SAT-175 TRAJECTORIES OF KIDNEY FUNCTION AND THEIR PREDICTORS: FINDINGS FROM THE CHRONIC RENAL INSUFFICIENCY COHORT (CRIC) STUDY

Abstract

Change in kidney function is not uniform across patients with chronic kidney disease (CKD). Kidney function trajectories and their predictors are poorly characterized and likely differ by diabetes status. The Chronic Renal Insufficiency Cohort (CRIC) Study is an on-going, multi-center, prospective cohort study of men and women with moderate to advanced CKD implemented in seven centers across the United States. Data from 3,939 CRIC participants (51% with diabetes and 44% non-Hispanic black) over the first five years of follow-up was used to determine groupings based on estimated glomerular filtration rate (eGFR) trajectories from group-based trajectory models using Stata version 14.2. Predictors of kidney function trajectories were selected from among 30 established and novel sociodemographic, clinical and biochemical factors from study baseline. Final models were generated after step-wise selection using multinomial logistic regression models for those with and without diabetes, separately. Three cubic kidney function trajectory groupings best fit the longitudinal pattern of eGFR change among CRIC participants with and without diabetes (Figure). The stable trajectories centered around 72 mL/min/1.73m2 for those without diabetes, around 62 mL/min/1.73m2 for those with diabetes, and included 21.1% and 14.9% of those participants, respectively. The middle trajectories included 42.3% of those without diabetes and 42.6% with diabetes, and reflected, on average, less than a one mL/min/1.73m2 decline per year. The steepest trajectories included 36.6% and 42.5% of those without and with diabetes, respectively. Predictors of steeper eGFR trajectories included established risk factors such as baseline eGFR and albuminuria for both diabetics and non-diabetics. Novel risk factors included NTproBNP and PTH regardless of diabetes status, smoking among non-diabetics, and serum bicarbonate, CX3CL1 (an inflammatory chemokine) and NGAL among those with diabetes. Between 15-20% of those with and without diabetes appear to have stable kidney function over five years of follow-up, while the remainder of diabetics have slightly steeper kidney function trajectories, on average, compared to those without diabetes in the CRIC Study. Shared predictors of steeper eGFR trajectories across diabetes status included a cardiac marker and a marker of mineral metabolism. Among diabetics, inflammatory and kidney injury markers are also significantly associated with faster CKD progression.