The ecacy of serum copeptin and zinc-alpha2-glycoprotein for predictors of tolvaptan treatment in decompensated cirrhotic patients

The ecacy of tolvaptan, an orally active vasopressin V2-receptor antagonist, has recently been reported in patients with refractory ascites unresponsive to conventional diuretics. However, the effect of tolvaptan varies among patients. Recently, the prognostic role of tolvaptan response in decompensated liver cirrhosis (LC) is also attracts attention. Our aim is to elucidate predictive factors using serum copeptin (vasopressin precursor), zinc-α2-glycoprotein (ZAG), cystatin C (renal biomarker), neutrophil gelatinase-associated lipocalin (NGAL) and liver-type fatty acid-binding protein (L-FABP) that portend a good response to tolvaptan in LC patients with ascites. levels predict patient response to tolvaptan even when measured prior to treatment. Non-responders group when compared to the Responders group. In addition, we indicated that serum BUN, copeptin and ZAG levels were independent predictors of the overall response to tolvaptan therapy. We further showed that serum creatinine and cystatin C levels were not predictive of patient response to tolvaptan, thus conrming the ndings reported by others 14,15 . In LC, arteriolar vasodilation causes underlling of the systemic arterial vascular space, and the decrease in the effective blood volume leads to a decrease in arterial pressure 21 . Consequently, activation of the renin-angiotensin-aldosterone system, sympathetic nervous system and the release of antidiuretic hormone take place as the body attempts to restore normal blood pressure homeostasis 22 . The main physiological function of AVP is antidiuresis, thereby regulating systemic osmotic pressure. Plasma AVP levels are normally regulated by plasma osmotic pressure and have been found to be elevated in LC patients with ascites, as reported by Pérez-Ayuso RM 23 . However, there has been no study to date exploring whether the effect of tolvaptan is related to plasma AVP or pro-AVP (copeptin) levels in ascites patients. The Non-responder group is considered to be in a state of relative vascular underlling with an increase in BUN and copeptin, thus suggesting intravascular dehydration 14 . Furthermore, serum copeptin also correlated with multiple factors including renal function and CRP, resulting in copeptin is not only a biomarker of renal function. Although NGAL and L-FABP are useful markers than copeptin for acute-on-chronic liver failure 24 and acute kidney injury 25 , respectively, this study indicated that serum NGAL and L-FABP were not useful in assessing the ecacy of response to tolvaptan in LC patients with ascites.


Introduction
Arginine vasopressin (AVP) is a potent antidiuretic hormone in the human body. Despite the clinical relevance of AVP in maintaining uid balance and vascular tone, measurement of the mature form of AVP is di cult due to small size (9 amino acids), short half-life, and its ability to bind platelets 1 . Copeptin, a 39-amino acid glycopeptide that comprises the C-terminal part of the AVP precursor, was found to be a stable and sensitive surrogate marker for AVP release 2 . Copeptin would be particularly helpful in diseases whereby primary or secondary (e.g. cardiovascular instability, chronic heart failure, sepsis) disturbances of the vasopressinergic system contribute to disease pathogenesis 3 . Current reports have demonstrated that a high serum copeptin concentration predicts survival in a heterogeneous population of hospitalized liver cirrhosis (LC) patients, independent of liver-speci c scoring systems 4 , suggesting that copeptin has the potential to be a biomarker of disease progression and prognosis in LC 5 .
Zinc-alpha-2-glycoperotein (ZAG) is a 41-kDa glycoprotein assigned to the major histocompatibility complex (MHC) class I family of proteins. ZAG is a secretory protein synthesized by epithelial cells and adipocytes and plays a role in lipid metabolism, cell cycling and cancer progression. ZAG has been reported as a biomarker of renal injury based on microarray analyses showing an increase of ZAG expression in aged proximal tubular cells 6 . In addition, a clinical investigation showed serum ZAG levels can predict mortality in hemodialysis patients 7 . However, these changes in serum ZAG levels have not been reported in LC patients with refractory ascites.
There are several established markers of renal function, such as serum cystatin C, neutrophil gelatinase-associated lipocalin (NGAL), and liver-type fatty acid binding protein (L-FABP). Serum cystatin C is a cysteine protease inhibitor produced in nearly all nucleated cells. As opposed to creatinine, cystatin C is not in uenced by muscle mass, thus permitting an accurate evaluation of renal function even in patients with low muscle mass. Urea NGAL is a binding partner with matrix metalloproteinase-9 in neutrophils 8 and is established as a clinical biomarker for acute renal injury 9 based on NGAL function of a kidney protector 10 . Urine L-FABP levels derived from proximal tubular epithelial cells are elevated in renal tubular injury episodes and are therefore used as an established marker of several kidney diseases, including acute kidney injury and chronic kidney diseases 12 . The primarily expression of NGAL and L-FABP in the liver led us to investigate whether serum NGAL and serum L-FABP re ect liver and kidney function. We have reported that serum NGAL and serum L-FABP is are prognostic factors for survival in chronic liver diseases 11,13 .
Ascites is the most frequent complication of LC and refractory ascites has a negative prognostic implication in the natural progression of LC. Tolvaptan is a novel, nonpeptide, orally administered, selective vasopressin V 2 -receptor antagonist that downregulates the expression of aquaporin-2 in the renal collecting duct. Current research has uncovered a promising role for tolvaptan as an add-on treatment in patients with hepatic ascites resistant to furosemide and/or spironolactone, as it is able to decrease body weight and alleviate edema 14,15 . However, the effect of tolvaptan is highly variable among patients with approximately one half of the patient cohort responding to the drug 14 . Blood urea nitrogen (BUN) 14 and C-reactive protein (CRP) 15 are recognized for differentiating diagnoses between Responders and Non-responders for tolvaptan, but these factors are insu cient, thus additional biomarkers are needed. Recently, in addition to the aforementioned studies, several studies have also reported Responders to tolvaptan led to improvement of long-term survival rates in cirrhotic patients with ascites 16,17 . Therefore, the aim of this study is to assess the correlation between the e cacy of tolvaptan and treatment related factors, including serum copeptin, ZAG, cystatin C, NGAL and L-FABP levels in LC patients with ascites.

Human samples
The study protocol was approved by the Clinical Research Ethics Review Committee of Mie University Hospital. This study was performed retrospectively on stored samples, and patients could opt out of their data. In this study, we enrolled 113 LC patients (69 males and 44 females) with a mean age of 67.5 ± 10.9 years who could be analyzed between November 2013 and September 2016 at Mie University hospital. We divided the 113 LC patients into tolvaptan treatment group or non-treatment group. The tolvaptan treatment group was consisted of 45 patients. LC was diagnosed based on morphologic changes of the liver such as hypertrophy of the left lateral and caudate lobes, or atrophy of the right posterior hepatic lobe on ultrasonography and through blood tests, and/or computed tomography, magnetic resonance imaging, FibroScan (Echosens, French) results, and esophageal varix by endoscopy, as is the general protocol. HCC was diagnosed based on histological ndings or typical imaging characteristics. Body mass index (BMI) was calculated as weight (kg)/height (m) squared. Using cross-sectional CT at the level of transverse process of lumbar vertebra L3, the bilateral psoas muscle area was identi ed. Psoas muscle index (PMI) (cm 2 /m 2 ) was de ned by normalizing psoas muscle area (cm 2 )/height (m) squared. Patients who had other malignancies within the past 3 years, spontaneous bacterial peritonitis, hepatic encephalopathy (coma scale score ≥ II), heart failure (the New York Heart Association de ned category ≥class II), human immunode ciency virus infection, pregnancy, or psychiatric problems were deemed to be unsuitable for clinical study.

Statistical analysis
All data are expressed as mean ± SD. Data were analyzed using the Mann-Whitney U test in two groups and one-way analysis of variance for comparison of continuous variables. The relationship between serum copeptin, ZAG, cystatin C, NGAL, L-FABP levels and clinical data were examined using Spearman's rank correlation coefficient. For each continuous variable, the optimal cutoff value that maximized the sum of sensitivity and speci city was selected using receiver operating characteristic (ROC) analysis for survival. A logistic regression analysis was utilized for the multivariate analysis in order to evaluate the relationship between effect of tolvaptan and clinical data. Only variables deemed to be significant (p<0.1) in the univariate analysis were included in the subsequent multivariate analysis. The statistical analyses were performed using JMP software program (SAS Institute, Cary, NC, USA) for univariate and multivariate logistic regression analysis. Differences were considered to be signi cant at p<0.05.

Results
Clinical characteristics of patients with or without tolvaptan In this study, we divided the 113 LC patients into tolvaptan treatment group or non-treatment group. Table 1 shows the comparison of baseline clinical characteristics and laboratory variables between patients with and without tolvaptan treatment. In the tolvaptan treatment group, dose of diuretic drugs, the Child-Pugh score, ALBI score, and serum copeptin, NGAL and L-FABP levels were signi cantly increased, while BMI, PMI, serum albumin levels were reduced, suggesting that patients treated with tolvaptan demonstrated more advanced liver diseases. In contrast, there were no obvious differences in age, gender, prevalence of HCC, creatinine, BUN, eGFR, ZAG and cystatin C between the two groups ( Table 1).

Correlation of copeptin, ZAG, cystatin C, NGAL and L-FABP to clinical parameters in all patients
The correlations between copeptin, ZAG, cystatin C, NGAL or L-FABP and clinical parameters in LC patients are shown in Table 2.

Background comparison between Responders and Non-responders to tolvaptan
In this study, we excluded 7 from 45 cases for which tolvaptan e cacy could not be determined because of transferring hospitals, lack of weight data or with other treatment such as albumin transfusion, ascites puncture, and cell-free and concentrated ascites reinfusion therapy (Supplementary Figure 1). We enrolled 38 decompensated LC patients with ascites (24 males and 14 females) with a mean age of 67.1 ± 9.7 years. We divided the 38 patients into two tolvaptan treatment groups: Responders and Nonresponders. The Non-responder group was de ned as patients with weight loss of <1.5 kg/week after receiving tolvaptan or performing paracentesis within the rst week 12 . All patients continued to take the same prescribed doses of furosemide and spironolactone within the rst week. There were no obvious differences in age, gender, bodyweight, BMI, PMI, presence of HCC, dose of diuretic drugs, albumin, total bilirubin, Child-Pugh score, ALBI score, FIB4-index, MELD score, creatinine, eGFR, serum sodium, cystatin C, NGAL and L-FABP between the two groups (Table 3; Figure 2-D, E and F). In contrast, BUN, copeptin and ZAG levels were signi cantly higher in Non-responders when compared to the Responders group (respectively, p=0.0014, p=0.0265, p=0.0142) ( Table   3; Figure 2-A, B and C). We calculated the cutoff values, area under the ROC curve, sensitivity and specificity of BUN, copeptin and ZAG using ROC analysis. The cutoff values ascertained from our analyses of BUN, copeptin and ZAG were 18.3mg/dL, 10.1pmol/L and 32.4 μg/mL, respectively (Table 3).
Predictors contributing to the effect of tolvaptan in treatment for ascites Using multivariate logistic regression analysis, we found BUN (odds ratio 7.43, p=0.0306), serum copeptin levels (odds ratio 9.12, p=0.013) and ZAG (odds ratio 7.43, p=0.0306) to be the significant predictors contributing to the e cacy of tolvaptan in the treatment for ascites (Table 4).

Discussion
In this study, we revealed for the rst time that serum copeptin and ZAG levels were signi cantly increased in the tolvaptan treatment Non-responders group when compared to the Responders group. In addition, we indicated that serum BUN, copeptin and ZAG levels were independent predictors of the overall response to tolvaptan therapy. We further showed that serum creatinine and cystatin C levels were not predictive of patient response to tolvaptan, thus con rming the ndings reported by others 14,15 . In LC, arteriolar vasodilation causes under lling of the systemic arterial vascular space, and the decrease in the effective blood volume leads to a decrease in arterial pressure 21 . Consequently, activation of the renin-angiotensin-aldosterone system, sympathetic nervous system and the release of antidiuretic hormone take place as the body attempts to restore normal blood pressure homeostasis 22 . The main physiological function of AVP is antidiuresis, thereby regulating systemic osmotic pressure. Plasma AVP levels are normally regulated by plasma osmotic pressure and have been found to be elevated in LC patients with ascites, as reported by Pérez-Ayuso RM 23 . However, there has been no study to date exploring whether the effect of tolvaptan is related to plasma AVP or pro-AVP (copeptin) levels in ascites patients. The Non-responder group is considered to be in a state of relative vascular under lling with an increase in BUN and copeptin, thus suggesting intravascular dehydration 14 . Furthermore, serum copeptin also correlated with multiple factors including renal function and CRP, resulting in copeptin is not only a biomarker of renal function. Although NGAL and L-FABP are useful markers than copeptin for acute-on-chronic liver failure 24 and acute kidney injury 25 , respectively, this study indicated that serum NGAL and L-FABP were not useful in assessing the e cacy of response to tolvaptan in LC patients with ascites.
Serum ZAG, it has been determined, is increased in acute kidney injury 26 and in minor kidney injury caused by normo-albuminuric diabetic kidney disease -patients presenting with renal insu ciency, but no signi cant proteinuria 27 . On the other hand, serum cystatin C accurately re ects renal function, while ZAG re ects lipolysis and renal function. Serum ZAG has been expected as a biomarker for cachexia 28 . Indeed, in this study, ZAG correlated with body weight, BMI and CRP. Bellos also reported that body weight loss and elevated CRP were the predictors for Non-responders to tolvaptan 15 . We assumed that association of serum ZAG with multiple factors including renal function, cachexia (relating to body weight and BMI) and/or in ammation (relating to CRP) led to the result.
Based on the evidence outlined above, serum copeptin may reveal the underlying pathological condition of decreased osmotic pressure, intravascular dehydration and in ammation, whereas serum ZAG may re ect minor kidney injury, cachexia and in ammation, suggesting that copeptin and ZAG are independent predictors of patient response to tolvaptan. This study has several limitations. This was a retrospective and single-center study with a small sample size and a short follow-up period. Further study using a larger cohort is required to investigate the interaction of copeptin, ZAG, and BUN and the possible predictors of patient survival. In conclusion, serum BUN > 18.3 mg/dL, copeptin < 10.1 pmol/L and ZAG levels of < 32.4 µg/ml appear to be good predictors of overall patient response to tolvaptan treatment.

Conclusions
Serum copeptin, ZAG and BUN levels can be used as novel biomarkers to determine overall response to tolvaptan in patients presenting with LC and ascites. Future study is required to develop a new clinical prediction model using serum copeptin, ZGA and BUN that would help clinicians determine the effectiveness of tolvaptan treatment.

Declarations Acknowledgement
There was no acknowledgement.

Statement of Ethics
The study protocol was approved by the Clinical Research Ethics Review Committee of Mie University Hospital (approval no. H2019-190). This research comply with the guidelines for human studies and should include evidence that the research was conducted ethically in accordance with the World Medical Association Declaration of Helsinki. The subjects (or their parents or legal guardians) have given their written informed consent.

Con ict of Interest Statement
Authors does not have any con ict of interests, except Motoh Iwasa received a lecture fee from Otsuka Pharmaceutical Co., Ltd, which manufactures tolvaptan.