Barbara Davis Center for Childhood Diabetes, University of Colorado Health Sciences Center, P.O. Box 6511, Mail Stop A140, Aurora, CO 80045, USA. david.maahs@uchsc.edu
OBJECTIVE: Renal function is an important determinant of coronary atherosclerosis, and serum cystatin C is a novel accurate measure of glomerular filtration rate (GFR) and a predictor of cardiovascular events and mortality. We hypothesized that in individuals with type 1 diabetes, cystatin C would 1) predict progression of subclinical coronary atherosclerosis (SCA) and 2) be a stronger predictor of SCA than serum creatinine, GFR (estimated by the Cockcroft-Gault [GFRCG] and Modification of Diet in Renal Disease [GFRMDRD] formulas), and albumin excretion rate. RESEARCH DESIGN AND METHODS: Coronary artery calcification was measured twice, using Imatron C-150 Ultrafast CT, over a 2.5 +/- 0.4-year interval in 509 adults with type 1 diabetes (42% male, age 36 +/- 9 years, duration 23 +/- 9 years). SCA progression (n = 131) was defined as a >2.5 increase in square root calcium volume score or development of clinical coronary artery disease. Predictors of SCA progression were examined in a model selected by stepwise logistic regression and an a priori-determined model. Next, each measure of renal function was inserted into the stepwise model, one at a time, and Akaike information criterion was used to compare the fit of the competing models. RESULTS: The stepwise model included cystatin C (odds ratio 1.44, 95% CI 1.00-2.18, P = 0.048), age, baseline coronary artery calcification, sex, diabetes duration, systolic blood pressure, and HDL. The stepwise model had a better fit than any of the competing models with serum creatinine, GFRCG, GFRMDRD, or albumin excretion rate replacing cystatin C. CONCLUSIONS: In individuals with type 1 diabetes, cystatin C modestly predicts SCA.