Background: Inflammatory markers have been used as prognostic factors in multiple malignancies.In cancer patients, critically ill, the utility of these have limited data.The aim of our study is to determine whether neutrophil lymphocyte ratio (NLR) or lymphocyte platelet ratio(PLR) are prognostic factors for mortality in critically ill patients. Methods: We retrospectively analyzed data of 79 patients with solid tumors admitted to ICU at Sanna-Aliada Clinic between January 2018 to December 2018. Inflammatory markers results were obtained from laboratory tests performed during the first 24h of admission to ICU. Receiving operating characteristic (ROC) curves were constructed and the sensitivity, specificity, predictive values and probability indicators for the NLR and PLR. Results: A total of 79 patients were assessed, 39 women and 40 men. The average age was 60.28 years, median of 61 ( 18 to 91). 51.9% had metastatic disease. The most frequent places were lung 12 (15.2 %) and brain 9 (11,4%) . The main cause for admission to ICU was infectious disease (40.5%). The analysis of normality (Kolmogorov-Smirnov test) indicates that the variables age, hemoglobin, leukocytes, platelets, neutrophils, lymphocytes, have a normal deviation while the other variables: lactate, PCR, neutrophil to lymphocyte ratio (NLR) , Platelet to lymphocyte ratio (PLR) are not distributed normally. Regarding mortality, 44 patients were alive at 30 days (66.7%), and 30 (45.5%) were alive at 90 days. The average stay in the ICU was 8.43 days, with a median of 6, (SD 7.17, 1 to 40 days), 22.8% died in the ICU. The evaluation of PLR and NLR as a mortality marker is significant for the group of patients admitted to the ICU due to a noninfectious pathology, generating an area under the curve (AUC) of 0.706 for NLR (95% CI, 0.535 - 0.876, p-value = 0.035) and 0.767 for PLR (95% CI, 0.615-0.918; p-value = 0.006); the optimal cut point by Youden’s index for NLR was 8.29 and 267.94 for PLR (Sensitivity: 76%, Specificity: 67%). In contrast, the group with infectious pathology, the AUC was 0.47 for NLR (p = 0.78) and 0.42 for PLR (p = 0.44). The relationship of the biomarkers with stay in ICU was also evaluated, finding a statistically significant association with the lactate value (p = 0.024, Kruskal-Wallis) Conclusions:

Inflammatory markers are useful as predictive markers of mortality in critically ill patients due to non-infectious causes.

The lactate value serves as a predictive factor of stay in the ICU for all the patients.

We suggest carrying out prospective studies to confirm the validity of our findings.