Purpose: The present study was to investigate if five values that are part of the hemogram analysis routinely checked before heart surgeries can be used as a high-quality, quick, low-cost, and easy-to-use outcome predictor.
Methods: This investigation was a retrospective, observational, cross-sectional study. Univariate and multivariate logistic regression was used to identify independent predictors for combined adverse events. We enrolled 1500 consecutive patients who underwent elective, on-pump, open-heart surgery from 2011 to 2014. Preoperative hemogram evaluation, red cell distribution width (RDW), mean platelet volume (MPV), platelet-to-lymphocyte ratio (PLR), and neutrophil-to-lymphocyte ratio (NLR) were recorded. We classified combined adverse events (CAE) as (1) myocardial infarction, (2) cardiac reoperation, (3) prolonged mechanical ventilation, (4) prolonged hospital stay, (5) rehospitalization, or (6) mortality.
Results: It was found that several parameters obtained as part of the hemogram, namely RDW, MPV, PLR, and NLR, can predict, individually or in combinations, the outcomes in open-heart surgery patients. It was found that the prediction success of NLR (4.8 fold) was higher compared to RDW (1.8 fold) and MPV (1.5 fold). When the prediction success of the combined parameters was investigated, the NLR-RDW (4.7 fold) pair was found higher in the prediction of CAE occurrence. The predictive success of the triple combination of NLR-MPV-RDW (5.5 fold) was higher than other combinations.
Conclusions: The triple combination of parameters obtained as part of the hemogram, NLR-RDW-MPV, indicated a much more predictive power than two parameters coupled. This combination of three parameters, NLR-RDW-MPV, is to be considered as a sensitive, high-quality, low-cost outcome prediction marker for cardiac surgery patients that is less time consuming and easy to use.
Keywords: Anesthesia; Mean platelet volume; Neutrophil-to-lymphocyte ratio; Platelet-to-lymphocyte ratio; Predictive data after cardiac surgery; Red cell distribution width.