Development and External Verification of a Nomogram for Patients with Persistent Acute Kidney Injury in the Intensive Care Unit

Albumin corrected anion gap for predicting in-hospital death among patients with acute myocardial infarction: A retrospective cohort study

OBJECTIVE

To explore the relationship between Anion Gap (AG), Albumin Corrected AG (ACAG), and in-hospital mortality of Acute Myocardial Infarction (AMI) patients and develop a prediction model for predicting the mortality in AMI patients.

METHODS

This was a retrospective cohort study based on the Medical Information Mart for Intensive Care (MIMIC)-Ⅲ, MIMIC-IV, and eICU Collaborative Study Database (eICU). A total of 9767 AMI patients who were admitted to the intensive care unit were included. The authors employed univariate and multivariable cox proportional hazards analyses to investigate the association between AG, ACAG, and in-hospital mortality; p < 0.05 was considered statistically significant. A nomogram incorporating ACAG and clinical indicators was developed and validated for predicting mortality among AMI patients.

RESULTS

Both ACAG and AG exhibited a significant association with an elevated risk of in-hospital mortality in AMI patients. The C-index of ACAG (C-index = 0.606) was significantly higher than AG (C-index = 0.589). A nomogram (ACAG combined model) was developed to predict the in-hospital mortality for AMI patients. The nomogram demonstrated a good predictive performance by Area Under the Curve (AUC) of 0.763 in the training set, 0.744 and 0.681 in the external validation cohort. The C-index of the nomogram was 0.759 in the training set, 0.756 and 0.762 in the validation cohorts. Additionally, the C-index of the nomogram was obviously higher than the ACAG and age shock index in three databases.

CONCLUSION

ACAG was related to in-hospital mortality among AMI patients. The authors developed a nomogram incorporating ACAG and clinical indicators, demonstrating good performance for predicting in-hospital mortality of AMI patients.

A model for predicting postoperative persistent acute kidney injury (AKI) in AKI after cardiac surgery patients with normal baseline renal function

BACKGROUND

Persistent acute kidney injury (AKI) after cardiac surgery is not uncommon and linked to poor outcomes.

HYPOTHESIS

The purpose was to develop a model for predicting postoperative persistent AKI in patients with normal baseline renal function who experienced AKI after cardiac surgery.

METHODS

Data from 5368 patients with normal renal function at baseline who experienced AKI after cardiopulmonary bypass cardiac surgery in our hospital were retrospectively evaluated. Among them, 3768 patients were randomly assigned to develop the model, while the remaining patients were used to validate the model. The new model was developed using logistic regression with variables selected using least absolute shrinkage and selection operator regression.

RESULTS

The incidence of persistent AKI was 50.6% in the development group. Nine variables were selected for the model, including age, hypertension, diabetes, coronary heart disease, cardiopulmonary bypass time, AKI stage at initial diagnosis after cardiac surgery, postoperative serum magnesium level of <0.8 mmol/L, postoperative duration of mechanical ventilation, and postoperative intra-aortic balloon pump use. The model's performance was good in the validation group. The area under the receiver operating characteristic curve was 0.761 (95% confidence interval: 0.737-0.784). Observations and predictions from the model agreed well in the calibration plot. The model was also clinically useful based on decision curve analysis.

CONCLUSIONS

It is feasible by using the model to identify persistent AKI after cardiac surgery in patients with normal baseline renal function who experienced postoperative AKI, which may aid in patient stratification and individualized precision treatment strategy.

A prediction model for assessing hypoglycemia risk in critically ill patients with sepsis

BACKGROUND

Few studies have reported the risk factors or developed a risk predictive model of hypoglycemia patients with sepsis.

OBJECTIVE

To develop a predictive model to assess the hypoglycemia risk in critically ill patients with sepsis.

METHODS

For this retrospective study, we collected the data from the Medical Information Mart for Intensive Care III and IV (MIMIC-III and MIMIC-IV). All eligible patients from the MIMIC-III were randomly divided into the training set for development of predictive model and testing set for internal validation of the predictive model at a ratio of 8:2. Patients from the MIMIC-IV database were used as the external validation set. The primary endpoint was the occurrence of hypoglycemia. Univariate and multivariate logistic model was used to screen predictors. Adopted receiver operating characteristics (ROC) and calibration curves to estimate the performance of the nomogram.

RESULTS

The median follow-up time was 5.13 (2.61-9.79) days. Diabetes, dyslipidemia, mean arterial pressure, anion gap, hematocrit, albumin, sequential organ failure assessment, vasopressors, mechanical ventilation and insulin were identified as the predictors for hypoglycemia risk in critically ill patients with sepsis. We constructed a nomogram for predicting hypoglycemia risk in critically ill patients with sepsis based on these predictors. An online individualized predictive tool: https://ghongyang.shinyapps.io/DynNomapp/. The established nomogram had a good predictive ability by ROC and calibration curves in the training set, testing set and external validation cohort.

CONCLUSION

A predictive model of hypoglycemia risk was constructed, with a good ability in predicting the risk of hypoglycemia in critically ill patients with sepsis.

Independent risk factors of acute kidney injury among patients receiving extracorporeal membrane oxygenation

OBJECTIVE

Acute kidney injury (AKI) is one of the most frequent complications in patients treated with extracorporeal membrane oxygenation (ECMO) support. The aim of this study was to investigate the risk factors of AKI in patients undergoing ECMO support.

METHODS

We performed a retrospective cohort study which included 84 patients treated with ECMO support at intensive care unit in the People's Hospital of Guangxi Zhuang Autonomous Region from June 2019 to December 2020. AKI was defined as per the standard definition proposed by the Kidney Disease Improving Global Outcome (KDIGO). Independent risk factors for AKI were evaluated through multivariable logistic regression analysis with stepwise backward approach.

RESULTS

Among the 84 adult patients, 53.6% presented AKI within 48 h after initiation of ECMO support. Three independent risk factors of AKI were identified. The final logistic regression model included: left ventricular ejection fraction (LVEF) before ECMO initiation (OR, 0.80; 95% CI, 0.70-0.90), sequential organ failure assessment (SOFA) score before ECMO initiation (OR, 1.41; 95% CI, 1.16-1.71), and serum lactate at 24 h after ECMO initiation (OR, 1.27; 95% CI, 1.09-1.47). The area under receiver operating characteristics of the model was 0.879.

CONCLUSION

Severity of underlying disease, cardiac dysfunction before ECMO initiation and the blood lactate level at 24 h after ECMO initiation were independent risk factors of AKI in patients who received ECMO support.

Prediction of persistent acute kidney injury in postoperative intensive care unit patients using integrated machine learning: a retrospective cohort study

Acute kidney injury (AKI) often occurs in patients in the intensive care unit (ICU). AKI duration is closely related to the prognosis of critically ill patients. Identifying the disease course length in AKI is critical for developing effective individualised treatment. To predict persistent AKI at an early stage based on a machine learning algorithm and integrated models. Overall, 955 patients admitted to the ICU after surgery complicated by AKI were retrospectively evaluated. The occurrence of persistent AKI was predicted using three machine learning methods: a support vector machine (SVM), decision tree, and extreme gradient boosting and with an integrated model. External validation was also performed. The incidence of persistent AKI was 39.4-45.1%. In the internal validation, SVM exhibited the highest area under the receiver operating characteristic curve (AUC) value, followed by the integrated model. In the external validation, the AUC values of the SVM and integrated models were 0.69 and 0.68, respectively, and the model calibration chart revealed that all models had good performance. Critically ill patients with AKI after surgery had high incidence of persistent AKI. Our machine learning model could effectively predict the occurrence of persistent AKI at an early stage.