Association of loss of muscle mass with mortality in liver cirrhosis without or before liver transplantation: A systematic review and meta-analysis.

BACKGROUND: Liver cirrhosis is a risk factor for the loss of muscle mass, which is associated with numerous adverse health outcomes. This meta-analysis aimed to examine whether loss of muscle mass was a predictor of increased mortality in cirrhotic patients without or before liver transplantation.
METHODS: Without language restriction, PubMed and Embase were searched for articles published from the earliest records to December 2018 investigating the influence of loss of muscle mass on survival of cirrhotic patients. Those who had undergone liver transplantation and had hepatocellular carcinoma were excluded. The main outcome was the hazard ratio (HR) for the association of mortality with loss of muscle mass, and the secondary outcome was the association of loss of muscle mass with Child-Pugh class and death caused by severe infection.
RESULTS: The meta-analysis included 16 observational studies, comprising 4070 participants. The pooled crude and adjusted HRs for the association of mortality with loss of muscle mass were 2.05 (95% confidence interval [CI], 1.51-2.78) and 2.36 (95% CI, 1.61-3.46). Using Child-Pugh Class A as reference, the odds ratios (ORs) for the association of loss of muscle mass with Child-Pugh Class B and Class C were 1.68 (95% CI, 0.96-2.92) and 1.94 (95% CI, 0.66-5.65). Patients with loss of muscle mass were likely to have infection-related mortality (OR = 3.38, 95% CI, 0.61-18.88) but the association did not reach statistical significance.
CONCLUSIONS: Loss of muscle mass is associated with mortality in cirrhotic patients without or before liver transplantation. Future studies should be conducted to explore whether exercise and nutritional supplementation can reverse muscle mass loss and improve long-term survival.

Introduction

Liver cirrhosis is a diffuse parenchymal hepatic disease characterized by extensive fibrosis and formation of irreversible nodules. Cirrhosis results in derangement of the hepatic vascular architecture and leads to life-threatening complications such as portal hypertension, gastroesophageal varices, and hepatic encephalopathy. Alcoholism and hepatitis B and C are common causes of liver cirrhosis and also exert a substantial influence on long-term outcomes. The conventional prognostic systems for predicting survival in cirrhotic patients are the Child-Pugh and the Model of End-Stage Liver disease (MELD) Scores.[ The main limitation of both tools is the lack of nutritional assessment.[ Malnutrition is prevalent in the cirrhotic population and results from an altered metabolism and decreased protein intake.[ A compromised nutritional status is associated with a poor prognosis in patients with chronic liver disease and can be reflected in changes in body composition, such as loss of skeletal muscle mass[ (Table 1).

Table 1

Summary of the included studies.

Loss of skeletal muscle mass may be accompanied by decreased muscle strength and impaired physical performance,[ and is linked to several adverse health outcomes, including cognitive impairment,[ depression,[ and mortality.[ Cirrhotic patients are vulnerable to loss of muscle mass due to reduced nutritional intake required for muscle generation, increase in myostatin, which inhibits muscle growth, and abnormal consumption of protein for energy production.[ Recently, a meta-analysis explored the relationship between sarcopenia and mortality in cirrhotic patients.[ The study was composed of a variety of patient groups including those not eligible for liver transplantation, those on the transplantation waiting list, those who underwent liver transplantation, and with hepatocellular carcinoma. However, the heterogeneity in the severity of underlying disease and subsequent treatments might cloud the association between loss of muscle mass with long-term survival in cirrhotic patients. Therefore, this meta-analysis focused on cirrhotic patients without or before liver transplantation and examined whether loss of muscle mass was a predictor of increased mortality in the target population.

Methods

This systematic review was performed in accordance with the Preferred Reporting Items for Systematic Reviews and Meta Analyses and the Meta-analysis of Observational Studies in Epidemiology statements and was based on a predefined but unpublished protocol.[ As the work is a review article, no ethics committee or institutional review board approval is needed.

Data sources and literature search strategy

Two investigators independently searched electronic databases (PubMed and Embase) without language restriction for articles published from the earliest records to December 2018. Clinical studies investigating the association between loss of muscle mass and survival of patients with liver cirrhosis were included. The following strategy was used for literature searches: (sarcopenia, OR frailty, OR skeletal muscle) AND (liver cirrhosis OR liver fibrosis) (Appendix 1). Reference lists of the retrieved studies were also manually searched to identify relevant articles.

Inclusion and exclusion criteria

The inclusion criteria for this meta-analysis were use of a case-control or a cohort study design, recruitment of patients with liver cirrhosis, evaluation of skeletal muscle mass, with or without physical performance, and reporting of data on all-cause mortality. Studies were excluded if they were case series or case reports that included patients with hepatocellular carcinoma on the initial enrollment following liver transplantation, if they lacked a clear definition of loss of muscle mass based on the measurement of skeletal muscle mass and/or physical performance, or if a hazard ratio (HR) or an odds ratio (OR) analyzing the association between loss of muscle mass and mortality could not be computed from the available data.

Assessment of study quality

Two investigators independently investigated the quality of included studies using the Newcastle–Ottawa Scale.[ The following aspects were evaluated: participant selection (representativeness of patients and controls, validated measurements of exposure, and assurance of outcome of interest not present at the start of the research), study comparability, and outcome assessment (appropriateness of outcome measurement, adequate follow-up period, and adequate number of participants being followed up). Discrepancy of opinion between reviewers was resolved through discussion or was settled by the corresponding author.

Data synthesis and statistical analysis

The following details were extracted from the included studies: author name, year of publication, patient characteristics, enrolled numbers, sex ratio, methods of muscle mass measurement, definition of loss of muscle mass, outcome variables, and adjusted confounders. The main outcomes were the crude and adjusted HRs investigating the association between loss of muscle mass and mortality. The summary adjusted HR was derived from pooling each adjusted HR reported in the included studies.

Since the mortality rate differs in patients according to the severity of liver cirrhosis and is unlikely to be a fixed number, a random-effects model was applied to pool the effect sizes.[ A stratified analysis was conducted based on patient characteristics. Heterogeneity across studies was assessed using Cochran's Q test and I statistics and was considered significant with I > 50%.[ Publication bias was determined by visually inspecting the symmetry of the effect size distribution on funnel plots and using Egger's test.[ We also employed the Duval and Tweedie trim and fill procedure to observe the change in summary effects following imputation of potential unpublished literature.[ All calculations were performed using Comprehensive Meta-analysis Software version 3 (Biostat, Englewood, NJ), with a P < .05 considered statistically significant.

Results

Search results

The literature search initially yielded 989 non-duplicated articles. After the titles and abstracts were reviewed, 27 articles were retrieved for full text review. Among these, 11 were excluded: 5 used the same patient group as the included trials,[ 2 evaluated skeletal muscle mass loss in cirrhotic patients but lacked survival analysis,[ only 1 included cirrhotic patients with different rates of muscle mass loss,[ and 3 explored the association between loss of muscle mass and hepatic encephalopathy but not all-cause mortality.[ The final meta-analysis included 16 studies (Fig. 1).[

Figure 1

Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram for the study selection process.

Study and participant characteristics

The meta-analysis consisted of 4070 participants, 29.8% (n = 1215) of whom were females. The age range in the selected citations was between 49.7 and 74 years. The 16 retrieved studies included 7 that enrolled patients on the waiting list for liver transplantation,[ 5 that enrolled patients with different severities of liver cirrhosis,[ 1 that enrolled patients with concomitant ascites,[ 1 that enrolled patients with cirrhotic admitted for treatment of sepsis,[ 1 that enrolled patients with cirrhotic receiving transjugular intrahepatic portosystemic shunts,[ and 1 that enrolled patients with acute variceal bleeding. [ Based on the definition of loss of muscle mass, 13 retrieved studies used the cross-sectional area of the psoas muscle on computed tomography (CT), with or without adjacent paraspinal and abdominal muscle imaging,[ 1 used psoas muscle thickness assessed with CT,[ 1 used mid-arm muscle circumference and handgrip strength,[ and 1 used skeletal muscle mass estimated using bioelectrical impedance analysis (BIA).[ Most of the selected studies retrospectively analyzed the medical records from cohorts of liver cirrhotic patients to define loss of muscle mass. The result of quality assessment of the selected studies is shown in Table 2. The included studies fulfilled most of the items evaluated except for comparability of cohorts and adequate follow-up period.

Table 2

Quality assessment by using the Newcastle–Ottawa Scale for the included studies.

Loss of muscle mass and all-cause mortality

The crude HRs for the association between loss of muscle mass and all-cause mortality were reported in 14 of the 16 retrieved studies,[ while the adjusted HRs were available in 11.[ The pooled crude and adjusted HRs were 2.05 (95% confidence interval [CI], 1.51–2.78; I = 87.5, P < .001) and 2.36 (95% CI, 1.61–3.46; I = 89.4, P < .001), respectively (Fig. 2). Sensitivity analysis was performed by eliminating studies not using CT to estimate skeletal muscle mass,[ and the pooled crude HR changed to 2.11 (95% CI, 1.50–2.95; I = 89.0, P < .001). Another sensitivity analysis was conducted by excluding the studies only reporting in-hospital mortality,[ and the pooled crude HR and adjusted HR changed to 2.07 (95% CI, 1.48–2.90, I = 88.9, P < .001) and 2.28 (95% CI, 1.54–3.35, I = 89.9, P < .001).

Figure 2

Forest plot of the crude (A) and adjusted (B) hazard ratios for the association between loss of muscle mass and all-cause mortality in patients with liver cirrhosis.

We conducted a stratified analysis based on patient characteristics. In the group of cirrhotic patients on the waiting list for liver transplantation, the crude and adjusted HRs were 1.62 (95% CI, 1.09–2.39) and 1.86 (95% CI, 1.23–2.84), respectively. In the group not qualified or specified as candidates for liver transplantation, there was an insignificant increase in the crude and adjusted HRs (2.50, with 95% CI, 1.82–3.43 and 3.27, with 95% CI, 2.10–5.11, respectively). Significant publication biases were detected in the crude and adjusted HRs (both P < .001) with Egger's test (Fig. 3). Nevertheless, after imputation of potentially unpublished studies using the Duval and Tweedie trim and fill function, the level of statistical significance remained consistent with the pooled crude HR (1.68, 95% CI, 1.29–2.19) and adjusted HR (1.80, 95% CI, 1.33–2. 44) for the association between loss of muscle mass and all-cause mortality.

Figure 3

Funnel plot of the crude (A) and adjusted (B) hazard ratios for the association between loss of muscle mass and all-cause mortality in patients with liver cirrhosis.

Loss of muscle mass and Child-Pugh class

Among the 16 included studies, 3 had available data that could be used to estimate the association between loss of muscle mass and Child-Pugh class in patients with liver cirrhosis.[ Using Child-Pugh Class A as a reference, the ORs for the association of loss of muscle mass with Child-Pugh Class B and Class C were 1.68 (95% CI, 0.91–2.66; I = 0, P = .573) and 1.94 (95% CI, 0.66–5.65; I = 45.1, P = .162), respectively. Our analysis did not reveal significant associations between loss of muscle mass and Child-Pugh class (Fig. 4).

Figure 4

Forest plot of the association between loss of muscle mass and Child-Pugh Class in patients with liver cirrhosis.

Loss of muscle mass and infection

The number of deaths related to infection or sepsis was reported in 3 of the included studies.[ Mortality in patients with loss of muscle mass was likely to be associated with infection (OR = 3.38, 95% CI, 0.61–18.88; I = 0, P = .542) but the association did not reach statistical significance (Fig. 5).

Figure 5

Forest plot of the association between loss of muscle mass and infection-associated mortality in patients with liver cirrhosis.

Discussion

The present meta-analysis explored the association between loss of muscle mass and overall survival in liver cirrhotic patients. Loss of skeletal muscle mass was associated with an increase in all-cause mortality, and the association remained statistically significant after adjusting for common confounders. Loss of muscle mass could not be predicted by the severity of chronic liver disease, categorized by Child-Pugh class. Severe infection might play a role in the increased mortality in cirrhotic patients associated with loss of muscle mass.

Recently, loss of muscle mass has been recognized in several systematic reviews and meta-analyses to have a negative effect on health-related outcomes. Chang et al identified 10 studies that investigated the association of sarcopenia with mortality in a group mainly consisting of geriatric patients.[ Beaudart et al included 17 studies that defined sarcopenia by using the protocol from the European Working Group on Sarcopenia in Older People and found a higher rate of mortality, functional decline, falls, and hospitalization among sarcopenic adults.[ Shachar et al summarized 38 studies enrolling patients with solid tumors and demonstrated that the loss of skeletal muscle mass determined by CT cross-sectional imaging was detrimental to overall survival.[ In 2015 and 2016, Kim et al and Sinclair et al published 2 review articles, summarizing available evidence on the definition, etiology, prevalence, mechanism, and clinical impact of loss of muscle mass in patients with liver cirrhosis or following liver transplantation.[

Regarding the impact of sarcopenia in a population with chronic liver disease, Kim et al found that cirrhotic patients with sarcopenia tended to have higher mortality rates, regardless of the status of liver transplantation.[ Yu et al found that sarcopenia was associated with steatohepatitis or advanced liver fibrosis in patients with nonalcoholic fatty liver disease.[ Chang et al demonstrated that sarcopenia increased the mortality rate in hepatocellular carcinoma,[ as well as the incidence of hepatic encephalopathy in cirrhotic patients.[ However, a detailed pooled quantitative analysis exploring the association of loss of muscle mass with survival in patients with cirrhosis before or without liver transplantation has been lacking, increasing the importance of our meta-analysis. The comparison of the present and previous meta-analyses related to sarcopenia in patients with chronic liver disease is summarized in Table 3.

Table 3

Comparisons of the previous and present meta-analyses related to sarcopenia in patients with chronic liver disease.

Our study showed that loss of muscle mass was associated with all-cause mortality in patients with cirrhosis. Previous studies revealed that loss of muscle mass is prevalent in the population with liver cirrhosis, and the mechanism is multidimensional, comprising malnutrition, abnormal use of protein as an energy resource, increased production of proinflammatory cytokines to accelerate muscle breakdown, and inhibition of muscle growth through elevated myostatin and decreased testosterone.[ The literature has shown that loss of muscle mass was independently associated with severe hepatic fibrosis, a predictor of mortality in patients with non-alcoholic liver disease.[ Therefore, with a growing amount of evidence referring to loss of muscle mass as a byproduct of liver cirrhosis, it is anticipated that the loss of muscle mass varies according to severity of cirrhosis and is subsequently related to mortality.

The Child-Pugh and MELD scores are commonly used for predicting prognosis in chronic liver disease.[ The former assesses total bilirubin, serum albumin, prothrombin time, ascites, and hepatic encephalopathy while the latter evaluates serum bilirubin, serum creatinine, and the international normalized ratio for prothrombin time. However, neither measurement tool includes a detailed evaluation of physical performance, body composition, and nutritional status. Thus, these prognostic tools present some limitations, with decreased predictive power in patients with lower MELD scores and inclusion of subjective assessment (hepatic encephalopathy and ascites) in the Child-Pugh score.[ This meta-analysis explored the association between loss of muscle mass and the severity of disease as categorized by the Child-Pugh class (not the MELD score due to data availability). Our results revealed that loss of muscle mass was not associated with Child-Pugh class B and C. Therefore, loss of muscle mass should be treated as an adjuvant prognostic factor along with these commonly employed scores to predict the survival of patients with liver cirrhosis.

Our pooled analysis also revealed a likely higher rate of infection-related death in patients with loss of muscle mass. It is well-known that immune responses such as phagocytosis and migration of neutrophils, natural killer cell activity, and activation of the complement system are impaired in cirrhotic patients.[ The skeletal muscles are considered to be an endocrine organ secreting and mediating various kinds of cytokines. Several studies reported that patients with sarcopenia had an increase in interleukin-6 and tumor necrosis factor-α and a decrease in interleukin 10, and were more vulnerable to significant systemic inflammatory reactions, which might be elicited by infection.[ Another possible mechanism is that muscle mass loss reflects the status of malnutrition, a well-known factor in immunodeficiency.[ Although the exact association between muscle volume loss and severe infection is still unknown, this connection might partially explain why patients with cirrhosis might have a shorter survival period in association with loss of muscle mass.

Most of the included studies used CT images obtained at the L3 or L4 vertebral level to define skeletal muscle mass loss, while 1 study employed BIA and 1 employed mid-arm muscle circumference and handgrip strength. The BIA evaluates body composition by applying a low amplitude electrical current on the skin to estimate intra- and intercellular fluid. This estimation may be biased in patients with local or generalized edema, which is a common finding in patients with cirrhosis.[ In contrast, the use of CT imaging at a fixed vertebral level to compute the skeletal muscle area allows the measurements to be more comparable among studies and is also less influenced by edema and ascites. Nevertheless, sensitivity analysis performed by excluding studies not using CT yielded no significant change in the association between loss of muscle mass and mortality. In recent years, various imaging tools have been used in evaluation of sarcopenia in patients waiting for liver transplantation, including CT, magnetic resonance imaging, dual-energy X-ray absorptiometry and ultrasonography. Although different methods have been used across studies, sarcopenia remains an independent factor predicting the prognosis waiting for or receiving liver transplantation.[ Until now, there is lack of studies comparing the prognostic accuracy among various imaging tools. Therefore, when interpreting sarcopenia in the population with liver cirrhosis, we need to be cautious that different methods for measuring muscle mass loss might lead to different conclusions.

Another point worth noting is that there was an insignificant higher association between loss of muscle mass and mortality in patients who were not specific or qualified candidates for liver transplantation compared to those on the transplantation waiting list. Since there are strict enrollment criteria for liver transplantation,[ we speculated that a discrepancy existed between both groups in terms of cirrhotic severity, which leads to a difference in the magnitude of association. In addition, the association between liver transplant contraindications and malnutrition (i.e., active excessive alcohol intake and noncompliance with therapy) and more intensive management of malnutrition before liver transplantation are possible factors that may modify the relationship between sarcopenia and mortality in patients with liver cirrhosis.

The present study has several limitations. First, all the enrolled studies used retrospective analyses, which were more subject to selection bias. Second, not all the outcomes were reported in each retrieved article, which could lead to potential publication bias. Third, the causes of death were limited in the selected studies and infection was the only item universally recorded. Therefore, the mortality in patients with cirrhosis are associated with loss of muscle mass may be mediated by other pathologic processes, such as hepatic encephalopathy, but these could not be identified through this meta-analysis. Fourth, the cut-off values for lower muscle mass varied among the included articles and all diagnostic protocols lacked assessment of muscle strength and physical performance. Future prospective studies using the criteria proposed by the European Working Group on Sarcopenia in Older People[ or the Asian Working Group for Sarcopenia[ should be conducted to determine whether the predictive power is improved after employing a more comprehensive algorithm to diagnose loss of muscle mass. Fifth, our meta-analysis did not separately report mortality according to the duration of follow-up. As the majority of the included studies used retrospective analysis and lacked uniform follow-up periods, we were unable to specify whether the mortality data were based on short-term or long-term follow-up. Sixth, the etiology of liver cirrhosis is an important factor influencing mortality. However, we could not analyze the impact of etiology of liver cirrhosis on our results because the HR in each study was not reported separately according to etiology.

In conclusion, loss of muscle mass was associated with all-cause mortality in patients with liver cirrhosis without or before liver transplantation. Loss of muscle mass in this population could not be predicted by the severity of cirrhosis categorized by the Child-Pugh class. The increased mortality rate in patients with loss of muscle mass might be related to severe infection. Future prospective studies should be conducted to explore whether exercise and nutritional supplementation can reverse muscle mass loss and improve long-term survival in patients with liver cirrhosis.

Author contributions

Conceptualization: Ke-Vin Chang.

Data curation: Ke-Vin Chang.

Formal analysis: Ke-Vin Chang.

Investigation: Ke-Vin Chang, Jin-De Chen.

Methodology: Ke-Vin Chang, Der-Sheng Han.

Writing – original draft: Ke-Vin Chang, Wei-Ting Wu, Der-Sheng Han.

Writing – review & editing: Ke-Vin Chang, Kuo-Chin Huang, Der-Sheng Han.