Serum Zinc Concentration and Sarcopenia: A Close Linkage in Chronic Liver Diseases

We sought to investigate the influence of serum zinc (Zn) concentration on sarcopenia in chronic liver diseases (CLDs, n = 372, median age = 65 years, 147 liver cirrhosis (LC) cases (39.5%)). Sarcopenia was defined by low grip strength and low skeletal muscle mass. Study subjects were divided into the following three groups (High-, Intermediate-, and Low-Zn groups) based on the baseline serum Zn level. The impacts of serum Zn concentration on sarcopenia were examined. The median (interquartile range) serum Zn concentration for all cases was 72.85 (63.7, 81.45) μg/dL. The proportions of sarcopenia in the High-Zn, Intermediate-Zn, and Low-Zn groups were 10.75% (10/93), 11.23% (21/187), and 27.17% (25/92), respectively (P = 0.9046 (High vs. Intermediate), P = 0.0007 (Intermediate vs. Low), P = 0.0044 (High vs. Low), overall P value = 0.0009). The median serum Zn concentrations in patients with sarcopenia, pre-sarcopenia, and control were 66.35, 73.1 and 73.8 μg/dL, respectively (P = 0.0234 (sarcopenia vs. pre-sarcopenia), P = 0.2116 (pre-sarcopenia vs. control), P = 0.0002 (sarcopenia vs. control), overall P value = 0.0016). In the multivariate analyses of factors linked to the presence of sarcopenia, Low-Zn was an independent predictor for all cases (P = 0.0236) and LC cases (P = 0.0082). In conclusion, Zn deficiency can be an independent predictor for sarcopenia in patients with CLDs.


Introduction
Zinc (Zn) is an important trace element that is needed for normal cell development, proliferation, and differentiation, and it is also known to be crucial to ensure an appropriate immunological reaction, such as anti-inflammatory effects, anti-oxidant effects, or autophagy [1][2][3][4]. Zn deficiency can cause a wide spectrum of clinical presentations, including appetite loss, body hair loss, impaired taste and smell, atrophy of testis, cerebral and immune dysfunction, and impairment of drug excretion ability, and they are frequently observed in chronic liver diseases (CLDs) as Zn homeostasis is primarily regulated in the liver [3,[5][6][7][8][9]. Albumin synthesis disability can cause Zn deficiency in patients with liver cirrhosis (LC) [3,10,11]. Serum Zn concentrations had the inverse correlation with serum ammonia levels in LC patients [12][13][14]. The degree of Zn deficiency is reported to correlate well with the severity

Patients
A total of 378 CLD patients with data for grip strength (GpS), SkMM using bioimpedance analysis (BIA), and serum Zn concentration were admitted to our institution between November 2013 and August 2018. Because overestimates could occur for the calculation of skeletal muscle mass index (SMI) using BIA in patients with massive ascites, 6 subjects with massive ascites were excluded from the study [19]. Three-hundred and seventy-two patients were consequently included in the current analysis. CLD was determined according to patient medical record, laboratory data, histological findings, and imaging findings. LC diagnosis was also based on histological findings (F4) or imaging findings (presence of varices, deformity of the liver, splenomegaly, etc.).

Measurement of GpS and SMI for Evaluating Sarcopenia
GpS was measured according to the current Japanese guidelines [18]. SMI was defined as "appendicular SkMM/(height (m)) 2 " using BIA. According to the current Japanese guidelines, patients with decreased GpS were defined as those with GpS <26 kg for male and <18 kg for female. Similarly, patients with decreased SkMM were defined as those with SMI <7.0 kg/m 2 for male and <5.7 kg/m 2 for female [18]. In males, patients with GpS <26 kg and SMI <7.0 kg/m 2 were classified as having sarcopenia, those with GS <26 kg and SMI ≤7.0 kg/m 2 as pre-sarcopenia, those with GS ≤26 kg and SMI <7.0 kg/m 2 as pre-sarcopenia, and those with GS ≤26 kg and SMI ≤7.0 kg/m 2 as control. In female, patients with GS <18 kg and SMI <5.7 kg/m 2 were classified as sarcopenia, those with GS <18 kg and SMI ≤5.7 kg/m 2 as pre-sarcopenia, those with GS ≤18 kg and SMI <5.7 kg/m 2 as pre-sarcopenia, and those with GS ≤18 kg and SMI ≤5.7 kg/m 2 as control.
Firstly, the impacts of serum Zn concentration on sarcopenia were examined for all cases and several subgroups according to the LC status. Secondly, factors associated with the presence of sarcopenia were studied using univariate and multivariate analyses. Correlation between serum Zn concentration and baseline characteristics was also examined. The ethics committee of our hospital acknowledged this study (number: 2296). The protocol in the study strictly observed all regulations of the Declaration of Helsinki.

Statistical Considerations
As for continuous parameters, Student's t test, Mann-Whitney U test, Pearson's correlation coefficient r, analysis of variance, or Kruskal-Wallis test were employed to assess group difference, as applicable. In categorical parameters, Pearson χ 2 test was employed to assess group difference, as applicable. Factors with P < 0.05 linked to the presence of sarcopenia in the univariate analysis were subjected to the multivariate logistic regression analysis to identify candidate parameters. Baseline factors significantly correlated with serum Zn concentration in the univariate analysis were also subjected to the multivariate logistic regression analysis to identify candidate parameters. Unless otherwise mentioned, data were indicated as median values (IQR). The threshold for statistical significance was set at P < 0.05. The JMP 13.2 (SAS Institute Inc., Cary, NC, USA) was employed to carry out statistical analysis.

Patient Baseline Characteristics
Baseline characteristics in our study (n = 372, 171 males and 201 females, median age (IQR) = 65 (54, 71.75) years) were shown in Table 1. In terms of liver disease etiology, hepatitis C virus was in the majority (60.2%, 224/372). The median (IQR) serum Zn concentration for all cases was 72.85

Uni-and Multivariate Analyses of Factors Associated with the Presence of Sarcopenia for All Cases (n = 372)
In all cases, univariate analysis identified nine factors to be significantly associated with the presence of sarcopenia: age (P < 0.0001), body mass index, (BMI, P < 0.0001), presence of LC (P = 0.0416), total bilirubin (P = 0.0276), serum albumin (P = 0.0008), alkaline phosphatase (ALP, P = 0.0143), estimated glomerular filtration rate (eGFR, P = 0.0178), serum sodium (P = 0.0164), and our classification of serum zinc concentration (P = 0.0009) ( Table 2). Multivariate analysis for the nine factors showed that age (P = 0.0005), BMI (P = 0.0003), serum albumin (P = 0.0011), and L-Zn group (P = 0.0236, I-Zn group as a reference) were found to be significant factors linked to the presence of

Uni-and Multivariate Analyses of Factors Associated with the Presence of Sarcopenia for All Cases (n = 372)
In all cases, univariate analysis identified nine factors to be significantly associated with the presence of sarcopenia: age (P < 0.0001), body mass index, (BMI, P < 0.0001), presence of LC (P = 0.0416), total bilirubin (P = 0.0276), serum albumin (P = 0.0008), alkaline phosphatase (ALP, P = 0.0143), estimated glomerular filtration rate (eGFR, P = 0.0178), serum sodium (P = 0.0164), and our classification of serum zinc concentration (P = 0.0009) ( Table 2). Multivariate analysis for the nine factors showed that age (P = 0.0005), BMI (P = 0.0003), serum albumin (P = 0.0011), and L-Zn group (P = 0.0236, I-Zn group as a reference) were found to be significant factors linked to the presence of sarcopenia (Table 3). Hazard ratios (HRs) and 95% confidence intervals (CIs) for these variables were listed in Table 3.

Uni-and Multivariate Analyses of Factors Associated with the Presence of Sarcopenia for LC Cases (n = 147)
In LC-cases, univariate analysis identified four factors to be significantly associated with the presence of sarcopenia: age (P = 0.0026), BMI (P = 0.0010), serum albumin (P = 0.0035), and our classification of serum zinc concentration (P = 0.0016) ( Table 4). Multivariate analysis for the four factors showed that age (P = 0.0363), BMI (P = 0.0014), and L-Zn group (P = 0.0082, I-Zn group as a reference) were found to be significant factors linked to the presence of sarcopenia (Table 5). HRs and 95% CIs for these variables were listed in Table 5.

Uni-and Multivariate Analyses of Factors Associated with the Presence of Sarcopenia for Non-LC Cases (n = 225)
In non-LC cases, univariate analysis identified four factors to be significantly associated with the presence of sarcopenia: age (P < 0.0001), BMI (P = 0.0003), eGFR (P = 0.0063), and HbA1c (P = 0.0091) ( Table 6). Multivariate analysis for the four factors showed that age (P = 0.0249) and BMI (P = 0.0021) were found to be significant factors linked to the presence of sarcopenia (Table 7). HRs and 95% CIs for these variables were listed in Table 7.

Correlation between Serum Zinc Concentration and Baseline Characteristics
Correlation coefficients and P values between serum zinc concentration and baseline characteristics were listed in Table 8. Serum albumin level had the strongest positive correlation with serum Zn concentration (r = 0.654, P < 0.0001) followed by branched-chain amino acid to tyrosine ratio (r = 0.496, P < 0.0001), while FIB-4 index had the strongest negative correlation with serum Zn concentration (r = −0.360, P < 0.0001) followed by total bilirubin (r = −0.355, P < 0.0001). Multivariate analysis of factors significantly correlated with serum Zn concentration revealed that only serum albumin level was a significant factor linked to serum Zn concentration (P < 0.0001).

Discussion
Chronic liver injury results in the impaired Zn homeostasis and eventually Zn deficiency, and most CLD patients have Zn deficiency [3][4][5]10,13,14,32]. Zn deficiency can present numerous symptoms, which can be associated with decreased QOL and worse clinical outcomes [3,5,33]. However, as mentioned earlier, there is little data currently available regarding relationship between serum Zn concentration and sarcopenia in CLDs. To clarify these issues is clinically of great importance due to the increasing interest for sarcopenia in CLDs these days. To the best of our knowledge, this is the first study elucidating the association between serum Zn level and sarcopenia in CLDs using large cohort (n = 372).
First of all, the validity of our classification of H-Zn, I-Zn, and L-Zn groups needs discussion. As mentioned above, CLD patients have tendency for lower serum Zn levels and only 111 of our analyzed subjects (29.8%) had normal serum Zn value (reference range, 80 µg/dL ≤serum Zn concentration ≤130 µg/dL in our institution) [3,5,7]. Therefore, we classified them into three groups based on quartiles in our cohort for the appropriate analysis and investigated the impact of serum Zn concentration on sarcopenia. In our analysis, the proportion of sarcopenia in the L-Zn group was significantly higher than that in the I-Zn group or the H-Zn group. Additionally, our multivariate analyses revealed that L-Zn was an independent predictor linked to the presence of sarcopenia for all cases and LC cases. These results suggest the clinical significance of serum Zn concentration for sarcopenia in CLD patients, even considering the non-significant relation between serum Zn concentration and sarcopenia in non-LC cases. Zn deficiency induces a number of physiological problems, which can be associated with sarcopenia [34]. In CLD patients, close monitoring for serum Zn concentration may be required for identifying sarcopenic subjects. While patients with pre-sarcopenia had similar serum Zn concentration to those with control for all cases, LC cases, and non-LC cases. Reviewing these results, sarcopenia staging (i.e., distinguishing between sarcopenia and pre-sarcopenia) seems to be an important strategy for CLD patients. A recent study reported that a stepwise increase of the proportion of fallers according to the severity of sarcopenia was observed in older populations, which was in line with our current results [35]. On the other hand, appropriate timing of Zn supplementation therapy should also be considered in CLD patients from the clinicians' perspective. In addition, whether Zn supplementation therapy in sarcopenic CLD patients can reverse sarcopenia needs to be clarified in future studies.
In comparison between L-Zn group and H-Zn group, the impact on sarcopenia was not prominent in the multivariate analysis for all cases (HR = 1.672, P = 0.3881, H-Zn as a reference). Although the reasons for these are unclear, maintaining serum Zn concentration in a certain range may be essential for avoiding sarcopenia in CLD patients.
Serum albumin level had the strongest positive correlation with serum Zn concentration (r = 0.654, P < 0.0001) and it was an independent predictor associated with serum Zn concentration in the multivariate analysis, while FIB-4 index had the strongest negative correlation with serum Zn concentration (r = −0.360, P < 0.0001) in our data. Because protein synthesis ability reflects serum Zn concentration, these results are not so surprising [3,5,7]. More importantly, age was correlated inversely with serum Zn concentration, although the r value was not impressive (r = −0.186, P = 0.0003). In aging process, a lot of mediators can be associated with inflammation and oxidative stress, and imbalance of Zn homeostasis is a common hallmark of aging [36][37][38]. The high prevalence of hypozincemia in our LC patients (84.3%) may be partly attributed to the higher age of our LC patients (median age = 68.0 years), as well as protein synthesis inability and other metabolic disorders [3,5,7]. On the other hand, clinicians should be aware that even in non-LC patients, a considerable number of patients with hypozincemia are present (60.9% in our non-LC patients).
Both age and BMI were independent predictors for all cases, LC cases, and non-LC cases in our multivariate analyses. CLD patients appears to suffer from sarcopenia based on the following two types of sarcopenia: aging-related sarcopenia (primary sarcopenia) and CLD specific abnormalities-related sarcopenia (secondary sarcopenia) [18]. Our current results may be attributed to the increase of aging CLD patients in our country [39][40][41]. As CLD patients age, the incidence of physical performance limitation will increase and sarcopenia may be an inevitable consequence with aging [42]. While lower BMI was shown to have strong influence on sarcopenia in our CLD patients, the impact of obesity on sarcopenia (sarcopenic obesity) cannot be ignored, since there were only 3 patients (0.8%) with BMI >35 kg/m 2 in our study subjects [43,44].
We acknowledge several limitations to this study. First, the study was a retrospective single-center observational study without non-CLD patients as a control group. Second, GpS and serum Zn concentration can vary depending on patient daily life activities or dietary habits. Third, patients with massive ascites potentially involved in liver disease-related sarcopenia were excluded because of the lack of reliability in BIA, creating bias. This was the major limitation for the evaluation of muscle mass using BIA compared with using computed tomography or magnetic resonance imaging, however, BIA may be cost-effective. Fourth, data for other trace elements such as calcium, iron, phosphorus, magnesium, and selenium were not included in this study, leading to bias. Fifth, due to the small number of cases with sarcopenic obesity in our study, the relationship between serum Zn concentration and sarcopenic obesity remains unknown. Finally, it was uncertain as to whether sarcopenia caused Zn deficiency or whether Zn deficiency caused sarcopenia in this cross-sectional study. Consequently, caution should be exercised for the interpretation of the current study data and larger prospective studies will be needed to confirm these results. Despite these limitations, our study results denoted that Zn deficiency in CLDs was closely associated with sarcopenia.

Conclusions
In conclusion, Zn deficiency can be an independent predictor for sarcopenia in patients with CLDs.