Serum creatinine-based estimates of glomerular filtration rate (GFR) are widely used to stage chronic kidney disease (CKD) in diabetics and other high-risk individuals, but variables such as the patient’s muscle mass and diet can affect the accuracy of the staging. Cystatin C has been proposed as an alternative marker that might improve CKD staging, and researchers recently evaluated a cystatin C-based estimated GFR (eGFR) equation in diabetics. Their results are the subject of this issue of Strategies.
The two widely-used equations to estimate GFR, MDRD and CKD-EPI both incorporate serum creatinine, a standard component of basic metabolic profile testing. This has facilitated routine reporting of eGFR, a simple clinical tool to identify patients with CKD. However, owing to the well-described limitations of serum creatinine, researchers and clinicians have long-desired more robust eGFR equations. Recently a consortium of U.S.- and Finland-based researchers evaluated eGFR based on cystatin C (Diabetes Care 2012;35:2311–6).
A protein freely filtered at the glomerulus and almost completely reabsorbed in the proximal tubules, cystatin C has been proposed as an alternative marker for CKD. Unlike serum creatinine, it is not influenced by variables such as muscle mass and diet, and eGFR based on cystatin C (eGFRcyst) has been shown to improve identification of nondiabetic patients with CKD stage 3. Studies also have demonstrated that eGFRcyst predicts death from cardiovascular disease in the elderly more accurately than the CKD-EPI eGFR equation (eGFRcreat). In addition, research has shown that neither eGFRcreat nor eGFRcyst are superior when compared against the gold standard measurement for GFR using iothalamate clearance.
Given the growing evidence base around the utility of eGFRcyst in identifying individuals with CKD at risk for complications, the authors sought to evaluate how this measure would perform against eGFRcreat in predicting risk of end stage renal disease (ESRD) and death in diabetics, a population particularly vulnerable to developing ESRD.
“In our studies of early diabetic nephropathy we have been using cystatin C to look at the changes in renal function among patients with microalbuminuria, and we published several papers on this subject,” explained first author Adrzej Krolewski, MD, PhD. “This study was a natural extension of that work. We were aware that creatinine had some limitations, so we wanted to test whether we could use cystatin C to stratify patients with diabetes according to their risk of end stage renal disease.” He is head of the section on genetics and epidemiology at the Joslin Diabetes Center in Boston and associate professor of medicine at Harvard Medical School.
Krolewski and his co-authors compared the eGFRcreat and eGFRcyst using data from three long-term follow-up studies of the risk of ESRD, including two conducted at the Joslin Diabetes Center and one in Finland. The Joslin populations included one cohort of 364 type 1 diabetics and another of 402 type 2 diabetics. The 399 type 1 diabetic Finnish subjects all were participants in the Finnish Diabetic Nephropathy (FinnDiane) cohort. All patients had been previously diagnosed with CKD stages 1-3, had baseline concentrations of creatinine and cystatin C measured, and were followed for an average of 8–10 years. Patients in both the type 1 diabetes cohorts were similar, but the type 2 diabetics were on average 15–17 years older at enrollment and had higher mean eGFRcreat, as well as a lower incidence of ESRD.
The authors found that overall, CKD staging by eGFRcyst agreed with that of eGFRcreat in 62% of Joslin patients and 73% of FinnDiane subjects. However, patients with a higher CKD stage by eGFRcyst than by eGFRcreat had a significantly higher risk of ESRD in comparison to those who had concordant CKD staging by the two equations, with a hazard ratio of 2.3. Similarly, subjects with a lower CKD stage by eGFRcyst than by eGFRcreat had a lower risk of ESRD in comparison to those who had concordant staging by the two equations, with a hazard ratio of 0.30.
“We showed that cystatin C has better diagnostic performance than serum creatinine as far as predicting time to end stage renal disease in patients who had type 1 diabetes and proteinuria. We found this to be true in U. S. as well as Finnish subjects. Furthermore, we showed exactly the same thing in a third cohort of type 2 diabetics,” explained Krolewski.
The findings add to prior research indicating that cystatin C is a robust marker of CKD, and should put to rest notions that it in some way captures risk for ESRD and CVD that is unrelated to kidney disease, according to Michael Shlipak, MD, MPH, chief of general internal medicine at the San Francisco VA Medical Center and professor-in-residence for the departments of medicine, epidemiology, and biostatistics at the University of California, San Francisco.
“The most important contribution of this study is really in honing the argument that cystatin C is a better marker of kidney function in populations who are not healthy volunteers, in the real world,” he said. “The main controversy in cystatin C has been how can it be so strong prognostically in longitudinal studies and yet in cross-sectional studies not be superior to creatinine like we would expect it to. I think that’s mostly the populations being studied, but many people like to focus on the possibility that cystatin C predicts longitudinal outcomes like death, cardiovascular disease, and end stage renal disease in a pathway that’s not related to kidney disease.”
The latter theory has hindered the use of cystatin C, contended Shlipak. “Every time we get close to identifying a clinical indication for cystatin C, this alternate hypothesis has really stymied the field. If you can dismiss the marker as not related to kidney function then you lose momentum for it as a test of kidney function. This study showing in a longitudinal analysis that it predicts end stage renal disease sort of takes that alternate pathway off the map.”
Krolewski emphasized that the findings show cystatin C’s utility in a select group of diabetics. “If a diabetic has proteinuria or if eGFR calculated using creatinine indicates they have chronic kidney disease, then that patient should have baseline measurement of cystatin C and follow-up measurements every couple of years, to help the doctor assess the patient’s real risk of end stage renal disease. For example, if eGFR using creatinine shows that individual has chronic kidney disease stage 3 but the eGFR using cystatin C comes out as stage 1 or 2, this person is at low-risk of progressing to end stage renal disease,” he said. “If a diabetic doesn’t have proteinuria and his eGFR calculated with serum creatinine is in the normal range, it is unlikely that cystatin C will help predict end stage renal disease. Only those persons whose eGFR calculated using cystatin C shows chronic kidney disease stage 3 or 4 will be at really high risk of progressing to ESRD.”
Krolewski suggested this pragmatic approach would reserve relatively more costly cystatin C measurements for patients whose test results would be more likely to inform their care. Shlipak pointed out that while the charges for cystatin C testing tend to be considerably higher than for serum creatinine, the incremental costs of measuring cystatin C are not much more expensive than for serum creatinine. He urged labs to offer the test in-house, particularly using assays calibrated to the international reference material. “There’s currently a lot of inertia in which nephrologists or other clinicians say they’d like to have a cystatin C test but it’s not available in their hospital lab. I also talk with clinical chemists who say they’d love to offer it in-house but no one seems to order it. So these groups are not really communicating with one another,” he said.
Shlipak added that a committee he participated in for Kidney Disease Improving Global Outcomes soon would issue updated guidelines on CKD classification and management. The document calls on hospitals with adequate resources to offer readily available cystatin C testing to meet the optimal care described in the guidelines.