ABSTRACT
Random survival forests (RSF) can be applied to many time-to-event research questions and are particularly useful in situations where the relationship between the independent variables and the event of interest is rather complex. However, in many clinical settings, the occurrence of the event of interest is affected by competing events, which means that a patient can experience an outcome other than the event of interest. Neglecting the competing event (i.e., regarding competing events as censoring) will typically result in biased estimates of the cumulative incidence function (CIF). A popular approach for competing events is Fine and Gray's subdistribution hazard model, which directly estimates the CIF by fitting a single-event model defined on a subdistribution timescale. Here, we integrate concepts from the subdistribution hazard modeling approach into the RSF. We develop several imputation strategies that use weights as in a discrete-time subdistribution hazard model to impute censoring times in cases where a competing event is observed. Our simulations show that the CIF is well estimated if the imputation already takes place outside the forest on the overall dataset. Especially in settings with a low rate of the event of interest or a high censoring rate, competing events must not be neglected, that is, treated as censoring. When applied to a real-world epidemiological dataset on chronic kidney disease, the imputation approach resulted in highly plausible predictor-response relationships and CIF estimates of renal events.
Subject(s)
Biometry , Humans , Biometry/methods , Survival Analysis , Models, Statistical , Proportional Hazards ModelsABSTRACT
RATIONALE & OBJECTIVE: Biomarkers that enable better identification of persons with chronic kidney disease (CKD) who are at higher risk for disease progression and adverse events are needed. This study sought to identify urine and plasma metabolites associated with progression of kidney disease. STUDY DESIGN: Prospective metabolome-wide association study. SETTING & PARTICIPANTS: Persons with CKD enrolled in the GCKD (German CKD) study with metabolite measurements, with external validation within the ARIC (Atherosclerosis Risk in Communities) Study. EXPOSURES: 1,513 urine and 1,416 plasma metabolites (Metabolon Inc) measured at study entry using untargeted mass spectrometry. OUTCOMES: Main end points were kidney failure (KF) and a composite kidney end point (CKE) of KF, estimated glomerular filtration rate<15mL/min/1.73m2, or a 40% decrease in estimated glomerular filtration rate. Death from any cause was a secondary end point. After a median of 6.5 years of follow-up, 500 persons had experienced KF, 1,083 had experienced the CKE, and 680 had died. ANALYTICAL APPROACH: Time-to-event analyses using multivariable proportional hazard regression models in a discovery-replication design with external validation. RESULTS: 5,088 GCKD study participants were included in analyses of urine metabolites, and 5,144 were included in analyses of plasma metabolites. Among 182 unique metabolites, 30 were significantly associated with KF, 49 with the CKE, and 163 with death. The strongest association with KF was observed for plasma hydroxyasparagine (HR, 1.95; 95% CI, 1.68-2.25). An unnamed metabolite measured in plasma and urine was significantly associated with KF, the CKE, and death. External validation of the identified associations of metabolites with KF or the CKE revealed directional consistency for 88% of observed associations. Selected associations of 18 metabolites with study outcomes have not been previously reported. LIMITATIONS: Use of observational data and semiquantitative metabolite measurements at a single time point. CONCLUSIONS: The observed associations between metabolites and KF, the CKE, or death in persons with CKD confirmed previously reported findings and also revealed several associations not previously described. These findings warrant confirmatory research in other study cohorts. PLAIN-LANGUAGE SUMMARY: Incomplete understanding of the variability of chronic kidney disease (CKD) progression motivated the search for new biomarkers that would help identify people at increased risk. We explored metabolites in plasma and urine for their association with unfavorable kidney outcomes or death in persons with CKD. Metabolomic analyses revealed 182 metabolites significantly associated with CKD progression or death. Many of these associations confirmed previously reported findings or were validated by analysis in an external study population. Our comprehensive screen of the metabolome serves as a valuable foundation for future investigations into biomarkers associated with CKD progression.
ABSTRACT
BACKGROUND: Population-wide research on potential new imaging biomarkers of the kidney depends on accurate automated segmentation of the kidney and its compartments (cortex, medulla, and sinus). METHODS: We developed a robust deep-learning framework for kidney (sub-)segmentation based on a hierarchical, three-dimensional convolutional neural network (CNN) that was optimized for multiscale problems of combined localization and segmentation. We applied the CNN to abdominal magnetic resonance images from the population-based German National Cohort (NAKO) study. RESULTS: There was good to excellent agreement between the model predictions and manual segmentations. The median values for the body-surface normalized total kidney, cortex, medulla, and sinus volumes of 9934 persons were 158, 115, 43, and 24 mL/m2. Distributions of these markers are provided both for the overall study population and for a subgroup of persons without kidney disease or any associated conditions. Multivariable adjusted regression analyses revealed that diabetes, male sex, and a higher estimated glomerular filtration rate (eGFR) are important predictors of higher total and cortical volumes. Each increase of eGFR by one unit (i.e., 1 mL/min per 1.73 m2 body surface area) was associated with a 0.98 mL/m2 increase in total kidney volume, and this association was significant. Volumes were lower in persons with eGFR-defined chronic kidney disease. CONCLUSION: The extraction of image-based biomarkers through CNN-based renal sub-segmentation using data from a population-based study yields reliable results, forming a solid foundation for future investigations.
Subject(s)
Kidney , Magnetic Resonance Imaging , Humans , Male , Female , Magnetic Resonance Imaging/methods , Magnetic Resonance Imaging/statistics & numerical data , Kidney/diagnostic imaging , Middle Aged , Aged , Adult , Germany , Glomerular Filtration Rate/physiology , Biomarkers/analysis , Neural Networks, Computer , Deep Learning , Cohort StudiesABSTRACT
Diabetic kidney disease (DKD) is the main cause of chronic kidney disease (CKD) and progresses faster in males than in females. We identify sex-based differences in kidney metabolism and in the blood metabolome of male and female individuals with diabetes. Primary human proximal tubular epithelial cells (PTECs) from healthy males displayed increased mitochondrial respiration, oxidative stress, apoptosis, and greater injury when exposed to high glucose compared with PTECs from healthy females. Male human PTECs showed increased glucose and glutamine fluxes to the TCA cycle, whereas female human PTECs showed increased pyruvate content. The male human PTEC phenotype was enhanced by dihydrotestosterone and mediated by the transcription factor HNF4A and histone demethylase KDM6A. In mice where sex chromosomes either matched or did not match gonadal sex, male gonadal sex contributed to the kidney metabolism differences between males and females. A blood metabolomics analysis in a cohort of adolescents with or without diabetes showed increased TCA cycle metabolites in males. In a second cohort of adults with diabetes, females without DKD had higher serum pyruvate concentrations than did males with or without DKD. Serum pyruvate concentrations positively correlated with the estimated glomerular filtration rate, a measure of kidney function, and negatively correlated with all-cause mortality in this cohort. In a third cohort of adults with CKD, male sex and diabetes were associated with increased plasma TCA cycle metabolites, which correlated with all-cause mortality. These findings suggest that differences in male and female kidney metabolism may contribute to sex-dependent outcomes in DKD.