RESUMO
OBJECTIVE: Whether early stages of kidney dysfunction assessed by the estimated glomerular filtration rate from cystatin C measurements (eGFRCysC) rather than from creatinine measurements (eGFRCr) would more precisely reflect the risk of developing type 2 diabetes (T2D) has not been clarified. We compared the risk of developing T2D associated with renal dysfunction indicated by eGFRCysC or eGFRCr measurements. METHODS: Studied were 2131 Japanese individuals without diabetes. Hazard ratios (HRs) for the development of T2D over 3-5 y were calculated across categories of eGFRCysC and eGFRCr, respectively. RESULTS: Reduced levels of eGFRCysC were associated with a step-wise increase in the cumulative incidence rate of T2D (p=0.007). In comparison with the eGFRCysC >85th percentile group (≥ 117.4 ml/min/1.73 m(2)), the lowest group, which was the eGFRCysC <15th percentile group (<86.2 ml/min/1.73 m(2)), had an adjusted HR of 2.30 (95% CI 1.13, 4.68) for T2D. Compared with the eGFRCr >85th percentile group, the lowest eGFRCr group (<15th percentile) had an HR of 1.19 (0.63, 2.24) for T2D. However, individuals with eGFRCr <60 ml/min/1.73 m(2) had a significantly increased risk of T2D. Clustering of both low eGFRCysC and low eGFRCr further elevated the HR for T2D compared with the presence of either. CONCLUSIONS: Although eGFRCr in ranges indicating chronic kidney disease reflected an elevated risk of developing diabetes, earlier stages of kidney dysfunction indicated by reduced eGFRCysC, which could not be captured by reduced eGFRCr, would be a marker for an elevated risk of developing T2D.