Abstract
Accurate prediction of diabetic kidney disease progression is challenging, but mandatory. Urinary Dickkopf-3 (uDKK3), a tubular, epithelial-derived glycoprotein and marker of tubular injury, is a promising biomarker for kidney function decline. We explored the clinical utility of uDKK3 to predict kidney function decline and adverse cardiovascular events in patients with type 2 diabetes mellitus (T2DM) in a primary health care setting. In this cohort study, 3,232 patients with T2DM were analyzed. The primary end point was a composite of a sustained estimated glomerular filtration rate (eGFR) decline ≥40%; a sustained increase in albuminuria of at least 30%, including a transition in albuminuria class; progression to end-stage kidney disease; and death from kidney failure. After adjustment for confounding variables, uDKK3 values >200 pg/mg creatinine were associated with a higher risk of the composite kidney end point during a median follow-up of 4.26 years. Furthermore, uDKK3 improved the prediction of the 1-year eGFR decline on top of albuminuria. Individuals with high uDKK3 levels also had an increased risk for adverse cardiovascular events and all-cause mortality. uDKK3 identifies patients with T2DM at high risk for kidney function decline on top of established biomarkers (albuminuria and eGFR). In primary care, uDKK3 may help to identify high-risk patients who might benefit from intensified treatment and/or referrals to specialists.
Article Highlights
Prediction of kidney function decline is challenging in patients with type 2 diabetes mellitus (T2DM). Urinary Dickkopf-3 (uDKK3), a profibrotic tubular protein, is a promising biomarker for detecting tubular injury and predicting the progression of chronic kidney disease.
This study assessed whether uDKK3 measurements improve risk prediction in patients with T2DM treated at the primary care level.
Elevated uDKK3 levels were associated with kidney function decline, on top of established biomarkers (estimated glomerular filtration rate and albuminuria). uDKK3 also identified patients at increased risk for cardiovascular events.
uDKK3 may help identify high-risk patients in primary care who could benefit from intensified treatment and/or referrals to specialists.
Graphical Abstract
Introduction
Recently, three major nephrology societies, the International Society of Nephrology, the American Society of Nephrology, and the European Renal Association, published a joint consensus statement on the worldwide impact of kidney diseases, with an estimate of >700 million people experiencing the medical and socioeconomic consequences of chronic kidney disease (CKD) (1). Importantly, this initiative highlighted that many aspects of CKD have so far been underestimated since most people with CKD are not aware of their impaired kidney function. The latter may be because progressive CKD often remains clinically silent for a long time. This may have serious consequences for the patients since the deterioration of kidney function is associated not only with the risk of reaching end-stage kidney disease (ESKD), necessitating kidney replacement therapy, but also with substantial comorbidity, which parallels high health care resource utilization (1,2). Thus, reliable identification of patients with ongoing CKD progression has broad consequences not only for patient well-being but also for saving of health care resources.
Currently, the major global causes of CKD are diabetes and/or hypertension. Particularly in individuals with type 2 diabetes mellitus (T2DM), progressive CKD is a global public health problem accompanied by significant comorbidity, especially cardiovascular disease, and reduced life expectancy (1–3). However, the individual course of diabetic kidney disease (DKD) is challenging to predict, and patterns of progression in persons with T2DM include linear and nonlinear trajectories of glomerular filtration rate (GFR) decline. Moreover, in a substantial portion of patients with T2DM, loss of kidney function occurs in the absence of gross albuminuria (so-called nonproteinuric CKD) (4). Thus, biomarkers are needed that may improve the detection of incipient kidney injury and the prediction of the individual course of CKD in patients with T2DM.
The glycoprotein Dickkopf-3 (DKK3) is a family member of the Dickkopf proteins, which regulate the essential-for-life Wnt/β-catenin signaling pathway (5). The Wnt/β-catenin signaling pathway, besides being crucial in embryogenesis and having an immunomodulatory function in cancer biology, plays an important role in kidney disease (6–8). DKK3 is upregulated by stressed tubular cells and, via the Wnt/β-catenin signaling pathway, promotes interstitial fibrosis and tubular atrophy (6). An upregulation of DKK3 in tubular epithelial cells was also found in patients with diabetes (9). DKK3 is secreted into the urine (urinary DKK3 [uDKK3]) where its concentration correlates with the extent of kidney tissue injury independent of its cause (6). In prospective observational studies in children and adults, the measurement of uDKK3 predicted acute kidney injury (AKI), AKI-CKD transition, and short-term kidney function loss in the setting of progressive CKD (including patients with diabetes) (10–16). In different CKD animal models and in humans with progressive CKD of various etiologies, increased uDKK3 levels correlated with the degree of future decline in estimated GFR (eGFR) (6,13–16). For example, in a Spanish cohort of patients with overt DKD, the measurement of uDKK3 clearly identified patients at high risk for more rapid CKD progression (14). Even more pertinent, in patients with chronic obstructive pulmonary disease without clinical signs of kidney injury (i.e., eGFR >90 mL/min/1.73 m2 and proteinuria <30 mg/g creatinine), the measurement of uDKK3 identified individuals with a significantly higher risk for declining eGFR in the subsequent 18-month period (17). Similar results were obtained in patients with incipient heart failure (i.e., New York Heart Association class I) (18). These data highlight uDKK3 as a tubular marker with potential for the recognition of patients with (inapparent) progressive CKD and adverse kidney outcomes on top of established biomarkers (i.e., albuminuria, eGFR). In this study, we clinically validated uDKK3 in a large multinational cohort of patients with T2DM treated at the primary care level. We further assessed the ability of uDKK3 to predict adverse cardiovascular events and all-cause mortality.
Research Design and Methods
The multinational, noninterventional Prospective Cohort Study in Patients with T2DM for Validation of Biomarkers (PROVALID) study included patients with incident or prevalent T2DM, irrespective of the presence of CKD. T2DM was defined according to the American Diabetes Association or as current treatment with glucose-lowering agents. Between 2011 and 2014, general practitioners and other facilities involved in the primary level of health care recruited 4,000 adults in five European countries (Austria, Hungary, the Netherlands, Scotland, and Poland). In-person follow-up took place annually, and at each visit, fasting morning blood and urine samples for biobanking, clinical, and laboratory data, as well as information on medication and potential study outcomes, were collected. Samples were stored in a central biobank at −80°C. We excluded only patients with active malignant tumors treated with chemotherapy; further study information has been published previously (19). Participants were treated according to local practice, and patient management was not affected by study participation. The final population of the current study comprised 3,232 individuals, as all patients from Poland (n = 768) had to be excluded because of missing follow-up information (Supplementary Fig. 1). The PROVALID study protocol was approved in each participating country by the responsible local institutional review board. Signing an informed consent was a prerequisite for study participation in all countries. This particular substudy was approved by the institutional review board of the Innsbruck Medical University, Innsbruck, Austria (EK no. 1363/2021).
All blood creatinine measurements were traceable to an isotope dilution mass spectrometry. In all participants, eGFR was calculated using the creatinine-based Chronic Kidney Disease Epidemiology Collaboration equation (20). The second morning spot urine was used to determine urinary albumin-to-creatinine ratio (UACR) and DKK3 measurements. CKD was defined according to Kidney Disease: Improving Global Outcomes (KDIGO) criteria. Sex was determined by self-report. DKK3 was measured in the second morning spot urine samples using a commercially available ELISA (DiaRen UG, Homburg, Germany), as described previously (13). The interassay test variability was 4.7% in the lower and 5.1% in the higher detection range. Cross reactivity with other Dickkopf proteins was excluded, and urinary concentrations were normalized to urinary creatinine concentrations to account for urine dilution. The laboratory personnel performing the uDKK3 measurements were blinded with respect to patient’s clinical data. In line with previous studies, we used a cutoff of 200 pg/mg creatinine to separate a low (n = 2,826) from a high (n = 406) uDKK3 group (13,17).
The primary end point was a composite of a sustained decline in the eGFR of at least 40%; a sustained increase in UACR of at least 30%, including a transition in albuminuria class; progression to ESKD requiring replacement therapy; and death from kidney failure. A decline in eGFR and an increase in albuminuria were considered as sustained if they persisted over a least two consecutive follow-ups. The secondary end point was a composite of death from cardiovascular causes, nonfatal myocardial infarction, or nonfatal stroke. All-cause mortality was analyzed as a tertiary end point.
Statistical Analysis
Discrete variables were described using absolute and relative frequencies, while continuous variables were summarized using median and first and third quartiles (interquartile range [IQR]). To test for differences in patient characteristics between groups, Wilcoxon-Mann-Whitney tests were used, and for categorical variables, χ2 tests were performed. The distribution of uDKK3 was described using a histogram on a log scale.
We used box plots to visualize the interrelation of UACR, eGFR, and uDKK3. The relative changes in eGFR for different UACR and uDKK3 groups from baseline to follow-up were described robustly with the median of the percentage change. As a descriptive measure of risk for reaching an end point, the incidence rate per 1,000 person-years with respective asymptotic 95% CIs was used. Longitudinally, the crude risk of reaching an end point was analyzed using Kaplan-Meier survival estimates, and cumulative incidences are displayed. The hypothesis of no differences in risk was tested using log-rank tests. For each end point, the adjusted relative risk was estimated using Cox proportional hazard regressions. For the composite kidney end point, interaction effects of uDKK3 with eGFR, UACR, and CKD were analyzed. All hazard ratios (HRs) were adjusted for age and sex, as well as for eGFR, UACR, and uDKK3. All analyses were complete case analyses.
Estimates are presented with 95% CIs. To evaluate the established threshold of 200 pg/mg creatinine uDKK3, a restricted cubic spline was added to a hazard regression adjusted for age, sex, eGFR, and UACR. As additional sensitivity analyses, Cox regressions adjusted for age, sex, eGFR, UACR, smoking, mean arterial blood pressure, BMI, HbA1c, LDL, coronary artery disease, peripheral artery disease, cerebrovascular artery disease, heart failure, diabetic retinopathy, lipid-lowering drugs, renin-angiotensin system (RAS) inhibitors, glucose-lowering drugs, glucagon-like peptide 1 (GLP-1) receptor agonists, and sodium–glucose cotransporter 2 (SGLT2) inhibitors were performed. Therefore, missing values were imputed using a k-nearest neighbor algorithm. Based on these regressions, respective adjusted cumulative incidences were derived via g-computation. We allowed for a type I error of 5%, with all hypotheses being two-sided. All analyses were performed using R 4.3.0 statistical software (21).
Data and Resource Availability
Data and resources are available from the corresponding author upon request and after approval by the PROVALID steering committee and the institutional review board.
Results
The data of 3,232 patients were analyzed. A total of 2,826 (87%) patients had low (≤200 pg/mg) and 406 (13%) high (>200 pg/mg creatinine) uDKK3 levels. The distribution is displayed in Supplementary Fig. 2. The median follow-up time was 4.26 (IQR 2.91–5.56) years and was slightly longer in patients with low levels of uDKK3 (4.33 [IQR 2.95–5.62] years) compared with patients with high uDKK3 levels (4.01 [IQR 2.39–5.21] years). Table 1 shows the details of the study population at baseline. Twenty percent of the participants had an eGFR <60 mL/min/1.73 m2, 20% had an UACR >30 mg/g creatinine, and 34% fulfilled the KDIGO diagnostic criteria of CKD. In general, patients with uDKK3 levels >200 pg/mg creatinine were more likely to be older, smokers, and men. Their eGFR was lower, UACR and HbA1c were higher, and they had more cardiovascular comorbidities. Accordingly, patients with uDKK3 levels >200 pg/mg creatinine were more often treated with lipid-lowering drugs at baseline. The number of patients taking RAS inhibitors, SGLT2 inhibitors, or GLP-1 receptor agonists were similar between groups. Due to the recruitment period between 2011 and 2014, none of the study participants were taking an SGLT2 inhibitor, and only a few participants were treated with a GLP-1 receptor agonist.
Table 1.
Clinical and laboratory data of the study population at baseline
| Baseline characteristic | n | Overall (n = 3,232) | uDKK3 ≤200 pg/mg creatinine (n = 2,826) | uDKK3 >200 pg/mg creatinine (n = 406) |
|---|---|---|---|---|
| Age (years) | 3,232 | 64 (58–70) | 64 (58–70) | 66 (60–72) |
| Sex (female) | 3,232 | 1,387 (43) | 1,268 (45) | 119 (29) |
| BMI (kg/m2) | 2,897 | 30.4 (27.3–34.0) | 30.4 (27.4–34.0) | 30.4 (27.1–34.1) |
| Mean arterial pressure (mmHg) | 3,182 | 97 (92–104) | 97 (92–104) | 97 (93–105) |
| eGFR (mL/min/1.73 m2) | 3,199 | 79 (64–94) | 80 (66–94) | 70 (53–88) |
| UACR (mg/g creatinine) | 3,145 | 8 (4–23) | 7 (4–19) | 20 (7–101) |
| eGFR | 3,199 | |||
| eGFR ≤60 mL/min/1.73 m2 | 627 (20) | 488 (17) | 139 (35) | |
| eGFR >60 mL/min/1.73 m2 | 2,572 (80) | 2,309 (83) | 263 (65) | |
| UACR | 3,145 | |||
| UACR ≤30 mg/g creatinine | 2,519 (80) | 2,294 (84) | 225 (56) | |
| UACR >30 mg/g creatinine | 626 (20) | 452 (16) | 174 (44) | |
| CKD | 3,155 | |||
| No | 2,090 (66) | 1,925 (70) | 165 (41) | |
| Yes | 1,065 (34) | 831 (30) | 234 (59) | |
| HbA1c (%) | 3,169 | 6.80 (6.30–7.50) | 6.80 (6.30–7.50) | 6.91 (6.30–7.80) |
| HbA1c (mmol/mol) | 3,169 | 51 (45–58) | 51 (45–58) | 52 (45–62) |
| LDL cholesterol (mg/dL) | 2,888 | 94 (73–120) | 95 (73–121) | 92 (70–120) |
| HDL cholesterol (mg/dL) | 3,143 | 46 (39–57) | 46 (39–58) | 44 (39–53) |
| Triglycerides (mg/dL) | 2,864 | 142 (99–204) | 142 (98–201) | 159 (106–221) |
| Total cholesterol (mg/dL) | 3,196 | 174 (147–205) | 174 (147–205) | 166 (147–205) |
| CRP (mg/dL) | 2,183 | 0.30 (0.10–0.70) | 0.30 (0.10–0.67) | 0.30 (0.11–0.70) |
| History of diabetic retinopathy | 3,232 | 413 (13) | 334 (12) | 79 (19) |
| History of coronary artery disease | 3,232 | 625 (19) | 510 (18) | 115 (28) |
| History of peripheral artery disease | 3,232 | 209 (6.5) | 175 (6.2) | 34 (8.4) |
| History of cerebral artery disease | 3,232 | 191 (5.9) | 154 (5.4) | 37 (9.1) |
| History of heart failure | 3,232 | 73 (2.3) | 61 (2.2) | 12 (3.0) |
| RAS inhibitor | 3,232 | 2,166 (67) | 1,890 (67) | 276 (68) |
| SGLT2 inhibitor | 3,232 | 0 (0) | 0 (0) | 0 (0) |
| GLP-1 receptor agonist | 3,232 | 44 (1.4) | 39 (1.4) | 5 (1.2) |
| Lipid-lowering drugs | 3,232 | 2,165 (67) | 1,860 (66) | 305 (75) |
| Glucose-lowering drugs | 3,232 | 2,794 (86) | 2,441 (86) | 353 (87) |
Data are median (IQR) or n (%). CRP, C-reactive protein.
Figure 1 shows the association between UACR and uDKK3 when stratified by eGFR categories G1–G4. Even though uDKK3 levels >200 pg/mg creatinine were not directly associated with the increase in UACR, on a cross-sectional level, UACR, and particularly uDKK3, were higher with increasing eGFR categories. In Table 2 the crude proportions of the composite kidney and cardiovascular end points, as well as all-cause mortality, are presented; they were all almost two times higher in individuals with an elevated uDKK3 concentration. Figure 2 provides data on the crude incidence of the same end points per 1,000 person-years. Additional information on the crude incidence rates per 1,000 patient-years for individual components in the total population and separately by uDKK3 levels are provided in Supplementary Table 1. In line with the 200 pg/mg uDKK3 threshold, the crude incidence of the composite kidney end point was significantly higher in patients in the top two quintiles of positive uDKK3 values than in those with nondetectable uDKK3 expression (uDKK3 = 0) (Supplementary Fig. 3). When adjusted for age, sex, eGFR, and UACR, uDKK3 values >200 pg/mg creatinine were still associated with a significantly higher risk of reaching the composite kidney end point (HR 1.52 [95% CI 1.13–2.04]). Similarly, adding a restricted cubic spline of uDKK3 to the regression showed an elevated hazard above the threshold (Supplementary Fig. 4). The effect on the composite kidney outcome was driven by a higher incidence of eGFR loss and initiation of kidney replacement therapy (Table 3). Death from kidney failure was not significantly different, and high uDKK3 levels did not predict an increase in UACR (Table 3). In addition, individuals with high uDKK3 concentrations had an increased risk for the cardiovascular composite end point (HR 1.52 [95% CI 1.07–2.16]), as well as all-cause mortality (HR 1.44 [95% CI 1.03–1.99]). Results did not significantly change when analyses were further adjusted for smoking, mean arterial blood pressure, BMI, HbA1c, LDL, cardiovascular disease (coronary artery disease, peripheral artery disease, cerebrovascular artery disease, heart failure), diabetic retinopathy, lipid-lowering drugs, RAS inhibitors, glucose-lowering drugs, GLP-1 receptor agonists, and SGLT2 inhibitors (composite kidney end point, HR 1.57 [95% CI 1.16–2.12]; composite cardiovascular end point, HR 1.53 [95% CI 1.08–2.17]; all-cause mortality, HR 1.44 [95% CI 1.04–2.00]). Further details are presented in Supplementary Table 2.
Figure 1.
Distribution of UACR stratified by urinary DKK3 excretion within eGFR categories G1–G4. uDKK3 separates higher UACR levels with more impaired kidney function, i.e., with higher eGFR category. For eGFR values, an open parenthesis indicates that the number is excluded from the range, and a closed square bracket indicates that the number is included. Notches around the median indicate 95% CIs.
Table 2.
Crude proportions of the composite kidney and cardiovascular end points and all-cause mortality in patients with T2DM
| Composite outcome | N | Overall (N = 3,232) | uDKK3 ≤200 pg/mg creatinine (n = 2,826) | uDKK3 >200 pg/mg creatinine (n = 406) | P |
|---|---|---|---|---|---|
| Composite kidney end point | 3,232 | 292 (9.0) | 231 (8.2) | 61 (15) | <0.001 |
| Composite cardiovascular end point | 3,232 | 209 (6.5) | 166 (5.9) | 43 (11) | <0.001 |
| Death from any cause | 3,232 | 229 (7.1) | 179 (6.3) | 50 (12) | <0.001 |
Data are n (%).
Figure 2.
Incidence of the composite kidney and cardiovascular end points and all-cause mortality per 1,000 person-years. Whiskers indicate 95% CIs.
Table 3.
HRs for end points compared with the respective reference group
| End point | Reference group† | Effect† | Adjusted HR | 95% CI | P |
|---|---|---|---|---|---|
| Composite kidney end point* | uDKK3 ≤200 | uDKK3 >200 | 1.52 | 1.13–2.04 | 0.006 |
| Composite cardiovascular end point* | uDKK3 ≤200 | uDKK3 >200 | 1.52 | 1.07–2.16 | 0.018 |
| Sustained increase in albuminuria* | uDKK3 ≤200 | uDKK3 >200 | 1.19 | 0.81–1.74 | 0.372 |
| Sustained decline in eGFR* | uDKK3 ≤200 | uDKK3 >200 | 2.04 | 1.11–3.77 | 0.022 |
| KFRT* | uDKK3 ≤200 | uDKK3 >200 | 3.53 | 1.49–8.37 | 0.004 |
| Death from kidney failure* | uDKK3 ≤200 | uDKK3 >200 | 4.40 | 0.96–20.12 | 0.056 |
| Death from any cause* | uDKK3 ≤200 | uDKK3 >200 | 1.44 | 1.03–1.99 | 0.031 |
| Composite kidney end point | eGFR >60 | eGFR ≤60 | 1.96 | 1.50–2.56 | 0.000 |
| Composite kidney end point | eGFR >60 | eGFR ≤60 and uDKK3 ≤200 | 1.75 | 1.29–2.36 | 0.000 |
| Composite kidney end point | eGFR >60 | eGFR ≤60 and uDKK3 >200 | 4.08 | 2.84–5.86 | 0.000 |
| Composite kidney end point | UACR ≤30 | UACR >30 | 1.62 | 1.25–2.11 | 0.000 |
| Composite kidney end point | UACR ≤30 | UACR >30 and uDKK3 ≤200 | 1.58 | 1.17–2.14 | 0.003 |
| Composite kidney end point | UACR ≤30 | UACR >30 and uDKK3 >200 | 2.99 | 2.08–4.29 | 0.000 |
| Composite kidney end point | CKD no | CKD yes | 2.05 | 1.61–2.59 | 0.000 |
| Composite kidney end point | CKD no | CKD yes and uDKK3 ≤200 | 1.74 | 1.34–2.26 | 0.000 |
| Composite kidney end point | CKD no | CKD yes and uDKK3 >200 | 3.23 | 2.31–4.53 | 0.000 |
All analyses were adjusted for age and sex. KFRT, kidney failure replacement therapy.
*Additionally adjusted for eGFR and UACR.
†Units of measurement for uDKK3 are pg/mg creatinine; eGFR, mL/min/1.73 m2; and UACR, mg/g creatinine.
uDKK3 levels >200 pg/mg creatinine, when used on top of known kidney risk factors and established biomarkers such as eGFR <60 mL/min/1.73 m2 or UACR >30 mg/g creatinine, significantly increased the risk prediction for the composite kidney end point (Table 3). Elevated uDKK3 levels further allowed for risk stratification of patients with CKD (Table 3). Figure 3 displays the respective Kaplan-Meier curves. As a robustness analysis of the Kaplan-Meier estimates, Supplementary Fig. 5 displays confirmatory results with the respective cumulative incidences g-formula adjusted for the full set of the above-mentioned variables.
Figure 3.
Cumulative incidence of the composite kidney end point. Kaplan-Meier curves for uDKK3 levels ≤200 pg/mg creatinine and >200 pg/mg creatinine, alone and on top of established kidney risk factors: eGFR <60 mL/min/1.73 m2, UACR >30 mg/g creatinine, or the presence of CKD according to the KDIGO classification. uDKK3 levels >200 pg/mg creatinine significantly increased the risk for the composite kidney end point compared with the respective reference group (eGFR >60 mL/min/1.73 m2, normoalbuminuria, or no CKD).
We descriptively assessed the relationship between uDKK3 and the course of eGFR within the 1st observational year. In this early study period, uDKK3 levels predicted the 1-year eGFR decline in each albuminuria class, indicating that uDKK3 can improve risk prediction on top of albuminuria. As can be seen in Fig. 4, the median annual loss of eGFR in patients with uDKK3 levels >200 pg/mg creatinine was almost twice as high compared with those with low uDKK3 levels (≤200 pg/mg creatinine) in both patients with moderately and with severely increased UACR.
Figure 4.
Descriptive analysis of the relationship between uDKK3 levels and the decline of eGFR within the 1st observational year, i.e., from baseline (BL) to the first annual follow-up visit (FU1). In patients with moderately or severely increased albuminuria and uDKK3 >200 pg/mg creatinine, eGFR decline was almost twice as high as in patients with moderately or severely increased albuminuria and uDKK3 ≤200 pg/mg creatinine. For UACR values, an open parenthesis indicates that the number is excluded from the range, and a closed square bracket indicates that the number is included.
Discussion
In this well-characterized prospective primary care cohort of patients with T2DM, high uDKK3 levels (>200 pg/mg creatinine) were significantly associated with kidney function decline across different strata of eGFR and albuminuria. The cutoff value of 200 pg/mg creatinine has previously been used in other studies (13,17). Importantly, the consistent association of high uDKK3 levels with the risk of adverse kidney events was independent of baseline eGFR and albuminuria, which to date are the only routinely available prognostic biomarkers. These findings are in line with and validate results from previous studies (13,15–17). Moreover, elevated uDKK3 identified patients at an increased risk of cardiovascular events and all-cause mortality.
In a substantial proportion of patients with DKD (4,22–24), as well as in patients with other causes of kidney injury (15,17), progressive CKD may develop in the absence of higher-grade albuminuria, or significant albuminuria only appears when kidney function already declines. This implies that in certain groups of patients, CKD progression and the evolution of albuminuria are the consequences of two separate pathophysiological pathways that may not necessarily act concomitantly, limiting the predictive value of albuminuria for CKD progression. In the current study, we document that uDKK3 improves prediction of kidney function loss in all strata of UACR, even though this was particularly visible in patients with higher levels of urinary albumin excretion. Thus, our data highlight uDKK3 as a widely applicable, noninvasive, proteinuria-independent marker for loss of kidney function in patients with T2DM. It further supports the concept that biomarkers from different kidney tissue compartments may synergistically improve the prognostication of CKD progression risk, i.e., a glomerular damage marker such as albuminuria in combination with a tubular cell-derived biomarker such as DKK3. In this respect, the measurement of uDKK3 may be particularly helpful for the prediction of short-term (e.g., yearly) kidney function loss as shown in previous studies (13,15–17) and observed again in the current study in patients with T2DM in a primary health care setting. In primary care, uDKK3 might thus be helpful to identify patients at immediate risk for kidney function decline who could benefit from nephroprotective interventions and specialized care.
The Dickkopf protein family represents regulators of the Wnt/β-catenin signaling pathway, which plays a major role in organ development during embryogenesis (5). While Wnt/β-catenin signaling is generally suppressed in postnatal life, it can be reactivated during organ injury, and the pathway has been shown to play a pivotal role in the development of AKI, cystic kidney diseases, and progressive CKD (8). Upon injury, DKK3 is released by tubular cells into the urine and drives the development of interstitial fibrosis, the common pathological finding in progressive CKD (6). DKK3 was found to be upregulated in tubular epithelial cells of patients with diabetes (9). The concept that DKK3 is secreted in a paracrine manner only by stressed/injured tubular cells is corroborated by observations from two independent cohorts of patients with coronavirus disease 2019 (COVID-19) infection (25,26). In the first cross-sectional study, kidney outcome was assessed in 1,328 patients with mild to moderate COVID-19 infection and compared with 443 matched control individuals without the infection (25). CKD or higher-grade albuminuria were equally present in patients after COVID-19 and individuals without COVID-19. Importantly, uDKK3 levels were also not increased in the COVID-19 cohort, indicating no systematic risk for ongoing GFR decline. In the second cohort, 55 patients were hospitalized for severe COVID-19 infection and accompanying AKI, and followed for up to 6 months thereafter (26). Persistently elevated uDKK3 levels in these patients indicated ongoing kidney injury and eGFR decline in the sense of AKI-CKD transition. In contrast, commonly used urinary markers, such as α-1-microglobulin and albuminuria, were less suitable for distinguishing patients with progressive CKD after severe COVID-19 infection (26).
The first hint that DKK3 may also be involved in the development of cystic kidney disease in humans came from a high-throughput single nucleotide polymorphism genotyping association study of 173 biological candidate genes in 794 White patients from 227 families with autosomal dominant polycystic kidney disease (ADPKD) due to a PKD1 mutation (27). The results suggested that genetic variation of DKK3 may modify cystic growth and, thus, the severity of ADPKD. In line with this finding, a recent cross-sectional study revealed that uDKK3 concentrations show a strong correlation with the Mayo classification in patients with ADPKD (28). uDKK3 levels also correlated with kidney function, which indicates that uDKK3 may predict the disproportionate loss of kidney function in this collective. Interestingly, an interaction between copeptin and uDKK3 was found in the prediction models, and the best model containing both variables and their interaction term resulted in a reasonably good explanation of variance in eGFR slope compared with previous models. This also highlights the concept of a multiple biomarker approach for a better estimation of CKD progression.
The observation that elevated uDKK3 levels also identify patients with T2DM at an increased risk of adverse cardiovascular outcomes is not surprising. In a large cohort of patients with various types of CKD (the CARE FOR HOMe [Cardiovascular and Renal Outcome in CKD 2–4 Patients–The Fourth Homburg Evaluation] study [13]), increased uDKK3 levels were significantly associated not only with short-term kidney function loss but also with the risk of atherosclerotic cardiovascular events (29). Taken together, these findings corroborate the known close relationship between progressive CKD and cardiovascular comorbidity and are in line with our results (1–3).
The current study is not without limitations. Analyses were performed in a cohort of patients with T2DM who were mainly of Caucasian ancestry. Therefore, generalizability of the findings to other ethnicities has to be determined. Furthermore, specific prospective interventional trials will be necessary to corroborate the concept of a personalized, uDKK3-guided management and treatment strategy of adult patients with T2DM.
In summary, uDKK3 identifies individuals with T2DM at high risk for kidney function loss and adverse cardiovascular outcomes, in addition to, but also independent of, the measurement of UACR. It thus may represent a novel tool for personalized management in these patients.
This article contains supplementary material online at https://doi.org/10.2337/figshare.29941115.
Article Information
Acknowledgments. The authors thank the study personnel of the PROVALID study and the general practitioners who collaborated with the PROVALID study.
The views and opinions expressed are those of the authors only and do not necessarily reflect those of the European Union or the granting authority. Neither the European Union nor the granting authority can be held responsible for them. The funders played no role in the conduct of the study, collection of data, management of the study, analysis of data, interpretation of data, or preparation of the manuscript.
Duality of Interest. S.D. has received congress support from AstraZeneca and Chiesi and a lecture fee from Boehringer Ingelheim outside of the submitted work. H.J.L.H. has received consulting fees (paid to employer) from AstraZeneca, Alexion, Bayer, Boehringer Ingelheim, CSL Behring, Dimerix, Eli Lilly, Gilead, Janssen, Merck, Novartis, Novo Nordisk, Roche, and Travere Therapeutics; research support for clinical trials from AstraZeneca, Boehringer Ingelheim, Bayer, and Novo Nordisk; and payments for lectures from AstraZeneca, Bayer, Novo Nordisk, and Novartis outside the submitted work. P.B.M. received lecture and/or consulting honoraria from AstraZeneca, Pharmacomsos, Astellas, GSK, Bayer, and Vifor and grants from Boehringer Ingelheim and AstraZeneca outside the submitted work. G.M. is the principal investigator and study coordinator of DC-ren. D.F. is associated with DiaRen UG. No other potential conflicts of interest relevant to this article were reported.
Author Contributions. F.K. planned and conducted the statistical analysis. F.K., S.Sc., S.Sh., S.T., S.E., J.L., H.J.L.H., P.B.M., L.R., A.W., and D.F. provided valuable input and reviewed and edited the manuscript. F.K., S.Sc., G.M., and D.F. conceptualized and planned the study. S.Sc. and D.F. conducted the uDKK3 analyses. S.D., S.Sh., S.T., and J.L. provided substantial contributions to the conception and design of the work and interpretation of the data. S.D. and G.M. wrote the original draft of the manuscript. S.Sh. assisted with the statistical analysis. S.E., H.J.L.H., P.B.M., L.R., A.W., and G.M. are investigators of the PROVALID study and provided substantial contributions to the acquisition and interpretation of the data. All authors approved the final draft of the manuscript. G.M. and D.F. are the guarantors of this work and, as such, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Prior Presentation. Parts of this study were presented in abstract form at the 62nd European Renal Association Congress, Vienna, Austria, 4–7 June 2025.
Funding Statement
This study was funded by the European Union’s Horizon 2020 research and innovation program under grant 848011 (DC-ren). S.Sc. and D.F. are supported by the Deutsche Forschungsgemeinschaft (SFB TRR 219, Project-ID 322900939). S.D. has received support from ERA PerMed/Österreichischer Wissenschaftsfonds (KidneySign, grant 779282/FWF, Project-ID I 6470-B) outside of the submitted work.
Footnotes
See accompanying article, p. 17.
Supporting information
References
- 1. Francis A, Harhay MN, Ong ACM, et al.; International Society of Nephrology . Chronic kidney disease and the global public health agenda: an international consensus. Nat Rev Nephrol 2024;20:473–485 [DOI] [PubMed] [Google Scholar]
- 2. Ortiz A, Covic A, Fliser D, et al.; Board of the EURECA-m Working Group of ERA-EDTA . Epidemiology, contributors to, and clinical trials of mortality risk in chronic kidney failure. Lancet 2014;383:1831–1843 [DOI] [PubMed] [Google Scholar]
- 3. Jankowski J, Floege J, Fliser D, Böhm M, Marx N. Cardiovascular disease in chronic kidney disease: pathophysiological insights and therapeutic options. Circulation 2021;143:1157–1172 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Porrini E, Ruggenenti P, Mogensen CE, et al.; ERA-EDTA diabesity working group . Non-proteinuric pathways in loss of renal function in patients with type 2 diabetes. Lancet Diabetes Endocrinol 2015;3:382–391 [DOI] [PubMed] [Google Scholar]
- 5. Glinka A, Wu W, Delius H, Monaghan AP, Blumenstock C, Niehrs C. Dickkopf-1 is a member of a new family of secreted proteins and functions in head induction. Nature 1998;391:357–362 [DOI] [PubMed] [Google Scholar]
- 6. Federico G, Meister M, Mathow D, et al. Tubular Dickkopf-3 promotes the development of renal atrophy and fibrosis. JCI Insight 2016;1:e84916. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Mourtada J, Thibaudeau C, Wasylyk B, Jung AC. The multifaceted role of human Dickkopf-3 (DKK-3) in development, immune modulation and cancer. Cells 2023;13:75. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Schunk SJ, Floege J, Fliser D, Speer T. WNT-β-catenin signalling - a versatile player in kidney injury and repair. Nat Rev Nephrol 2021;17:172–184 [DOI] [PubMed] [Google Scholar]
- 9. Fu K, Jing C, Shi J, et al. WTAP and METTL14 regulate the m6A modification of DKK3 in renal tubular epithelial cells of diabetic nephropathy. Biochem Biophys Res Commun 2024;738:150524. [DOI] [PubMed] [Google Scholar]
- 10. Schunk SJ, Zarbock A, Meersch M, et al. Association between urinary Dickkopf-3, acute kidney injury, and subsequent loss of kidney function in patients undergoing cardiac surgery: an observational cohort study. Lancet 2019;394:488–496 [DOI] [PubMed] [Google Scholar]
- 11. Roscigno G, Quintavalle C, Biondi-Zoccai G, et al. Urinary Dickkopf-3 and contrast-associated kidney damage. J Am Coll Cardiol 2021;77:2667–2676 [DOI] [PubMed] [Google Scholar]
- 12. Hu J, Zhou Y, Huang H, et al. Prediction of urinary dickkopf-3 for AKI, sepsis-associated AKI, and PICU mortality in children. Pediatr Res 2023;93:1651–1658 [DOI] [PubMed] [Google Scholar]
- 13. Zewinger S, Rauen T, Rudnicki M, et al. Dickkopf-3 (DKK3) in urine identifies patients with short-term risk of eGFR loss. J Am Soc Nephrol 2018;29:2722–2733 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Sánchez-Álamo B, García-Iñigo FJ, Shabaka A, et al. Urinary Dickkopf-3: a new biomarker for CKD progression and mortality. Nephrol Dial Transplant 2021;36:2199–2207 [DOI] [PubMed] [Google Scholar]
- 15. Speer T, Schunk SJ, Sarakpi T, et al.; 4C Study Investigators ; ESCAPE Trial Investigators. Urinary DKK3 as a biomarker for short-term kidney function decline in children with chronic kidney disease: an observational cohort study. Lancet Child Adolesc Health 2023;7:405–414 [DOI] [PubMed] [Google Scholar]
- 16. Schäfer A-KC, Pieper D, Dihazi H, et al. Urinary Dickkopf-3 (DKK3) is associated with greater eGFR loss in patients with resistant hypertension. J Clin Med 2023;12:1034. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Schunk SJ, Beisswenger C, Ritzmann F, et al. Measurement of urinary Dickkopf-3 uncovered silent progressive kidney injury in patients with chronic obstructive pulmonary disease. Kidney Int 2021;100:1081–1091 [DOI] [PubMed] [Google Scholar]
- 18. Pieper D, Sandek A, Schäfer A-K, et al. Urinary Dickkopf-3 as a potential marker for estimated glomerular filtration rate decline in patients with heart failure. J Am Heart Assoc 2024;13:e036637. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Eder S, Leierer J, Kerschbaum J, et al. A Prospective Cohort Study in Patients with Type 2 Diabetes Mellitus for Validation of Biomarkers (PROVALID) - study design and baseline characteristics. Kidney Blood Press Res 2018;43:181–190 [DOI] [PubMed] [Google Scholar]
- 20. Stevens PE, Ahmed SB, Carrero JJ, et al. KDIGO 2024 clinical practice guideline for the evaluation and management of chronic kidney disease. Kidney Int 2024;105:S117–S314 [DOI] [PubMed] [Google Scholar]
- 21. R Core Team . R: a language and environment for statistical computing. Vienna, R Foundation for Statistical Computing, 2021 [Google Scholar]
- 22. Oshima M, Shimizu M, Yamanouchi M, et al. Trajectories of kidney function in diabetes: a clinicopathological update. Nat Rev Nephrol 2021;17:740–750 [DOI] [PubMed] [Google Scholar]
- 23. Penno G, Solini A, Orsi E, et al.; Renal Insufficiency and Cardiovascular Events (RIACE) Study Group . Non-albuminuric renal impairment is a strong predictor of mortality in individuals with type 2 diabetes: the Renal Insufficiency and Cardiovascular Events (RIACE) Italian multicentre study. Diabetologia 2018;61:2277–2289 [DOI] [PubMed] [Google Scholar]
- 24. Viazzi F, Russo GT, Ceriello A, et al. Natural history and risk factors for diabetic kidney disease in patients with T2D: lessons from the AMD-annals. J Nephrol 2019;32:517–525 [DOI] [PubMed] [Google Scholar]
- 25. Schmidt-Lauber C, Hänzelmann S, Schunk S, et al. Kidney outcome after mild to moderate COVID-19. Nephrol Dial Transplant 2023;38:2031–2040 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Husain-Syed F, Villa G, Wilhelm J, et al. Renal markers for monitoring acute kidney injury transition to chronic kidney disease after COVID-19. Nephrol Dial Transplant 2021;36:2143–2147 [DOI] [PubMed] [Google Scholar]
- 27. Liu M, Shi S, Senthilnathan S, et al. Genetic variation of DKK3 may modify renal disease severity in ADPKD. J Am Soc Nephrol 2010;21:1510–1520 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Arjune S, Späth MR, Oehm S, et al. DKK3 as a potential novel biomarker in patients with autosomal polycystic kidney disease. Clin Kidney J 2024;17:sfad262. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Fliser D. Urinary Dickkopf-3 and the evaluation of chronic kidney disease progression. Nephrol Dial Transplant 2021;36:2161–2163 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.