Prevalence of Sarcopenia and Its Impact on Cardiovascular Events and Mortality among Dialysis Patients: A Systematic Review and Meta-Analysis

Sarcopenia in end-stage kidney disease patients requiring dialysis is a frequent complication but remains an under-recognized problem. This meta-analysis was conducted to determine the prevalence of sarcopenia and explored its impacts on clinical outcomes, especially cardiovascular events, and mortality in dialysis patients. The eligible studies were searched from PubMed, Scopus, and Cochrane Central Register of Controlled trials up to 31 March 2022. We included studies that reported the interested outcomes, and the random-effects model was used for analysis. Forty-one studies with 7576 patients were included. The pooled prevalence of sarcopenia in dialysis patients was 25.6% (95% CI 22.1 to 29.4%). Sarcopenia was significantly associated with higher mortality risk (adjusted OR 1.83 (95% CI 1.40 to 2.39)) and cardiovascular events (adjusted OR 3.80 (95% CI 1.79 to 8.09)). Additionally, both low muscle mass and low muscle strength were independently related to increased mortality risk in dialysis patients (OR 1.71; 95% CI (1.20 to 2.44), OR 2.15 (95% CI 1.51 to 3.07)), respectively. This meta-analysis revealed that sarcopenia was highly prevalent among dialysis patients and shown to be an important predictor of cardiovascular events and mortality. Future intervention research to alleviate this disease burden in dialysis patients is needed.


Introduction
Sarcopenia, a disorder defined as having a reduction in skeletal muscle mass with declining muscle strength and function [1], is a frequent complication in patients with chronic kidney disease (CKD), especially in end-stage kidney disease (ESKD) patients undergoing kidney replacement therapy (KRT). It has been shown to be an important predictor of patients' falls and fractures, mobility disorders, dependency, low quality of life, and death [2,3]. Interest in sarcopenia increased after the European Working Group for Sarcopenia for Older People (EWGSOP) established the Sarcopenia Consensus in 2010 [4]. There were exponentially increased publications of this disorder in various populations [5], including chronic dialysis patients [6,7]. Moreover, many recent studies have indicated that there is an observed association between sarcopenia and various clinical outcomes.
Beyond the aging-related process, ESKD-related sarcopenia was categorized as secondary sarcopenia due to several vulnerable factors, including the accumulation of uremic toxins, metabolic acidosis, malnutrition, loss of amino acids in dialysis therapy, and typical low-grade chronic inflammatory status [6,7]. These elements together result in a final process of increasing protein degradation and reducing protein synthesis, leading to a negative nitrogen balance [8,9]. Finally, dialysis patients tend to have low physical activity, resulting in aggravation of muscle loss [10].
Indeed, studies regarding the prevalence of sarcopenia among dialysis patients yield widely varying results, depending on the populations, the methods used to measure muscle mass, and the diagnosis criteria [2,4,[10][11][12][13][14][15]. As a result of the large discrepancies in definition, sarcopenia prevalence in patients undergoing KRT is, to date, poorly defined. It remains an under-recognized problem that may hinder the early intervention, leading to poor clinical outcome. This systematic review and meta-analysis aimed to determine the global prevalence of sarcopenia in patients with ESKD undergoing both hemodialysis (HD) and peritoneal dialysis (PD) modalities and explore whether it has significant association with the clinical outcomes, especially cardiovascular (CV) events and mortality. Such information may give growing attention to sarcopenia and enhance future intervention research to address effective clinical service programs in dialysis patients.

Methods
We performed this systematic review and meta-analysis following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 [16] guidelines for reporting systematic reviews of interventions and a prespecified registered protocol in the PROSPERO database (registration number CRD 42022324618).

Searching Strategy
We conducted a systematic search of relevant articles in the database PubMed incorporating Medical Subject Heading Indexation (MeSH) terms, Scopus, and Cochrane Central Register of Controlled trials up to 31 March 2022. Keywords were detailed in Supplementary Material Table S1. The search had no language restriction and focused on adults (age ≥ 18 years).

Study Outcomes
The primary outcome was the prevalence of sarcopenia in dialysis patients. We decided to focus on the prevalence estimates derived from validated and well-established diagnostic criteria widely used internationally, including the European Working Group on Sarcopenia in Older People (EWGSOP), Asian Working Group for Sarcopenia (AWGS), Foundation for the National Institutes of Health Sarcopenia Projects (FNIH), and International Working Group on Sarcopenia (IWGS). We also planned to explore the association of sarcopenia with clinical outcomes. The clinical outcomes of interest were CV events (defined as any event related to the cardiovascular system including myocardial infarction, stroke, hypertensive crisis, heart failure, or unspecified cardiovascular serious adverse events reported according to the definition of each trial) and all-cause mortality.

Eligibility Criteria
Studies that were eligible for inclusion were: (1) the population consisting of adults (≥18 years) with ESKD on dialysis (HD or PD); (2) the prevalence of sarcopenia was reported; (3) sarcopenia was defined by assessing both low muscle mass (LMM) and low muscle strength (LMS); (4) the study reported baseline characteristics of both sarcopenic and non-sarcopenic patients. Exclusion criteria were as follows: (1) the study was an animal study, a review article, case report, or case series; (2) the study focused only on a specific group of sarcopenias (e.g., sarcopenic obesity); (3) the study did not stratify dialysis patients from patients with other spectra of kidney diseases: non-dialysis-dependent CKD, and patients undergoing kidney transplantation. ment between the first two reviewers by discussion. The largest sample size report was selected if the reports resulted from the same cohort.

Data Extraction and Risk of Bias Assessment
Two authors (AB and WW) independently carried out data extraction. The data tables were categorized in topics, including author, year of publication, study design, number of the study population, baseline characteristics of patients (age, sex, dialysis vintage, mode of KRT), operational criteria for the diagnosis of sarcopenia, measurement tools (for assessing muscle mass, muscle strength and/or physical performance), duration of follow-up, the prevalence of sarcopenia as well as CV events, and mortality rate in cohort studies.
The risk of bias was assessed using the Newcastle-Ottawa Scale (NOS) for cohort studies, and the NOS was adapted for cross-sectional studies [17,18]. The NOS includes a series of questions used to assess the choice of study participants, the comparability of the population, and the determination of exposure or outcomes. The NOS for cohort studies was converted to AHRQ standards as good, fair, or poor quality. The NOS was adapted for cross-sectional studies with a maximum score of 10 points. The 9-10 points study was categorized as very good studies; 7-8 points as good studies; 5-6 points as satisfactory studies and 0-4 points as unsatisfactory studies. Using these checklists, two authors (AB and WW) evaluated each of the included articles for their quality. Divergent views were resolved by consultation with a third author (PS) (Supplementary Tables S2 and S3).

Statistical Analysis
The results of the systematic review were qualitatively tabulated and synthesized. For a subset of studies with analyzable and comparable data, the results were synthesized quantitatively by performing random-effects model meta-analyses to compute absolute net differences in continuous variables (i.e., age, body mass index (BMI), serum albumin, phosphate, parathyroid hormone (PTH), creatinine (Cr), C-reactive protein (CRP), 25-OH vitamin D, hemoglobin (Hb), and Kt/V) and pooled odd ratios (OR) for binary variables (i.e., presence versus absence of sarcopenia) from univariate and multivariate analyses. The data from both analyses were retrieved for meta-analysis. All pooled estimates were displayed with a percentage and 95% confidence interval (CI). Existence of heterogeneity among effect sizes of individual studies was determined by using the Cochrane's Q test and the I 2 index, with a value of 75% or greater indicating medium-to-high heterogeneity. To explore sources of heterogeneity, we performed subgroup meta-analyses according to KRT modalities (HD or PD), race (Asian or non-Asian), time of muscle mass measurement (pre-HD, intra-HD, or post-HD), the tool of muscle mass measurement (dual-energy X-ray absorptiometry (DXA), bioelectrical impedance spectroscopy (BIS), bioelectrical impedance analysis (BIA) or calf circumference (CC)), follow-up duration (less or more than 2 years), and components of diagnostic criteria (LMM or LMS). Forest plot was used as a graphical representation of heterogeneity among the included studies. Publication bias was formally evaluated by Funnel plots and the Egger test. The analyses were performed using Comprehensive Meta-Analysis version 2.0 (www.meta-analysis.com (accessed on 4 August 2022); Biostat, Englewood, NJ, USA).

Operational Criteria for the Diagnosis of Sarcopenia
The diagnostic methods and measurement tools for muscle assessment of the individual study are shown in Table 1. All studies assessed muscle mass plus muscle strength and/or physical performance to fulfill the validated operational diagnostic criteria of sarcopenia. The EWGSOP 2010 criteria (n = 17:41.5%) was the most utilized. Regarding measurement tools for assessing muscle mass, BIA (n = 26:63.4%) and DXA (n = 10:24.4%) were the two mainly selected. All muscle strength was assessed by handgrip test (n = 41:100%). Physical performance was evaluated in nearly half of all studies (n = 19:46.3%) mainly by gait speed test.

Methodological Quality
Regarding the Newcastle-Ottawa Scale (NOS) for cohort studies and the NOS adapted for cross-sectional studies [17,18], all cohort studies (n = 16:100%) were considered as good quality. Most of the cross-sectional studies (n = 24:96%) were considered as good quality (score of 7-8), while the remaining study (n = 1:4%) was considered as satisfactory quality (score of 5-6) (see Supplementary Tables S2 and S3).

Variables
No. Studies

Cardiovascular Events
CV events were reported in two studies [23,37] (175 patients) and were included in the meta-analysis. According to adjusted OR, sarcopenia was significantly associated with increased CV events in dialysis patients (adjusted OR 3.80, 95%CI: 1.79 to 8.09, p = 0.001, I 2 0%) ( Table 4).

Hospitalization
Two studies [25,50] followed dialysis patients to assess the impact of sarcopenia on hospitalization outcomes. The first study [25] included 126 chronic HD patients aged 63.2 ± 13 years with 3-year follow-up. The authors reported no significant difference in the incidence of hospitalization between sarcopenic and non-sarcopenic patients without absolute reported number. In contrast, the second study [50] followed 170 patients on maintenance HD for 3 years. The risk of hospitalization was significantly higher in sarcopenic patients with a crude RR of 1.80 (95% CI 1.35-2.41, p < 0.001) and a fully adjusted RR (adjusted for age, gender, dialysis vintage, and DM) of 2.07 (95%CI 1.48-2.88, p < 0.001).

Dependency
Only one study [29] estimated the association of sarcopenia with dependency in activities of daily living (ADL). This study was performed in a total of 238 patients with maintenance HD with an average age of 60.9 ± 13.2 years. Univariate analysis showed that the presence of sarcopenia was significantly associated with dependency in both basic ADL (OR 2.69, 95% CI 1.39-5.20, p = 0.003) and instrumental ADL (OR 3.33, 95% CI 1.94-5.72, p < 0.001). However, after adjusting for clinical covariates, the OR for neither basic ADL nor instrumental ADL was significant (p > 0.05).

Frailty
Only one study [21] reported the association of sarcopenia with frailty based on the Clinical Frailty Scale (CFS). This study evaluated 119 PD patients with the mean age of 66.8 ± 13.2 years and the mean follow-up period of 589.2 days. According to the multivariate logistic regression model, after adjusting for age, sex, BMI, nPNA, and Charlson comorbidity index (CCI), sarcopenia was significantly correlated with frailty (adjusted OR 12.2, 95% CI 2.27-65.5, p = 0.003). Table 5 details the results of subgroup analyses comparing all-cause mortality between sarcopenic and non-sarcopenic dialysis patients as stratified by dialysis modalities (HD or PD), race (Asian or non-Asian), time of muscle mass measurement (pre-HD, intra-HD, or post-HD), tool of muscle mass measurement (DXA, BIS, BIA, or CC), follow-up duration (less or more than 2 years), components of diagnostic criteria (low muscle mass or low muscle strength), and five adjustment variables (age, sex, comorbidity, nutritional indices, and inflammatory markers). In brief, there was a significant association between dialysis patients with sarcopenia and all-cause mortality relative to non-sarcopenic dialysis patients across all subgroups and adjustment variables as mentioned above. However, there was significant heterogeneity based on the Q-test p-value and I 2 index in some subgroups of the included studies, including Asian ethnicity and tools of muscle mass assessment by BIA.

Meta-Regression Model
The meta-regression modeling, adjusted for sample size, was applied to mean ages, mean serum albumin, mean dialysis vintage, and crude prevalence rates from all studies, as these generated more data points. The modeling revealed that the prevalence of sarcopenia was not influenced by age (p = 0.63), serum albumin (p = 0.64), or dialysis vintage (p = 0.59).

Assessment of Publication Bias
The funnel plot for the outcome of both sarcopenic prevalence and all-cause mortality in the studies included in the meta-analysis was asymmetrical (Figures S2 and S3). According to publication bias on all-cause mortality outcome, the sensitivity analysis was conveyed to identify the most influential study on the pooled adjusted odds ratio. We found that the pooled adjusted OR of all-cause mortality depended on a single study by Kim et al. [23] conducted in patients who had a high burden of co-morbid disease, espe-cially cardiovascular disease. After excluding this study, the all-cause mortality outcome did not have a significant publication bias (p = 0.21) ( Figure S4).

Discussion
ESKD patients undergoing KRT are at high risk of sarcopenia owing to the trend toward increased dialysis among the aging population incorporated with several ESKDrelated sarcopenic factors. From this systematic review and meta-analysis, it was found that sarcopenia was highly prevalent among dialysis patients (overall pooled prevalence 25.6%) and widely varying, ranged from 1.5% to 68% (Figure 1), mainly depending on the applied diagnostic criteria, tools and time of muscle assessment, dialysis modalities, and abnormal fluid status in dialysis patients. Furthermore, this condition was also associated with adverse clinical outcomes, especially CV events and all-cause mortality. Dialyzed patients with sarcopenia were associated with increased risk of CV events (adjusted OR 3.80) and death (adjusted OR 1.83) compared with non-sarcopenic individuals ( Table 4).
In ESKD patients undergoing KRT, both HD and PD, the frequency of sarcopenia is even greater than among subjects with normal renal function [61] because uremic-induced anorexia, metabolic disorder, and hormonal derangements inhibit muscle synthesis and accelerate muscle wasting [62]. Regarding dialysis modalities, our study found that PD patients seemed to have a significantly lower prevalence of sarcopenia than maintenance HD patients ( Table 2). First of all, ESKD patients who chose PD rather than HD modality seemed to have a better health status as a result of younger age, less comorbidity and greater physical independence [63]. Moreover, the influence of each KRT modality on sarcopenia might be related to several issues including daily routine and the patient's level of physical activity [64]. Most PD patients spend most of their time out of the healthcare center. They have more available time during the day, especially patients treated with nocturnal intermittent peritoneal dialysis (NIPD) mode, due to requiring only one or two dialysate exchanges per day at nighttime. On the other hand, HD patients must spend at least four hours, twice or thrice a week, with limited activity during transit to the HD center and during the HD session. In addition, after the HD session some HD patients may experience troublesome symptoms related to the HD procedure, such as fatigue, dizziness, cramping, etc. Therefore, on HD day, HD patients are often at increased need of rest and consequently favor a more sedentary lifestyle.
Another issue related to KRT modality is the assessment timing. This meta-analysis revealed that sarcopenia tended to be under-diagnosed by pre-HD compared with post-HD measurement of muscle mass (21.5% pre-HD to 27.8% post-HD), but with no statistical significance (p for interaction = 0.21) (Table 2). Similarly, a previous study [26] illustrated that post-dialysis measurement of muscle mass provided greater reliability. Although pre-dialysis measurement is more convenient, it certainly causes an error of muscle mass overestimation from a relatively high amount of water-contained muscle [65,66]. In addition, Yilmaz et al. [67] also revealed that over-hydration was statistically more frequent in PD than in post-HD patients (30.3% vs. 11.6%, p = 0.043). This factor may lead to the underestimation of the prevalence rate of sarcopenia in PD patients. Besides muscle mass assessment, measurement of muscle strength was also affected. Pinto et al. [68] demonstrated the dramatic reduction in HGS after HD compared with before HD sessions. The results showed a significant increase in the number of patients with HGS below the 30th percentile (44.9% before HD to 55.1% after HD; p < 0.01), particularly as it was related to post-dialysis fatigue caused by rapid removal of water and solutes during the procedure.
The last issue between HD and PD patients is that there is a strong body of evidence suggesting that PD better preserves residual kidney function (RKF) when compared with conventional HD [69]. The benefit of well-preserved RKF factor is greater elimination of protein-bound uremic toxins (PBUTs). Alcalde-Estévez. et al. [70] showed that PBUTs (combination of indoxyl sulphate and p-cresol) impaired the skeletal muscular regeneration process by inhibiting myoblast proliferation, reducing myogenic differentiation, and promoting muscular fibrosis, even at low concentrations, in a uremic rat model. From all the above factors including limited physical activity, the assessment timing, HD-related factors, and less-preserved RKF, the higher prevalence of muscle wasting and weakness in HD patients compared with PD patients was illustrated in our study (Table 5).
Diabetes mellitus (DM), one of the major health burdens and causes of KRT initiation worldwide [71], was shown to be negatively affecting various aspects of muscle health through impairments in protein metabolism, vascular and mitochondrial dysfunction by several different mechanisms, including inflammation, insulin resistance, advanced glycation end-product accumulation, and increased oxidative stress [72]. Mori. et al. [24] demonstrated that among 308 patients undergoing HD, DM as a comorbid disease was significantly associated with a higher rate in sarcopenic patients than non-sarcopenia individuals (41% vs. 27%, p = 0.015). Moreover, the presence of DM was an independent contributor to sarcopenia and an independent predictor of all-cause mortality in this population. In contrast, we found that diabetic patients were not significantly associated with greater risk of sarcopenia relative to non-diabetic individuals (OR 1.20, 95% CI 0.72 to 1.57, p = 0.18). This finding may be potentially explained by the better glycemic control in advanced CKD progress to ESKD requiring KRT from spontaneous resolution of hyperglycemia and normalization of glycated hemoglobin (HbA1C) levels designed, known as burn-out diabetes phenomenon [73]. In addition, DM is only one of the several ESKDrelated factors in the development of sarcopenia. It is possible that the impact of DM might be overlaid by the other factors, including advanced age, long dialysis vintage, malnourishment, and CKD-Mineral and Bone Disorder (CKD-MBD) ( Table 3).
Cardiovascular disease (CVD) is a highly common complication and the first cause of death in patients undergoing dialysis. Moreover, mortality due to CVD in this population is twenty times higher than in the general population [74]. Our meta-analysis demonstrated that sarcopenia in dialysis patients was one of the most important predictors of CV events as well as mortality outcome, and this was independent of study design, population, sex, continent, dialysis method, sarcopenia definition, and study quality (Table 5).
Regarding its components of diagnostic criteria, our meta-analysis illustrated that both LMM and LMS were independently related to increased mortality risk in dialysis patients. However, the magnitude of associated risks tended to be greater in the LMS component compared with the LMM component (Table 4). Several cohort studies [25,31,40,50,57] showed that dynapenia (i.e., LMS) could be superior to LMM for better prediction of adverse clinical outcomes, including quality of life, hospitalization, CV events, and all-cause mortality in ESKD patients. Consistent with these findings, Goodpaster et al. postulated that during the muscle wasting process, muscle strength decline occurs at a faster rate than muscle mass loss [75]. The proposed mechanism is that initial muscle weakness leads to decreased muscle function, diminished physical activity, and sometimes ends up with immobilization, as well as muscle disuse atrophy. Thus, decreased muscle mass and muscle strength are likely to be both the cause and the sequelae of each other. However, muscle mass in dialyzed patients is not the only determinant of muscle strength, but other factors such as muscle relaxation and fiber muscle atrophy can also explain LMS [76]. Indeed, the fulfilled diagnostic criteria of sarcopenia should be regarded as a late stage of muscle wasting since LMS alone (probable sarcopenia) is enough to trigger assessment of cause and provide early effective therapeutic intervention [2].
Although there are two previous meta-analyses on this topic (one that included 30 studies [11] and another that included 50 studies [77]), one of these reports [11] included several studies that defined sarcopenia as the presence of low muscle mass alone (or probable sarcopenia). In addition, not all studies used the validated sarcopenia diagnostic criteria. Thus, it could not be specified whether the prevalence and clinical outcomes exactly occurred from sarcopenia. In another previous report [77], the overall prevalence of sarcopenia was not explored. Moreover, the subgroup analyses according to dialysis modalities were limited due to a very small number of PD patients. Lastly, none of these reports explored the association of sarcopenia with CV events, one of the most important adverse clinical outcomes in dialysis patients.
The strength of the present study is that this is the first systematic review and metaanalysis of observational studies that demonstrates the worldwide pooled prevalence of sarcopenia and explores an association between sarcopenia and adverse clinical outcomes, especially cardiovascular events and mortality, among ESKD patients on dialysis. We included reports that performed multivariable analyses to account for potential confounders of these associations. In addition, the search included all studies published up to March 2022. More recent studies, using the EWGSOP 2019 as well as the AWGS that updated its sarcopenia definition and its cut-off points in March 2020, have been included. This might impact the recently updated overall prevalence of all classifications and subgroups by cut-off points. Despite these strengths, some limitations are worth mentioning. Firstly, our systematic review was limited to observational studies, and in the absence of randomized controlled trials, the cause-and-effect relation between sarcopenia and clinical outcomes remains speculative. Secondly, it is important to note the heterogeneity of the included studies due to the variety of population, the applied diagnostic criteria of sarcopenia, tools and time of muscle assessment, and dialysis modalities. Lastly, some clinical outcomes were reported by a small number of trials and more sample size is required to show a significant association.

Conclusions
In conclusion, our systematic review revealed that sarcopenia was highly prevalent among dialysis patients and demonstrated to be an important predictor of cardiovascular events and mortality. Additionally, both low muscle mass and low muscle strength were independently related to increased mortality risk. Of note, these findings could increase awareness of musculoskeletal health and encourage nephrologists who treat patients at risk of sarcopenia, leading to early detection and prompt implementation of beneficial therapeutic strategies. To optimize individual therapeutic approach and alleviate this burden of disease in dialysis patients, more future intervention research is needed.  Data Availability Statement: The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.