jfa journal

AND option

OR option



H. Liu1,#, W. Li2,#, M. Zhu3, X. Wen4, J. Jin5, H. Wang6, D. Lv7, S. Zhao8, X. Wu9, J. Jiao10


1. School of Nursing, Peking University, Beijing 100191, China; 2. Department of Clinical Nutrition, Chinese Academy of Medical Sciences – Peking Union Medical College, Peking Union Medical College Hospital (Dongdan campus), Beijing 100730, China; 3. Department of Geriatrics, Chinese Academy of Medical Sciences – Peking Union Medical College, Peking Union Medical College Hospital (Dongdan campus), Beijing 100730, China; 4. Department of Nursing, Sichuan Provincial People’s Hospital, Chengdu 610072, China; 5. Department of Nursing, The Second Affiliated Hospital Zhejiang University School of Medicine, Hangzhou 310009, China; 6. Department of Nursing, Tongji Hospital, Tongji medical college, Huazhong University of Science and Technology, Wuhan 430074, China; 7. Department of Nursing, The Second Affiliated Hospital of Haerbin medical University, Haerbin 150081, China; 8. Department of Nursing, Qinghai Provincial People’s Hospital, Xining 810007, China; 9. Chinese Nursing Association, Beijing 100035, China; 10. Department of Nursing, Chinese Academy of Medical Sciences – Peking Union Medical College, Peking Union Medical College Hospital (Dongdan campus), Beijing 100730, China. #Contributed equally.

Corresponding Author: Hongpeng Liu, Peking University School of Nursing, Beijing, China, liuhongpeng12@sina.com; Xinjuan Wu, wuxinjuan@sina.com

J Frailty Aging 2024;in press
Published online January 31, 2024, http://dx.doi.org/10.14283/jfa.2024.9



BACKGROUND: Population aging might increase the prevalence of undernutrition in older people, which increases the risk of frailty. Numerous studies have indicated that myokines are released by skeletal myocytes in response to muscular contractions and might be associated with frailty. This study aimed to evaluate whether myokines are biomarkers of frailty in older inpatients with undernutrition.
METHODS: The frailty biomarkers were extracted from the Gene Expression Omnibus and Genecards datasets. Relevant myokines and health-related variables were assessed in 55 inpatients aged ≥ 65 years from the Peking Union Medical College Hospital prospective longitudinal frailty study. Serum was prepared for enzyme-linked immunosorbent assay using the appropriate kits. Correlations between biomarkers and frailty status were calculated by Spearman’s correlation analysis. Multiple linear regression was performed to investigate the association between factors and frailty scores.
RESULTS: The prevalence of frailty was 13.21%. The bioinformatics analysis indicated that leptin, adenosine 5‘-monophosphate-activated protein kinase (AMPK), irisin, decorin, and myostatin were potential biomarkers of frailty. The frailty group had significantly higher concentrations of leptin, AMPK, and MSTN than the robust group (p < 0.05). AMPK was significantly positively correlated with frailty (p < 0.05). The pre-frailty and frailty groups had significantly lower concentrations of irisin than the robust group (p < 0.05), whereas the DCN concentration did not differ among the groups. Multiple linear regression suggested that the 15 factors influencing the coefficients of association, the top 50% were the ADL score, MNA-SF score, serum albumin concentration, urination function, hearing function, leptin concentration, GDS-15 score, and MSTN concentration.
CONCLUSIONS: Proinflammatory myokines, particularly leptin, myostatin, and AMPK, negatively affect muscle mass and strength in older adults. ADL and nutritional status play major roles in the development of frailty. Our results confirm that identification of frailty relies upon clinical variables, myokine concentrations, and functional parameters, which might enable the identification and monitoring of frailty.

Key words: Frailty, myokines, biomarkers, skeletal muscle, undernutrition.

Abbreviations: MNA-SF: Mini Nutritional Assessment Short-Form; SD: standard deviation; CI: confidence interval; ADL: activities of daily living; WBC: white blood cell; AMPK: adenosine 5‘-monophosphate-activated protein kinase; DCN: decorin; MSTN: myostatin; LEP: Leptin.



The substantial increase in life expectancy and rapid expansion of the aging population place increased pressure on healthcare systems globally (1). Population aging might increase the prevalence of undernutrition in older people due to age-related pathophysiological and psychosocial factors, protein intake, or medications that cause changes in dietary habits, leading to specific nutritional deficits, especially among older inpatients (2-4). Nutritional status often deteriorates further during hospitalization, and its negative impact on health substantially increases the risk of frailty by decreasing the protein intake and accelerating sarcopenia (5, 6).
Frailty is characterized by a decline in functioning across multiple physiological systems, accompanied by an increased vulnerability to stressors, and is associated with negative health-related events (6, 7). We previously developed a frailty prediction model that integrated patient demographic characteristics, physical and psychological factors, and routine laboratory blood test results to predict the 30-day frailty risk among Chinese older adults with undernutrition, underpinned by the biological principles of causality (8). Interestingly, the sensitivity and area under the receiver operating characteristic curve of the model increased when routine laboratory test results (clinical biomarkers including serum albumin and hemoglobin) were added; the same effect was found in a prospective validation dataset of 268 older adults hospitalized in Peking Union Medical College Hospital (9).
Biomarkers are objectively measured and evaluated as indicators of normal biological processes, pathogenic processes, or pharmacologic responses to interventions (10). Although some candidate biomarkers of frailty have been identified, none of them has yet been incorporated into the assessment or monitoring of the condition (11). In recent years, increasing research interest has been focused on myokines, a set of peptides or cytokines synthesized by skeletal muscle cells and subsequently released into the circulation, where they exert autocrine, paracrine, or endocrine effects in other cells, tissues, or organs (12-14). Previous studies have suggested that myostatin (MSTN), brain-derived neurotrophic factor, interleukin 6 and 7, follicle statin-associated protein 1, and irisin may be closely involved in the regulation of muscle production and loss (10, 11, 15). As the release of myokines from skeletal muscle may change with the progression of physical frailty and sarcopenia, these biomolecules might be useful as biomarkers of muscle (dys)function (10). Therefore, the addition of frailty biomarkers into the frailty prediction model might enable the identification of individuals with frailty and allow better tracking of frailty over time, optimization of treatments and interventions, and monitoring of the efficacy of interventions (10, 16).
In the present study, we extracted the biomarkers of frailty from the Gene Expression Omnibus (GEO) and Genecards datasets. We then prospectively collected blood samples from older inpatients to determine whether these blood biomarkers were frailty myokines, and combined this information with the patient’s demographic characteristics, clinical examination findings, nutritional status, psychological factors, and routine laboratory blood test results to examine the associations with frailty, in order to objectively enrich the frailty prediction model in the future.




The present study used data from a prospective validation dataset for the frailty prediction model integrating physical factors, psychological variables and routine laboratory test parameters to predict the 30-day frailty risk in older adults with undernutrition. All included participants were 65 years old or older, with no frailty according to the FRAIL scale (scores range from 0 to 2) and with undernutrition (scores range from 0 to 11) according to the Mini-Nutritional Assessment-Short Form (MNA-SF). The study flowchart is shown in Figure S1. In total, 2842 participants (Development cohort and Validation cohort 1) were enrolled from six provinces in China between October 2018 and February 2019 and used for the modeling and internal validation. Samples collected from 268 patients in Peking Union Medical College Hospital between August 2021 and November 2021 were used for the external validation of the frailty model (Validation cohort 2). Among the 268 patients, a total of 55 blood samples were collected but two were lost to follow-up. Therefore, 53 inpatients were analyzed in the current study. Details of the frailty prediction model can be found elsewhere (17).

Definitions of health-related variables

We developed a multiple linear regression including the following factors potentially associated with frailty: age, sex, routine laboratory blood test results (serum albumin and hemoglobin concentrations), vision function, hearing function, urination function, activities of daily living (ADL) score, depression score, and nutritional status. The routine laboratory blood test results were extracted from the medical records. The sensory impairment was identified by self-reported assessment of visual and hearing functions. Similar to the Survey of Health, Aging and Retirement in Europe, and the China Health and Retirement Longitudinal Study (CHARLS), we collected self-reported data on visual functions using 2 questions: “Is your eyesight for seeing things at a distance excellent (1), very good (2), good (3), fair (4), or poor (5)?” and “How good is your eyesight for seeing things up close, like reading ordinary newspaper print? Would you say your eyesight for seeing things up close is excellent (1), very good (2), good (3), fair (4), or poor (5)?” Data on hearing functions were also collected using the question: “Is your hearing excellent (1), very good (2), good (3), fair (4), or poor (5)?” We identified participants as having visual or hearing impairment if they reported fair or poor vision (for either long distance or near vision) or hearing and then categorized these measures as follows: no visual impairment, visual impairment, no hearing impairment, and hearing impairment. The urinary function was assessed whether or not influenced normal life and was dichotomized as normal function or dysfunction (such as frequent micturition, urgent micturition, and urinary retention) (18, 19). Ability to perform ADL was evaluated using the Barthel Index (BI), which is a 10-item instrument measuring disability in terms of a person’s level of functional independence in personal ADL. The BI consists of 10 ADL, each graded as 0, 5, or 10 with a maximum total score of 100 (20, 21). A higher BI indicates better capacity to perform ADL. The BI has been validated for use in Chinese older adults (22). Nutritional status was measured by the Mini Nutritional Assessment—Short Form (MNA-SF), which is a six-item scale that assesses nutritional risk among the geriatric population (23). The MNA-SF score ranges from 0 to 14, with a lower score denoting worse undernutrition. For the purpose of our study, MNA- SF scores range from 0 to 11 referring to undernutrition, and undernutrition includes malnutrition risk and malnourished. Thus, the nutritional status means at risk of malnutrition (scores ranging from 8 to 11 points) and malnourished (scores ranging from 0 to 7 points) (9). The MNA-SF has been validated for use in Chinese older adults (17, 24). Depression was assessed using the 15-item Geriatric Depression Scale 15 (GDS-15) (25). The total GDS-15 score ranges from 0 to 15, with a higher score denoting more severe depression. The GDS-15 has been validated for use in Chinese older adults (26).

Assessment of frailty

The dependent variable (outcome) was frailty. Frailty status was divided into the robust, pre-frailty, and frailty groups of the inpatients who met 0, 1–2, or ≥ 3 of the FRAIL criteria (27). A higher total FRAIL score indicates worse frailty.

Bioinformatics Analysis

Data Sources

To construct the biological network of frailty, the frailty-related genes were collected from the GEO database (http://www.ncbi.nlm.nih.gov/geo/) and GeneCards (https://www.genecards.org/). In the GEO database, the GSE1428 series (platform: GPL96) of 12 specimens with skeletal muscle sarcopenia and 10 normal specimens was used. Data for 127 “frailty” gene sets from the GeneCards database. Finally, 167 frailty-related genes were obtained.

Data Processing

We compared different sample groups in the GEO series and screened differentially expressed genes by the GEO query (2.54.1 version) and limma (3.42.2 version) R packages. The recognition threshold was set as differentially expressed genes (DEGs) with a false discovery rate of < 0.05 and |log2 multiple changes| of > 1. The upregulated and downregulated genes were analyzed. Hierarchical cluster analysis was then performed on the GEO series. The extracted and analyzed genetic data were visualized using the ggplot2 (3.3.3 version) and ComplexHeatmap (2.2.0 version) R packages.
The DEG list combined the extracted GeneCards and GSE1428 series for gene ontology (GO) and protein interaction analyses. GO functional enrichment analysis is a process of classifying genes or proteins. GO analysis and visual processing were performed using Metascape (http://metascape.org). Disease enrichment analysis was performed using the DisGeNET platform (http://www.disgenet.org) (28). Protein interaction analysis aims to establish a protein–protein interaction (PPI) network. The PPI network was established through the STRING11.5 (https://string-db.org/) online analysis platform. The potential targets of frailty were introduced into the STRING database with the species limited to “Homo sapiens” to obtain the PPI data. The data extracted from the PPI network were analyzed and visualized by R. All bioinformatics analyses were performed using R version 3.6.3.

Serum Sample Collection and Testing

Blood collection

In all patients, venous blood samples were taken in the morning following an overnight fast and after at least 15 minutes of supine rest. Blood samples were collected in vacutainer tubes (5 mL and serum samples were produced by centrifugation at 3000 rpm for 10 minutes at 4℃ (K15, Sigma, Osterode am Harz, Germany). Serum samples were stored at -80℃ until analysis.

Measurements of frailty biomarkers and myokines

Based on bioinformatics data and analysis results, the following five potential frailty biomarkers and myokines were identified: leptin, adenosine 5‘-monophosphate-activated protein kinase (AMPK), decorin (DCN), myostatin (MSTN), and irisin. The serum concentrations of these potential frailty markers and myokines were assessed. Serum was prepared for enzyme-linked immunosorbent assay using the appropriate kits (MSTN, irisin, leptin, DCN: Wuhan Huamei Biotech Co., Ltd., Wuhan, China; AMPK: Shanghai Jianglai Industrial Limited by Share Ltd., Shanghai, China). The lower detection limits for MSTN, irisin, leptin, DCN, and AMPK were 0.625 ng/mL, 3.12 ng/mL, 0.156 ng/mL, 3.9 pg/mL, and 0.312 ng/mL, respectively. The coefficients of variation for the five assays ranged from 8% to 15% during the sample analysis period. Results were expressed as pg/mL (DCN) or ng/mL (MSTN, leptin, irisin, AMPK). Each sample was run twice, and the mean for each sample was used as the index value. Two kit-supplied quality controls were run on each plate in duplicate and confirmed to fall within the expected range. Myokine concentrations were calculated by reference to an eight-point five-parameter logistic standard curve for each myokine.

Biochemical determinations

Serum biochemical parameters were measured in duplicate. Routine blood testing included hemoglobin and serum albumin concentrations. To minimize variation, analyses were performed by trained technicians using standard operating procedures on a single day using the same calibration and set-up.


This study was approved by the Ethics Committee of Peking Union Medical College Hospital (S-K540 and JS-2781). All study procedures were conducted in accordance with the principles of the Declaration of Helsinki. Written informed consent was provided by all participants or their legal guardians.

Statistical analysis

We used SPSS 25.0, Stata 15.0, and R version 3.6.3 for all statistical analyses. Continuous variables were summarized as mean ± standard deviation or median (25th–75th percentile), while categorical variables were summarized as number (percentage). Comparisons between groups were compared using the t-test, Mann-Whitney U-test, χ2 test, or Kruskal-Wallis test, as appropriate. Correlations between biomarkers and frailty status (robust, pre-frailty, and frailty) were calculated by Spearman’s correlation analysis. Multiple linear regression was performed to investigate the association between risk factors and frailty scores (scores ranging from 0 to 5). All P values were two-sided and statistical significance was set at 0.05.



Prevalence of frailty, ADL score, nutritional status, and blood test results in each group

Table 1 shows the baseline data of the included patients. During 30 days of follow-up, the frailty, pre-frailty, and robust groups comprised 13.21% (n=7), 69.81% (n=37), and 16.98% (n=9) of patients, respectively. The mean age was lower in the pre-frailty group (69.00 ± 4.69 years) than the frailty group (72.14 ± 7.40 years) and robust group (71.44 ± 5.10 years). The respective prevalence of frailty and pre-frailty were 10.0% (1/10) and 80.0% (8/10) among inpatients in the cardiology department and 13.95% (6/43) and 67.44% (29/43) in the gastroenterology department.
The frailty group had a lower ADL score (85.71 ± 8.86) than the pre-frailty and robust groups (97.57 ± 4.19 and 97.22 ± 8.33, respectively). The frailty group had more severe depression and poorer nutritional status compared with the prefrailty and robust groups. The frailty group had lower hemoglobin (112.71 ± 19.41 g/L) and serum albumin concentrations (35.14 ± 7.80 g/L) than the pre-frailty group (hemoglobin: 131.16 ± 18.59 g/L; albumin: 41.35 ± 5.50 g/L) and robust group (hemoglobin: 139.78 ± 15.56 g/L; albumin: 49.44 ± 22.97 g/L).

Table 1. Baseline data of the participants (n=53)

Values are presented as number (percentage) unless stated otherwise. ADL, activities of daily living; GDS-15, 15-item Geriatric Depression Scale 15; MNA-SF, the Mini Nutritional Assessment—Short Form; SD, standard deviation.


Frailty biomarkers from bioinformatics analysis

Data preprocessing and screening of DEGs in the frailty group

After data preprocessing, we extracted a total of 40 DEGs from the gene expression matrix, as shown in the volcano map and heatmap (Figure S2). A search of the Genecards database on November 16, 2021 using the keyword “frailty” retrieved 127 genes after deleting duplicate values. In addition to the genes extracted from the GEO database, we obtained a final total of 167 frailty-related genes.

Functional Correlation Analysis

The 167 genes underwent functional annotation and disease enrichment analysis to determine the biological significance of frailty. The GO enrichment analysis showed that DEGs were mainly related to nutritional metabolism functions (glucose metabolic process, cellular response to lipids, organic hydroxy compound metabolic process, orexin receptor pathway), muscle growth functions (muscle cell proliferation, AMPK cascade), and biosynthesis functions (regulation of growth, regulation of the nitric oxide biosynthetic process, regulation of hormone levels, negative regulation of cell proliferation, positive regulation of peptidyl-tyrosine phosphorylation, positive regulation of protein phosphorylation, age-related functions (cancer pathways), and immune-related functions (response to extracellular stimulus, multicellular organismal homeostasis, regulation of defense response, network map of SARS-CoV-2 signaling pathways) (Figure 1A). Frailty-related genes were analyzed using the DisGeNET platform (Figure 1B), including the investigation of the molecular underpinnings of specific human diseases and the analysis of the properties of disease genes. Diseases enriched by DEGs were mainly age-related chronic diseases, such as sarcopenia, malnutrition, presenile dementia, dyslipidemia, diabetes mellitus, diabetic retinopathy, amyloidosis, myocardial ischemia, heart disease, sleep apnea, vascular disease, chronic kidney disease, renal insufficiency, kidney failure, periodontitis, and endothelial dysfunction. These results showed that muscle function and aging play an important role in the occurrence and development of frailty.

Figure 1. Frailty biomarkers extracted by GO and DO analyses

A. GO analysis of frailty-related genes. B. DO analysis of frailty-related genes. GO, gene ontology; DO, disease ontology.


PPI Network Analysis

The PPI information of the STRING database was used to remove incomplete and self-interaction information. A PPI network related to frailty was constructed (Figure 2A). The network contained 163 nodes and 1244 edges, with an average node degree of 346, expected number of edges of 15.3, average node degree of 15.3, local clustering coefficient of 0.559, and number of edges of 488. The number of edges was greater than the expected number of edges, and the PPI enrichment P value of the PPI network was less than 1.0 × 10−16. The top 50 key genes identified based on the number of edges connected in the network where the protein was located (Figure 2B) included the nutrition metabolism genes ALB and INS, myokine gene MSTN, and adipokine gene LEP (leptin). These genes are all essential nodes in the PPI network.

Figure 2. Network of PPIs related to frailty and the top 50 key proteins

A. GO analysis of frailty-related genes. B. DO analysis of frailty-related genes. GO, gene ontology; DO, disease ontology.


Frailty biomarkers and myokines

Based on the GO and disease ontology analyses with PPI as the mainline, five biomarkers related to muscle formation and aging were selected (Table S1). Based on the top 50 targets of the PPI network, the proteins selected in the present study were leptin, MSTN, DCN (an MSTN-inhibitory binding protein), downstream irisin (FNDC5), and the regulatory target AMPK.

Concentrations of frailty markers and myokines in peripheral blood

Table S2 compares the serum concentrations of the markers of frailty (leptin, AMPK, irisin, DCN, and MSTN) in the robust, pre-frailty, and frailty groups. Irisin concentrations significantly differed between all groups (P<0.05). The frailty group had significantly higher concentrations of leptin, AMPK, and MSTN than the robust group (all P<0.05). The pre-frailty and frailty groups had significantly lower irisin concentrations than the robust group (both P<0.05). However, the DCN concentration did not differ among groups (P=0.287).

Correlation analysis of frailty status and blood biomarkers across all participants

As shown in Table 2, AMPK was significantly positively correlated with frailty status (p<0.01). Leptin and Irisin showed a negative correlation with frailty status, DCN and MSTN showed a positive correlation with frailty status.

Table 2. Correlation analysis of frailty status vs. blood biomarkers across all participants


Multivariate analysis of factors associated with frailty scores

Table S3 shows the results of multivariate analysis of factors potentially associated with frailty scores. The multivariate analysis results (adjusted R2=0.244) indicated that the factors influencing coefficients of association, from highest to lowest, were the ADL score (0.429), MNA-SF score (0.416), serum albumin concentration (0.317), urination function (0.243), hearing function (0.214), leptin concentration (0.187), GDS-15 score (0.164), MSTN concentration (0.135), vision function (0.118), sex (0.089), AMPK concentration (0.072), irisin concentration (0.069), age (0.066), DCN concentration (0.036), and hemoglobin concentration (0.020).



This prospective study aimed to explore the association between myokines and frailty according to the GEO and Genecards datasets, and then prospectively collected blood samples from older inpatients to examine the correlations between the blood myokines and frailty. The prevalence of frailty among inpatients aged ≥ 65 years during 30 days of follow-up was 13.21%. The frailty group had significantly higher concentrations of leptin, AMPK, and MSTN than the robust group. Additionally, AMPK concentration was significantly positively correlated with frailty. The pre-frailty and frailty groups had significantly lower concentrations of irisin than the robust group, whereas the DCN concentration did not differ among groups. Among the 15 factors influencing the coefficients of association, the top 50% were the ADL score, MNA-SF score, serum albumin concentration, urination function, hearing function, leptin concentration, GDS-15 score, and MSTN concentration.
In the current study, the prevalence of frailty was 13.21% and the prevalence of pre-frailty was 69.81%. Similarly, a systematic review and meta-analysis of 46 observational studies from 28 countries or regions showed that the prevalence of frailty in community-dwelling individuals aged ≥ 60 years was 13.6% (29). In contrast, a cross-sectional survey of the relationship between nutritional status and frailty using the MNA-SF and Fried phenotype indicated that the prevalence of frailty was as high as 26.5% in community-dwelling undernourished older adults (mean age 66.6 ± 7.8 years) (30). The differences between studies in the prevalence of frailty reflects the differences in frailty measurement tools, theoretical basis, study design, and study participants.
The prevalence of pre-frailty was 80% among cardiology inpatients, and 13.95% of frailty among gastroenterology inpatients. The prevalence was higher than the pooled prevalence of pre-frailty (69.81%) and frailty (13.21%) among the entire study cohort. A recent review article indicated that frailty is a cardiovascular syndrome, both of them are interdependent and have the same physiological underpinning that predisposes to the progression of both disease processes (31). The aging gastrointestinal system and microbiome undergo numerous compositional changes that adversely affect digestive health and absorption, which may contribute to frailty (32). Therefore, it is important for cardiovascular and gastrointestinal practices to actively identify patients with frailty who could benefit from frailty interventions.
Myokines regulate several processes associated with sarcopenia, such as slowness, muscle dynapenia and wasting, through autocrine, paracrine, and endocrine mechanism s(12-14). Sarcopenia is regarded as a key component of frailty (8), thus, we speculate that myokines may be the biomarkers of frailty among older adults. The current study indicated that the AMPK concentration was significantly positively correlated with frailty. AMPK is a trimeric complex that is widely distributed in the human body and is a serine/threonine-protein kinase, known as an energy switch (33). AMPK is widely expressed and highly conserved, and is involved in the regulation of various metabolic reactions and signaling pathways that are closely related to the courses of various diseases (34). Recent studies have evaluated the relationships between AMPK and diabetes, obesity, and cancer. The AMPK signaling pathway is involved in the occurrence and development of sarcopenia and frailty (35, 36). Although AMPK activation is not directly involved in muscle contraction or stretch-stimulated increases in protein synthesis, it regulates autophagy to reduce muscle atrophy caused by dysfunctional organelles (35). Dysregulation of autophagic flux with age inhibits lysosomal storage processes involved in muscle biogenesis. Exercise-induced autophagy may have interventional effects on the progression of frailty and sarcopenia by regulating Akt/mTOR and Akt/FoxO3a signaling pathways and AMPK-mediated mitochondrial quality control (37). Monitoring of serum AMPK concentrations before and after intervention may provide evidence for the effect of exercise intervention on frailty and sarcopenia.
Leptin is an appetite-suppressing adipokine and a hormone that suppresses hunger to maintain energy balance (38), which links food intake with energy expenditure and body composition, and increases muscle mass by reducing the expression of muscle atrophy-related factors, such as MSTN, MuRF1, and F-box. The leptin concentration was significantly higher in the frailty group than the robust group. Relatively high concentrations of leptin may be associated with chronic inflammation in debilitated patients, as subclinical inflammation increases leptin concentrations (38). When the aging immune function fails to respond appropriately to stress, this manifests as a chronic excessive response to inflammatory stimuli, even after the stimulus is eliminated, and inflammation is related to the catabolism of skeletal muscle and fat. This further aggravates the clinical manifestations of weakness, such as sarcopenia and weight loss (39), and promotes a relatively high expression of leptin (40, 41). Leptin is a confirmed marker of aging; therefore, research on the changes in the sensitivity to leptin with age may help to understand the age-related degeneration of multiple organs, including skeletal muscle, and to understand age-related factors such as frailty and disease (39).
MSTN is a peptide that is produced, expressed, and released by myocytes, affects the physiology of muscle and other organs and tissues, and is a negative regulator of skeletal muscle (42). MSTN inhibits muscle cell proliferation, is involved in the regulation of energy and glucose and lipid metabolism, and may have autocrine, paracrine, and endocrine effects on metabolic processes involving lipids (e.g., lipolysis, browning of fat cells, lack of fatty acid oxidation), muscle (e.g., glucose uptake), liver (e.g., glycogenolysis and glycogen metabolism), insulin sensitivity, and stimulation of angiogenesis (43). From embryonic development to adulthood, MSTN negatively regulates mass, impairs muscle synthesis, and increases muscle catabolism (44-46). MSTN also induces muscle loss by stimulating and activating the ubiquitin-proteasome system and attenuating the activation of protein kinase B (Akt), satellite cells, and myogenic factors such as MyoD, leading to decreased muscle mass, sarcopenia, and exacerbation of debilitating clinical manifestations (47).
The pre-frailty and frailty older people showed significantly lower serum irisin concentrations than the robust people. A previous study with older women also reported similar results (48). Thus, the development of frailty may influence the synthesis and release of myokines. DCN is thought to act as a counter-regulator of myostatin by binding and inactivating myostatin (49), but there is no significant difference between the robust, pre-frailty, and frailty groups. These results may be because sarcopenia (changes in muscle mass, power, and strength) is regarded as a key component of frailty (8), thus there are significant changes in the protein secretion level or mRNA of DCN in muscles, and the changes of serum DCN are not as significant as those in the muscles (50). Additionally, even though previous research suggested that plasma DCN is increased after acute and chronic exercise (49), Willoughby et. al conducted nutritional intervention combined with exercise in older women whereas no significant difference in the serum DCN levels compared to the control group (51), emphasizing that serum DCN might not be the frailty serum biomarkers. The difference in these results may be due to the limited sample sizes, studies with larger sample sizes and more sophisticated analysis are needed in the future.
Our study also examined the factors influencing the coefficients of association. The serum albumin concentration was among the top three factors influencing the development of frailty, which is consistent with a previous cross-sectional study that identified clinical markers including serum albumin concentration as frailty biomarkers (16). Additionally, the results of this study also suggested that ADL and nutritional status play major roles in the development of frailty. This support the results that physical exercise combined with nutritional intervention is a critical strategy to improve the frailty status among the elderly (52).
Previous research suggested that the identification of frailty relies upon multivariate modeling of clinical variables, myokine concentrations, functional variables, and imaging parameters (16). This study attempted to use the multi-marker approaches combined with multivariate modeling to show great potential for addressing the complexity of the pathophysiology of frailty and unveiling novel targets for interventions (53).

Study limitations

This study has some limitations. The ability of the identified myokines to predict frailty needs to be verified in a longitudinal study of the effects on long-term adverse clinical outcomes. Second, well-designed longitudinal studies are needed to enable the incorporation of reliable biomarkers into clinical practice and identify novel targets for frailty interventions. Third, the small sample size might result in low statistical power and limit comparisons with other studies. We previously found that 297 frailty events occurred during the first 4 weeks of follow-up with an absolute rate of 13.54% (N=2194) (9), and the incidence of frailty was 13.21% in this study, which is slightly lower than the previous study. As a result, further large-sample longitudinal studies should be conducted to validate the findings.



A chronic subclinical proinflammatory state leads to a variety of complex immune and hormonal changes in the geriatric population. Proinflammatory myokines, particularly leptin, MSTN, and AMPK, exert detrimental effects on muscle mass and strength during the development of frailty in older adults. Our results support the idea that identification of frailty relies upon multivariate modeling of clinical variables, myokine concentrations, and functional variables (16, 53). Deeper understanding of these frailty-related factors will reveal the pathophysiology of frailty and unveil novel targets for interventions.


Acknowledgments: The authors are grateful to all the persons who participated in the collection of the data. We thank Kelly Zammit, BVSc, from Liwen Bianji (Edanz) (www.liwenbianji.cn/), for editing the English text of a draft of this manuscript. We thank Wei Chen, PhD, from Peking Union Medical College Hospital for helping to give directions and analyze the myokines.

Funding: This paper was supported by the Project funded by China Postdoctoral Science Foundation (grant number 2022TQ0016 and 2022M720298).

Conflict of interest: The authors have no conflicts of interest to declare.

Consent for publication: Not applicable.

Availability of data and materials: The datasets generated for this study are available on request to the corresponding author.

Authors’ contributions: HL and XWu conceived and designed this study and reviewed the manuscript, HL and WL prepared and edited the manuscript. HL, WL, and MZ performed statistical analyses and drafted the tables. XWen, JJin, HW, DL, SZ, and JJiao recruited participants, collected data, and edited the manuscript.

Ethical guidelines statement: This study was approved by the Ethics Committee of Peking Union Medical College Hospital (S-K540 and JS-2781). All procedures were conducted in accordance with the principles of the Declaration of Helsinki. Written informed consent was provided by all participants or their legal guardians.

Trial registration: Chinese Clinical Trial Registry (ChiCTR1800017682, August 9, 2018; ChiCTR2100044148, March 11, 2021).





1. Kemoun P, Ader I, Planat-Benard V, et al. A gerophysiology perspective on healthy ageing. Ageing Res Rev 2022;73:101537. doi: 10.1016/j.arr.2021.101537
2. Martínez-Reig M, Gómez-Arnedo L, Alfonso-Silguero SA, et al. Nutritional risk, nutritional status and incident disability in older adults. The FRADEA study. J Nutr Health Aging 2014;18(3):270-6. doi: 10.1007/s12603-013-0388-x
3. Inzitari M, Doets E, Bartali B, et al. Nutrition in the age-related disablement process. J Nutr Health Aging 2011;15(8):599-604. doi: 10.1007/s12603-011-0053-1
4. Dent E, Hoogendijk EO, Visvanathan R, et al. Malnutrition Screening and Assessment in Hospitalised Older People: a Review. J Nutr Health Aging 2019;23(5):431-41. doi: 10.1007/s12603-019-1176-z
5. Morley JE, Malmstrom TK. Frailty, sarcopenia, and hormones. Endocrinol Metab Clin North Am 2013;42(2):391-405. doi: 10.1016/j.ecl.2013.02.006
6. Hoogendijk EO, Afilalo J, Ensrud KE, et al. Frailty: implications for clinical practice and public health. Lancet 2019;394(10206):1365-75. doi: 10.1016/s0140-6736(19)31786-6 [published Online First: 2019/10/15]
7. Dent E, Martin FC, Bergman H, et al. Management of frailty: opportunities, challenges, and future directions. Lancet 2019;394(10206):1376-86. doi: 10.1016/s0140-6736(19)31785-4
8. Clegg A, Young J, Iliffe S, et al. Frailty in elderly people. The Lancet 2013;381(9868):752-62. doi: 10.1016/s0140-6736(12)62167-9
9. Liu H, Li C, Jiao J, et al. Development and validation of risk prediction model for identifying 30-day frailty in older inpatients with undernutrition: A multicenter cohort study. Frontiers in Nutrition 2023;9 doi: 10.3389/fnut.2022.1061299
10. Calvani R, Marini F, Cesari M, et al. Biomarkers for physical frailty and sarcopenia: state of the science and future developments. J Cachexia Sarcopenia Muscle 2015;6(4):278-86. doi: 10.1002/jcsm.12051
11. Coelho-Junior HJ, Picca A, Calvani R, et al. If my muscle could talk: Myokines as a biomarker of frailty. Exp Gerontol 2019;127:110715. doi: 10.1016/j.exger.2019.110715
12. Pedersen BK. Physical activity and muscle-brain crosstalk. Nat Rev Endocrinol 2019;15(7):383-92. doi: 10.1038/s41574-019-0174-x
13. Pedersen BK, Steensberg A, Fischer C, et al. The metabolic role of IL-6 produced during exercise: is IL-6 an exercise factor? Proc Nutr Soc 2004;63(2):263-7. doi: 10.1079/pns2004338
14. Pedersen L, Hojman P. Muscle-to-organ cross talk mediated by myokines. Adipocyte 2012;1(3):164-67. doi: 10.4161/adip.20344
15. Colaianni G, Cinti S, Colucci S, et al. Irisin and musculoskeletal health. Annals of the New York Academy of Sciences 2017;1402(1):5-9. doi: 10.1111/nyas.13345
16. Kane AE, Sinclair DA. Frailty biomarkers in humans and rodents: Current approaches and future advances. Mech Ageing Dev 2019;180:117-28. doi: 10.1016/j.mad.2019.03.007
17. Liu H, Jiao J, Zhu C, et al. Associations Between Nutritional Status, Sociodemographic Characteristics, and Health-Related Variables and Health-Related Quality of Life Among Chinese Elderly Patients: A Multicenter Prospective Study. Frontiers in Nutrition 2020;7(192) doi: 10.3389/fnut.2020.583161
18. Jiao J, Y W, C Z, et al. Prevalence and associated factors for frailty among elder patients in China: a multicentre cross-sectional study. BMC geriatrics 2020;20(1):100. doi: 10.1186/s12877-020-1496-1
19. T X, J J, C Z, et al. Prevalence and Potential Associated Factors of Depression among Chinese Older Inpatients. The journal of nutrition, health & aging 2019;23(10):997-1003. doi: 10.1007/s12603-019-1270-2
20. Pascual JC, Belinchon I, Ramos JM. Use of the Barthel index, activities of daily living, in dermatologic surgery in patients aged 80 years and older. Int J Dermatol 2015;54(2):222-6. doi: 10.1111/ijd.12680
21. Shiao CC, Hsu HC, Chen IL, et al. Lower Barthel Index Is Associated with Higher Risk of Hospitalization-Requiring Pneumonia in Long-Term Care Facilities. Tohoku J Exp Med 2015;236(4):281-8. doi: 10.1620/tjem.236.281
22. Liu H, Jiao J, Zhu C, et al. Potential associated factors of functional disability in Chinese older inpatients: a multicenter cross-sectional study. BMC Geriatr 2020;20(1):319. doi: 10.1186/s12877-020-01738-x
23. Liu H, Song B, Jin J, et al. Nutritional Risk, Health Outcomes, and Hospital Costs Among Chinese Immobile Older Inpatients: A National Study. Frontiers in Nutrition 2021;8(1026) doi: 10.3389/fnut.2021.758657
24. Liu H, Jiao J, Zhu M, et al. Nutritional Status According to the Short-Form Mini Nutritional Assessment (MNA-SF) and Clinical Characteristics as Predictors of Length of Stay, Mortality, and Readmissions Among Older Inpatients in China: A National Study. Frontiers in Nutrition 2022;9 doi: 10.3389/fnut.2022.815578
25. Arthur AJ, Jagger C, Lindesay J, et al. Evaluating a mental health assessment for older people with depressive symptoms in general practice: a randomised controlled trial. Br J Gen Pract 2002;52(476):202-7.
26. Zhao H, He J, Yi J, et al. Factor Structure and Measurement Invariance Across Gender Groups of the 15-Item Geriatric Depression Scale Among Chinese Elders. Front Psychol 2019;10:1360. doi: 10.3389/fpsyg.2019.01360
27. Lopez D, Flicker L, Dobson A. Validation of the frail scale in a cohort of older Australian women. J Am Geriatr Soc 2012;60(1):171-3. doi: 10.1111/j.1532-5415.2011.03746.x
28. Piñero J, Bravo À, Queralt-Rosinach N, et al. DisGeNET: a comprehensive platform integrating information on human disease-associated genes and variants. Nucleic acids research 2017;45(D1):D833-d39. doi: 10.1093/nar/gkw943
29. Ofori-Asenso R, Chin KL, Mazidi M, et al. Global Incidence of Frailty and Prefrailty Among Community-Dwelling Older Adults: A Systematic Review and Meta-analysis. JAMA Netw Open 2019;2(8):e198398. doi: 10.1001/jamanetworkopen.2019.8398
30. Wei K, Nyunt MSZ, Gao Q, et al. Frailty and Malnutrition: Related and Distinct Syndrome Prevalence and Association among Community-Dwelling Older Adults: Singapore Longitudinal Ageing Studies. Journal of the American Medical Directors Association 2017;18(12):1019-28. doi: 10.1016/j.jamda.2017.06.017
31. Ijaz N, Buta B, Xue QL, et al. Interventions for Frailty Among Older Adults With Cardiovascular Disease: JACC State-of-the-Art Review. J Am Coll Cardiol 2022;79(5):482-503. doi: 10.1016/j.jacc.2021.11.029
32. Haran JP, McCormick BA. Aging, Frailty, and the Microbiome-How Dysbiosis Influences Human Aging and Disease. Gastroenterology 2021;160(2):507-23. doi: 10.1053/j.gastro.2020.09.060
33. Janovska A, Hatzinikolas G, Staikopoulos V, et al. AMPK and ACC phosphorylation: effect of leptin, muscle fibre type and obesity. Mol Cell Endocrinol 2008;284(1-2):1-10. doi: 10.1016/j.mce.2007.12.013
34. Bermeo K, Castro H, Arenas I, et al. AMPK mediates regulation of voltage-gated calcium channels by leptin in isolated neurons from arcuate nucleus. Am J Physiol Endocrinol Metab 2020;319(6):E1112-E20. doi: 10.1152/ajpendo.00299.2020
35. Higashida K, Inoue S, Nakai N. Iron deficiency attenuates protein synthesis stimulated by branched-chain amino acids and insulin in myotubes. Biochem Biophys Res Commun 2020;531(2):112-17. doi: 10.1016/j.bbrc.2020.07.041
36. Garcia D, Shaw RJ. AMPK: Mechanisms of Cellular Energy Sensing and Restoration of Metabolic Balance. Molecular cell 2017;66(6):789-800. doi: 10.1016/j.molcel.2017.05.032
37. Zeng Z, Liang J, Wu L, et al. Exercise-Induced Autophagy Suppresses Sarcopenia Through Akt/mTOR and Akt/FoxO3a Signal Pathways and AMPK-Mediated Mitochondrial Quality Control. Front Physiol 2020;11:583478. doi: 10.3389/fphys.2020.583478
38. Li CW, Yu K, Shyh-Chang N, et al. Circulating factors associated with sarcopenia during ageing and after intensive lifestyle intervention. J Cachexia Sarcopenia Muscle 2019;10(3):586-600. doi: 10.1002/jcsm.12417
39. Steinberg GR, Parolin ML, Heigenhauser GJ, et al. Leptin increases FA oxidation in lean but not obese human skeletal muscle: evidence of peripheral leptin resistance. Am J Physiol Endocrinol Metab 2002;283(1):E187-92. doi: 10.1152/ajpendo.00542.2001
40. Hamrick MW, Herberg S, Arounleut P, et al. The adipokine leptin increases skeletal muscle mass and significantly alters skeletal muscle miRNA expression profile in aged mice. Biochem Biophys Res Commun 2010;400(3):379-83. doi: 10.1016/j.bbrc.2010.08.079
41. Pereira S, Cline DL, Glavas MM, et al. Tissue-Specific Effects of Leptin on Glucose and Lipid Metabolism. Endocrine reviews 2021;42(1):1-28. doi: 10.1210/endrev/bnaa027
42. Suh J, Kim NK, Lee SH, et al. GDF11 promotes osteogenesis as opposed to MSTN, and follistatin, a MSTN/GDF11 inhibitor, increases muscle mass but weakens bone. Proc Natl Acad Sci U S A 2020;117(9):4910-20. doi: 10.1073/pnas.1916034117
43. Diaz BB, Gonzalez DA, Gannar F, et al. Myokines, physical activity, insulin resistance and autoimmune diseases. Immunol Lett 2018;203:1-5. doi: 10.1016/j.imlet.2018.09.002
44. Tobin JF, Celeste AJ. Myostatin, a negative regulator of muscle mass: implications for muscle degenerative diseases. Curr Opin Pharmacol 2005;5(3):328-32. doi: 10.1016/j.coph.2005.01.011
45. Carnac G, Ricaud S, Vernus B, et al. Myostatin: biology and clinical relevance. Mini Rev Med Chem 2006;6(7):765-70. doi: 10.2174/138955706777698642
46. Guo A, Li K, Xiao Q. Sarcopenic obesity: Myokines as potential diagnostic biomarkers and therapeutic targets? Experimental gerontology 2020;139:111022. doi: 10.1016/j.exger.2020.111022
47. Durieux AC, Amirouche A, Banzet S, et al. Ectopic expression of myostatin induces atrophy of adult skeletal muscle by decreasing muscle gene expression. Endocrinology 2007;148(7):3140-7. doi: 10.1210/en.2006-1500
48. Park HS, Kim HC, Zhang D, et al. The novel myokine irisin: clinical implications and potential role as a biomarker for sarcopenia in postmenopausal women. Endocrine 2019;64(2):341-48. doi: 10.1007/s12020-018-1814-y
49. Kanzleiter T, Rath M, Görgens SW, et al. The myokine decorin is regulated by contraction and involved in muscle hypertrophy. Biochem Biophys Res Commun 2014;450(2):1089-94. doi: 10.1016/j.bbrc.2014.06.123
50. Liu X, Zhang N, Sun B, et al. Time-specific effects of acute eccentric exercise on myostatin, follistatin and decorin in the circulation and skeletal muscle in rats. Physiol Res 2022;71(6):783-90. doi: 10.33549/physiolres.934833
51. Willoughby DS, Cardaci TD, Machek SB, et al. Resistance Exercise-Induced Increases in Muscle Myostatin mRNA and Protein Expression Are Subsequently Decreased in Circulation in the Presence of Increased Levels of the Extracellular Matrix Stabilizing Protein Decorin. J Sports Sci Med 2022;21(4):616-24. doi: 10.52082/jssm.2022.616
52. Hsieh TJ, Su SC, Chen CW, et al. Individualized home-based exercise and nutrition interventions improve frailty in older adults: a randomized controlled trial. Int J Behav Nutr Phys Act 2019;16(1):15. doi: 10.1186/s12966-019-0855-9
53. Picca A, Calvani R, Cesari M, et al. Biomarkers of Physical Frailty and Sarcopenia: Coming up to the Place? Int J Mol Sci 2020;21(16) doi: 10.3390/ijms21165635

© Serdi 2024