Urinary uromodulin independently predicts end-stage renal disease and rapid kidney function decline in a cohort of chronic kidney disease patients

Supplemental Digital Content is available in the text


Introduction
Chronic kidney disease (CKD) represents one of the major medical burdens in Western countries. Health care costs associated with CKD are high and further increase when endstage renal disease (ESRD) is reached. [1][2][3] In addition, morbidity and mortality are significantly elevated, predominantly due to cardiovascular complications. [4] Therefore, it is crucial to diagnose patients with CKD early and to identify those who have rapid CKD progression to potentially intervene or prepare them for renal replacement therapy. [5][6][7] Biomarkers appear to be an attractive diagnostic approach to early identify CKD patients and those who are at risk for rapid CKD progression. [8,9] Different parameters for the prediction of ESRD or decline of estimated glomerular filtration rate (eGFR) and death in the long term (>3 years of follow-up) have been evaluated. [10][11][12][13][14][15][16] However, markers that predict kidney function decline in the short term are needed in order to take measures such as hemodialysis access. Recently, urinary uromodulin (uUMOD) has been identified as a valuable parameter for the prediction of ESRD and progression of CKD in a large cohort over a period >9 and >3 years, respectively. [17,18] In this study, we evaluated whether uUMOD is associated with rapid loss of eGFR and ESRD in CKD patients within 1 year of follow-up.

Patients and methods
The final cohort consisted of 230 patients at stages I-V of CKD who presented to the outpatient clinic of a tertiary care university hospital. The study was approved by the local ethics committee of Klinikum rechts der Isar, Technische Universität, Munich, Germany, and adheres to the declaration of Helsinki. All patients enrolled in this study gave their informed consent. The only inclusion criteria followed the definitions for CKD according to the last KDIGO guidelines [19] : "CKD is defined as abnormalities of kidney structure or function, present for >3 months, with implications for health." Therefore, we established the diagnosis of CKD when either eGFR was <60 mL/min and/or apparent signs of kidney damage were present over a period of 3 months. As apparent signs of kidney damage, we considered proteinuria with a cut-off >150 mg/g creatinine on spot urine specimen and/ or histologically proven kidney disease and/or abnormalities detected in imaging techniques (ultrasound, computed tomography, magnetic resonance imaging, or nuclear imaging). Calculation of eGFR was based on both serum creatinine and cystatin C concentrations (CKD-EPI crea-cystatin ). [20] Exclusion criteria were age <18 years, psychiatric comorbidities that would not allow written informed consent, and lack of serum/urine sample at the time of potential enrollment. Furthermore, patients with symptomatic or asymptomatic urinary tract infection (UTI), defined as detection of leucocytes and/or bacteria in the urinary sediment at the time of enrollment, were excluded, as it is unclear how acute UTI affects uUMOD secretion. The following parameters were assessed, as they were shown to be relevant markers for CKD progression [21][22][23][24] : eGFR, systolic blood pressure (SBP), diastolic blood pressure (DBP), spot proteinuria (calculated as mg/g urine creatinine), and C-reactive protein (CrP). We compiled the following demographic variables: age, gender, body mass index (BMI), accompanying coronary heart disease (CHD), peripheral artery disease (PAD), and concomitant diabetes mellitus (data were obtained from electronic chart review). Prevalence of CHD and PAD was recorded according to the last medical report, concomitant diabetes in view of the antidiabetic medication and/or HbA1c levels above the cut-off of 5.9%. All laboratory measurements, except uUMOD, had been performed at the day of enrollment. Medications with a renoprotective effect were recorded, including angiotensinconverting enzyme (ACE)-inhibitors/AT1-antagonists, aldosterone antagonists, bicarbonate, erythropoietin, uric acid lowering agents, active vitamin D, and phosphate binding agents.
Primary outcome was reaching ESRD or a 25% decline in eGFR as a composite endpoint within 1 year of follow-up (ascertained through chart review). We chose the 25% cut-off relying on recent studies that demonstrated the ability of a lower eGFR decline to predict renal outcome. [25,26] Follow-up was obtained 12 months after inclusion of the last patient into the study.
Patients' demographics, medication, and predictive parameters are presented in Table 1.

Measurement of urinary uromodulin
All urine samples were stored at -80°C before measurements were performed. Urinary uromodulin measurements were performed using a commercially available assay (Euroimmun AG, Lübeck, Germany). Short performance characteristics of the ELISA for plasma samples given by the manufacturer are as follows: detection limit for plasma samples 2 ng/mL; mean linearity recovery 97% (83-107% at 59-397 ng/mL); intra-assay precision 1.8-3.2% (at 30-214 ng/mL), inter-assay precision 6.6-7.8% (at 35-228 ng/mL), and inter-lot precision 7.2-10.1% (at 37-227 ng/mL). Urine samples were diluted 1 : 101 using dilution buffer. One hundred microliter of calibrators, controls, or diluted samples were pipetted into coated wells of the microtiter plate (MTP); subsequently, 100 mL of biotinylated detection antibody (final concentration 50 ng/mL) were added. The MTP was covered with foil and incubated for 2 hours at 450 rotations per minute (rpm) and room temperature on a rotary shaker. After 2 hours, the MTP was washed 3 times using 300 mL washing buffer, and then the wells were tapped gently. One hundred microliter of Steptavidin-Polyperoxidase (SPO, final concentration 67 ng/mL) were pipetted into each well followed by another incubation for 30 minutes at 450 rpm. Subsequently, the SPO was soaked and the MTP washed 3 times with 300 mL of washing buffer. Consequently, 100 mL of substrate solution (containing the chromogen tetramethylbenzidin and hydrogen peroxide as the substrate for SPO) were pipetted into each well. The MTP was incubated in the dark for 15 minutes at room temperature. The reaction was terminated by adding 100 mL of stop solution. This causes a color change from blue to yellow. Finally, the substrate solution was measured using a photometer at a wavelength of 450 nm and reference wavelength of 620 nm. Data analysis was performed using the program Magellan (Tecan Group Ltd., Männedorf, Switzerland).

Statistics
Due to skewed distribution, data are presented as median with minimum and maximum. Categorical variables are reported in absolute numbers and percentages. We evaluated the correlations of uUMOD, eGFR, and proteinuria using linear regression modeling adjusted for age, gender, and BMI. To better fit the model, we log transformed uUMOD, eGFR serum concentrations, and proteinuria. Subsequently, we divided the cohort into quartiles according to uUMOD concentrations for further analysis. For comparison of demographic data, medication, and laboratory parameters between the quartiles, exact Fisher test for categorical variables and Kruskal-Wallis test for continuous variables were used. Univariable Cox regression analysis was calculated for each variable with the composite endpoint being the dependent variable and the predictor being the independent one. For further work-up, we chose a stepwise approach: parameters significantly associated with the endpoint in univariable analysis (P < .05) were included in the multivariable analysis using forward inclusion. uUMOD was further evaluated in receiver-operating curve (ROC)-analysis to assess the cut-off point (OCO) with optimal sensitivity and specificity to predict the composite endpoint. Kaplan-Meier analysis was performed to illustrate the association between uUMOD and the composite endpoint.
All reported P values are 2-sided, with a significance level of .05 and have not been adjusted for multiple testing. For statistical analysis, SPSS 23 (IBM, Armonk, NY) was used.

Patients' demographics
Three hundred five patients were initially included in the study. At the time of follow-up assessment, 75 (24.6%) patients were lost to follow-up. The patients did not differ substantially from the remaining 230 patients included in the final analysis (Suppl. Table 1 vs Table 1, http://links.lww.com/MD/D7). The mean age of the subjects included was 60 (minimum 18; maximum 92) years, and 152 (66%) were male. Glomerulonephritis was the most frequent underlying disease (UD) with 87 of 230 patients Table 2 Causes for end-stage renal disease in the total cohort and urinary uromodulin quartiles.  Values in median (minimum;maximum) for continuous variables; ACEI = angiotensin-converting enzyme inhibitor, ARB = angiotensin receptor blocker, BMI = body mass index, BP = blood pressure, CHD = coronary heart disease, Cr = creatinine, eGFR = estimated glomerular filtration rate (CKD-EPI creatinine/cystatinC ), PAD = peripheral arterial disease, UA = uric acid, UMOD = uromodulin. Significant differences among the groups are indicated in bold. Detailed baseline characteristics of the participants are presented in Table 1. The classification of UD is reported in Table 2.

Disease
Forty-seven (20.4%) patients reached the composite endpoint, of whom 33 patients reached ESRD and 14 experienced at least 25% decrease in eGFR but not ESRD (Table 1). Of the patients reaching ESRD, 2 were stage CKD III, 12 CKD IV, and 19 CKD V. Among the patients only experiencing at least 25% decrease in eGFR but not ESRD, the patients were widely distributed among all stages of CKD: 2 patients stage I, 3 patients stage II, 3 patients stage III, 5 patients stage IV, and 1 patient stage V.

Univariable analysis of differences between uUMOD quartiles
Demographic parameters did not differ significantly between the quartiles ( Table 1). The quartile with the lowest uUMOD concentrations had the lowest eGFR and the highest degree of proteinuria (P < .001), the latter decreasing to the quartile with the highest uUMOD concentrations (Table 1). CRP was not different between the groups. The quartile with the lowest uUMOD concentrations had a significantly higher proportion of bicarbonate (P = .005), active vitamin D, and phosphate binding medication (P < .001, Table 1). ACE-inhibitors/ARBs, erythropoiesis-stimulating, and uric acid lowering agent prescription were not statistically different within the quartiles (Table 1).

Discussion
In order to take necessary steps to treat CKD patients (e.g., to prepare the patient for renal replacement therapy), biomarkers that predict rapid deterioration of kidney function are needed, but data on this topic are very rare. uUMOD has been shown to predict development of CKD over a period of >9 years. [18] To our knowledge, we demonstrated here for the first time that uUMOD is also associated with rapid progression to ESRD and/or rapid decline of kidney function within 1 year. uUMOD is discussed to play a pathogenic role in CKD. [27] uUMOD has been only moderately well correlated to eGFR previously. [18,28] We detected a moderate association between logarithmic uUMOD and eGFR in our cohort. This suggests that uUMOD potentially represents tubular mass independently from glomerular function, as urinary uromodulin excretion has been shown to correlate with tubular mass. [29] Furthermore, luminal secretion of UMOD into the urine appears to be differently regulated from apical release, as circulatory UMOD was shown to nominally correlate stronger to eGFR than uUMOD. [30] Besides this, tubular mass appears to be important for the preservation of overall kidney function, as we detected that loss of kidney function was predicted by uUMOD independently from eGFR. The question if the predictive value of uUMOD is based on pathophysiologic mechanisms or simply by reflection of tubular mass is beyond the scope of this article.
Other urinary markers were evaluated with regard to their predictive value for loss of kidney function. Urinary neutrophil gelatinase associated lipocalin (NGAL) concentrations in combination with urinary creatinine concentrations were associated with rapid loss of renal function and ESRD in a cohort of 158 patients at stage 3 and 4 of CKD. [31] However, a larger study on >3000 patients did not show a substantial benefit of urinary NGAL concentrations as a predictor when added to known parameters such as proteinuria in CKD patients for adjustment. [16] Urinary cystatin C has not been extensively studied in this context. A Korean study proved its value only in normoalbuminuric diabetic patients. [32] Urinary kidney injury molecule 1 (KIM-1) was also evaluated for predicting CKD progression. [33][34][35] Bhavsar et al [33] and Nielsen et al [35] could not demonstrate a benefit using KIM-1 for risk stratification. Similarly, the value of KIM-1 as a predictor was also rather limited in the study of Peralta et al, [34] only showing a significant difference when comparing the highest decile with the lower 90% of patients. Another problem is that urinary KIM-1 is significantly influenced by medication and sodium restriction. [36] As KIM-1 is suggested to be a marker for acute tubular injury (e.g., prolonged ischemia), KIM-1 appears to be rather useful in the setting of acute kidney failure. [37,38] Recent research focused on a urinary proteomic analysis approach to predict the risk of CKD progression. [39] Although this appears to be a promising way to identify alterations of renal cell activity, interaction, and loss of renal tissue in a dynamic analysis, we currently consider this method a long way off clinical practice due to very high costs. In addition, further reliable data are needed.
Wilson et al [40] suggested a simple approach to estimate the risk for ESRD in CKD patients from urinary creatinine adjusted for fat-free mass. Although a significant predictive value of urinary creatinine was seen, the study did not adjust for parameters that we would assume to be of relevance such as concomitant pharmacologic treatment. Di Micco et al [41] also suggested that lower urinary creatinine concentration predicted ESRD in stage 3 Table 3 Univariable and multivariable cox proportional hazard regression analysis with the composite endpoint ESRD/25% eGFR decline.    to 5 of CKD. However, a very moderate relationship was seen in multivariable analysis with an increased risk of 2% with every 20 mg/dL reduction of urinary creatinine concentration. Furthermore, no significant differences of urinary creatinine concentrations were seen in patients at CKD stage 5, hampering the use of urinary creatinine as a tool for risk evaluation within this important subcohort. eGFR and albuminuria were evaluated in a large meta-analysis with over 20,000 patients. [42] Both parameters were predictive for ESRD, but the heterogeneity concerning eGFR was quite large between the studies. As our data suggest, eGFR might only be of value for risk stratification over a longer period. Proteinuria/albuminuria is indeed a helpful parameter that can also be influenced by clinical measures, but in the study of Astor et al, [42] an 8-fold elevation of proteinuria was needed to reach a HR of 3.04. As the range in uUMOD was smaller in our study, uUMOD might be more promising, as also more subtle differences might indicate a change in risk.
Our study has limitations: we only analyzed the short-term outcomes, so no data on long-term relevance of uUMOD can be provided. However, this was already demonstrated by Garimella et al. [18] Furthermore, data cannot be generalized, as we predominantly involved Caucasian patients. Also, a disproportional large percentage of patients included had glomerulonephritis as the UD, which does not fully represent the overall CKD population. Furthermore, over 30% of the patients included were at CKD stage IV-V, limiting the generalizability of the results to earlier CKD stages. Data were assessed in a single center, so local particularities could have an impact on the results. Samples were stored at -80°C before measurements were performed.
In conclusion, uUMOD appears to be a promising independent biomarker for risk stratification of rapid disease progression in CKD patients.