Online Program Home
My Program

Abstract Details

Activity Number: 187 - Contributed Poster Presentations: Section on Nonparametric Statistics
Type: Contributed
Date/Time: Monday, July 29, 2019 : 10:30 AM to 12:20 PM
Sponsor: Section on Nonparametric Statistics
Abstract #304828
Title: Modeling Kidney Function Decline via Functional Principal Components Analysis (FPCA)
Author(s): Brian Kwan* and Loki Natarajan and Jing Zhang and Tobias Fuhrer and Daniel Montemayor
Companies: University of California, San Diego and University of California, San Diego and Moores Cancer Center, University of California, San Diego and Institute of Molecular Systems Biology, ETH Zurich and University of Texas Health Science Center at San Antonio
Keywords: diabetic kidney disease; eGFR; functional principal components analysis; functional data analysis; non-linear; trajectory modeling

Patients with Type 2 diabetes (T2DM) are at high risk for diabetic kidney disease (DKD). Past studies have shown that annual estimated glomerular filtration rate (eGFR) change is a good measure of longitudinal DKD and eGFR curves are natural markers for DKD progression. Linear mixed-effects models are often used to estimate subject-specific eGFR trajectories; however, non-linear trends may exist. Curve estimation is further complicated by sparse and irregular spaced eGFR time series data. FPCA is a novel approach to exploring dominant modes of variation of eGFR curves while overcoming sparsity and irregular spaced data. Also, FPCA can project trajectories onto subject-specific scores, which can be used to delineate clinical patterns of DKD and cluster patients. Utilizing a cohort of 1003 T2DM patients from the Chronic Renal Insufficiency Cohort with up to 10 years of annual follow-up, we implemented FPCA to model eGFR trajectories. Our results uncovered dominant modes of variation and novel patterns of kidney function decline. This work has the potential to inform personalized DKD treatments and patient selection for clinical trials.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2019 program