Clinical significance of sarcopenia in the treatment of patients with primary hepatic malignancies, a systematic review and meta-analysis

Background The impact of sarcopenia on outcomes following treatment for primary liver tumors remains contentious. Therefore, we performed a systematic literature review and meta-analysis to evaluate the clinical significance of sarcopenia in the treatment of patients with primary liver tumors. Data sources A systematic literature search was performed in English through February 1, 2017 in databases. Results There were significant differences between patients with and without sarcopenia in overall 1- and 3-year survival (1 year: OR: 0.43; 95% CI: 0.27-0.68; P=0.0004; 3 year: OR: 0.67; 95% CI: 0.47-0.96; P=0.03). However, overall 5-year survival showed no significant difference between the groups (OR: 0.61; 95% CI: 0.35-1.07; P=0.08). Patients with sarcopenia showed a significant 53% reduction in disease-free survival within 5 years (OR: 0.47; 95% CI: 0.28-0.79; P=0.005). Also, sarcopenia had a significantly negative impact on recurrence in patients with primary liver tumors (RR: 2.71; 95% CI: 1.46-5.05; P=0.002). Regarding complications rate, we concluded that there was a statistically significant difference between two groups in overall complications rate (RR: 2.52; 95% CI: 1.50-4.22; P=0.0005). However, the major complications rate showed no significant difference between the groups (RR: 1.19; 95% CI: 0.65-2.20; P=0.57). Conclusions Sarcopenia seemed to have a negative effect on overall survival in patients with primary liver tumors in the early phase post-treatment, but further research is needed to investigate the prognostic impact on overall survival over the longer term. Moreover, sarcopenia could significantly increase the incidence rates of post-treatment recurrence and overall complications in patients with primary liver tumors.


INTRODUCTION
Sarcopenia, which is defined as the loss of muscle mass and function [1], is critically involved not only in aging but also in a variety of chronic diseases, such as tuberculosis infection, chronic obstructive pulmonary disease (COPD), diabetes mellitus (DM), advanced organ failure and other wasting conditions. In recent years, studies have shown that sarcopenia is prevalent in patients with certain cancers [2,3], and evidence has revealed its prognostic significance in oncologic patients [4][5][6][7][8][9][10]. A study by Chindaprasirt J [11] demonstrated that patients who suffered from sarcopenia had shorter median overall www.impactjournals.com/oncotarget/ Oncotarget, 2017, Vol. 8, (No. 60), pp: 102474-102485 Meta-Analysis www.impactjournals.com/oncotarget survival among cancer patients. More specifically, the frequency of sarcopenia varied greatly from 11.1% to 76% in patients with primary liver tumors (hepatocellular carcinoma (HCC) and intrahepatic cholangio-carcinoma (ICC)) [12][13][14], and the demonstration of the prognostic significance of sarcopenia has been reported relatively frequently in patients with primary liver tumors. To the best of our knowledge, several reviews [15][16][17] have shown that sarcopenia can be associated with impaired overall survival and increased postoperative morbidity in primary liver tumors. However, these studies have had several limitations. First, the reviews included studies evaluating different tumor types (including gastrointestinal and hepatopancreatobiliary malignancies). Second, the relationships between sarcopenia and the recurrence of malignancy or post-treatment complications (overall complications and major complications) were not described.
To date, it has been frustrating that there have been no systematic reviews specifically identifying sarcopenia in people with primary liver tumors, and the impact of sarcopenia on outcomes following treatment for primary liver tumors remains contentious. This systematic literature review and meta-analysis were performed to evaluate the clinical significance of sarcopenia in the treatment of patients with primary liver tumors. Outcomes of interest included overall survival, disease-free survival, recurrence of malignancy and post-treatment complications (overall complications and major complications).

Literature selection
Database searches yielded 71 entries, of which 52 were excluded because of duplications (29 trials) or irrelevance (23 trials). Of the 19 publications that qualified for abstract review, 1 was excluded because it was not a comparative trial, 1 trial was excluded because it included other malignancies than primary liver tumors, 4 trials were excluded because sarcopenia was not evaluated by the L3 skeletal muscle index (L3 SMI) or L3 total psoas area (L3 TPA), 2 trials were excluded because necessary data were not available, 2 trials were excluded because medical treatment was evaluated, and 1 trial was excluded because it was about postoperative sarcopenia. The Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) flow diagram for study selection is shown in Figure 1.

Included trials and quality assessment
Detailed characteristics of the included patients are listed in Table 1. Of these 8 comparative trials, the Newcastle-Ottawa Scale (NOS) [18] was applied to determine the risk of bias in this literature. According to the scores (Table 2), all the included trials were considered high-quality (score ≥7) trials.

Post-treatment complication rate
Complications were evaluated according to the Clavien-Dindo classification, and major complications (MCs) were defined as a Clavien-Dindo grade greater than 3. In our study, the impact of sarcopenia on overall complications (OCs) and major complications was examined.

DISCUSSION
Numerous studies have demonstrated that there are correlations between sarcopenia and patients' basic characteristics, such as age [20-22, 24, 26, 27], sex [4,12,23,27], liver function(serum albumin levels) [4,22,23], and BMI values [4,12,19,[22][23][24][26][27][28]. As reported, sarcopenia occurs during aging, and in patients with primary hepatic malignancies, this is possibly due to cachexia-associated processes. Accordingly, in patients with primary hepatic malignancies, sarcopenia is prevalent across all ages but particularly in the elderly [20-22, 24, 26, 27]. Moreover, because of faster deterioration of  muscle mass in men than in women, at older ages, men experience greater losses of muscle mass than women [29,30]; thus, with regard to sex differences, sarcopenia is more common in men than women [4,12]. However, a study by Yabusaki et al. [23] showed that the results for sex differences seemed to contradict the present findings. Furthermore, patients with sarcopenia have significantly lower BMI values than patients without sarcopenia, and the serum albumin levels of these patients are generally lower [4,22,23]. In limited studies, sarcopenia was correlated with number of tumors [23], micro-vascular invasion [24] and tumor stage [24], which would seem to contradict the present findings. We found no correlations between sarcopenia and tumor size, Child-Pugh grade, Model for End-stage Liver Disease (MELD) score, liver cirrhosis, Barcelona Clinic Liver Cancer (BCLC) classification or tumor differentiation. There have been no reports concerning the relationships between sarcopenia and lymph node metastasis, distant metastasis or positive surgical margin. Additionally, skeletal muscle protein breakdown occurs mainly because of age-related low-grade systemic inflammation, alongside physical inactivity and malnutrition in the elderly. This low-grade systemic inflammation is characterized by elevated proinflammatory cytokines and is caused by age-related cell damage and mitochondrial dysfunction [3]. However, there have been no reports concerning the relationship between sarcopenia and white blood cell counts, neutrophil-tolymphocyte ratios (NLRs) or C-reaction protein (CRP). Sarcopenia has been revealed to be an unfavorable prognostic factor for clinical outcomes in cancer. However, whether this finding also holds true in patients with primary hepatic malignancies has increasingly been explored recently. Unfortunately, this research field has been severely hampered by a lack of consensus on the definition of sarcopenia, and the definition of sarcopenia has varied even among studies conducted by researchers in the same country [19,21,22]. Consequently, the prevalence of sarcopenia has varied widely, from 11.1% to 76% across published studies. CT scanning is the gold standard tool to quantify skeletal muscle mass [31] and hence constitutes a good resource for objective identification of sarcopenia. To avoid intrinsic bias, we chose trials in which sarcopenia was defined using a L3 SMI/TPA cut-off rather than intramuscular adipose tissue content (IMAC) or visceral fat area (VFA) in our meta-analysis. However, the included studies' evaluations of different cut-off values for sarcopenia undoubtedly led to bias between the associations of sarcopenia with clinical outcomes.
In recent years, the clinical significance of sarcopenia has been reported relatively frequently in patients with primary hepatic malignancies, and concerns about the impact of sarcopenia on outcomes following treatment could not be confirmed in various studies. Levolger et al. [19] showed that sarcopenia was associated with impaired survival (P=0.002) in patients with HCC, and MCs were more frequent (P=0.033) in sarcopenic patients. However, a study by Yabusaki et al. [23] showed that there was no significant correlation between OS (P=0.72), and there was no difference in the incidence of postoperative MCs (P=0.62) between the two groups. The difference might have derived from the inclusion criteria because Levolger et al. included patients with hepatocellular carcinoma undergoing treatment with curative intent (hepatectomy or RFA), while Yabusaki et al. enrolled patients who underwent primary hepatectomy for hepatocellular carcinoma. Also, the relatively small number of participants could have biased the results. Our systematic review and meta-analysis confirmed that patients with sarcopenia had a statistically significantly poorer prognosis (overall 1-and 3-year survival, 5-year DFS, incidence of recurrence and post-treatment OCs) than those without sarcopenia. In contrast, there were no significant differences in overall 5-year survival, 1-and 3-year DFS or post-treatment MC. Yabusaki et al. [23] reported that sarcopenia impairs OS, mainly due to an increase in treatmentrelated deaths (P=0.029). Evidence has shown that sarcopenic patients with primary hepatic malignancies seem to be less able to recover from treatment. Although the mechanisms by which sarcopenia affects patients' recovery are not fully understood, differences in the severity of the underlying liver disorders impacting skeletal muscle mass might very well play a part.
Studies have shown that patients with sarcopenia are more susceptible to bacterial infection, slow wound healing and longer length of hospital stays [32,33], which might explain why sarcopenia has an adverse impact on short-term overall survival, consistent with our study that patients with sarcopenia had poorer overall 1-and 3-year survival but better overall 5-year survival. However, the available data from our meta-analysis suggested that sarcopenia increases the incidence of recurrence significantly with poor 5-year DFS. As reported, recurrence of HCC after curative treatment remains a major challenge for the management of liver malignancy, and recent studies have widely shown that sarcopenia was a strong and independent prognostic factor for recurrence after liver resection for (c) www.impactjournals.com/oncotarget HCC [22,34], which was in agreement with our finding. Therefore, whether the reduction in OS found in patients with sarcopenia is due to increased recurrent malignancy or due to increased treatment-related mortality remains controversial. Combining the evidence revealed by Yabusaki et al. [23] that sarcopenia impairs OS mainly due to an increase in treatment-related deaths, we can conclude that sarcopenia impairs OS due to a combination of treatment-related deaths and recurrence of malignancy. That is, treatment-related death plays a primary role in the early phase (within 3 years) posttreatment, while recurrence plays a primary role over the longer term (perhaps more than 5 years) after treatment, which could explain our results that sarcopenia resulted in poorer 5-year DFS but better 1-and 3-year DFS. The reduction in OS (1-and 3-year) in patients with sarcopenia and lack of associations with DFS (1-and 3-year) between groups also suggested that there are  perhaps other factors associated with poor survival in patients with sarcopenia. As reported, sarcopenia can be merely a secondary finding of longstanding chronic disease such as diabetes, gastrointestinal disorders, and cardiovascular diseases. This finding is in agreement with our results that sarcopenia was not associated with DFS (1-and 3-year) between groups because non-liver-or non-HCC/ICC-related factors are also related to prognosis. Interestingly, we noted that patients with sarcopenia had poorer 5-year DFS but better overall 5-year survival. The reason for this inconsistency is undoubtedly multi-factorial. As in our previous descriptions, sarcopenia increases the incidence of recurrence significantly, so more patients die from recurrence of malignancy in patients with sarcopenia on the basis of similar OS to that of patients without sarcopenia, which results in poorer 5-year DFS in patients with sarcopenia. Also, the available data suggested that the number of patients/trials on DFS and recurrence might have been too small, so bias cannot be excluded.
Regarding post-treatment complications, they remain controversial. Valero et al. [21] showed that the presence of sarcopenia was an independent predictive factor of OCs and MCs. However, evidence [19] has also shown sarcopenia did not increase the incidence of OCs. The available data from our meta-analysis suggested that sarcopenia increased the incidence of OCs significantly. However, our study provided no evidence of an increased risk of MCs in sarcopenic patients. The reasons below might account for the nonstatistically significant difference in MCs: 1) patients with sarcopenia might receive more intensive care before, during and after treatment; and 2) the absence of a sub-analysis of each complication might have caused potential bias because each complication that develops after HCC/ICC treatment could have a different relationship or causality with sarcopenia and a different impact of prognosis.
Our present meta-analysis had several limitations. In particular, the absence of randomized, controlled clinical trials and the preponderance of retrospective studies could have biased the results to some extent. Second, the relatively small number of participants could have biased the results. Third, several trials were excluded because sarcopenia was not evaluated by L3 SMI/TPA, which could have caused potential bias. Fourth, no subgroup analysis could be conducted because the data extracted from the included studies were not disaggregated by groups. Fifth, the absence of a consistent duration between CT scans and treatment might have created potential bias. Sixth, we failed to elucidate whether sarcopenia was truly a causal factor of poor prognosis or merely a concomitant finding of primary hepatic malignancies or a type of comorbidity.

Search strategy and study selection
This systematic literature review and meta-analysis were performed using the methodology suggested by the PRISMA guidelines. A systematic literature search was performed in English through February 1, 2017, in the following databases: Medline, PubMed, Medline, Embase, Web of Science and the Cochrane Library. The key words used were "sarcopenia," "hepatocellular carcinoma," "intrahepatic cholangio-carcinoma," "primary hepatic malignancy," "third lumbar skeletal muscle index," "third lumbar total psoas area," and abbreviations thereof. The key words were combined with appropriate Boolean operators, and for further relevant articles, we also checked the reference lists of all the identified trials. After completing the literature searches, titles and abstracts of the studies were screened by two authors, and any disagreement was resolved by discussion or, if necessary, adjudicated by a third author.

Inclusion and exclusion criteria
The inclusion criteria were as follows: 1) a prospective or retrospective cohort study was adopted; 2) the L3 SMI/TPA was measured by CT scan; and 3) patients were examined regularly and had a minimum follow-up duration of 12 months. When the same patient cohort overlapped among different publications, only the latest or complete study was considered.
The following studies were considered to be ineligible: 1) studies associated with animals; 2) case reports or studies in which the control groups were not well designed; 3) post-treatment sarcopenia; 4) and sarcopenia evaluated by intramuscular adipose tissue content (IMAC)/visceral fat area (VFA) rather than by L3 SMI/TPA.

Data extraction and study quality
Data extraction was performed independently by two authors using a standard form. The following data were extracted from each study: the basic information of the study (surname of the first author and year of publication, country of the procedure performed, study design, group assignment, histopathological type, treatment, CT scan parameters, definition of sarcopenia, number of patients and ages, follow-up duration) and the clinical significance of sarcopenia in the treatment of HCC/ICC (overall survival, disease-free survival, recurrence of malignancy, overall complications and major complications). The methodological quality of each trial was assessed according to the NOS, which evaluates included studies based on three broad perspectives: 1) selection; 2) comparability; and 3) outcome. The NOS assigns a maximum score of 4, 2 and 3 for selection, comparability and outcome respectively. High-quality trials scored ≥7, and moderate-quality trials scored ≥5 (maximum possible score equal to 9).

Statistical analysis
The statistical analysis was performed independently by two authors according to recommendations from the PRISMA statement and the Cochrane handbook from the Cochrane Collaboration. Pooled relative risk (RR) with 95% confidence intervals (95% CIs) were calculated for each principal dichotomous variable outcome using either a fixed effects model or a random effects model. Values <1 did not favor the sarcopenia group and values >1 favored the sarcopenia group for overall survival and disease-free survival rate (it should have been the opposite for the parameters such as recurrence and complications rate). We analyzed heterogeneity among the studies using Cochrane's Q test and by calculating I 2 , with P < 0.05 used to denote statistical significance and with I 2 calculated to measure the proportion of total variation in the estimates of treatment effect due to heterogeneity beyond chance. No subgroup analysis could be conducted in our meta-analysis because the data extracted from the included studies were not disaggregated by groups. All the statistical analyses were performed using RevMan software, version 5.3, provided by Cochrane Collaboration.

CONCLUSION
In conclusion, sarcopenia seemed to have a negative effect on OS in patients with primary hepatic malignancies in the early phase post-treatment, but further research is needed to investigate the prognostic impact on OS over the longer term. Also, we concluded that sarcopenia could significantly increase the incidence rates of post-treatment recurrence and overall complications in patients with primary hepatic malignancies. However, due to a lack of the subgroup analysis, sufficient RCTs and prospective cohort studies, the prognostic impact of sarcopenia remains contentious; therefore, future investigations are needed to evaluate the prognostic impact of sarcopenia in primary hepatic malignancies.

Author contributions
ZGQ participated in design, data collection, studies selection, data analysis/interpretation, and statistics and wrote the first draft. MSF and LRF performed research and wrote the first draft. MSF and YJW collected and analyzed the data. All authors contributed to the design and interpretation of the study and to further drafts. ZLS is the guarantor.