Predict models for prolonged ICU stay using APACHE II, APACHE III and SAPS II scores: A Japanese multicenter retrospective cohort study

Prognostic value of lactic dehydrogenase-to-albumin ratio in critically ill patients with acute respiratory distress syndrome: a retrospective cohort study

Background

Lactic dehydrogenase (LDH)-to-albumin ratio (LAR) was an independent risk factor for mortality in the patients with acute respiratory distress syndrome (ARDS) due to coronavirus disease 2019 (COVID-19), while the relationship among LAR and short-term, long-term, in-hospital mortalities of ARDS remains unclear. The current study aims to investigate the association between LAR and significant prognosis in patients with ARDS.

Methods

We conducted a retrospective cohort study and analyzed patients with ARDS on the Medical Information Mart for Intensive Care IV (MIMIC-IV) version 2.0 database. In the current study, 30-day mortality was defined as the primary outcome; 90-day mortality and in-hospital mortality were defined as secondary outcomes. Multivariate regression analysis, Kaplan-Meier curve analysis and subgroup analysis were performed to research the association between LAR and prognosis in patients with ARDS.

Results

A total of 358 critically ill patients with ARDS were enrolled in the current study. The mean age of the participants was 62.6±16.0 and the median of LAR was 14.3. According to the Kaplan-Meier curve analysis, the higher LAR group had a higher 30-day, 90-day and in-hospital mortalities. We also analyzed the 30-day mortality to receiver operating characteristic (ROC) curves by comparing the value between LAR and LAR + simplified acute physiology score II (SAPS II). The area under the curve (AUC) of the LAR group was 0.694 [95% confidence interval (CI): 0.634-0.754, P<0.001], and 0.661 for the LAR + SAPS II (95% CI: 0.599-0.722, P<0.001). For 30-day mortality, after adjusting for covariates, hazard ratios (HRs) (95% CIs) for tertile 2 (LAR 8.7-30.9) and tertile 3 (LAR >30.9) were 2.00 (1.37, 2.92) and 2.50 (1.50, 4.15), respectively. Similar results were also observed for 90-day mortality and in-hospital mortality.

Conclusions

Elevated LAR levels are associated with increased 30- and 90-day mortalities, as well as in-hospital mortality in patients with ARDS, which means LAR levels may predict the mortalities of ARDS patients.

Factor analysis based on SHapley Additive exPlanations for sepsis-associated encephalopathy in ICU mortality prediction using XGBoost - a retrospective study based on two large database

Objective

Sepsis-associated encephalopathy (SAE) is strongly linked to a high mortality risk, and frequently occurs in conjunction with the acute and late phases of sepsis. The objective of this study was to construct and verify a predictive model for mortality in ICU-dwelling patients with SAE.

Methods

The study selected 7,576 patients with SAE from the MIMIC-IV database according to the inclusion criteria and randomly divided them into training (n = 5,303, 70%) and internal validation (n = 2,273, 30%) sets. According to the same criteria, 1,573 patients from the eICU-CRD database were included as an external test set. Independent risk factors for ICU mortality were identified using Extreme Gradient Boosting (XGBoost) software, and prediction models were constructed and verified using the validation set. The receiver operating characteristic (ROC) and the area under the ROC curve (AUC) were used to evaluate the discrimination ability of the model. The SHapley Additive exPlanations (SHAP) approach was applied to determine the Shapley values for specific patients, account for the effects of factors attributed to the model, and examine how specific traits affect the output of the model.

Results

The survival rate of patients with SAE in the MIMIC-IV database was 88.6% and that of 1,573 patients in the eICU-CRD database was 89.1%. The ROC of the XGBoost model indicated good discrimination. The AUCs for the training, test, and validation sets were 0.908, 0.898, and 0.778, respectively. The impact of each parameter on the XGBoost model was depicted using a SHAP plot, covering both positive (acute physiology score III, vasopressin, age, red blood cell distribution width, partial thromboplastin time, and norepinephrine) and negative (Glasgow Coma Scale) ones.

Conclusion

A prediction model developed using XGBoost can accurately predict the ICU mortality of patients with SAE. The SHAP approach can enhance the interpretability of the machine-learning model and support clinical decision-making.

Standardized pharmacological management of delirium after on-pump cardiac surgery reduces ICU stay and ventilation in a retrospective pre-post study

Cardiac surgery patients not only undergo a highly invasive procedure but are at risk for a diversity of postoperative complications. Up to 53% of these patients suffer from postoperative delirium (POD). This severe and common adverse event increases mortality and prolonged mechanical ventilation and extends the intensive care unit stay. The objective of this study was to test the hypothesis that standardized pharmacological management of delirium (SPMD) may reduce the length of stay in the intensive care unit (ICU), duration of postoperative mechanical ventilation, and the incidence of postoperative complications such as pneumonia or bloodstream infections in on-pump cardiac surgery ICU patients. In this retrospective, single-center observational cohort study, 247 patients were examined between May 2018 to June 2020, who underwent on-pump cardiac surgery, suffered from POD, and received pharmacological POD treatment. 125 were treated before and 122 after SPMD implementation in the ICU. The primary endpoint was a composite outcome, including the length of ICU stay, postoperative mechanical ventilation time, and ICU survival rate. The secondary endpoints were complications including postoperative pneumonia and bloodstream infections. Although the ICU survival rate was not significantly different between both groups, the length of ICU stay (control group: 23 ± 27 days; SPMD group: 16 ± 16 days; p = 0.024) and the duration of mechanical ventilation were significantly reduced in the SPMD-cohort (control group: 230 ± 395 h; SPMD group: 128 ± 268 h; p = 0.022). Concordantly, the pneumonic risk was reduced after SPMD introduction (control group: 44.0%; SPMD group: 27.9%; p = 0.012) as well as the incidence for bloodstream infections (control group: 19.2%; SPMD group: 6.6%; p = 0.004). Standardized pharmacological management of postoperative delirium in on-pump cardiac surgery ICU patients reduced the length of ICU stay and duration of mechanical ventilation significantly, leading to a decrease in pneumonic complications and bloodstream infections.