Causes of Death in End-Stage Kidney Disease: Comparison Between the United States Renal Data System and a Large Integrated Health Care System Original paper
Researched by:
-
Dr. Umar
ID
Dr. Umar
Read MoreClinical Pharmacist and Clinical Pharmacy Master’s candidate focused on antibiotic stewardship, AI-driven pharmacy practice, and research that strengthens safe and effective medication use. Experience spans digital health research with Bloomsbury Health (London), pharmacovigilance in patient support programs, and behavioral approaches to mental health care. Published work includes studies on antibiotic use and awareness, AI applications in medicine, postpartum depression management, and patient safety reporting. Developer of an AI-based clinical decision support system designed to enhance antimicrobial stewardship and optimize therapeutic outcomes.
-
Chronic Kidney Disease (CKD)
Chronic Kidney Disease (CKD)
Dysbiosis in chronic kidney disease (CKD) reflects a shift toward reduced beneficial taxa and increased pathogenic, uremic toxin-producing species, driven by a bidirectional interaction in which the uremic environment disrupts microbial composition and dysbiotic metabolites accelerate renal deterioration.
-
End-Stage Renal Disease (ESRD)
End-Stage Renal Disease (ESRD)
End-stage renal disease is the irreversible loss of kidney function marked by uremic toxin accumulation, systemic complications, and the need for dialysis or transplantation. Its pathophysiology involves nephron loss, inflammation, metabolic disruption, and microbiome-derived toxins that accelerate cardiovascular and immune dysfunction.
-
Dr. Umar
Read More
Researched by:
-
Dr. Umar
ID
Dr. Umar
Read More
Microbiome Signatures identifies and validates condition-specific microbiome shifts and interventions to accelerate clinical translation. Our multidisciplinary team supports clinicians, researchers, and innovators in turning microbiome science into actionable medicine.
Dr. Umar
What was studied?
This study investigated the concordance of causes of death recorded for end-stage renal disease (ESRD) patients when comparing two major data sources: the United States Renal Data System (USRDS) national registry and the Kaiser Permanente Southern California (KPSC) integrated health system. Using mortality data from 2007–2016, researchers quantified how often the underlying cause of death matched across these systems, assessed agreement using Cohen’s weighted kappa statistics, and explored subcategory-specific concordance. Although the investigation did not address microbiome metrics or host–microbe interactions, the study’s findings indirectly inform microbiome-oriented clinical databases by revealing the limitations of mortality attribution data that are often used to correlate microbiome signatures with clinical outcomes.
Who was studied?
The cohort included 4,118 adults with ESRD whose deaths were recorded in both USRDS and KPSC databases. The mean age was 71 years, 41.2% were women, and the population was racially diverse: White (38.2%), Black (21%), Hispanic (28.8%), and Asian (9.1%). Most patients (90.1%) received hemodialysis, with 9.7% on peritoneal dialysis and fewer than 1% post-transplant. Deaths occurred across a decade and reflected the broad demographic composition of Southern California. No microbial sequencing, stool sampling, or infection-specific microbiome characterization was performed, and therefore, microbial signatures cannot be inferred from the dataset.
Most important findings
The study found only slight agreement (overall 36.4%, kappa = 0.20) between the underlying causes of death recorded by USRDS and KPSC. The most common KPSC causes were circulatory (35.7%), endocrine/metabolic (24.2%), and genitourinary (12.9%), while USRDS most frequently reported cardiac disease (46.9%), withdrawal from dialysis (12.6%), and infection (10.1%). Importantly for microbiome-related interpretations, infection-related deaths—a category often relevant for microbial signature studies—showed weak concordance (kappa = 0.20) and low positive agreement (26%), meaning infection-attributed deaths may be inconsistently classified across systems. This variability limits the reliability of linking microbiome patterns to infection-related mortality outcomes when using registry data alone. Variability in categorization, absence of ICD-10 categories in USRDS, and inconsistent coding practices contributed to discordance.
A condensed table summarizing key cross-source patterns:
| Category | KPSC most common (%) / USRDS most common (%) |
|---|---|
| Circulatory/Cardiac | 35.7 / 46.9 |
| Endocrine-metabolic | 24.2 / 0.4 |
| Genitourinary | 12.9 / not listed |
| Infection | 3.0 / 10.1 |
Key implications
The study underscores substantial limitations in using registry-reported causes of death to interpret clinical outcomes, especially for mechanistic studies that require precise attribution of mortality categories, such as microbiome–mortality correlation research. Inconsistent categorization, particularly for infections, metabolic causes, and chronic disease contributions, means that downstream analyses linking microbial biomarkers to death mechanisms may be confounded by misclassification bias. Improving coding harmonization, integrating standardized ICD-10 categories into registry systems, and ensuring consistent adjudication across care settings would enhance the interpretability of mortality data and improve the accuracy of microbiome-clinical associations in translational research.
Citation
Bhandari SK, Zhou H, Shaw SF, Shi J, Tilluckdharry NS, Rhee CM, Jacobsen SJ, Sim JJ. Causes of Death in End-Stage Kidney Disease: Comparison Between the United States Renal Data System and a Large Integrated Health Care System. American Journal of Nephrology. 2022;53(1):32-40. doi:10.1159/000520466