Estimating patient risk before heart surgery (HS) is crucial. Perioperative inflammation is associated with several complications and mortality. This study investigated blood cell count inflammatory indices (BCCII) to predict risks, including neutrophil-to-lymphocyte ratio (NLR), derivate NLR (DNLR), neutrophil-to-platelet-lymphocyte ratio (NLPR), lymphocyte-to-monocyte ratio, platelet-to-lymphocyte ratio (PLR), Systemic Inflammatory Index (SII), Systemic Inflammatory Reaction Index (SIRI), and Aggregate Index of Systemic Inflammation (AISI). Data from a cohort of patients undergoing on-pump HS at a single center in Brazil were retrospectively analyzed. Data were obtained from medical records and a laboratory analyzer, and SPSS version 20.0 was used for index calculations and statistical analyses. In total, 444 surgeries were performed, and 40 in-hospital deaths occurred. Except for PLR, all other indexes were independent predictors of death after multivariate adjustment (all p < 0.05). Discrimination performance was absent for PLR and AISI, and NLR, NLPR, and DNLR demonstrated the best area under the receiver operating characteristic curve (AUC > 0.7; all p < 0.0001). For survivors (n = 404), all indexes exhibited a correlation with the length of hospital stay (all p < 0.03), and NLR, NLPR, and DNLR were predictors (p < 0.026) of poor operative outcomes (acute myocardial infarction, cerebrovascular attack, cardiac arrest, low cardiac output, prolonged mechanical ventilation, renal failure, and sepsis). All BCCII scores were associated with length of hospital stay. Apart from PLR, all indexes were independent predictors of in-hospital mortality. Accuracy was highest for NLR, NLPR, and DNLR; for survivors, these three factors were good predictors of poor operative outcomes.
Read full abstract