Platelet to lymphocyte and neutrophil to lymphocyte ratios as strong predictors of mortality in intensive care population.

OBJECTIVE
Patients in intensive care units (ICU) have greater morbidity and mortality. We aimed to study neutrophil to lymphocyte ratio (NLR) and platelet to lymphocyte ratio (PLR) in the ICU population.


METHODS
Medical and laboratory data of patients treated in ICU were retrospectively analyzed. Patients were divided into deceased and survived groups.


RESULTS
The NLR of survived and deceased groups were 3.6 (0.2-31) and 9.5 (1-40), respectively (p<0.001). The PLR of the survived group (111 [16-537]) was significantly lower than the PLR of the deceased (209 [52-1143]), (p<0.001). An NLR higher than 4.9 had 84% sensitivity and 67% specificity is selecting deceased patients (AUC:0.80, p<0.001). A PLR higher than 112 had 83% sensitivity and 52% specificity in predicting deadly cases (AUC:0.76, p<0.001). Both PLR and NLR were significantly and positively correlated with c reactive protein levels.


CONCLUSION
We suggest that physicians should pay particular attention to the treatment of patients in ICU with elevated NLR and PLR.


INTRODUCTION
Patients with organ dysfunction and failure, subjects that need to be followed-up closely, and patients who need urgent treatment usually require treatment in Intensive Care Units (ICU). Since mortality and morbidity rates are much higher in the ICU population compared to other patients who do not require treatment in ICU, surrogate markers for predicting mortality in ICU subjects have been established. C-reactive protein (CRP) is one of these markers and usually increases in conditions associated with inflammation or infection 1 .
The utility of hemogram-derived inflammatory markers is on the rise recently. Several of hemogram markers have been suggested to be related to inflammatory situations and outcomes in the literature 2-5 .
square test was used to analyze categorical variables. The correlation between variables were conducted by Pearson's correlation analyze test. A ROC analysis was performed to define the cut off values of CRP, NLR, and PLR in predicting mortality. Statistical significance was set on a p-value lower than 0.05.

RESULTS
A total of 172 subjects were enrolled in the study; 95 in the survived group and 77 in the deceased. The mean age of the survived group was 63.3 ± 20.4 years and 71.5 ± 14.5 in the deceased group. Subjects in the deceased group were significantly older than those in the survived group (p=0.003). There were 41 women and 51 men in the survived group and 32 women and 45 men in the deceased. Gender was not significantly different between the survived and deceased groups (p=0.83).
Treatment durations in the intensive care unit for patients in the survived and deceased groups were 3 (1-118) and 6 (0-97) days, respectively. The difference was statistically significant (p=0.003).
Levels of WBC, plt, and plasma glucose were not significantly different between study groups (p>0.05 for all). Serum CRP (p<0.001) was significantly higher and Hb (p=0.02) and Htc (p=0.02), were significantly lower in the deceased group compared with the survived group.
Neutrophil to lymphocyte ratio (NLR) and platelet to lymphocyte ratio (PLR) are two novel markers reproduced from routine hemogram tests. The association between NLR and inflammatory conditions is reported in a recent study 6 . Moreover, Salciccioli et al. 7 proposed that NLR was related to the outcomes of critically ill subjects. Similarly, PLR was also regarded as a prognostic and inflammatory predictor in the ICU population 8 . However, further studies are needed to confirm the association between mortality and these markers and investigate the possible correlation between CRP and PLR and NLR.
In the present retrospective study, we aimed to compare the NLR and PLR values of deceased ICU subjects with those of patients who survived. We also aimed to find out whether PLR and NLR were correlated to the CRP.

METHODS
The medical records of patients treated in the intensive care unit of our institution between April 2017 and January 2018, were enrolled in the study. An approval of institutional directorate was obtained. The data of the subjects were recorded from an institutional computerized database and patient files. Pregnant patients, subjects with known hematologic disorders, on treatment that may interfere with platelet functions, and older than 18 years were excluded from the study. Patients were grouped into two groups according to the outcome, survived or deceased.
Age, gender, reason for ICU admission, duration of intensive care treatment, and laboratory parameters, including, white blood cell count (WBC), neutrophil count (neu), lymphocyte count (lym), hemoglobin (Hb), hematocrit (Htc), platelet count (plt), and c-reactive protein (cRP) were recorded. The NLR was calculated by merely dividing the neu by the lym, and PLR was calculated by dividing the plt by the lym.
Statistical analysis was done using SPSS software (SPSS 10.0 for Windows, IBM Inc., Chicago, Il, USA). Kolmogorov-Smirnov test was conducted to observe the distribution of parameters in the study groups. Homogenous variables were analyzed by independent samples t-test and expressed as mean +/-Standard Deviation, while non-homogenous variables were analyzed by Mann Whitney U test and expressed as median (min-max). Chi-

DISCUSSION
The results of the present study showed that the elevation of both PLR and NLR was associated with mortality in intensive care patients. Moreover, PLR and NLR were better than CRP in predicting mortality. Finally, both PLR and NLR were significantly correlated with serum CRP levels.
The proportion of NLR to lymphocyte count, namely NLR, can be used as an inflammatory marker in intensive care practice since the physiologic response of circulating leukocytes to inflammatory stress cause an elevation in neutrophil count and decrement in lymphocyte count 9 . Activated neutrophils secrete enzymes such as acid phosphatase, myeloperoxidase, and elastase, which cause tissue destruction 10 . Thus, NLR is proposed as a marker that reflects the inflammatory burden by elevated neutrophil count and indicates physiological stress and poor health condition by decreased lymphocyte count 11 . Indeed, mortality due to coronary heart disease increased with higher NLR levels in studies in the literature 12, 13 .
The NLR has been introduced as a simple but effective marker of inflammatory condition, since it was found to be correlated with the severity and prognosis of the subjects with sepsis when compared and evaluated along with prognostic scores, such as APACHE 2 (Acute Physiology and Chronic Health Evaluation II) and SOFA (Sepsis-related Organ Failure Assessment) 14 . Similar to the knowledge found in the literature, the present study suggests that NLR could serve as a prognostic factor in the ICU population.
The association between PLR and inflammatory conditions is relatively a more recent finding then the association between inflammation and NLR. Besides NLR, PLR also predicted the prognosis of subjects with hepatic cancer 15 . In another study from 2014, authors reported that both NLR and PLR elevated with the increasing stage of ovarian carci-nomas 16 . Moreover, it was found that NLR and PLR were higher in subjects with non-dipper hypertension compared with patients with dipper hypertension 17 . Shimoyama et al. 18 proposed that NLR and PLR were better than all other inflammation-related prognostic scores in predicting outcomes of subjects with gastrointestinal perforation. NLR and PLR were suggested as reliable prognostic indexes in renal cell carcinoma 19 , in advanced pancreas carcinoma 20 , in mesenteric arterial embolism and thrombosis 21 , and in the severity of gallstone pancreatitis 22 . Qi et al. 23 reported that PLR was an independent prognostic marker for survival in melanoma patients in a recently published study. The results of the present study emphasize the significant association between mortality of ICU subjects and elevated PLR suggested by the data in the literature.
Even though CRP is the most important and commonly used inflammatory marker in clinical practice, our results show that both PLR and NLR predicted mortality better than CRP. This is an outstanding result which could add a lot to current literature knowledge.
The inevitable limitation of our study is its retrospective design, which could cause selection bias. Another limitation could be the relatively small study population. However, the results of the present study found a significant association between the prognosis of ICU subjects and PLR and NLR and a significant correlation between CRP and both NLR and PLR. CONCLUSION We suggest that automatic blood count devices should be arranged to calculate PLR and NLR levels since higher PLR and NLR levels predict increased mortality in the ICU population. We also suggest that physicians should pay particular attention in the treatment of patients in ICU with elevated NLR and PLR.