Glomerular filtration rate and albuminuria: twin manifestations of nephropathy in diabetes.

I n the past three decades, urinary albumin excretion has assumed a central role in the diagnosis and management of kidney disease among people with diabetes, both type 1 and type 2. Microalbuminuria was initially found to predict subsequent overt albuminuria (dipstick positive, or 300 mg/24 h), which in turn predicted loss of GFR (1–3). From the strength of these relationships, it has frequently been assumed that microalbuminuria and overt albuminuria are requisite first and second steps along a single pathway that leads to loss of GFR and ESRD. In recent work, Perkins et al. (4) challenged these assumptions by demonstrating that microalbuminuria can regress to normoalbuminuria among people with type 1 diabetes. In this issue of the Journal of the American Society of Nephrology (JASN), these authors build on their previous work by examining change in GFR (estimated as 100/serum concentration of cystatin C) among patients from the same cohort (5); 578 patients with urine albumin excretion rates that were consistently 300 g/min were followed for 8 to 12 yr. Loss of GFR was defined as an average change in estimated GFR (eGFR; “slope”) that exceeded 3.3%/yr, a threshold that represents the 2.5th percentile of age-standardized eGFR decline among participants without diabetes in the Baltimore Aging Study. Persistent microalbuminuria was a strong risk factor for subsequent loss of GFR, reemphasizing earlier work that established the importance of sustained increases in urine albumin excretion in the pathogenesis and diagnosis of diabetic kidney disease. However, patients who lost GFR at a high rate did not have overt albuminuria, by study design, and some had “normal” urinary excretion of albumin. This study contributes to a growing literature that suggests that overt albuminuria does not always precede a significant loss of GFR in the setting of diabetes and that measuring albuminuria alone does not fully capture the scope of early diabetic kidney disease (6–8). Instead, albuminuria and GFR loss may represent complementary, if overlapping, manifestations of kidney damage. This supposition raises important questions. Are the underlying pathophysiologic processes and clinical risk factors of albuminuria and loss of GFR different in the early stages of diabetic kidney disease? How do albuminuria and loss of GFR each affect prognosis in terms of future decline in kidney function and the concurrent increasing risk for other complications such as cardiovascular disease? In particular, can early loss of GFR be slowed or even reversed, as we now believe microalbuminuria can regress? Do albuminuria and loss of GFR warrant different targeted therapeutic interventions? As Perkins et al. suggest, might successful treatment to improve values for glycemia, BP, and lipid concentrations reduce the rate of decline or halt the fall in GFR? To answer these questions, we are in critical need of a reliable marker of early GFR loss. The development of new biomarkers in nephrology has lagged behind that in other disciplines. For instance, whereas we now can diagnose with great specificity and sensitivity myocardial infarction with troponin concentrations in serum, we can only crudely estimate GFR using serum creatinine concentrations. The National Kidney Disease Education Program (http://www.nkdep.nih.gov) and the National Kidney Foundation (9) now recommend the use of estimating equations to improve the diagnostic accuracy of serum creatinine. These recommendations constitute a large step forward, with GFR most often estimated using the abbreviated Modification of Diet in Renal Disease (MDRD) equation (10). This works reasonably well for patients whose eGFR is 60 ml/min per 1.73 m. However, the lack of accuracy of the MDRD equation above an eGFR of 60 ml/min per 1.73 m constitutes a significant limitation to its application in many clinical scenarios, including early diabetic kidney disease. Even a careful protocol to standardize creatinine measurements across all instruments in clinical laboratories and likely a more consistently applied MDRD equation (11) are still unlikely to permit sufficient accuracy in the “subnormal” range of kidney function (i.e., GFR from 60 to 90 ml/min per 1.73 m or higher). GFR estimates that are based on serum creatinine concentration are further limited by variation in creatinine production on the basis of age, gender, race, and body composition. Therefore, while continuing to encourage the use of creatinine-based estimating equations, the nephrology community should eagerly seek other methods to measure or better estimate GFR. Measurement of serum cystatin C concentration has attracted increasing attention as an alternative method to assess GFR in clinical medicine. Most reports suggest that the relationship of Published online ahead of print. Publication date available at www.jasn.org.