Abstract: FR-PO427

Hyperfiltration Predicts Rapid GFR Decline in a General Non-Diabetic Population and in Type 2 Diabetes

Session Information

Category: Chronic Kidney Disease (Non-Dialysis)

  • 301 CKD: Risk Factors for Incidence and Progression

Authors

  • Melsom, Toralf, UiT The Arctic University of Norway, Tromso, Norway
  • Looker, Helen C, National Institutes of Health, Phoenix, Arizona, United States
  • Kretzler, Matthias, University of Michigan, Ann Arbor, Michigan, United States
  • Nelson, Robert G., National Institutes of Health, Phoenix, Arizona, United States
  • Eriksen, Bjorn Odvar, UiT The Arctic University of Norway, Tromso, Norway
  • Nair, Viji, University of Michigan, Ann Arbor, Michigan, United States
  • Schei, Jørgen, UiT The Arctic University of Norway, Tromso, Norway
  • Mariani, Laura H., University of Michigan, Ann Arbor, Michigan, United States
  • Stefansson, Vidar T. N., UiT The Arctic University of Norway, Tromso, Norway
  • Harder, Jennifer L., University of Michigan, Ann Arbor, Michigan, United States
  • Jenssen, Trond G., UiT The Arctic University of Norway, Tromso, Norway
  • Solbu, Marit D., UiT The Arctic University of Norway, Tromso, Norway
  • Norvik, Jon V., UiT The Arctic University of Norway, Tromso, Norway
Background

An abnormally high glomerular filtration rate (GFR), or renal hyperfiltration, may predispose individuals to subsequent rapid GFR decline in diabetes, prediabetes, obesity, and hypertension. This hypothesis remains controversial; however, in diabetes, it is supported by clinical trials showing that treatments that acutely reduce the GFR, such as ACE inhibitors and sodium-glucose cotransport inhibitors, may reduce medium-term GFR decline.

Methods

We investigated whether a higher GFR predicts a steeper long-term GFR decline in two diverse populations, Pima Indians with type 2 diabetes (N=319) and non-diabetic Caucasians in Norway (the Renal Iohexol Clearance Survey [RENIS], N=1594). Because spurious correlations between initial values (e.g., GFR level) and subsequent changes may bias ordinary regression methods, we assessed this relationship as the correlation between the random intercept and random slope in a linear mixed model. This method separately estimates the error term (e.g., the day-to-day variation in the GFR) and random effects, eliminating bias because of regression to the mean.

Results

The mean (SD) baseline GFRs were 149.4 (43.3) and 104.0 (20.1) ml/min, and the median (IQR) follow-up times were 9.1 (4.0-15.0) and 5.6 (5.2–6.0) years in the Pima and RENIS cohorts, respectively. Higher body weight and higher fasting plasma glucose concentration were associated with a higher baseline GFR (a higher intercept) in both cohorts in multivariable adjusted linear mixed regression models with a random intercept and slope (p<0.001). The adjusted correlation between the random intercept and random slope was -0.42 (95% confidence interval, -0.55 to -0.26) in the Pima cohort and -0.32 (-0.40 to -0.23) in the RENIS cohort, demonstrating that higher baseline GFRs were associated with a steeper GFR decline.

Conclusion

Renal hyperfiltration predicts accelerated long-term GFR decline in type 2 diabetes and in the general non-diabetic population.

Funding

  • NIDDK Support