Diabetologia (2018) 61:996–1011

https://doi.org/10.1007/s00125-018-4567-5

REVIEW

Biomarkers of diabetic kidney disease

Helen M. Colhoun 1 & M. Loredana Marcovecchio 2

Received: 24 October 2017 / Accepted: 3 January 2018 / Published online: 8 March 2018
# The Author(s) 2018. This article is an open access publication

Abstract
Diabetic kidney disease (DKD) remains one of the leading causes of reduced lifespan in diabetes. The quest for both prognostic
and surrogate endpoint biomarkers for advanced DKD and end-stage renal disease has received major investment and interest in
recent years. However, at present no novel biomarkers are in routine use in the clinic or in trials. This review focuses on the
current status of prognostic biomarkers. First, we emphasise that albuminuria and eGFR, with other routine clinical data, show at
least modest prediction of future renal status if properly used. Indeed, a major limitation of many current biomarker studies is that
they do not properly evaluate the marginal increase in prediction on top of these routinely available clinical data. Second, we
emphasise that many of the candidate biomarkers for which there are numerous sporadic reports in the literature are tightly
correlated with each other. Despite this, few studies have attempted to evaluate a wide range of biomarkers simultaneously to
define the most useful among these correlated biomarkers. We also review the potential of high-dimensional panels of lipids,
metabolites and proteins to advance the field, and point to some of the analytical and post-analytical challenges of taking initial
studies using these and candidate approaches through to actual clinical biomarker use.

Keywords Biomarker . Diabetic kidney disease . Epidemiology . Nephropathy . Review

Abbreviations CVD Cardiovascular disease

ACR Albumin to creatinine ratio DKD Diabetic kidney disease
ADMA Asymmetric dimethylarginine ESRD End-stage renal disease
ApoA4 Apolipoprotein A4 FGF Fibroblast growth factor
B2M β2-Microglobulin KIM-1 Kidney injury molecule-1
C1QB Complement C1q subcomponent subunit B L-FABP Liver-type fatty acid-binding protein
CD5L CD5 antigen-like MCP-1 Monocyte chemoattractant protein-1
CKD Chronic kidney disease MDRD Modification of Diet in Renal Disease
CKD273 CKD classifier based on 273 miRNA MicroRNA
urinary peptides MR-proADM Mid-regional fragment of
CKD-EPI Chronic Kidney Disease proadrenomedullin
Epidemiology Collaboration NGAL Neutrophil gelatinase-associated lipocalin
NT-proNBP N-terminal pro-B-type natriuretic peptide
PRIORITY Proteomic Prediction and Renin
Electronic supplementary material The online version of this article Angiotensin Aldosterone System
(https://doi.org/10.1007/s00125-018-4567-5) contains a slideset of the Inhibition Prevention Of Early
figures for download, which is available to authorised users. Diabetic nephRopathy In TYpe 2 Diabetic
Patients With Normoalbuminuria
* Helen M. Colhoun
SBP Systolic BP
[email protected]
SDMA Symmetric dimethylarginine
1
SUMMIT SUrrogate markers for Micro- and
MRC Institute of Genetics & Molecular Medicine, The University of Macro-vascular hard endpoints for
Edinburgh, Western General Hospital, Crewe Road,
Edinburgh EH4 2XU, UK Innovative diabetes Tools
2
Department of Paediatrics, University of Cambridge,
Cambridge, UK
Diabetologia (2018) 61:996–1011 997

SYSKID Systems biology towards novel underestimated GFR in a large study [6]. Others have found
chronic kidney disease diagnosis that creatinine agrees more closely than cystatin C with direct-
and treatment ly measured GFR [7]. In those with and without diabetes,
TNFR TNF receptor cystatin C predicts CVD mortality and ESRD better than
VEGF Vascular endothelial growth factor eGFR does [8, 9]. However, this may be because factors other
than renal function that affect ESRD risk, including diabetes,
might also affect serum cystatin C levels, rather than because
Introduction cystatin C-based eGFR is more accurately measuring GFR
itself [10].
Diabetic kidney disease (DKD) and its most severe manifes- Albuminuria strongly predicts progression of DKD but it
tation, end-stage renal disease (ESRD), remains one of the lacks specificity and sensitivity for ESRD and progressive
leading causes of reduced lifespan in people with diabetes decline in eGFR. In type 2 diabetes a large proportion of those
[1]. Even early stages of DKD confer a substantial increase who have renal disease progression are normoalbuminuric
in the risk of cardiovascular disease (CVD) [1, 2], so the [11, 12]. It has been shown that the coexistence of albuminuria
therapeutic goal should be to prevent these earlier stages, not makes DKD rather than non-diabetic CKD more likely in
just ESRD. However, there has been an impasse in the devel- people with type 2 diabetes [13]. However, even in type 1
opment of drugs to reverse DKD, with many Phase 3 clinical diabetes, where non-diabetic CKD is much less common, al-
trial failures [3]. The current hard endpoints for the licencing buminuria was reported to have a poor positive predictive
of drugs for chronic kidney disease (CKD) or DKD approved value for DKD as only about a third of those with
by most authorities, including the US Food and Drug microalbuminuria had progressive renal function decline
Administration, are a doubling of serum creatinine or the onset [14]. Albumin excretion also had low sensitivity, as only about
of ESRD or renal death. Some of the trial failures are due to half of those with progressive renal function decline were
insufficient power, with low overall rates of progression to albuminuric [14]. Clearly, in evaluating the predictive perfor-
these hard endpoints during the typical trial duration of 3– mance of novel biomarkers, investigators should adjust for
7 years. As a result, there is increasing interest in the develop- baseline eGFR and albuminuria. Historical eGFR data are
ment of prognostic or predictive biomarkers to allow for risk not always routinely available. Nonetheless, it is important
stratification into clinical trials, as well as eventually for where possible to evaluate whether biomarkers improve pre-
targeting preventive therapy. There is also interest in the de- diction on top of historical eGFR.
velopment of biomarkers of drug response that are surrogates
for these harder endpoints. Here we review some of the larger
studies published in the last 5 years on prognostic or predictive Clinical predictors of DKD in type 1 and type 2
biomarkers for DKD. Our emphasis is on illustrating some diabetes
key aspects of the approaches being used recently and what
further improvements are needed, rather than systematically Apart from albuminuria and eGFR, other risk factors routinely
reviewing every sporadic biomarker report. captured in clinical records can predict GFR decline. These
have been systematically well reviewed elsewhere [15]. In
brief, established clinical risk factors include age, diabetes
Biomarkers currently in use duration, HbA1c, systolic BP (SBP), albuminuria, prior
eGFR and retinopathy status. However, there have been rela-
It is well established that the best predictor of future ESRD is tively few attempts to build and validate predictive equations
the current GFR and past GFR trajectory [4]. Thus, GFR is the using clinical data that would form the basis for evaluating the
most common prognostic biomarker being used for predicting marginal improvement in prediction with biomarkers
ESRD in both clinical practice and in trials. The Chronic [16–18]. Those that have attempted this reported C statistics
Kidney Disease Epidemiology Collaboration (CKD-EPI) for ESRD or renal failure death or prediction of incident albu-
and Modification of Diet in Renal Disease (MDRD) equa- minuria in the range 0.85–0.90 in type 2 diabetes [17, 18]. In
tions, both based on serum creatinine, are commonly used to the Joslin cohorts with type 1 diabetes, eGFR slope, albumin
estimate GFR. The difference in accuracy for staging between to creatinine ratio (ACR) and HbA1c had a C statistic (not
CKD-EPI and MDRD is slight, with 69% vs 65% overall cross-validated) for ESRD of 0.80 [19–21]. In the FinnDiane
accuracy for given stages being found in one study [5]. cohort the best model had a C statistic of 0.67 for ESRD [22].
Serum cystatin C-based eGFR has been proposed as advanta- In the Steno Diabetes Center cohort, HbA1c, albuminuria,
geous since, unlike creatinine, it is not related to muscle mass. haemoglobin, SBP, baseline eGFR, smoking, and low-
Equations based on cystatin C overestimated directly mea- density lipoprotein/high-density lipoprotein ratio explained
sured GFR, while equations based on serum creatinine 18–25% of the variability in decline [23]. In the
998 Diabetologia (2018) 61:996–1011

EURODIAB cohort predictive models for albuminuria includ- 2 (sTNFR1 and sTNFR2) [28–33], fibroblast growth factors 21
ed HbA1c, AER, waist-to-hip ratio, BMI and ever smoking and 23 (FGF21, FGF23) [25, 34–41] and pigment epithelium-
with a non-cross-validated C statistic of 0.71 [24]. derived factor (PEDF) [42]. Positive associations have also been
In summary, most studies have reported at least modest C found for biomarkers of endothelial dysfunction, including mid-
statistics for models that contain clinical risk factors beyond regional fragment of proadrenomedullin (MR-proADM) [43],
eGFR, albuminuria status and age for renal outcomes in type 1 and cardiac injury, including N-terminal pro-B-type natriuretic
and 2 diabetes. However, despite this, very few biomarker peptide (NT-proBNP) [43]. Copeptin, a surrogate marker for
studies have evaluated the marginal improvement in predic- arginine vasopressin, was associated with albuminuria progres-
tion beyond such factors. In the SUrrogate markers for Micro- sion and incident ESRD independently of baseline eGFR in four
and Macro-vascular hard endpoints for Innovative diabetes studies [44–47]. Proximal tubular proteins, such as urinary
Tools (SUMMIT) study, for example, while forward selection KIM-1, NGAL [48–50] and liver-type fatty acid-binding protein
of biomarkers on top of a limited set of clinical covariates (L-FABP) [51–53] have been associated with a faster decline in
selected a panel of 14 biomarkers as predictive, increasing eGFR [48]. The data are most consistent for KIM-1, a protein
the C statistic from 0.71 to 0.89, a more extensive clinical risk expressed on the apical membrane of renal proximal tubule
factor model already had a C statistic of 0.79 and a panel of cells, with urinary concentrations rising in response to acute
only seven biomarkers showed an improvement in prediction renal injury [49, 54–56]. Urinary and blood levels of KIM-1
beyond this [25]. increased across CKD stages and were associated with eGFR
slopes and progression to ESRD during follow-up in some stud-
ies [57, 58], but it has not always been a strong independent
Novel biomarker studies predictor of progression [59, 60]. There are reports of its asso-
ciation with regression of microalbuminuria in type 1 diabetes
Ideally, we seek predictive or prognostic biomarkers of the [61]. That these associations could reflect a causal role for KIM-
hard endpoint demanded by drug regulatory agencies (i.e. 1 was suggested by an analysis of the FinnDiane cohort with
doubling of serum creatinine or the onset of ESRD or renal type 1 diabetes [62]. In this analysis, KIM-1 did not predict
death). In practice, since many cohorts do not have the neces- progression to ESRD independently of AER. However, using
sary length of follow-up or numbers of incident hard end- a Mendelian randomisation approach, based on genome-wide
points, many studies have sought biomarkers of intermediate association study data for the KIM-1 gene, an inverse association
phenotypes such as incident albuminuria, DKD stage 3 or of increased KIM-1 levels with lower eGFR emerged, suggest-
eGFR slopes above a certain threshold (Table 1). ing a causal link with renal function.

Studies testing single biomarkers or small sets Panels of candidate biomarkers

of biomarkers
Each of the above biomarkers have some evidence supporting
Most biomarker reports in the literature are of single candidate their prediction of renal function decline or other DKD-related
biomarkers or small sets of candidate biomarkers that may be phenotypes. However, although they have been investigated as
assayed in single assays, usually ELISAs, or on multiplexed reflecting specific pathways or processes, in reality there are
platforms, such as the Myriad RBM KidneyMAP panel very strong correlations between these biomarkers, even be-
(https://myriadrbm.com/, accessed 17 October 2017). Until tween different pathways. Figure 2 shows the correlation ma-
recently, most of these studies have taken as their starting trix for some of these from the SUMMIT study [25]. Yet, rel-
point molecules identified from in vitro studies, cell-based atively few studies have assayed many of these candidates
studies or animal models. For example, animal models identi- together to allow the marginal gain in prediction with each
fied kidney injury molecule-1 (KIM-1) [26] and neutrophil additional biomarker to be evaluated. Of those that have, some
gelatinase-associated lipocalin (NGAL) [27]. Candidates stud- used a hybrid of discovery and candidate approaches
ied to date probe pathways thought causal in DKD, such as harnessing bioinformatics and systems biology modelling tech-
inflammation, glycation or glycosylation, or endothelial dys- niques [63]. So, for example, in the SUMMIT study [25], we
function. Others focus on glomerular features, such as glyco- conducted both data mining and literature review to arrive at
calyx abnormalities, extracellular matrix deposition, podocyte sets of candidates that several pathophysiological processes
damage or glomerular fibrosis. Others focus on acute or chron- considered relevant for DKD. We assayed these but also a
ic proximal or distal tubular dysfunction (Fig. 1). larger set of biomarkers (207 in total) that were already
As detailed in Table 1, among these studies of single or few multiplexed with these candidates in the most efficient analysis
biomarkers, some of the most frequently reported associations platforms that were Luminex and mass spectrometry-based.
with DKD-relevant phenotypes are for biomarkers of inflamma- Altogether, 30 biomarkers had highly significant evidence of
tion and fibrosis pathways, such as soluble TNF receptors 1 and association with renal function decline when examined singly
Table 1 Main studies on biomarkers and DKD published between 2012 and 2017

Author, ref. Sample size and Study design DKD stage Biomarkers Main results Adjustments
population

Single biomarkers or several biomarkers not as a panel

Burns et al [102] N = 259 (n = 194 Cross-sectional Normoalbuminuria; Urinary angiotensinogen Urinary angiotensinogen and ACE No adjustments
T1D, n = 65 varying levels of GFR and ACE2 levels, activity associated with ACR
controls) activity of ACE and
ACE2
Diabetologia (2018) 61:996–1011

Velho et al [44] N = 986 Prospective Varying levels of albumin Plasma copeptin Upper tertiles of copeptin associated Baseline sex, age, and duration of
T1D excretion and GFR with a higher incidence of ESRD diabetes
Carlsson et al N = 607 Prospective Varying levels of albumin Plasma endostatin Endostatin levels associated with Baseline age, sex, eGFR and ACR
[103] T2D excretion increased risk of GFR decline and
mortality
Dieter et al [104] N = 135 Prospective Proteinuria Serum amyloid A Higher serum amyloid A levels UACR, eGFR, age, sex and ethnicity
T2D predicted higher risk of death and
ESRD
Wang et al [105] N = 100 (n = 80 with Cross-sectional Varying levels of eGFR Serum and urinary ZAG Serum and urinary ZAG associated with No adjustments
T2D, n = 20 and ACR eGFR and UACR, respectively
healthy controls)
Pikkemaat et al N = 161 T2D Prospective eGFR >60 ml min−1 Copeptin Copeptin predicted development of Age, sex, diabetes duration,
[47] 1.73 m−2 CKD stage 3, borderline significant antihypertensive treatment,
on adjustment for baseline eGFR HbA1c, BMI, SBP
Garg et al [50] N = 91 Cross-sectional Varying levels of albumin Urinary NGAL and NGAL and cystatin C were significantly No adjustments
T2D (including excretion cystatin C higher in participants with vs those
n = 30 with without microalbuminuria
prediabetes)
Viswanathan et al N = 78 (n = 65 T2D, Cross-sectional Varying degrees of Urinary L-FABP L-FABP inversely associated with No adjustments
[52] n = 13 controls) albuminuria eGFR and positively associated with
protein to creatinine ratio
Panduru et al [62] N = 1573 Prospective Varying degrees of Urinary KIM-1 KIM-1 did not predict progression to HbA1c, triacylglycerols, AER
T1D + Mendelian albuminuria ESRD independently of AER
randomisation Mendelian randomisation supported a
causal link between KIM-1 and eGFR
Pavkov et al [31] N = 193 Prospective Varying levels of albumin Serum TNFR1 and Elevated concentrations of TNFR1 or Age, sex, HbA1c, MAP, ACR and
T2D excretion, TNFR2 TNFR2 associated with increased risk GFR
eGFR: ≥60 ml/min in of ESRD
89% participants
Fufaa et al [106] N = 260 Prospective Varying levels of albumin Urinary KIM-1, L-FABP, NGAL and L-FABP independently Baseline age, sex, diabetes duration,
T2D excretion and eGFR NAG and NGAL associated with ESRD and mortality hypertension, HbA1c, GFR, ACR
Bouvet et al N = 36 Cross-sectional Normoalbuminuria and Urinary NAG Higher NAG levels associated with No adjustments
[107] T2D macroalbuminuria microalbuminuria
Har et al [40] N = 142 Cross-sectional Varying levels of eGFR Urinary Increased urinary cytokine/chemokine Glycaemia
T1D Normoalbuminuria cytokines/chemokines excretion according to filtration status
with highest levels in hyperfiltering
individuals, although not significant
after adjustments
999
Table 1 (continued)
1000

Author, ref. Sample size and Study design DKD stage Biomarkers Main results Adjustments
population

Petrica et al [108] N = 91 (n = 70 T2D, Cross-sectional Normoalbuminuria and Urinary Significant association between UACR, cystatin C, CRP
n = 21 controls) microalbuminuria α1-microglobulin and biomarkers of proximal tubule
KIM-1 (proximal dysfunction and podocyte biomarkers
tubule markers), (independently of albuminuria and
nephrin and VEGF renal function)
(podocyte markers),
AGE, UACR and
serum cystatin C
Wu et al [109] N = 462 Cross-sectional Varying levels of albumin Serum Klotho, NGAL, Klotho and NGAL associated with ACR No adjustments
T2D excretion 8-iso-PGF2α, MCP-1,
TNF-α, TGF-β1
Sabbisetti et al N = 124 Prospective Proteinuria Serum KIM-1 KIM-1 associated with eGFR slopes and Baseline ACR, eGFR, and HbA1c
[58] T1D CKD 1-5 progression to ESRD
Velho et al [45] N = 3101 Prospective Albuminuria Plasma copeptin Copeptin independently associated with Baseline sex, age, diabetes duration,
T2D renal events (doubling of creatinine or hypertension, diuretics use,
ESRD) HbA1c, eGFR, triacylglycerols,
HDL-cholesterol, AER
do Nascimento N = 101 Cross-sectional Varying levels of albumin Urinary mRNA levels of Urinary nephrin discriminated between No adjustments
et al [110] (n = 19 prediabetes, excretion podocyte-associated the different stages of DKD and
n = 67 diabetes proteins (nephrin, predicted increases in albuminuria
[T1D, T2D] and podocin, podocalyxin,
n = 15 controls) synaptopodin, TRPC6,
α-actinin-4 and
TGF-β1)
Boertien et al [46] N = 1328 Prospective Varying degrees of Copeptin Copeptin associated with change in Age, sex, diabetes duration,
T2D albuminuria and eGFR eGFR independently of baseline antihypertensive use, HbA1c,
eGFR. This association not present in cholesterol, BP,BMI, smoking
those on RASi
Lopes-Virella N = 1237 Prospective Normoalbuminuria Serum E-selectin, IL-6, TNFR1 and TNFR2 and E-selectin best Treatment allocation, baseline AER,
et al [33] T1D PAI-1, sTNFR1, predictors of progression to ACEi/ARB use, retinopathy cohort,
TNFR2 macroalbuminuria sex, age, HbA1c, diabetes duration
Panduru et al N = 2454 (n = 2246 Prospective Varying degrees of Urinary L-FABP L-FABP was an independent predictor Baseline WHR, HbA1c,
[111] T1D, n = 208 albuminuria of progression at all stages of DKD, triacylglycerols, ACR
controls) but L-FABP did not significantly
improve risk prediction above AER
Araki et al [53] N = 618 Prospective Varying levels of albumin Urinary L-FABP L-FABP associated with decline in Age, sex, BMI, HbA1c, cholesterol,
T2D excretion, serum eGFR triacylglycerols,
creatinine ≤ HDL-cholesterol, hypertension,
8.8×10−2 mmol/l RASi use, BP
Lee et al [112] N = 380 Prospective Varying levels of albumin Plasma TNFR1 and FGF-23 was associated with increased Sex, baseline diabetes duration,
T2D excretion FGF-23 risk of ESRD, only in unadjusted HbA1c, eGFR, AER
model
Cherney et al [41] N = 150 Cross-sectional Normoalbuminuria 42 urinary IL-6, IL-8, PDGF-AA and RANTES No adjustments
T1D cytokines/chemokines levels differed across ACR tertiles
Diabetologia (2018) 61:996–1011
Table 1 (continued)

Author, ref. Sample size and Study design DKD stage Biomarkers Main results Adjustments
population

Conway et al [60] N = 978 Prospective Varying degrees of Urinary KIM-1 and KIM-1 and GPNMB associated with Baseline eGFR, ACR, sex, diabetes
T2D albuminuria and eGFR GPNMB faster eGFR decline, only in duration, HbA1c, BP
unadjusted models
Higher KIM-1 associated with mortality
risk, only in unadjusted models
Diabetologia (2018) 61:996–1011

Nielsen et al [48] N = 177 Prospective Proteinuria Urinary NGAL and Higher levels of the biomarkers Age, sex, HbA1c, SBP and urinary
T2D KIM1 and plasma associated with a faster decline in albumin
FGF23 eGFR, although this was not
independent of known promoters
Jim et al [113] N = 76 (n = 66 T2D, Cross-sectional Normoalbuminuria and Urinary nephrin levels Nephrinuria occurred before the onset of No adjustments
n = 10 controls) microalbuminuria microalbuminuria
Gohda et al [30] N = 628 Prospective Normal renal function; TNFR1 and TNFR2 TNFR1 and TNFR2 strongly associated HbA1c, AER, and eGFR
T1D normoalbuminuria and with risk for early renal decline
microalbuminuria
Niewczas et al N = 410 Prospective CKD 1-3 Plasma TNF-α, TNFR1, TNFR1 and TNFR2 were strongly Age, HbA1c, AER, and eGFR
[29] T2D and TNFR2, ICAM-1, associated with risk of ESRD
VCAM-1, PAI-1, IL-6
and CRP
Fu et al [49] N = 112 (n = 88 with Cross-sectional Varying degrees of Urinary KIM-1, NAG, Higher levels of the three markers in No adjustments
T2D, n = 24 albuminuria NGAL T2D than controls.
controls) Positive association of NGAL and NAG
with ACR; negative association of
NGAL and eGFR
Nielsen et al [59] N = 63 Prospective Varying levels of albumin Urinary NGAL, KIM-1 Elevated NGAL and KIM-1 were Age, sex, diabetes duration, BP,
T1D excretion and GFR and L-FABP associated with faster decline in GFR, HbA1c, AER
but not after adjustments for known
progression promoters
Kamijo-Ikemori N = 552 (n = 140 Cross-sectional and Varying degrees of Urinary L-FABP L-FABP associated with progression of Age, sex, HbA1c, albuminuria status
et al [51] T2D and n = 412 prospective albuminuria and GFR nephropathy at baseline, BP
controls)
Vaidya et al [61] N = 697 (n = 659 Cross-sectional and Varying levels of albumin Urinary IL-6, CXCL10/ KIM-1 and NAG both individually and Age, sex, AER, HbA1c, SBP,
T1D, n = 38 prospective excretion IP-10, NAG and collectively were significantly renoprotective treatment and
controls) KIM-1 associated with regression of cholesterol
microalbuminuria
Panel of biomarkers /proteomics signatures
Coca et al [114] N = 1536 (n = 1346 Nested case–control CKD at various stages TNFR1, TNFR2 and Higher levels of the three biomarkers Clinical variables
T2D, n = 190 study and KIM-1 associated with higher risk of eGFR
controls) prospective decline in persons with early or
advanced DKD
Bjornstad et al N = 527 Prospective Varying levels of albumin Plasma biomarkers B2M, cystatin C, NGAL and Age, sex, HbA1c, SBP,
[69] T1D excretion and eGFR osteopontin predicted impaired eGFR LDL-cholesterol, baseline log
ACR and eGFR
1001
Table 1 (continued)
1002

Author, ref. Sample size and Study design DKD stage Biomarkers Main results Adjustments
population

Peters et al [70] N = 354 Prospective Varying levels of albumin Plasma ApoA4, ApoA4, CD5L, C1QB and IBP3 Age, diabetes duration, diuretic use,
T2D excretion and eGFR ApoC-III, CD5L, improved the prediction of rapid HDL-cholesterol
C1QB, complement decline in renal function
factor H-related independently of recognised clinical
protein 2, IGFBP3 risk factors
Mayer et al [66] N = 1765 Prospective CKD at various stages YKL-40, GH-1, HGF, Biomarkers explained variability of Sex, age, smoking, baseline eGFR,
T2D matrix annual eGFR loss by 15% and 34% ACR, BMI, total cholesterol, BP
metalloproteinases: (adj R2) in patients with eGFR ≥60 and HbA1c
MMP2, MMP7, and <60 ml min−1 1.73 m−2
MMP8, MMP13, respectively.
tyrosine kinase and A combination of molecular and clinical
TNFR1 predictors increased the adjusted R2 to
35% and 64% in these two groups,
respectively.
Saulnier et al N = 1135 Prospective Varying levels of albumin Serum TNFR1, TNFR1, MR-proADM and NT-proBNP Age, sex, diabetes duration, HbA1c,
[115] T2D excretion and eGFR MR-proADM and improved risk prediction for renal BP, baseline eGFR and ACR
NT-proBNP function decline
Looker et al [25] N = 307 Nested case–control CKD 3 207 serum biomarkers Panel of 14 biomarkers improved Age, sex, eGFR, albuminuria,
(n = 154 T2D, clinical prediction (from 0.706 to HbA1c, ACEi and ARB use, BP,
n = 153 controls) 0.868) weighted average of past eGFRs,
diabetes duration, BMI, prior
CVD, insulin use,
antihypertensive drugs
Pena et al [116] N = 82 Prospective Normoalbuminuria and Plasma peptides 18 peptides (related to PI3K-Akt, Baseline albuminuria status, eGFR,
T2D macroalbuminuria VEGF, mTOR, MAPK, and p38 RASi use
MAPK, Wnt signalling) improved
risk prediction for transition from
micro to macroalbuminuria (C
statistic from 0.73 to 0.80)
Pena et al [64] N = 82 Prospective Varying levels of albumin 28 biomarkers MMPs, tyrosine kinase, podocin, CTGF, Baseline smoking, sex, SBP, eGFR,
T2D excretion and eGFR TNFR1, sclerostin, CCL2, YKL-40, use of oral diabetic medication
and NT-proCNP improved prediction
of eGFR decline when combined with
established risk markers
Foster et al [117] N = 250 Prospective Unselected but 54% β-Trace protein and B2M β-Trace protein associated with ESRD GFR, albuminuria, age, sex, diabetes
T2D albuminuric duration, hypertension,
cholesterol
Agarwal et al [67] N = 87 (n = 67 T2D, Prospective CKD 2-4 17 urinary and 7 plasma Urinary C-terminal FGF-2: strongest Baseline albuminuria and eGFR
n = 20 controls) Varying levels of albumin biomarkers association with ESRD
excretion Plasma VEGF associated with the
composite outcome of death and
ESRD
Siwy et al [75] N = 165 Prospective Wide ranges of eGFR Urinary CDK273 Validation of this urinary Albuminuria
T2D and urinary albumin proteome-based classifier in a
multicentre prospective setting
Diabetologia (2018) 61:996–1011
Table 1 (continued)

Author, ref. Sample size and Study design DKD stage Biomarkers Main results Adjustments
population

Verhave et al [68] N = 83 Prospective Overt diabetic Urinary IL-1β, IL-6, MCP-1 and TGF-β1 were independent Albuminuria
T1D and T2D nephropathy IL-8, MCP-1, TNF-α, and additive to proteinuria in
TGF-β1, and PAI-1 predicting the rate of renal function
decline
Bhensdadia et al N = 204 Prospective eGFR stage 1-2 and Urine peptides Haptoglobin to creatinine ratio: best Albuminuria, ACEi use
Diabetologia (2018) 61:996–1011

[84] T2D normo-/ predictor of early renal function

macroalbuminuria decline
Merchant et al N = 33 Prospective Microalbuminuria Small (<3 kDa) plasma Plasma kininogen and kininogen No adjustments but stratum matched
[82] T1D peptides fragments associated with renal for eGFR and albuminuria
function decline
Roscioni et al N = 88 Prospective Normoalbuminuria and CKD273 (urine) Able to detect progression from normo- Baseline albuminuria status, eGFR,
[78] T2D microalbuminuria to micro- and micro- to RASi use
macroalbuminuria
Zürbig et al [76] N = 35 Prospective Normoalbuminuria; Urinary CKD273 Early detection of progression to Albuminuria
T1D and T2D normal eGFR macroalbuminuria: AUC 0.93 vs 0.67
for urinary albumin
Titan et al [118] N = 56 Prospective Macroalbuminuria Urinary RBP and serum Urinary RBP and MCP-1: Creatinine clearance, proteinuria, BP
T2D and urinary cytokines independently related to the risk of
(TGF-β, MCP-1 and CKD progression
VEGF)
Schlatzer et al N = 465 Nested case–control CKD 1 Panel of 252 urine A panel including Tamm–Horsfall Age, diabetes duration, HbA1c,
[83] T1D Normoalbuminuria peptides protein, progranulin, clusterin, and BMI, WHR, smoking, total and
α-1 acid glycoprotein improved the HDL-cholesterol, SBP, ACR, uric
AUC from 0.841 (clinical variables) acid, cystatin C, BP/lipid
to 0.889 treatment
Metabolomics
Niewczas et al N = 158 Prospective Proteinuria and CKD 3 Global serum 7 modified metabolites were associated Baseline HbA1c, ACR, eGFR, BP,
[119] T1D metabolomic profiling with renal function decline and time BMI, smoking, uric acid levels,
to ESRD RASi use, other antihypertensive
treatment, and statins
Klein et al [120] N = 497 Prospective Normoalbuminuria Multiple plasma Increased plasma levels of very long Treatment group, baseline
T1D ceramide species and chain ceramide species associated retinopathy, sex, HbA1c, age,
individual sphingoid with reduced macroalbuminuria risk AER, lipid levels, diabetes
bases and their duration, ACEi/ARB use
phosphates
Pena et al [121] N = 90 Case–control and Normoalbuminuria and Plasma and urinary Urine hexose, glutamine and tyrosine Albuminuria, eGFR, RASi use
T2D prospective macroalbuminuria metabolomics and plasma histidine and
butenoylcarnitine associated with
progression from micro- to
macroalbuminuria
Niewczas et al N = 80 Prospective CKD 1-3 78 plasma metabolites Abnormal levels of uremic solutes and Albuminuria, eGFR, HbA1c
[122] T2D nested case–control (uremic solutes) and essential amino acids associated with
study essential amino acids progression to ESRD
1003
Table 1 (continued)
1004

Author, ref. Sample size and Study design DKD stage Biomarkers Main results Adjustments
population

Sharma et al N = 181 (n = 114 Cross-sectional Different CKD stages 13 urine metabolites of Differences in urine metabolome Age, race, sex, MAP,BMI, HbA1c,
[123] T2D, n = 44 T1D, mitochondrial between healthy controls and diabetes diabetes duration
n = 23 control) metabolism mellitus and CKD cohorts
Hirayama et al N = 78 Cross-sectional Varying levels of albumin 19 serum metabolites Able to discriminate presence or No adjustments
[124] T2D excretion absence of diabetic nephropathy
Van der Kloet N = 52 Prospective Normoalbuminuria Metabolite profiles of Acylcarnitines, acylglycines and No adjustments
et al [125] T1D 24 h urines metabolites related to tryptophan
metabolism were discriminating
metabolites for progression to micro
or macroalbuminuria
Ng et al [126] N = 90 Cross-sectional Varying levels of eGFR Octanol, oxalic acid, Able to discriminate low vs normal Age at diagnosis, age at
T2D phosphoric acid, eGFR examination, baseline serum
benzamide, creatinine, creatinine
3,5-dimethoxymandelic
amide and
N-acetylglutamine
Han et al [127] N = 150 (n = 120 Cross-sectional Varying levels of albumin 35 plasma non-esterified Able to discriminate albuminuria status No adjustments
T2D, n = 30 excretion and 32 esterified fatty
controls) acids

8-iso-PGF2α, 8-iso-prostaglandin F2α; ACEi, ACE inhibitors; ACR, albumin-creatinine ratio; Apo, apolipoprotein; ARB, angiotensin receptor blockers; B2M; β2-microglobulin; C1QB, complement C1q
subcomponent subunit B; CD5L, CD5 antigen-like; CCL2, chemokine ligand 2; CKD, chronic kidney disease; CRP, C-reactive protein; CTGF, connective tissue growth factor; CVD, cardiovascular
disease; CXCL10, CXC chemokine ligand-10; DKD, diabetic kidney disease; ESRD, end-stage renal disease; FGF, fibroblast growth factor; GPNMB, glycoprotein non-metastatic melanoma protein B;
GH, growth hormone; HGF, hepatocyte growth factor; IGFBP3, insulin-like growth factor binding protein 3; ICAM-1, intercellular adhesion molecule-1; IP-10, inducible protein 10; L-FABP, liver-type
fatty acid-binding protein; MAP, mean arterial blood pressure; MAPK, mitogen-activated protein kinases; MCP-1, monocyte chemoattractant protein-1; MMP, matrix metalloproteinase; MR-proADM,
mid-regional pro-adrenomedullin; mTOR, mechanistic target of rapamycin; NAG, N-acetylglucosamine; NGAL, neutrophil gelatinase-associated lipocalin; NT-proBNP, N-terminal pro-B-type natriuretic
peptide; NT-proCNP, N-terminal pro-C-type natriuretic peptide; P13K-Akt, phosphatidylinositol-3-kinase and protein kinase B; PAI-1, plasminogen activator inhibitor-1; PDGF-AA, platelet-derived
growth factor-AA; RANTES, regulated on activation, normal T cell expressed and secreted; RASi, renin–angiotensin system inhibitor; RBP, retinol binding protein; SBP, systolic BP; sTNFR1, soluble
TNF receptor-1; T1D, type 1 diabetes; T2D, type 2 diabetes; TNFR, TNF receptor; TRPC6, transient receptor potential cation channel subfamily member 6; UACR, urine albumin-to-creatinine ratio;
VCAM-1, vascular cell adhesion molecule 1; VEGF, vascular endothelial growth factor; YKL-40, chitinase-3-like protein 1; ZAG, zinc α2-glycoprotein
Diabetologia (2018) 61:996–1011
Diabetologia (2018) 61:996–1011 1005

Glomerulus
Cystatin C Other such approaches are detailed in Table 1. Of particular
α1-Microglobulin note, the Systems biology towards novel chronic kidney dis-
β2-Microglobulin
Albumin
ease diagnosis and treatment (SYSKID) consortium used data
mining and de novo omics profiling to construct a molecular
process model representation of CKD in diabetes [64], choos-
Inflammation/endothelial
ing ultimately to measure 13 candidates that represented the
damage/fibrosis
α1-Antitrypsin
Distal tubule four largest processes of the model [65]. The panel that gave
Osteopontin
TGF-β1
NGAL
an increase in prediction of renal disease progression was then
MCP-1
VEGF
Copeptin reported (C statistic increased from 0.835 to 0.896). In a recent
MR-proADM validation study of nine of the biomarkers, the investigators
Proximal tubule
NT-proBNP
KIM-1
reported that the panel was useful in prediction based on an
TNFR1, TNFR2
SDMA/ADMA NGAL increase in the adjusted r2 for the prediction model for eGFR
FGF21 L-FABP progression from 29% and 56% for those with a baseline
eGFR above and below 60 ml min 1.73 m−2, respectively, to
CKD273 FGF23
CD5L
Collecting duct 35% and 64%, respectively, for the biomarker panel on top of
MMPs
Loop of Henle Copeptin
Endostatin clinical variables [66].
Osteopontin
Uromodulin In a study exploring 17 candidate urinary and seven plasma
(Tamm–Horsfall protein) biomarkers in 67 participants with type 2 diabetes, Agarwal et al
Fig. 1 Presumed site of origin of commonly associated biomarkers pre- [67] found that urinary C-terminal FGF-2 showed the strongest
dictive of DKD. MMPs, matrix metalloproteases. This figure is available association with ESRD, whereas plasma vascular endothelial
as part of a downloadable slideset
growth factor (VEGF) was associated with the composite out-
come of death and ESRD. The analysis was adjusted for baseline
and adjusted for historical and baseline eGFR, albuminuria and
eGFR only and ACR. Of a panel of seven candidates, Verhave
other covariates. In forward selection, 14 biomarkers were se-
et al found that urinary monocyte chemoattractant protein-1
lected adjusting for this basic set of covariates (Table 1). On top
(MCP-1) and TGF-β1 predicted renal function decline indepen-
of a more extensive set of covariates, seven biomarkers were
dently of albuminuria. Adjustment for baseline eGFR was not
selected: KIM-1, symmetric dimethylarginine/asymmetric
made as it surprisingly did not predict decline in univariate test-
dimethylarginine (SDMA/ADMA) ratio, β2-microglobulin
ing [68]. In the Coronary Artery Calcification in Type 1 Diabetes
(B2M), α1-antitrypsin, C16-acylcarnitine, FGF-21 and uracil.
(CACTI) study using Kidney Injury Panels 3 and 5, (Meso Scale
Diagnostics, www.mesoscale.com/en/products/kidney-injury-
panel-3-human-kit-k15189d/ accessed 08 January 2018)
containing seven biomarkers, component 2 of a principal
component analysis containing B2M, cystatin C, NGAL and
osteopontin predicted incident impaired eGFR [69]. Recently,
of eight candidate biomarkers studied after adjustment for
clinical predictors, apolipoprotein A4 (ApoA4), CD5 antigen-
like (CD5L), and complement C1q subcomponent subunit B
(C1QB) independently predicted rapid decline in eGFR in 345
people with type 2 diabetes. A notable feature of this study was
the adjustment for extensive clinical covariates [70].
Thus, there is some, but not complete, overlap in the ex-
plored and selected biomarkers in these panel studies so that
further optimisation of a panel of the best reported biomarkers
could be considered, especially if it focused on including bio-
markers with low correlation with each other. It is also the case
that all of the studies, including our own, are too small and there
is a need for a large-scale collaboration to increase power,
Fig. 2 Correlation matrix of biomarker measures in the SUMMIT project quantify prediction and to demonstrate generalisability [25].
(www.imi-summit.eu/) showing there is high correlation between
biomarkers that are of interest because of different pathway Discovery ‘omic’ approaches
involvement. ADM, adrenomedullin; FABP, fatty acid-binding protein;
LAP TGF-β1, latency-associated-peptide; OPN, osteopontin; THP,
Tamm–Horsfall urinary protein; VWF, von Willebrand factor. This Apart from candidate biomarkers on multiplexed panels, glob-
figure is available as part of a downloadable slideset al discovery or ‘hypothesis-free’ approaches measuring large
1006 Diabetologia (2018) 61:996–1011

sets of lipids, metabolites and amino acids, peptides and pro- Other proteomics A nested case–control plasma proteomics
teins are increasingly used [71]. The assay methods have most study yielded kininogen and kininogen fragments as predic-
commonly used mass spectrometry-based approaches, but tors of renal function decline. No adjustment was made for
other proteomic methods are now also used [72, 73]. Here baseline eGFR but stratum matching was used [82]. Using a
we describe some of the main ‘omic’ studies, focusing on mass spectrometry approach on 252 urine peptides followed
whether associations are prospective and whether they have by ELISA validation in a nested case–control design, a panel
adjusted for baseline eGFR and other relevant covariates. including Tamm–Horsfall protein (also known as
uromodulin), progranulin, clusterin and α-1 acid glycoprotein
CKD273 This mass spectrometry-based method combines data improved prediction of early decline in eGFR in a cohort of
on 273 urinary peptides into a score that has high accuracy in 465 adults with type 1 diabetes, but no adjustment was made
the cross-sectional classification of eGFR status [74] and has for baseline eGFR [83]. In another urinary proteomics study
been developed as a commercial test by Mosaique Diagnostics with a very small initial discovery step and then single bio-
(http://mosaiques-diagnostics.de/mosaiques-diagnostics/, marker validation in 204 participants, haptoglobin emerged to
accessed 18 October 2017). Most (74%) of the peptides are be the best predictor of early renal functional decline but no
collagen fragments, with polymeric-immunoglobulin recep- adjustment for baseline eGFR was made [84].
tor, uromodulin (Tamm–Horsfall protein), clusterin, CD99 an-
tigen, albumin, B2M, α1-antitrypsin and others comprising Metabolomics Several studies have also assessed the poten-
the remainder. The collagens, polymeric-immunoglobulin re- tial of metabolomics in the context of DKD. A recent sys-
ceptor, clusterin, CD99 antigen and uromodulin were lower tematic review [85] considered 12 studies (although all
with worse renal function, whereas the others were higher. included control groups, most were cross-sectional), where
CKD273 was cross-sectionally associated with having al- a metabolomics-based approach was applied to identify
buminuria or/and eGFR <45 ml min−1 1.73 m−2 in individuals potential biomarkers of DKD. The main metabolites were
with type 2 diabetes [75]. In a small study (n = 35) of people products of lipid metabolism (such as esterified and non-
with type 1 and type 2 diabetes the CKD273 score improved esterified fatty acids, carnitines, phospholipids), branch-
the C statistic for progression to albuminuria to 0.93 compared chain amino acid and aromatic amino acid metabolism, carni-
with 0.67 when using AER, but these data were not fully tine and tryptophan metabolism, nucleotide metabolism
adjusted for baseline eGFR [76]. In 2672 participants from (purine, pyrimidine), the tricarboxylic acid cycle or uraemic
nine different cohorts, 76.3% with diabetes, CKD273 predict- solutes. The meta-analysis highlighted differences in the results
ed rapid progression of eGFR better than AER [77]. In a from studies included and this might be related to differences in
nested case–control analysis, Roscioni et al reported a signif- study population, sample selection, analytical platform.
icant but smaller increase in C statistic for albuminuria inci- In the SUMMIT study we used mass spectrometry to mea-
dence that was robust to adjustment for eGFR [78]. The most sure low-molecular-weight metabolites, peptide and proteins
convincing data to date on the utility of CKD273 come from a (144 in all) as well as 63 proteins by ELISA and Luminex in a
subset of 737 samples obtained at baseline in the Diabetic prospective design. Adjusted for extensive covariates, the ar-
Retinopathy Candesartan Trials (DIRECT)-Protect 2. The ginine methylated derivatives of protein turnover ADMA and
CKD273 score was strongly associated with incident SDMA, and more strongly their ratio, were independently
microalbuminuria independently of baseline AER, eGFR predictive of rapid progression of eGFR. This ratio, along with
and other variables. In this study, higher baseline eGFR was metabolites uracil, α1-antitrypsin and C-16 acylcarnitine,
associated with incident microalbuminuria, an unusual find- were included in the final panel of seven biomarkers [25].
ing, and CKD273 did not show the expected cross-sectional In summary, there are too many global discovery studies in
association with baseline eGFR [79]. Higher CKD273 score at which prediction has not been properly assessed on top of
baseline was associated with a larger reduction in ACR in the available clinical data, such that replication of findings with
spironolactone group vs placebo (p = 0.026 for interaction) proper adjustments is warranted.
[80]. However, after adjustment for baseline ACR, the inter-
action between treatment and CKD273 was not statistically Genetic biomarkers Detailed reviews of the literature on genet-
significant (p = 0.12). The concept that CKD273 will be use- ic biomarkers of DKD have been recently published and are not
ful in determining risk of disease progression and may also the focus of this review [86]. In brief, a review of genetic
stratify treatment response to spironolactone is being more discovery for DKD concluded that “the search for specific var-
definitively tested in the ongoing Proteomic Prediction and iants that confer predisposition to DKD has been relatively
Renin Angiotensin Aldosterone System Inhibition unrewarding” [86]. The effect sizes of the reported loci are very
Prevention Of Early Diabetic nephRopathy In TYpe 2 small in type 1 [87] and type 2 diabetes [88]. While interna-
Diabetic Patients With Normoalbuminuria (PRIORITY) trial, tional meta-analysis of data from the SUMMIT and other con-
of 3280 participants with type 2 diabetes [81]. sortia are underway, given the effect sizes, it seems very
Diabetologia (2018) 61:996–1011 1007

unlikely that genetic risk scores for DKD will contribute use- been clearly shown to substantially increase prediction of
fully as biomarkers for use in the clinical prediction of DKD, DKD-related phenotypes beyond known predictors. Few
even if they may reveal useful insights into pathogenesis. studies have attempted to estimate the marginal improvement
in prediction beyond historical eGFR readings that can be
MicroRNAs (miRNAs) MiRNAs are small non-coding RNA, expressed as the within-person slope or weighted average past
that block protein translation and can induce messenger RNA eGFR, as we did in the SUMMIT study [25]. This is an im-
degradation, thereby acting as regulators of gene expression portant omission given the increasing availability of electronic
[89]. Several studies have assessed urinary and serum healthcare records and potential for applying algorithms to
miRNA in participants with type 1 and type 2 diabetes in rela- such longitudinal clinical data more easily than measuring
tion to different DKD stages [90–97]. These studies are mostly biomarkers. Even where some consistency in findings is ob-
very small [95] and most have reported simply cross-sectional served, the extent of publication bias is unknown. Most im-
associations of urinary miRNAs with albuminuria status [91, portantly, biomarkers other than ACR and eGFR are not being
93–96]. Three studies have used a nested case–control within routinely used to risk stratify individuals into trials or in clin-
prospective cohort design, one of which was in pooled samples ical practice, despite considerable research investment into
[90, 92, 97]. However, there is no overlap in the specific DKD biomarkers in recent years.
miRNAs being reported as being relevant to DKD. Taken al- Large discovery panels have the potential to yield novel
together there is not convincing evidence as yet for a clinically biomarkers, but progress has been hampered by small sample
useful role for miRNAs in the prediction of DKD progression. sizes, inadequate data analysis approaches (including failure
to test the marginal increase beyond established risk factors)
and lack of samples for replication. Futhermore, discovery
Are any novel biomarkers actually being used approaches that yield panels of biomarkers measured on dif-
yet? ferent platforms do not lend themselves to an easily imple-
mented single panel in the clinical setting.
In reality, despite all the attempts to develop novel prognostic If this field is to be advanced, there is a need for a concerted
biomarkers, few current trials use biomarkers other than albu- effort to (1) generate and share data on the correlation between
minuria or eGFR as stratification variables or entry criteria. existing candidate biomarkers and biomarkers generated from
An exception is the PRIORITY trial [81], in which the available discovery platforms; (2) generate replication and vali-
CKD273 panel is being used to risk stratify people into a dation sample and data sets that allow the best panel from avail-
spironolactone vs placebo arm. able data to be defined; (3) harness the predictive information that
Biomarkers as surrogates of drug response is not the focus of exists in clinical records in the era of electronic health record
this review but we note that there are also few trials using surro- data. Future discoveries should then be evaluated for their mar-
gate biomarkers as endpoints. One ongoing trial is using urinary ginal prediction on top of clinical data and validated biomarkers.
proteomic panels as a surrogate outcome measure [98]. Another
study includes urinary NGAL and KIM-1 as secondary outcome Duality of interest HMC’s institution has a patent co-filed for some of the
biomarkers mentioned in this article.
measures [99], and another is using N-acyl-β-D-glucosidase,
B2M and cystatin C [100]. The SYSKID consortium have ar- Contribution statement Both authors were responsible for drafting the
gued that past trials have shown that albuminuria/eGFR are in- article and revising it critically for important intellectual content. Both
sufficient to predict the individual’s response to renoprotective authors approved the version to be published.
treatments in DKD, and that biomarkers more closely
Open Access This article is distributed under the terms of the Creative
representing molecular mechanisms involved in disease progres- Commons Attribution 4.0 International License (http://
sion and being targeted by therapies are needed [64]. Recently, creativecommons.org/licenses/by/4.0/), which permits unrestricted use,
Pena et al found that urinary metabolites previously shown to be distribution, and reproduction in any medium, provided you give appro-
at lower levels in those with DKD than without, decreased in the priate credit to the original author(s) and the source, provide a link to the
Creative Commons license, and indicate if changes were made.
placebo arm of a trial but remained stable in the arm treated with
the endothelin A receptor blocker atrasentan over a short,
12 week trial [101]. Further such studies of changes in bio-
markers over time and in response to treatment are needed. References

1. Livingstone SJ, Levin D, Looker HC et al (2015) Estimated life

Future perspectives expectancy in a Scottish cohort with type 1 diabetes, 2008-2010.
JAMA 313:37–44
2. Livingstone SJ, Looker HC, Hothersall EJ et al (2012) Risk of
In summary, despite the large number of reports in the litera- cardiovascular disease and total mortality in adults with type 1
ture, at present there are few validated biomarkers that have diabetes: Scottish registry linkage study. PLoS Med 9:e1001321
1008 Diabetologia (2018) 61:996–1011

3. Chan GCW, Tang SCW (2016) Diabetic nephropathy: landmark 24. Vergouwe Y, Soedamah-Muthu SS, Zgibor J et al (2010)
clinical trials and tribulations. Nephrol Dial Transplant 31:359–368 Progression to microalbuminuria in type 1 diabetes: development
4. Jones RH, Hayakawa H, Mackay JD, Parsons V, Watkins PJ and validation of a prediction rule. Diabetologia 53:254–262
(1979) Progression of diabetic nephropathy. Lancet 1:1105–1106 25. Looker HC, Colombo M, Hess S et al (2015) Biomarkers of rapid
5. Michels WM, Grootendorst DC, Verduijn M, Elliott EG, Dekker chronic kidney disease progression in type 2 diabetes. Kidney Int
FW, Krediet RT (2010) Performance of the Cockcroft-Gault, 88:888–896
MDRD, and new CKD-EPI formulas in relation to GFR, age, 26. Ichimura T, Bonventre JV, Bailly V et al (1998) Kidney injury
and body size. Clin J Am Soc Nephrol 5:1003–1009 molecule-1 (KIM-1), a putative epithelial cell adhesion molecule
6. Stevens LA, Coresh J, Schmid CH et al (2008) Estimating GFR containing a novel immunoglobulin domain, is up-regulated in
using serum cystatin C alone and in combination with serum cre- renal cells after injury. J Biol Chem 273:4135–4142
atinine: a pooled analysis of 3,418 individuals with CKD. Am J 27. Mishra J, Ma Q, Prada A et al (2003) Identification of neutrophil
Kidney Dis 51:395–406 gelatinase-associated lipocalin as a novel early urinary biomarker
7. Barr EL, Maple-Brown LJ, Barzi F et al (2017) Comparison of for ischemic renal injury. J Am Soc Nephrol 14:2534–2543
creatinine and cystatin C based eGFR in the estimation of glomer- 28. Pavkov ME, Weil EJ, Fufaa GD et al (2016) Tumor necrosis factor
ular filtration rate in indigenous Australians: the eGFR Study. Clin receptors 1 and 2 are associated with early glomerular lesions in
Biochem 50:301–308 type 2 diabetes. Kidney Int 89:226–234
8. Menon V, Shlipak MG, Wang X et al (2007) Cystatin C as a risk 29. Niewczas MA, Gohda T, Skupien J et al (2012) Circulating TNF
factor for outcomes in chronic kidney disease. Ann Intern Med receptors 1 and 2 predict ESRD in type 2 diabetes. J Am Soc
147:19–27 Nephrol 23:507–515
9. Krolewski AS, Warram JH, Forsblom C et al (2012) Serum con- 30. Gohda T, Niewczas MA, Ficociello LH et al (2012) Circulating
centration of cystatin C and risk of end-stage renal disease in TNF receptors 1 and 2 predict stage 3 CKD in type 1 diabetes. J
diabetes. Diabetes Care 35:2311–2316 Am Soc Nephrol 23:516–524
10. Stevens LA, Schmid CH, Greene T et al (2009) Factors other than 31. Pavkov ME, Nelson RG, Knowler WC, Cheng Y, Krolewski AS,
glomerular filtration rate affect serum cystatin C levels. Kidney Int Niewczas MA (2015) Elevation of circulating TNF receptors 1
75:652–660 and 2 increases the risk of end-stage renal disease in American
11. Macisaac RJ, Jerums G (2011) Diabetic kidney disease with and Indians with type 2 diabetes. Kidney Int 87:812–819
without albuminuria. Curr Opin Nephrol Hypertens 20:246–257 32. Yamanouchi M, Skupien J, Niewczas MA et al (2017) Improved
clinical trial enrollment criterion to identify patients with diabetes
12. Retnakaran R, Cull CA, Thorne KI, Adler AI, Holman RR (2006)
at risk of end-stage renal disease. Kidney Int 92:258–266
Risk factors for renal dysfunction in type 2 diabetes: U.K.
33. Lopes-Virella MF, Baker NL, Hunt KJ, Cleary PA, Klein R,
Prospective Diabetes Study 74. Diabetes 55:1832–1839
Virella G (2013) Baseline markers of inflammation are associated
13. Ekinci EI, Jerums G, Skene A et al (2013) Renal structure in
with progression to macroalbuminuria in type 1 diabetic subjects.
normoalbuminuric and albuminuric patients with type 2 diabetes
Diabetes Care 36:2317–2323
and impaired renal function. Diabetes Care 36:3620–3626
34. Antonellis PJ, Kharitonenkov A, Adams AC (2014) Physiology
14. Krolewski AS (2015) Progressive renal decline: the new paradigm
and Endocrinology Symposium: FGF21: insights into mechanism
of diabetic nephropathy in type 1 diabetes. Diabetes Care 38:954–
of action from preclinical studies. J Anim Sci 92:407–413
962
35. Stein S, Bachmann A, Lossner U et al (2009) Serum levels of the
15. Radcliffe NJ, Seah J-M, Clarke M, MacIsaac RJ, Jerums G, Ekinci adipokine FGF21 depend on renal function. Diabetes Care 32:
EI (2017) Clinical predictive factors in diabetic kidney disease 126–128
progression. J Diabetes Investig 8:6–18 36. Han SH, Choi SH, Cho BJ et al (2010) Serum fibroblast growth
16. Keane WF, Brenner BM, de Zeeuw D et al (2003) The risk of factor-21 concentration is associated with residual renal function
developing end-stage renal disease in patients with type 2 diabetes and insulin resistance in end-stage renal disease patients receiving
and nephropathy: the RENAAL study. Kidney Int 63:1499–1507 long-term peritoneal dialysis. Metabolism 59:1656–1662
17. Elley CR, Robinson T, Moyes SA et al (2013) Derivation and 37. Jian W-X, Peng W-H, Jin J et al (2012) Association between
validation of a renal risk score for people with type 2 diabetes. serum fibroblast growth factor 21 and diabetic nephropathy.
Diabetes Care 36:3113–3120 Metabolism 61:853–859
18. Jardine MJ, Hata J, Woodward M et al (2012) Prediction of 38. Fon Tacer K, Bookout AL, Ding X et al (2010) Research resource:
kidney-related outcomes in patients with type 2 diabetes. Am J comprehensive expression atlas of the fibroblast growth factor
Kidney Dis 60:770–778 system in adult mouse. Mol Endocrinol 24:2050–2064
19. Rosolowsky ET, Skupien J, Smiles AM et al (2011) Risk for 39. Kim HW, Lee JE, Cha JJ et al (2013) Fibroblast growth factor 21
ESRD in type 1 diabetes remains high despite renoprotection. J improves insulin resistance and ameliorates renal injury in db/db
Am Soc Nephrol 22:545–553 mice. Endocrinology 154:3366–3376
20. Skupien J, Warram JH, Smiles AM, Stanton RC, Krolewski AS 40. Har RLH, Reich HN, Scholey JW et al (2014) The urinary
(2016) Patterns of estimated glomerular filtration rate decline lead- cytokine/chemokine signature of renal hyperfiltration in adoles-
ing to end-stage renal disease in type 1 diabetes. Diabetes Care 39: cents with type 1 diabetes. PLoS One 9:e111131
2262–2269 41. Cherney DZI, Scholey JW, Daneman D et al (2012) Urinary
21. Skupien J, Warram JH, Smiles AM et al (2012) The early decline markers of renal inflammation in adolescents with type 1 diabetes
in renal function in patients with type 1 diabetes and proteinuria mellitus and normoalbuminuria. Diabet Med 29:1297–1302
predicts the risk of end stage renal disease. Kidney Int 82:589–597 42. Hui E, Yeung C-Y, Lee PCH et al (2014) Elevated circulating
22. Forsblom C, Moran J, Harjutsalo V et al (2014) Added value of pigment epithelium-derived factor predicts the progression of di-
soluble tumor necrosis factor-α receptor 1 as a biomarker of abetic nephropathy in patients with type 2 diabetes. J Clin
ESRD risk in patients with type 1 diabetes. Diabetes Care 37: Endocrinol Metab 99:E2169–E2177
2334–2342 43. Bidadkosh A, Lambooy SPH, Heerspink HJ et al (2017)
23. Andrésdóttir G, Jensen ML, Carstensen B et al (2015) Improved Predictive properties of biomarkers GDF-15, NTproBNP, and
prognosis of diabetic nephropathy in type 1 diabetes. Kidney Int hs-TnT for morbidity and mortality in patients with type 2 diabe-
87:417–426 tes with nephropathy. Diabetes Care 40:784–792
Diabetologia (2018) 61:996–1011 1009

44. Velho G, El Boustany R, Lefèvre G et al (2016) Plasma copeptin, 62. Panduru NM, Sandholm N, Forsblom C et al (2015) Kidney injury
kidney outcomes, ischemic heart disease, and all-cause mortality molecule-1 and the loss of kidney function in diabetic nephropa-
in people with long-standing type 1 diabetes. Diabetes Care 39: thy: a likely causal link in patients with type 1 diabetes. Diabetes
2288–2295 Care 38:1130–1137
45. Velho G, Bouby N, Hadjadj S et al (2013) Plasma copeptin and 63. Heinzel A, Muhlberger I, Fechete R, Mayer B, Perco P (2014)
renal outcomes in patients with type 2 diabetes and albuminuria. Functional molecular units for guiding biomarker panel design.
Diabetes Care 36:3639–3645 Methods Mol Biol 1159:109–133
46. Boertien WE, Riphagen IJ, Drion I et al (2013) Copeptin, a surro- 64. Pena MJ, Heinzel A, Heinze G et al (2015) A panel of novel
gate marker for arginine vasopressin, is associated with declining biomarkers representing different disease pathways improves pre-
glomerular filtration in patients with diabetes mellitus (ZODIAC- diction of renal function decline in type 2 diabetes. PLoS One 10:
33). Diabetologia 56:1680–1688 e0120995
47. Pikkemaat M, Melander O, Bengtsson Boström K (2015) 65. Heinzel A, Mühlberger I, Stelzer G et al (2015) Molecular disease
Association between copeptin and declining glomerular filtration presentation in diabetic nephropathy. Nephrol Dial Transplant
rate in people with newly diagnosed diabetes. The Skaraborg 30(Suppl 4):iv17–iv25
Diabetes Register. J Diabetes Complicat 29:1062–1065 66. Mayer G, Heerspink HJL, Aschauer C et al (2017) Systems
48. Nielsen SE, Reinhard H, Zdunek D et al (2012) Tubular markers biology-derived biomarkers to predict progression of renal func-
are associated with decline in kidney function in proteinuric type 2 tion decline in type 2 diabetes. Diabetes Care 40:391–397
diabetic patients. Diabetes Res Clin Pract 97:71–76 67. Agarwal R, Duffin KL, Laska DA, Voelker JR, Breyer MD,
49. Fu W-J, Li B-L, Wang S-B et al (2012) Changes of the tubular Mitchell PG (2014) A prospective study of multiple protein bio-
markers in type 2 diabetes mellitus with glomerular markers to predict progression in diabetic chronic kidney disease.
hyperfiltration. Diabetes Res Clin Pract 95:105–109 Nephrol Dial Transplant 29:2293–2302
50. Garg V, Kumar M, Mahapatra HS, Chitkara A, Gadpayle AK, 68. Verhave JC, Bouchard J, Goupil R et al (2013) Clinical value of
Sekhar V (2015) Novel urinary biomarkers in pre-diabetic ne- inflammatory urinary biomarkers in overt diabetic nephropathy: a
phropathy. Clin Exp Nephrol 19:895–900 prospective study. Diabetes Res Clin Pract 101:333–340
51. Kamijo-Ikemori A, Sugaya T, Yasuda T et al (2011) Clinical sig- 69. Bjornstad P, Pyle L, Cherney DZI et al (2017) Plasma biomarkers
nificance of urinary liver-type fatty acid-binding protein in diabet- improve prediction of diabetic kidney disease in adults with type 1
ic nephropathy of type 2 diabetic patients. Diabetes Care 34:691– diabetes over a 12-year follow-up: CACTI study. Nephrol Dial
696 Transplant. https://doi.org/10.1093/ndt/gfx255
52. Viswanathan V, Sivakumar S, Sekar V, Umapathy D, Kumpatla S 70. Peters KE, Davis WA, Ito J et al (2017) Identification of novel
(2015) Clinical significance of urinary liver-type fatty acid binding circulating biomarkers predicting rapid decline in renal function in
protein at various stages of nephropathy. Indian J Nephrol 25:269– type 2 diabetes: the Fremantle Diabetes Study Phase II. Diabetes
273 Care 40:1548–1555
53. Araki S, Haneda M, Koya D et al (2013) Predictive effects of 71. Pena MJ, Mischak H, Heerspink HJL (2016) Proteomics for pre-
urinary liver-type fatty acid-binding protein for deteriorating renal diction of disease progression and response to therapy in diabetic
function and incidence of cardiovascular disease in type 2 diabetic kidney disease. Diabetologia 59:1819–1831
patients without advanced nephropathy. Diabetes Care 36:1248– 72. Gold L, Ayers D, Bertino J et al (2010) Aptamer-based
1253 multiplexed proteomic technology for biomarker discovery.
54. Yin C, Wang N (2016) Kidney injury molecule-1 in kidney dis- PLoS One 5:e15004
ease. Ren Fail 38:1567–1573 73. Carlsson AC, Ingelsson E, Sundstrom J et al (2017) Use of prote-
55. Zhao X, Zhang Y, Li L et al (2011) Glomerular expression of omics to investigate kidney function decline over 5 years. Clin J
kidney injury molecule-1 and podocytopenia in diabetic glomer- Am Soc Nephrol 12:1226–1235
ulopathy. Am J Nephrol 34:268–280 74. Argiles A, Siwy J, Duranton F et al (2013) CKD273, a new pro-
56. Alter ML, Kretschmer A, Von Websky K et al (2012) Early uri- teomics classifier assessing CKD and its prognosis. PLoS One 8:
nary and plasma biomarkers for experimental diabetic nephropa- e62837
thy. Clin Lab 58:659–671 75. Siwy J, Schanstra JP, Argiles A et al (2014) Multicentre prospec-
57. Waikar SS, Sabbisetti V, Arnlov J et al (2016) Relationship of tive validation of a urinary peptidome-based classifier for the di-
proximal tubular injury to chronic kidney disease as assessed by agnosis of type 2 diabetic nephropathy. Nephrol Dial Transplant
urinary kidney injury molecule-1 in five cohort studies. Nephrol 29:1563–1570
Dial Transplant 31:1460–1470 76. Zürbig P, Jerums G, Hovind P et al (2012) Urinary proteomics for
58. Sabbisetti VS, Waikar SS, Antoine DJ et al (2014) Blood kidney early diagnosis in diabetic nephropathy. Diabetes 61:3304–3313
injury molecule-1 is a biomarker of acute and chronic kidney 77. Pontillo C, Jacobs L, Staessen JA et al (2017) A urinary proteome-
injury and predicts progression to ESRD in type I diabetes. J based classifier for the early detection of decline in glomerular
Am Soc Nephrol 25:2177–2186 filtration. Nephrol Dial Transplant 32:1510–1516
59. Nielsen SE, Andersen S, Zdunek D, Hess G, Parving H-H, 78. Roscioni SS, de Zeeuw D, Hellemons ME et al (2013) A urinary
Rossing P (2011) Tubular markers do not predict the decline in peptide biomarker set predicts worsening of albuminuria in type 2
glomerular filtration rate in type 1 diabetic patients with overt diabetes mellitus. Diabetologia 56:259–267
nephropathy. Kidney Int 79:1113–1118 79. Lindhardt M, Persson F, Zürbig P et al (2017) Urinary proteomics
60. Conway BR, Manoharan D, Manoharan D et al (2012) Measuring predict onset of microalbuminuria in normoalbuminuric type 2
urinary tubular biomarkers in type 2 diabetes does not add prog- diabetic patients, a sub-study of the DIRECT-Protect 2 study.
nostic value beyond established risk factors. Kidney Int 82:812– Nephrol Dial Transplant 32:1866–1873
818 80. Lindhardt M, Persson F, Oxlund C et al (2018) Predicting albu-
61. Vaidya VS, Niewczas MA, Ficociello LH et al (2011) Regression minuria response to spironolactone treatment with urinary proteo-
of microalbuminuria in type 1 diabetes is associated with lower mics in patients with type 2 diabetes and hypertension. Nephrol
levels of urinary tubular injury biomarkers, kidney injury mole- Dial Transplant. 33:296–303
cule-1, and N-acetyl-β-D-glucosaminidase. Kidney Int 79:464– 81. Lindhardt M, Persson F, Currie G et al (2016) Proteomic predic-
470 tion and renin angiotensin aldosterone system inhibition
1010 Diabetologia (2018) 61:996–1011

prevention of early diabetic nephropathy in type 2 diabetic patients 101. Pena MJ, de Zeeuw D, Andress D et al (2017) The effects of
with normoalbuminuria (PRIORITY): essential study design and atrasentan on urinary metabolites in patients with type 2 diabetes
rationale of a randomised clinical multicentre trial. BMJ Open 6: and nephropathy. Diabetes Obes Metab 19:749–753
e010310 102. Burns KD, Lytvyn Y, Mahmud FH et al (2017) The relationship
82. Merchant ML, Niewczas MA, Ficociello LH et al (2013) Plasma between urinary renin-angiotensin system markers, renal function,
kininogen and kininogen fragments are biomarkers of progressive and blood pressure in adolescents with type 1 diabetes. Am J
renal decline in type 1 diabetes. Kidney Int 83:1177–1184 Physiol Renal Physiol 312:F335–F342
83. Schlatzer D, Maahs DM, Chance MR et al (2012) Novel urinary 103. Carlsson AC, Östgren CJ, Länne T, Larsson A, Nystrom FH,
protein biomarkers predicting the development of Ärnlöv J (2016) The association between endostatin and kidney
microalbuminuria and renal function decline in type 1 diabetes. disease and mortality in patients with type 2 diabetes. Diabetes
Diabetes Care 35:549–555 Metab 42:351–357
84. Bhensdadia NM, Hunt KJ, Lopes-Virella MF et al (2013) Urine 104. Dieter BP, McPherson SM, Afkarian M et al (2016) Serum amy-
haptoglobin levels predict early renal functional decline in patients loid a and risk of death and end-stage renal disease in diabetic
with type 2 diabetes. Kidney Int 83:1136–1143 kidney disease. J Diabetes Complicat 30:1467–1472
85. Zhang Y, Zhang S, Wang G (2015) Metabolomic biomarkers in 105. Wang Y, Li Y-M, Zhang S, Zhao J-Y, Liu C-Y (2016) Adipokine
diabetic kidney diseases—a systematic review. J Diabetes zinc-alpha-2-glycoprotein as a novel urinary biomarker presents
Complicat 29:1345–1351 earlier than microalbuminuria in diabetic nephropathy. J Int Med
86. Ahlqvist E, van Zuydam NR, Groop LC, McCarthy MI (2015) Res 44:278–286
The genetics of diabetic complications. Nat Rev Nephrol 11:277– 106. Fufaa GD, Weil EJ, Nelson RG et al (2015) Association of urinary
287 KIM-1, L-FABP, NAG and NGAL with incident end-stage renal
87. Sandholm N, Salem RM, McKnight AJ et al (2012) New suscep- disease and mortality in American Indians with type 2 diabetes
tibility loci associated with kidney disease in type 1 diabetes. mellitus. Diabetologia 58:188–198
PLoS Genet 8:e1002921 107. Bouvet BR, Paparella CV, Arriaga SMM, Monje AL, Amarilla
88. Teumer A, Tin A, Sorice R et al (2016) Genome-wide association AM, Almará AM (2014) Evaluation of urinary N-acetyl-beta-D-
studies identify genetic loci associated with albuminuria in diabe- glucosaminidase as a marker of early renal damage in patients with
tes. Diabetes 65:803–817 type 2 diabetes mellitus. Arq Bras Endocrinol Metabol 58:798–
89. Simpson K, Wonnacott A, Fraser DJ, Bowen T (2016) 801
MicroRNAs in diabetic nephropathy: from biomarkers to therapy. 108. Petrica L, Vlad A, Gluhovschi G et al (2014) Proximal tubule
Curr Diab Rep 16:35 dysfunction is associated with podocyte damage biomarkers
90. Argyropoulos C, Wang K, Bernardo J et al (2015) Urinary nephrin and vascular endothelial growth factor in type 2 diabetes
MicroRNA profiling predicts the development of mellitus patients: a cross-sectional study. PLoS One 9:e112538
microalbuminuria in patients with type 1 diabetes. J Clin Med 4:
109. Wu C, Wang Q, Lv C et al (2014) The changes of serum sKlotho
1498–1517
and NGAL levels and their correlation in type 2 diabetes mellitus
91. Argyropoulos C, Wang K, McClarty S et al (2013) Urinary
patients with different stages of urinary albumin. Diabetes Res
microRNA profiling in the nephropathy of type 1 diabetes.
Clin Pract 106:343–350
PLoS One 8:e54662
110. do Nascimento JF, Canani LH, Gerchman F et al (2013)
92. Pezzolesi MG, Satake E, McDonnell KP, Major M, Smiles AM,
Messenger RNA levels of podocyte-associated proteins in sub-
Krolewski AS (2015) Circulating TGF-β1-regulated miRNAs and
jects with different degrees of glucose tolerance with or without
the risk of rapid progression to ESRD in type 1 diabetes. Diabetes
nephropathy. BMC Nephrol 14:214
64:3285–3293
93. Peng H, Zhong M, Zhao W et al (2013) Urinary miR-29 correlates 111. Panduru NM, Forsblom C, Saraheimo M et al (2013) Urinary
with albuminuria and carotid intima-media thickness in type 2 liver-type fatty acid-binding protein and progression of diabetic
diabetes patients. PLoS One 8:e82607 nephropathy in type 1 diabetes. Diabetes Care 36:2077–2083
94. Zhou J, Peng R, Li T et al (2013) A potentially functional poly- 112. Lee JE, Gohda T, Walker WH et al (2013) Risk of ESRD and all
morphism in the regulatory region of let-7a-2 is associated with an cause mortality in type 2 diabetes according to circulating levels of
increased risk for diabetic nephropathy. Gene 527:456–461 FGF-23 and TNFR1. PLoS One 8:e58007
95. Delic D, Eisele C, Schmid R et al (2016) Urinary exosomal 113. Jim B, Ghanta M, Qipo A et al (2012) Dysregulated nephrin in
miRNA signature in type II diabetic nephropathy patients. PLoS diabetic nephropathy of type 2 diabetes: a cross sectional study.
One 11:e0150154 PLoS One 7:e36041
96. Eissa S, Matboli M, Aboushahba R, Bekhet MM, Soliman Y 114. Coca SG, Nadkarni GN, Huang Y et al (2017) Plasma biomarkers
(2016) Urinary exosomal microRNA panel unravels novel bio- and kidney function decline in early and established diabetic kid-
markers for diagnosis of type 2 diabetic kidney disease. J ney disease. J Am Soc Nephrol 28:2786–2793
Diabetes Complicat 30:1585–1592 115. Saulnier P-J, Gand E, Velho G et al (2017) Association of circu-
97. Barutta F, Bruno G, Matullo G et al (2017) MicroRNA-126 and lating biomarkers (adrenomedullin, TNFR1, and NT-proBNP)
micro-/macrovascular complications of type 1 diabetes in the with renal function decline in patients with type 2 diabetes: a
EURODIAB Prospective Complications Study. Acta Diabetol French prospective cohort. Diabetes Care 40:367–374
54:133–139 116. Pena MJ, Jankowski J, Heinze G et al (2015) Plasma proteomics
98. Effects of dapagliflozin treatment on urinary proteomic patterns in classifiers improve risk prediction for renal disease in patients with
patients with type 2 diabetes (DapKid). https://clinicaltrials.gov/ hypertension or type 2 diabetes. J Hypertens 33:2123–2132
ct2/show/NCT02914691. Accessed 17 Oct 2017 117. Foster MC, Inker LA, Hsu C-Y et al (2015) Filtration markers as
99. Renoprotective effects of dapagliflozin in type 2 diabetes (RED) predictors of ESRD and mortality in Southwestern American
https://clinicaltrials.gov/ct2/show/NCT02682563. Accessed 17 Indians with type 2 diabetes. Am J Kidney Dis 66:75–83
Oct 2017 118. Titan SM, Vieira JM, Dominguez WVet al (2012) Urinary MCP-1
100. Ivabradine to treat microalbuminuria in patients with type 2 dia- and RBP: independent predictors of renal outcome in
betes and coronary heart disease (BENCH) https://clinicaltrials. macroalbuminuric diabetic nephropathy. J Diabetes Complicat
gov/ct2/show/NCT03105219. Accessed 17 Oct 2017 26:546–553
Diabetologia (2018) 61:996–1011 1011

119. Niewczas MA, Mathew AV, Croall S et al (2017) Circulating 124. Hirayama A, Nakashima E, Sugimoto M et al (2012) Metabolic
modified metabolites and a risk of ESRD in patients with type 1 profiling reveals new serum biomarkers for differentiating diabetic
diabetes and chronic kidney disease. Diabetes Care 40:383–390 nephropathy. Anal Bioanal Chem 404:3101–3109
120. Klein RL, Hammad SM, Baker NL et al (2014) Decreased plasma 125. van der Kloet FM, Tempels FWA, Ismail N et al (2012) Discovery
levels of select very long chain ceramide species are associated of early-stage biomarkers for diabetic kidney disease using ms-
with the development of nephropathy in type 1 diabetes. based metabolomics (FinnDiane study). Metabolomics 8:109–119
Metabolism 63:1287–1295 126. Ng DPK, Salim A, Liu Y et al (2012) A metabolomic study of low
121. Pena MJ, Lambers Heerspink HJ, Hellemons ME et al (2014) estimated GFR in non-proteinuric type 2 diabetes mellitus.
Urine and plasma metabolites predict the development of diabetic Diabetologia 55:499–508
nephropathy in individuals with type 2 diabetes mellitus. Diabet 127. Han L-D, Xia J-F, Liang Q-L et al (2011) Plasma esterified and
Med 31:1138–1147 non-esterified fatty acids metabolic profiling using gas
122. Niewczas MA, Sirich TL, Mathew AV et al (2014) Uremic solutes chromatography-mass spectrometry and its application in the
and risk of end-stage renal disease in type 2 diabetes: metabolomic study of diabetic mellitus and diabetic nephropathy. Anal Chim
study. Kidney Int 85:1214–1224 Acta 689:85–91
123. Sharma K, Karl B, Mathew AV et al (2013) Metabolomics reveals
signature of mitochondrial dysfunction in diabetic kidney disease.
J Am Soc Nephrol 24:1901–1912