Associations of high protein supplements with gut microbiota and skeletal muscle mass in hospitalized older people

Abstract

Background/objective

Prolonged bed rest is highly prevalent among hospitalized older adults and markedly accelerates the loss of muscle mass and physical function. Currently, there are no effective interventions to counteract this decline, and the underlying mechanisms remain poorly characterized. This study aimed to investigate whether high protein intake can simultaneously modulate muscle mass and the gut microbiota, and whether gut microbial composition mediates muscle regulation in hospitalized older people.

Methods

A self-controlled study was conducted on 43 older patients aged 60 to 90 years old with low skeletal muscle mass. During the 3-month intervention phase, all participants received approximately 36 g of high-protein supplementation daily, comprising both casein and whey proteins. This was followed by a 3-month control phase in which participants received standard nursing care without protein supplementation.

Results

A significant increase in skeletal muscle mass index from baseline was seen in male group at 3 months (6.0–6.3 kg/m²) but declined to 6.1 kg/m² at 6 months (P < 0.05). No significant changes were observed in females (P > 0.05). Gut microbiota analysis revealed that bacterial diversity and microbial structure were affected by protein supplementation and differed by sex. Males exhibited a greater abundance of SMI- and SMM-associated beneficial bacteria following protein intake. Furthermore, metabolic pathway analysis indicated that microbial functions related to amino acid synthesis were positively correlated with SMI-linked species such as Blautia wexlerae and Corynebacterium dentalis.

Conclusions

High-protein supplementation may promote muscle anabolism in hospitalized older males by modulating the composition and metabolic function of the gut microbiota, specifically by enhancing microbial pathways related to amino acid synthesis. These results suggest the presence of a gut-muscle axis and highlight the potential of targeted protein interventions to counteract inactivity-related muscle loss in older patients.

Trial registration

The trial was registered at Chinese Clinical Trial Registry with identifier ChiCTR2400085432 on 07/06/2024.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12877-025-06546-9.

Introduction

Hospitalized older adults are particularly vulnerable to muscle loss due to prolonged bed rest during their stay [1]. This frequently leads to sarcopenia, a geriatric syndrome marked by progressive decline in both skeletal muscle mass and function, establishing a vicious cycle of functional impairment that substantially diminishes quality of life and elevates risks of morbidity and mortality [2, 3].

The pathophysiology of sarcopenia is multifactorial, involving not only low physical activity but also age-related anabolic resistance, chronic low-grade inflammation, and frequently coexisting protein-energy malnutrition [4]. Although resistance training and gait/balance training appear most effective for anabolic stimulus [1, 5], it is often impractical or inadequate in hospitalized older patients. Therefore, nutritional supplementation, particularly with high-quality protein, emerges as a critical therapeutic strategy to mitigate muscle breakdown and stimulate muscle protein synthesis (MPS) [3, 6]. Previous studies have demonstrated that whey protein has superior effects in promoting MPS for older adults [7, 8]. In contrast, casein has traditionally been characterized as a “slow” protein due to its delayed digestion and absorption kinetics, resulting in a less pronounced but prolonged increase in postprandial muscle protein synthesis rates [9]. Nevertheless, casein intake may offer benefits by sustaining elevated plasma amino acid concentrations [10] and promoting postprandial protein deposition through the suppression of proteolysis [11]. Despite these insights, evidence regarding the efficacy of blended whey and casein supplements for preserving muscle mass remains limited, and the mechanistic underpinnings of protein-mediated effects on sarcopenia in bedridden older adults are still not well-defined.

The human gut microbiota consists of 10–100 trillion microorganisms, playing a vital role in intestinal immune and endocrine functions, energy homeostasis, nutritional status, and health maintenance [12]. Recently, emerging evidence suggests that the gut microbiota is also associated with muscle breakdown and frailty in older individuals, raising the interests in concept of the ‘gut–muscle axis’ [12–14]. Gut microbiota dysbiosis could reduce protein absorption and lead to anabolic resistance [15]. Compared to non-sarcopenia controls, the sarcopenic patients were generally poor in bacteria producing SCFAs, which indicated that decreased abundance of butyrate-producing bacteria may exacerbate the muscle loss in aging [14]. Nutritional supplements regulate muscle and gut microbes simultaneously, but whether the gut bacteria are mediators in muscle regulation still remains unclear [12]. Thus, exploring the interaction of gut microbiota with high protein intake and the host muscle mass will provide us new foundations for the design of microbe-based therapies.

Therefore, this study aimed to explore the effects of blended whey and casein protein supplementation on gut microbiota and muscle health in hospitalized older adults. It was hypothesized that higher intakes of blended protein have a positive effect on muscle mass, perhaps via modulation of gut microbiota.

Methods

Sample size

We based our sample size calculation from published study [16] and considered an expected mean ± SD increase in muscle mass of 1.1 ± 1.2 kg after high-protein supplements, with a power of 90% and a two-sided alpha of 5%, as well as 20% attrition. This gave a sample size of 19 patients.

Participants

This study recruited 48 patients aged 60 to 90 years from the inpatient department of Shanghai Ideal Nursing Hospital. All participants had low skeletal muscle mass, defined according to the Asian Working Group for Sarcopenia (AWGS) 2019 consensus criteria [17] as a skeletal muscle index (SMI) of < 5.7 kg/m² for women and < 7.0 kg/m² for men, which was measured using bioelectrical impedance analysis (BIA; InBody 720). Participants were excluded if they were with recent hospitalizations (within 1 month) for acute condition; had mental disorders, had severe kidney or liver failure, or malignancies, using medication interfering with nutrients absorption or gut microbiota (e.g., PPIs, antibiotics, probiotics, corticosteroids and anticoagulants) within the past month; had severe gastrointestinal diseases or gastrointestinal surgery in the past 3 months. Before intervention, five of the 48 participants were excluded because of declining to receive high-protein oral nutritional supplements (ONS) or collect blood, or refusing to participate, thus 43 participants were included in the intervention period. Of the 43 participants receiving intervention, 37 (89.6%) were followed up and received the whole intervention at 3 months, and 28 (65.1%) were followed up without intervention at 6 months. The detailed flowchart is shown in Fig. 1.

Fig. 1.

Flow Diagram of Participants Through the Study

The study was conducted in accordance with the ethical principles described in the Declaration of Helsinki and all procedures involving human subjects were approved by the Research Ethics Board of Shanghai Fourth People’s Hospital, Tongji University School of Medicine. Written informed consent was obtained from all subject before the study.

Standard hospital care

All patients received standard hospital care, which consisted of routine nursing support—including monitoring of vital signs, assistance with activities of daily living (ADLs), and medication administration, as well as physical rehabilitation comprising simple, non-resistance-based exercises.

Diet and ONS intervention

This was a self-controlled study. The research period was divided into an interventional stage from December to February (0–3 months) and a control stage from March to May (4–6 months). During the intervention, all participants received identical, standardized meals (approximately 1800 kcal/day, with 30%, 16%, and 54% of energy from fat, protein, and carbohydrates, respectively), prepared by the hospital dietary service according to a fixed menu, and were instructed to consume two servings of ONS daily for three consecutive months. To account for variations in appetite, nurses took photographs of participants’ plates after each meal to visually document food leftovers. The nutritionist estimated approximate intakes based on these images. This assessment confirmed that all subjects consumed at least 80% of the provided meals. The ONS used in this study was a high-protein supplement composed of 84% protein (from a 4:1 ratio of casein to hydrolyzed whey protein), providing approximately 85 kcal and 18 g of protein per serving. The remaining components comprised 5% moisture, 3% carbohydrate, 1.5% fat, and 6% minerals, which were primarily included for product stability and palatability rather than for targeted nutritional intervention. The detailed ingredient list is provided in Supplementary Table 3. The ONS was administered twice daily (at 10:00 AM and 3:00 PM) as between-meal snacks under the supervision of nurses to ensure compliance. In the control stage, participants received only standardized meals.

Outcome measures

For all enrolled participants, we collected data on sex, age, medical history, and medication use both at baseline and throughout follow-up visits. Body composition, blood parameters, and fecal samples were assessed during each designated visit.

Measurement of body composition

Body weight and height were measured for each participant while wearing light clothing and no shoes, and body mass index (BMI) was calculated as the weight in kilograms divided by the square of the height in meters. Whole-body composition including body fat mass, skeletal muscle mass (SMM) and appendicular skeletal muscle mass (ASMM) were assessed by using bioelectrical impedance analysis (BIA) with a Tanita BF-300 device at baseline (T0), 3 months (T3) and 6 months (T6). Skeletal muscle mass indices (SMI) were calculated as below: SMI = ASMM (kg) ∕height (m)².

Measurement of blood parameters

All blood samples were drawn under fasting conditions and collected through two vacutainer tubes (one for plasma and the other for serum) containing EDTA. After extraction, the samples were centrifuged at 1500 rpm for 10 min at room temperature. Then the hemogram and several biochemical parameters related to nutritional level or inflammatory status or safety assessment (including urea, creatinine, total proteins, albumin, globulin, total cholesterol (TC), triglyceride (TG), high-density lipoprotein (HDL), low-density lipoprotein (LDL), vitamin D, vitamin C, vitamin E, CRP and IL-6) were measured immediately in clinical laboratories belonging to the hospital.

Measurement of gut microbiota composition

Stool samples were collected from participants at each visit and transported to the research facility in coolers with ice packs within 12 h. The samples were frozen at − 80 °C within 10 min of collection and stored until DNA extraction. The gut microbiota composition was determined via 16 S rRNA gene sequencing analysis as described previously [18]. Briefly, the Magic Pure Stool and Soil Genomic DNA Kit (TransGen) was used to extract genomic DNA from stool samples. The V3-V4 region of 16 S rRNA gene was amplified using primers 338 F(ACTCCTACGGGAGGCAGCAG) and 806R (GGACTACHVGGGTWTCTAAT). The PCR program for target genes was optimized as follows: denaturation at 95 °C for 2 min, 20 cycles of amplification (45 s at 95 °C, 30 s at 55 °C, and 30 s at 72 °C), and extension 72 °C for 5 min. The PCR products were detected via 2% agarose gel electrophoresis and were subsequently purified with an AxyPrep DNA Gel Extraction Kit (AXYGEN). The amplicons were then equally pooled and sequenced on an Illumina NextSeq2000 instrument (2 × 300 cycles) according to standard protocols.

Statistical analysis

Descriptive statistics were calculated, and values were presented as means and standard deviations if normally distributed. The paired-samples t-test was used to assess changes in blood parameters and body composition over time (baseline to post-intervention) between groups. Then participants who received at least 1 follow-up visit were included in the linear mix-effects model to examine repeated measures of ASMM and SMI at baseline (T0), 3 months (T3) and 6 months (T6). In fitting the model, we scaled each variable and measured time as fixed effects. Participant ID was entered as a random effect.

For gut microbiota analysis, normalized amplicon sequence variants (ASVs) were used to assess microbial richness (ACE index), evenness (Shannon evenness), and diversity (Shannon index). Differences in bacterial community composition over time were evaluated using permutational multivariate analysis of variance (PERMANOVA). Microbial functional profiling was inferred with PICRUSt2. Hierarchical all-against-all association analysis (HALLA) was applied to identify correlations between functional profiles and microbiome data, with false discovery rate (FDR)-adjusted P < 0.05 considered significant. Differential abundance of microbial taxa and functional pathways was determined using Microbiome Multivariable Associations with Linear Models (MaAsLin2), adjusting for gender, age, and BMI. Features with an adjusted P < 0.25 (default significance threshold) and minimum abundance of 0.0005 were retained. All analyses were performed using R Studio (version 4.0; Posit Software, PBC) and SPSS (version 27.0; IBM Corp., Armonk, NY, USA).

Results

Patient’s characteristics

The baseline and post-intervention characteristics by sex are presented in Table 1 and Supplementary Table 1. The cohort comprised 47% males and 53% females, with a mean age of 76.7 ± 10.1 years and a mean BMI of 20.8 ± 2.8 kg/m². As measured by BIA, all participants had low SMI, with average values of 5.9 ± 1.0 kg/m² and 4.8 ± 1.1 kg/m² for males and females, respectively. Elevated mean levels of IL-6 and CRP, which were outside the normal reference range (Supplementary Table 2), indicated a chronic inflammatory state at both baseline and post-intervention.

Table 1.

Baseline characteristics of the participants divided by sex

Characteristic	Total (n = 43)	Male (n = 20)	Female (n = 23)	P-value
Age, y	76.7 ± 10.1	75.9 ± 11.0	77.4 ± 11.4	0.635
Weight, kg	56.5 ± 9.8	62.7 ± 7.8	51.7 ± 8.5	< 0.001
Height, cm	164.3 ± 7.4	170.1 ± 5.6	159.9 ± 5.3	< 0.001
BMI, kg/m2	20.8 ± 2.8	21.7 ± 2.6	20.1 ± 2.8	0.415
Fat mass, kg	22.8 ± 7.8	21.3 ± 6.7	24.5 ± 8.7	< 0.001
SMM, kg	19.5 ± 5.4	22.3 ± 4.6	15.6 ± 3.8	< 0.001
ASMM, kg	14.1 ± 4.0	17.1 ± 3.2	11.6 ± 2.5	< 0.001
SMI, kg/m2	5.3 ± 1.1	5.9 ± 1.0	4.8 ± 1.1	< 0.001
Total protein, g/L	62.7 ± 6.4	63.5 ± 4.9	62.0 ± 7.5	0.454
Albumin, g/L	37.9 ± 7.6	37.2 ± 4.4	38.4 ± 9.6	0.601
Globulin, g/L	25.8 ± 6.3	26.3 ± 5.0	25.3 ± 7.3	0.611
Prealbumin, mg/L	188.7 ± 80.3	188.8 ± 68.8	188.7 ± 61.0	0.997
IL-6, pg/mL	17.1 ± 17.7	19.9 ± 21.0	14.8 ± 14.5	0.361
CRP, mg/L	14.8 ± 20.0	12.6 ± 14.5	16.7 ± 24.1	0.572
Vitamin D, nmol/mL	75.9 ± 13.3	75.8 ± 13.9	76.1 ± 13.1	0.945
HDL-C, mmol/L	1.2 ± 0.5	1.1 ± 0.3	1.3 ± 0.6	0.181
LDL-C, mmol/L	2.5 ± 1.0	2.3 ± 0.8	2.6 ± 1.0	0.335
TC, mmol/L	3.7 ± 1.0	3.5 ± 0.8	3.8 ± 1.1	0.342
TG, mmol/L	1.1 ± 0.4	1.1 ± 0.4	1.1 ± 0.3	0.372
Urea, mmol/L	5.5 ± 1.8	5.9 ± 1.7	5.1 ± 1.9	0.942
Creatinine, µmol/L	57.1 ± 21.7	65.5 ± 22.5	46.4 ± 15.6	0.048

Paired samples t-test analysis

The changes in body composition and blood parameters before and after the 3-month intervention are displayed in Table 2. Significant improvements in the mean SMM, ASMM and SMI from baseline were observed only in the male group at 3 months (22.4–24.0 kg, 17.3–18.5 kg and 6.0–6.3 kg/m², respectively). Furthermore, blood indicators including total protein, globulin, and Vitamin D also increased significantly, while IL-6, CRP and LDL-C decreased in the male group (P < 0.05). In the female group, functional outcomes showed no improvement from baseline to 3 months, except for significant increases in total protein and globulin (P < 0.05). Analysis of prespecified safety outcomes showed that the high-protein intervention did not significantly affect urea or creatinine levels in either the male or female groups (P > 0.05). Overall, these results suggest a more favorable response to the intervention in males compared to females.

Table 2.

Within-Group changes in the outcomes during the 3-months intervention period

Characteristic	Total			Male			Female
	before	after	P	before	After	P	Before	after	p
Fat mass, kg	22.9 ± 7.3	21.5 ± 8.3	0.220	21.3 ± 6.7	19.6 ± 7.8	0.191	24.5 ± 8.7	23.4 ± 10.5	0.595
SMM, kg	19.5 ± 5.4	20.7 ± 5.5	0.092	22.4 ± 4.6	24.0 ± 3.9	0.048	15.6 ± 3.8	16.2 ± 3.8	0.650
ASMM, kg	15.0 ± 3.6	15.6 ± 4.1	0.064	17.3 ± 2.5	18.5 ± 2.7	0.033	12.2 ± 2.5	12.2 ± 2.7	0.978
SMI, kg/m2	5.4 ± 1.0	5.6 ± 1.1	0.077	6.0 ± 0.7	6.3 ± 0.6	0.032	4.7 ± 0.9	4.7 ± 0.9	0.982
Total protein, g/L	62.2 ± 7.6	67.0 ± 7.8	0.003	63.8 ± 5.8	67.6 ± 8.0	0.046	60.0 ± 9.4	66.2 ± 8.0	0.008
Albumin, g/L	37.8 ± 9.5	36.5 ± 6.6	0.448	36.9 ± 4.8	36.4 ± 8.1	0.662	39.0 ± 13.5	36.7 ± 4.3	0.541
Globulin, g/L	26.0 ± 7.6	31.3 ± 5.9	< 0.001	26.8 ± 5.7	31.0 ± 4.5	0.013	24.8 ± 9.7	31.6 ± 7.6	0.002
Prealbumin, mg/L	188.7 ± 80.3	190.8 ± 87.2	0.365	188.8 ± 68.8	198.1 ± 87.2	0.147	188.7 ± 61.0	186.6 ± 81.3	0.439
IL-6, pg/mL	16.5 ± 17.2	9.6 ± 5.0	0.068	18.2 ± 20.0	8.2 ± 4.9	0.036	14.3 ± 13.8	11.4 ± 4.7	0.533
CRP, mg/L	14.4 ± 17.4	8.1 ± 10.6	0.145	17.2 ± 19.5	7.7 ± 12.2	0.046	9.4 ± 13.8	8.8 ± 0.7	0.942
Vitamin D, nmol/L	73.8 ± 13.4	101.4 ± 31.2	< 0.001	73.6 ± 13.3	106.8 ± 35.1	0.002	74.0 ± 14.3	94.5 ± 25.3	0.052
HDL-C, mmol/L	1.1 ± 0.3	1.0 ± 0.3	0.026	1.1 ± 0.3	1.0 ± 0.4	0.071	1.1 ± 0.3	1.0 ± 0.2	0.209
LDL-C, mmol/L	2.4 ± 1.0	2.0 ± 0.7	0.004	2.2 ± 0.7	1.9 ± 0.7	0.015	2.8 ± 1.3	2.2 ± 0.8	0.117
TC, mmol/L	3.7 ± 1.1	3.3 ± 0.7	0.048	3.4 ± 0.8	3.2 ± 0.8	0.125	4.1 ± 1.4	3.5 ± 0.7	0.178
TG, mmol/L	1.1 ± 0.4	1.0 ± 0.4	0.167	1.1 ± 0.4	1.0 ± 0.4	0.138	1.1 ± 0.4	1.1 ± 0.5	0.647
Urea, mmol/L	5.5 ± 1.8	7.1 ± 4.1	0.015	5.9 ± 1.7	6.8 ± 3.0	0.144	5.1 ± 1.9	7.5 ± 5.3	0.059
Creatinine, µmol/L	57.1 ± 21.7	54.7 ± 25.1	0.263	65.5 ± 22.5	62.6 ± 27.7	0.367	46.4 ± 15.6	44.7 ± 17.9	0.549

Linear mixed-effect models

The data distribution and changes in mean SMI and ASMM using linear mixed-effect models are displayed in Fig. 2. For SMI, A significant increase in SMI from baseline was seen in male group at 3 months (6.0–6.3 kg/m2) but falling to 6.1 kg/m2 at 6 months (P < 0.05). Similarly, ASMM significantly increased from 17.7 kg to 19.0 kg at 3 months in males, before decreasing to 17.5 kg at 6 months (P < 0.05). While for female, no significant changes were observed in either SMI or ASMM throughout the study period (P > 0.05).

Fig. 2.

Data distribution for SMI and ASMM throughout the 6-month study

Relationship between ONS intake and gut microbial diversity

To evaluate the dynamic changes in the gut microbiota following ONS supplementation, we conducted 16 S rRNA gene sequencing on fecal samples collected from participants before and after the intervention. As shown in Fig. 3, microbial richness decreased significantly after ONS intake (T3 vs. T0). Similarly, microbial evenness and diversity also exhibited a significant reduction following ONS consumption but increased notably after the intervention was discontinued (T6 vs. T3). Beta-diversity, assessed by PERMANOVA and PCoA, showed significant differences only between baseline and the 3-month time point (P < 0.05, Fig. 3).

Fig. 3.

Diversity measures of microbial communities. A, richness index; B, evenness index; C, diversity index; D, community structure shown by Principal Coordinate Analysis (PCoA)

Microbiota composition and abundance of participants with low skeletal muscle mass at baseline

The distributions of bacterial taxa at phylum, family, and genus levels were calculated from the samples at baseline. The results revealed that the gut microbiota was dominated by four phyla (average > 10%, Fig. 4A), including Bacillota (average 53.89%), Bacteroidota (17.07%), Pseudomonadota (14.02%) and Actinomycetota (11.84%). At the family level (Fig. 4B), Lachnospiraceae (average 18.59%), Enterobacteriaceae (13.03%), Oscillospiraceae (11.09%), Bifidobacteriaceae (10.36%), Bacteroidaceae (9.31%) and Lactobacillaceae (5.10%) presented relatively high abundances (average > 5%), and Lachnospiraceae. Oscillospiraceae and Bacteroidaceae were conserved in all the samples. At the genus level (Fig. 4B), 13 genera exhibited mean relative abundances exceeding 2%, with Bacteroides (5.77%) being the only genus universally present in all individuals. When comparing the gut microbiota composition between male and female subjects, no significant differences were observed at different taxa level (P > 0.05).”

Fig. 4.

Gut microbiota composition of participants with low skeletal muscle mass at baseline. A, phylum level abundance; B, family and genus level abundance

Relationship between gut microbiota and skeletal muscle mass

The association between gut microbiota and skeletal muscle mass was assessed by HALLA and shown in Fig. 5. A relatively high abundance of genera Blautia wexlerae, Corynebacterium dentalis and Blautia obeum had positive associations with SMI. Blautia wexlerae and Streptococcus ilei were positively correlated with SMI/SMM, while Clostridioides difficile and Eubacterium callanderi were negatively associated with SMM. After three months of ONS consumption, sex-specific differences in bacterial genera emerged. Those taxa showing positive associations with SMI and SMM were significantly more prevalent in male participants than in females (Figure 5G).

Fig. 5.

Abundance of species associated with SMI and SMM. A-D, positive associations; E-F, negative associations; G, abundance at T3 time point

Predicted microbiota functional profiling and associations with species

With respect to functional profiling, we predicted a total of 405 pathways associated with the gut microbiota, and 147 pathways had significant correlations with six species that corelated with skeletal muscle mass (Supplementary Table 4, P < 0.05). Compared to baseline, the pathways were enriched after ONS intake (P < 0.25), and the SMI-related species were associated with 19 pathways involved in amino acid biosynthesis and degradation (Supplementary Table 4). As is shown in Fig. 6, three of SMI-related species that had positive associations with SMI/SMM were positively correlated with six pathways, including S-adenosyl-L-methionine cycle I (PWY-6151), L-glutamate and L-glutamine biosynthesis (PWY-5505), L-lysine biosynthesis I (DAPLYSINESYN-PWY), glycogen biosynthesis I (GLY2COGENSYNTH-PWY), peptidoglycan maturation (PWY0-1586), and methanogenesis from acetate (METH-ACETATE-PWY). Clostridioides difficile correlated positively with three pathways involving in degradation.

Fig. 6.

The functional profiling correlations with the species that associated with low skeletal muscle mass. The listed pathways were correlations with at least three species

Discussion

In the study, we found that supplementation with a blend of casein and hydrolyzed whey protein significantly improved the skeletal muscle index and reduced serum levels of inflammatory markers such as IL-6 and CRP in hospitalized older males. Furthermore, we observed that the gut microbiota composition was associated with high protein intake and skeletal muscle mass. Protein supplementation markedly influenced bacterial diversity and microbial structure, with these effects exhibiting sex-specific differences. The gut microbes that had positive associations with SMI and SMM were more abundant in males than in females after high-protein consumption.

Muscle mass is maintained by a tightly controlled balance between muscle protein synthesis and breakdown [19]. Despite the impart of dietary protein on sarcopenia remains controversial, with some studies reporting no significant benefits on protein synthesis or muscle strength, the majority of evidence supports the advantages of high-quality protein supplementation for muscle mass preservation [9]. Increasing daily protein intake to 1.2–1.5 g/kg/day, preferably of whey protein, containing large amounts of essential amino acids (EAAs) such as leucine, is recommended for sarcopenic population according to nutritional guidelines [20]. A meta-analysis encompassing over 20 randomized controlled trials concluded that muscle-targeted oral nutritional supplements, primarily based on whey protein, improved muscle-related outcomes regardless of concomitant exercise training [21]. As for casein protein, it is classified as a slowly digestible protein because of a slower gastric emptying [22, 23]. However, casein protein may confer benefits through the inhibition of protein breakdown [11]. According to Zanini et al.’s, cow-milk proteins (comprising whey and casein in a ratio of approximately 20:80) seemed to exert favorite effects on maintenance of skeletal muscle mass in ageing [24], which was consistent with the findings of our study. Another analysis integrating data from 18 randomized controlled trials involving over 600 participants suggested that larger protein doses and milk protein intake were associated with greater postprandial availability of plasma phenylalanine [23]. Tyler, et al. also found that muscle protein synthetic response is equal after milk compared to whey protein ingestion when co-ingestion with carbohydrate [25]. Therefore, blended protein formulations may represent a preferable option for promoting muscle maintenance in vulnerable populations.

Consistent with previous observation, we also found that higher intakes of protein could maintain muscle mass in older men [26], while the significant association was not shown in females [27]. This discrepancy may be attributed to postmenopausal hormonal changes or generally higher adiposity in women [28]. Hormonal decline associated with aging is also likely to impact the loss of muscle mass, with reduced amounts of testosterone and estrogen in men and women, respectively [29]. The rapid decline in estrogen levels among postmenopausal women results in a less anabolic environment, thereby reducing muscle protein synthesis rates. Testosterone exerts a direct and dose-dependent effect on stimulating muscle protein synthesis and hypertrophy in men [30]. Although testosterone levels typically decline with age, residual concentrations in older men may still provide a more favorable endocrine background for responding to protein supplementation. These observations collectively suggest that nutritional interventions for sarcopenia may need to be sex-specific, and future studies should consider incorporating baseline hormonal profiles as potential predictors of treatment response.

Interestingly, we also observed that improvements of muscle mass in males were along with reductions in systemic inflammatory markers, including IL-6 and CRP. Chronic inflammation is considered a major risk factor in age-related diseases. It was widely demonstrated that the pathophysiology of sarcopenia is relevant with chronic low-grade inflammation due to ageing [31]. Circulating proinflammatory cytokines are known to modulate key signaling pathways that regulate both protein synthesis and degradation [32]. A systematic review has summarized studies between inflammatory cytokines and sarcopenia, and found that individuals with sarcopenia frequently exhibit elevated levels of IL-6 and CRP [32], and reductions in these markers have been associated with improvements in metabolic health [33] and muscle mass [32]. Our findings revealed that higher protein intake may help decrease muscle depletion, thereby mitigate chronic low-grade inflammation. In addition, vitamin D was also seen elevated following intervention in our study. Previous studies have established an independent association between vitamin D status and both muscle mass and strength [34], particularly in men [35]. Thus, the observed muscle gains may also be linked to improved vitamin D levels, highlighting the potential multifactorial benefit of nutritional supplementation.

Regarding the gut microbiota, this study revealed a negative correlation between high-protein ONS supplementation and bacterial diversity. This reduction may be attributed to the lower proportion of dietary carbohydrates in high-protein diet, as carbohydrates serve as the fundamental substrate for microbial fermentation and energy production [36]. Correlation analysis showed the species C.dentalis, B.obeum, B.wexlerae, S. ilei had significantly positive relations with SMI/SMM, and C. difficile and E.callanderi correlated significantly negatively with SMM, which were similar with previous studies. Blautia wexlerae has been shown to exert beneficial effects against obesity and type 2 diabetes in both human and animal studies [37]. The Potential mechanism included the production of amino acid metabolites (e.g., S-adenosylmethionine, acetylcholine, L-ornithine) with anti-adipogenic and anti-inflammatory properties, as well as short-chain fatty acids (SCFAs) such as succinate, lactate, and acetate, which collectively contribute to a healthier gut environment [37]. According to previous studies, SCFAs are known to positively regulate skeletal muscle metabolism by enhancing muscle glycogen storage, improving metabolic efficiency of muscle fibers, and activating the mTOR/IGF-1 pathway to promote protein synthesis [38].C.dentalis was human homolog of Dubosiella newyorkensis. Zhang et al. have found that both of them were robust SCFA-producing commensal bacterium and had probiotic immunomodulatory effects to maintain intestinal immune homeostasis [39]. B.obeum was positively associated with multiple obesity phenotypic indicators [36], while it can produce lantibiotic peptides against C.difficile [40], and was one component of fecal microbiota transplants to treat recurrent C. difficile infection [41]. As the major cause of healthcare-associated infections, C. difficile was associated with fragility and functional loss in older people [42]. Interestingly, E.callanderi has previously been associated with anti-inflammatory activity [43], yet our data indicate a negative correlation with SMM, warranting further investigation into its role in muscle metabolism.

We further analyzed predicted metabolic pathways in the gut microbiota before and after ONS intervention. Three SMI-positively correlated species were also associated with six pathways involved in amino acid biosynthesis, including those producing glutamate, lysine, and S-adenosylmethionine (SAM). The skeletal muscle is the main tissue for glutamine storage, and glutamine is made via the action of glutamine synthetase from glutamate and ammonia, primarily in skeletal muscle ([44]. Lysine can mediate the satellite cells activation to govern skeletal muscle growth [45]. SAM is the biosynthetic precursor for homocysteine and cysteine ([46], and acts as a methyl donor for biologic methylation to prevent diabetes, obesity, and inflammation ([37]. Among different metabolic pathways, the PWY-5705 and P341-PWY were significantly enriched after ONS intervention. The P341-PWY pathway can produce pyruvate and ATPs, which could supply energy for muscle contraction, and provide energy for the active-transport Ca⁺⁺ pumps in the sarcoplasmic reticulum ([47]. These findings suggested that a high-protein diet may facilitate muscle growth partly through modulation of the gut microbiota and its metabolic activities, particularly those related to energy supply, anti-inflammation, and biosynthetic pathways critical for muscle protein regulation.

Of note, while the study yielded several important findings, it is important to acknowledge its potential limitations. Firstly, this was a self-controlled study with a limited sample size. Although our study was sufficiently powered to detect the overall treatment effect, the sample size limited robust sex-stratified analyses. Nevertheless, the inclusion of both male and female participants strengthens the scientific validity and clinical applicability, providing a more holistic perspective on sarcopenia management [48]. Future randomized clinical trials with larger sample sizes are warranted to confirm and explore these sex-based differences more definitively. Second, while sarcopenia is diagnostically defined by the presence of both low muscle mass and low muscle strength and/or impaired physical performance, this study focused specifically on muscle mass. This focus was necessitated by practical constraints in assessing grip strength or gait speed in hospitalized older adults with chronic bed rest. It is important to note, however, that low muscle mass alone is a well-established indicator of sarcopenia and is independently associated with serious clinical outcomes including fractures, frailty, physical disability, and increased mortality [49]. Moreover, muscle mass is widely accepted as a primary endpoint in trials targeting sarcopenia and muscle-related disorders [50]. Thus, we believe the enrolled subjects with low muscle mass could also be representative and valuable for research. Finally, Actual individual dietary intake was not quantitatively measured. To mitigate this limitation, nursing staff estimated intake via photographic documentation of meals, and the standardized meal preparation provided by the hospital ensured a high degree of dietary consistency across participants.

Conclusions

In conclusion, our study establishes a compelling link between high-protein nutritional intervention, gut microbiota remodeling, and the preservation of skeletal muscle mass in hospitalized older adults. We demonstrated that supplementation with a blend of casein and hydrolyzed whey protein significantly improved muscle mass in older males and induced distinct shifts in microbial diversity and community structure. Mechanistically, we identified bacterial amino acid metabolic pathways—particularly those involved in the production of glutamate, lysine, and S-adenosylmethionine—as potential key mediators supporting muscle anabolism. These findings reinforce the concept of a microbiota–muscle axis and underscore the clinical value of targeted protein supplementation in attenuating muscle loss associated with chronic immobility in older patients.

Authors’ contributions

Y.W designed and developed the research, revised the manuscript and is the corresponding author. HJ.Z critically revised the manuscript and is the other corresponding author. Y.C contributed to the methodology, statistical analysis and writing of the manuscript. YZ.W performed microbiota analysis and data visualization, and revised the manuscript. W.R and Z.Y conducted the research and collected data. XN.L and YQ.Z provided data, developed the methodology and revised the manuscript.

Funding

This work was supported by the Discipline Promotion Programme of Shanghai Fourth People’s Hospital (SY-XKZT-2020-1010 and SY-XKZT-2024-2003) and Startup Foundation for Scientific Research of Shanghai Fourth People’s Hospital(sykyqd01601).

Data availability

The sequence data have been deposited in the GenBank Sequence Read Archive with the accession code PRJNA1255383.

Declarations

Ethics approval and consent to participate

This study was conducted in accordance with the principles of the Helsinki Declaration and approved by ethics committee of Shanghai Fourth People’s Hospital (2021114-001). Informed consent was obtained from all individual participants included in the study.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.