The Relationship between Controlling Nutritional (CONUT) Score and Clinical Markers among Adults with Hepatitis C Virus Related Liver Cirrhosis

Aims: To identify the relationship between the Controlling Nutrition Status (CONUT) score and clinical parameters among adults with hepatitis C virus (HCV)-related liver cirrhosis (LC) (n = 264, 141 males and 123 females). Methods: The relationship between the CONUT score and clinical variables such as Child-Pugh classification were investigated. We also examined factors linked to poor nutritional state as determined by CONUT score. Results: According to the CONUT score, normal nutritional state was found in 57 patients, mild malnutrition state in 132, moderate malnutrition state in 68 and severe malnutrition state in 7. The CONUT score ranged from 0 to 9 (median = 2) in Child-Pugh A (n = 198), 0 to 10 (median = 6) in Child-Pugh B (n = 62) and 6 to 9 (median = 7.5) in Child-Pugh C (n = 4) (overall, p < 0.00001). Multivariate analysis revealed that FIB-4 index, branched-chain amino acid to tyrosine ratio and extracellular water to total body water ratio in bioimpedance analysis were significant for both CONUT score 2 or more, and 5 or more. FIB-4 index had the highest predictability for both CONUT score 2 or more and 5 or more among three parameters. Conclusion: The CONUT score well reflects liver functional reserve among adults with HCV-related LC. FIB-4 index can be useful for malnutrition.


Introduction
The liver is the target organ for the metabolism of three major classes of molecules (fat, protein and carbohydrate) and it exerts a unique role in carbohydrate metabolism by preserving glucose concentration levels within the normal range [1][2][3][4][5]. Hepatitis C virus (HCV) is a major cause of chronic hepatitis worldwide and is the leading cause of liver cirrhosis (LC) in Japan [6,7]. LC, which develops over a long period of time due to the chronic inflammation in the liver, is well accepted to be an end-stage form of chronic liver disease and is accompanied by numerous nutrition disorders [8]. Of these, protein-energy malnutrition (PEM) is one of the most common complications seen in LC patients, which is associated with high morbidity and mortality in LC patients [9][10][11][12][13]. Thus, appropriate nutritional evaluation is essential for the management of LC patients.

Study Design and Inclusion Criteria
We retrospectively analyzed a total of 264 patients with HCV-related LC who were admitted to Division of Hepatobiliary and Pancreatic disease, Department of Internal Medicine, Hyogo College of Medicine, Hyogo, Japan between February 2006 and November 2015. The diagnosis of LC was determined based on clinical data, including liver biopsy specimens, laboratory data, clinical characteristics of portal hypertension, and/or radiological imaging such as computed tomography and ultrasonography. All analyzed patients had available data for body composition analysis using bioimpedance analysis (BIA). All analyzed patients had detection of HCV antibody and had no evidence of concurrent hepatitis B virus infection, and no clear evidence of drug-induced or alcoholic liver disease.

Exclusion Criteria
We excluded patients with massive ascites requiring abdominal paracentesis from this study as body composition analyses in BIA can be challenging in LC subjects with severe fluid retention. In other words, body weight, skeletal muscle mass index (SMI) and body mass index in BIA can be overestimated in subjects with massive ascites [13]. HCC patients were also excluded from this analysis. SMI in BIA was defined as "appendicular skeletal muscle mass/(height (m)) 2 " [28] Upper-SMI was defined as "skeletal muscle mass of upper extremities/(height (m)) 2 ". Lower-SMI was defined as "skeletal muscle mass of lower extremities/(height (m)) 2 ".

Our Objectives and Ethical Approval
We aimed to elucidate the relationship between the CONUT score and liver function markers, liver fibrosis markers and body composition data. We also examined factors associated with poor nutritional state as determined by CONUT score using univariate and multivariate analyses. The current study protocol strictly adhered to all regulations of the Declaration of Helsinki, and was approved by the institutional review board of Hyogo College of Medicine, Nishinomiya, Hyogo, Japan (approval no. 2117). We have received written inform consent from all participant patients.

Statistical Analysis
For quantitative variables, the statistical analysis among groups was carried out using Student's t test, Mann-Whitney U test, Kruskal-Wallis test, Fisher's exact test or Spearman's rank correlation coefficient r s after assessing the normality of their distribution. Variables with p value < 0.05 in the univariate analysis were subjected into the multivariate analysis using the logistic regression analysis. Receiver operating characteristics (ROC) curve analysis and area under the ROC curve (AUC) results were presented along with the corresponding optimal cutoff point that maximized the sum of specificity and sensitivity, sensitivity and specificity. Data are presented as median (range) unless otherwise stated. Statistical significance was set at p < 0.05. Statistical analysis was carried out with the JMP 13 (SAS Institute Inc., Cary, NC, USA).

Relationship between the CONUT Score and Other Clinical Variables (Spearman's Rank Correlation Coefficient r s )
Relationship between the CONUT score and other clinical variables for all cases are presented in Table 2. Significant variables with positive correlation with CONUT score were ECW to TBW ratio, total bilirubin, aspartate aminotransferase (AST), FIB-4 index, hyaluronic acid and tyrosine. Significant variables with negative correlation with CONUT score were prothrombin time (PT), platelet count, triglyceride, branched-chain amino acid (BCAA) to tyrosine ratio (BTR) and BCAA concentration. The r s values and p values of those parameters are listed in Table 3. Univariate analysis identified nine factors to be significantly associated with the presence of CONUT score ≥ 2 (p < 0.05): ECW to TBW ratio, total bilirubin, PT, platelet count, triglyceride, AST, BTR, FIB-4 index and hyaluronic acid (Table 4). Multivariate analysis for the seven factors (Platelet count and AST were excluded, because FIB-4 index includes platelet count and AST [30][31][32][33]) showed that ECW to TBW ratio, FIB-4 index, and BTR were significant factors linked to CONUT score ≥ 2 ( Table 5). Odds ratios and 95% confidence intervals (CIs) of these factors are listed in Table 5.  Univariate analysis identified eight factors to be significantly associated with the presence of CONUT score ≥ 5 (p < 0.05): ECW to TBW ratio, total bilirubin, PT, platelet count, triglyceride, BTR, FIB-4 index and hyaluronic acid ( Table 6). Multivariate analysis for the seven factors (Platelet count was excluded, because FIB-4 index includes platelet count [30][31][32][33]) showed that ECW to TBW ratio, FIB-4 index, and BTR were significant factors linked to CONUT score ≥ 5 (Table 7). Odds ratios and 95% CIs of these factors are listed in Table 7.  When the continuous variables changed over the entire range. BTR; branched-chain amino acid to tyrosine ratio, ECW; extracellular water, TBW; total body water.
3.5. ROC Analyses for Predicting CONUT Score ≥ 2 or CONUT Score ≥ 5 in FIB-4 Index, BTR and ECW to TBW Ratio Since FIB-4 index, BTR and ECW to TBW ratio were significant predictors linked to the presence of both CONUT score ≥ 2 and CONUT score ≥ 5, we further performed ROC analyses for those factors.
Corresponding AUC, optimal cutoff point, sensitivity (%) and specificity (%) of FIB-4 index, BTR and ECW to TBW ratio for predicting CONUT score ≥ 2 or CONUT score ≥ 5 were listed in Table 8. For both predicting CONUT score ≥ 2, and CONUT score ≥ 5, FIB-4 index had the highest AUC among three parameters (Table 8 and Figure 2A-F). Data are expressed as number or median (range). ECW; extracellular water, TBW; total body water, SMI; skeletal muscle mass index, AST; aspartate aminotransferase, ALT; alanine aminotransferase, eGFR; estimated glomerular filtration rate, BTR; branched-chain amino acid to tyrosine ratio. When the continuous variables changed over the entire range. BTR; branched-chain amino acid to tyrosine ratio, ECW; extracellular water, TBW; total body water.

ROC Analyses for Predicting CONUT Score ≥ 2 or CONUT Score ≥ 5 in FIB-4 Index, BTR and ECW to TBW Ratio
Since FIB-4 index, BTR and ECW to TBW ratio were significant predictors linked to the presence of both CONUT score ≥ 2 and CONUT score ≥ 5, we further performed ROC analyses for those factors.

Discussion
To the best of our knowledge, this is one of the largest studies assessing the relationship between the CONUT score and clinical parameters in HCV-related LC patients. Although the CONUT scoring system may be established in surgically treated patients, we believe our current results are worth reporting as our data may shed some lights on the relationship between the CONUT score and clinical parameters in adult HCV-related LC patients.
In our results, the CONUT score well correlated with Child-Pugh classification and other liver functional parameters, which agree with previous reports, and the significant relationship between the CONUT score and liver functional reserve was well validated in HCV-related LC [18,34]. On the other hand, the multivariate analyses revealed that FIB-4 index, BTR and ECW to TBW ratio were independent predictors linked to the presence of both CONUT score ≥ 2 and CONUT score ≥ 5. These results denoted that FIB-4 index, BTR and ECW to TBW ratio are helpful for predicting malnutrition status in HCV-related LC patients.
It is of note that in the ROC analyses, FIB-4 index had the highest AUC for predicting both CONUT score ≥ 2 and CONUT score ≥ 5. Considering our results, in the clinical settings, in HCV-related LC patients with FIB-4 index > 5.60, some nutritional interventions may be recommended, and in HCV-related LC patients with FIB-4 index > 7.89, intensive nutritional therapies should be performed. While FIB-4 index is a well established liver fibrosis marker, this marker can also be a nutritional marker [30][31][32][33]. On the other hand, BTR reflects protein synthesis ability and its statistical significance in the multivariate analysis is not so surprising [34]. However, considering our ROC analyses, at least in HCV-related LC patients with BTR < 4.0, BCAA supplementation therapy should be considered [1,2]. ECW to TBW ratio reflects edematous status [29]. Its significance in the multivariate analysis suggest the importance of fluid management for HCV-related LC patients with malnutrition.
Advanced LC with overhydrated status can cause malnutrition and muscle wasting and lower SMI was expected to be linked to higher CONUT score, but such results were not obtained in our analysis. Overhydrated status may result in the overestimation of skeletal muscle mass and this may cause our current results [35,36].
Serum hyaluronic acid had the well correlation with the CONUT score, although it did not reach significance in the multivariate analysis. In our previous study, we reported that serum hyaluronic acid well predicts PEM in HCV-related liver disease [9]. In that study, serum hyaluronic acid level yielded the AUC (0.849) at an optimal cutoff value of 151.0 ng/mL for the presence of PEM [9]. While in this study, optimal cutoff points for the presence of CONUT score ≥ 2 and CONUT score ≥ 5 were 165 ng/mL and 295 ng/mL, respectively. Reviewing these results, PEM can develop even in patients with mild malnutrition status. In LC patients with higher CONUT score, clinicians should be aware of the presence of PEM.
Several limitations related to this study warrant mention. First, this is a retrospective observational study. Second, the study was based on a Japanese HCV-related liver disease population, and additional studies on different liver disease etiology and ethnic backgrounds are necessary to further validate and extrapolate to other backgrounds. Third, the numbers of Child-Pugh A, B or C patients were not well balanced for analysis. Fourth, dietary intake in our analyzed subjects were unclear, potentially leading to bias. However, our current results demonstrated that the CONUT score well correlates with liver function and laboratory parameters such as FIB-4 index and BTR are useful for predicting malnutrition as defined by the CONUT score.
In conclusion, the CONUT score well reflects liver functional reserve, and in particular, FIB-4 index, can be a useful marker for the presence of malnutrition in adult patients with HCV-related LC.