Dataset Information

Combining renal cell arrest and damage biomarkers to predict progressive AKI in patient with sepsis.

ABSTRACT: BACKGROUND:Sepsis is the most common trigger for AKI and up to 40% of mild or moderate septic AKI would progress to more severe AKI, which is associated with significantly increased risk for death and later CKD/ESRD. Early identifying high risk patients for AKI progression is a major challenge in patients with septic AKI. METHODS:This is a prospective, multicenter cohort study which enrolled adult patients with sepsis and initially developed stage 1 or 2 AKI in the intensive care unit from January 2014 to March 2018. AKI was diagnosed and staged according to 2012 KDIGO-AKI guidelines. Renal cell arrest biomarkers (urinary TIMP2*IGFBP7, u[TIMP-2]*[IGFBP7]) and renal damage biomarkers (urinary KIM-1[uKIM-1] and urinary IL-18 [uIL-18]) were measured at time of AKI clinical diagnosis, and the performance of biomarkers for predicting septic AKI progression alone or in combination were evaluated. The primary outcome was AKI progression defined as worsening of AKI stage. The secondary outcome was AKI progression with subsequent death during hospitalization. RESULTS:Among 433 screened patients, 149 patients with sepsis and stage 1 or 2 AKI were included, in which 63 patients developed progressive AKI and 49 patients subsequently died during hospitalization. u[TIMP-2]*[IGFBP7], uKIM-1 and uIL-18 independently predicted the progression of septic AKI in which u[TIMP-2]*[IGFBP7] showed the greatest AUC (0.745; 95%CI, 0.667-0.823) as compared to uKIM-1 (AUC 0.719; 95%CI 0.638-0.800) and uIL-18 (AUC 0.619; 95%CI 0.525-0.731). Combination of u[TIMP-2]*[IGFBP7] with uKIM-1 improved the performance of predicting septic AKI progression with AUC of 0.752. u[TIMP-2]*[IGFBP7], alone or combined with uKIM-1/uIL-18, improved the risk reclassification over the clinical risk factor model alone both for the primary and secondary outcomes, as evidenced by significant category-free net reclassification index. CONCLUSIONS:Combination of renal cell arrest and damage biomarkers enhanced the prediction of AKI progression in patients with sepsis and improved risk reclassification over the clinical risk factors.

SUBMITTER: Tao X

PROVIDER: S-EPMC8672478 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:BackgroundDeep learning algorithms have achieved human-equivalent performance in image recognition. However, the majority of clinical data within electronic health records is inherently in a non-image format. Therefore, creating visual representations of clinical data could facilitate using cutting-edge deep learning models for predicting outcomes such as in-hospital mortality, while enabling clinician interpretability. The objective of this study was to develop a framework that first transforms longitudinal patient data into visual timelines and then utilizes deep learning to predict in-hospital mortality.Methods and findingsAll adult consecutive patient admissions from 2008-2016 at a tertiary care center were included in this retrospective study. Two-dimensional visual representations for each patient were created with clinical variables on one dimension and time on the other. Predictors included vital signs, laboratory results, medications, interventions, nurse examinations, and diagnostic tests collected over the first 48 hours of the hospital stay. These visual timelines were utilized by a convolutional neural network with a recurrent layer model to predict in-hospital mortality. Seventy percent of the cohort was used for model derivation and 30% for independent validation. Of 115,825 hospital admissions, 2,926 (2.5%) suffered in-hospital mortality. Our model predicted in-hospital mortality significantly better than the Modified Early Warning Score (area under the receiver operating characteristic curve [AUC]: 0.91 vs. 0.76, P < 0.001) and the Sequential Organ Failure Assessment score (AUC: 0.91 vs. 0.57, P < 0.001) in the independent validation set. Class-activation heatmaps were utilized to highlight areas of the picture that were most important for making the prediction, thereby providing clinicians with insight into each individual patient's prediction.ConclusionsWe converted longitudinal patient data into visual timelines and applied a deep neural network for predicting in-hospital mortality more accurately than current standard clinical models, while allowing for interpretation. Our framework holds promise for predicting several important outcomes in clinical medicine.

Project description:CONTEXT Slowly progressive chronic tubulo-interstitial damage jeopardizes long-term renal allograft survival. Both immune and non-immune mechanisms are thought to contribute, but the most promising targets for timely intervention have not been identified. OBJECTIVE In the current study we seek to determine the driving force behind progressive histological damage of renal allografts, without the interference of donor pathology, delayed graft function and acute graft rejection. DESIGN We used microarrays to examine whole genome expression profiles in renal allograft protocol biopsies, and analyzed the correlation between gene expression and the histological appearance over time. The gene expression profiles in these protocol biopsies were then compared with gene expression of biopsies with acute T-cell mediated rejection. PATIENTS Human renal allograft biopsies (N=120) were included: 96 rejection-free protocol biopsies and 24 biopsies with T-cell mediated acute rejection. RESULTS In this highly cross-validated study, we demonstrate the significant association of established, ongoing and future chronic histological damage with regulation of adaptive immune gene expression (T-cell and B-cell transcript sets) and innate immune response gene expression (dendritic cell, NK-cell, mast cell and granulocyte transcripts). We demonstrate the ability of gene expression analysis to perform as a quantitative marker for ongoing inflammation with a wide dynamic range: from subtle subhistological inflammation prior to development of chronic damage, over moderate subclinical inflammation associated with chronic histological damage, to marked inflammation of Banff-grade acute T-cell mediated rejection. CONCLUSION Progressive chronic histological damage after kidney transplantation is associated with significant regulation of both innate and adaptive immune responses, months before the histological lesions appear. This study therefore corroborates the hypothesis that quantitative inflammation below the diagnostic threshold of classic T-cell or antibody-mediated rejection is associated with early subclinical stages of progressive renal allograft damage. We used microarrays to examine whole genome expression profiles in renal allograft protocol biopsies, and analyzed the correlation between gene expression and the histological appearance over time. The gene expression profiles in these protocol biopsies were then compared with gene expression of biopsies with acute T-cell mediated rejection. Human renal allograft biopsies (N=120) were included: 96 rejection-free protocol biopsies and 24 biopsies with T-cell mediated acute rejection.

Project description:Sepsis is a life-threatening syndrome characterized by acute organ dysfunction due to infection. In particular, acute kidney injury (AKI) is common among patients with sepsis and is associated with increased mortality and morbidity. Oxidative stress is an important contributor to the pathogenesis of sepsis-related AKI. Plasma free thiols (R-SH) reflect systemic oxidative stress since they are readily oxidized by reactive species and thereby serve as antioxidants. Here, we aimed to assess the concentrations of serum free thiols in sepsis and associate these with major adverse kidney events (MAKE). Adult non-trauma patients who presented at the emergency department (ED) with a suspected infection were included. Free thiol levels and ischemia-modified albumin (IMA), a marker of oxidative stress, were measured in plasma at baseline, at the ward, and at three months, and one year after hospitalization. Plasma free thiol levels were lower at the ED visit and at the ward as compared to three months and one year after hospital admission (p < 0.01). On the contrary, plasma levels of IMA were higher at the ED and at the ward compared to three months and one year after hospital admission (p < 0.01). Furthermore, univariate logistic regression analyses showed that plasma free thiol levels at the ED were inversely associated with long-term renal function decline and survival at 90 days (MAKE90) and 365 days (MAKE365) (OR 0.43 per standard deviation [SD] [0.22-0.82, 95% CI], p = 0.011 and OR 0.58 per SD [0.34-0.96, 95% CI], p = 0.035, respectively). A multivariate regression analysis revealed an independent association of plasma free thiols at the ED (OR 0.52 per SD [0.29-0.93, 95% CI], p = 0.028) with MAKE365, even after adjustments for age, eGFR at the ED, SOFA score, and cardiovascular disease. These data indicate the clear role of oxidative stress in the pathogenesis of sepsis-AKI, as reflected in the lower plasma free thiol levels and increased levels of IMA.

Dataset Information

Combining renal cell arrest and damage biomarkers to predict progressive AKI in patient with sepsis.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets