Analysis of factors affecting the prognosis of ICU patients by multiple logistic regression model: a retrospective cohort study of 1 299 patients in 12 consecutive years

Abstract

To explore the prognostic factors of intensive care unit (ICU) patients. A retrospective cohort study was conducted. The clinical data of patients admitted to ICU of Beijing Geriatric Hospital from January 2005 to December 2016 were collected. According to the prognosis, the patients were divided into death group and survival group, and the mortality trend with time and age was observed. Survival and death were treated as dependent variables, while the gender, age, reason of ICU admission, acute physiology and chronic health evaluation II (APACHE II) score, whether undergoing invasive mechanical ventilation (MV), invasive MV reason, duration of invasive MV, whether successful weaning, whether ICU nosocomial infection, hospital acquired pneumonia (HAP), central line-associated bloodstream infection (CLABSI), acute kidney injury (AKI) occurred, whether undergoing continuous blood purification (CBP), whether septic shock occurred, whether undergoing invasive hemodynamic monitoring or blood transfusion and length of ICU stay were used as the independent variables. First, the clinical data was analyzed by univariate analysis. Second, the independent variables influencing the dependent variable were analyzed by logistic regression analysis to screen out prognostic factors of ICU patients. During the study period, 1 325 patients were treated, 26 patients with missing data, 1 299 patients in final group, 865 patients in survival group, and 434 in death group. The proportion of men in the patients was larger (67.7%), and the age span was larger (16-105 years old). The mortality rate in different years showed no significant difference (χ 2 = 16.712, P = 0.117), and with the increase of age, the mortality rate of ICU showed an upward trend (χ 2 = 16.399, P = 0.022). The univariate analysis showed that ICU deaths were unrelated to gender, but associated with age, APACHE II score, invasive MV, invasive MV reason, duration of invasive MV, successful weaning, ICU nosocomial infection, HAP, CLABSI, AKI, septic shock, whether CBP or invasive hemodynamic monitoring, blood transfusion, the reason of ICU admission and the length of ICU stay (all P < 0.05). The relevant factors with statistical difference found in univariate analysis were analyzed in logistic regression analysis, which showed that whether successful weaning [odds ratio (OR) = 0.016, 95% confidence interval (95%CI) = 0.010-0.025], and whether AKI (OR = 3.917, 95%CI = 2.331-6.582) or septic shock occurred (OR = 2.808, 95%CI = 1.604-4.915) were the dependent variables of death or survival (all P = 0.000). Regression coefficient (β value) of successful weaning was -4.155, suggesting that unsuccessful weaning patients were likely to die. The β value of AKI and septic shock was 1.365 and 1.033, suggesting that the patients with AKI or septic shock were more likely to die. Whether the success of weaning, whether the occurrence of AKI or septic shock are independent prognostic factors for the prognosis of ICU patients. Measures for the prevention of the above three aspects could improve the prognosis of ICU patients.