A Pilot Longitudinal 16S rRNA Gene Sequencing Study Exploring the Relationship Between Gut Microbiota and Body Composition in Healthy Adults

Introduction

The gut microbiome was proven to be important to human physiological functioning, including protection and immune function (Mills et al. 2019), synthesis and metabolism, and nutrition absorption (Gomaa 2020), and is even related to emotions and cognitive function (Liang et al. 2022). However, the diversity of the gut microbiome fosters the complexity of those mechanisms, and there are differences in the gut microbiome between individuals (Robles-Alonso and Guarner 2013; Ursell et al. 2012).

These inter-individual variations may be influenced by a variety of host-related factors, including body composition. The body’s composition is a vital marker for evaluating a person’s overall health. It is well-established that excessive body fat and obesity are correlated with numerous chronic diseases, including type 2 diabetes, cardiovascular diseases, stroke, malignancies, and arthritis (WHO 2000). These chronic conditions are increasingly supplanting infectious diseases as the primary threats to human health in today’s world (Maffetone et al. 2017). Moreover, muscle strength is profoundly influenced by skeletal muscle mass, which, in turn, affects motor performance and balance (Abramowitz et al. 2018). As individuals age, muscle mass can diminish by 30% to 40% compared to their younger years (Lexell et al. 1983). This reduction is particularly pronounced in sarcopenic patients, where the ratio of the limb skeletal muscle mass to body weight is frequently employed to predict the extent of functional impairment and mortality following an illness (Bijlsma et al. 2013). In recent years, gut microbiota has been extensively investigated as a critical factor closely linked to body composition.

The gut microbiome impacts skeletal muscle mass and body fat. Short-chain fatty acids (SCFAs) relate to fat synthesis and consumption (Morrison and Preston 2016), with A. muciniphila and Veillonella associated with SCFA synthesis from carbohydrates (Li et al. 2023; Scheiman et al. 2019). The Firmicutes phylum correlates with increased weight (Afzaal et al. 2022; Duranti et al. 2017), while overweight children often exhibit higher Staphylococcus aureus and lower Bifidobacterium (Kal-liomäki et al. 2008). Specifically, butyrate increases insulin sensitivity and triggers the mammalian target of rapamycin (mTOR) and AMP-activated protein kinase (AMPK) signaling, resulting in protein synthesis (Hawley 2020). Dietary unsaturated fats can increase butyrate-producing Faecalibacterium prausnitzii and Roseburia (Muralidharan et al. 2019). Furthermore, microbiome-mediated inflammation inhibition enhances muscle mass, while Proteobacteria can increase gut permeability and insulin resistance (Di Vincenzo et al. 2024), butyrate, A. muciniphila, and Bifidobacterium breve improve amino acid absorption and reduce muscle injury, promoting muscle mass (Di Vincenzo et al. 2024; Li et al. 2023).

Previous research identified certain phyla of the gut microbiome that influence body composition. However, most studies focused on specific phyletic interventions through biotics. Some cross-sectional studies found that Bacteroides, Blautia, A. muciniphi-la, Prevotella, Lactobacillus, Faecalibacterium, and Bifidobacterium were associated with better body composition, characterized by greater muscle mass and less body fat. In contrast, Firmicutes and Proteobacteria showed adverse effects (Afzaal et al. 2022; Chew et al. 2023; Di Vincenzo et al. 2024; Lustgarten 2019; Sugimura et al. 2022). Most of the existing research was either cross-sectional or involved biotic interventions, lacking long-term observations of individual subjects. Understanding the relationship between the fecal microbiome and body composition requires accounting for individual variabilities. In this study, we focused on comparing changes in body composition within the same person over time to explore the gut microbiome’s impacts.

Experimental

Results

The research ran from April 2022 to September 2023, and 15 participants were recruited to analyze their fecal microbiome and body composition. The age of participants ranged from 20 to 31 years. Ten males and five females participated in the study, and all exercised at least 150 min/week. Demographic information is included in Table I.

The body composition of subjects was measured a total of 4 times, with measurements taken at 3-month intervals. Heights ranged 158–185 cm, body weights ranged 50.9–106.9 kg, BMI ranged 17.68–33.74 kg/m², skeletal muscle percentages ranged 31.1%–50.1%, and body fat percentages ranged 11.6%–45.1%. According to WHO criteria, a BMI ≥ 25 kg/m² was classified as overweight. Sixteen measurements were defined as being overweight in the study.

Taxonomic comparisons were based on OTUs clustered at 99% similarity and annotated to the closest matching reference species in the SILVA v138.1 database. For readability, taxa are referred to by their representative species names throughout the Results section.

Stratifying participants by a BMI threshold of 25 kg/m² revealed significantly lower microbiota alpha diversity in the overweight group compared to the normal weight group (p < 0.01) (Fig. 1). No significant differences in alpha diversity were found between high and low body fat or skeletal muscle groups (p = 0.3–0.7). The beta diversity analysis using principal coordinates analysis (PCoA) based on Bray-Curtis distances showed significant differences in microbiota composition across all body composition groupings, including overweight vs. normal weight, high vs. low skeletal muscle percentage, and high vs. low body fat (PERMANOVA, p < 0.01) (Fig. 2).

Fig. 1.

Fig. 2.

A total of 16 measurement points (not necessarily from distinct participants), each corresponding to a BMI value ≥ 25 kg/m² across the four sampling time points, were classified as overweight.

LMM analyses incorporating time, group, and their interaction were first performed to account for repeated measures. Analyses of 139 core taxa (prevalence > 20%) revealed overall stability, with most taxa showing no significant time × group interactions (Supplementary Tables SI, SII), likely reflecting the small sample size and inter-individual variability. While Bifidobacterium adolescentis reached significance in the LMM (p = 0.04), it was excluded from primary discussion due to non-significant baseline-to-9-month changes in stratified analyses. Therefore, exploratory analyses using a stratified baseline-to-endpoint approach were conducted to visualize potential trends in microbiota related to body composition changes.

At the final measurement, nine participants showed an increase in the body fat percentage and BMI, while the remaining six exhibited a decrease. Similarly, the skeletal muscle percentage decreased in 10 participants, while it increased in the other five. After comparing different body compositions and their changes over time, several putative fecal microbial species were observed that showed consistent trends in relation to both the body composition and its changes.

In participants with an increased body fat percentage, the OTU annotated as Parasutterella excrementihominis showed a within-subject increase in abundance between baseline and 9 months, as supported by the Wilcoxon signed-rank test (Z = –0.94, p = 0.35, r = 0.39), indicating a medium effect. Conversely, in participants with lower body fat percentage, the abundance of this species declined over the same interval (Z = –0.36, p = 0.72, r = 0.18), indicating a small effect. While Spearman’s rank correlation analysis between changes in microbial abundance and body fat percentage showed only a weak association (p = 0.68, ρ = 0.12), the directional consistency observed here remains a preliminary observation and should be interpreted with caution, as it lacks robust statistical significance in this small cohort (Fig. 3a).

Fig. 3

The sequences assigned to A. muciniphila showed higher relative abundance in participants with lower body fat percentage. Additionally, in participants with increased body fat, the abundance of this species decreased at 9 months compared with baseline, as supported by the Wilcoxon signed-rank test (Z = –1.40, p = 0.16, r = 0.50), indicating a medium-to-large effect size. However, Spearman’s rank analysis showed only a weak and non-significant inverse relationship between the changes in A. muciniphila and body fat (p = 0.66, ρ = –0.12). This discrepancy suggests that while a potential trend may exist, it should be regarded strictly as a preliminary observation and interpreted with caution, particularly given the small sample size (Fig. 3b).

The OTU representative of Dialister invisus showed a higher relative abundance in participants with greater skeletal muscle mass, and the ANOVA-like model indicated a significant overall trend across the baseline and 9-month assessments. However, the paired comparison within individuals did not reach statistical significance (Z = 0.00, p = 1.00, r = 0.00). Consistent with this, Spearman’s rank analysis showed virtually no linear relationship between the changes in D. invisus and skeletal muscle percentage (p = 0.96, ρ = –0.02). The lack of consistency between group-level trends and individual longitudinal shifts suggests that the observed differences may primarily reflect baseline variations or overall temporal patterns rather than meaningful with-in-individual changes (Fig. 3c).

Conversely, participants with a lower muscle proportion had a higher relative abundance of the sequences assigned to Bifidobacterium pseudocatenu-latum. In addition, this putative species showed a 9-month increase compared with baseline among participants whose muscle proportion had decreased, as supported by the Wilcoxon signed-rank test (Z = 0.94, p = 0.35, r = 0.42), indicating a medium effect size. However, this trend was less evident in the continuous analysis; Spearman’s rank correlation showed only a weak and non-significant inverse relationship between the changes in B. pseudocatenulatum and muscle percentage (p = 0.72, ρ = -0.10). This weak correspondence suggests that while a potential trend may exist, it should be regarded as a preliminary observation and interpreted with caution (Fig. 3d).

To clarify potential sex-specific differences in fecal microbiota changes associated with body composition, we performed sex-stratified analyses. We found that a significant change in the OTU assigned as B. pseudoca-tenulatum (p = 0.01) was predominantly observed in females with decreased skeletal muscle mass. In contrast, significant alterations in putative D. invisus (p = 0.01) and P. excrementihominis (p < 0.01), associated with changes in body fat, were primarily observed in males.

Prior research suggested that the Firmic-utes-to-Bacteroidetes (F/B) ratio may be associated with obesity. In this study, the F/B ratio was calculated for each sample as the relative abundance of Firmicutes divided by that of Bacteroidetes, and the mean ratio was reported for each group. The mean F/B ratio was 0.39 in the overweight group and 0.54 in the normal-weight group (Fig. 4). The difference between the two groups reached marginal statistical significance (p = 0.052).

Fig. 4.

Discussion

Previous studies on the relationship between body composition and fecal microbiota have largely examined probiotic use, dietary interventions, or prescribed health programs. In contrast, this study investigated microbiota-body composition associations under natural lifestyle conditions and further tracked individuals over time. By classifying microbiota shifts according to longitudinal changes in body composition, we aimed to identify representative taxa potentially linked to body composition dynamics.

While most previous studies suggested that individuals with obesity exhibit lower gut microbiota diversity compared to the general population, a systematic review indicated that this trend was not statistically significant, which warrants further investigation (Pinart et al. 2022). Part of the results of this study aligned with our hypothesis, showing that the overweight group exhibited lower microbiota diversity compared to individuals with a normal weight, supporting the findings of most prior research. However, no clear trends were observed between groups with high and low body fat percentages or skeletal muscle percentages. The beta diversity analysis similarly demonstrated significant inter-sample biodiversity differences in the fecal microbiota among individuals with distinct body composition characteristics. This suggests that future studies should include a larger sample size and more-appropriate cutoff points for group classifications to further validate this hypothesis.

An optimal body composition is characterized by lower body fat, higher muscle mass, and a moderately maintained BMI. In this study, we observed that individuals with a greater muscle percentage tended to have higher abundances of the OTU representative of D. invisus, whereas lower body fat was associated with putative A. muciniphila, a bacterium known for its role in maintaining gut barrier integrity and metabolic health.

In our study, the taxon annotated as D. invisus showed a significantly higher abundance in participants with higher skeletal muscle percentage, whereas its within-subject temporal variation remained unclear. Consistent with this, Spearman’s rank analysis revealed virtually no linear relationship between changes in D. invisus and muscle percentage, suggesting that the observed link might be driven more by baseline differences between individuals rather than longitudinal shifts. This may be related to its role as a mediator of inflammatory responses in individuals with specific dietary habits (Shi et al. 2022). Additionally, D. invis-us was observed to decrease in patients with various autoimmune and inflammatory diseases, suggesting a potential role in reducing inflammatory responses (Joossens et al. 2011; Lee et al. 2019). A reduction in inflammation could support muscle synthesis, thereby influencing the body composition of participants.

Previous research showed a positive correlation between Parasutterella sp. and BMI, with a significant reduction in P. excrementihominis observed following a 12-week low-calorie, low-carbohydrate weight-loss intervention (Henneke et al. 2022). Our study further explored the relationship between the OTU assigned to P. excrementihominis and body composition, focusing on body fat percentage, and similarly supported the positive association between these putative species and being overweight or obese. Although the linear correlation was weak, the consistent directional trend across different statistical approaches strengthens the possibility of this link. Since SCFAs may have an inhibitory effect on fatty acid synthesis (He et al. 2020), and Parasutterella sp. and P. excrementihominis produce fewer SCFAs (Nagai et al. 2009), a higher relative abundance of this bacterium may promote fatty acid synthesis, potentially leading to an increased body fat percentage or body weight.

B. pseudocatenulatum has traditionally been regarded as an important probiotic in the gut, with benefits including restoration of the gut barrier function, reduction of inflammation, improvement of lipid profiles, and enhancement of insulin receptor sensitivity (Cano et al. 2013; Mauricio et al. 2017; Moya-Pérez et al. 2015; Sanchis-Chordà et al. 2019). However, in this study, an opposite trend was observed: the proportion of sequences representative of B. pseudocatenulatum increased as the muscle mass percentage decreased. Notably, while this association remained consistent in the continuous analysis, it was not supported by a strong linear correlation, suggesting that the observed trend may be highly variable among individuals in this small cohort. Most previous studies explored the relationship between representative taxon and related markers through animal models or interventional supplementation of B. pseudocatenulatum, whereas our study examined changes in the body composition and fecal microbiota under a natural diet and living conditions. The lack of a robust linear link indicates that this unexpected finding remains a preliminary observation. Whether this difference in study context accounts for the contrasting trends remains to be clarified through further research.

Another extensively studied probiotic, A. muciniph-ila, was similarly observed to have beneficial associations in this study. Previous animal studies showed that 4 weeks of supplementation with A. muciniphila improved muscle strength in mice with muscle atrophy, potentially through its influence on the ubiquitin-pro-teasome pathway (Byeon et al. 2022). Moreover, A. muciniphila has been reported to exhibit anti-inflammatory properties, enhance gut barrier function, and promote SCFA production, all of which contribute to improved health and body composition (Abuqwider et al. 2021; Depommier et al. 2019; Naito et al. 2018). This study also observed an increase in A. muciniphila in individuals with lower body fat or during periods of fat loss. While the continuous analysis using Spearman’s rank correlation showed only a weak inverse relationship between changes in A. muciniphila and body fat percentage, the direction of the trend remained consistent with our categorical findings. This discrepancy suggests that while a potential trend may exist, it should be regarded strictly as a preliminary observation and interpreted with caution. The significance of this bacterium in relation to body fat and muscle mass warrants further research to explore its quantitative and qualitative impacts, providing evidence for its potential use in exercise health or therapeutic interventions.

The sex-specific associations between fecal microbiota and body composition may reflect underlying hormonal, metabolic, and immune differences between males and females (Mayneris-Perxachs et al. 2020). Estrogen, for instance, affects gut permeability and systemic metabolism, potentially influencing microbiota-host interactions (Santos-Marcos et al. 2023). In our study, changes in B. pseudocatenulatum were observed in females with reduced muscle mass, whereas taxa such as Dialister and Parasutterella were linked to changes in fat in males. Although the causal role of these microbes remains unclear, previous studies suggested that bile acid metabolism and sex hormones may mediate such differences, and that microbial associations with body composition can be sex-specific (Park et al. 2022). These findings require further validation and mechanistic investigation.

Previous systematic reviews pointed out that the Firmicutes-to-Bacteroidetes ratio is typically higher in obese individuals (Pinart et al. 2022). However, this study observed an opposite trend. This discrepancy may be attributed to the small sample size of this study and to the fact that participants were selected based on engaging in at least 150 minutes of exercise per week, a characteristic that differs from those of the sample populations used in previous studies. Future research on the relationship between the fecal microbiota and body composition should take participant characteristics into account.

One limitation of this study is its relatively small sample size, which reflects the inherent inter-individual variability in the fecal microbiota and the challenges of longitudinal sampling. To better explore microbiota distribution patterns, we not only compared differences between individuals but also performed within-subject analyses before and after changes in body composition. Nevertheless, the post-hoc power analysis based on the Wilcoxon signed-rank test indicated limited statistical power (0.05–0.43 at α = 0.05), suggesting that the findings should be interpreted as exploratory. Despite this limitation, the longitudinal repeated-measures design may have helped reduce inter-individual variability and provided preliminary effect-size estimates to guide adequately powered future studies. Additionally, this study did not define clear cutoff points for certain body composition parameters, such as muscle mass and body fat percentage. Dietary intake and exercise levels were not recorded, as the study aimed to observe fecal microbiota and body composition under participants’ natural, real-world lifestyle conditions. This design allows for the capture of spontaneous variation and enhances ecological validity, though it introduces potential confounding effects from diet and physical activity. The gut microbiome is highly responsive to dietary fiber and caloric intake, which are known to modulate the abundance of key taxa identified in this study, such as Dialister spp. and P. excrementihominis (Ghosh et al. 2020). Similarly, the abundance of A. mu-ciniphila is closely linked to both dietary quality and host adiposity (Asnicar et al. 2021). Without detailed dietary logs, we cannot definitively determine whether the observed microbial shifts were solely related to changes in body composition or were partly driven by unrecorded dietary fluctuations. Previous studies have demonstrated that physical activities, particularly moderate-intensity continuous training, can lead to an increase in the abundance of A. muciniphila, Bifidobacterium spp., and Dialister spp. in various populations (Clarke et al. 2014; Kim et al. 2025; Torquati et al. 2023), highlighting the critical role of exercise in gut microbiome modulation. We acknowledge that this lack of control for major determinants of the microbiota limits the internal validity of our findings. Therefore, our results should be interpreted as associations rather than direct causal links. Future research could establish individualized cutoff points and incorporate standardized dietary assessments alongside objective physical activity monitoring (e.g., using accelerometers or wearable devices) to enable more precise analyses and isolate the effects of lifestyle variance from physiological adaptations.

In conclusion, this pilot study provided preliminary observations indicating that temporal shifts in the fecal microbiota may be associated with changes in body composition in a small, single-region cohort. Trends such as higher relative abundance of putative D. invisus in participants with greater skeletal muscle mass must be interpreted with extreme caution. These patterns remained strictly exploratory, particularly as they lacked robust statistical significance in continuous correlation models and required confirmation in larger and more diverse populations. While the directional consistency across different statistical approaches hints at potential longitudinal links, these observations remain preliminary and hypothesis-generating. Future multi-center studies with sufficient statistical power are essential to validate and extend these findings.