Study probes the link in patients with diabetes.
Glucose control influences glomerular filtration rate (GFR) and its estimations in patients with type 1 or 2 diabetes, a French study found.
Estimation of GFR using the Cockcroft-Gault (CG) formula or the Modification of Diet in Renal Disease (MDRD) study equation is recommended in patients with diabetes and CKD. Acute hyperglycemia is known to induce glomerular hyperfiltration in patients with and without diabetes, but the relationship between glycemic control as reflected by hemoglobin A1c levels and estimated GFR in patients with type 1 or type 2 diabetes is not well understood.
In this study, GFR was measured isotopically after injection of 51Cr-EDTA and predicted using
the CG and MDRD formulas in 193 patients (mean age 63 years) with type 1 or type 2 diabetes. The study group had a wide range of renal function and A1c levels.
The mean GFR as measured isotopically was 57.0 mL/min/1.73m2. The CG formula significantly overestimated GFR (61.7 mL/min/1.73m2), whereas the MDRD equation significantly underestimated it (51.3 mL/min/1.73m2). Additionally, A1c correlated positively with measured GFR and GFR estimated by CG or MDRD, and correlated negatively with serum creatinine level, the researchers reported in Diabetes Care 2006;29:1491-1495. The correlation between A1c and measured GFR was significant in patients with type 1 or 2 diabetes.
Each 1% increase in HbA1c was associated with 6 mL/min increase in measured GFR and a 5.6 and 5.3% mL/min/1.73m2 increase for GFR estimated by CG and MDRD.
GFR and its estimations were significantly higher in patients with poorly controlled diabetes (A1c greater than 8%) than in those with well controlled diabetes (A1c 8% or less). The mean measured GFR was 45.4 mL/min in patients with well controlled diabetes and 66.8 mL/min/1.73m2 in those with poorly controlled diabetes.
In patients with well controlled diabetes, the measured GFR was lower than the CG estimation. In patients with poorly controlled diabetes, the measured GFR was higher than the MDRD estimation.
In both patient groups, the correlation coefficients between GFR and its estimations were not influenced by glycemic control and were higher with the MDRD than with the CG. The maximal diagnostic accuracy was better for the MDRD than the CG, except for the diagnosis of moderate renal failure in patients with well controlled diabetes.
“The CG erroneously takes account of body weight to calculate GFR, and GFR does not directly change according to body weight in reality,” said Vincent Rigalleau, MD, of the University of Bordeaux. “This is the source of a demonstrated weight-related bias, and A1c may influence this bias.”
Why the CG equation is inaccurate at high A1c remains undetermined, but the researchers speculate that the complex relationship between body weight and A1c is probably a source of imprecision. Poor diabetes control may lead to glycosuria and loss of body weight in some patients, he explained, whereas other patients may experience weight gain and poor diabetes control in response to dietary nonadherence.
“The relations between GFR and its estimations were not affected by the degree of glucose control, and the precision and diagnostic accuracy of the CG formula were diminished for A1c >8%,” the authors wrote. “We think that the MDRD equation is preferable for the diagnosis of moderate and severe renal failure in diabetic patients, as it is more accurate than the CG and more robust when glucose control is not optimal.”