Prediction of elderly acute kidney injury (AKI) in intensive care units (ICU) based on machine learning model

Acute Kidney Injury (AKI) is a major complication in patients admitted to Intensive Care Units (ICU), causing both clinical and economic burden on the healthcare system. This study develops a novel machine-learning (ML) model to predict, with several hours in advance, the AKI episodes of stage 2 and 3 (according to KDIGO definition) acquired in ICU. A total of 16'760 ICU adult patients from 145 different ICU centers and 3 different countries (US, Netherland, Italy) are retrospectively enrolled for the study. Every hour the model continuously analyzes the routinely-collected clinical data to generate a new probability of developing AKI stage 2 and 3, according to KDIGO definition, during the ICU stay. The predictive model obtains an auROC of 0.884 for AKI (stage 2/3 KDIGO) prediction, when evaluated on the internal test set composed by 1'749 ICU stays from US and EU centers. When externally tested on a multi-centric US dataset of 6'985 ICU stays and multi-centric Italian dataset of 1'025 ICU stays, the model achieves an auROC of 0.877 and of 0.911, respectively. In all datasets, the time between model prediction and AKI (stage 2/3 KDIGO) onset is at least of 14 hours after the first day of ICU hospitalization. In this study, a novel ML model for continuous and early AKI (stage 2/3 KDIGO) prediction is successfully developed, leveraging only routinely-available data. It continuously predicts AKI episodes during ICU stay, at least 14 hours in advance when the AKI episode happens after the first 24 hours of ICU admission. Its performances are validated in an extensive, multi-national and multi-centric cohort of ICU adult patients. This ML model overcomes the main limitations of currently available predictive models. The benefits of its real-world implementation enable an early proactive clinical management and the prevention of AKI episodes in ICU patients. Furthermore, the software could be directly integrated with IT system of the ICU.

Urinary IL-18 is an early predictive biomarker of acute kidney injury after cardiac surgery

BackgroundAcute kidney injury (AKI) in pediatric critical care patients is diagnosed using elevated serum creatinine, which occurs only after kidney impairment. There are no treatments other than supportive care for AKI once it has developed, so it is important to identify patients at risk to prevent injury. This study develops a machine learning model to learn pre-disease patterns of physiological measurements and predict pediatric AKI up to 48 h earlier than the currently established diagnostic guidelines.MethodsEHR data from 16,863 pediatric critical care patients between 1 month to 21 years of age from three independent institutions were used to develop a single machine learning model for early prediction of creatinine-based AKI using intelligently engineered predictors, such as creatinine rate of change, to automatically assess real-time AKI risk. The primary outcome is prediction of moderate to severe AKI (Stage 2/3), and secondary outcomes are prediction of any AKI (Stage 1/2/3) and requirement of renal replacement therapy (RRT). Predictions generate alerts allowing fast assessment and reduction of AKI risk, such as: “patient has 90% risk of developing AKI in the next 48 h” along with contextual information and suggested response such as “patient on aminoglycosides, suggest check level and review dose and indication”.ResultsThe model was successful in predicting Stage 2/3 AKI prior to detection by conventional criteria with a median lead-time of 30 h at AUROC of 0.89. The model predicted 70% of subsequent RRT episodes, 58% of Stage 2/3 episodes, and 41% of any AKI episodes. The ratio of false to true alerts of any AKI episodes was approximately one-to-one (PPV 47%). Among patients predicted, 79% received potentially nephrotoxic medication after being identified by the model but before development of AKI.ConclusionsAs the first multi-center validated AKI prediction model for all pediatric critical care patients, the machine learning model described in this study accurately predicts moderate to severe AKI up to 48 h in advance of AKI onset. The model may improve outcome of pediatric AKI by providing early alerting and actionable feedback, potentially preventing or reducing AKI by implementing early measures such as medication adjustment.

Acute kidney injury (AKI) is a common complication in critically ill patients and is associated with long-term complications and an increased mortality. This work presents preliminary findings from the first freely available European intensive care database released by Amsterdam UMC. A machine learning (ML) model was developed to predict AKI in the intensive care unit 12 hours before the actual event. Main features of the model included medications and hemodynamic parameters. Our models perform with an accuracy of 81.8% on moderate to severe AKI and 79.8% on all AKI patients. Those results can compete with models reported in the literature and introduce an ML model for AKI based on European patient data.

Machine learning for dynamic and early prediction of acute kidney injury after cardiac surgery

Background: Recent attempts were made to identify early indicators of acute kidney injury (AKI) in order to accelerate treatment and hopefully improve outcomes. This study aims to assess the value of urinary neutrophil gelatinase-associated lipocalin (uNGAL) as a predictor of AKI, severe AKI, and the need for renal replacement therapy (RRT). Methods: We conducted a prospective study and included adults admitted to our intensive care unit (ICU) at King Abdulaziz University Hospital (KAUH), between May 2012 and June 2013, who had at least 1 major risk factor for AKI. They were followed up throughout their hospital stay to identify which potential characteristics predicted any of the above 3 outcomes. We collected information on patients’ age and gender, the Acute Physiology And Chronic Health Evaluation, version II (APACHE II) score, the Sepsis-Related Organ Failure Assessment (SOFA) score, serum creatinine and cystatin C levels, and uNGAL. We compared ICU patients who presented with any of the 3 outcomes with others who did not. Results: We included 75 patients, and among those 21 developed AKI, 18 severe AKI, and 17 required RRT. Bivariate analysis revealed intergroup differences for almost all clinical variables (e.g., patients with AKI vs. patients without AKI); while multivariate analysis identified mean arterial pressure as the only predictor for AKI (p < 0.001) and the SOFA score (p = 0.04) as the only predictor for severe AKI. For RRT, day 1 maximum uNGAL was the stronger predictor (p < 0.001) when compared to admission diagnosis (p = 0.014). Day 1 and day 2 maximum uNGAL levels were good and excellent predictors for future RRT, but only fair to good predictors for AKI and severe AKI. Conclusions: Maximum urine levels of uNGAL measured over the first and second 24 h of an ICU admission were highly accurate predictors of the future need for RRT, however less accurate at detecting early and severe AKI.

Derivation and validation of the renal angina index to improve the prediction of acute kidney injury in critically ill children

Background: The Dublin Acute Biomarker Group Evaluation (DAMAGE) Study is a prospective 2-center observational study investigating the utility of urinary biomarker combinations for the diagnostic and prognostic assessment of acute kidney injury (AKI) in a heterogeneous adult intensive care unit (ICU) population. The objective of this study is to evaluate whether serial urinary biomarker measurements, in combination with a simple clinical model, could improve biomarker performance in the diagnostic prediction of severe AKI and clinical outcomes such as death and need for renal replacement therapy (RRT). Methods: Urine was collected daily from patients admitted to the ICU, for a total of 7 post-admission days. Urine biomarker concentrations (neutrophil gelatinase-associated lipocalin [NGAL], α-glutathione S-transferase [GST], π-GST, kidney injury molecule-1 [KIM-1], liver-type fatty acid-binding protein [L-FABP], Cystatin C, creatinine, and albumin) were measured. Urine biomarkers were combined with a clinical prediction of AKI model, to determine ability to predict AKI (any stage, within 2 days or 7 days of ICU admission), or a ­30-day composite clinical outcome (RRT – or death). Results: A total of 257 (38%) patients developed AKI within 7 days of ICU admission. Of those who developed AKI, 106 (41%) patients met stage 3 AKI within 7 days of ICU admission and 208 patients of the entire study cohort (31%) met the composite clinical endpoint of in-hospital mortality or RRT within 30 days of ICU admission. The addition of urinary NGAL/albumin to the clinical model modestly improved the prediction of AKI, in particular severe stage 3 AKI (area under the curve [AUC] of 0.9 from 0.87, p = 0.369) and the prediction of 30-day RRT or death (AUC 0.83 from 0.79, p = 0.139). Conclusion: A clinical model incorporating severity of illness, patient demographics, and chronic illness with currently available clinical biomarkers of renal function was strongly predictive of development of AKI and associated clinical outcomes in a heterogeneous adult ICU population. The addition of urinary NGAL/albumin to this simple clinical model improved the prediction of severe AKI, need for RRT and death, but not at a statistically or clinically significant level, when compared to the clinical model alone.

Acute kidney injury (AKI) frequently occurs in critically ill patients with coronary heart disease (CHD), and its development markedly elevates mortality rates and prolongs hospitalization duration. Early AKI prediction is crucial for timely intervention and amelioration of patient outcomes. This study aimed to develop and verify a clinical prediction model for the occurrence of AKI upon admission in the critically ill population with CHD through machine learning (ML). Data from the MIMIC-IV (Medical Information Mart for Intensive Care IV) version 2.2 database were gathered and included information about critically ill individuals with CHD in the intensive care unit (ICU). The dataset was randomized into a training set (70%) and a testing set (30%). Least absolute shrinkage and selection operator (LASSO) regression was used for feature variable selection. ML models, including logistic regression (LR), decision tree (DT), naive Bayes (NB), random forest (RF), extreme gradient boosting (XGBoost), and support vector machine (SVM), were constructed using 13 variables in the training set. The 6 models were compared in the testing set to identify the best-performing model. Subsequently, the model was assessed using calibration curve analysis and decision curve analysis (DCA). External validation was conducted using data from the Second Affiliated Hospital of Zhengzhou University. Ultimately, the predictive model was interpreted via Shapley Additive Explanation (SHAP) values. In total, 2711 patients with CHD admitted to the ICU were selected, with 1809 (66.7%) having AKI. XGBoost exhibited the best performance regarding discrimination (area under the receiver operating characteristic curve [AUROC]=0.765, 95% CI 0.731-0.800), accuracy (0.725), and sensitivity (0.759). External validation using a cohort of 226 patients confirmed the strong generalizability of the XGBoost model (AUROC=0.835, 95% CI 0.782-0.887). Feature importance analyses derived from SHAP values, DT, RF, and XGBoost consistently identified 5 key predictors associated with the development of AKI: mechanical ventilation, use of antiplatelet agents, age, N-terminal pro-B-type natriuretic peptide (NT-proBNP) levels, and acute physiology score III (APSIII). ML models can serve as reliable tools for forecasting AKI in the critically ill population with CHD. The XGBoost model is highly accurate and may aid doctors in identifying high-risk individuals for early intervention to lower mortality.

Acute kidney injury (AKI) is a common complication of hospitalization that greatly and negatively affects the short-term and long-term outcomes of patients. Current guidelines use serum creatinine level and urine output rate for defining AKI and as the staging criteria of AKI. However, because they are not sensitive or specific markers of AKI, clinicians find it difficult to predict the occurrence of AKI and prescribe timely treatment. Advances in computing technology have led to the recent use of machine learning and artificial intelligence in AKI prediction, recent research reported that by using electronic health records (EHR) the AKI prediction via machine-learning models can reach AUROC over 0.80, in some studies even reach 0.93. Our review begins with the background and history of the definition of AKI, and the evolution of AKI risk factors and prediction models is also appraised. Then, we summarize the current evidence regarding the application of e-alert systems and machine-learning models in AKI prediction.

Acute kidney injury (AKI) is common and linked to poor outcomes, but early detection remains challenging. Previous research identified urinary trace elements (TE) as early AKI biomarkers in intensive care unit (ICU) or cardiac surgery patients. We aimed to explore whether urinary TE enhance machine learning (ML) models for AKI prediction. We constructed ML models using the ICU cohort. We filtered the variables and optimized hyperparameters before predicting Kidney Disease: Improving Global Outcomes stage 2-3 AKI using eight ML classifiers: light gradient boosting machine (LightGBM), random forest (RF), ML logistic regression, support vector machine, multilayer perceptron, eXtreme gradient boosting (XGBoost), Gaussian Naive Bayes and k-nearest neighbors. External validation was performed in the cardiac surgery cohort. Among 149 ICU patients (median age 56.0 [interquartile range (IQR): 43.5-67.0], 63.1% male), 25 developed stage 2-3 AKI; among 144 cardiac surgery patients (median age 70.0 [IQR: 62.0-76.0], 72.9% male), 12 developed stage 2-3 AKI. Each ML in the internal validation had area under the curve (AUC) above .7, with XGBoost having the highest (.813); LightGBM had the second highest AUC (.799), highest G-mean (.567) and F1-score (.545). In external validation, RF had the highest AUC (.740), XGBoost had the highest G-mean (.289) and F1-score (.286). Age, strontium and boron were consistently ranked among the top five most important features in LightGBM, RF and XGBoost. ML models primarily based on urinary TE can identify AKI risk in both clinical groups (ICU and cardiac surgery), with LightGBM, RF and XGBoost serving as high-performance models for early prediction of stage 2-3 AKI.

Despite advancements in medical care, acute kidney injury (AKI) remains a major contributor to adverse patient outcomes and presents a significant challenge due to its associated morbidity, mortality, and financial cost. Machine learning (ML) is increasingly being recognized for its potential to transform AKI care by enabling early prediction, detection, and facilitating an individualized approach to patient management. This scoping review aims to provide a comprehensive analysis of externally validated ML models for the prediction, detection, and management of AKI. We systematically searched for relevant literature from inception to 15 February 2024, using four databases—MEDLINE, EMBASE, Web of Science, and Scopus. We focused solely on models that had undergone external validation, employed Kidney Disease Improving Global Outcomes (KDIGO) definitions for AKI, and utilized ML models (excluding logistic regression models). A total of 44 studies encompassing 161 ML models for AKI prediction, severity assessment, and outcomes in both adult and pediatric populations were included in the review. These studies encompassed 4,153,424 patient admissions, with 1,209,659 in the development and internal validation cohorts and 2,943,765 in the external validation cohorts. The ML models demonstrated significant variability in performance owing to differing clinical settings, populations, and predictors used. Most of the included models were developed in specialized patient populations, such as those in intensive care units, post-surgical settings, and specific disease states (e.g. congestive heart failure, traumatic brain injury, etc.). Moreover, only a few models incorporated dynamic predictors of AKI which are crucial for improving clinical utility in rapidly evolving clinical conditions like AKI. The variable performance of these models when applied to external validation cohorts highlights the challenges of reproducibility and generalizability in implementing ML models in AKI care. Despite acceptable performance metrics, none of the models assessed in this review underwent validation or implementation in real-world clinical workflows. These findings underscore the need for standardized performance metrics and validation protocols to enhance the generalizability and clinical applicability of these models. Future efforts should focus on enhancing model adaptability by incorporating dynamic predictors and unstructured data and by ensuring that models are developed in diverse patient populations. Moreover, collaboration between clinicians and data scientists is critical to ensure the development of models that are clinically relevant, fair, and tailored to real-world healthcare environments.

This systematic review evaluates the clinical utility of a novel biomarker kidney injury molecule 1 (Kim-1) in the prediction, diagnosis and prognosis of acute kidney injury (AKI). We searched literature in electronic databases from January 2002 to December 2009 by the key words "kidney injury molecule 1" or "Kim-1" and "acute kidney injury" or "acute renal failure". Studies were eligible for inclusion if they were primary studies published in English, in which Kim-1 was measured for the purpose of prediction, diagnosis or prognosis of AKI in patients. Eight articles met the selection criteria for inclusion in the study. Compared to non AKI patients, Kim-1 increased significantly (at least p<0.05) in AKI patients by 2 hours after cardiac surgery. In the prediction of AKI in patients within 24 hours of cardiac surgery, the sensitivity of Kim-1 ranged from 92% to 100% and AUC between 0.78 and 0.91. Kim-1 increased significantly (at least p<0.05) in AKI established patients, especially in patients with acute tubular necrosis (ATN). The AUC of Kim-1 in the diagnosis of AKI was from 0.9 to 0.95. However, Kim-1 showed weak association with the need of renal replacement therapy and death of AKI patient. Kim-1 is a potential novel urinary biomarker in the early detection of AKI within 24 hours after kidney insult. It might be especially beneficial in the diagnosis of ischemic ATN.

Perioperative Acute Kidney Injury

BackgroundThe development of acute kidney injury (AKI) during an intensive care unit (ICU) admission is associated with increased morbidity and mortality.MethodsOur objective was to develop and validate a data driven multivariable clinical predictive model for early detection of AKI among a large cohort of adult critical care patients. We utilized data form the Medical Information Mart for Intensive Care III (MIMIC-III) for all patients who had a creatinine measured for 3 days following ICU admission and excluded patients with pre-existing condition of Chronic Kidney Disease and Acute Kidney Injury on admission. Data extracted included patient age, gender, ethnicity, creatinine, other vital signs and lab values during the first day of ICU admission, whether the patient was mechanically ventilated during the first day of ICU admission, and the hourly rate of urine output during the first day of ICU admission.ResultsUtilizing the demographics, the clinical data and the laboratory test measurements from Day 1 of ICU admission, we accurately predicted max serum creatinine level during Day 2 and Day 3 with a root mean square error of 0.224 mg/dL. We demonstrated that using machine learning models (multivariate logistic regression, random forest and artificial neural networks) with demographics and physiologic features can predict AKI onset as defined by the current clinical guideline with a competitive AUC (mean AUC 0.783 by our all-feature, logistic-regression model), while previous models aimed at more specific patient cohorts.ConclusionsExperimental results suggest that our model has the potential to assist clinicians in identifying patients at greater risk of new onset of AKI in critical care setting. Prospective trials with independent model training and external validation cohorts are needed to further evaluate the clinical utility of this approach and potentially instituting interventions to decrease the likelihood of developing AKI.