Acute Kidney Injury Sub-phenotypes Research Articles

Methods for phenotyping adult patients with acute kidney injury: a systematic review.

Acute kidney injury (AKI) is a multifaceted disease characterized by diverse clinical presentations and mechanisms. Advances in artificial intelligence have propelled the identification of AKI subphenotypes, enhancing our capacity to customize treatments and predict disease trajectories. We conducted a systematic review of the literature from 2017 to 2022, focusing on studies that utilized machine learning techniques to identify AKI subphenotypes in adult patients. Data were extracted regarding patient demographics, clustering methodologies, discriminators, and validation efforts from selected studies. The review highlights significant variability in subphenotype identification across different populations. All studies utilized clinical data such as comorbidities and laboratory variables to group patients. Two studies incorporated biomarkers of endothelial activation and inflammation into the clinical data to identify subphenotypes. The primary discriminators were comorbidities and laboratory trajectories. The association of AKI subphenotypes with mortality, renal recovery and treatment response was heterogeneous across studies. The use of diverse clustering techniques contributed to variability, complicating the application of findings across different patient populations. Identifying AKI subphenotypes enables clinicians to better understand and manage individual patient trajectories. Future research should focus on validating these phenotypes in larger, more diverse cohorts to enhance their clinical applicability and support personalized medicine in AKI management.

Advances in critical care nephrology through artificial intelligence.

This review explores the transformative advancement, potential application, and impact of artificial intelligence (AI), particularly machine learning (ML) and large language models (LLMs), on critical care nephrology. AI algorithms have demonstrated the ability to enhance early detection, improve risk prediction, personalize treatment strategies, and support clinical decision-making processes in acute kidney injury (AKI) management. ML models can predict AKI up to 24-48 h before changes in serum creatinine levels, and AI has the potential to identify AKI sub-phenotypes with distinct clinical characteristics and outcomes for targeted interventions. LLMs and generative AI offer opportunities for automated clinical note generation and provide valuable patient education materials, empowering patients to understand their condition and treatment options better. To fully capitalize on its potential in critical care nephrology, it is essential to confront the limitations and challenges of AI implementation, including issues of data quality, ethical considerations, and the necessity for rigorous validation. The integration of AI in critical care nephrology has the potential to revolutionize the management of AKI and continuous renal replacement therapy. While AI holds immense promise for improving patient outcomes, its successful implementation requires ongoing training, education, and collaboration among nephrologists, intensivists, and AI experts.

Incidence, Outcomes, and Predictors of Subphenotypes of Acute Kidney Injury among Acute Respiratory Distress Syndrome Patients: A Prospective Observational Study.

Acute kidney injury (AKI) is a heterogeneous syndrome with subphenotypes. Acute kidney injury is one of the most common complications in acute respiratory distress syndrome (ARDS) patients, which influences mortality. It was a single-center observational study on 266 ARDS patients on invasive mechanical ventilation (IMV) to determine the subphenotypes of AKI associated with ARDS. Subphenotyping was done based on the serum creatinine (SCr) trajectories from day 1 to day 5 of IMV into resolving (subphenotype 1) or non-resolving (subphenotype 2) AKI. Out of 266 ARDS patients, 222 patients were included for data analysis. 141 patients (63.51%) had AKI. The incidence of subphenotype 2 AKI among the ARDS cohort was 78/222 (35.13%). Subphenotype 2 AKI was significantly more among the non-survivors (87.7% vs 36.2 %, p < 0.001). Subphenotype 2 AKI was an independent predictor of mortality among ARDS patients (p < 0.001, adjusted odds ratio 8.978, 95% CI [2.790-28.89]. AKI subphenotype 1 had higher median day 1 SCr than subphenotype 2 but lower levels by day 3 and day 5 of IMV. The median time of survival was 8 days in AKI subphenotype 2 vs 45 days in AKI with subphenotype 1 [Log-Rank (Mantel-Cox) p < 0.001]. The novel DRONE score (Driving pressure, Oxygenation, and Nutritional Evaluation) ≥ 4 predicted subphenotype 2 AKI. The incidence of subphenotype 2 (non-resolving) AKI among ARDS patients on IMV was about 35% (vs 20% subphenotype 1 AKI), and it was an independent predictor of mortality. The DRONE score ≥4 can predict the AKI subphenotype 2. The serum creatinine trajectory-based subphenotype of AKI (resolving vs non-resolving) determines survival in ARDS patients. Non-resolving AKI subphenotype 2 is an independent predictor of mortality in ARDS. The novel DRONE score (driving pressure, oxygenation, and nutritional evaluation) ≥ 4 within 48 hours of IMV predicted the AKI subphenotype 2 among ventilated ARDS patients. Todur P, Nileshwar A, Chaudhuri S, Srinivas T. Incidence, Outcomes, and Predictors of Subphenotypes of Acute Kidney Injury among Acute Respiratory Distress Syndrome Patients: A Prospective Observational Study. Indian J Crit Care Med 2023;27(10):724-731.

Precision management of acute kidney injury in the intensive care unit: current state of the art.

Acute kidney injury (AKI) is a prototypical example of a common syndrome in critical illness defined by consensus. The consensus definition for AKI, traditionally defined using only serum creatinine and urine output, was needed to standardize the description for epidemiology and to harmonize eligibility for clinical trials. However, AKI is not a simple disease, but rather a complex and multi-factorial syndrome characterized by a wide spectrum of pathobiology. AKI is now recognized to be comprised of numerous sub-phenotypes that can be discriminated through shared features such as etiology, prognosis, or common pathobiological mechanisms of injury and damage. The characterization of sub-phenotypes can serve to enable prognostic enrichment (i.e., identify subsets of patients more likely to share an outcome of interest) and predictive enrichment (identify subsets of patients more likely to respond favorably to a given therapy). Existing and emerging biomarkers will aid in discriminating sub-phenotypes of AKI, facilitate expansion of diagnostic criteria, and be leveraged to realize personalized approaches to management, particularly for recognizing treatment-responsive mechanisms (i.e., endotypes) and targets for intervention (i.e., treatable traits). Specific biomarkers (e.g., serum renin; olfactomedin 4 (OLFM4); interleukin (IL)-9) may further enable identification of pathobiological mechanisms to serve as treatment targets. However, even non-specific biomarkers of kidney injury (e.g., neutrophil gelatinase-associated lipocalin, NGAL; [tissue inhibitor of metalloproteinases 2, TIMP2]·[insulin like growth factor binding protein 7, IGFBP7]; kidney injury molecule 1, KIM-1) can direct greater precision management for specific sub-phenotypes of AKI. This review will summarize these evolving concepts and recent innovations in precision medicine approaches to the syndrome of AKI in critical illness, along with providing examples of how they can be leveraged to guide patient care.

Acute kidney injury in patients receiving immune checkpoint inhibitors: a retrospective real-world study

BackgroundImmune checkpoint inhibitors (ICPi) can cause immune-related adverse events (irAEs) including acute kidney injury (AKI). We investigated the incidence of ICPi-associated AKI (ICPi-AKI) and AKI from other causes (non-ICPi-AKI) in cancer patients treated with ICPi. MethodsThis was a single-centre retrospective cohort study of patients receiving ICPi therapy between December 2011 and August 2020. AKI was defined and staged by the Kidney Disease Improving Global Outcomes creatinine criteria. The primary outcome was the incidence of AKI and ICPi-AKI. ResultsA total of 1037 patients were included in the final analysis. The median age was 63 years, 60% were male, and 22% had pre-existing chronic kidney disease. Overall, 189 patients (18.2%) developed AKI of whom 37 patients (3.6%) had ICPi-AKI. In patients with progressive cancer, AKI was not associated with increased mortality. In treatment responders, non-ICPi-AKI was associated with an increased risk of mortality (adjusted hazard ratio [HR] 2.03; 95% confidence interval [CI] 1.12–3.67), whereas ICPi-AKI was not linked to an increased risk of death (adjusted HR 0.60; 95% CI 0.18–1.96). Patients with ICPi-AKI were more likely to have higher AKI stages and less likely to have complete kidney recovery compared with non-ICPi-AKI (54% versus 79%, p = 0.01). ConclusionAKI was common in cancer patients treated with ICPi. Patients with ICPi-AKI had worse kidney outcomes compared to those with AKI from other causes. However, non-ICPi-AKI was associated with a higher risk of death. These findings emphasise the importance of identifying different sub-phenotypes of AKI.

Identifying acute kidney injury subphenotypes using an outcome-driven deep-learning approach

ObjectiveAcute kidney injury (AKI), a common condition on the intensive-care unit (ICU), is characterized by an abrupt decrease in kidney function within a few hours or days, leading to kidney failure or damage. Although AKI is associated with poor outcomes, current guidelines overlook the heterogeneity among patients with this condition. Identification of AKI subphenotypes could enable targeted interventions and a deeper understanding of the injury’s pathophysiology. While previous approaches based on unsupervised representation learning have been used to identify AKI subphenotypes, these methods cannot assess time series or disease severity. MethodsIn this study, we developed a data- and outcome-driven deep-learning (DL) approach to identify and analyze AKI subphenotypes with prognostic and therapeutic implications. Specifically, we developed a supervised long short-term memory (LSTM) autoencoder (AE) with the aim of extracting representation from time-series EHR data that were intricately correlated with mortality. Then, subphenotypes were identified via application of K-means. ResultsIn two publicly available datasets, three distinct clusters were identified, characterized by mortality rates of 11.3%, 17.3%, and 96.2% in one dataset and 4.6%, 12.1%, and 54.6% in the other. Further analysis demonstrated that AKI subphenotypes identified by our proposed approach were statistically significant on several clinical characteristics and outcomes. ConclusionIn this study, our proposed approach could successfully cluster the AKI population in ICU settings into 3 distinct subphenotypes. Thus, such approach could potentially improve outcomes of AKI patients in the ICU, with better risk assessment and potentially better personalized treatment.

Integrated Analysis of Blood and Urine Biomarkers to Identify Acute Kidney Injury Subphenotypes and Associations With Long-term Outcomes

Acute kidney injury (AKI) is a heterogeneous clinical syndrome with varying causes, pathophysiology, and outcomes. We incorporated plasma and urine biomarker measurements to identify AKI subgroups (subphenotypes) more tightly linked to underlying pathophysiology and long-term clinical outcomes. Multicenter cohort study. 769 hospitalized adults with AKI matched with 769 without AKI, enrolled from December 2009 to February 2015 in the ASSESS-AKI Study. 29 clinical, plasma, and urinary biomarker parameters used to identify AKI subphenotypes. Composite of major adverse kidney events (MAKE) with a median follow-up period of 4.7 years. Latent class analysis (LCA) and k-means clustering were applied to 29 clinical, plasma, and urinary biomarker parameters. Associations between AKI subphenotypes and MAKE were analyzed using Kaplan-Meier curves and Cox proportional hazard models. Among 769 AKI patients both LCA and k-means identified 2 distinct AKI subphenotypes (classes 1 and 2). The long-term risk for MAKE was higher with class 2 (adjusted HR, 1.41 [95% CI, 1.08-1.84]; P=0.01) compared with class 1, adjusting for demographics, hospital level factors, and KDIGO stage of AKI. The higher risk of MAKE among class 2 was explained by a higher risk of long-term chronic kidney disease progression and dialysis. The top variables that were different between classes 1 and 2 included plasma and urinary biomarkers of inflammation and epithelial cell injury; serum creatinine ranked 20th out of the 29 variables for differentiating classes. A replication cohort with simultaneously collected blood and urine sampling in hospitalized adults with AKI and long-term outcomes was unavailable. We identify 2 molecularly distinct AKI subphenotypes with differing risk of long-term outcomes, independent of the current criteria to risk stratify AKI. Future identification of AKI subphenotypes may facilitate linking therapies to underlying pathophysiology to prevent long-term sequalae after AKI. Acute kidney injury (AKI) occurs commonly in hospitalized patients and is associated with high morbidity and mortality. The AKI definition lumps many different types of AKI together, but subgroups of AKI may be more tightly linked to the underlying biology and clinical outcomes. We used 29 different clinical, blood, and urinary biomarkers and applied 2 different statistical algorithms to identify AKI subtypes and their association with long-term outcomes. Both clustering algorithms identified 2 AKI subtypes with different risk of chronic kidney disease, independent of the serum creatinine concentrations (the current gold standard to determine severity of AKI). Identification of AKI subtypes may facilitate linking therapies to underlying biology to prevent long-term consequences after AKI.

Definitions, phenotypes, and subphenotypes in acute kidney injury-Moving towards precision medicine.

The current definition of acute kidney injury (AKI) is generic and, based only on markers of function, is unsuitable for guiding individualized treatment. AKI is a complex syndrome with multiple presentations and causes. Targeted AKI management will only be possible if different phenotypes and subphenotypes of AKI are recognised, based on causation and related pathophysiology. Molecular signatures to identify subphenotypes are being recognised, as specific biomarkers reveal activated pathways. Assessment of individual clinical risk needs wider dissemination to allow identification of patients at high risk of AKI. New and more timely markers for glomerular filtration rate (GFR) are available. However, AKI diagnosis and classification should not be limited to GFR, but include tubular function and damage. Combining damage and stress biomarkers with functional markers enhances risk prediction, and identifies a population enriched for clinical trials targeting AKI. We review novel developments and aim to encourage implementation of these new techniques into clinical practice as a strategy for individualizing AKI treatment akin to a precision medicine-based approach.

A universal predictive and mechanistic urinary peptide signature in acute kidney injury

BackgroundThe delayed diagnosis of acute kidney injury (AKI) episodes and the lack of specificity of current single AKI biomarkers hamper its management. Urinary peptidome analysis may help to identify early molecular changes in AKI and grasp its complexity to identify potential targetable molecular pathways.MethodsIn derivation and validation cohorts totalizing 1170 major cardiac bypass surgery patients and in an external cohort of 1569 intensive care unit (ICU) patients, a peptide-based score predictive of AKI (7-day KDIGO classification) was developed, validated, and compared to the reference biomarker urinary NGAL and NephroCheck and clinical scores.ResultsA set of 204 urinary peptides derived from 48 proteins related to hemolysis, inflammation, immune cells trafficking, innate immunity, and cell growth and survival was identified and validated for the early discrimination (< 4 h) of patients according to their risk to develop AKI (OR 6.13 [3.96–9.59], p < 0.001) outperforming reference biomarkers (urinary NGAL and [IGFBP7].[TIMP2] product) and clinical scores. In an external cohort of 1569 ICU patients, performances of the signature were similar (OR 5.92 [4.73–7.45], p < 0.001), and it was also associated with the in-hospital mortality (OR 2.62 [2.05–3.38], p < 0.001).ConclusionsAn overarching AKI physiopathology-driven urinary peptide signature shows significant promise for identifying, at an early stage, patients who will progress to AKI and thus to develop tailored treatments for this frequent and life-threatening condition. Performance of the urine peptide signature is as high as or higher than that of single biomarkers but adds mechanistic information that may help to discriminate sub-phenotypes of AKI offering new therapeutic avenues.

Advances in laboratory detection of acute kidney injury

Recent advances have improved our understanding of the epidemiology and pathophysiology of acute kidney injury (AKI). So far, the Kidney Disease: Improving Global Outcome guidelines define and stratify kidney injury based on increases in serum creatinine level and/or decreases in urine output. Although the term AKI acknowledges the existence of cellular injury, its diagnosis is still only defined by the reduced excretory function of the kidney. New biomarkers that aid a better understanding of the relationship between acute tubular injury and kidney dysfunction have been identified, reflecting the advances in molecular biology. The expression of some of these novel biomarkers precedes changes in conventional biomarkers or can increase their predictive power. Therefore, they might enhance the clinical accuracy of the definition of AKI. This review summarizes the limitations of the current AKI classification and a panel of candidate biomarkers for augmenting AKI classification and recognition of AKI subphenotypes. We expect that the integration of appropriately selected biomarkers in routine clinical practice can improve AKI care.

Association of latent class analysis-derived subphenotypes of acute kidney injury with mortality in critically ill patients with cardiovascular disease: a retrospective cohort study

BackgroundTo explore the potential heterogeneity of acute kidney injury (AKI) and evaluate the prognostic differences among AKI subphenotypes in critically ill patients with cardiovascular diseases.MethodsData were extracted from the Medical Information Mart for Intensive Care (MIMIC)-III database. Latent class analysis (LCA) was used to explore the potential subphenotypes of AKI in critically ill patients with cardiovascular diseases. The number of classes was identified by the Bayesian information criterion and entropy. The differences in prognostic ability among the AKI subphenotypes were evaluated by logistic regression analysis.ResultA total of 7738 AKI patients were enrolled in this study. Using LCA, AKI patients were divided into 4 heterogeneous subphenotypes, which were obviously different from the Kidney Disease: Improving Global Outcomes (KDIGO) stages. Interestingly, class 3 classified by LCA was dominated by stage 2, while the mortality rate in class 3 was significantly different from that in class 1 (15.2% vs. 1.6%, p < 0.05). After further adjustment, the mortality rate in class 3 remained higher than that in class 1, with an odds ratio of 12.31 (95% confidence interval, 8.96–16.89).ConclusionsLCA was feasible for AKI classification in critically ill patients with cardiovascular disease, and 4 distinct subphenotypes of AKI patients with different prognoses were identified. Our results highlighted the potential heterogeneity of AKI patients, which is worthy of further investigation.

Abstract 9592: Acute Kidney Injury Subphenotypes in Critically Ill Patients with Cardiovascular Diseases by Latent Class Analysis: A Retrospective Cohort Study

Introduction: According to the potential heterogeneity of acute kidney injury (AKI) in critically ill patients with cardiovascular diseases, we explored the subphenotypes of AKI and evaluated the prognostic differences among AKI subphenotypes. Hypothesis: We hypothesized that patients with different AKI subphenotypes would suffer different prognosis. Methods: Clinical information from the Medical Information Mart for Intensive Care (MIMIC)-III database was enrolled. Latent class analysis (LCA) was used to explore the potential subphenotypes of AKI in critically ill patients with cardiovascular diseases. The number of classes was identified using the Bayesian information criterion and entropy. The differences in prognostic ability among the AKI subphenotypes were evaluated by logistic regression analysis. Results: In this study, a total of 7738 AKI patients were enrolled and divided into 4 heterogeneous subphenotypes by using LCA. The subphenotypes showed remarkable dofference from kidney disease classified by improving Global Outcomes (KDIGO) stages. Notedly, though both dominated by stage 2, the mortality between class 4 and class 1 classified by LCA were significantly different (15.2% vs. 1.6%, p &lt;0.05). The trend remained the same after further adjustment, with an odds ratio of 12.31 (95% confidence interval, 8.96-16.89). Conclusions: LCA was feasible for AKI classification and different subphenotypes prognosis prediction in critically ill patients with cardiovascular disease. Our study revealed the potential heterogeneity of AKI patients, which is worthy of further investigation.

Definitions of acute renal dysfunction: an evolving clinical and biomarker paradigm.

The current definition and classification of acute kidney injury (AKI) has limitations and shortcomings, which impact clinical management. The aim of this review is to highlight recent advances in our understanding of the pathophysiology and epidemiology of AKI, which impacts management and offers opportunities. Kidney damage varies according to the type of primary insult, secondary effects and mitigating responses and leads to distinct molecular, cellular and functional changes. Different sub-types of AKI with varying clinical phenotypes, recovery patterns and responses to therapeutic interventions have been identified. New tools to identify and characterize these AKI sub-types are available with the potential opportunity for individualized timely aetiology-based management of AKI. The identification of different sub-phenotypes of AKI based on genetic, molecular, cellular and functional pathophysiological changes following potential nephrotoxic exposures is possible with new technologies. This offers opportunities for personalized management of AKI and supports the call for a refinement of the existing AKI criteria.

Data Driven Analysis of Molecular Data Classifies AKI Patient and Predicts Clinical Outcomes

Background: Acute kidney injury (AKI) is a heterogeneous clinical syndrome with varying causes, pathophysiology and outcomes; but staging AKI is predominantly by serum creatinine. We sought to identify AKI sub-groups (sub-phenotypes) more tightly linked to underlying pathophysiology and long-term clinical outcomes. Methods: We independently applied latent class analysis (LCA) and k-Means clustering, to 29 clinical, plasma and urinary biomarker parameters measured during hospitalization to identify AKI sub-phenotypes in the ASSESS-AKI Study. AKI sub-phenotype associations were examined with the composite of major adverse kidney events (MAKE), defined as incident or progressive chronic kidney disease, long-term dialysis, or all-cause death during study follow-up. Findings: Among 769 AKI patients both LCA and k-Means clustering identified two AKI sub-phenotypes. Class 1 was characterized by a higher prevalence of prior congestive heart failure and favorable blood inflammatory and urinary tubular injury biomarkers, while class 2 was characterized by higher rates of prior chronic kidney disease and less favorable biomarkers. After a median follow-up of 4.7 years, the risk for MAKE was higher with class 2 (adjusted hazard ratio 1.41; 95% CI, 1.08 to 1.84; p=0.01) compared with class 1, adjusting for demographics, hospital level factors and KDIGO Stage of AKI. The higher risk of MAKE among class 2 was explained by a higher risk of chronic kidney disease progression and dialysis. The top five variables that were different between class 1 and 2 included plasma and urinary biomarkers of inflammation and epithelial cell injury, while serum creatinine ranked 20th out of the 29 variables for differentiating class 1 and 2. Interpretation: In this analysis, we identify two molecularly distinct AKI sub-phenotypes with differing risk of long-term outcomes, independent of current criteria to risk stratify AKI. Future identification of AKI sub-phenotypes may facilitate linking therapies to underlying pathophysiology to prevent long-term sequalae after AKI. Funding: National Institute of Diabetes, Digestive and Kidney Diseases Declaration of Interest: None to declare. Ethical Approval: The study was approved by the Yale University, Vanderbilt University, Kaiser Permanente, and University of Washington Institutional Review Boards. Written informed consent was obtained from participants.

Identification of acute kidney injury subphenotypes.

AKI is a complex clinical syndrome with many causes and there is a broad range of clinical presentations that vary according to duration, severity and context. Established consensus definitions of AKI are nonspecific and limited to kidney function. This reduces treatment options to generic approaches rather than individualized, cause-based strategies that have limited both understanding and management of AKI. The context and the temporal phase of kidney injury are critical features in the course of AKI and critical to timing-relevant intervention. These features are missing in generic definitions and terms used to describe AKI. Subphenotypes of AKI can be identified from novel damage biomarkers, from functional changes including creatinine trajectories, from the duration of change and from associated clinical characteristics and comorbidities. Subphenotype parameters can be combined in risk scores, or by association strategies ranging from a simple function-damage matrix to complex methods, such as machine learning. Examples of such strategies are reviewed along with tentative proposals for a revised nomenclature to facilitate description of AKI subphenotypes. Appropriate intervention requires refinement of the nomenclature of AKI to identify subphenotypes that facilitate correctly timed and selectively targeted intervention.

Genetic variation implicates plasma angiopoietin-2 in the development of acute kidney injury sub-phenotypes

BackgroundWe previously identified two acute kidney injury (AKI) sub-phenotypes (AKI-SP1 and AKI-SP2) with different risk of poor clinical outcomes and response to vasopressor therapy. Plasma biomarkers of endothelial dysfunction (tumor necrosis factor receptor-1, angiopoietin-1 and 2) differentiated the AKI sub-phenotypes. However, it is unknown whether these biomarkers are simply markers or causal mediators in the development of AKI sub-phenotypes.MethodsWe tested for associations between single-nucleotide polymorphisms within the Angiopoietin-1, Angiopoietin-2, and Tumor Necrosis Factor Receptor 1A genes and AKI- SP2 in 421 critically ill subjects of European ancestry. Top performing single-nucleotide polymorphisms (FDR < 0.05) were tested for cis-biomarker expression and whether genetic risk for AKI-SP2 is mediated through circulating biomarkers. We also completed in vitro studies using human kidney microvascular endothelial cells. Finally, we calculated the renal clearance of plasma biomarkers using 20 different timed urine collections.ResultsA genetic variant, rs2920656C > T, near ANGPT2 was associated with reduced risk of AKI-SP2 (odds ratio, 0.45; 95% CI, 0.31–0.66; adjusted FDR = 0.003) and decreased plasma angiopoietin-2 (p = 0.002). Causal inference analysis showed that for each minor allele (T) the risk of developing AKI-SP2 decreases by 16%. Plasma angiopoietin-2 mediated 41.5% of the rs2920656 related risk for AKI-SP2. Human kidney microvascular endothelial cells carrying the T allele of rs2920656 produced numerically lower levels of angiopoietin-2 although this was not statistically significant (p = 0.07). Finally, analyses demonstrated that angiopoietin-2 is minimally renally cleared in critically ill subjects.ConclusionGenetic mediation analysis provides supportive evidence that angiopoietin-2 plays a causal role in risk for AKI-SP2.

Identifying sub-phenotypes of acute kidney injury using structured and unstructured electronic health record data with memory networks

Acute Kidney Injury (AKI) is a common clinical syndrome characterized by the rapid loss of kidney excretory function, which aggravates the clinical severity of other diseases in a large number of hospitalized patients. Accurate early prediction of AKI can enable in-time interventions and treatments. However, AKI is highly heterogeneous, thus identification of AKI sub-phenotypes can lead to an improved understanding of the disease pathophysiology and development of more targeted clinical interventions. This study used a memory network-based deep learning approach to discover AKI sub-phenotypes using structured and unstructured electronic health record (EHR) data of patients before AKI diagnosis. We leveraged a real world critical care EHR corpus including 37,486 ICU stays. Our approach identified three distinct sub-phenotypes: sub-phenotype I is with an average age of 63.03±17.25 years, and is characterized by mild loss of kidney excretory function (Serum Creatinine (SCr) 1.55±0.34 mg/dL, estimated Glomerular Filtration Rate Test (eGFR) 107.65±54.98 mL/min/1.73 m2). These patients are more likely to develop stage I AKI. Sub-phenotype II is with average age 66.81±10.43 years, and was characterized by severe loss of kidney excretory function (SCr 1.96±0.49 mg/dL, eGFR 82.19±55.92 mL/min/1.73 m2). These patients are more likely to develop stage III AKI. Sub-phenotype III is with average age 65.07±11.32 years, and was characterized moderate loss of kidney excretory function and thus more likely to develop stage II AKI (SCr 1.69±0.32 mg/dL, eGFR 93.97±56.53 mL/min/1.73 m2). Both SCr and eGFR are significantly different across the three sub-phenotypes with statistical testing plus postdoc analysis, and the conclusion still holds after age adjustment.

Identification of Acute Kidney Injury Subphenotypes with Differing Molecular Signatures and Responses to Vasopressin Therapy.

Currently, no safe and effective pharmacologic interventions exist for acute kidney injury (AKI). One reason may be that heterogeneity exists within the AKI population, thereby hampering the identification of specific pathophysiologic pathways and therapeutic targets. The aim of this study was to identify and test whether AKI subphenotypes have prognostic and therapeutic implications. First, latent class analysis methodology was applied independently in two critically ill populations (discovery [n = 794] and replication [n = 425]) with AKI. Second, a parsimonious classification model was developed to identify AKI subphenotypes. Third, the classification model was applied to patients with AKI in VASST (Vasopressin and Septic Shock Trial; n = 271), and differences in treatment response were determined. In all three populations, AKI was defined using serum creatinine and urine output. A two-subphenotype latent class analysis model had the best fit in both the discovery (P = 0.004) and replication (P = 0.004) AKI groups. The risk of 7-day renal nonrecovery and 28-day mortality was greater with AKI subphenotype 2 (AKI-SP2) relative to AKI subphenotype 1 (AKI-SP1). The AKI subphenotypes discriminated risk for poor clinical outcomes better than the Kidney Disease: Improving Global Outcomes stages of AKI. A three-variable model that included markers of endothelial dysfunction and inflammation accurately determined subphenotype membership (C-statistic 0.92). In VASST, vasopressin compared with norepinephrine was associated with improved 90-day mortality in AKI-SP1 (27% vs. 46%, respectively; P = 0.02), but no significant difference was observed in AKI-SP2 (45% vs. 49%, respectively; P = 0.99) and the P value for interaction was 0.05. This analysis identified two molecularly distinct AKI subphenotypes with different clinical outcomes and responses to vasopressin therapy. Identification of AKI subphenotypes could improve risk prognostication and may be useful for predictive enrichment in clinical trials.

Circulating levels of soluble Fas (sCD95) are associated with risk for development of a nonresolving acute kidney injury subphenotype

BackgroundCritically ill patients with acute kidney injury (AKI) can be divided into two subphenotypes, resolving or nonresolving, on the basis of the trajectory of serum creatinine. It is unknown if the biology underlying these two AKI recovery patterns is different.MethodsWe measured eight circulating biomarkers in plasma obtained from a cohort of patients admitted to an intensive care unit (ICU) (n = 1241) with systemic inflammatory response syndrome. The biomarkers were representative of several biologic processes: apoptosis (soluble Fas), inflammation (soluble tumor necrosis factor receptor 1, interleukin 6, interleukin 8) and endothelial dysfunction, (angiopoietin 1, angiopoietin 2, and soluble vascular cell adhesion molecule 1). We tested for associations between biomarker levels and AKI subphenotypes using relative risk regression accounting for multiple hypotheses with the Bonferroni correction.ResultsDuring the first 3 days of ICU admission, 868 (70%) subjects developed AKI; 502 (40%) had a resolving subphenotype, and 366 (29%) had a nonresolving subphenotype. Hospital mortality was 12% in the resolving subphenotype and 21% in the nonresolving subphenotype. Soluble Fas was the only biomarker associated with a nonresolving subphenotype after adjustment for age, body mass index, diabetes, and Acute Physiology and Chronic Health Evaluation III score (p = 0.005).ConclusionsIdentifying modifiable targets in the Fas-mediated pathway may lead to strategies for prevention and treatment of a clinically important form of AKI.

Acute kidney injury subphenotypes based on creatinine trajectory identifies patients at increased risk of death.

BackgroundAcute kidney injury (AKI) is common among intensive care unit (ICU) patients. AKI is highly heterogeneous, with variable links to poor outcomes. Current approaches to classify AKI severity and identify patients at highest risk for poor outcomes focus on the maximum change in serum creatinine (SCr) values. However, these scores are hampered by the need for a reliable baseline SCr value and the absence of a component differentiating transient from persistent rises in SCr. We hypothesized that identification of resolving or nonresolving AKI subphenotypes based on the early trajectory of SCr values in the ICU would better differentiate patients at risk of hospital mortality.MethodsWe performed a secondary analysis of two prospective studies of ICU patients admitted to a trauma ICU (group 1; n = 1914) or general medical-surgical ICUs (group 2; n = 1867). In group 1, we tested definitions for resolving and nonresolving AKI subphenotypes and selected the definitions resulting in subphenotypes with the greatest separation in risk of death relative to non-AKI controls. We applied this definition to group 2 and tested whether the subphenotypes were independently associated with hospital mortality after adjustment for AKI severity.ResultsAKI occurred in 46% and 69% of patients in groups 1 and 2, respectively. In group 1, a resolving AKI subphenotype (defined as a decrease in SCr of 0.3 mg/dl or 25% from maximum in the first 72 h of study enrollment) was associated with a low risk of death. A nonresolving AKI subphenotype (defined as all AKI cases not meeting the “resolving” definition) was associated with a high risk of death. In group 2, the resolving AKI subphenotype was not associated with increased mortality (relative risk [RR] 0.86, 95% CI 0.63–1.17), whereas the nonresolving AKI subphenotype was associated with higher mortality (RR 1.68, 95% CI 1.15–2.44) even after adjustment for AKI severity stage.ConclusionsThe trajectory of SCr levels identifies AKI subphenotypes with different risks for death, even among AKI cases of similar severity. These AKI subphenotypes might better define the patients at risk for poor outcomes who might benefit from novel interventions.Electronic supplementary materialThe online version of this article (doi:10.1186/s13054-016-1546-4) contains supplementary material, which is available to authorized users.