In our community-based sample not selected for kidney-related outcomes, a broad array of inflammatory markers was associated with CKD status, cystatin C concentration and urinary albumin excretion. Even among individuals without CKD by creatinine-based eGFR, markers of inflammation were strongly associated with higher cystatin C quartiles. Most notably, a significant proportion of variability in TNFR2 concentration was explained by CKD status and higher cystatin C quartiles.
Our data show that higher concentrations of TNF-alpha, interleukin-6, TNFR2, osteoprotegerin and fibrinogen are associated with CKD status, higher cystatin C quartiles and higher UACR quartiles. In addition, higher concentrations of MCP-1 and myeloperoxidase correlate with CKD status; higher levels of CRP, ICAM-1, P-selectin, MCP-1 and myeloperoxidase correlate with higher cystatin C quartiles; and higher concentration of ICAM-1 correlates with higher UACR quartiles.
Inflammation, oxidative stress and a pro-coagulant state are thought to be important mechanisms for CVD in patients with CKD [4
]. CRP is a well-established marker of inflammation, and TNF-alpha along with interleukin-6 is a key cytokine that mediates both acute and chronic inflammation [23
]. TNFR2, a soluble TNF-alpha receptor, likely plays an important role in downstream signalling of inflammatory and proliferative mediators [24
]. ICAM-1 and P-selectin facilitate leucocyte recruitment and vascular transmigration, and MCP-1 recruits monocytes during the inflammatory process [23
]. CD40 ligand plays a role in inducing cytokine production, fibroblast proliferation, lymphocyte proliferation and endothelial cell activation [25
]. Osteoprotegerin facilitates monocyte adhesion and sensitizes endothelial cells to TNF-alpha [26
]. Myeloperoxidase and urinary isoprostanes are markers of oxidative stress [27
], and fibrinogen is an indicator of a pro-coagulant state [28
Our findings add to the existing literature on inflammatory biomarkers and kidney disease. Multiple reports have suggested associations between kidney measures and one or small panel of inflammatory markers [5,6,8–12
]. More recently, a report from the Multi-Ethnic Study of Atherosclerosis (MESA) cohort showed significant associations between a panel of six inflammatory markers that included CRP, interleukin-6, TNF-alpha receptor-1 (TNFR1), ICAM-1, fibrinogen, and factor VIII and both CKD and cystatin C [7
]. The strongest association was observed with TNFR1 for both CKD and cystatin C. The MESA Study, however, did not test the associations with urinary albumin excretion, and their panel did not include markers of oxidative stress and osteoprotegerin pathways. Our study is unique for using a wide-breadth of biomarkers involved in multiple pathways of inflammatory response as well as for robust multivariable adjustments.
We found TNFR2 to be the marker most associated with measures of kidney function and albuminuria. Our results are in agreement with the Health Professionals’ Follow-up Study’s findings which showed that men with TNFR2 concentrations in the highest as compared with the lowest quartile have eight times more likelihood of having CKD [12
]. Similarly, the post hoc analysis of the Cholesterol and Recurrent Events trial also showed that participants with the highest tertile compared with the lowest tertile of TNFR2 have the higher rate of loss of kidney function [11
]. TNF-alpha receptors are cleared by kidneys [29
], and elevated levels seen with lower kidney function may be a reflection of decreased clearance. Our findings with TNFR2, the MESA Study’s findings with TNFR1, and the strong associations we observed between TNF-alpha and kidney function suggest that TNF-alpha pathway may potentially be a key player in the mediation of inflammation in kidney disease.
CRP, in our study, was not related to CKD after accounting for CVD risk factors. This finding is consistent with the results from post hoc analysis of the MDRD Study which showed that CRP did not correlate with worsening kidney function over a 2.2-year follow-up period [30
]. Our finding, however, is in contrast with the observations in the MESA Study, the Cardiovascular Health Study, and the report from the National Health and Nutrition Examination Survey III dataset that showed higher CRP in individuals with CKD [7–9
]. Difference in the definition of CKD and the use of more limited covariate models in other reports may account for the inconsistent results. It is, therefore, uncertain if CRP is associated with mild-to-moderate CKD in the community after accounting for multiple CVD risk factors.
Our results for markers of oxidative stress were variable. Unlike a prior report of higher plasma F2 isoprostane concentration in individuals on haemodialysis [31
], we observed lower urinary isoprostane levels in participants with CKD. We also did not observe significant relations between urinary isoprostane concentration and higher cystatin C quartiles. Higher myeloperoxidase levels, on the other hand, were associated with both CKD and higher cystatin C quartiles. Future studies are needed to better delineate roles of specific oxidative stress markers in kidney disease.
Our study has a number of strengths. We examined a broad panel of biomarkers representing multiple pathways involved in the inflammatory process. Our assessment of kidney disease was comprehensive with the use of serum creatinine, cystatin C and UACR. Our multivariable adjustments were robust, and our sample size provided adequate power to detect modest associations.
The most important limitation of our study is our inability to infer causality from the observed associations. This limitation is inherent in cross-sectional and observational studies and is particularly important in our study as some inflammatory biomarkers may have shown associations with kidney disease simply as a result of lower renal clearance. Higher biomarker concentration in these cases may be the result of rather than the cause of lower kidney function. This may be an important consideration for markers like TNF-alpha receptors which are cleared by kidneys but not for biomarkers such as CRP and interleukin-6 which are non-renally cleared. [29
Our study has some other important limitations. Our samples were composed entirely of individuals of European ancestry, and it is uncertain how our findings might apply to other racial and ethnic groups or to a group selected for kidney pathology. We defined CKD using a one-time measurement of serum creatinine, and this could have led to misclassifications as the MDRD equation underestimates kidney function in individuals without CKD [32
]. Some individuals who were classified as having CKD may have had reversible causes for lower kidney function as well. It is also possible that albuminuria and reduced kidney function may just be markers of systemic vascular disease, and the associations between inflammatory biomarkers and kidney measures may have reflected associations between inflammation and vascular diseases. Our primary findings, however, did not change when individuals with prevalent CVD or diabetes were excluded from the analysis. Finally, not accounting for the presence of chronic inflammatory conditions in our participants may have potentially confounded our analyses as well.
Our study has important implications for future research. Our findings show that inflammation is significantly associated with mild-to-moderate CKD in community-dwelling individuals. Our results also suggest that numerous pathways are likely involved in mediating inflammation in CKD. The strong results for TNF-alpha and TNFR2 indicate the particular importance of the TNF-alpha pathway in kidney disease. Future research from well-powered longitudinal studies is needed to better characterize the potential role of various inflammatory markers in assessing for adverse CVD outcomes in individuals with mild-to-moderate CKD.