Evaluating Kidney Function Decline in Children with Chronic Kidney Disease Using a Multi-Institutional Electronic Health Record Database

Background: identifying risk factors for decline of kidney function in children with chronic kidney disease

Chronic kidney disease (CKD) in children is a rare, understudied disease, which limits available evidence to aid in clinical decision making for these patients.

What’s known about CKD:

• May result from congenital anomalies of the kidney and urinary tract (CAKUT) or acquired disease
• Glomerular disease most common form of acquired CKD, leads to kidney failure
• Recurring acute kidney infections (AKI) linked to higher risk, progression of CKD
• Children with CKD may have other serious conditions, including
  • Cardiovascular disease
    • 10- to 100-fold increase of occurrence compared to general population
    • 25% of children with kidney failure and cardiovascular disease will die
• Metabolic bone disease
• Neurocognitive dysfunction

Previous prospective cohorts studies of CKD in children had small sample sizes, possible sample bias, and excluded other comorbid conditions like malignancy. Those studies are resource-intensive and may not fully represent the source population.

The purpose of this study was to use large-scale, real-world, longitudinal data from clinical settings to:

• Identify children with CKD
• Evaluate CKD progression
  • Including prior underrepresented populations with a history of malignancy
• Stratify clinical risk factors for kidney function decline
  • Glomerular disease etiology
• Hypertension
• Proteinuria

Methods: EHR data for children with CKD collected from PEDSnet

Data was collected from PEDSnet, a multi-institutional clinical research network of aggregated inpatient and outpatient EHR data cross six U.S. children’s health care organizations including Children’s Hospital Colorado. Data used in this study spanned from Jan. 1, 2009 to Feb. 28, 2022.

Bradley Dixon, MD, pediatric nephrologist in the Kidney Center at Children’s Colorado and associate professor of Pediatrics-Nephrology at the University of Colorado School of Medicine, contributed to the study. Dr. Dixon has been involved in several other large scale nephrology research projects leveraging PEDSnet data through Children’s Colorado’s participation in Glomerular Learning Network (GLEAN) collaborative.

Retrospective cohort study criteria:

• Confirmed CKD
  • 2 estimated glomerular filtration rate (eGFR) values
    • <90 ml/min per 1.73 m² (risk factor for CKD progression early in disease course)
    • ≥15 ml/min per 1.73 m²
  • Separated by ≥90 days without an intervening value ≥90 ml/min per 1.73 m²
  • All patients must have been seen by nephrologist in an outpatient setting on or after cohort entry to ensure clinically-significant CKD
• 18 months old to 18 years old
• No history of long-term dialysis or a kidney transplant

Subcohorts

Defined by CKD etiology:

• Glomerular
• Nonglomerular
• Malignancy associated

Outcomes

CKD progression was defined by composite outcome:

• eGFR <15 ml/min per 1.73 m², ≥50%
• Decline in eGFR
• Long-term dialysis, or
• Kidney transplant

Covariates

The association of hypertension and proteinuria was assessed with composite outcomes overall and across sub-cohorts.

• Hypertension (≥2 physician visits with hypertension diagnosis code within 2 years of cohort entrance)
• Proteinuria (≥1 least one urinalysis with ≥11 proteinuria within 2 years of cohort entrance)

Results: risk factors associated with CKD progression in children

Among the 7,148,875 children in the PEDSnet database, 11,240 children met all inclusion criteria for this study.

Characteristics of children with CKD at time of cohort entrance

Note: Continuous data is presented as median and interquartile range and categorical data as number or percent.

Nonglomerular subcohort

• 8,987 children with CAKUT
  • Hydronephrosis
  • Renal dysplasia
  • Renal agenesis
  • Obstructive and reflux nephropathy
• 10 years old
• 5 years follow up
• 76 eGFR value
• 31% had proteinuria
• 32% had hypertension

Glomerular subcohort

• 1,314 children
• 13 years old
• 9 years follow up
• 69 eGFR value
• 81% had proteinuria
• 60% had hypertension

Other findings

• Children in the malignancy subcohort were younger at cohort entrance, had longer follow-up time than other subcohorts
• No significant differences in sex across cohorts
• Majority had eGFR 60-89 ml/min per 1.73 at cohort entrance
  • Glomerular subcohort had lower median eGFR, greater proportion of 15 to 59 ml/min per 1.73 m² than other subcohorts
• ≥31% of children had proteinuria
• ≥32% of children had hypertension within 2 years of cohort entrance
• Glomerular CKD had highest incidence of proteinuria and hypertension
• 61% hypertension patients treated with an antihypertensive within first 2 years
• 59% patients with hypertension and proteinuria treated with reninangiotensin-aldosterone system blockade

Overall outcomes

• 17% of children progressed to the composite kidney failure outcome (eGFR ,15 ml/min per 1.73 m, ≥50% decline in eGFR, long-term dialysis, or kidney transplant)
  • 15% experienced ≥50% decline in eGFR
  • 10% reached eGFR ,15 ml/min per 1.73 m²
  • 4% required long-term dialysis
  • 4% required kidney transplant

Key study findings

Factors associated with CKD progression:

• Lower eGFR category
  • Adjusted hazard ratio (aHR) 1.44 [95% confidence interval (95% CI), 1.23 to 1.69]
  • aHR 2.38 [95% CI, 2.02 to 2.79]
  • 75 [95% CI,5.05 to 6.55] for eGFR45–59ml/min per 1.73m²
  • 30 to 44ml/min per 1.73m², 15–29 ml/min per 1.73m² at cohort entrance, respectively, when compared with eGFR60–89 ml/min per 1.73m²
• Glomerular disease
  • aHR 2.01 [95% CI, 1.78 to 2.28]
  • Most rapid progression to kidney failure among subcohorts
• Malignancy
  • aHR 1.79 [95% CI, 1.52 to 2.11]
• Proteinuria
  • aHR 2.23 [95% CI, 1.89 to 2.62]
  • More rapid decline in kidney function
• Hypertension
  • aHR 1.49 [95% CI, 1.22 to 1.82]
  • More rapid decline in kidney function
• Both proteinuria and hypertension
  • aHR 3.98 [95% CI, 3.40 to 4.6]
  • Fastest progression to kidney failure
• Number of complex chronic comorbidities at start of follow-up
  • 07 [95% CI,1.05 to1.10] per additional comorbid body system
• Male sex
  • aHR 1.16 [95% CI, 1.05 to 1.28]
• Younger age at cohort entrance
  • aHR 0.95 [95% CI, 0.94 to 0.96] per year older

Discussion and conclusion: previously established risk factors for CKD progression confirmed

Study authors found disease etiology, albuminuria, hypertension, age, male sex, lower eGFR, and greater medical complexity at start of follow-up were associated with more rapid decline in kidney function in children with CKD.

This study leveraged large-scale, multi-institutional EHR data collected in real-word settings to study pediatric CKD over an extended period. It included children with a history of cancer, a previously underrepresented population.

Findings confirm several previously established risk factors for CKD progression from prior observational studies and applied face validity to this novel approach to studying pediatric CKD.

• Replicated key findings from the comprehensive Chronic Kidney Disease in Children prospective cohort study
• Comparable to findings from adult cohort studies assessing kidney disease progression

The study could offer the foundation for future pragmatic clinical trials in children with CKD and may serve as a roadmap for using EHR data networks for effective study of rare diseases.