Branched-chain amino acids (BCAAs), including isoleucine, leucine, and valine, are essential nutrients primarily derived from dietary sources due to the lack of de novo synthesis in humans. Beyond their structural role in muscle, circulating BCAAs are increasingly recognized as active metabolic modulators^[1] and important intermediate phenotypes linking food quality analysis and post-consumption metabolism^[2,3]. Notably, emerging evidence links plasma BCAAs to cardiometabolic health (CMH), serving as early biomarkers of increased risks for cardiovascular disease and type 2 diabetes^[4].

Given their implications for CMH, identifying modifiable determinants of circulating BCAAs is of substantial public health importance. As systemic BCAAs reflect the metabolic processing of food rather than just direct dietary ingredients, their circulating concentrations are highly responsive to dietary quality and nutritional modifications. For instance, altering macronutrient composition or specific amino acid ratios in protein-restricted models can partially restore BCAA homeostasis^[5], and structured dietary interventions have been shown to improve circulating BCAA profiles^[6]. In addition to dietary intake, plasma BCAA levels are regulated by a constellation of host factors. Longitudinal studies reveal that BCAA levels tend to increase from young adulthood to midlife, although trajectories vary; males and individuals with obesity or dyslipidemia are more likely to exhibit sustained increases or persistently elevated levels^[7]. Lifestyle behaviors also play an important role. Persistent smoking and alcohol consumption may perturb BCAA levels by disrupting mitochondrial energy metabolism^[8,9], whereas regular physical activity promotes BCAA catabolism^[10]. Despite these observations, population-level evidence evaluating these determinants remains limited. Most previous studies have considered a narrow range of host factors^[11], and dietary analyses have largely emphasized dietary patterns or protein intake, failing to characterize contributions of specific food groups to BCAA homeostasis^[6,12].

Although BCAAs have been implicated in the development of cardiometabolic diseases, systematic evaluations encompassing a broad panel of CMH biomarkers, spanning glycemic control, lipid metabolism, and systemic inflammation, remain scarce. In addition, while current research suggests that plasma BCAAs influence CMH through specific cellular mechanisms, including platelet activation or mitochondrial dysfunction^[13,14], the broader systemic molecular pathways in humans remain poorly characterized. There remains limited evidence from large-scale population cohorts utilizing high-throughput proteomics to systematically evaluate associations between plasma BCAAs and circulating proteomic biomarkers. This gap hinders a comprehensive understanding of the biological pathways and potential mechanisms underlying the associations between BCAAs and CMH.

To address these knowledge gaps, we utilized the large-scale UK Biobank (UKB) cohort to systematically examine the population-level associations of dietary and non-dietary factors with plasma BCAA levels, and evaluated the relationships between plasma BCAA levels and a broad panel of CMH biomarkers. High-throughput proteomics was further leveraged to explore key biological pathways and identify central hub protein markers within these networks.

A total of 28,273 participants were included in the primary analyses of determinants of plasma BCAAs. Baseline characteristics were stratified by quartiles of total BCAA concentrations [Table 1]. In multivariable-adjusted models, plasma BCAAs exhibited specific association patterns with dietary macronutrients and food groups [Figure 1A and Supplementary Tables 4-7]. Overall, plasma BCAA levels were inversely associated with carbohydrate intake and positively associated with animal protein and MUFA consumption (all P < 0.05). At the food group level, positive associations were observed between plasma BCAAs and red meat, poultry, eggs, dairy products, and non-oily fish. In contrast, plant-based food showed mostly weak or non-significant associations with plasma BCAAs. Notably, individual BCAA species displayed unique dietary signatures. Isoleucine demonstrated a distinctive positive relationship with PUFA intake, whereas leucine concentrations were negatively associated with the consumption of plant protein, dietary fiber, oily fish, and fruits. Regarding non-dietary factors [Figure 1B and Supplementary Tables 8-11], older age was associated with lower plasma BCAA levels. Marked sex differences were observed, with males exhibiting higher levels than females. Among females, postmenopausal status and hormone therapy use were both positively associated with plasma BCAAs. Adiposity showed robust associations with BCAAs: underweight individuals had lower levels, whereas participants with obesity, defined by either BMI or waist circumference (central obesity), had substantially higher concentrations. Compared with other ethnic groups, White participants had lower levels of isoleucine and leucine. For lifestyle behaviors, lower levels of physical activity were associated with higher plasma BCAA concentrations. Current smoking was associated with lower plasma BCAAs. Although associations across alcohol consumption categories were generally not significant, BCAA levels tended to decrease with increasing alcohol intake, and high-risk drinking was significantly associated with lower isoleucine levels. These associations remained highly consistent in sensitivity analyses. Excluding participants with atypical dietary reports or further adjusting for eGFR did not materially alter the effect estimates [Supplementary Tables 12-21].

Cross-sectional analyses revealed that elevated plasma BCAAs were consistently associated with unfavorable glycemic and lipid profiles [Figure 2 and Supplementary Table 22]. The strongest positive associations were observed for TG (total BCAA: β = 0.288, 95%CI: 0.276-0.300; isoleucine: β = 0.249, 95%CI: 0.237-0.261; leucine: β = 0.234, 95%CI: 0.222-0.247; valine: β = 0.307, 95%CI: 0.295-0.319; all P < 0.001). Weaker, yet statistically significant, positive associations were identified for TC, ApoB, and LDL cholesterol, whereas associations with ApoA1 and HDL-C were largely null. Notably, plasma BCAAs were inversely associated with eGFR, suggesting an association between higher BCAA levels and compromised renal function. In the linear regression models, plasma BCAA levels were inversely associated with both SBP and DBP. However, restricted cubic spline (RCS) models revealed a non-linear inverted U-shaped association between plasma BCAAs and blood pressure metrics [Supplementary Figures 2 and 3]. While the three individual BCAAs exhibited broadly concordant associations across CMH domains, species-specific nuances were evident: isoleucine and leucine showed more pronounced associations with blood pressure and renal function, whereas valine demonstrated stronger associations for lipid metabolism markers. Sensitivity analyses confirmed the robustness of these findings, as excluding participants with atypical dietary reports and hypertension, and further adjusting for eGFR, did not materially attenuate the observed associations [Supplementary Tables 23-25].

Subgroup analyses revealed significant effect modification [Figure 3 and Supplementary Table 26]. Associations between plasma BCAAs and various CMH markers were generally stronger in older adults, individuals with a higher BMI, and participants with hypertension (P for interaction < 0.05). Furthermore, the relationships between plasma BCAAs and biomarkers of glucose metabolism, renal function, and inflammation were more pronounced in participants with diabetes than those without (P for interaction < 0.05). Associations of plasma BCAAs with glucose and renal biomarkers were also stronger among participants with hyperlipidemia or MASLD (P for interaction < 0.05). Sex also emerged as a consistent effect modifier in the associations between plasma BCAAs and CMH markers (P for interaction < 0.05). Specifically, positive associations of plasma BCAAs with glycemic markers (blood glucose and HbA1c) and lipid markers (TG and ApoA1), as well as inverse associations with blood pressure, were more pronounced in males.

We further examined association patterns between plasma BCAAs and CMH-related circulating proteins. Overall, the protein profiles associated with individual BCAA species were highly concordant [Figure 4A and B, Supplementary Table 27]. Higher plasma BCAA levels were associated with increased abundance of proteins involved in energy balance and metabolic signaling, including GIP, PRAP1, AGRP, PYY, FST, GHR, and FGF19. In contrast, plasma BCAAs were inversely associated with proteins implicated in lipid metabolism and insulin sensitivity regulation, including ADIPOQ, LPL, IGFBP1, IGFBP2, SHBG, and CKB. Although the number of significant proteins differed across individual BCAAs, we identified 172 proteins consistently associated with both total BCAA and all three individual species, which were defined as consensus proteins. In addition, 14 proteins were uniquely associated with isoleucine, 23 with leucine, and 207 with valine [Figure 4C].

This shared, species-specific proteomic signature indicates the coexistence of common and distinct biological functions and pathways that are involved in the interplay between BCAA homeostasis and CMH. GO enrichment analysis based on the consensus proteins yielded 301 significant terms, including 246 biological processes (BP), 13 cellular components (CC), and 42 molecular functions (MF) [Figure 5A and Supplementary Table 28]. The most prominently enriched BP terms highlighted processes related to immune and inflammatory responses [e.g., positive regulation of chemokine (C-X-C motif) ligand 2 production; FDR < 0.01, fold enrichment = 28.67] and lipid metabolic homeostasis (e.g., high-density lipoprotein particle remodeling; FDR < 0.001, fold enrichment = 27.30). MF enrichment primarily involved binding functions, notably transforming growth factor-beta (TGF-β) receptor binding (type I and type II), and apolipoprotein binding, as well as structural and enzymatic functions, such as extracellular matrix structural constituents conferring compression resistance, and carbonate dehydratase activity (all FDR < 0.05). CC enrichment localized these proteins to extracellular structures and intracellular compartments, including the plasma lipoprotein particle and the lysosomal lumen (FDR < 0.05). Further analysis of BCAA species-specific proteins revealed distinct functional profiles. Isoleucine-specific proteins were primarily enriched in MF terms, including hormone activity and integrin binding (all FDR < 0.01) [Supplementary Figure 4 and Supplementary Table 29]. Leucine-specific proteins predominantly highlighted chemotaxis and chemokine receptor binding, and were primarily localized to structures like platelet alpha granules and the lysosomal lumen (all FDR < 0.05) [Supplementary Figure 5A and Supplementary Table 29]. Valine-specific proteins were strongly enriched in pathways related to coagulation and wound healing (e.g., platelet activation and fibrinolysis), as well as amino sugar metabolism, characteristically localizing to platelet dense granules and the CD40 receptor complex (all FDR < 0.05) [Supplementary Figure 6A and Supplementary Table 29]. KEGG pathway enrichment analysis of the 172 consensus proteins identified five significantly enriched pathways, comprising 28 proteins: nitrogen metabolism, pantothenate and CoA biosynthesis, fatty acid degradation, cytokine-cytokine receptor interaction, and hormone signaling (all FDR < 0.1) [Figure 5B and Supplementary Table 30]. PPI network analysis of these 28 proteins delineated three major functional modules centered on the positive regulation of the receptor signaling pathway via STAT, nitrogen metabolism, and branched-chain amino acid catabolism. Notably, LEP emerged as the primary hub, occupying central positions within the network [Figure 5C]. KEGG enrichment analysis of isoleucine-specific proteins did not identify any significantly enriched pathways. Conversely, leucine-specific proteins were enriched in 14 pathways related to immune responses, inflammation, and cell migration, such as the IL-17 signaling pathway and integrin signaling (all FDR < 0.05) [Supplementary Figure 5B and Supplementary Table 30]. Valine-specific proteins were significantly enriched in a single pathway, the complement and coagulation cascades (FDR < 0.01) [Supplementary Figure 6B and Supplementary Table 30]. Subsequent PPI network analyses of these pathway-enriched, BCAA species-specific proteins revealed CXCL8 and SDC1 as key hubs for leucine, and F10 and PLAT as hubs for valine [Supplementary Figures 5C and 6C]. Overall, among the 49 proteins mapped to significant KEGG pathways across these analyses, 29 proteins showed a positive correlation with plasma BCAAs, while 20 proteins showed a negative correlation [Supplementary Figure 7 and Supplementary Table 27].

In this large-scale study, we demonstrated that plasma BCAA concentrations were potentially shaped by dietary intake, particularly diets rich in animal protein and low in carbohydrates, alongside key non-dietary factors such as adiposity and male sex. Elevated BCAAs were robustly associated with adverse cardiometabolic profiles. Beyond clinical markers, we identified core proteomic signatures that were either shared across all BCAA species or specific to individual species. Pathway and network analyses indicated that consensus proteins converge on diverse metabolic and endocrine pathways, with LEP acting as a central hub. Collectively, these findings underscore that BCAAs function as systemic biomarkers associated with diet and lifestyle factors, as well as CMH.

Our study identified distinct dietary determinants of plasma BCAAs, with higher levels strongly associated with the consumption of animal-based foods (e.g., red meat and dairy products) and total animal protein. Conversely, we observed inverse associations with plant-based dietary components encompassing fruits, carbohydrates, plant protein, and dietary fiber. These findings indicate that the source of dietary protein plays a fundamental role in shaping BCAA homeostasis, consistent with previous work^[30]. This divergence likely reflects the inherently higher BCAA content of animal proteins compared with plant sources^[31]. However, from a food metabolomics perspective, circulating BCAAs should be interpreted not merely as direct dietary components, but as dynamic indicators of dietary quality and post-consumption metabolism. Owing to their limited hepatic extraction following intestinal absorption^[32], circulating BCAA concentrations are highly responsive to the dietary matrix, which governs nutrient bioavailability^[33]. Accordingly, beyond differences in amino acid composition, plant-based diets may further modulate BCAA homeostasis through multiple secondary metabolites. For instance, higher fiber intake has been shown to downregulate microbial genes responsible for BCAA biosynthesis in animal models, thereby improving host metabolic profiles^[34]. Plant-derived bioactive compounds, including flavonoids and polyphenols, undergo extensive biotransformation by host and microbial enzymes. The resulting metabolites may indirectly participate in the regulation of BCAA metabolism by modulating gut microbiota function, influencing microbiota-derived metabolites, and activating the AMPKα pathway^[35,36]. Accumulating evidence indicates that replacing animal-based diets with plant-based dietary patterns reduces mortality from cardiovascular and cerebrovascular diseases^[37]. Our findings suggest that lower plasma BCAAs are potentially involved in these cardioprotective effects. Accordingly, adherence to plant-based or Mediterranean dietary patterns may represent a practical strategy to improve CMH through modulation of BCAA homeostasis.

We also observed several non-dietary host factors that contributed to plasma BCAAs. Consistent with prior evidence^[38], we observed elevated plasma BCAAs in individuals with insufficient physical activity. Skeletal muscle serves as the primary site for BCAA storage and catabolism, meaning that physical inactivity directly diminishes catabolic efficiency and contributes to elevated circulating levels^[39]. The reduced plasma BCAAs observed in smokers and alcohol consumers likely reflect underlying physiological impairment rather than a protective adaptation. These reductions may result from tobacco-induced disruption of insulin signaling and mitochondrial function^[40] and alcohol-related hepatic injury^[41], both of which also contribute to cardiometabolic disturbances. Lower BCAA levels in older adults and in females, particularly premenopausal females, were observed. These patterns may reflect sex-specific hormonal differences and dynamic changes in hormonal and metabolic levels across physiological phases^[42]. Postmenopausal females experience a decline in estrogen, which is often accompanied by an imbalance in energy and metabolic homeostasis; hormone therapy has been shown to improve the relevant metabolic markers^[43]. Furthermore, males typically have greater skeletal muscle mass^[44], which may contribute to higher BCAA levels. The liver, as another major metabolic organ in BCAA metabolism, likely contributes to these sex differences through hormone-regulated metabolic processes and cross-talk with gut, muscle, and adipose tissues^[45]. Plasma BCAAs were also higher in individuals with obesity in our analyses. Obesity may elevate plasma BCAAs through a combination of excessive dietary BCAA intake and impaired BCAA catabolism within adipose tissue^[46]. The resulting accumulation of circulating BCAAs persistently activates the mTOR pathway, thereby inhibiting insulin signaling and reducing peripheral insulin sensitivity^[39,46]. This exacerbates dyslipidemia and broader metabolic dysfunction, perpetuating obesity in a vicious cycle.

Our observation that elevated BCAAs are associated with adverse CMH biomarker profiles aligns with their established role in cardiometabolic pathology^[47]. By systematically evaluating a comprehensive panel of biomarkers, we extended existing evidence beyond glucose and lipid homeostasis. Notably, the worsened renal function associated with BCAAs in our study suggests that plasma BCAAs may impair renal function by activating mTOR signaling and promoting fibrosis-related epithelial-mesenchymal transition^[48]. Although linear models in our study indicated an overall inverse association between plasma BCAAs and blood pressure, RCS analyses revealed a concentration-dependent, inverted U-shaped relationship. Mechanistically, our KEGG enrichment analyses of differential proteins offer a potential explanation. Among the 49 identified network proteins, both positively and negatively BCAA-associated proteins included blood pressure‑protective and risk‑associated factors, suggesting that multiple, potentially competing pathways may underlie the observed association^[49]. In addition, the overall negative association may be partly attributable to reverse causality or residual confounding inherent to the cross-sectional study design. We also observed significant sexual heterogeneity in the associations between plasma BCAAs and CMH biomarkers, potentially reflecting hormone-related differential regulation of BCAA metabolism and downstream pathways. Due to the absence of the protective effects of estrogen, men may be more susceptible to glucose and lipid metabolism disorders at elevated BCAA levels^[50].

Subsequent proteomic analyses revealed a distinct circulating signature associated with BCAAs. Elevated levels of BCAAs were positively correlated with proteins associated with atherosclerotic or cardiometabolic risk, including AGRP, FGF19, and PRAP1^[51]. Conversely, they were inversely associated with proteins involved in metabolic regulation and the maintenance of vascular function, including LPL, ADIPOQ, IGFBP1, IGFBP2, and SHBG^[51]. The observed upregulation of cardioprotective proteins like PYY, GIP, and FST may reflect compensatory endocrine and metabolic responses to early systemic stress, potentially capturing a subclinical stage of cardiometabolic dysregulation^[52-54]. Consistent with this interpretation, pathway enrichment analysis revealed that the mechanism underlying the association between plasma BCAAs and CMH may involve a multilevel molecular network. The shared pathways, including nitrogen metabolism, pantothenate and CoA biosynthesis, fatty acid degradation, cytokine-cytokine receptor interaction, and hormone signaling, are anchored by the consensus hub protein LEP (i.e., leptin), which is a key regulator of energy and metabolic homeostasis^[55]. Previous studies suggest that high BCAA levels often coexist with elevated leptin levels^[56]. Pathologically, aberrant leptin signaling induces insulin resistance, dyslipidemia, and elevated inflammatory markers, accelerating atherosclerosis^[57]. It also impairs renal hemodynamics and glomerular structure, leading to reduced eGFR^[58]. Collectively, these effects contribute to the onset and progression of cardiometabolic diseases. Beyond these shared mechanisms, we identified pathways specific to individual BCAA species. Notably, leucine-specific proteins localize to secretory granules and the extracellular matrix, where they regulate immune cell chemotaxis and migration through interactions involving chemokines, cytokines, receptors, and pattern recognition receptors, thereby contributing to tissue morphogenesis and repair. The leucine-specific network is centered on CXCL8 and SDC1. CXCL8 is a key chemokine that regulates the directed migration and activation of neutrophils and is involved in angiogenesis^[59]. SDC1 is a key regulator of signal transduction and lipid transport^[60] and may contribute to atherosclerosis. In contrast, valine is potentially involved in inflammatory injury and aberrant repair, concurrently with alterations in the immune-coagulation-fibrinolysis-apoptosis-remodeling pathway. Specifically, valine-specific protease F10 and PLAT are key components of coagulation and fibrinolysis, respectively, with antagonistic functions. Activated F10 serves as the core of the prothrombinase complex, catalyzing the conversion of prothrombin to thrombin and functioning as a critical convergence point of the intrinsic and extrinsic coagulation pathways^[61]. PLAT, a physiological activator of fibrinolysis, catalyzes plasminogen conversion to plasmin, initiating fibrin clot degradation^[62]. The coagulation-fibrinolysis balance is essential for vascular homeostasis, and valine may regulate this equilibrium, thereby influencing CMH-related pathophysiological processes, though the precise molecular mechanisms remain to be elucidated.

Several strengths support the robustness of our observations, including the large sample size, standardized metabolomic profiling, comprehensive dietary and biomarker assessment, and integration of proteomic data. Nevertheless, our findings require cautious interpretation. First, the cross-sectional design precludes causal inference. We cannot determine whether higher BCAAs are a cause, consequence, or correlate of cardiometabolic dysfunction. Second, diet was assessed using a baseline 24-h recall, which quantifies estimation of habitual intake but remains susceptible to dietary fluctuations and reporting bias. Third, given the consumption of mixed diets and the lack of external validation, associations between BCAAs and specific food categories should be interpreted with caution. Fourth, although we adjusted for major demographic, lifestyle, and dietary covariates, residual confounding due to food processing, cooking practices, underlying health conditions, and medication use cannot be ruled out.

In conclusion, this large-scale study demonstrates that plasma BCAAs are strongly influenced by modifiable host factors, particularly dietary factors, and are robustly associated with adverse cardiometabolic and proteomic profiles. These findings support a role for BCAAs as biomarkers associated with lifestyle factors and the development of cardiometabolic diseases. Future longitudinal studies and precision nutrition intervention trials are needed to establish causality and to determine whether dietary modulation of the BCAAs can mitigate cardiometabolic risk.