In a recent systematic review and meta-analysis of 12 studies with 13,262 cardiac surgery patients, Perry et al [22] conclude that the perioperative NLR value is an independent predictor of short-term and long-term postoperative mortality, besides a considerable between-study statistical heterogeneity (I2 = 94.39%), explained basically by the study-level variables. In the light of their findings, we decided to perform the present post hoc analysis of data collected by our group for a primary cohort survey [17] where we have shown that cardiac surgery patients are at risk of nutritional status deterioration – as assessed by means of BIA, mainly phase angle and fat-free mass – positively related to morbidity and mortality; our data regarding NLR and PLR were thus evaluated as predictive indices for 90day mortality, trying to avoid the bias reported in the aforementioned meta-analysis.
Among all the parameters analyzed the NLR5 and NLR7, as well as PLR3, were found to exhibit good discriminatory performance for predicting 90day mortality. The multivariate analysis performed thereafter showed that NLR7 together with the ICU length of stay were independent risk factors for death, adjusted for age, gender, and MUST score. Prolonged ICU length of stay has also been confirmed as an independent risk factor in other studies, which reported that postoperative morbidity and mortality are increased in patients with prolonged ICU stay after cardiac surgery. Moreover, prolonged hospitalization is mainly associated with respiratory events and prolonged ventilator intubation time [23]. Thus, in our study, most of the patients who finally died had remained intubated for more than 48hours in the ICU. At this point, it should be also noted that in the multivariate analysis model, the preoperative MUST score, which is an indicator of patient's nutritional status, was not associated with increased mortality, because there was no difference detected in the median MUST score between the patients died and them who survived.
Furthermore, when dividing the patients according to “points” received – one point per each positive predictor NLR5, NLR7, and PLR3 – most of the patients having the highest score of 3 points, eventually died (9 out of 11). Moreover, PLR3, NLR5, and NLR7 have also good performance for predicting prolongation of hospital stay more than 7 days.
It is well known that the contribution of NLR and PLR in evaluating the immune status of the patients, and consequently their inflammatory response, has been studied thoroughly during the last two decades [24]. Moreover, during an overwhelming inflammatory response, lymphocytopenia and lymphocyte hypoactivity occur, due to B-cells and T-cells apoptosis, both contributing to greater mortality [25, 26]. In addition to lymphocytopenia, neutrophilia and inappropriate systemic neutrophil activation and migration within the microvasculature contribute to tissue damage and multiple organ failure [27]. Moreover, in cardiac surgery patients, besides the operational stress itself, the additional use of the stressful extracorporeal circuit triggers an unavoidable major immune response, accelerated through the contact of the blood products and the surfaces of the CPB tubes.
NLR and PLR have been evaluated, either separately or together, as predictive factors for mortality related to cardiovascular outcomes [28, 29]. The prognostic performance of these indices has been also evaluated by others on cardiac surgery patients [30–33]. In a recent retrospective study of 1,694 patients divided into two groups according to their preoperative NLR optimal cut-off point of 3.23, authors conclude that patients having an NLR value greater than 3.23 experienced greater mortality (OR:3.36, 95% CI: 1.63–6.91); other parameters, as the ICU stay were also affected [34]. The predictive value of preoperative NLR and PLR values for major adverse cardiovascular and cerebrovascular events has also been confirmed by Larmann et al. [35], who reported a cut-off point of > 204.4 for PLR and > 3.1 for NLR. However, it should be highlighted that both studies referred to the preoperative NLR and PLR values, this probably being the reason they report lower cut-off values compared to our results. In our study, we evaluated the postoperative values of NLR and PLR, which are definitely affected by the inflammatory response elicited by operational stress. Close similar are the results of Zhu et al. [36] who found the critical postoperative NLR value to be 7.23 when they correlated NLR to mortality after cardiac surgery. In the same manner, the optimal cut-off point for predicting mortality with NLR on postoperative day 1 was 7.28, in a recent retrospective study including 2,707 cardiac surgery patients.
Our post-hoc analysis also reveals that the lower the PLR value the higher the possibility of death. This could be easily explained by the fact that thrombocytopenia is a sign of both infection and inflammatory response; platelets interact with white blood cells or vascular endothelial cells directly, based on a contact-dependent mechanism, and indirectly through the secretion of inflammatory cytokines [37]. Thus, the involvement of platelets in the inflammatory process is noted both locally and systemically [38]. This recruitment of platelets in combination with their adhesion to white blood cells to enhance their effect reduces the absolute number of circulating platelets which is then reflected in the decrease of PLR [39]. This is why we demonstrate a reversed probability – the lower the PLR the higher the probability of death - while others by using preoperative values support a positive correlation [35].
It is of interest to mention that when the patients were divided according to the points received – one point per each positive predictor NLR5, NLR7, and PLR3 – we found that 9 out of 11 patients who died presented with all the parameters (PLR3, NL5, and NLR7) positive, whereas when none of the indexes was abnormal all the patients survived. This finding is in accordance with a previous study, where the predictive power increased in parallel with the number of abnormal parameters measured (NLR, red cell distribution width, and mean platelet volume) [15]. Based on all the aforementioned findings we could support the option that these parameters could be highly informative for the postoperative monitoring of cardiac surgery patients, easy to perform, and of a very low cost; they consist of a routine, daily practice and calculating them during the first 7 postoperative days, may let the patients who are at higher risk for adverse outcomes, including mortality, to be discriminated, even from the 3rd postoperative day.
It is commonly accepted that doctors have a perpetual desire and will to be able to predict the prognosis of their patients. However, this commitment should not let them forget that their role is to provide the best possible care to critically ill patients, regardless of whether the odds are in their favor or whether the patient's survival would be a minor miracle. Healthcare professionals should follow the “deontological theory” and try to gain the greatest good for the patient and act for the patients’ benefit [40]. Given the above, the strategy of predicting as soon as sooner which patients have a higher chance of experiencing worse morbidity or higher mortality aims at triggering the reflexes of physicians in time and alerting them to possible complications that may arise in the future. In this way, it will be possible to ensure maximum patient-centered medical care delivery with the lowest possible consumption of resources, as these are consumed on the basis of need rather than horizontally.
There are some limitations in this study. First, it is a post-hoc analysis of retrospectively collected data for the primary work. Second, based on the variation of hematological parameters, our sample size could be considered small. However, the conducted post-hoc power analysis has revealed a power of 81.2%; thus, the results are reliable and safe conclusions can be drawn.