Our analysis demonstrated that cystatin C was significantly correlated with all six procoagulant and inflammatory biomarkers across a broad range of kidney function, even after adjustment for age, gender, race/ethnicity, and BMI. While creatinine-based eGFR had significant correlations with all biomarkers except ICAM-1 in subjects with CKD, it was only associated with TNF-αR1 in participants without CKD. Adjusted mean levels of all biomarkers increased significantly across each quintile of cystatin C, and all biomarkers except ICAM-1 were elevated in persons with eGFR < 60 compared with eGFR ≥ 60. In general, TNF-αR1 had the strongest correlations with both cystatin C and eGFR in all groups.
Both creatinine-based eGFR and cystatin C correlated with most inflammatory markers in subjects with chronic kidney disease. However, in patients without CKD, only cystatin C had significant correlations with all markers of inflammation. There are several possible explanations for the association of cystatin C with procoagulant and inflammatory biomarkers in patients without chronic kidney disease. One possibility is that GFR is linearly associated with inflammation, and using cystatin C, a marker of renal function that is less dependent on muscle mass or age, reveals the true association between kidney function and markers of inflammation when GFR is greater than 60 mL/min/1.73 m2. Prior studies have shown that inflammatory markers are not associated with creatinine-based eGFR above 60 [8, 9, 16]. However, this absence of association may be due to imprecision of eGFR in the normal range. A second explanation is that cystatin C is associated with inflammation independent of kidney function [23, 24]. One study found that cystatin C was associated with CRP independent of creatinine clearance; however, that study did not have the gold standard of measured GFR . Another study, also without a gold standard for GFR, found that the association between cystatin C and CRP disappeared after adjustment for 24-hour urine creatinine clearance . Our observation in this current study, that creatinine-based eGFR and cystatin C have similar associations with inflammatory markers among persons with CKD, makes it seem less likely that cystatin C has a direct association with inflammation that is independent of kidney function. However, our study also lacks a gold standard measurement of GFR, and therefore cannot be conclusive.
Both cystatin C and eGFR had substantial correlations with TNF-αR1, while the associations with the other inflammatory and procoagulant markers were more modest. In a murine model, one study demonstrated that 125I-labeled soluble TNF receptors were primarily cleared by the mouse kidney . In contrast, other inflammatory markers, such as CRP and IL-6, are primarily cleared by the liver [27, 28]. While the strong association between kidney function and TNF-αR1 levels may be simply attributable to renal clearance of TNF-αR1, TNF-α itself may also play a more complex role in the mediation of kidney damage. A future study should evaluate whether TNF-α and its soluble receptors predict the longitudinal progression of kidney disease.
A secondary goal of our analysis was to assess the role of race/ethnicity in the association between cystatin C and markers of inflammation. The association of four biomarkers – CRP, IL-6, ICAM-1, and factor VIII – with cystatin C had statistically significant interactions by race/ethnicity. Specifically, these four biomarkers had higher beta coefficients in whites compared with other races/ethnicities. One important issue is that, according to recent data, the assay used in MESA is only able to detect certain polymorphisms of ICAM-1 . One allele in particular, ICAM-1 RS5491-T, is more common in African-Americans and not detected by the MESA assay. Further research is ongoing to evaluate the importance of polymorphisms on detection of circulating markers of inflammation. In the MESA study, we are not aware of other assay issues that would affect the interpretation of race/ethnicity interactions with biomarkers.
One possible explanation for the interactions by race/ethnicity in cystatin C models is that regulation of inflammatory cytokines is more dependent on intact kidney function in whites than in other races/ethnicities. Prior studies have shown that CRP and fibrinogen levels are lower in whites than in African-Americans, and that whites at a given baseline level of CRP seem to have slower rises in serum creatinine over time compared with African-Americans [30–32]. These findings would suggest that whites may have a propensity for greater renal excretion of cytokines, although additional studies using urine measurements of cytokines may be helpful to evaluate these associations more effectively. Another explanation for these findings is that cystatin C is a better marker of GFR in whites versus other races/ethnicities. Overall, the association of cystatin C with GFR in non-white groups has not been well studied.
Our analysis has several limitations. First, while cystatin C has been conclusively demonstrated to be a reliable marker of kidney function, it may have associations with inflammation that are dependent of kidney function. Such associations would not be supported by this study, however, since we found that cystatin C and creatinine-based eGFR had equally strong associations with inflammatory markers for subjects when eGFR < 60. Second, as a cross-sectional study, we are unable to determine temporality in the association between cystatin C and multiple biomarkers. For example, inflammation may lead to declining kidney function, or reduced kidney function may lead to elevated inflammatory biomarkers. As stated above, we did not have a gold standard measurement for kidney function, such as iothalamate clearance. We also assumed that eGFR < 60 was the appropriate cutpoint for chronic kidney disease for all subjects, although some data suggests that the established MDRD equation may need to be modified to more accurately characterize CKD race/ethnicity groups other than whites and African-Americans .