New Potential Biomarkers for Chronic Kidney Disease Management—A Review of the Literature

The prevalence of chronic kidney disease (CKD) is increasing worldwide, and the mortality rate continues to be unacceptably high. The biomarkers currently used in clinical practice are considered relevant when there is already significant renal impairment compromising the early use of potentially successful therapeutic interventions. More sensitive and specific biomarkers to detect CKD earlier on and improve patients’ prognoses are an important unmet medical need. The aim of this review is to summarize the recent literature on new promising early CKD biomarkers of renal function, tubular lesions, endothelial dysfunction and inflammation, and on the auspicious findings from metabolomic studies in this field. Most of the studied biomarkers require further validation in large studies and in a broad range of populations in order to be implemented into routine CKD management. A panel of biomarkers, including earlier biomarkers of renal damage, seems to be a reasonable approach to be applied in clinical practice to allow earlier diagnosis and better disease characterization based on the underlying etiologic process.


Introduction
Chronic kidney disease (CKD) is a long-term condition characterized by a progressive and irreversibly loss of kidney function or persisting renal damage [1]. CKD has been recognized as a global public health problem. Estimation reports on prevalence, as well as on related morbidity and mortality, confirmed the high socioeconomic burden of this disease [2], particularly due to progression to end-stage renal disease (ESRD) and association with cardiovascular disease (CVD).
Irrespective of the etiology of CKD, various structural and functional changes within the kidney will develop during the disease course, resulting in glomerular, tubular and vascular injuries [3]. The progression phase of the disease is characterized by a persistent state of inflammation, hypoxia and oxidative stress that contribute to the development of renal fibrosis [4] (Figure 1). In clinical practice, glomerular filtration rate (GFR) estimation and albuminuria are widely used for CKD diagnosis and prognosis. GFR estimation correlates with the degree of kidney dysfunction, while albuminuria identifies the presence of renal damage. However, these traditional biomarkers only increase when a significant filtration capacity has been lost and kidney damage is advanced [5]. Thus, they increase when several injuries have already occurred in renal cells [6,7]. Early identification of CKD is an important unmet medical need, not only to predict and prevent CKD progression, but also to further improve patients' survival and reduce associated morbidities. Hence, more sensitive and early biomarkers are needed to achieve this goal.
Over the last twenty years, several new biomarkers were identified as promising candidates for CKD management. In this review, we selected some of the most well-studied and well-defined biomarkers, which are associated with different pathophysiological mechanisms underlying CKD initiation and progression. We summarize the more important results from published studies on selected biomarkers of renal function, tubular lesions, endothelial dysfunction, and inflammation. A section on novel metabolomic biomarkers is also included due to the increasing attention towards metabolomic analysis in this field. A search on PubMed of human studies published between 2015 and 2020 was conducted using the keywords "CKD", "chronic kidney disease" and "chronic renal insufficiency"; the biomarker name (s) to search the title or abstract; or the MeSH term "Metabolomics". From the retrieved articles, and after title and abstract screening, we selected studies that evaluate the validity of the new biomarkers on CKD diagnosis and prognosis in patients from different backgrounds. Additionally, we searched for additional publications in the references of the selected articles.

New Biomarkers for Chronic Kidney Disease Management
Renal disease is caused by an initial loss of nephrons; then, as a consequence of the kidney functional adaptations to the damage, disease progresses sequentially through different pathophysiological processes, leading to an irreversible state of fibrosis. Moreover, tubulointerstitial hypoxia, inflammation and oxidative stress are simultaneously a cause and effect of renal injury and form a vicious cycle in CKD progression.
The search for new biomarkers should focus on better indicators of renal dysfunction than GFR, and on markers of specific types of kidney injury, assessed in serum and/or urine. The study of specific biomarkers would allow the identification of kidney damage, and they should reflect the underlying pathophysiological processes of kidney damage [8], namely, changes in renal function, tubulointerstitial damage, endothelial dysfunction and inflammation, and/or cardiovascular risk ( Figure 2). The assessment of renal function relies mainly on the estimation of GFR, using creatinine or creatinine and cystatin C-based equations. Because of the known limitations of these markers, several alternative markers have been studied, namely beta trace protein (BTP) and β2-microglobulin (B2M) ( Table 1).
Beta trace protein and β2-microglobulin are low molecular weight proteins that are filtered by the glomeruli and almost completely reabsorbed by the proximal tubules. Since its urinary excretion is residual, the increase in these proteins has been proposed as potential serum markers of decreased GFR, and also as markers of tubular damage if its urinary excretion is elevated [9]. According to Inker et al. (2016) and Foster et al. (2017), BTP and B2M are not affected by ethnicity, and less influenced by age and gender than serum creatinine [10,11].
BTP, also known as prostaglandin D2 synthase, catalyzes the conversion of prostaglandin H2 to prostaglandin D2. It has been used as a marker of cerebrospinal fluid leakage since it is an important constituent of cerebral spinal fluid and is in much lower concentrations in blood [12]. Many studies have described and compared the diagnostic performance of BTP with the traditional markers of CKD [13][14][15], reporting that increased urinary and systemic BTP concentrations were highly correlated with creatinine and cystatin C concentrations [16,17]; moreover, the urinary concentration of BTP progressively increased, along with GFR reduction, and with a rise in its serum concentration [16]. A study conducted by Motawi et al. (2017) showed that serum BTP was higher in type 2 diabetes (T2D) CKD patients with microalbuminuria than in patients with normoalbuminuria, suggesting the potential of BTP in the early diagnosis of diabetic nephropathy [18]. These results are in accordance with previous findings from Dajak et al. (2011) showing high diagnostic accuracy for BTP to detect albuminuria (>30 mg/g) [19].
B2M is a component of the major histocompatibility class I molecule family that is expressed on the surface of most nucleated cells and is present in most biological fluids [20]. Serum and urinary concentrations of B2M increase with the progression of CKD and are particularly high in ESRD patients [20,21].
Several studies have also documented the association of both BTP and B2M with progression to ESRD, CVD development, and mortality [22][23][24]. In spite of the association of BTP and B2M with ESRD, only B2M levels were associated with mortality in a 14-year follow-up study in 250 T2D Pima Indians [22]. However, a larger study, comprising 3613 adults from the Chronic Renal Insufficiency Cohort study, showed that both BTP and B2M were independent predictors of ESRD and all-cause mortality. A cohort study carried out in 9703 participants from the Atherosclerosis Risk in Communities study confirmed the potential utility of measuring changes in B2M to predict worsening to ESRD. A 30% decline in kidney function, assessed using this novel filtration marker, was strongly associated with ESRD [24].
Some researchers used BTP and B2M to develop newer estimated GFR (eGFR) equations. However, in most studies, no advantages were found over the traditional equations used in traditional practice. Pottel et al. (2018) and White et al. (2019) described no additional benefits of adding BTP and B2M to creatinine and cystatin C equations [25,26]. However, a meta-analysis that included 23,318 individuals from six different studies compared the improvements in reclassification of eGFR, risk prediction of ESRD and mortality using BTP-and B2M-based eGFR equations. Compared to the traditional creatinine-based equations, eGFR BTP-and B2M-based equations improved the risk associations and fairly improved the prediction for ESRD and mortality (after adjustment for established risk factors). Moreover, the best performance was found using an eGFR equation that included a combination of both markers [27].

Klotho
Klotho is a transmembrane protein, mainly expressed in the proximal and distal tubular cells [28]. CKD patients present reduced levels of klotho. Klotho deficiency is positively correlated with kidney function decline, even in CKD patients at stages 1 and 2 [29,30]. Therefore, klotho was previously described as an extremely sensitive and early marker in CKD [29] since its levels reflect the degree of renal insufficiency. In accordance with such findings, Qian et al. (2018) reported that changes in soluble klotho could be used as indicators of CKD progression, since they correlated with changes in eGFR [31]. Likewise, a meta-analysis by  reported a significant positive correlation between soluble klotho levels and eGFR in CKD patients [32].
In CKD patients, decreased klotho levels were also associated with increased albumin excretion [33], higher risk of CVD [34], mortality [35], and CKD related inflammation [36]. Moreover, klotho has a pathophysiological role in ion disorders, and might be used as a marker of abnormal phosphate and bone metabolism in these patients. Zheng et al. (2018) demonstrated that serum klotho levels were associated with the degree of mineral bone disorder in a group of 125 hemodialyzed patients. Patients with osteoporosis presented lower klotho levels than those with normal bone mass or with osteopenia [37]. In addition, a study conducted on 152 CKD patients showed that klotho levels were negatively correlated with serum phosphate levels, suggesting that decreased klotho aggravates the urinary phosphate excretion disorder [38].
The importance of neutrophil gelatinase-associated lipocalin (NGAL), kidney injury molecule-1 (KIM-1) and N-acetyl-β-D-glucosaminidase (NAG) as early indicators of tubular lesions has been documented, but some conflicting results were found. NGAL and KIM-1 are expressed in tubular epithelial cells in response to injury and have been proposed as early biomarkers of CKD [47,48]; their values are increase before the development of irreversible tubular atrophy and fibrosis.
NGAL is a ubiquitous lipocalin iron-carrying protein, mainly secreted by activated neutrophils in response to bacterial infection. It is also expressed in the renal tubular epithelial cells and released in case of cell damage [47]. Both systemic and urinary NGAL concentrations increase when tubular kidney damage occurs [49]. It is believed that NGAL has a protective role in case of renal damage, since it contributes to the recovery of the epithelium after injury [50]; it appears as a biomarker of tubular renal injury, and not of renal (dys)function. NGAL appears as a promising biomarker for early stages of CKD, since it has been associated with early decline in eGFR and albuminuria [49]. The findings of Abbasi et al. (2020) suggest that urinary NGAL levels increase before the onset of traditional markers (eGFR and albuminuria) in T2D patients [51], which indicates that urinary NGAL could be a complementary measurement in the early diagnosis of diabetic nephropathy. In line with this, a recent meta-analysis conducted by Kapoula et al. (2019) aimed to evaluate the diagnostic accuracy of NGAL for early predicting diabetic nephropathy. This analysis enrolled 22 studies, comprising a total of 683 healthy individuals and 3249 diabetic patients (488 with type 1 diabetes (T1D) and 2761 with T2D). The authors found that both serum and urinary NGAL showed an increasing trend alongside with albuminuria and eGFR aggravation. The highest concentrations of NGAL were found in patients with the highest disease severity. Moreover, NGAL showed a moderate to high capacity for early detection of diabetic nephropathy. Most previous published studies reported a similar tendency regarding NGAL for the diagnosis of diabetic nephropathy [52].
KIM-1 is a cellular receptor responsible for regulating immune cells activity in response to viral infections. It is not detectable in the normal kidney, but elevated levels were found in experimental and clinical kidney damage [48]. KIM-1 is a recognized biomarker for acute kidney injury (AKI), and it is also upregulated in CKD [53]. Zhang et al. (2018) reported that higher baseline concentrations of KIM-1 were associated with higher odds of incident CKD in a cohort of 324 adults with hypertension but without baseline kidney disease [54]. In patients with T2D and early stages of CKD (G1 and G2), serum and urinary NGAL, as well as urinary KIM-1, were increased in association with increased urine albumin-to-creatinine ratio, and appeared as significant predictors of albuminuria [55]. In a follow-up study of 527 adults with T1D, Bjornstad et al. (2018) reported that only NGAL (and not KIM-1) was associated with urine albumin-to-creatinine ratio and incident impaired GFR [56].
NAG is a glycosidase found mainly in the lysosomes of proximal tubular cells. Due to its molecular weight (130,000 Dalton), NAG cannot be filtered at the glomerulus; thus, increased urinary concentrations of NAG are signs of proximal tubular damage [57].
In order to predict progression of CKD in patients with chronic heart failure, Jungbauer et al. (2016) evaluated urinary NAG, NGAL and KIM-1 concentrations in 149 patients, who were followed for a period of 5 years. Strong associations of CKD progression with NAG and KIM-1, but not with NGAL, were found in these patients [58], suggesting their usefulness as cardiorenal markers. However, in a prospective cohort study comprising 250 patients at all CKD stages, NGAL was more strongly correlated with disease progression in more advanced CKD stages than KIM-1 and NAG. In addition, NGAL was the only predictor of ESRD and death [59].
The conflicting results may suggest that NGAL, KIM-1 and NAG have different behaviors depending on the CKD cause. Further studies should focus on the kinetics of NGAL, KIM-1 and NAG in CKD with different etiologies to identify their specific field as renal disease biomarkers.

Liver-Type Fatty Acid Binding Protein (L-FABP)
The liver-type fatty acid binding protein (L-FABP) is abundantly expressed in hepatocytes and in proximal renal tubular cells. Injury of the proximal tubular cells induces upregulation of the L-FABP gene, leading to increased L-FABP expression by these cells and, consequently, to an increase in the urinary L-FABP excretion [60]. The urinary levels of L-FABP have been correlated with the degree of tubulointerstitial damage in renal biopsies [61].
Urinary L-FABP seems to be useful in predicting AKI [62] and also AKI-to-CKD transition [63]. In T2D patients, L-FABP appears to be a more sensitive marker than proteinuria to predict CKD progression. Accordingly, Kathir et al. (2017) reported that urinary L-FABP was associated with a decline in GFR in CKD patients without albuminuria [64]. A prospective, observational, multicenter study, comprising 244 Japanese patients with CKD, correlated higher urinary L-FABP levels with the development of ESRD and CVD, irrespective of diabetes. Non-fatal or fatal CVD events and progression to ESRD were associated with higher L-FABP levels and low eGFR [65]. Similarly, Maeda et al. (2015) showed that L-FABP, as well as urinary albumin-to-creatinine ratio, could be useful in assessing cardiovascular damage in T2D patients at CKD stages 1 and 2, since they correlated with the elevation of cardiac markers and electrocardiogram abnormalities [66].

Uromodulin (UMOD)
Uromodulin (UMOD), also known as Tamm-Horsfall protein, is a kidney-specific protein, exclusively produced by the renal tubules, and it is involved in the control of water-electrolyte balance and in the defense against urinary tract bacterial infections [67]. Under normal conditions, UMOD is the most abundant protein in urine. In the case of CKD patients, with tubular atrophy and interstitial fibrosis, the urinary and serum UMOD concentrations are reduced. Rare mutations in the UMOD gene were associated with hereditary autosomal-dominant tubulointerstitial diseases [67].
An observational study that included 170 CKD patients at stages 1 to 5 pre-dialysis and 30 healthy individuals showed that serum UMOD levels were significantly lower in CKD patients, and correlated negatively with serum creatinine and cystatin C levels and positively with eGFR. Moreover, UMOD circulating levels were highly accurate in the assessment of CKD stages, showing a gradual decline with the progression of kidney disease [68]. Accordingly, Steubl et al. (2016) showed that plasma UMOD was capable of differentiating patients without CKD and with stage 1 CKD, contrarily to creatinine and cystatin C [69]. Furthermore, Scherberich et al. (2018) reported that declining serum UMOD concentrations are associated with loss of kidney function in patients with CKD at stages 1 to 5 and autoimmune kidney diseases, even in the absence of creatinine changes. Decreased levels of serum UMOD were further associated with the structural integrity of the renal parenchyma [70].

Biomarkers of Endothelial Dysfunction
Endothelial dysfunction begins in the early stages of CKD and progresses with renal disease severity [83][84][85][86], suggesting that it may be the key element linking CKD to the increased CV risk in these patients. The mechanisms underlying endothelial dysfunction in CKD include reduced nitric oxide (NO) production due to increased levels of endogenous NO synthase inhibitors, oxidative stress, inflammation, and vascular calcification [87,88]. Several endothelial dysfunction markers have been studied in the context of CKD. In this study, we focused on asymmetric dimethylarginine (ADMA) and fetuin-A, since these biomarkers reflect distinct pathophysiological alterations affecting endothelium integrity (Table 3).

Asymmetric Dimethylarginine (ADMA)
Asymmetric dimethylarginine (ADMA) is the most effective endogenous inhibitor of NO synthase, and its accumulation contributes to endothelial dysfunction and, therefore, to atherosclerotic changes [89]. In fact, ADMA is an independent risk marker for CVD development and for all-cause mortality [90].
In a recently published study, ADMA was positively associated with endothelial dysfunction and with CKD duration and severity in pre-dialysis patients [86]. Other studies found similar associations between ADMA levels and eGFR. A cross-sectional study conducted on 176 CKD patients showed a correlation between increasing plasma levels of ADMA and kidney function deterioration; moreover, stage 5 patients registered the highest elevation on ADMA plasma levels. The increase in ADMA was associated with by both eGFR < 60 mL/min/1.73 m 2 and anemia, two risk factors for CVD [84]. Inverse associations between eGFR and ADMA concentrations were also found in an elderly Korean cohort; in this study, the mean ADMA levels were significantly higher in subjects with eGFR < 60 mL/min/1.73 m 2 than in those with a higher eGFR [91]. A systematic review and meta-analysis conducted by  that analyzed six prospective and cross-sectional studies correlated circulating ADMA concentrations with carotid intimamedia thickness in CKD patients. Since carotid intima-media thickness is an important predictor of atherosclerosis, the researchers proposed that elevated circulating ADMA levels are predictors of atherosclerotic disease development in these patients, and not only a marker of subclinical atherosclerosis [92]. Furthermore, Bartnicki et al. (2016) found that left ventricular structural and functional pathological changes occur in parallel with elevated serum ADMA concentrations [93]. Triches et al. (2018) showed that higher circulating ADMA levels were associated with new-onset microalbuminuria and/or with a progression in initial albuminuria in at least 30% in a follow-up study of hypertensive and both hypertensive and diabetic patients [85]. An increase in ADMA may precede albuminuria development, suggesting that ADMA levels are an earlier biomarker of CKD in this population. Subjects who presented increased ADMA values during the follow-up period progressed to the later stages of CKD [85]. Moreover, the magnitude of the ADMA increase might be a marker of the rate of kidney disease progression. Additionally, Seliger et al. (2016) demonstrated in a group of older hypertensive patients with and without CKD, that laser Doppler flowmetry-based measures of endothelial dysfunction and microvascular reactivity are the strongest determinants of albuminuria, irrespective of diabetes status [94]. This study provides further evidence that endothelial function is correlated with albuminuria and that albuminuria may reflect glomerular microvascular dysfunction.

Fetuin-A
Fetuin-A, a vascular calcification inhibitor, is also a risk factor for the development of endothelial dysfunction in CKD patients [95]. Reduced serum levels of fetuin-A have been associated with increased CV mortality in maintenance hemodialysis patients [96]. However, the relationship between vascular calcification and endothelial dysfunction is not so clear in patients with early CKD.

Mutluay et al. (2019) conducted a study on 238 CKD patients at stages 3 and 4 and with
ESRD in order to compare fetuin-A levels between them. Fetuin-A levels were significantly lower in ESRD patients than in patients with stages 3 or 4 of CKD. Moreover, in ESRD patients, the lower levels of fetuin-A were associated with higher vascular calcification scores and carotid intima-media thickness [97]. A meta-analysis of 13 studies comprising 5169 patients at all stages of CKD aimed to determine the relationship between circulating fetuin-A levels and the risk for all-cause mortality. Data collected for this meta-analysis by Zhou et al. (2019) suggest that there is a significant association between low fetuin-A levels and higher risk of mortality, independent of diabetes and inflammation, in dialysis patients, but not in non-dialysis patients. In dialyzed patients, a 0.1 g/L increase in fetuin-A levels lowers the risk for all-cause mortality in 8% [98].
In addition to its role in endothelial dysfunction, fetuin-A has also been associated with inflammation [96,99] and nutritional status [100,101], which is a common complication associated with increased mortality in CKD patients [102]. Dialyzed patients have poorer nutritional status than non-dialyzed patients [103], which might explain the differences in fetuin-A levels found between these populations.

Biomarkers of Inflammation
Several studies in the literature suggest that activation of the inflammatory processes in the early stages of CKD drives kidney function impairment and propose that the assessment of inflammatory markers might help in earlier CKD diagnosis. Increased inflammatory markers and changes in GFR have been widely reported [108][109][110]. Moreover, inflammation is a risk factor for CKD-associated morbidity and may contribute to cardiovascular mortality in CKD patients [111] (Table 4).
Increased levels of interleukin-6 (IL-6) and tumor necrosis factor-alpha (TNF-α) involved in the inflammatory response predict poor outcome in patients with renal disease [112]. A proteomic approach identified an increased expression of serum IL-6 and TNF-α in the early stages of CKD, which highlights the importance of these inflammatory biomarkers for CKD patients' diagnoses and management [113]. An observational study conducted by Kamińska et al. (2019), associated IL-6, but not TNF-α, to coronary artery calcification, a risk factor for cardiovascular mortality, in CKD and ESRD patients [114].
The overexpression of other pro-inflammatory cytokines, such as interleukin-8 (IL-8) and interleukin-18 (IL-18), has also been linked to renal function decline [113,115,116]. Likewise, soluble TNF receptors 1 and 2 (sTNFR1 and sTNFR2) showed an important role in the progression of atherosclerosis and kidney diseases [117,118].
Pentraxin 3 (PTX3) is produced by resident and innate immunity cells in peripheral tissues and increases rapidly in response to primary local activation of inflammation [119,120]. A cross-sectional study comprising two community-based cohorts of elderly subjects without CKD found that increased levels of PTX3 were associated with decreased eGFR and could independently predict incident CKD [121]. In ESRD patients undergoing dialysis, plasma PTX3 levels were shown to be more accurate than C-reactive protein (CRP), a traditional inflammatory marker, in predicting all-cause mortality. Moreover, plasma PTX3 levels were associated with other inflammatory markers (IL-6 and TNF-α), with adipokines disturbances and with the worst lipid risk profile [122]. This recent study highlights the association of PTX3 with multiple risk factors, such as inflammation and malnutrition. Furthermore, PTX3 was postulated to be a predictive marker of mortality in patients with advanced CKD [122,123].
Growth differentiation factor-15 (GDF-15) is a member of the transforming growth factor beta cytokine superfamily, and its expression may be induced in response to ischemia. GDF-15 was significantly associated with increased risk of CKD progression in several studies [124][125][126][127]. Nair et al. (2017) showed that circulating GDF-15 levels were strongly correlated with intrarenal expression of GDF-15, stating that GDF-15 may be a marker for intrarenal signaling pathways associated with CKD development and progression [124]. The Framingham study found associations between increased GDF-15 levels and the development of incident CKD with rapid decline of kidney function [128]. This study was limited to participants without CKD and suggests that increases in GDF-15 may help to identify higher risk for CKD development. Tuegel et al. (2018) found that elevated serum levels of GFD-15 were associated with an increased rate of heart failure in adults with CKD, but had no associations with atherosclerotic CVD events [126].
Inflammation and oxidative stress may contribute to DNA damage. In fact, DNA damage has been correlated with enhanced inflammatory states in different conditions, including CKD [129]. The quantification of circulating cell-free DNA (cfDNA), a measure of DNA damage, was proven to be helpful in detecting tumors in patients with renal cell carcinoma [130], and in predicting AKI after cardiac surgery [131].
Patients with CKD present different types of DNA damage and increased circulating levels of cfDNA [129]. Coimbra et al. (2017) analyzed the degree of genomic damage in ESRD patients undergoing hemodialysis (HD), and found that patients presented higher levels of CRP and cfDNA, correlating the persistent inflammatory state of those patients with the degree of DNA injury [132]. Furthermore, in non-dialyzed patients, urinary levels of cfDNA were predictors of unfavorable renal outcomes [133]. These findings were consistent with the results of Li et al. (2020), which demonstrated that increased serum cfDNA levels were associated with an increased risk of diabetic nephropathy in a cohort of diabetic patients [134], suggesting the potential of cfDNA as a biomarker of disease progression.
Watson et al. (2019) developed a Kidney Injury Test (KIT) based on the composite measurement and validation of six biomarkers across a set of 397 urine samples from patients with CKD, or with an increased risk for CKD (diabetic or hypertensive patients). The KIT analyzed DNA, protein, and metabolite markers, including cfDNA. In patients with normal renal function (eGFR) ≥ 90 mL/min/1.73 m 2 ), the KIT score clearly identified those with predisposing risk factors for CKD, which could not be detected by eGFR or proteinuria [135]. Higher PTX3 levels are associated with lower GFR and independently predict incident CKD in the elderly [121] 2016 Prospective cohort 3430 patients with reduced eGFR from the CRIC study plasma IL-6, TNF-α Elevated plasma levels of TNF-α were associated with rapid loss of kidney function in CKD patients [108] 2016 Prospective cohort 543 patients with stage 5 CKD plasma IL-6, TNF-α IL-6 and TNF-a could predict all-cause mortality risk; only IL-6 could classify clinical CVD [112] 2017 Prospective cohort 521 adults with CKD from the C-PROBE and the SKS studies plasma or serum GDF-15 Circulating GDF-15 levels were strongly correlated with intrarenal expression of GDF15 and significantly associated with increased risk of CKD progression [124] 2017 Prospective cohort 984 CKD patients stages 1-5 serum TNFR1, TNFR2 TNFR1 and 2 were associated with CVD and other risk factors in CKD, independently of eGFR [117] 2017 Cross-sectional 1816 community residents randomly selected from the Dong-gu study plasma PTX3 A significantly higher risk of CKD was found in the group with the highest plasma levels of PTX3 when compared to the group with the lowest levels [136] 2017  Adults with CKD and higher circulating levels of GDF-15 presented greater mortality; elevated GDF-15 was also associated with an increased rate of HF [126] 2018 Prospective cohort 200 patients with T2D plasma GDF-15 Higher GDF-15 improved risk prediction of decline in kidney function; in patients with T2D and microalbuminuria, higher GDF-15 was independently associated with all-cause mortality [137] 2018 Cross-sectional 201 patients with CKD and 201 controls plasma PTX3 Plasma PTX3 levels were increased in patients with CKD when compared to controls [138] 2019 Prospective cohort 3664 participants with CKD from the CRIC study plasma GDF-15 GDF-15 was significantly associated with an increased risk of CKD progression [125] 2019 Prospective cohort 57 CKD patients at stages 3-5 and 19 healthy controls serum IL-6, TNF-α TNF and IL-6 were significantly higher in more advanced CKD stages; IL-6, but not TNF-α, was associated with 5-year risk of all-cause mortality in CKD patients [114] 2019 Prospective cohort 318 ESRD patients, undergoing HD and 22 healthy controls plasma PTX3, IL-6, TNFα, CRP When comparing inflammatory mediators, the increase in PTX3 levels was the only predictor of all-cause mortality in dialysis patients [122] 2019 Prospective cohort 124 patients with CKD (stages 1-5) plasma and urinary cfDNA No correlations were found between cfDNA levels and CKD staging; higher urinary levels of cfDNA were associated with worse renal outcomes at 6 months [133]

Metabolomic Studies on CKD Biomarkers
In patients with CKD, metabolomic profiling identified several metabolites associated with alterations in carbohydrates, amino acids, nucleotides, and lipids metabolism. Since the progression of CKD is concomitant with metabolism alterations, these metabolites are being studied as potential biomarkers ( Table 5).
The identification of these small molecules could help in assessing the multiple pathophysiological changes in CKD, since they might reflect early impairments in specific pathways. Lee et al. (2016) demonstrated that the presence of diabetes in CKD patients induces metabolic changes that are reflected in the levels of serum metabolites, such as arginine, N-acetyl-glycoprotein, glutamine, alanine and leucine [139]. CKD patients with diabetes present more severe metabolic changes than those without diabetes. However, these effects are particularly enhanced in the early stages rather than in advanced CKD [139]. A prospective study that enrolled 193 patients who developed CKD during the follow-up period and 193 matched controls found that decreased urinary levels of glycine and histidine were associated with incident CKD [140]. Additionally, Guo et al. (2019) identified 5-methoxytryptophan, canavaninosuccinate, acetylcarnitine, tiglylcarnitine and taurine as potential predictors of early CKD in a study population of 587 adults at all stages of CKD and of 116 healthy controls [141].
Metabolomic approaches identified that kynurenine, a metabolite of the amino acid L-tryptophan and an intermediate in the formation of NAD+, was associated with decreased eGFR and that alterations in the kynurenine/tryptophan ratio were associated with the presence of renal disease [142]. Additionally, kynurenine increases along with CKD progression, since patients with higher CKD stages exhibit higher kynurenine levels [143]. Debnath et al. (2017) showed that kynurenine was positively associated with TNF-α, suggesting that increased kynurenine might result from the ongoing inflammatory process [144].
Moreover, Ma et al. (2019) demonstrated that abnormalities in triacylglycerols and cardiolipins-phosphatidylethanolamines could discriminate CKD progression, preceding ESRD by several years [145]. Furthermore, Hu et al. (2018) identified three serum metabolites (fumarate, allantoin, and ribonate) associated with all-cause mortality from 299 CKD patients from the Modification of Diet in Renal Disease study, and 963 patients with CKD from the African American Study of Kidney Disease and Hypertension cohort [146].
Besides metabolomic studies, there are other omic studies, namely proteomics, that are useful for the study of CKD. A comparison of CKD patients' proteomic urinary profiles against those of healthy controls demonstrated that using a panel of 273 CKD-specific peptides could discriminate between the two groups [153]. The urinary peptide-based classifier CKD273 has been validated in different longitudinal and cross-sectional studies [154] and has sufficient sensitivity and specificity to allow differential diagnosis of CKD different etiologies [153]. The CKD273 represents one of most notable advances from proteomic studies of CKD biomarkers. The CKD273 is now commercially available as a non-invasive diagnostic test, allowing not only an earlier detection of CKD, but also patient stratification. Accordingly, in a cohort of patients with T2D and normoalbuminuria, the CKD273 proteomic biomarker panel was able to predict early diabetic nephropathy [155].   [150] 2017 Cross-sectional 20 CKD patients at stage 3 and 20 at stage 5, and 20 healthy controls Glycoursodeoxycholic acid and 2-hydroxyethane sulfonate were downregulated in the urine of patients, and pregnenolone sulfate was also found to be decreased in plasma when compared to controls [160] 2018 Prospective cohort 56 Brazilian macroalbuminuric CKD patients Lower levels of 1,5-AG, norvaline and l-aspartic acid were significantly associated with the risk of a combined outcome of mortality, dialysis need or creatinine doubling [148] 2018 Retrospective cohort 227 patients with CKD and a nested subgroup of 57 for follow up Eleven metabolites from various metabolic pathways were associated with reduced eGFR; increased urinary concentrations of betaine and myo-inositol were found to be prognostic markers of CKD progression [161]

Future Perspectives and Conclusions
There is an emerging need for the identification of reliable early biomarkers of kidney injury, of progression of the disease and of morbidity and mortality risk. The reliance on only traditional biomarkers may result in a long time lapse, along which successful interventions could be applied. Most of these potentially new biomarkers identified from experimental studies may detect renal injury earlier than traditional biomarkers. Moreover, newer biomarkers can provide information about the pathophysiological mechanism (s) underlying renal disease, predict disease progression, severity and associated cardiovascular and/or all-cause mortality.
Creatinine limitations have been well known for more than thirty years [168]. However, creatinine is still the standard biomarker used for CKD evaluation in clinical practice. Newer glomerular filtration markers, such as BTP and B2M, have proven their potential to improve the accuracy and the predictive value of GFR estimation. With the increased use of cystatin C, BTP and B2M, GFR estimation is likely to undergo further improvements [18,26]. Other markers, such as NGAL, KIM-1 and L-FABP might be helpful in identifying early tubular damage, especially in the creatinine-blind range, and before pathological and irreversible changes occur. Moreover, their specificity allows the recognition of kidney damage separately from changes in kidney function. The initiation and the progression of endothelial dysfunction, closely related to the inflammatory response, is simultaneously a cause and a consequence of both CKD and CVD [110], which highlights the importance of early detection of such pathological pathways through early endothelial dysfunction and/or inflammatory markers assessment.
Advancing laboratory techniques that allow the concomitant analysis of multiple biomarkers are currently being used and have largely impacted the discovery of new biomarkers. Dozens of dysregulated peptides and metabolites were identified by proteomic and metabolomic studies. Furthermore, microRNA analysis studies have highlighted its implication in the pathogenesis of CKD [169][170][171], particularly in kidney fibrotic transformation [172]. Different microRNA expression profiles were identified under different renal disease conditions [173]. Due to microRNAs functional promiscuity, since they usually target proteins from various pathways, confirmation studies are still required to clarify its role in the clinical diagnostic setting. Despite their promising potential, metabolite profiling technologies and microRNA analysis are still expensive and time consuming, explaining why their application is not widely available.
It is unlikely that a single biomarker can predict CKD progression and identify the multiple pathophysiological processes involved in CKD progression or the underlying primary renal disease. Instead, a panel measuring multiple biomarkers seems to be a more reasonable approach to better predict CKD development and to access the outcomes of the disease, since there are several different mechanisms by which CKD can result. Additionally, the incorporation of multiple serum and/or urinary biomarkers is likely to be synergetic in predicting renal failure while decreasing the impact of non-renal determinants for each biomarker alone [174]. Panels of biomarkers for earlier detection of CKD are expected to include biomarkers related to the primary disease cause, while biomarkers for predicting disease progression are more likely to be related to the sequentially reduced renal function. Although it seems reasonable, the availability of such tools depends on the characterization of patients' profiles against healthy controls and the validation of methods for analytical procedures.
CKD burden is not limited to the disease itself, extending to ESRD-and CKD-related CVD and other complications. Identification of CKD at early stages allows more rapid interventions in order to control disease progression and management of risk factors. This will translate in better patients' prognoses and economic benefits related to decreased healthcare expenses.
In summary, further validation for most of these new potential biomarkers requires larger studies with standardized methodologies in order to be implemented in routine CKD management, either for an early diagnosis or for the detection of disease worsening.
Moreover, the search for panel (s) of biomarkers that can synergistically detect renal disease or a poor outcome for renal patients is important.  Figure 2 was created with biorender.com.

Conflicts of Interest:
The authors declare no conflict of interest.