Defining Kidney Health Dimensions and Their Associations with Adverse Outcomes in Persons with Diabetes and CKD.

KEY POINTS

We identified three kidney health dimensions using 17 urine and plasma biomarkers across two cohorts of persons with diabetes and CKD. Worse scores for tubule injury, tubule function, and systemic inflammation/filtration were associated with a higher risk of CKD progression and death. A multibiomarker approach could help capture tubulointerstitial health in persons with diabetes and CKD.

BACKGROUND

Individual kidney tubule biomarkers are associated with risks of CKD progression and mortality in persons with diabetes. Integrating multiple kidney biomarkers using a latent variable method of exploratory factor analysis could define distinct dimensions of kidney health and their associations with adverse outcomes.

METHODS

We conducted a factor analysis of 17 candidate urine and plasma biomarkers in 1256 participants with diabetes and eGFR <60 ml/min per 1.73 m from the Chronic Renal Insufficiency Cohort (CRIC; =701) and the REasons for Geographic And Racial Differences in Stroke (REGARDS; =555) studies. We used Cox proportional hazards models to evaluate the associations of identified factors with CKD progression and mortality, adjusting for baseline clinical risk factors, eGFR, and albuminuria.

RESULTS

Three factor scores comprising ten biomarkers were identified: (plasma TNF receptor-1 and TNF receptor-2, plasma soluble urokinase plasminogen activator receptor, and plasma symmetric dimethylarginine), (urine EGF, urine asymmetric dimethylarginine, and urine symmetric dimethylarginine), and (urine -1 microglobulin, urine kidney injury molecule-1, and urine monocyte chemoattractant protein-1). In CRIC, there were 244 incident ESKD events, 102 with ≥40% eGFR decline from baseline, and 259 deaths; in REGARDS, there were 121 incident ESKD events and 462 deaths. In CRIC, lower (hazard ratio per 1-SD, 0.36; 95% confidence interval, 0.25 to 0.52) and higher (1.45; 1.18 to 1.78) scores were independently associated with higher CKD progression risk. Associations in REGARDS were weaker but directionally consistent ( score [0.81; 0.47 to 1.39] and score [1.12; 0.73 to 1.72]). Higher (1.47; 1.15 to 1.87) scores were associated with higher mortality risk in CRIC, but not REGARDS (1.15; 0.96 to 1.38). Higher factor scores were associated with higher mortality risk in both cohorts (CRIC: 1.35; 1.07 to 1.71; REGARDS: 1.41; 1.20 to 1.65).

CONCLUSIONS

Three distinct kidney health dimensions were identified, and each associated with CKD progression and/or all-cause mortality in persons with diabetes and CKD.