Association between Albumin-Bilirubin Grade and Non-Protein Respiratory Quotient in Patients with Chronic Liver Diseases

We sought to elucidate the relationship between albumin-bilirubin (ALBI) grade and non-protein respiratory quotient (npRQ) calculated by indirect calorimetry in chronic liver disease (CLD) patients (n = 601, median age = 63 years). Factors linked to npRQ < 0.85, which is reported to be an optimal cutoff point for the prognosis in liver cirrhosis (LC) patients, were also investigated using univariate and multivariate analyses. The median npRQ for all cases was 0.86. In total, 253 patients (42.1%) had npRQ < 0.85. The proportions of patients with npRQ < 0.85 in LC and non-LC patients were 51.9% (166/320) in LC patients and 31.0% (87/281) in non-LC patients (p < 0.0001). The median npRQ in ALBI grades 1, 2, and 3 for all cases were: 0.89, 0.85, and 0.82 (overall p < 0.0001). The proportions of patients with npRQ < 0.85 were 31.0% (71/229) in ALBI grade 1, 46.34% (152/328) in ALBI grade 2, and 68.18% (30/44) in ALBI grade 3 (overall p < 0.0001). In multivariate analyses of factors linked to npRQ < 0.85, ALBI grade 3 (p = 0.0095, hazard ratio = 3.242, ALBI grade 1 as a reference) was an independent predictor along with prothrombin time (p = 0.0139). In conclusion, ALBI grade can be a useful marker for npRQ in patients with CLDs.


Introduction
The liver plays a pivotal role in the nutrition metabolism and the storage of nutrients; it is deeply involved in the maintenance of serum blood glucose, gluconeogenesis from glycogen, amino acids and lactic acid, and the production of ketone bodies from fatty acids [1,2]. Liver cirrhosis (LC), which occurs after a long period of time due to persistent inflammation in the liver, is often complicated with protein-energy malnutrition (PEM) [1][2][3][4][5]. Protein malnutrition can be evaluated by serum albumin level, and energy malnutrition can be evaluated by testing the non-protein respiratory quotient (npRQ) using indirect calorimetry [6]. PEM is one of the representative complications seen in LC patients, and it can be linked to high morbidity and mortality in LC patients [1][2][3][4][5][6][7]. A previous report showed that protein malnutrition as defined by serum albumin < 3.5 g/dL and energy malnutrition as defined by npRQ < 0.85 were found in 61% and 43%, respectively, and PEM was found in 27% of LC patients

Patients
A total of 601 CLD patients with data for npRQ available were admitted to our hospital between October 2005 and August 2018, and they were analyzed in this study. All analyzed patients had liver histological data (F0-F4), and 594 patients (98.8%) had bioimpedance analysis (BIA) data (skeletal muscle index (SMI) or extracellular water (ECW) to total body water (TBW) ratio). ALBI score was calculated and classified into three grades (ALBI grades 1, 2, and 3) as reported previously [12].

Measurement of npRQ by Indirect Calorimetry
Carbohydrates, lipids, and proteins taken into the body are metabolized to produce energy. Nutrients that are components in the body are also broken down as needed and eventually metabolized as energy. In the process of this metabolic reaction, oxygen is consumed, and carbon dioxide, water, and heat are produced. When carbohydrates, lipids, and proteins are metabolized, the ratio between the amount of oxygen consumed and the amount of carbon dioxide produced is different, which is called "RQ" [25]. Carbon dioxide production per minute (VCO 2 ) and oxygen consumption per minute (VO 2 ) were tested by indirect calorimetry. Total urinary excretion of nitrogen (UN) was measured as described elsewhere [6,25]. npRQ was calculated using the following formula: npRQ = (1.44VCO 2 -4.890UN) divided by (1.44VO 2 -6.04UN) [6,25].

SMI and ECW to TBW Ratio Using BIA
The definition of SMI was "appendicular skeletal muscle mass (kg) divided by (height (m)) 2 " using BIA. Based on the current Japanese criteria, the definition of decreased SMM was: SMI < 7.0 kg/m 2 in men and < 5.7 kg/m 2 in women [26]. Based on the concept that excessive ECW leads to edematous state, extracellular fluid (ECF) status was defined as the ECW to TBW ratio. In healthy persons, ECW to TBW ratio can be maintained at a constant value (ECW to TBW ratio around 0.38). ECF excess was classified as follows: normal (ECW to TBW ratio < 0.390), mild overhydrated state (ECW to TBW ratio 0.390-0.399), and moderate to severe overhydrated state (ECW to TBW ratio ≥ 0.400) (Biospace Co. Ltd., Seoul, Korea) [27].
We examined the relationship between npRQ and ALBI grade and factors linked to npRQ < 0.85, which is reported to be an optimal cutoff point for the prognosis in LC patients, using univariate and multivariate analyses [6,23]. The ethics committee of our hospital (Hyogo College of Medicine Hospital) acknowledged the study (approval no. 1831). The protocol in the current study strictly observed the regulations of the Declaration of Helsinki.

Statistical Considerations
In continuous variables, in order to estimate between-group difference, Student's t test, Mann-Whitney U test, Spearman's rank coefficient r s , analysis of variance, or Kruskal-Wallis tests were employed as appropriate. Normality of data distribution to analyze the continuous variables was tested by the Shapiro-Wilk test. In categorical variables, in order to estimate between-group difference, Fisher's exact tests or Pearson χ 2 tests were employed as appropriate. Factors with p < 0.05 linked to npRQ < 0.85 in the univariate analysis were entered into the multivariate logistic regression analysis to select candidate parameters. Unless otherwise stated, data are presented as interquartile range (IQR). The threshold for statistical significance was set at p < 0.05. The JMP 14 (SAS Institute Inc., Cary, NC, USA) was employed to perform statistical analyses.

Correlation between npRQ and ALBI Score According to Liver Disease Etiology
Correlation between npRQ and ALBI score according to liver disease etiology was shown in Figure

Discussion
The npRQ calculated from indirect calorimetry represents the ratio of carbohydrate to fat oxidation, and lower npRQ value is reported to be an indicator of poor prognosis in patients with LC [6]. However, as mentioned above, indirect calorimetry is time-consuming and has high cost [23,24]. Not all clinicians are familiar with the indirect calorimetry used, and it therefore does not spread to the clinical settings. Thus, identifying markers related to npRQ level in patients with CLDs seems to be of clinical importance, while the ALBI grading system has spread to clinical settings due to its convenience of use and good predictability [13][14][15][16][17][18][19][20]. We therefore conducted the current analysis. The strong point in this study was its large sample size (n = 601). To the best of our knowledge, this is the first study examining the relationship between npRQ and ALBI grade in CLD patients. In our data, npRQ level was well stratified according to ALBI grade for all cases, although such tendencies were not observed in several subgroups. In addition, ALBI score significantly correlated with npRQ value, and ALBI grade was an independent predictor linked to npRQ < 0.85 along with PT in the multivariate analysis. The correlation coefficient between ALBI score and npRQ (r s = −0.2476) was stronger than that between Child-Pugh score and npRQ (r s = −0.2003). These results denoted that the ALBI grading system can be helpful for predicting npRQ to some extent. The ALBI grading system includes serum albumin [12]. Saito et al. reported that supplementation with BCAA granules improved energy metabolism and serum albumin level after radiofrequency ablation therapy in HCC patients [28]. Our results can be associated with their results. In non-LC patients, however, the ALBI grading system cannot be helpful for predicting npRQ. The fact that most non-LC patients had normal albumin and bilirubin levels may be linked to our results. In the comparison of npRQ level among ALBI grades 1, 2, and 3 in HCC patients, the overall P value also did not reach significance. In patients with cancer, tumor-secreted or tumor-host interactions cause imbalance of energy requirements and energy uptake [29]. In HCC patients, the presence of HCC itself may be associated with energy metabolism regardless of the severity of ALBI grade [30][31][32][33].
Hanai et al. reported that plasma FFA level was significantly correlated with npRQ value (r = −0.39, p < 0.0001) [11]. In our subgroup analyses, according to liver disease etiology, the strongest correlation between ALBI score and npRQ was found in patients with AIH or PBC (r s = −0.5290, p < 0.0001), which had a stronger impact on npRQ than their data [11]. In autoimmune liver diseases, ALBI grade may be helpful for predicting energy malnutrition.
As shown in Figure 1A, the difference of npRQ level in patients with F0-2 and F3 did not reach significance, whereas npRQ levels in patients with F3 and F4 were quite different (p < 0.0001). This appears to be an important message for clinicians. Additionally, the proportion of patients with npRQ < 0.85 in LC patients (51.9%, 166/320) was significantly higher than that in non-LC patients (31.0%, 87/281) (p < 0.0001). These were expected results; more importantly, more than 30% non-LC patients had npRQ < 0.85. Clinicians should be aware of these, although a decrease in npRQ level in non-LC patients was not be able to be revealed by ALBI grade, as shown in Figure 3B. PEM was identified in 97 patients (30.3%) in our LC patients. A previous report presented that PEM was observed in 27% of LC patients (n = 294), which was similar to our data [8].
Patients with higher SMI may exercise well, and such patients may have higher npRQ value; however, SMI did not affect npRQ in our data [9]. The clinical implication of muscle mass or muscle function on energy metabolism in patients with CLDs should be further investigated in future studies. On the other hand, npRQ level was well stratified according to the edematous state defined by ECW to TBW ratio, as demonstrated in Figure 1C. Additionally, ECW to TBW ratio had the significant negative correlation with ALBI score (r s = −0.1261, p = 0.0021), although it did not reach significance in the univariate analysis linked to npRQ < 0.85. Body composition analysis using BIA is particularly attractive because of its noninvasiveness [34][35][36]. ECW to TBW ratio using BIA may provide useful information for clinicians [37].
Several limitations warrant mention in this study. First, this study was a single-center cross-sectional study with a retrospective nature. Second, the number of patients with ALBI grade 3 was relatively small for analysis, although the major strength of our study was its large sample size (n = 601). Third, npRQ level can vary according to patient activities, creating bias. Fourth, the correlation between npRQ level and ALBI score was relatively weak, although it reached statistical significance. Caution should therefore be applied for the interpretation of our data. Nevertheless, our study results demonstrated that npRQ level in CLDs was closely associated with ALBI grade. In conclusion, ALBI grade can be a useful marker for npRQ in patients with CLDs. In CLD patients with severe ALBI grade, presence of poor npRQ should be noted.