In a large, multiethnic cohort free of cardiovascular disease and with a spectrum of kidney function ranging from normal to moderate CKD, even small decrements in eGFR were associated with significantly higher levels of sTM, sTF, D-Dimer, vWF, factor VIII, PAP and TFPI. Our results suggest that dysregulation of hemostasis could play an important pathologic role in CKD. In particular, these associations may help to explain the consistently strong and independent associations of cystatin C with cardiovascular and overall mortality, despite adjustment for known risk factors including CRP [5
], IL-6 and TNF-α [6
Soluble thrombomodulin and soluble tissue factor had the largest associations with eGFRcys. sTM is a receptor for thrombin found on vascular and lymphatic endothelial cells. Injured endothelial cells release sTM into the circulation, where it binds thrombin, inhibiting it from forming fibrin and activating platelets. sTM has additional anti-thrombotic activity as a cofactor for the thrombin-catalyzed activation of protein C [23
]. Elevated sTM is associated with lower eGFR in CKD and end stage renal disease ESRD [24
], and our results extend these findings to a broader spectrum of eGFR. The association of sTM and cardiovascular disease varies by patient population: in healthy individuals, lower sTM predicts coronary heart disease (CHD), but in patients with established CHD, higher sTM levels are associated with recurrence of CHD and with higher mortality [26
]. Whether higher levels of sTM are associated with CVD or mortality in our cohort would require further investigation.
Tissue factor (TF) initiates the extrinsic coagulation cascade. Smooth muscle cells, adventitial fibroblasts, and pericytes constitutively express TF, but endothelial cells express TF only after injury [27
]. TF has a prominent role in atherosclerotic plaque rupture and acute coronary syndrome (ACS). TF is concentrated in the core of the atherosclerotic plaque; when the plaque ruptures, TF comes into contact with blood, binds factor VII and initiates the extrinsic coagulation cascade [28
]. Higher levels of sTF were reported previously in CKD as compared to healthy subjects [29
]. Epidemiological studies of sTF and CVD occurrence in healthy individuals are lacking, but in patients with ACS, high levels of sTF on admission are associated with poor outcome [30
]. To our knowledge this paper is the first to report an association between higher sTF and lower eGFR in patients without CKD. Whether high levels of sTF contribute to the high cardiovascular mortality of persons with kidney disease is unknown.
D-Dimer, vWF, Factor VIII, and PAP showed moderate associations with lower eGFRcys. Elevated levels of D-Dimer are associated with lower eGFR [12
] and CVD [31
] across the spectrum of eGFR. vWF is increased in subjects with CKD [32
], and it is considered to be a marker for endothelial dysfunction [33
] and CVD risk [34
]. Factor VIII is increased in subjects with CKD [12
]. In the Atherosclerosis Risk in Communities (ARIC) Study, factor VIII was associated with progression of CKD and with fatal MI [36
]. In MESA, Factor VIII was associated with overall but not cardiovascular mortality [37
]. Plasmin, an enzyme that increases fibrinolysis, is inactivated when complexed with a2-antiplasmin to form PAP [11
]. Subjects with CKD have higher levels of PAP than subjects without CKD [12
]. In a cohort with CAD, PAP was associated with higher CVD risk [38
]; in the Cardiovascular Health Study (CHS), PAP was associated with incident myocardial infarction [39
Several of the hemostatic markers evaluated in this study are strong risk factors for venous thomboembolic disease, including D-dimer, factor VIII, and vWF [40
]. Recent epidemiological studies have linked CKD with risk of venous thrombosis [41
]. Thus, our findings may be applicable to both arterial and venous thrombosis, and these associations may extend into higher levels of kidney function than previously appreciated.
There are several possible mechanisms to explain the association of lower eGFR and higher levels of hemostatic factors. A direct effect of decreased renal clearance may explain an increase in levels of smaller molecular weight hemostatic markers such as sTF and sTM, as these may be filtered at the glomerulus. Fragments of sTM ranging from 28 to 108 kDa can be detected in blood and urine [43
]. Soluble tissue factor has a MW of 29 kDa. Urinary tissue factor, with MW of 34 kDa, is elevated in glomerulonephritis and various malignancies, but it is thought to originate from the kidneys, rather than being blood borne and filtered at the glomerulus [44
]. Whether sTF and sTM are filtered at the glomerulus would have to be tested in animal studies. Given the higher molecular weight of D-Dimer, vWF, factor VIII and PAP, it seems unlikely that these markers are elevated as a direct result of decreased renal clearance. It is possible that elevations in these markers relate to processes initiated by smaller molecules. Conversely, kidney dysfunction may generate a thrombotic milieu indirectly through electrolyte or acid-base abnormalities, which may alter activities of enzymes involved in coagulation. Inflammation associated with lower eGFR may cause elevations in markers that are acute phase reactants. Alternatively, microvascular disease may promote the development of CKD, and in parallel, may cause elevations in these hemostatic biomarkers; in this scenario, the markers would be neither cause nor effect of lower eGFR.
A strength of this study is the unique setting, as MESA is a large, well-characterized cohort including four major race/ethnic groups. The use of cystatin C to measure eGFR increases our ability to detect associations of biomarkers across a broader spectrum of kidney function. However, there are important limitations to our study. As is typical for most large epidemiological studies, we did not use a gold standard measure of eGFR, such as iothalamate or iohexol clearance. Due to the cross-sectional design, we cannot ascertain the direction of association; it is equally possible that increased levels of thrombotic markers causes decreased eGFR, or that decreased eGFR causes the elevated markers. Finally, residual confounding may exist in our analysis, which may account for some of our findings, but would be less likely to account for the strong associations of eGFRcys with sTM and sTF.