Abstract: TH-PO1048
Can Estimated Glomerular Filtration Rate (eGFR) Slope Be a Good Outcome Measure for Glomerular Disease?
Session Information
- Glomerular Diseases: Epidemiology, Mechanisms, Complications, Outcomes
November 07, 2019 | Location: Exhibit Hall, Walter E. Washington Convention Center
Abstract Time: 10:00 AM - 12:00 PM
Category: Glomerular Diseases
- 1203 Glomerular Diseases: Clinical, Outcomes, and Trials
Authors
- Gillespie, Brenda W., University of Michigan, Ann Arbor, Michigan, United States
- Smith, Abigail R., Arbor Research Collaborative for Health, Ann Arbor, Michigan, United States
- Zee, Jarcy, Arbor Research Collaborative for Health, Ann Arbor, Michigan, United States
- Gipson, Debbie S., University of Michigan Mott Children's Hospital, Ann Arbor, Michigan, United States
- Mansfield, Sarah, Arbor Research Collaborative for Health, Ann Arbor, Michigan, United States
- Mariani, Laura H., University of Michigan, Ann Arbor, Michigan, United States
- Helmuth, Margaret, Arbor Research Collaborative for Health, Ann Arbor, Michigan, United States
- Robinson, Bruce M., Arbor Research Collaborative for Health, Ann Arbor, Michigan, United States
Background
Unlike chronic kidney disease (CKD), where trajectories of eGFR over time are commonly linear, trajectories in glomerulonephropathy (GN) are often variable, with excursions into high eGFR (hyperfiltration), acute kidney injury (AKI) and sudden drops to kidney failure. This makes the use of eGFR slope, a statistically powerful progression outcome, challenging in this population. Our objective was to assess the prevalence of nonlinear features and propose adjustments to facilitate use of eGFR slope as an outcome measure.
Methods
Using data from CureGN, patients with at least one year of follow-up, at least 5 eGFR measurements, and a diagnosis of Minimal Change Disease (MCD) were included. Analysis was restricted to MCD due to high prevalence of high eGFR (>120 mL/min/1.73 m2). Linear regression models were fitted to each participant’s data with and without winsorizing (capping) eGFR values at 120 mL/min/1.73m2 with the rationale that values above 120 represented high filtration rates but not clinically improved eGFR. Trajectories were deemed linear if root mean squared error (RMSE) was ≤20; we examined individual nonlinear trajectories to assess deviation from linearity.
Results
74 adult (mean±SD age = 43.2±19.1, range 18-82) and 116 pediatric (mean±SD age = 7.7±3.6, range 2-16) participants were included. At least one eGFR>120 was present in 22% of adults and 83% of children. Linear fits were improved after winsorizing: 76% (95% adult, 64% pediatric) had good fits prior to, and 96% (96% adult, 96% pediatric) had good fits after winsorizing. Based on visual assessment of trajectories, most cases with nonlinear fits had apparent AKI, and a few had sudden kidney failure. In AKI, allowing the fit to include outlying point(s) provided a reasonable overall slope estimate. In kidney failure, the linear fits did not follow each curve but often captured the overall trend very well (Figure).
Conclusion
Linear models of eGFR have previously been used in analyses of CKD data; the added features of winsorizing and data examination will enhance their use for GN patients with MCD.
Funding
- NIDDK Support