Introduction
Biomarkers currently in use
Clinical predictors of DKD in type 1 and type 2 diabetes
Novel biomarker studies
Author, ref. | Sample size and population | Study design | DKD stage | Biomarkers | Main results | Adjustments |
---|---|---|---|---|---|---|
Single biomarkers or several biomarkers not as a panel | ||||||
Burns et al [102] | N = 259 (n = 194 T1D, n = 65 controls) | Cross-sectional | Normoalbuminuria; varying levels of GFR | Urinary angiotensinogen and ACE2 levels, activity of ACE and ACE2 | Urinary angiotensinogen and ACE activity associated with ACR | No adjustments |
Velho et al [44] | N = 986 T1D | Prospective | Varying levels of albumin excretion and GFR | Plasma copeptin | Upper tertiles of copeptin associated with a higher incidence of ESRD | Baseline sex, age, and duration of diabetes |
Carlsson et al [103] | N = 607 T2D | Prospective | Varying levels of albumin excretion | Plasma endostatin | Endostatin levels associated with increased risk of GFR decline and mortality | Baseline age, sex, eGFR and ACR |
Dieter et al [104] | N = 135 T2D | Prospective | Proteinuria | Serum amyloid A | Higher serum amyloid A levels predicted higher risk of death and ESRD | UACR, eGFR, age, sex and ethnicity |
Wang et al [105] | N = 100 (n = 80 with T2D, n = 20 healthy controls) | Cross-sectional | Varying levels of eGFR and ACR | Serum and urinary ZAG | Serum and urinary ZAG associated with eGFR and UACR, respectively | No adjustments |
Pikkemaat et al [47] | N = 161 T2D | Prospective | eGFR >60 ml min−1 1.73 m−2 | Copeptin | Copeptin predicted development of CKD stage 3, borderline significant on adjustment for baseline eGFR | Age, sex, diabetes duration, antihypertensive treatment, HbA1c, BMI, SBP |
Garg et al [50] | N = 91 T2D (including n = 30 with prediabetes) | Cross-sectional | Varying levels of albumin excretion | Urinary NGAL and cystatin C | NGAL and cystatin C were significantly higher in participants with vs those without microalbuminuria | No adjustments |
Viswanathan et al [52] | N = 78 (n = 65 T2D, n = 13 controls) | Cross-sectional | Varying degrees of albuminuria | Urinary L-FABP | L-FABP inversely associated with eGFR and positively associated with protein to creatinine ratio | No adjustments |
Panduru et al [62] | N = 1573 T1D | Prospective + Mendelian randomisation | Varying degrees of albuminuria | Urinary KIM-1 | KIM-1 did not predict progression to ESRD independently of AER Mendelian randomisation supported a causal link between KIM-1 and eGFR | HbA1c, triacylglycerols, AER |
Pavkov et al [31] | N = 193 T2D | Prospective | Varying levels of albumin excretion, eGFR: ≥60 ml/min in 89% participants | Serum TNFR1 and TNFR2 | Elevated concentrations of TNFR1 or TNFR2 associated with increased risk of ESRD | Age, sex, HbA1c, MAP, ACR and GFR |
Fufaa et al [106] | N = 260 T2D | Prospective | Varying levels of albumin excretion and eGFR | Urinary KIM-1, L-FABP, NAG and NGAL | NGAL and L-FABP independently associated with ESRD and mortality | Baseline age, sex, diabetes duration, hypertension, HbA1c, GFR, ACR |
Bouvet et al [107] | N = 36 T2D | Cross-sectional | Normoalbuminuria and macroalbuminuria | Urinary NAG | Higher NAG levels associated with microalbuminuria | No adjustments |
Har et al [40] | N = 142 T1D | Cross-sectional | Varying levels of eGFR Normoalbuminuria | Urinary cytokines/chemokines | Increased urinary cytokine/chemokine excretion according to filtration status with highest levels in hyperfiltering individuals, although not significant after adjustments | Glycaemia |
Petrica et al [108] | N = 91 (n = 70 T2D, n = 21 controls) | Cross-sectional | Normoalbuminuria and microalbuminuria | Urinary α1-microglobulin and KIM-1 (proximal tubule markers), nephrin and VEGF (podocyte markers), AGE, UACR and serum cystatin C | Significant association between biomarkers of proximal tubule dysfunction and podocyte biomarkers (independently of albuminuria and renal function) | UACR, cystatin C, CRP |
Wu et al [109] | N = 462 T2D | Cross-sectional | Varying levels of albumin excretion | Serum Klotho, NGAL, 8-iso-PGF2α, MCP-1, TNF-α, TGF-β1 | Klotho and NGAL associated with ACR | No adjustments |
Sabbisetti et al [58] | N = 124 T1D | Prospective | Proteinuria CKD 1-5 | Serum KIM-1 | KIM-1 associated with eGFR slopes and progression to ESRD | Baseline ACR, eGFR, and HbA1c |
Velho et al [45] | N = 3101 T2D | Prospective | Albuminuria | Plasma copeptin | Copeptin independently associated with renal events (doubling of creatinine or ESRD) | Baseline sex, age, diabetes duration, hypertension, diuretics use, HbA1c, eGFR, triacylglycerols, HDL-cholesterol, AER |
do Nascimento et al [110] | N = 101 (n = 19 prediabetes, n = 67 diabetes [T1D, T2D] and n = 15 controls) | Cross-sectional | Varying levels of albumin excretion | Urinary mRNA levels of podocyte-associated proteins (nephrin, podocin, podocalyxin, synaptopodin, TRPC6, α-actinin-4 and TGF-β1) | Urinary nephrin discriminated between the different stages of DKD and predicted increases in albuminuria | No adjustments |
Boertien et al [46] | N = 1328 T2D | Prospective | Varying degrees of albuminuria and eGFR | Copeptin | Copeptin associated with change in eGFR independently of baseline eGFR. This association not present in those on RASi | Age, sex, diabetes duration, antihypertensive use, HbA1c, cholesterol, BP,BMI, smoking |
Lopes-Virella et al [33] | N = 1237 T1D | Prospective | Normoalbuminuria | Serum E-selectin, IL-6, PAI-1, sTNFR1, TNFR2 | TNFR1 and TNFR2 and E-selectin best predictors of progression to macroalbuminuria | Treatment allocation, baseline AER, ACEi/ARB use, retinopathy cohort, sex, age, HbA1c, diabetes duration |
Panduru et al [111] | N = 2454 (n = 2246 T1D, n = 208 controls) | Prospective | Varying degrees of albuminuria | Urinary L-FABP | L-FABP was an independent predictor of progression at all stages of DKD, but L-FABP did not significantly improve risk prediction above AER | Baseline WHR, HbA1c, triacylglycerols, ACR |
Araki et al [53] | N = 618 T2D | Prospective | Varying levels of albumin excretion, serum creatinine ≤ 8.8×10−2 mmol/l | Urinary L-FABP | L-FABP associated with decline in eGFR | Age, sex, BMI, HbA1c, cholesterol, triacylglycerols, HDL-cholesterol, hypertension, RASi use, BP |
Lee et al [112] | N = 380 T2D | Prospective | Varying levels of albumin excretion | Plasma TNFR1 and FGF-23 | FGF-23 was associated with increased risk of ESRD, only in unadjusted model | Sex, baseline diabetes duration, HbA1c, eGFR, AER |
Cherney et al [41] | N = 150 T1D | Cross-sectional | Normoalbuminuria | 42 urinary cytokines/chemokines | IL-6, IL-8, PDGF-AA and RANTES levels differed across ACR tertiles | No adjustments |
Conway et al [60] | N = 978 T2D | Prospective | Varying degrees of albuminuria and eGFR | Urinary KIM-1 and GPNMB | KIM-1 and GPNMB associated with faster eGFR decline, only in unadjusted models Higher KIM-1 associated with mortality risk, only in unadjusted models | Baseline eGFR, ACR, sex, diabetes duration, HbA1c, BP |
Nielsen et al [48] | N = 177 T2D | Prospective | Proteinuria | Urinary NGAL and KIM1 and plasma FGF23 | Higher levels of the biomarkers associated with a faster decline in eGFR, although this was not independent of known promoters | Age, sex, HbA1c, SBP and urinary albumin |
Jim et al [113] | N = 76 (n = 66 T2D, n = 10 controls) | Cross-sectional | Normoalbuminuria and microalbuminuria | Urinary nephrin levels | Nephrinuria occurred before the onset of microalbuminuria | No adjustments |
Gohda et al [30] | N = 628 T1D | Prospective | Normal renal function; normoalbuminuria and microalbuminuria | TNFR1 and TNFR2 | TNFR1 and TNFR2 strongly associated with risk for early renal decline | HbA1c, AER, and eGFR |
Niewczas et al [29] | N = 410 T2D | Prospective | CKD 1-3 | Plasma TNF-α, TNFR1, and TNFR2, ICAM-1, VCAM-1, PAI-1, IL-6 and CRP | TNFR1 and TNFR2 were strongly associated with risk of ESRD | Age, HbA1c, AER, and eGFR |
Fu et al [49] | N = 112 (n = 88 with T2D, n = 24 controls) | Cross-sectional | Varying degrees of albuminuria | Urinary KIM-1, NAG, NGAL | Higher levels of the three markers in T2D than controls. Positive association of NGAL and NAG with ACR; negative association of NGAL and eGFR | No adjustments |
Nielsen et al [59] | N = 63 T1D | Prospective | Varying levels of albumin excretion and GFR | Urinary NGAL, KIM-1 and L-FABP | Elevated NGAL and KIM-1 were associated with faster decline in GFR, but not after adjustments for known progression promoters | Age, sex, diabetes duration, BP, HbA1c, AER |
Kamijo-Ikemori et al [51] | N = 552 (n = 140 T2D and n = 412 controls) | Cross-sectional and prospective | Varying degrees of albuminuria and GFR | Urinary L-FABP | L-FABP associated with progression of nephropathy | Age, sex, HbA1c, albuminuria status at baseline, BP |
Vaidya et al [61] | N = 697 (n = 659 T1D, n = 38 controls) | Cross-sectional and prospective | Varying levels of albumin excretion | Urinary IL-6, CXCL10/IP-10, NAG and KIM-1 | KIM-1 and NAG both individually and collectively were significantly associated with regression of microalbuminuria | Age, sex, AER, HbA1c, SBP, renoprotective treatment and cholesterol |
Panel of biomarkers /proteomics signatures | ||||||
Coca et al [114] | N = 1536 (n = 1346 T2D, n = 190 controls) | Nested case–control study and prospective | CKD at various stages | TNFR1, TNFR2 and KIM-1 | Higher levels of the three biomarkers associated with higher risk of eGFR decline in persons with early or advanced DKD | Clinical variables |
Bjornstad et al [69] | N = 527 T1D | Prospective | Varying levels of albumin excretion and eGFR | Plasma biomarkers | B2M, cystatin C, NGAL and osteopontin predicted impaired eGFR | Age, sex, HbA1c, SBP, LDL-cholesterol, baseline log ACR and eGFR |
Peters et al [70] | N = 354 T2D | Prospective | Varying levels of albumin excretion and eGFR | Plasma ApoA4, ApoC-III, CD5L, C1QB, complement factor H-related protein 2, IGFBP3 | ApoA4, CD5L, C1QB and IBP3 improved the prediction of rapid decline in renal function independently of recognised clinical risk factors | Age, diabetes duration, diuretic use, HDL-cholesterol |
Mayer et al [66] | N = 1765 T2D | Prospective | CKD at various stages | YKL-40, GH-1, HGF, matrix metalloproteinases: MMP2, MMP7, MMP8, MMP13, tyrosine kinase and TNFR1 | Biomarkers explained variability of annual eGFR loss by 15% and 34% (adj R2) in patients with eGFR ≥60 and <60 ml min−1 1.73 m−2 respectively. A combination of molecular and clinical predictors increased the adjusted R2 to 35% and 64% in these two groups, respectively. | Sex, age, smoking, baseline eGFR, ACR, BMI, total cholesterol, BP and HbA1c |
Saulnier et al [115] | N = 1135 T2D | Prospective | Varying levels of albumin excretion and eGFR | Serum TNFR1, MR-proADM and NT-proBNP | TNFR1, MR-proADM and NT-proBNP improved risk prediction for renal function decline | Age, sex, diabetes duration, HbA1c, BP, baseline eGFR and ACR |
Looker et al [25] | N = 307 (n = 154 T2D, n = 153 controls) | Nested case–control | CKD 3 | 207 serum biomarkers | Panel of 14 biomarkers improved clinical prediction (from 0.706 to 0.868) | Age, sex, eGFR, albuminuria, HbA1c, ACEi and ARB use, BP, weighted average of past eGFRs, diabetes duration, BMI, prior CVD, insulin use, antihypertensive drugs |
Pena et al [116] | N = 82 T2D | Prospective | Normoalbuminuria and macroalbuminuria | Plasma peptides | 18 peptides (related to PI3K-Akt, VEGF, mTOR, MAPK, and p38 MAPK, Wnt signalling) improved risk prediction for transition from micro to macroalbuminuria (C statistic from 0.73 to 0.80) | Baseline albuminuria status, eGFR, RASi use |
Pena et al [64] | N = 82 T2D | Prospective | Varying levels of albumin excretion and eGFR | 28 biomarkers | MMPs, tyrosine kinase, podocin, CTGF, TNFR1, sclerostin, CCL2, YKL-40, and NT-proCNP improved prediction of eGFR decline when combined with established risk markers | Baseline smoking, sex, SBP, eGFR, use of oral diabetic medication |
Foster et al [117] | N = 250 T2D | Prospective | Unselected but 54% albuminuric | β-Trace protein and B2M | β-Trace protein associated with ESRD | GFR, albuminuria, age, sex, diabetes duration, hypertension, cholesterol |
Agarwal et al [67] | N = 87 (n = 67 T2D, n = 20 controls) | Prospective | CKD 2-4 Varying levels of albumin excretion | 17 urinary and 7 plasma biomarkers | Urinary C-terminal FGF-2: strongest association with ESRD Plasma VEGF associated with the composite outcome of death and ESRD | Baseline albuminuria and eGFR |
Siwy et al [75] | N = 165 T2D | Prospective | Wide ranges of eGFR and urinary albumin | Urinary CDK273 | Validation of this urinary proteome-based classifier in a multicentre prospective setting | Albuminuria |
Verhave et al [68] | N = 83 T1D and T2D | Prospective | Overt diabetic nephropathy | Urinary IL-1β, IL-6, IL-8, MCP-1, TNF-α, TGF-β1, and PAI-1 | MCP-1 and TGF-β1 were independent and additive to proteinuria in predicting the rate of renal function decline | Albuminuria |
Bhensdadia et al [84] | N = 204 T2D | Prospective | eGFR stage 1-2 and normo-/macroalbuminuria | Urine peptides | Haptoglobin to creatinine ratio: best predictor of early renal function decline | Albuminuria, ACEi use |
Merchant et al [82] | N = 33 T1D | Prospective | Microalbuminuria | Small (<3 kDa) plasma peptides | Plasma kininogen and kininogen fragments associated with renal function decline | No adjustments but stratum matched for eGFR and albuminuria |
Roscioni et al [78] | N = 88 T2D | Prospective | Normoalbuminuria and microalbuminuria | CKD273 (urine) | Able to detect progression from normo- to micro- and micro- to macroalbuminuria | Baseline albuminuria status, eGFR, RASi use |
Zürbig et al [76] | N = 35 T1D and T2D | Prospective | Normoalbuminuria; normal eGFR | Urinary CKD273 | Early detection of progression to macroalbuminuria: AUC 0.93 vs 0.67 for urinary albumin | Albuminuria |
Titan et al [118] | N = 56 T2D | Prospective | Macroalbuminuria | Urinary RBP and serum and urinary cytokines (TGF-β, MCP-1 and VEGF) | Urinary RBP and MCP-1: independently related to the risk of CKD progression | Creatinine clearance, proteinuria, BP |
Schlatzer et al [83] | N = 465 T1D | Nested case–control | CKD 1 Normoalbuminuria | Panel of 252 urine peptides | A panel including Tamm–Horsfall protein, progranulin, clusterin, and α-1 acid glycoprotein improved the AUC from 0.841 (clinical variables) to 0.889 | Age, diabetes duration, HbA1c, BMI, WHR, smoking, total and HDL-cholesterol, SBP, ACR, uric acid, cystatin C, BP/lipid treatment |
Metabolomics | ||||||
Niewczas et al [119] | N = 158 T1D | Prospective | Proteinuria and CKD 3 | Global serum metabolomic profiling | 7 modified metabolites were associated with renal function decline and time to ESRD | Baseline HbA1c, ACR, eGFR, BP, BMI, smoking, uric acid levels, RASi use, other antihypertensive treatment, and statins |
Klein et al [120] | N = 497 T1D | Prospective | Normoalbuminuria | Multiple plasma ceramide species and individual sphingoid bases and their phosphates | Increased plasma levels of very long chain ceramide species associated with reduced macroalbuminuria risk | Treatment group, baseline retinopathy, sex, HbA1c, age, AER, lipid levels, diabetes duration, ACEi/ARB use |
Pena et al [121] | N = 90 T2D | Case–control and prospective | Normoalbuminuria and macroalbuminuria | Plasma and urinary metabolomics | Urine hexose, glutamine and tyrosine and plasma histidine and butenoylcarnitine associated with progression from micro- to macroalbuminuria | Albuminuria, eGFR, RASi use |
Niewczas et al [122] | N = 80 T2D | Prospective nested case–control study | CKD 1-3 | 78 plasma metabolites (uremic solutes) and essential amino acids | Abnormal levels of uremic solutes and essential amino acids associated with progression to ESRD | Albuminuria, eGFR, HbA1c |
Sharma et al [123] | N = 181 (n = 114 T2D, n = 44 T1D, n = 23 control) | Cross-sectional | Different CKD stages | 13 urine metabolites of mitochondrial metabolism | Differences in urine metabolome between healthy controls and diabetes mellitus and CKD cohorts | Age, race, sex, MAP,BMI, HbA1c, diabetes duration |
Hirayama et al [124] | N = 78 T2D | Cross-sectional | Varying levels of albumin excretion | 19 serum metabolites | Able to discriminate presence or absence of diabetic nephropathy | No adjustments |
Van der Kloet et al [125] | N = 52 T1D | Prospective | Normoalbuminuria | Metabolite profiles of 24 h urines | Acylcarnitines, acylglycines and metabolites related to tryptophan metabolism were discriminating metabolites for progression to micro or macroalbuminuria | No adjustments |
Ng et al [126] | N = 90 T2D | Cross-sectional | Varying levels of eGFR | Octanol, oxalic acid, phosphoric acid, benzamide, creatinine, 3,5-dimethoxymandelic amide and N-acetylglutamine | Able to discriminate low vs normal eGFR | Age at diagnosis, age at examination, baseline serum creatinine |
Han et al [127] | N = 150 (n = 120 T2D, n = 30 controls) | Cross-sectional | Varying levels of albumin excretion | 35 plasma non-esterified and 32 esterified fatty acids | Able to discriminate albuminuria status | No adjustments |