Admission plasma uromodulin and the risk of acute kidney injury in hospitalized patients with cirrhosis: a pilot study

Abstract

Acute kidney injury (AKI) is a common complication in hospitalized patients with cirrhosis. Uromodulin, a protein uniquely produced by the kidney and released both in the urine and circulation, has been shown to regulate AKI and is linked to tubular reserve. Although low levels of urine uromodulin are associated with AKI after cardiac surgery, it is unclear whether circulating uromodulin can stratify the risk of AKI, particularly in a susceptible population such as hospitalized patients with cirrhosis. Thus, we investigated whether plasma uromodulin measured at the time of admission is associated with subsequent hospital-acquired AKI (defined by a rise in serum creatinine >0.3mg/dL within 48 h or ≥ 1.5 times baseline) in patients with cirrhosis. A total of 98 patients [mean age 54 yr, Model for Endstage Liver Disease Sodium (MELD-Na) score 19, and baseline creatinine of 0.95 mg/dL] were included, of which 13% (n = 13) developed AKI. Median uromodulin levels were significantly lower in patients who developed AKI compared with patients who did not (9.30 vs. 13.35 ng/mL, P = 0.02). After adjusting for age, sex, diabetes, hypertension, albumin, and MELD-Na score as covariates on multivariable logistic regression, uromodulin was independently associated with AKI [odd ratios of 1.19 (95% confidence interval 1.02, 1.37; P = 0.02)]. Lower uromodulin levels on admission are associated with increased odds of subsequent AKI in hospitalized patients with cirrhosis. Further studies are needed to better understand the role of uromodulin in the pathogenesis and as a predictive biomarker of AKI in this population.

NEW & NOTEWORTHY In this study, we found that admission plasma uromodulin levels are significantly lower in patients who developed subsequent acute kidney injury (AKI) during their hospital stay compared with patients who did not. Additionally, uromodulin is independently associated with AKI development after adjusting for clinically relevant parameters such as age, sex, diabetes, hypertension, severity of cirrhosis, and kidney function. To our knowledge, this is the first study linking plasma uromodulin with AKI development in patients with cirrhosis.

INTRODUCTION

Acute kidney injury (AKI) is a common complication that occurs in patients with decompensated cirrhosis (13) and is included in the definition of acute on chronic liver failure, a syndrome characterized by acute decompensation and extra-hepatic organ failures driven by excessive systemic inflammation (30, 32). In patients admitted with decompensated cirrhosis, AKI is present in 20% at the time of admission to the hospital (19) and develops in up to 19% after admission (19). Development of in-hospital AKI is independently associated with high short-term mortality (44) and poor outcomes after liver transplantation (31). AKI is also associated with progressive loss of kidney function, which can lead to future episodes of AKI (i.e., AKI on top of chronic kidney disease); thus, it contributes to significant morbidity (21). To improve outcomes related to hospital-acquired AKI in the setting of cirrhosis, identifying patients at high risk is essential. A reliable approach for AKI prediction and risk stratification could be valuable in this context, as it may allow for timely preventative and possible therapeutic care (2) and appropriate allocation of hospital resources.

Uromodulin, also known as Tamm-Horsfall protein, is a glycoprotein exclusively expressed in the thick ascending limb and distal convoluted tubule of the kidneys. This protein is excreted as the most abundant urinary protein in healthy humans (10) but is also released by the kidney into the circulation through the peritubular capillary network (8, 28). Because of its unique production in thick ascending limb cells, uromodulin, unlike other conventional markers for kidney function, is linked to tubular health and reserve (23, 29, 33, 34, 39). In fact, measurement of uromodulin in the urine and the serum has been positively correlated with kidney function (6, 23, 33, 37, 41). Furthermore, a protective role of uromodulin in experimental models of AKI has been demonstrated (11, 28). This positive effect against injury occurs through an immunomodulatory function of uromodulin that downregulates inflammation both in the kidney and systemically. Therefore, it is not surprising to observe emerging interest in the value of measuring uromodulin to prognosticate kidney disease and/or cardiovascular events (5, 14, 15, 23, 29).

Indeed, many recent studies linked low levels of urinary or circulating uromodulin with the risk of chronic kidney disease progression and mortality (5, 6, 14, 16, 23, 39). In AKI, available data are limited to cardiac surgery, showing that low urine uromodulin is associated with higher odds of AKI and higher peak serum creatinine after cardiac surgery (4, 15). However, it may be even more impactful to determine whether the prognostic value of uromodulin could be extended to other clinical settings. Furthermore, there is no available data on the value of measuring circulating uromodulin in predicting the risk of AKI. This could be quite relevant because serum or plasma measurements may be easier to handle and store and could be less prone to enzymatic degradation by proteases compared with the urine (46). Additionally, serum or plasma levels may be easier to interpret, as compared with urinary levels whereby a correction for urinary concentration may be required (27). Thus, the primary objective of this study was to evaluate whether admission plasma uromodulin is associated with AKI development within 7 days of hospitalization in cirrhotic patients. Other relevant biomarkers were also assessed and compared with the performance of uromodulin. Additionally, given the link of uromodulin and inflammation in patients without cirrhosis (22, 23, 39), we evaluated the association of uromodulin with inflammatory cytokines relevant to the pathophysiology of AKI and cirrhosis (1), such as tumor necrosis α (TNF-α), interleukin (IL)-6, and IL-10.

MATERIALS AND METHODS

Study design.

The study population comprised hospitalized cirrhotic patients who were nonconsecutively enrolled in a study evaluating urea metabolism (17). The enrollment period was from September 2010 to April 2016 and included patients who were nonelectively admitted to Indiana University Hospital. The diagnosis of cirrhosis was made based on clinical parameters involving laboratory tests, endoscopic or radiologic evidence of cirrhosis, evidence of decompensation (hepatic encephalopathy, ascites, variceal bleeding, jaundice), and liver biopsy if available. Inclusion criteria included a known diagnosis of cirrhosis, age ≥ 18 yr, and plasma samples obtained within 24 h of admission. Exclusion criteria included prior kidney or liver transplant, history of active cancer, AKI on admission, hemodialysis at the time of admission, and confirmed pregnancy. Patients received written information about the study and a signed consent form. All study protocols were submitted to and approved by the institutional review board at our center.

Data collection.

For patients who met inclusion criteria, data on etiology of cirrhosis (hepatitis C, alcohol, nonalcoholic steatohepatitis, hepatitis C with current alcohol use, and other), demographics, presence of infection on admission, history of complications related to cirrhosis (ascites, hepatic encephalopathy, history of portal hypertension related bleed), history of diabetes, history of hypertension, baseline creatinine, baseline estimated glomerular filtration rate (eGFR) calculated by Chronic Kidney Disease Epidemiology Collaboration creatinine equation (24), admission laboratory data (creatinine, eGFR, albumin, sodium, total bilirubin, internal normalized ratio, and white blood cell count), and mean arterial pressure were collected. Additionally, the severity of cirrhosis was recorded on admission through calculation of the Model for End-stage Liver Disease Sodium (MELD-Na) (12) and Child Turcotte Pugh (CTP) (35) scores. The MELD-Na score includes total bilirubin, international normalized ratio, creatinine, and sodium. The CTP score includes total bilirubin, albumin, international normalized ratio, and severity of ascites and hepatic encephalopathy.

Definition of AKI.

AKI was defined as a rise in creatinine of 0.3 mg/dL or 50% increase from baseline, as recommended by Kidney Disease Improving Global Outcomes (18). Baseline creatinine was defined per the Kidney Disease Improving Global Outcomes (18) as the availability of outpatient serum creatinine within the past year. When more than one creatinine value was available, the closest to admission time to the hospital was used. The median [interquartile range (IQR)] time between baseline creatinine and admission was 15 days (3, 42). In patients who did not have a baseline creatinine (n = 10), the admission creatinine was used. Furthermore, to ensure that these 10 patients had stable kidney function at the time of admission (i.e., did not already have AKI), we excluded any of them whose serum creatinine changed by > 0.30 mg/dL within 48 h from the time of admission, similar to what Belcher et al. (3) used previously .

Sample collection and biomarker measurement.

Blood samples (30 mL) were obtained via venipuncture or drawn from an established peripheral or central line. Samples were immediately refrigerated and then centrifuged for 10 min at −4°C for plasma. Aliquots of 0.5 mL of supernatant were subsequently stored within 6 h of collection in cryovials at −80°C. For the purpose of this investigation, in addition to uromodulin, the following kidney biomarkers and inflammatory cytokines were measured at the time of admission: cystatin-C (CysC), neutrophil gelatinase-associated lipocalin (NGAL), kidney injury molecule-1 (KIM-1), osteopontin, IL-18, IL-10, IL-6, and TNF-α. One aliquot was needed for uromodulin, CysC, NGAL, KIM-1, osteopontin, IL-18, IL-10, IL-6, and TNF-α. Samples were stored for a median (IQR) of 1,883 (1,669–2,164) days until biomarker measurements were made. All biomarkers were measured from frozen aliquots that did not undergo any freeze-thaw cycles. Laboratory measurements were performed by personnel blinded to patient information at the Multiplex Analysis Core at the Indiana University Melvin and Bren Simon Cancer Center.

Cytokine and kidney biomarker concentrations in the plasma were assayed using MilliporeSigma multiplex kits (Human kidney Injury Panel 4-2-Plex, Human Kidney Injury Panel 5-2-Plex, Human Cytokine/Chemokine Panel 6-Plex, and Human Cytokine 5-Plex). The kits contain spectrally distinct antibody-immobilized beads (2 or 5 bead sets specifically for the above biomarkers), cytokine standard cocktail, cytokine quality control I and II, detection antibody cocktail, streptavidin-phycoerythrin, assay buffer, wash buffer, serum matrix, and a microtiter plate. The assay was performed according to the manufacturer’s protocol. After preparation, samples were processed (50 beads per bead set in 25 microliter sample size) on a Bio-Plex 200 System with High Throughput Fluidics (HTF) Multiplex Array System (Bio-Rad Laboratories, Hercules, CA). All samples were run in duplicate. The detection ranges for the biomarkers were as follows: osteopontin 1.172–1,284.08 ng/mL; KIM-1 0.18–223.37 ng/mL; NGAL 0.05–48.36 ng/mL; uromodulin 0.05–50.00 ng/mL; IL-18 8.04–24,703.45 pg/mL, and CysC 0.06–61.03 ng/mL. The detection ranges for inflammatory biomarkers were as follows: IL-10 3.17–10,021.54 pg/mL; IL-6 3.21–9,872.59 pg/mL; and TNF-α 3.18–10,031.73 pg/mL.

Statistical analysis.

Patients were stratified based on the development of AKI or not, and their characteristics on admission were compared. Continuous variables were presented as means ± SD and median ± IQR where deemed appropriate. Categorical variables were presented as frequencies and percentages. Bivariate differences across groups with respect to categorical variables were analyzed using chi-square tests, using Fisher’s Exact tests when cell counts were small, whereas continuous variables were analyzed using Wilcoxon rank-sum tests when data were nonlinear and Student’s t tests when data were linear. Correlations between uromodulin with clinical variables and inflammatory cytokines were analyzed through Spearman nonparametric correlations because of data skewness. A two-sided nominal P value ≤ 0.05 was considered significant. Multivariable logistic regression models were used to determine whether uromodulin was independently associated with AKI development. Variables of clinical significance were chosen for the multivariable model. The final list of covariates was also determined by removing variables that caused high collinearity, as accessed by Variance Inflation Factors. We reported odds ratios and their corresponding 95% confidence intervals. All analytic assumptions were verified, and all analyses were performed using SAS v9.4 (SAS Institute, Cary, NC).

RESULTS

A total of 160 patients with cirrhosis were enrolled during the study period. Of these, 62 patients were excluded (21 with hepatocellular cancer, 5 on hemodialysis, and 36 with AKI on admission), leaving 98 patients for inclusion in this study analysis.

Patient population.

The mean age, MELD-Na score, and CTP score were 54.3 ± 8.4 yr, 19.4 ± 6.2, and 9.4 ± 2.1, respectively. The mean baseline and admission creatinine were 0.95 ± 0.34 and 0.99 ± 0.39 mg/dL, respectively. Among those patients who did not have a baseline creatinine (n = 10), the mean admission creatinine was 0.83 mg/dL and the mean change of creatinine within 48 h was 0.13 mg/dL. Of these 10 patients without prior baseline creatinine, only 1 patient developed AKI on day 4 of hospitalization. The median (IQR) of uromodulin was 12.71 ng/mL (8.31–18.31 ng/mL). Nonalcoholic steatohepatitis (35%), alcohol (22%), and hepatitis C with concurrent alcohol (17%) were the most common etiologies of cirrhosis. The most common reasons for admission were hepatic encephalopathy (36%), infection (26%), ascites and anasarca management (16%), and portal hypertension-related bleeding (11%).

Development of AKI.

Thirteen patients (13%) developed AKI and 85 patients did not develop AKI. There were no statistical differences in age, gender, race, etiology of cirrhosis, history of complications related to cirrhosis, and history of diabetes or hypertension between the two groups at the time of admission (Table 1). Additionally, there were no significant differences in baseline creatinine and eGFR, admission creatinine and eGFR, admission mean arterial pressure, presence of infection on admission, admission white blood cell count, and admission MELD-Na and CTP scores between the two groups. However, patients who developed AKI had a significantly lower albumin compared with those who did not develop AKI (2.5 vs. 2.9 g/dL, P = 0.02). Median uromodulin levels were found to be significantly lower in patients who developed AKI compared with patients who did not develop AKI [9.30 (6.31–13.22) vs. 13.35 (8.55–18.96) ng/mL, P = 0.02] (Fig. 1).

Table 1.

Baseline patient demographics stratified by AKI status

	No AKI Development	AKI Development	P
n	85	13
Age	54.02 (8.38)	55.77 (8.32)	0.49
Gender, n (%) male	46 (54)	6 (46)	0.78
Race, n (%) white	77 (91)	12 (92)	0.52
Etiology of cirrhosis, n (%)
HCV	10 (12)	3 (23)
HCV + alcohol	16 (19)	1 (8)
NASH	29 (34)	5 (38)	0.75
Alcohol	19 (22)	3 (23)
Other	11 (13)	1 (8)
Infection on admission, n (%)	18 (21)	5 (38)	0.18
History of ascites, n (%)	61 (72)	12 (92)	0.17
History of HE, n (%)	58 (68)	11 (85)	0.19
History of PHTN bleed, n (%)	13 (15)	1 (8)	0.69
DM, n (%)	35 (41)	6 (46)	0.77
HTN, n (%)	29 (34)	5 (38)	0.76
MAP	84.08 (12.90)	80.38 (11.78)	0.33
Baseline creatinine, mg/dL	0.94 (0.35)	0.99 (0.30)	0.60
Baseline eGFR, mL·min−1·1.73 m−2*	86.06 (26.86)	76.77 (21.34)	0.23
Admit creatinine, mg/dL	0.98 (0.38)	1.08 (0.41)	0.39
Admit eGFR, mL·min−1·1.73 m−2*	82.82 (27.01)	73.62 (26.82)	0.26
Admit sodium, mmol/L	133.56 (12.00)	132.54 (6.62)	0.65
Admit albumin, g/dL	2.94 (0.62)	2.51 (0.57)	0.02
Admit total bilirubin, mg/dL	4.19 (4.95)	2.84 (2.60)	0.14
Admit WBC, 109/L	6.98 (4.50)	7.92 (3.84)	0.48
Admit INR	1.65 (0.58)	1.61 (0.57)	0.84
CTP score	9.31 (2.10)	9.77 (1.88)	0.45
CTP class, n (%)
A	10 (12)	0 (0)
B	33 (39)	6 (46)	0.57
C	42 (49)	7 (54)
MELD-sodium score	19.27 (6.22)	20.23 (6.37)	0.61

AKI, acute kidney injury; CTP, Child Turcotte Pugh; DM, diabetes mellitus; eGFR, estimated glomerular filtration rate; HCV, hepatitis C; HE, hepatic encephalopathy; HTN, hypertension; INR, international normalized ratio; MAP, mean arterial pressure; MELD, Model for Endstage Liver Disease; NASH, nonalcoholic steatohepatitis; PTHN, portal hypertension; WBC, white blood cell.

^*Based on Chronic Kidney Disease Epidemiology Collaboration creatinine equation (24).

Fig. 1.

Comparison of uromodulin stratified by acute kidney injury (AKI) status. Bar graph showing the median interquartile ranges of uromodulin for patients who developed AKI compared with patients who did not.

The median time (IQR) for AKI development was 4 (2, 5) hospital days, and all 13 patients were categorized as stage 1 at the time of diagnoses. Of these, 4 and 5 patients progressed to stage 2 and 3, respectively, and 4 patients subsequently required hemodialysis. The most common precipitants identified for AKI development were infection (31%), aggressive diuresis (27%), and large volume paracentesis (27%).

Clinical factors associated with uromodulin levels at admission.

There were no significant differences in uromodulin levels between race, sex, etiology of cirrhosis, and CTP classes. Patients with diabetes (n = 41) had significantly lower median uromodulin levels compared with those without diabetes [9.30 (6.89–14.91) vs. 15.45 (10.94–21.43) ng/mL, P < 0.01]. Similarly, patients with hypertension (n = 34) had a significantly lower median uromodulin level compared with patients without hypertension [9.76 (7.04–18.31) vs. 14.48 (10.20–19.31) ng/mL, P = 0.02]. Furthermore, uromodulin was significantly and positively correlated with admission eGFR (r = 0.40, P < 0.01). There were no significant correlations between uromodulin and MELD-Na score, CTP score, MAP, albumin, or white blood cell count.

Correlations between uromodulin and kidney biomarkers and inflammatory cytokines.

Uromodulin was significantly correlated inversely with CysC (r = −0.46, P < 0.01) and directly with osteopontin (r = 0.23, P = 0.03) and IL-18 (r = 0.22, P = 0.03). There were no significant correlations between uromodulin and NGAL and KIM-1 (Table 2). There were no significant correlations between uromodulin and the inflammatory cytokines, IL-10, IL-6, and TNF-α (Table 2).

Table 2.

Correlations between uromodulin and kidney biomarkers and inflammatory cytokines

	Uromodulin	P
Kidney biomarkers
NGAL	−0.17	0.10
Cystatin-C	−0.46	<0.01
KIM-1	0.03	0.76
Osteopontin	0.23	0.03
IL-18	0.22	0.03
Inflammatory cytokines
TNF-α	−0.08	0.41
IL-10	−0.01	0.90
IL-6	0.03	0.77

IL, interleukin; KIM-1, kidney injury molecule-1; NGAL, neutrophil gelatinase-associated lipocalin; TNF, tumor necrosis factor.

Differences in kidney biomarkers and inflammatory molecules stratified by AKI status.

Median NGAL (203.69 vs. 126.67 ng/mL, P = 0.01) and osteopontin (83.89 vs. 55.05 ng/mL P < 0.01) levels were found to be significantly higher in patients who developed AKI compared with patients who did not. There were no differences between the two groups for KIM-1, CysC, and IL-18 (Table 3). Of the inflammatory cytokines, median levels for IL-6 and TNF-α were found to be statistically higher in the AKI development group versus no AKI development group [23.76 vs. 16.24 pg/mL (P = 0.02) and 46.33 vs. 26.6 pg/mL (P < 0.01), respectively] (Table 3).

Table 3.

Comparison of kidney biomarkers and inflammatory cytokines stratified by AKI status

	No AKI Development	AKI Development	P
n	85	13
Kidney biomarkers, median (IQR)
NGAL, ng/mL	126.67 (99.49–173.45)	203.69 (155.53–284.98)	0.01
Cystatin C, ng/mL	843.59 (652.42–1,144.28)	997.08 (897.15–1,322.00)	0.14
KIM-1, ng/mL	1.19 (0.47–2.05)	0.97 (0.66–2.68)	0.10
Osteopontin, ng/mL	55.05 (35.57–76.14)	83.89 (57.25–94.85)	<0.01
IL-18, pg/mL	77.03 (46.67–171.52)	99.94 (84.99–110.68)	0.59
Inflammatory cytokines, median (IQR)
IL-10, pg/mL	11.27 (7.10–19.33)	15.39 (12.21–25.45)	0.12
IL-6, pg/mL	16.24 (8.54–24.23)	23.76 (20.72–63.65)	0.02
TNF-α, pg/mL	26.60 (17.79–38.08)	46.33 (39.02–55.11)	<0.01

n, number of patients. AKI, acute kidney injury; IL, interleukin; IQR, interquartile range; KIM-1, kidney injury molecule-1; NGAL, neutrophil gelatinase-associated lipocalin; TNF, tumor necrosis factor.

Association of uromodulin with AKI.

Each 1 ng/mL lower, uromodulin was associated with 19% increased odds for AKI development [1.19 (95% confidence interval 1.02, 1.37); P = 0.02] after adjusting for age, sex, diabetes, hypertension, serum albumin, and MELD-Na score. We found that none of the other kidney biomarkers were associated with AKI after similar adjustment for confounders.

DISCUSSION

In this study, we found that admission plasma uromodulin levels are significantly lower in patients who developed subsequent AKI during their hospital stay compared with patients who did not. Although other commonly used biomarkers such as NGAL and osteopontin also showed differences between the two groups, only uromodulin remained independently associated with AKI development after adjusting for clinically relevant parameters such as age, sex, diabetes, hypertension, severity of cirrhosis, and kidney function. To our knowledge, this is the first study linking plasma uromodulin with AKI development, albeit in a unique population of patients with cirrhosis.

The function of uromodulin is not fully understood; however, recent discoveries have underscored its importance as a regulatory protein for renal function and its association with functional renal reserve (5, 33, 34, 39). In the setting of AKI, the expression of uromodulin has been found to be significantly decreased at the onset of injury in experimental models, suggesting that early AKI leads to uromodulin deficiency in the kidney (10, 11). These data have been supported by findings in kidney transplant patients showing that serum uromodulin levels correlate with graft function after transplant (39). As far as uromodulin and the risk of AKI, the available data are on urine uromodulin levels in clinical studies, whereby low urine uromodulin levels have been linked to increased AKI risk after cardiac surgery (4, 15) and liver transplantation (38). Our study is the first linking circulating uromodulin measured on admission to hospital-acquired AKI. Therefore, measurement of circulating uromodulin on admission may allow for early risk assessment for timely prevention and intervention (2). Nevertheless, further prospective validation studies in larger populations are needed to validate its potential use as a theranostic tool to guide prevention and management of patients with cirrhosis.

Interestingly, we found no correlation between uromodulin and TNF-α. This is in contrast to a previous study that found associations of low urine uromodulin levels with TNF-α(22). The cause of this discrepancy is likely due to the small sample size of our population and the bias introduced by cirrhosis, itself a state of systemic inflammation (1). However, albeit weak direct associations, we found IL-18 and osteopontin to be correlated to uromodulin. IL-18 is a proinflammatory cytokine that is released by the proximal tubules in response to inflammation and kidney injury and has been linked with AKI in patients with cirrhosis (36). Osteopontin is an extracellular matrix protein involved in the inflammatory response (20) that has been found to be significantly increased in all tubular segments in AKI (45). Given its possible protective, anti-inflammatory functions and role in maintaining homeostasis (29), uromodulin’s correlation with IL-18 and osteopontin is perhaps due to a compensatory increase in response to a systemic illness (9), such as a state of cirrhosis. However, larger studies in patients with cirrhosis are required to substantiate this correlation. It is interesting to note the diverging association of NGAL, TNF-α, and IL-6 to the risk of AKI but not to uromodulin. This could be related to the strength of the association of uromodulin to the risk of AKI, whereby this link remains uniquely significant even after adjustment for potential confounders. Of course, we cannot rule out the presence of biological heterogeneity in the clinical context of cirrhosis (3, 7, 40), which could weaken the correlation of inflammatory and injury markers on admission to the risk of AKI in this small cohort.

Higher serum concentrations of CysC have been associated with AKI in patients with cirrhosis (25, 26). However, in comparison to these studies, we did not find CysC (or any of the other biomarkers except for uromodulin) to be associated with AKI. Possible reasons for these differences may be attributed to differences in study population, impact and prevalence of bacterial infections, differences in baseline values of CysC, and adjusting for known influences of CysC such as serum albumin (42). It is important to note that in regard to the latter two, our patients had a much lower baseline CysC compared with aforementioned studies and our multivariable model was adjusted for serum albumin. Furthermore, these differences may be attributable to the different assay platform for CysC. It is also not surprising to find an association between uromodulin and CysC, which reflects the known association between uromodulin levels in the urine and circulation with kidney function (41).

There are several limitations to our study. This is an exploratory study performed at a single center. The small sample size may have reduced our ability to link more variables on admission to the odds of developing subsequent AKI. Therefore, our findings will need to be validated in a larger study before generalization. It is also unclear if the findings could be extended to an outpatient setting of stable decompensated cirrhotic patients, which is also very clinically relevant given their risk for AKI (43). Furthermore, serial measurements of uromodulin may add significant insights to its usefulness in predicting the risk of AKI and will likely need to be incorporated in future studies. Lastly, although no harmful effects have been reported on the storage time for uromodulin, we cannot rule out possible degradation of the protein over time, which may lead to dilution and therefore potential misclassification.

Despite these limitations, our study also had a number of strengths. First, all patients included in the study had detailed laboratory information, which allowed for serial creatinine monitoring and discernment for AKI development during the course of the hospitalization. Furthermore, we were able to obtain baseline creatinine in 90% of our cohort, which was not available in prior studies linking low uromodulin to AKI. Lastly, we have included hard outcomes such as AKI stage 3 and hemodialysis use, which were unavailable in the aforementioned prior studies due to mild AKI.

In conclusion, in this study, we report that low plasma uromodulin levels at the time of admission are independently associated with higher odds of AKI in hospitalized patients with cirrhosis. If validated, the measurement of uromodulin could enhance our clinical decision making for risk assessment of AKI in patients with cirrhosis.

GRANTS

This study was supported by the National Institute of Diabetes Digestive and Kidney Disease Grants 1R01DK111651 (to T. M. El-Achkar), K23 DK114556 (to P. S. Garimella), and K23 DK109202 (to E. S. Orman) and Veterans Affairs Merit Award (T. M. El-Achkar).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

K.R.P. and T.M.E.-A. conceived and designed research; K.R.P., J.E.S., and T.M.E.-A. analyzed data; K.R.P., P.S.G., E.M., and T.M.E.-A. interpreted results of experiments; K.R.P. and T.M.E.-A. prepared figures; K.R.P., J.E.S., and T.M.E.-A. drafted manuscript; K.R.P. and T.M.E.-A. edited and revised manuscript; K.R.P., P.S.G., E.M., J.E.S., M.S.G., R.E.W., M.D.A., E.S.O., L.D.N., A.P.D., N.C., and T.M.E.-A. approved final version of manuscript.