[Monocyte/lymphocyte ratio as a predictor of 30-day mortality and adverse events in critically ill patients: analysis of the MIMIC-III database]
Yanxiu Li,Yun Liu,Chunlei Zhou,Zhongwen Zhang,Xiangrong Zuo,Jinghang Li,Quan Cao
DOI: https://doi.org/10.3760/cma.j.cn121430-20201223-00772
Abstract:Objective: To investigate the correlation of monocyte/lymphocyte ratio (MLR) with the prognosis and adverse event in critically ill patients. Methods: Basic information of patients were extracted from Medical Information Mart for Intensive Care-III (MIMIC-III), including demographics, blood routine, biochemical indexes, systemic inflammatory response syndrome score (SIRS), sequential organ failure assessment (SOFA) score, and outcome, etc. MLR on the first day of intensive care unit (ICU) admission was calculated. The receiver operating characteristic curve (ROC curve) was applied to evaluate the prognostic value of MLR on the 30-day mortality and its cut-off value. According to the cut-off value, the patients were divided into two groups, and the differences between the groups were compared. Logistic regression model was used to analyze the relationship of MLR with 30-day mortality, continuous renal replacement therapy (CRRT), mechanical ventilation, the length of ICU stay, and total hospitalization time. Results: (1) A total of 43 174 critically ill patients were included. ROC curve showed that area under ROC curve (AUC) of MLR in predicting 30-day mortality was 0.655 [95% confidence interval (95%CI) was 0.632-0.687]. The cut-off value of MLR calculated according to the maximum Yoden index was 0.5. There were 16 948 patients with MLR ≥ 0.5 (high MLR group) and 26 226 patients with MLR < 0.5 (low MLR group). (2) Compared with the low MLR group, the high MLR group had higher age, proportion of male, body mass index (BMI) [age (years old): 66.0 (51.7, 78.4) vs. 57.6 (27.1, 74.6), proportion of male: 57.2% vs. 52.5%, BMI (kg/m2): 26.5 (22.5, 31.1) vs. 24.7 (14.3, 29.7)]. The high MLR group also had higher incidence of complications (hypertension: 49.2% vs. 44.6%, chronic heart failure: 32.6% vs. 21.7%, diabetes mellitus: 27.0% vs. 23.4%, chronic obstructive pulmonary disease: 21.5% vs. 16.1%, renal insufficiency: 19.3% vs. 13.1%), and higher white blood cell count (WBC), blood glucose, lactate (Lac), serum creatinine (SCr), SIRS score and SOFA score [WBC (×109/L): 13.8 (9.6, 19.2) vs. 11.5 (8.4, 15.6), blood glucose (mmol/L): 8.66 (6.88, 11.49) vs. 8.27 (6.55, 10.88), Lac (mmol/L): 2.2 (1.5, 3.7) vs. 2.1 (1.4, 3.3), SCr (μmol/L): 106.1 (70.7, 176.8) vs. 88.4 (70.7, 132.6), SIRS score: 3 (2, 4) vs. 2 (2, 3), SOFA score: 4 (2, 7) vs. 3 (1, 5)]. The 30-day mortality, and the proportion of patients with length of ICU stay > 5 days, total hospitalization time > 14 days, CRRT and mechanical ventilation > 5 days were significantly higher in high MLR group (30-day mortality: 20.0% vs. 8.3%, length of ICU stay > 5 days: 33.2% vs. 20.4%, total hospitalization time > 14 days: 33.7% vs. 16.2%, CRRT: 3.6% vs. 0.7%, mechanical ventilation > 5 days: 18.4% vs. 5.7%), with statistically significant differences (all P < 0.05). (3) After adjusted with the related factors, multivariate Logistic regression analysis showed that elevated MLR was an independent risk factor for increased 30-day mortality [odd ratio (OR) = 1.54, 95%CI was 1.37-1.72, P < 0.001]. Moreover, the increased MLR was independently associated with the increased risk of usage of CRRT (OR = 2.77, 95%CI was 2.18-3.51), mechanical ventilation > 5 days (OR = 2.45, 95%CI was 2.21-2.72), the length of ICU stay > 5 days (OR = 2.29, 95%CI was 2.10-2.49), and total hospitalization time > 14 days (OR = 2.28, 95%CI was 2.08-2.49), all P < 0.001. Conclusions: Retrospective analysis of large sample shows that MLR elevation is an independent risk factor for 30-day mortality, usage of CRRT, prolonged mechanical ventilation time, prolonged hospitalization, prolonged length of ICU stay. MLR can be used for risk stratification of severe patients.