Diabetes Mellitus as a Risk Factor for Progression from Acute Kidney Injury to Acute Kidney Disease: A Specific Prediction Model

Application of an interpretable machine learning method to predict the risk of death during hospitalization in patients with acute myocardial infarction combined with diabetes mellitus

BACKGROUND

Predicting the prognosis of patients with acute myocardial infarction (AMI) combined with diabetes mellitus (DM) is crucial due to high in-hospital mortality rates. This study aims to develop and validate a mortality risk prediction model for these patients by interpretable machine learning (ML) methods.

METHODS

Data were sourced from the Medical Information Mart for Intensive Care IV (MIMIC-IV, version 2.2). Predictors were selected by Least absolute shrinkage and selection operator (LASSO) regression and checked for multicollinearity with Spearman's correlation. Patients were randomly assigned to training and validation sets in an 8:2 ratio. Seven ML algorithms were used to construct models in the training set. Model performance was evaluated in the validation set using metrics such as area under the curve (AUC) with 95% confidence interval (CI), calibration curves, precision, recall, F1 score, accuracy, negative predictive value (NPV), and positive predictive value (PPV). The significance of differences in predictive performance among models was assessed utilising the permutation test, and 10-fold cross-validation further validated the model's performance. SHapley Additive exPlanations (SHAP) and Local Interpretable Model-agnostic Explanations (LIME) were applied to interpret the models.

RESULTS

The study included 2,828 patients with AMI combined with DM. Nineteen predictors were identified through LASSO regression and Spearman's correlation. The Random Forest (RF) model was demonstrated the best performance, with an AUC of 0.823 (95% CI: 0.774-0.872), high precision (0.867), accuracy (0.873), and PPV (0.867). The RF model showed significant differences (p < 0.05) compared to the K-Nearest Neighbours and Decision Tree models. Calibration curves indicated that the RF model's predicted risk aligned well with actual outcomes. 10-fold cross-validation confirmed the superior performance of RF model, with an average AUC of 0.828 (95% CI: 0.800-0.842). Significant Variables in RF model indicated that the top eight significant predictors were urine output, maximum anion gap, maximum urea nitrogen, age, minimum pH, maximum international normalised ratio (INR), mean respiratory rate, and mean systolic blood pressure.

CONCLUSION

This study demonstrates the potential of ML methods, particularly the RF model, in predicting in-hospital mortality risk for AMI patients with DM. The SHAP and LIME methods enhance the interpretability of ML models.

Representation of time-varying and time-invariant EMR data and its application in modeling outcome prediction for heart failure patients

OBJECTIVE

To represent a patient record with both time-invariant and time-varying features as a single vector using an end-to-end deep learning model, and further to predict the kidney failure (KF) status and mortality of heart failure (HF) patients.

MATERIALS AND METHODS

The time-invariant EMR data included demographic information and comorbidities, and the time-varying EMR data were lab tests. We used a Transformer encoder module to represent the time-invariant data, and refined a long short-term memory (LSTM) with a Transformer encoder attached to the top to represent the time-varying data, taking the original measured values and their corresponding embedding vectors, masking vectors, and two types of time intervals as inputs. The proposed representations of patients with time-invariant and time-varying data were used to predict KF status (949 out of 5268 HF patients diagnosed with KF) and mortality (463 in-hospital deaths) for HF patients. Comparative experiments were conducted between the proposed model and some representative machine learning models. Ablation experiments were also performed around the time-varying data representation, including replacing the refined LSTM with the standard LSTM, GRU-D and T-LSTM, respectively, and removing the Transformer encoder and the time-varying data representation module, respectively. The visualization of the attention weights of the time-invariant and time-varying features was used to clinically interpret the predictive performance. We used the area under the receiver operating characteristic curve (AUROC), the area under the precision-recall curve (AUPRC), and the F1-score to evaluate the predictive performance of the models.

RESULTS

The proposed model achieved superior performance, with average AUROCs, AUPRCs and F1-scores of 0.960, 0.610 and 0.759 for KF prediction and 0.937, 0.353 and 0.537 for mortality prediction, respectively. Predictive performance improved with the addition of time-varying data from longer time periods. The proposed model outperformed the comparison and ablation references in both prediction tasks.

CONCLUSIONS

Both time-invariant and time-varying EMR data of patients could be efficiently represented by the proposed unified deep learning model, which shows higher performance in clinical prediction tasks. The way to use time-varying data in the current study is hopeful to be used in other kinds of time-varying data and other clinical tasks.

Update on prognosis driven classification of pediatric AKI

Acute kidney injury (AKI) affects a large proportion of hospitalized children and increases morbidity and mortality in this population. Initially thought to be a self-limiting condition with uniformly good prognosis, we now know that AKI can persist and progress to acute kidney disease (AKD) and chronic kidney disease (CKD). AKI is presently categorized by stage of injury defined by increase in creatinine, decrease in eGFR, or decrease in urine output. These commonly used biomarkers of acute kidney injury do not change until the injury is well established and are unable to detect early stage of the disease when intervention is likely to reverse injury. The kidneys have the ability to compensate and return serum creatinine to a normal or baseline level despite nephron loss in the setting of AKI possibly masking persistent dysfunction. Though these definitions are important, classifying children by their propensity for progression to AKD and CKD and defining these risk strata by other factors besides creatinine may allow for better prognosis driven discussion, expectation setting, and care for our patients. In order to develop a classification strategy, we must first be able to recognize children who are at risk for AKD and CKD based on modifiable and non-modifiable factors as well as early biomarkers that identify their risk of persistent injury. Prevention of initial injury, prompt evaluation and treatment if injury occurs, and mitigating further injury during the recovery period may be important factors in decreasing risk of AKD and CKD after AKI. This review will cover presently used definitions of AKI, AKD, and CKD, recent findings in epidemiology and risk factors for AKI to AKD to CKD progression, novel biomarkers for early identification of AKI and AKI that may progress to CKD and future directions for improving outcome in children with AKI.

Predictors of Acute Kidney Disease Severity in Hospitalized Patients with Acute Kidney Injury

Acute kidney disease (AKD) forms part of the continuum of acute kidney injury (AKI) and worsens clinical outcomes. Currently, the predictors of AKD severity have yet to be established. We conducted a retrospective investigation involving 310 hospitalized patients with AKI and stratified them based on the AKD stages defined by the Acute Dialysis Quality Initiative criteria. Demographic, clinical, hematologic, and biochemical profiles, as well as 30-day outcomes, were compared between subgroups. In the analysis, the use of offending drugs (odds ratio, OR (95% confidence interval, CI), AKD stage 3 vs. non-AKD, 3.132 (1.304−7.526), p = 0.011, AKD stage 2 vs. non-AKD, 2.314 (1.049−5.107), p = 0.038), high AKI severity (OR (95% CI), AKD stage 3 vs. non-AKD, 6.214 (2.658−14.526), p < 0.001), and early dialysis requirement (OR (95% CI), AKD stage 3 vs. non-AKD, 3.366 (1.008−11.242), p = 0.049) were identified as independent predictors of AKD severity. Moreover, a higher AKD severity was associated with higher 30-day mortality and lower dialysis-independent survival rates. In conclusion, our study demonstrated that offending drug use, AKI severity, and early dialysis requirement were independent predictors of AKD severity, and high AKD severity had negative impact on post-AKI outcomes.

Advances in Biosensor Technologies for Acute Kidney Injury

Acute kidney injury (AKI) is one of the most prevalent and complex clinical syndromes with high morbidity and mortality. The traditional diagnosis parameters are insufficient regarding specificity and sensitivity, and therefore, novel biomarkers and their facile and rapid applications are being sought to improve the diagnostic procedures. The biosensors, which are employed on the basis of electrochemistry, plasmonics, molecular probes, and nanoparticles, are the prominent ways of developing point-of-care devices, along with the mutual integration of efficient surface chemistry strategies. In this manner, biosensing platforms hold pivotal significance in detecting and quantifying novel AKI biomarkers to improve diagnostic interventions, potentially accelerating clinical management to control the injury in a timely manner. In this review, novel diagnostic platforms and their manufacturing processes are presented comprehensively. Furthermore, strategies to boost their effectiveness are also indicated with several applications. To maximize these efforts, we also review various biosensing approaches with a number of biorecognition elements (e.g., antibodies, aptamers, and molecular imprinting molecules), as well as benchmark their features such as robustness, stability, and specificity of these platforms.

Application of Machine Learning to Predict Acute Kidney Disease in Patients With Sepsis Associated Acute Kidney Injury

Background: Sepsis-associated acute kidney injury (AKI) is frequent in patients admitted to intensive care units (ICU) and may contribute to adverse short-term and long-term outcomes. Acute kidney disease (AKD) reflects the adverse events developing after AKI. We aimed to develop and validate machine learning models to predict the occurrence of AKD in patients with sepsis-associated AKI. Methods: Using clinical data from patients with sepsis in the ICU at Beijing Friendship Hospital (BFH), we studied whether the following three machine learning models could predict the occurrence of AKD using demographic, laboratory, and other related variables: Recurrent Neural Network-Long Short-Term Memory (RNN-LSTM), decision trees, and logistic regression. In addition, we externally validated the results in the Medical Information Mart for Intensive Care III (MIMIC III) database. The outcome was the diagnosis of AKD when defined as AKI prolonged for 7-90 days according to Acute Disease Quality Initiative-16. Results: In this study, 209 patients from BFH were included, with 55.5% of them diagnosed as having AKD. Furthermore, 509 patients were included from the MIMIC III database, of which 46.4% were diagnosed as having AKD. Applying machine learning could successfully achieve very high accuracy (RNN-LSTM AUROC = 1; decision trees AUROC = 0.954; logistic regression AUROC = 0.728), with RNN-LSTM showing the best results. Further analyses revealed that the change of non-renal Sequential Organ Failure Assessment (SOFA) score between the 1st day and 3rd day (Δnon-renal SOFA) is instrumental in predicting the occurrence of AKD. Conclusion: Our results showed that machine learning, particularly RNN-LSTM, can accurately predict AKD occurrence. In addition, Δ SOFAnon-renal plays an important role in predicting the occurrence of AKD.