliable predictor of patients’ outcomes in terms of length of stay and mortality.