In the ICU environment, the stable commensal microbiota is rapidly replaced by a ‘dysbiosis signature’ characterized by a loss of diversity and the dominance of opportunistic pathogens. This shift is frequently driven by necessary but disruptive clinical interventions. For instance, Shao et al (27) recently emphasized that antibiotic-induced dysbiosis notably aggravates cerebral injury, particularly in patients undergoing extracorporeal membrane oxygenation. Such disruptions lead to a marked depletion of protective taxa, such as Bifidobacterium and Lactobacillus, which are key in maintaining immune homeostasis (28).

The resulting microbial landscape is often dominated by Proteobacteria, a shift that is associated with increased systemic inflammation. In patients with neuro-critical illness, this gastrointestinal dysbiosis is also a prognostic indicator of clinical severity (29). Furthermore, the loss of beneficial metabolites, such as SCFAs, impairs the ability of the host to regulate inflammatory responses within the gut lumen (30). From a mechanistic perspective, this loss of SCFAs represents not just the deficiency of metabolic fuel. SCFAs, particularly butyrate and valerate, act as endogenous histone deacetylase inhibitors (HDACs) that selectively target class I HDACs (such as HDAC3), thereby modulating the transcription of pro-inflammatory genes (including IL-1β, IL-6, TNF-α and IFN-γ) and reinforcing the intestinal epithelial barrier through epigenetic mechanisms, such as increased histone H3 acetylation at promoters of antiinflammatory genes (for example, Foxp3 and occludin) and reduced histone acetylation at promoters of proinflammatory genes (8,12). Consequently, the ‘ICU microbiome’ becomes a reservoir for virulence, where the overgrowth of pathobionts forms the basis of systemic translocation. In summary, the dysbiosis signature in critical illness represents a key transition from a symbiotic ecosystem to a pathogenic state that fuels systemic and organ-specific complications (29,31).

Serving as the primary defense against the translocation of luminal contents, the physical barrier of the gut is composed of a single layer of enterocytes joined by tight junctions. In critical illness, this barrier is compromised through a number of pathways. Alhasson et al (32) demonstrated that altered gut microbiomes in stress models cause intestinal injury through the activation of toll-like receptor 4 (TLR4), leading to increased permeability, often referred to as ‘leaky gut’. This process is further exacerbated by splanchnic hypoperfusion and the surge of inflammatory cytokines such as TNF-α, which promote enterocyte apoptosis and the disassembly of tight junction proteins (33).

In addition, specific molecular channels serve a role in this disintegration. Previous research by Meena et al (34) identified that the transient receptor potential vanilloid 6 channel mediates gut barrier dysfunction and systemic responses during metabolic stress, highlighting a potential target for barrier preservation. When these barriers fail, the influx of pathogen-associated molecular patterns into the portal and systemic circulation becomes uncontrolled (35). Emerging evidence has elucidated the precise molecular hierarchy associating barrier breakdown with neuroinflammation: Translocated lipopolysaccharides (LPS) activates circulating monocytes, triggering TNF-α and IL-1β release, which in turn compromise blood-brain barrier (BBB) integrity by downregulating claudin-5 and occluding (5,7,8,11). This is supported by multi-omics profiling demonstrating that fecal microRNA-gut microbiota interactions (such as miR-30e-3p) are associated with both gut barrier dysfunction and hippocampal neuroinflammation (5). This ‘leaky gut’ phenomenon is an important juncture in GBA disruption, as it converts a localized intestinal issue into a systemic inflammatory driver. In summary, the breakdown of intestinal integrity in the critically ill is a multifactorial process involving microbial signals, cytokine-mediated injury and ion channel dysregulation, all of which facilitate the dissemination of toxic substances (32,34).

Once the intestinal barrier is breached, gut-derived danger signals utilize both humoral and neural pathways to reach the brain. Metabolic endotoxemia, characterized by elevated levels of LPS in the blood, is a major driver of post-ischemic neuroinflammation and cognitive decline (36,37). Singer et al (38) provided key evidence that bacterial DNA and even viable bacteria can disseminate to the brain during sepsis, directly challenging the integrity of the BBB. These signals activate resident microglia, the primary immune cells of the brain, shifting them toward a pro-inflammatory M1 phenotype (39,40).

The neural pathway, primarily involving the vagus nerve, also serves as a rapid conduit for gut-derived signals. While the vagus nerve typically mediates anti-inflammatory effects, its afferent fibers can transmit distress signals to the nucleus tractus solitarius, which can either exacerbate neuroinflammation via downstream sympathetic and HPA axis activation or dampen it through the cholinergic anti-inflammatory pathway, depending on the context and the balance of efferent vagal activity (41,42). This neuro-immune crosstalk is particularly evident in sepsis-associated encephalopathy (SAE), whereby gut dysbiosis exacerbates hippocampal amyloid burden and microglial activation (43,44). Consequently, the brain is no longer an ‘immune-privileged’ site during critical illness (i.e., the blood-brain barrier disruption allows gut-derived inflammatory mediators and pathogens to access the central nervous system) but rather a direct recipient of intestinal distress. Overall, the afferent arm of GBA disruption involves a combination of circulating endotoxins and neural signaling that collectively trigger marked neuroinflammation and cognitive dysfunction (5,45).

GBA disruption is bidirectional, as central nervous system stress markedly impairs gastrointestinal function through efferent pathways. The activation of the hypothalamic-pituitary-adrenal (HPA) axis and the sympathetic nervous system leads to a state of ‘sympathetic overdrive’. Schmidt and Müller-Werdan (46) identified that autonomic dysfunction is a hallmark of patients with critical illness, manifesting as reduced heart rate variability and impaired gastrointestinal motility. This sympathetic hyperactivity, often seen following traumatic brain injury, redirects blood flow away from the mesenteric circulation, causing further mucosal ischemia (47,48).

In addition, the ENS is modulated by these central signals. Sepsis has been shown to induce cholinergic deficits and alter the behavior of calcium channels in mesenteric ganglion neurons, leading to gastroparesis and ileus (49,50). This impairment in motility not only prevents effective enteral feeding but also promotes further stasis and overgrowth of pathogenic bacteria, thus closing the pathogenic loop (41,51). The efferent arm therefore represents the ‘top-down’ contribution of the brain to gut failure, where stress-induced autonomic shifts exacerbate the dysbiosis and barrier breakdown that initiated the cycle. Recent research shows that efferent sympathetic signals directly suppress HDAC3 expression in intestinal epithelial cells, reducing antimicrobial peptide production and creating a feed-forward loop that impairs barrier function through the same HDAC pathway that SCFAs normally activate to maintain homeostasis (8). Ultimately, the efferent arm reinforces GBA disruption by compromising gut motility, mucosal blood flow and local immune defenses through sustained sympathetic and hormonal stress (52).

Afferent and efferent arms of GBA disruption do not operate in isolation; rather, they engage in a reciprocal amplification loop that transforms an acute insult into a self-sustaining pathological cascade. Gut-derived signals traversing the afferent arm (translocated LPS, pro-inflammatory cytokines and microbial metabolites) trigger neuroinflammation and HPA axis activation, which in turn potentiates the efferent arm through sympathetic overdrive and vagal dysfunction (5,45). This sympathetic hyperactivity exacerbates intestinal hypoperfusion, impairs mucosal barrier integrity and promotes dysmotility, thereby creating a permissive environment for further bacterial overgrowth and translocation (8). The clinical consequence of this bidirectional amplification is a ‘brain-gut-brain’ cycle, whereby the initial brain-directed inflammatory insult worsens gut function, which subsequently generates additional neurotoxic signals that perpetuate central nervous system injury (1). This self-reinforcing dynamic explains why GBA disruption, once established, often persists despite resolution of the original septic or traumatic trigger. Breaking this cycle therapeutically requires simultaneous targeting of both afferent signals (such as reducing endotoxemia and neuroinflammation) and efferent dysfunction (for example, restoring gut motility and barrier integrity), a principle that underpins the bundled nutritional approach.

ICU-acquired weakness (ICUAW) and profound muscle wasting are no longer viewed solely as consequences of disuse or malnutrition but are now recognized as products of the ‘gut-muscle axis’ (53,54). During sepsis and major trauma, systemic inflammation driven by intestinal barrier breakdown triggers skeletal muscle proteolysis through the activation of the ubiquitin-proteasome pathway. A study by Liu et al (53) specifically highlighted the potential role of the gut microbiota in sepsis-induced myopathy, suggesting that dysbiosis exacerbates muscle loss by modulating systemic inflammatory mediators, including TNF-α, IL-1β and IL-6 . This catabolic state is further worsened by the loss of beneficial microbial metabolites that normally support anabolic signaling in host tissues (18).

Beyond acute muscle loss, the persistence of a catabolic state contributes to the development of chronic critical illness. Recent experimental evidence has led to new animal models being established to evaluate persistent inflammation, immunosuppression and catabolism syndrome (PICS). Wang et al (54) developed a model demonstrating that gut microbiota dysbiosis and subsequent bacterial translocation are foundational to the development of PICS, associating intestinal failure directly with long-term catabolic outcomes. Consequently, the gut serves as a persistent ‘motor’ for systemic catabolism, hindering physical recovery even after the initial insult has resolved. In summary, ICUAW and catabolic failure in the ICU are rooted in the systemic inflammatory burden generated by a dysfunctional gut-brain-muscle network (53,54).

Clinical manifestation of neuroinflammation in the ICU often appears as delirium or SAE, which are associated with prolonged hospitalization and long-term cognitive decline. This neurological deterioration is closely associated with the influx of gut-derived neurotoxins and the disruption of the BBB. Giridharan et al (45) showed that the crosstalk between the gut and brain is a primary driver of sepsis-induced cognitive decline, where intestinal signals directly modulate hippocampal function. Furthermore, metabolic endotoxemia resulting from dietary or stress-induced gut permeability has been identified as a key mechanism associating high-fat intake or critical stress with human cognitive impairments (55,56).

The role of microbial metabolites in mitigating or exacerbating this cognitive decline has become a notable research focus. For instance, SCFAs have been shown to attenuate hippocampal neuroinflammation through the activation of NLRP6 inflammasome pathways, offering a protective effect against SAE (30,40). Conversely, the loss of these metabolites during critical illness leaves the brain vulnerable to microglial overactivation. Huang et al (39) recently suggested that gut microbiota-derived 3-indoleacetic acid provides protection against SAE by acting on microglial aryl hydrocarbon receptor (AhR), further proving that specific microbial signatures dictate neurological outcomes. Overall, cognitive dysfunction in patients with critical illness is a direct clinical reflection of GBA-mediated neuroinflammation and the loss of protective microbial signaling (39,45).

A notable proportion of ICU survivors suffer from a maladaptive immune state characterized by simultaneous hyper-inflammation and immunoparalysis. Approximately 30% of septic patients present with immunoparalysis, defined as failure of circulating mononuclear cells to produce cytokines upon ex vivo stimulation, often leading to T-cell exhaustion (2,7). Furthermore, among the ~7.6% of critically ill patients who develop chronic critical illness, 30–50% show evidence of PICS (5,6). The gut microbiota acts as a central regulator of this immunometabolic balance and its disruption often leads to clinical severity in conditions such as coronavirus disease-19. Cuenca et al (57) identified that specific dysbiosis patterns serve as clear indicators of severity in patients with critical illness, where the loss of microbial diversity is associated with uncontrolled inflammatory responses. This state of ‘septic instability’ is particularly pronounced in older populations, whereby the gut microbiota fails to maintain overall stability after a septic event compared with their younger counterparts (58).

Furthermore, the structural components of the intestinal environment, such as hyaluronan, have been shown to affect the progression of PICS by modulating the inflammatory milieu (59). When the gut remains in a dysbiotic state, it continuously primes the immune system toward a state of exhaustion or ‘paralysis’, rendering the patient susceptible to secondary healthcare-associated infections, such as ventilator-associated pneumonia, bloodstream infections, and urinary tract infections. This clinical trajectory is often driven by the overgrowth of Gram-negative pathogens, which can be predicted through structural equation modeling of gut overgrowth (60). In summary, the persistence of inflammation and the onset of immunosuppression in the ICU are largely dictated by the inability of the hosts to restore a homeostatic gut-microbiome environment (54,57).

‘Gut-lung axis’ represents a key clinical pathway where intestinal dysfunction complicates respiratory management. Patients who are mechanically ventilated experience rapid shifts in their pulmonary microbiome, which are often influenced by the translocation of bacteria from the oropharynx and gastrointestinal tract. Zakharkina et al (61) observed that the dynamics of the pulmonary microbiome during mechanical ventilation are associated with the occurrence of pneumonia, highlighting the lungs as a downstream victim of systemic dysbiosis. Respiratory tract dysbiosis itself has been associated with worse clinical outcomes and higher mortality rates in these populations (62,63).

In addition, the clinical success of weaning from mechanical ventilation is frequently hampered by systemic inflammation that weakens diaphragmatic function. The composition of the gut microbiome has been proposed as a notable predictor of 28-day mortality in patients who are mechanically ventilated, suggesting that the ‘health’ of the gut-lung axis determines the ability of the patients to liberate from life support (64,65). Interventions such as oral care or probiotic administration have been explored to mitigate these risks, although results vary across clinical trials: One trial found that probiotic oral care with Lactobacillus plantarum 299 was not superior to chlorhexidine in reducing oropharyngeal colonization or improving clinical outcomes in mechanically ventilated patients (66), while another evaluated the impact of a tooth-brushing-based oral hygiene protocol on oral bacteriota and healthcare-associated infections in ventilated COVID-19 patients, reporting changes in microbial composition and safety outcomes (67). Ultimately, prolonged mechanical ventilation and weaning failure are often the result of a feedback loop where gut-derived inflammation impairs respiratory drive and facilitates lung injury. In summary, the gut-lung axis is a decisive factor in the respiratory recovery of patients with critical illness (64).

Dietary fiber is frequently underutilized in the ICU due to concerns regarding bowel distension or incompatibility with enteral feeding tubes; however, the exclusion of fermentable substrates deprives the microbiota of key nutrients required for homeostasis. The primary mechanism by which fiber exerts neuroprotective effects is through the production of SCFAs, particularly butyrate and acetate. These metabolites serve as potent histone deacetylase inhibitors that downregulate inflammatory cytokines and reinforce the BBB. Fan et al (69) demonstrated that prebiotic fibers notably enhance SCFA production in endotoxemic models, leading to a reduction in gut commensal translocation and systemic inflammation. Consequently, the reintroduction of fermentable fibers acts as a key signal to suppress the systemic inflammatory response syndrome that drives encephalopathy.

Furthermore, non-fermentable fibers, such as cellulose, may possess distinct immunomodulatory properties beyond fermentation (70–72). Morowitz et al (70) showed that dietary supplementation with nonfermentable fiber alters the gut microbiota composition and confers notable protection in murine models of sepsis. This suggests that the physical structure of fiber may provide a scaffold for beneficial bacteria or mechanically stimulate mucus production, thereby enhancing the physical barrier against pathogen invasion. In addition, the administration of specific cellulose supplements has been observed to modulate immune responses in endotoxemia by reducing the production of pro-inflammatory cytokines (including TNF-α, IL-1β and IL-6), attenuating NF-κB pathway activation, and decreasing neutrophil infiltration, thereby preventing the apoptosis of intestinal epithelial cells (71,72). Therefore, fiber supplementation should be viewed as a ‘prebiotic pharmaceutical’ that maintains the functional integrity of the GBA rather than an optional additive for bowel regularity.

In addition to direct barrier support, the metabolic byproducts of fiber fermentation serve an important role in systemic immune regulation. Chancharoenthana et al (73) recently demonstrated that modulation of sepsis by Lacticaseibacillus rhamnosus is associated with the levels of SCFAs in feces and blood. At the molecular level, SCFAs exert effects through two complementary mechanisms: Activation of G-protein-coupled receptors (GPR41/43) on immune cells and inhibition of class I HDACs. Butyrate and valerate increase histone acetylation at promoters of anti-inflammatory genes (Foxp3 and occludin) and recent evidence from an SAE murine model has demonstrated that SCFA supplementation activates NOD-like receptor family pyrin domain containing 6 (NLRP6) inflammasomes, restores intestinal barrier [Zonula Occludens-1 (ZO-1): 1.8-fold increase] and ameliorates cognitive deficits [escape latency: cecal ligations and puncture (CLP) 48 sec vs. SCFA + CLP 32 sec; P<0.01] (11). Thus, providing adequate fermentable substrates represents the basic principle of GBA-targeted nutrition, ensuring that the microbiome retains the capacity to synthesize neuroprotective metabolites (74,75).

Restoring microbial diversity through the direct administration of live biotherapeutics represents a proactive strategy to reverse the ‘ICU dysbiosis signature’. While the use of probiotics in patients with critical illness has previously been controversial due to safety concerns, emerging evidence supports their efficacy in specific phenotypes to reduce infection and modulate immunity. Shimizu et al (76) reviewed the application of synbiotics in critical care, highlighting their role in normalizing gut immunity and preventing the overgrowth of pathogenic bacteria that trigger systemic inflammation. By competitively excluding pathogens and enhancing the expression of tight junction proteins, probiotics act as a ‘living shield’ for the intestinal mucosa.

The mechanistic basis for these benefits involves complex signaling pathways between the microbe and the host immune system. Chen et al (77) highlighted that Lactobacillus rhamnosus GG promotes the recovery of the colon barrier in septic mice specifically by accelerating the regeneration of intestinal stem cells. This regenerative capacity is important for patients suffering from ischemia-reperfusion injury or antibiotic-associated mucosal damage. Beyond established barrierrepair mechanisms, including upregulation of tight junction proteins (for example, occludin, claudins and ZO-1), suppression of pro-inflammatory cytokines, inhibition of NF-κB and MLCK signaling, activation of TLR-2, and modulation of autophagy, specific gut microbial metabolites have been shown to directly modulate neuroinflammation through the GBA (77). Indole derivatives, particularly 3-indoleacetic acid (IAA), produced by tryptophan-metabolizing bacteria, act as ligands for the AhR expressed on microglia. Huang et al (39) demonstrated that patients with sepsis and SAE exhibit markedly reduced fecal IAA levels. In a murine CLP model, exogenous IAA supplementation was shown to activate microglial AhR, suppress the secretion of proinflammatory cytokines (including TNF-α, IL-1β and IL-6), and improve cognitive performance. This protective effect was shown to be abolished by co-administration of the AhR antagonist CH223191, demonstrates that IAA confers neuroprotection in an AhR-dependent manner. In addition, recent trials have explored the use of synbiotics, which combine probiotics with prebiotic substrates, to enhance the survival and colonization of beneficial strains. Safabakhsh et al (78) reported preliminary results from a double-blind trial indicating that probiotic supplementation effectively reduces intestinal permeability in critically ill children with sepsis, suggesting that early intervention can preserve GBA integrity across different age groups.

Beyond established probiotics, the field is advancing toward precision engineering and postbiotics. Yao et al (79) developed an innovative approach using SpyTag-PEGylated probiotics to deliver IL-1 receptor antagonists, which successfully modulated gut-lung crosstalk and mitigated septic lung injury. This illustrates the potential of ‘pharmaconutrition’ where genetically or chemically modified microbes serve as vehicles for targeted anti-inflammatory therapy. Furthermore, the administration of specific microbial metabolites, or postbiotics, allows for the benefits of bacterial signaling without the risk of administering live organisms to patients who are immunocompromised (80). In summary, direct microbial modulation through probiotics and synbiotics offers a potent mechanism to stabilize the enteric environment and dampen the vagal and humoral signals that drive neuroinflammation (81,82).

Nutrient delivery route is a decisive factor in maintaining the functional architecture of the GBA. Total parenteral nutrition (TPN), while life-saving in cases of intestinal failure, is associated with mucosal atrophy and the loss of gut-associated lymphoid tissue function. Ikeda et al (83) demonstrated that the lack of enteral feeding leads to a pathological upregulation of intestinal TLRs and pro-inflammatory cytokines, sensitizing the gut to endotoxin. This ‘disuse atrophy’ creates a permissive environment for bacterial translocation, fueling the systemic inflammation that ultimately affects the brain. Conversely, the physical presence of nutrients in the lumen stimulates the release of incretins (for example, glucose-dependent insulinotropic polypeptide and glucagon-like peptide-1) and trophic factors (for example, glucagon-like peptide-2, insulin-like growth factor-1 and epidermal growth factor) that maintain enterocyte health.

A key concept in understanding TPN-induced dysfunction is the phenomenon of bacterial adaptation to starvation. Ralls et al (84) provided insight into the gut origin of sepsis by showing that during enteral nutrient deprivation, commensal bacteria switch their transcriptional profile to ‘forage’ on the host mucus layer. This degradation of the protective mucus barrier allows bacteria to come into direct contact with the epithelium, triggering inflammation and barrier failure. Therefore, even minimal amounts of EN, often termed ‘trophic feeding’, are key in satisfying the metabolic demands of the microbiota and prevent them from consuming the defensive layers of the host (85,86).

In addition, early EN has been associated with specific cellular protective mechanisms. Hu et al (87) found that early EN preserves intestinal barrier function by reducing the formation of neutrophil extracellular traps in critically ill surgical patients. By minimizing local neutrophil activation, enteral feeding breaks the cycle of local inflammation that leads to systemic organ dysfunction. This evidence supports the principle that the enteral route should be prioritized not just for caloric delivery, but as a therapeutic maneuver to maintain the physiological barrier between the gut lumen and the systemic circulation (87,88).

While enteral stimulation is important, the volume and composition of nutrition must be carefully titrated to avoid metabolic overload during the acute phase of critical illness. Aggressive full feeding can induce hyperglycemia and suppress autophagy, a cellular cleaning process important for clearing intracellular pathogens and damaged organelles. Hadley and Hinds (89) discussed anabolic strategies, noting that pushing calories in the early phase of critical illness often fails to halt catabolism and may exacerbate metabolic stress. A GBA-targeted approach advocates for trophic feeding, delivering small volumes to sustain mucosal integrity and microbial diversity, without overwhelming the compromised mitochondrial function of the host.

Specific nutrient compositions also serve a role in modulating the inflammatory response during feeding. Luyer et al (90) observed that pretreatment with high-fat EN reduces endotoxin and TNF-α levels, preserving gut barrier function after hemorrhagic shock. This suggests that lipid-rich formulas may stimulate the cholecystokinin-vagal pathway, triggering an anti-inflammatory reflex that protects the gut. However, the type of lipid is key; omega-3 fatty acids and whey peptides have been shown to offer improved protection against LPS-mediated sepsis compared with standard lipids (91).

The goal of titrating feeding strategies is to maintain a ‘metabolic peace’ between the host and the microbiome. Oami et al (24) reviewed critical care nutrition from a metabolic point of view, suggesting that nutritional interventions must align with the dynamic phases of critical illness. By initiating trophic feeding early, clinicians can prevent the ‘starving gut’ phenomenon and support the GBA, while gradually advancing to full caloric targets as the patient stabilizes. Consequently, the principle of GBA-targeted nutrition prioritizes the biological impact of the feed over the immediate achievement of caloric goals, ensuring that nutrition supports rather than hinders the resolution of inflammation (24,92).

Probiotics and synbiotics constitute a hallmark of the GBA-informed bundle by restoring microbial diversity and inhibiting the translocation of pathogens. Extensive research has focused on their role in preventing healthcare-associated infections, particularly ventilator-associated pneumonia (VAP) (93,94). For example, an early clinical trial has demonstrated that probiotic prophylaxis could markedly reduce the incidence of VAP in pediatric populations (93). Similarly, a meta-analysis of adult patients in the ICU have indicated that synbiotic therapy effectively modulates gut microbiota and reduces both enteritis and pneumonia rates (94).

However, evidence remains nuanced regarding mortality and universal application. Johnstone et al (95) conducted a large-scale randomized trial which demonstrated that Lactobacillus rhamnosus GG did not notably decrease VAP incidence compared with placebo in a broad critically ill population. This finding has been further determined in recent systematic reviews and trial sequential analyses, which suggest that while probiotics may reduce infection rates, their impact on overall mortality remains unremarkable (96,97). A systematic analysis of heterogeneity across these trials has revealed a number of key factors that may explain the conflicting results. Strain specificity is important: The negative trial used a single strain without prebiotics, whereas positive trials often used multi-strain synbiotics (94). Dosage and timing also vary markedly, with earlier initiation (within 48 h) exhibiting greater benefits. Patient population differences are also important, as benefits appear more notable in surgical patients compared with medical ICU populations (96). These sources of heterogeneity suggest that a standardized probiotic protocol is unlikely to succeed across all patients in the ICU. Despite these challenges, certain strains have shown specific promise. Panigrahi et al (98) demonstrated in a landmark study that a synbiotic preparation markedly reduced neonatal sepsis in a rural setting, highlighting the potential for targeted GBA modulation.

In addition to infectious outcomes, probiotics appear to influence the systemic inflammatory response and organ function. Probiotic administration has been associated with the modulation of serum cytokine levels, including proinflammatory cytokines (TNFα, IL6 and IL8) and antiinflammatory cytokines (IL10), and endotoxin concentrations in various critical and chronic illness models, such as clinical trials in pediatric sepsis, animal models of sepsis, and studies in acute pancreatitis and liver cirrhosis (99). Furthermore, multispecies probiotics have been shown to attenuate dysbiosis and improve clinical markers, including disease severity scores (SOFA and APACHE II), inflammatory markers (CRP and PCT), and endotoxin levels, in the early stages of sepsis (81). Collectively, these findings have suggested that while probiotics are a beneficial tool for GBA modulation, their efficacy is likely strain-specific and timing-dependent and systematic assessment of trial heterogeneity is key before generalizing results.

Prebiotics and dietary fibers serve as important substrates for the production of SCFAs, which are key in maintaining the intestinal epithelial barrier. Systematic reviews have highlighted that fiber and prebiotic supplementation in EN can improve gastrointestinal tolerance and reduce the duration of diarrhea in patients with critical illness (100,101). By fostering the growth of beneficial bacteria such as Bifidobacterium, prebiotics mitigate the metabolic endotoxemia often observed in systemic inflammation (102).

However, clinical evidence on fiber interventions shows marked heterogeneity. The type and fermentability of fiber differ: Soluble fiber blends reduce diarrhea more effectively compared with nonfermentable fibers. Doses range from 10–30 g/day, with higher doses increasing intolerance. Patient populations have been shown to respond differently; trauma patients show favorable microbial shifts (103), while medical patients in the ICU with ileus may not tolerate fiber. Timing is also important; early fiber may help stable patients but could exacerbate shock states. These factors may explain contradictory trial outcomes. Recent evidence has suggested that the inclusion of fiber in enteral formulas can dynamically shift the microbial community in patients with trauma (103). This shift is key in GBA health, as it promotes the production of butyrate, which has known neuroprotective and anti-inflammatory properties. A clinical trial has also indicated that prebiotic supplementation may improve sleep quality and metabolic markers through the GBA in patients with chronic metabolic disturbances, which may have implications for ICU delirium and recovery (102). In summary, prebiotics are not passive nutrients but active modulators of the gut environment. Despite this, heterogeneity in fiber type, dose, patient selection and timing must be addressed in future trials to establish standardized recommendations.

Resolution of inflammation is an active process mediated by specialized pro-resolving mediators (SPMs) derived from omega-3 polyunsaturated fatty acids. Dalli et al (104) identified novel pro-resolving pathways that are key in tissue regeneration and the termination of acute inflammatory responses. In the context of sepsis, the temporal profile of these lipid mediators is closely associated with clinical survival and the resolution of organ dysfunction (105).

Experimental and clinical studies have shown that omega-3 fatty acids such as docosahexaenoic acid and eicosapentaenoic acid can attenuate septic shock-induced arterial dysfunction and neuroinflammation (106,107). Furthermore, specific SPMs such as maresin-1 have been found to ameliorate microglial activation and cognitive decline in septic models by modulating inflammatory signaling pathways, including inhibition of p38 MAPK phosphorylation and activation of the PPARγ/STAT6 signaling pathway (108,109). These findings suggest that incorporating omega-3 fatty acids into the care bundle may facilitate the transition from systemic inflammation to homeostatic resolution through the GBA. Therefore, lipid-based GBA modulation represents an emerging frontier for protecting the central nervous system during critical illness.

Polyphenols such as curcumin and resveratrol offer potent antioxidant and anti-inflammatory benefits that reinforce the GBA. Research has shown that curcumin can alleviate LPS-induced lung and liver failure by suppressing oxidative stress and NF-κB signaling (110,111). In addition, nano-curcumin supplementation in patients with sepsis has been associated with improvements in endothelial function and the reduction of inflammatory biomarkers (112).

Resveratrol has similarly demonstrated protective effects against sepsis-induced cardiac and renal dysfunction in a number of experimental models, including cecal ligation and puncture-induced sepsis in rats and mice (113). These compounds appear to work by activating sirtuin-1 and nuclear factor erythroid 2-related factor 2 (Nrf2) pathways, which are key in maintaining mitochondrial dynamics and cellular homeostasis during severe stress (114,115). Recent advances have also explored the use of resveratrol-loaded nanoparticles to enhance efficacy in addressing acute lung injury (116). Ultimately, the integration of polyphenol-rich compounds into nutrition care provides an additional layer of neuro-metabolic protection for the critically ill.

Effectiveness of the GBA-informed bundle depends on the delivery method and the timing of EN. Research into feeding tolerance has indicated that the early initiation of EN, potentially supplemented with synbiotics, can improve protein homeostasis and reduce muscle wasting (117,118). Notably, the choice between sequential and continuous feeding may also influence the composition of the gut microbiota and the overall metabolic response (119).

Structured EN protocols have been shown to improve nutritional indicators and clinical prognosis in neurological patients in the ICU (118). Furthermore, the duration of oral care using maternal milk or other biological substrates can modulate the oral microbiota composition and reduce the incidence of healthcareassociated infections, thereby positively affecting the oralgut microbial axis and improving clinical outcomes in vulnerable populations (120). These strategies underscore the importance of not only what is delivered but how it is administered to optimize GBA communication. In summary, a standardized approach to feeding delivery is important in the successful implementation of the GBA-informed nutrition care bundle.

While the proposed GBA-informed bundle provides a standardized framework, minor evidence-based adjustments may enhance its applicability to specific critically ill subpopulations. For patients with sepsis, probiotic initiation should be deferred until hemodynamic stabilization (norepinephrine ≤0.5 µg/kg/min) due to translocation risks (96,121). For patients with major trauma, early synbiotic administration (within 24–48 h) combined with higher protein targets (1.5–2.0 g/kg/day) may better address hypercatabolism and reduce infectious complications. For patients with neurocritical illness, protocolized EN with prebiotic-fortified formulas has been associated with improved nutritional indicators, including serum albumin, prealbumin and nitrogen balance, as well as lower diarrhea rates (118). These adaptations do not alter the core components of the bundle but refine their timing and implementation across distinct pathophysiological contexts.

Adoption of standardized protocols is a key component of effective nutritional management in the ICU. Heyland et al (122) demonstrated that implementation of the Enhanced Protein-Energy Provision via the enteral Route (PEP uP) protocol, which emphasizes volume-based feeding (VBF) and nurse-driven adjustments, notably improves the delivery of EN compared with traditional hourly rate-based feeding methods. This approach has been further validated in surgical and neuroscience populations, where VBF protocols have been shown to effectively enhance protein and energy provision (123,124). A systematic review has further determined that such protocolized systems demonstrate an improved performance with regard to reaching nutritional goals and minimizing caloric deficits in patients who are mechanically ventilated (125).

Furthermore, the application of early EN protocols based on enhanced recovery after surgery concepts has shown promise in improving clinical outcomes and shortening ICU stays (23). Recent studies have highlighted that the success of these protocols depends on adherence and the use of clinical decision support tools (126,127). Ventura and Waitzberg (128) suggested that while protocols are necessary, their effectiveness is associated with the education of the multidisciplinary team. Collectively, these findings suggest that structured, protocol-driven strategies are key in the consistent delivery of GBA-targeted interventions across diverse clinical settings.

Despite the availability of protocols, achieving nutritional targets remains challenging due to frequent interruptions for medical procedures and perceived gastrointestinal intolerance. Kuslapuu et al (17) identified that a notable proportion of EN interruptions in the ICU lack a clear evidence-based justification, leading to suboptimal feeding. To mitigate this, specific protocols designed to avoid tube feed interruptions during procedures such as tracheostomies have been implemented with high success rates (129). In addition, managing common complications such as diarrhea is important in maintaining bundle adherence.

Research has indicated that the development of predictive models for diarrhea can help clinicians identify patients at high-risk and preemptively adjust nutritional strategies (130). Studies have also explored the impact of feeding patterns, noting that while intermittent feeding may more closely mimic physiological states, continuous feeding remains a standard in numerous protocols to ensure volume delivery (131,132). Furthermore, addressing constipation through prophylactic laxative regimens is an often overlooked but key aspect of GBA health (133). In summary, identifying and systematically removing practical barriers is a prerequisite for achieving the physiological benefits of GBA modulation at the bedside.

Safety is an important consideration when implementing specialized GBA-targeted nutrients, particularly in patients who are hemodynamically unstable. The results of large-scale trials such as REducing Deaths due to OXidative Stress (REDOXS) and METAPLUS have suggested that high-dose glutamine and antioxidant supplementation might increase mortality in patients with multi-organ failure (134,135). However, Wischmeyer (136) argued that these findings should not lead to the total abandonment of glutamine, especially in populations such as burn patients where recent evidence continues to show benefit (137). This underscores the necessity of a risk-stratified approach rather than a universal application of all bundle components.

To enhance clinical operability for high-risk populations, a refined risk stratification has been proposed. For probiotics, absolute contraindications include: Central venous catheter (CVC) in situ due to fungemia risk, particularly with Saccharomyces boulardii, with an associated mortality of 36.1% (138,139). Additional contraindications comprise neutropenia (<500/µl), short bowel syndrome and high-dose vasopressor support (norepinephrine >0.5 µg/kg/min) indicating hemodynamic instability (96,121). Patients who are immunocompromised (active chemotherapy, solid organ transplant on intensive immunosuppression) should generally not receive live probiotics (96). For glutamine and high-dose antioxidants, evidence from the REDOXS trial has demonstrated harm in patients with multiorgan failure (especially renal dysfunction) and those in shock (140). Thus, in patients with multiorgan failure or ongoing shock, high-dose glutamine should be withheld. For stable patients without these risk factors, the full GBA-informed bundle may be safely initiated.

The use of probiotics also requires careful safety monitoring, as rare cases of Saccharomyces fungemia have been reported in patients with critical illness receiving these supplements (138,139). Clinicians must weigh the proven benefits of probiotics in reducing VAP against the potential for infectious complications in severely immunocompromised individuals (96,121). Therefore, a nuanced understanding of patient-specific contraindications is important for the safe implementation of microbial modulation strategies. Ultimately, safety protocols must be integrated into the GBA-informed bundle to protect vulnerable ICU populations from potential adverse effects.

Modern implementation strategies have increasingly incorporated advanced technology to guide and monitor nutritional therapy. Ultrasound-guided protocols for assessing gastric residual volume and intestinal motility offer a non-invasive and reliable method to evaluate feeding tolerance at the bedside (141,142). These technological advances allow for real-time adjustments to the nutrition bundle, ensuring that delivery is optimized according to the changing physiological status of the patient. In addition, the involvement of a dedicated nutrition support team has been shown to notably improve the administration and monitoring of EN (143).

In the future, the integration of biomarkers and metabolomics may allow for even more precise GBA-informed care (144). For example, monitoring fecal calprotectin or microbial metabolites could provide early warnings of intestinal barrier failure or dysbiosis (145). The coordination between dietitians, nurses and physicians remains important in successful bundle implementation, ensuring that each component of the GBA-informed care is delivered accurately (22). Overall, the synergy between technological innovation and multidisciplinary collaboration is key in the global adoption and long-term success of the GBA-centric nutrition framework.

To address the lack of standardization and stratification, the present section provides a concise, ready-to-use protocol (Table IV). Population stratification is based on the modified Nutrition Risk in the Critically Ill (NUTRIC) score (≥5 indicates high risk) combined with disease category (sepsis/septic shock, trauma or neuro-critical illness) (21,118). Explicit contraindications for probiotics include CVC, neutropenia (<500/µl), short bowel syndrome and concurrent high-dose vasopressors (norepinephrine >0.5 µg/kg/min) (96,138). Dynamic adjustment follows three phases: i) Acute phase (days 1–3), trophic EN (10–20 ml/h) and no probiotics; ii) stabilization phase (days 3–7), gradual volume increase and initiate prebiotic fiber (10–15 g/day) and probiotics after vasopressor weaning; and iii) recovery phase (>day 7), full bundle including omega-3 (5–10 g/day) if tolerated (24,68). Hold parameters include new-onset diarrhea (>3 episodes/24 h), gastric residual volume >500 ml or signs of bowel ischemia. Standard operating procedures are summarized in Table IV.