Paraprobiotics and postbiotics: from mechanistic evidence to food applications and clinical translation

1 Introduction

Probiotics are live microorganisms that, when administered in adequate amounts, confer a health benefit on the host (Hill et al., 2014). Their use in fermented foods, supplements, and specialized nutrition is well established, but maintaining viability during processing, storage, distribution, and gastrointestinal transit remains a technological constraint. Safety also requires context-specific assessment because rare cases of invasive infection have been reported in vulnerable individuals receiving live microbial products (Churin et al., 2026). These limitations have stimulated interest in non-viable microbial preparations that can retain biologically active cellular structures or fermentation-derived components without requiring viable-cell maintenance.

Paraprobiotics generally refer to non-viable intact or fragmented microbial cells that confer a health benefit, whereas ISAPP defines a postbiotic as a preparation of inanimate microorganisms and/or their components that confers a health benefit on the host (Barros et al., 2020; Siciliano et al., 2021; Salminen et al., 2021). The distinction is important because purified metabolites, bacteriocins, peptides, EPS, cell-free supernatants, or extracellular vesicles do not automatically qualify as postbiotics when evaluated in isolation. Non-viable preparations may reduce selected viability-dependent safety concerns and can facilitate formulation in food systems, but safety, stability, dose standardization, and biological activity remain preparation- and matrix-specific rather than universal properties. In this review, paraprobiotics and postbiotics are treated as partially overlapping but not synonymous concepts. The term paraprobiotic is used operationally for non-viable, intact, or fragmented microbial cells associated with a health benefit, whereas the term postbiotic is reserved for preparations that satisfy the ISAPP definition, namely, preparations of inanimate microorganisms and/or their components that confer a demonstrated health benefit on the host. Accordingly, a paraprobiotic preparation may also qualify as a postbiotic when it fulfills these criteria, but the two terms should not be used interchangeably. This distinction is particularly important because the term “paraprobiotic” has not been defined through an international consensus framework comparable to that developed by ISAPP for postbiotics (Barros et al., 2020; Siciliano et al., 2021; Salminen et al., 2021).

This review critically evaluates paraprobiotics and postbiotics across molecular mechanisms, human-health evidence, food applications, production technologies, regulatory frameworks, and future translation. Particular emphasis is placed on distinguishing direct evidence for well-characterized non-viable preparations from findings obtained with viable probiotics, multicomponent formulations, or isolated microbial-derived products that do not necessarily meet the current ISAPP definition.

2 Literature search and evidence framework

This narrative review was developed through literature searches conducted in PubMed, Scopus, Web of Science, and Google Scholar for publications available up to July 2026. Search terms included combinations of “paraprobiotic,” “postbiotic,” “non-viable microorganism,” “inactivated probiotic,” “heat-killed probiotic,” “microbial-derived component,” “food application,” “inactivation,” “stability,” “sensory,” “clinical trial,” “regulation,” and “scale-up.” Peer-reviewed experimental and clinical studies, consensus statements, and relevant regulatory documents were considered. Priority was given to studies clearly reporting the microbial source or strain, preparation or inactivation procedure, experimental or clinical model, and biological or technological outcomes. Studies involving viable probiotics were retained only when they provided relevant comparative context. Studies focusing exclusively on isolated metabolites, bacteriocins, cell-free products, or extracellular vesicles were not classified as direct postbiotic evidence unless the preparation met the current ISAPP definition. Duplicate publications, studies outside the scope of the review, and reports lacking sufficient methodological or compositional information were excluded. Reference lists of key publications were also screened for relevant studies. The resulting evidence was synthesized as a critical narrative review rather than a systematic review or meta-analysis.

3 Comparative analysis of biotic concepts

Figure 1 illustrates selected microbial inactivation approaches; the processing conditions shown are examples rather than universal manufacturing specifications. Table 1 compares probiotics, paraprobiotics, and postbiotics. Throughout this review, isolated short-chain fatty acids (SCFAs), bacteriocins, purified proteins or peptides, EPS, cell-free supernatants, and bEVs (bacterial extracellular vesicles) are described as microbial-derived components or products unless they are part of a preparation that meets the ISAPP definition of a postbiotic. Biological effects are interpreted as preparation-specific and are not assigned universally to paraprobiotics or postbiotics.

TABLE 1. Comparative characteristics of probiotics, paraprobiotics, and postbiotics.

According to the ISAPP, consensus definition, purified microbial metabolites such as SCFAs, bacteriocins, peptides; EPS, or other cell-free compounds do not, by themselves, qualify as postbiotics.A postbiotic preparation must contain inanimate microbial cells and/or their components and must confera demonstrated health benefit on the host.

4 Biological mechanisms of action

Non-viable microbial preparations can influence host signaling, barrier function, immune responses, and microbiota-related processes through cell-associated structures and other preparation-specific components. The following mechanisms should therefore be interpreted in relation to the composition of the preparation studied.

4.1 Immunomodulation and pattern recognition receptor signaling

Non-viable microbial preparations can modulate innate immune signaling through cell-associated components such as peptidoglycan, lipoteichoic acid, S-layer proteins, and EPS, which may interact with TLR2, TLR4, DC-SIGN (dendritic cell-specific intercellular adhesion molecule-3-grabbing non-integrin), NF-κB (nuclear factor kappa B), and MAPK pathways (Kim et al., 2012; Konstantinov et al., 2008; Wachi et al., 2014). In Lactobacillus delbrueckii subsp. delbrueckii TUA4408L, bacterial cells and acidic or neutral EPS fractions differentially influenced Toll-like receptors (TRL) TLR2/TLR4-and TLR3-related responses in porcine intestinal epithelial cells, with effects depending on the EPS fraction and inflammatory or viral stimulus (Wachi et al., 2014; Kanmani et al., 2018). Heat-killed Lactiplantibacillus plantarum KU15122 reduced nitric oxide, reactive oxygen species, inflammatory mediators, and NF-κB/AP-1/MAPK signaling in LPS-stimulated RAW 264.7 macrophages (Lee et al., 2023). Heat-killed Lactiplantibacillus argentoratensis BBLB001 similarly reduced IL-8 in a Caco-2/RAW264.7 co-culture and attenuated inflammatory outcomes in dextran sulfate sodium (DSS)-induced colitis in mice (Itoh et al., 2024). All these studies show that immunomodulatory activity depends on strain, component, preparation, and model; universal percentage changes in cytokine production should therefore not be assigned to paraprobiotic or postbiotic preparations. Table 2 summarizes representative non-viable microbial preparations, microbial structural components, metabolites, bacteriocins, and other microbial-derived products evaluated in experimental and clinical studies.

TABLE 2. Representative non-viable microbial preparations, microbial-derived components, and related products evaluated in experimental and clinical studies.

The table includes defined non-viable microbial preparations along with isolated microbial structural components, metabolites, bacteriocins, and microbially derived products because they provide mechanistic, preclinical, technological, or clinical evidence relevant to this review. Inclusion in the table does not imply classification asa postbiotic. Purified metabolites, bacteriocins, isolated structural components, cell-free supernatants, and other microbially derived products do not independently meet the ISAPP, definition ofa postbiotic. Preparations evaluated only in preclinical models are described conservatively as non-viable microbial or paraprobiotic-type preparations unlessa demonstrated health benefit supports their classification asa postbiotic.

4.2 Intestinal barrier integrity: role of SCFAs

SCFAs, principally acetate, propionate, and butyrate, are microbial fermentation products that regulate intestinal epithelial and immune functions. Their effects include signaling through GPR41/FFAR3, GPR43/FFAR2, and GPR109A/HCAR2 (G protein-coupled receptors 41, 43, and 109A), with experimental studies showing modulation of inflammatory responses and regulatory T-cell pathways (Sun et al., 2017; Kim et al., 2012; Singh et al., 2014). Butyrate also acts directly on epithelial barrier function. In porcine jejunal organoid-derived monolayers, it attenuated deoxynivalenol-induced barrier disruption and partly counteracted changes in tight-junction-associated genes (Alberge et al., 2024). In Caco-2BBe cells and DSS-induced colitis, butyrate increased Slc26a3 expression through inhibition of the HDAC8 (histone deacetylase 8)/NF-κB pathway and improved barrier function (Peng et al., 2024). These findings support receptor-mediated and epithelial mechanisms, but effects depend on the specific SCFA, concentration, model, and pathological context. Purified SCFAs should therefore be described as microbial-derived metabolites rather than automatically classified as postbiotics.

4.3 Antimicrobial activity and microbiota modulation

Antimicrobial activity is preparation- and compound-specific. Thuricin CD showed minimum inhibitory concentrations below 3.1 μg/mL against Clostridioides difficile and Bacillus firmus and a narrower spectrum than nisin, fidaxomicin, or vancomycin in a 47-strain comparison and mini-fecal fermentation model, indicating that microbiota-sparing activity cannot be generalized to all bacteriocins (Walsh et al., 2024). Tong et al. reported bactericidal activity of a water extract derived from Bacillus amyloliquefaciens J and Lactiplantibacillus plantarum SN4 against Escherichia coli, Staphylococcus aureus, and Salmonella Typhimurium, with membrane disruption and organic acids contributing to the effect (Tong et al., 2025). These findings support antimicrobial potential but require preparation-specific validation.

4.4 Gut–brain axis: SCFA-mediated neural and enteroendocrine signaling

SCFAs may contribute to gut–brain communication through enteroendocrine, vagal, immune, and potentially direct central pathways, although human evidence remains limited (Mansuy-Aubert and Ravussin, 2023; Mathias et al., 2024). In mice, the intestinal microbiota increased colonic tryptophan hydroxylase 1 expression and peripheral serotonin production (Reigstad et al., 2015). In experimental neuropathic pain models, Roseburia intestinalis and butyrate attenuated pain-related responses through GPR41-expressing vagal sensory neurons and downstream brain signaling (Jiang et al., 2025). SCFAs may also influence blood–brain barrier and glial function through receptor-mediated and epigenetic mechanisms, but available evidence is predominantly experimental. These findings support biological plausibility rather than established clinical neurological benefits of non-viable microbial or microbial-derived interventions.

4.5 bEVs as microbial-derived bioactive effectors

bEVs are nanoscale membrane-derived particles that can carry proteins, lipids, nucleic acids, enzymes, and metabolites. Their size, composition, and biological activity depend on the producing strain, culture conditions, and isolation method. Preparations containing bEVs may fall within the ISAPP postbiotic definition only when they are part of a preparation of inanimate microorganisms and/or their components that confers a demonstrated health benefit; microbial origin alone is insufficient. Experimental studies indicate that bEVs from microorganisms such as Lacticaseibacillus rhamnosus GG and Akkermansia muciniphila can influence inflammatory signaling and intestinal barrier function, while selected vesicles may participate in inter-organ communication. All these observations remain predominantly preclinical. Translation will require standardized isolation, purification, quantification, cargo characterization, stability, potency, and safety assessment (Li et al., 2024; Ronacher et al., 2025; Tong et al., 2021; Chelakkot et al., 2018; Kang et al., 2020; Yaghoubfar et al., 2020; Liu et al., 2021; Pirolli et al., 2023).

5 Applications in human health and the food industry

In the following sections, evidence is differentiated according to study type. Direct clinical evidence refers to interventions using a defined non-viable microbial preparation; multicomponent evidence refers to formulations in which the independent contribution of the microbial preparation cannot be isolated; studies using viable probiotics are included only as comparative context; and cellular or animal studies are considered preclinical evidence.

5.1 Applications in human health

Non-viable microbial preparations have been investigated across gastrointestinal, metabolic, immune-related, dermatological, and other clinical contexts. The strength of evidence varies substantially among indications and preparations, and many reported effects remain preclinical or originate from multicomponent formulations. Compared with live probiotics, non-viable preparations eliminate the capacity for microbial replication and may therefore reduce selected viability-dependent risks, but non-viability should not be interpreted as evidence of complete safety. Compositional, toxicological, and clinical evaluation remains necessary for each preparation and intended population (Figure 2).

5.1.1 Gastrointestinal disorders

Data supporting the use of non-viable microbial preparations in gastrointestinal disorders remains strain-, preparation-, and disease-specific. In inflammatory bowel disease (IBD), most available studies are preclinical. Itoh et al. evaluated heat-killed Lactiplantibacillus argentoratensis BBLB001 in a Caco-2/RAW264.7 co-culture model and in mice with dextran sulfate sodium-induced colitis. In the co-culture model, the preparation reduced IL-8 secretion following lipopolysaccharide (LPS) stimulation and increased the expression of epithelial cell-adhesion-related factors. In the mouse model, treatment attenuated body-weight loss, colon shortening, myeloperoxidase activation, and inflammatory cytokine responses in serum and colonic tissues. Increased expression of epithelial adhesion molecules, mucin, and secretory IgA was also observed. These findings support the anti-inflammatory and barrier-protective potential of this specific heat-killed preparation, although its efficacy has not yet been established in patients with IBD (Itoh et al., 2024). For comparison, Wasilewska et al. reported beneficial effects of viable yogurt starter cultures in a DSS-induced colitis mouse model. Because the intervention used active cultures, the study provides comparative probiotic evidence rather than direct evidence for paraprobiotics or postbiotics (Wasilewska et al., 2019).

The strongest clinical observations for a non-viable bacterial preparation in gastrointestinal disease are currently available for IBS. In a multicenter, randomized, double-blind, placebo-controlled trial, 443 patients with IBS received either placebo or non-viable, heat-inactivated Bifidobacterium bifidum MIMBb75 cells once daily for 8 weeks. The composite primary endpoint—defined as at least a 30% improvement in abdominal pain together with adequate relief of overall IBS symptoms during at least four of the eight treatment weeks—was achieved by 34% of patients receiving the heat-inactivated preparation and 19% of those receiving placebo. The risk ratio was 1.7 (95% confidence interval, CI: 1.3–2.4). No serious adverse events occurred in the intervention group, and tolerability was rated as good or very good by 91% of treated patients (Andresen et al., 2020). Microbial-derived antimicrobial compounds may also have potential relevance to Clostridioides difficile infection. Walsh et al. compared thuricin CD, nisin, fidaxomicin, and vancomycin against a panel of bacterial strains and in a mini-fecal fermentation model. Thuricin CD exhibited minimum inhibitory concentrations below 3.1 μg/mL against C. difficile and Bacillus firmus and showed a narrower activity spectrum than nisin, fidaxomicin, and vancomycin. These findings indicate that selected bacteriocins may inhibit C. difficile while exerting more limited effects on other members of the intestinal microbiota. These results are preclinical and do not demonstrate therapeutic efficacy in patients with C. difficile infection (Walsh et al., 2024). Current research supports a clinically relevant, strain-specific effect of heat-inactivated B. bifidum MIMBb75 in IBS. By contrast, evidence for IBD and C. difficile infection remains predominantly experimental. The effects observed for individual strains, heat-killed preparations, active cultures, or purified bacteriocins should therefore not be generalized to all paraprobiotic and postbiotic products. Overall, direct clinical evidence beyond IBS remains limited. Evidence relevant to IBD is derived predominantly from cellular and murine colitis models, and controlled human trials evaluating defined nonviable microbial preparations are still lacking. Similarly, the evidence discussed for C. difficile is based mainly on in vitro or ex vivo antimicrobial studies, including purified bacteriocins, and does not demonstrate treatment efficacy in infected patients. Independent clinical replication is therefore required before these findings can support therapeutic conclusions. Consequently, the evidence for IBD and C. difficile infection should be regarded as mechanistic or preclinical rather than as evidence of established clinical efficacy.

5.1.2 Metabolic disorders

Findings regarding the metabolic effects of paraprobiotic and postbiotic preparations remain limited and heterogeneous. Available studies include a controlled human trial evaluating a multicomponent formulation containing prebiotic fibers and a Bifidobacterium breve-derived EPS preparation in adults with prediabetes and preclinical experiments investigating the effects of a heat-killed bacterial strain on lipid accumulation. Support obtained with viable probiotic strains may provide useful comparative information, but it should not be presented as direct evidence for paraprobiotics or postbiotics.

In a randomized, double-blind, placebo-controlled, parallel-group clinical trial, Beteri et al. assigned 53 adults with prediabetes to receive either a daily dietary fiber formulation or a microcrystalline cellulose placebo for 12 weeks, followed by a 4-week follow-up period. The intervention contained galacto-oligosaccharides, konjac glucomannan, and an EPS preparation derived from Bifidobacterium breve 091109. Compared with placebo, the intervention was associated with increased gut microbial alpha diversity and greater abundance of selected butyrate-producing bacteria. Reductions in glycated hemoglobin and fasting plasma glucose were also reported, with mean values in the intervention group falling below the thresholds used to define prediabetes (Beteri et al., 2024). A randomized clinical trial with viable Limosilactobacillus reuteri NCIMB 30242 reported improvements in several lipid parameters; but because the strain was viable, the study is included only as comparative probiotic evidence (Jones et al., 2012).

Preclinical evidence directly relevant to non-viable microbial preparations was reported by Choi et al. Heat-killed Lactiplantibacillus plantarum Q1 reduced lipid accumulation in cellular models, including differentiated 3T3-L1 adipocytes. In high-fat diet-fed C57BL/6 mice, oral administration of the heat-killed preparation attenuated body weight gain and hepatic lipid accumulation. Molecular analyses indicated that these effects were associated with regulation of adipogenic, lipogenic, and inflammatory pathways, including reduced expression of PPAR-γ (peroxisome proliferator-activated receptor gamma) and tumor necrosis factor alpha (TNF-α) (Choi et al., 2025). These findings demonstrate biological activity in cellular and animal models but do not yet establish efficacy in humans. The available studies suggest that selected nonviable microbial preparations may influence glycemic and lipid-related outcomes. But human evidence is limited to specific multicomponent formulations, whereas evidence for heat-killed strains remains predominantly preclinical. These effects are strain- and formulation-specific and should not be generalized to all paraprobiotic or postbiotic preparations.

5.1.3 Immune modulation and infection prevention

Clinical evidence for immune- or infection-related benefits of non-viable microbial preparations remains limited. Heat-killed preparations can modulate inflammatory mediators in experimental models, but these findings should not be extrapolated to infection prevention or vulnerable populations. Naito et al. reported fewer upper respiratory tract infection symptoms and changes in mononuclear phagocyte markers after viable Lacticaseibacillus paracasei strain Shirota; because the intervention used a live probiotic, it provides comparative context only (Naito et al., 2025). Non-viable preparations may reduce selected viability-dependent risks, but each product still requires preparation-specific compositional, toxicological, and clinical assessment (Churin et al., 2026).

5.1.4 Dermatological applications and the gut–skin axis

The skin microbiome contributes to cutaneous barrier homeostasis, colonization resistance, and local immune regulation. Alterations in its composition and functional activity are associated with inflammatory skin disorders, including atopic dermatitis, acne vulgaris, and psoriasis (Bay and Ring, 2022).

Inanimate microbial preparations may offer practical advantages for dermatological formulations because their biological activity does not depend on maintaining microbial viability during manufacture, storage, or application. In contrast, incorporating live microorganisms into topical formulations may be affected by the formulation composition, preservatives, pH, water activity, oxygen exposure, and storage conditions. Nevertheless, the stability and biological activity of inanimate preparations still depend on their microbial composition, production and inactivation methods, concentration, delivery vehicle, and storage conditions (De Almeida et al., 2023).

Clinical studies are currently limited to specific microbial preparations. In a prospective, randomized, double-blind, vehicle-controlled study, 75 participants with atopic dermatitis received either a topical formulation containing 5% Vitreoscilla filiformis lysate or the vehicle formulation once daily for 30 days. Compared with the vehicle, the lysate-containing formulation produced a significantly greater reduction in SCORAD (SCORing Atopic Dermatitis) scores and pruritus. The treatment was also associated with reduced cutaneous colonization by Staphylococcus aureus. On the other hand, transepidermal water loss improved with the vehicle formulation alone, and the study therefore does not demonstrate a specific effect of the lysate on this measure of skin-barrier function (Gueniche et al., 2008).

In another randomized, double-blind, vehicle-controlled trial, 66 women with reactive skin applied either a cream containing 10% Bifidobacterium longum lysate or a control cream to the face, arms, and legs twice daily for 2 months. By the end of the treatment period, the lysate-containing formulation significantly reduced skin sensitivity and, after 29 days, significantly decreased both clinical and participant-reported skin dryness. It also increased skin resistance to experimentally induced barrier disruption, as indicated by the greater number of tape stripping’s required to reach the predefined transepidermal water loss threshold (Guéniche et al., 2010).

In vitro studies also indicate that the effects of microbial lysates on epidermal barrier function are strain-dependent. Sultana et al. treated primary human epidermal keratinocytes with lysates derived from Bifidobacterium longum, Lactiplantibacillus plantarum, Limosilactobacillus reuteri, Limosilactobacillus fermentum, and Lacticaseibacillus rhamnosus GG. Except for the L. fermentum lysate, which reduced keratinocyte viability, the tested lysates enhanced tight-junction barrier function within 24 h, as assessed by transepithelial electrical resistance. The B. longum and L. rhamnosus GG lysates produced the strongest responses, which were concentration-dependent and persisted for up to 4 days (Sultana et al., 2013).

The potential contribution of orally administered inactivated microorganisms to skin hydration has also been investigated. In a 12-week randomized, placebo-controlled, double-blind trial, 80 healthy adults who were concerned about dry skin received either heat-killed Lactiplantibacillus plantarum L-137 or placebo. In the overall study population, forearm skin water content was significantly higher in the heat-killed L-137 group than in the control group at week 4. Nevertheless, the overall group effect across the study period and the difference in change from baseline did not reach statistical significance. Facial transepidermal water loss showed only a statistical trend toward improvement. These results suggest a possible effect on selected skin moisture and barrier parameters, but they do not establish a consistent clinical benefit across the entire intervention period (Yoshitake et al., 2022).

5.1.5 Attribution of clinical effects in defined and multicomponent interventions

Interpretation of the clinical literature is complicated by substantial differences in intervention composition. Some trials evaluated defined non-viable microbial preparations as the principal active intervention, allowing a relatively direct assessment of preparation-specific effects. By contrast, several studies evaluated fermented foods, infant formulas, or supplements containing prebiotics, human milk oligosaccharides, dietary fibers, or other functional ingredients together with non-viable microbial components. In these multicomponent interventions, the observed clinical outcomes cannot be attributed independently to the microbial-derived fraction. Studies involving viable probiotics provide an additional category of comparative evidence but should not be interpreted as direct evidence for paraprobiotics or postbiotics. Table 3 summarizes these distinctions and the corresponding level of attribution that can reasonably be assigned to representative clinical studies.

TABLE 3. Classification of representative clinical evidence by intervention composition and effect attribution.

“Relatively direct” indicates that the defined non-viable microbial preparation was the principal active intervention; it does not imply that the results are generalizable to other strains or preparations. Multicomponent interventions do not allow independent attribution of efficacy to the microbial-derived component.

5.2 Applications in the food industry

Non-viable microbial preparations are increasingly investigated as functional ingredients in food systems because their intended technological use does not depend on maintaining viable cell counts. This may expand formulation options in matrices or processes that are challenging for live probiotics. In this section, cell-free supernatants, purified bacteriocins, organic acids, and other microbially derived products are also discussed when they provide relevant evidence regarding food preservation, processing, sensory quality, or matrix interactions. However, these products are not classified as postbiotics unless the evaluated preparation contains inanimate microorganisms and/or their components and has been demonstrated to confer a health benefit in accordance with the ISAPP definition. They are therefore referred to by their specific preparation or product class whenever these criteria are not met. Technological performance, sensory compatibility, stability, safety, and biological activity remain preparation- and matrix-specific and must be demonstrated experimentally (Kye et al., 2022).

5.2.1 Technological advantages and processing stability

Ompared with live probiotics, non-viable microbial preparations may offer greater formulation flexibility because their intended technological functionality does not depend on maintaining a defined number of viable cells throughout food processing and storage. This characteristic may facilitate their incorporation into food matrices and processing conditions that are unsuitable for maintaining probiotic viability. By contrast, non-viability does not automatically imply complete thermal, chemical, or storage stability, and the retention of biological activity must be demonstrated for each preparation and final product (Barros et al., 2020; Siciliano et al., 2021; Lee et al., 2023; Pimentel et al., 2023).

Paraprobiotic preparations can be produced using thermal and non-thermal inactivation methods, including conventional heating, high-pressure treatment, ultrasound, irradiation, pulsed electric fields, and other physical or chemical approaches. These treatments differ in their mechanisms of microbial inactivation and may alter cell-envelope integrity, morphology, physiological status, and the accessibility of structural components involved in biological activity. Consequently, the effects of processing are both strain- and method-dependent, and the inactivation conditions must be optimized according to the microorganism, food matrix, and intended functional application (Lee et al., 2023; Pimentel et al., 2023). Inactivation technologies similarly highlight that excessive processing may alter the chemical composition or biological activity of the resulting preparation rather than simply eliminating microbial viability. Experimental data for this variability was provided by Almada et al., who applied six inactivation methods to Lactobacillus acidophilus, Lacticaseibacillus casei, and Bifidobacterium animalis. The treatments produced different effects on cultivability, cell integrity, physiological status, and cellular morphology, demonstrating that the properties of a paraprobiotic preparation cannot be predicted solely from the original probiotic strain or from the general category of inactivation treatment (Almada et al., 2021).

Quality control should therefore include confirmation of microbial inactivation together with characterization of cell integrity, morphology, composition, and relevant biological activity. Depending on the preparation, analytical markers may include cell equivalents, selected cell-wall components, EPS, or microbial metabolites. Despite this, these markers are suitable for dose standardization only when they are reproducibly quantified and shown to reflect the composition or functional activity of the specific product (Barros et al., 2020; Lee et al., 2023; Pimentel et al., 2023). Barros et al. specifically discuss the need for analytical methods and quality-control approaches for paraprobiotic and postbiotic products (Barros et al., 2020). Accordingly, thermal resistance, pH stability, storage conditions, and shelf life should be determined experimentally for each preparation in the intended food matrix. Universal stability limits cannot be assigned to non-viable microbial and microbial-derived ingredients because different microorganisms, cellular components, metabolites, inactivation methods, and food matrices may respond differently during processing and storage. Long-term storage remains substantially less characterized than short-term processing and refrigerated stability. Many food studies reviewed here evaluated products over periods of days to several weeks, while evidence covering the complete intended commercial shelf life remains limited. Non-viability should therefore not be interpreted as evidence of long-term stability. Future studies should combine real-time and, where appropriate, accelerated stability testing with measurement of residual viability; preparation-specific compositional or potency markers; microbiological quality; physicochemical deterioration; and sensory acceptance through the end of shelf life. Packaging material, storage atmosphere, temperature, humidity, and light exposure should also be reported because these factors may affect both the food matrix and the incorporated microbial-derived preparation. Consequently, commercially relevant shelf-life claims should be established for the final product rather than extrapolated from the stability of the isolated ingredient.

5.2.2 Applications in dairy products

Dairy products are among the most frequently investigated matrices for incorporating non-viable microbial and fermentation-derived preparations. Their established fermentation processes and physicochemical properties facilitate the addition of non-viable microbial cells and microbial-derived components, but technological effects remain strain-, preparation-, and process-dependent.

In yogurt, Molaee Parvarei et al. compared formulations containing either viable or inactivated Lactobacillus acidophilus ATCC SD 5221 and Bifidobacterium animalis subsp. lactis BB-12 during 28 days of refrigerated storage. Incorporation of inactivated cells increased the viability of the conventional yogurt starter cultures and resulted in lower syneresis and higher water-holding capacity than the corresponding probiotic formulations, whereas the instrumental color parameters were not significantly affected. The timing of inactivated-cell addition also influenced acidification and organic acid profiles, demonstrating that both the microbial strain and the stage of incorporation affect the technological properties of the final product (Molaee Parvarei et al., 2021a).

Sensory and rheological analyses of yogurts prepared with the same strains further demonstrated formulation-dependent effects. Yogurt containing heat-killed B. animalis subsp. lactis BB-12 added before fermentation achieved the highest overall sensory score among the evaluated formulations, whereas combinations of heat-killed L. acidophilus and B. animalis added before fermentation received lower overall scores. The addition of killed cells also modified rheological parameters, indicating that incorporation of heat-killed microbial cells cannot be assumed to preserve texture and sensory properties independently of strain and processing strategy (Molaee Parvarei et al., 2021b).

More recent evidence was provided by Kaan et al., who evaluated reduced-lactose fermented milk beverages supplemented with heat-inactivated Lacticaseibacillus rhamnosus GG or B. animalis subsp. lactis BB-12 during 30 days of storage at 4 °C. The bacterial suspensions were heat-inactivated at 95 °C for 10 min and incorporated either before or after fermentation. When the heat-inactivated BB-12 preparation was added before fermentation, the EPS concentration reached 1.64 g/kg on day 15. The same formulation showed the highest reported total phenolic content (21.67 mg GAE (gallic acid equivalents)/100 g), radical-scavenging activity (220.95 mg TE (Trolox equivalents)/kg), and antioxidant capacity (650.09 mg TEOC (Trolox equivalent antioxidant capacity)/kg). Probiotic counts remained at or above 8 log CFU/g in all fermented milk beverages. Across the formulations, overall acceptability scores ranged from 4.79 to 7.14, and the presence of heat-inactivated microbial preparations improved texture and overall acceptability scores. These findings provide direct evidence that selected heat-inactivated microbial preparations can modify the technological, functional, and sensory properties of fermented dairy beverages (Kaan et al., 2026).

Yet, beneficial technological effects are not universal. Eminoğlu incorporated heat-killed L. acidophilus at 1 × 10⁹ cells/mL into yogurt and stored the products at 4 °C ± 1 °C for 21 days. In this formulation, the inactivated bacterial cells did not significantly alter the physicochemical, microbiological, antioxidant, or antimicrobial characteristics of the yogurt. In the same study, supplementation with 0.5% autolyzed and dried Saccharomyces cerevisiae affected the physical, chemical, and sensory properties, illustrating the strong dependence of product performance on the type of microbial-derived preparation used (Eminoğlu, 2026).

Clinical indications also exist for heat-treated fermented dairy preparations, although such findings should be distinguished from technological studies. In a randomized, double-blind, placebo-controlled trial, 120 healthy adults with a tendency toward constipation consumed 100 mL/day of either heat-treated Lactobacillus helveticus CP790-fermented milk or a control beverage for 4 weeks. Intake of the heat-treated fermented milk significantly improved stool consistency and straining during defecation and was associated with a reduction in the relative abundance of Desulfobacterota. These results demonstrate that a dairy matrix can serve as a delivery vehicle for a clinically evaluated non-viable microbial preparation (Tanihiro et al., 2024).

These studies show that yogurt and fermented milk are practical matrices for the delivery of non-viable microbial preparations. Nevertheless, technological and functional outcomes depend on the strain, preparation, dose, and process, and improvements in texture, stability, sensory quality, or biological activity should not be generalized across dairy products.

5.2.3 Applications in bakery and cereal products

Bakery products represent an important application area for non-viable microbial preparations and fermentation-derived ingredients because conventional baking markedly reduces probiotic viability. For example, Zhang et al. evaluated Lactiplantibacillus plantarum supplemented into bread and reported reductions of approximately 4–5 log CFU/g during baking at 174, 200, or 235 °C from an initial level of approximately 8.8 log CFU/g in the dough. These findings illustrate the difficulty of maintaining probiotic viability in baked products and provide a technological rationale for exploring non-viable microbial preparations and fermentation-derived components in this food category (Zhang et al., 2018).

Recent experimental results support the feasibility of this approach in sourdough bread. Kwak et al. investigated Limosilactobacillus fermentum SPC L75-1 as a sourdough starter and found that the strain reached 9.64 log CFU/mL after the first sourdough inoculation and remained above 9 log CFU/mL during back-slopping. Sourdough fermented with SPC L75-1 reached a final pH of 3.50, compared with 4.06 for the reference L. fermentum KACC 11441 strain, and produced significantly higher lactic acid concentrations. The use of SPC L75-1 increased gas production by 18.5%, while the resulting bread showed a specific volume comparable to the yeast-fermented control and lower hardness after 3 days of storage at room temperature. The volatile profile was also broadly similar to that of the control bread (Kwak et al., 2026).

The same study separately evaluated the functional properties of heat-inactivated SPC L75-1 cells. Cells treated at 95 °C for 30 min significantly reduced nitric oxide production in LPS-stimulated RAW 264.7 macrophages from 34.07 ± 0.71 to 29.25 ± 2.28 μM and decreased COX-2 expression by 37.5% relative to the LPS-treated group. These results indicate that the strain retained measurable biological activity after controlled thermal inactivation. Despite this, the functional assays were performed using experimentally heat-killed cells rather than extracts obtained directly from the baked bread; therefore, they should not be interpreted as showing that the finished bread retained the same immunomodulatory activity (Kwak et al., 2026). This distinction is important because biological activity measured in an experimentally prepared microbial fraction cannot be assumed to persist after incorporation into, and processing within, a complex food matrix. Processing can alter microbial cell structures and other effector molecules, while interactions with proteins, polysaccharides, lipids, pH, water activity, and thermal history can further modify their accessibility and functional expression. Accordingly, biological activity should ideally be re-evaluated in, or recovered from, the final food product under the intended processing and storage conditions (Zhao et al., 2026).

Non-viable bacterial preparations may also affect microorganisms used in bakery fermentation. Pihurov et al. investigated non-viable Lacticaseibacillus paracasei MIUG BL80 cells produced by combined ultrasound and thermal treatment in a Saccharomyces cerevisiae MIUG D129 fermentation model. At 15% supplementation with the non-viable bacterial preparation, the specific growth rate of S. cerevisiae increased from 0.227 in the control to 0.507, while the generation time decreased from 4.403 to 1.972 h. The preparation also improved the stability of wet yeast biomass during storage at 4 °C. Although this study was not conducted in a finished bread product, it provides direct indications that non-viable bacterial preparations can influence the technological performance of yeast starter cultures relevant to bakery fermentation (Pihurov et al., 2025).

Fermentation-derived microbial metabolites may also contribute to bread preservation. Black et al. demonstrated that Levilactobacillus hammesii converted linoleic acid during sourdough fermentation into an antifungal monohydroxy C18:1 fatty acid, which accumulated at 0.73 ± 0.03 g/L. The antifungal monohydroxy fatty acid fraction showed minimum inhibitory concentrations of 0.1–0.7 g/L. Incorporation of 20% sourdough fermented with L. hammesii extended the mold-free shelf life of bread by 2–3 days, whereas addition of 0.15% coriolic acid extended mold-free storage from 2 days to more than 6 days (Black et al., 2013).

These studies support the technological potential of nonviable microbial preparations and fermentation-derived components in bakery applications, particularly in sourdough fermentation, starter performance, texture, and microbial stability. Nevertheless, evidence that specific microbial-derived components retain biological activity after conventional baking remains limited. Available reviews emphasize that the thermal stability of EPS, cell-surface components, SCFAs, biosurfactants, and other microbially derived molecules cannot yet be generalized across baking conditions. Product-specific characterization before and after thermal processing is therefore required (Pérez-Alvarado et al., 2022).

5.2.4 Applications in meat and seafood products

Meat and seafood products have been investigated as suitable matrices for the application of non-viable microbial preparations and fermentation-derived antimicrobial compounds, particularly as biopreservatives. Cell-free supernatants produced by lactic acid bacteria (LAB) contain organic acids and proteinaceous antimicrobial compounds whose activity depends on the producing strain, concentration, food matrix, and storage conditions. Therefore, their preservative effects should be evaluated directly in the target food rather than extrapolated from in vitro antimicrobial assays.

Wang et al. reported direct results in beef products, evaluating the cell-free supernatant of Lactiplantibacillus plantarum 90 in ground beef gel. Addition of 0.5% cell-free supernatant inhibited microbial growth during refrigerated storage and extended product shelf life without significantly affecting pH, cooking loss, instrumental color, or texture. Higher concentrations were not technologically equivalent: additions above 0.5% progressively altered several texture parameters, including hardness, springiness, adhesiveness, and chewiness. Mechanistic experiments showed that the antibacterial activity resulted from the combined effects of organic acids and proteinaceous antimicrobial substances that disrupted bacterial cell structures (Wang et al., 2023).

Comparable food-preservation effects have been reported with LAB-derived cell-free preparations in active beef packaging, poultry, and seafood systems. Whey-protein films containing Lactobacillus sakei cell-free supernatant reduced E. coli and Listeria monocytogenes on fresh beef (del Carmen Beristain-Bauza et al., 2017), while preparations derived from Pediococcus acidilactici, Lactiplantibacillus plantarum, and Latilactobacillus sakei reduced pathogen growth and extended refrigerated shelf life in chicken products (İncili et al., 2022; Serter et al., 2024). LAB cell-free supernatants also delayed microbial and biochemical spoilage of ribbonfish (Jo et al., 2021), whereas nisin extended the microbiological shelf life of vacuum-packaged rainbow trout (Behnam et al., 2015). These studies support food-preservation applications but do not imply that isolated bacteriocins or cell-free products independently meet the ISAPP definition of a postbiotic.

All these studies demonstrate that non-viable microbial preparations, cell-free fermentation fractions, and selected microbial antimicrobial compounds can reduce microbial growth and extend the storage stability of meat and seafood products. On the other hand, their effects are strongly dependent on the producing microorganism, preparation method, concentration, delivery system, food matrix, and storage conditions. Moreover, isolated bacteriocins and cell-free metabolites should not automatically be classified as postbiotics under the strict ISAPP definition unless the preparation itself has been demonstrated to confer a health benefit on the host.

5.2.5 Applications in infant nutrition and specialized foods

Infant formula represents one of the most extensively investigated specialized nutritional matrices for the incorporation of fermentation-derived, non-viable microbial preparations. Fermentation during formula manufacture can generate microbial metabolites and cellular components, while the microorganisms used in the process may subsequently be inactivated. By contrast, the resulting formulations are often multicomponent products containing prebiotics, human milk oligosaccharides, or other functional ingredients; consequently, clinical outcomes cannot always be attributed specifically to the fermentation-derived microbial fraction.

Clinical studies of fermented infant formulas generally evaluate multicomponent products rather than isolated non-viable microbial preparations. Rodriguez-Herrera et al. studied a partially fermented formula containing fermentation-derived components together with scGOS/lcFOS in 200 healthy term infants. Growth was comparable with the control formula, while softer stools and a lower frequency of investigator-reported colic were observed (Rodriguez-Herrera et al., 2019). Vandenplas et al. subsequently evaluated a formula containing fermented components, scGOS/lcFOS, 2′-FL, and milk fat in 215 healthy term infants and confirmed equivalent growth and good gastrointestinal tolerance (Vandenplas et al., 2020). Because both interventions contained several functional ingredients, their outcomes cannot be attributed independently to the fermentation-derived microbial fraction; moreover, 2′-FL is a human milk oligosaccharide and not a postbiotic.

Effects on mucosal immune and microbiota-related parameters have also been investigated. Béghin et al. randomized 280 healthy term infants to formulas containing fermentation-derived bioactive compounds produced by B. breve C50 and S. thermophilus 065, scGOS/lcFOS, their combination, or standard formula, while 70 breastfed infants were included as a reference group. At 4 months, fecal secretory IgA was significantly higher in infants receiving the combination of fermented formula components and scGOS/lcFOS than in the control group. The combination was also associated with microbiota composition and metabolic activity more closely resembling those of breastfed infants. Growth was normal, and the tested formulas were well tolerated (Béghin et al., 2021). Again, these findings support the complete formulation rather than demonstrating an isolated effect of the fermentation-derived components.

Recent studies are particularly relevant for preterm nutrition because non-viable preparations can be incorporated into formulas without administering replicating microorganisms. Campeotto et al. studied 58 preterm infants with a gestational age of 30–35 weeks who received either a preterm formula fermented with B. breve C50 and S. thermophilus 065 and heat-inactivated at the end of manufacture or a standard preterm formula. No significant differences in growth or overall digestive tolerance were observed. During weeks 3–4, abdominal distension was not reported in the fermented-formula group, compared with eight cases in the control group. Fecal calprotectin decreased significantly over time in the fermented-formula group, whereas the intervention did not significantly alter overall bacterial colonization. No adverse effects were reported during the study (Campeotto et al., 2011). These results demonstrate the feasibility of using a heat-inactivated fermented preparation in preterm formula, but the small sample size precludes broad conclusions regarding clinical protection in premature infants.

Longer-term clinical data have also been reported. In a randomized, double-blind, placebo-controlled trial involving 971 healthy infants aged 4–6 months, consumption for 5 months of a formula fermented with B. breve C50 and S. thermophilus 065 did not significantly reduce the incidence or duration of acute diarrhea or the number of hospital admissions. Conversely, diarrhea episodes were less severe: dehydration occurred in 2.5% of infants receiving the fermented formula compared with 6.1% receiving standard formula; medical consultations were required in 46.0% versus 56.6%; oral rehydration solutions were prescribed in 41.9% versus 51.9%; and formula changes occurred in 59.5% versus 74.9% (Thibault et al., 2004). Thus, the study supports a possible reduction in diarrhea severity rather than prevention of diarrhea itself.

Beyond infant nutrition, the use of postbiotic preparations in specialized medical nutrition remains an emerging field. In the 2025 GALA-POSTBIO phase 2 trial, 56 patients with advanced compensated alcohol-related liver disease were randomized to receive either ReFerm®, an oat-based preparation fermented with Lactobacillus plantarum DSM 9843, or standard nutritional support for 24 weeks. The primary endpoint—a ≥10% reduction in hepatic α-smooth muscle actin expression—was achieved by 29% of patients receiving ReFerm®and 14% receiving standard nutritional support, but the difference was not statistically significant (OR 2.40, 95% CI 0.63–9.16). No treatment-related serious adverse events occurred, while several secondary markers of liver fibrogenesis and intestinal barrier function showed favorable changes (Hansen et al., 2025). These findings illustrate the potential of fermented non-viable microbial preparations in specialized medical nutrition but also emphasize that clinical efficacy remains to be established in larger confirmatory trials.

5.2.6 Sensory properties and consumer acceptance

Sensory compatibility is an important determinant of the successful incorporation of non-viable microbial and fermentation-derived preparations into foods. Available studies indicate that their effects on appearance, flavor, texture, and overall acceptability are highly dependent on the microbial source, preparation method, concentration, food matrix, and storage conditions rather than being uniformly neutral or beneficial.

Direct sensory evidence is available from yogurts supplemented with microbial preparations derived from fermentation. Sadighbathi et al. evaluated low-fat yogurts fortified with 3% freeze-dried fermentation-derived microbial preparations produced by Streptococcus thermophilus or Lactobacillus delbrueckii subsp. bulgaricus in either cheese whey or skim milk. Fifteen semi-trained panelists assessed appearance, flavor, mouthfeel, body and texture, and overall acceptability on a 10-point hedonic scale after 11 days of refrigerated storage. Yogurt containing the L. delbrueckii subsp. bulgaricus-derived preparation produced in cheese whey achieved an overall acceptability score of 8.73 ± 1.49, compared with 7.40 ± 1.66 for the control and 6.66 ± 1.49 for yogurt containing the S. thermophilus-derived preparation produced in skim milk. Body and texture scores were 8.60 ± 1.56 and 6.06 ± 1.56, respectively, for these two fermentation-derived formulations, demonstrating that sensory performance depended markedly on the producing microorganism and production substrate (Sadighbathi et al., 2023).

More recent consumer data confirm this formulation-dependent response. Kaan et al. evaluated reduced-lactose fermented milk beverages containing heat-inactivated Lacticaseibacillus rhamnosus GG or Bifidobacterium animalis subsp. lactis BB-12 using 80 untrained consumers. Appearance, taste, texture, and overall acceptability were evaluated on a nine-point hedonic scale during 30 days of refrigerated storage. Across formulations, appearance scores ranged from 5.07 to 7.57, taste from 5.07 to 6.86, texture from 4.71 to 7.50, and overall acceptability from 4.79 to 7.14. Products prepared with B. animalis subsp. lactis BB-12 generally received higher sensory scores than those prepared with L. rhamnosus GG, and the presence of paraprobiotics significantly improved texture and overall acceptability irrespective of whether they were incorporated before or after fermentation (Kaan et al., 2026).

Nonetheless, sensory effects are not necessarily positive. Eminoğlu compared yogurts fortified with heat-killed Lactobacillus acidophilus at 1 × 10⁹ cells/mL with yogurts containing 0.5% autolyzed and dried Saccharomyces cerevisiae. Sensory evaluations were performed on days 1, 11, and 21 of storage. Heat-killed L. acidophilus did not significantly alter sensory characteristics relative to the corresponding control, whereas the S. cerevisiae-containing products received significantly lower taste scores on days 1 and 11 because panelists detected distinct “umami” or “bouillon-like” notes. Nevertheless, taste scores remained above 6 on the nine-point scale, indicating acceptable products, and the differences were no longer evident by day 21. These findings demonstrate that microbial-derived preparations may introduce characteristic flavor compounds and that their sensory impact can change during storage (Eminoğlu, 2026). Similar concentration-dependent effects have been demonstrated in meat products. Ozma et al. evaluated lamb slices coated with Malva sylvestris mucilage containing different concentrations of a cell-free preparation derived from Lacticaseibacillus paracasei ATCC 55544. Twenty-five trained panelists assessed odor, color, appearance, and overall acceptance on a nine-point hedonic scale during 10 days at 4 °C. On day 10, overall acceptance decreased to 2.47 ± 0.39 in the untreated control but remained at 8.57 ± 0.45 in samples treated with the highest concentration of the cell-free microbial preparation. Corresponding odor scores were 2.60 ± 0.42 versus 7.94 ± 0.59, while appearance scores were 3.95 ± 0.39 versus 6.76 ± 0.38. The authors reported no significant deterioration in color, appearance, or overall acceptance of the highest-concentration treatment during refrigerated storage (Ozma et al., 2022).

These studies demonstrate that incorporating non-viable microbial and fermentation-derived preparations can maintain or, in selected formulations, improve sensory acceptability, although such effects cannot be generalized across products. Sensory outcomes are strongly dependent on the microbial source, concentration, carrier matrix, processing conditions, and storage time. Consequently, no universal effective dose can currently be proposed from sensory studies, and each food formulation requires product-specific sensory and consumer evaluation before commercial application.

5.2.7 Cross-matrix comparison and industrial trade-offs

Comparison across food matrices reveals distinct technological opportunities and limitations. Dairy products offer relatively mild processing conditions and established fermentation infrastructure, but reported technological and sensory responses depend strongly on strain, inactivation procedure, and timing of incorporation, while most available storage studies remain relatively short. Bakery products provide a strong technological rationale for non-viable preparations because conventional baking markedly compromises probiotic viability; however, evidence that specific microbial-derived components retain biological activity after baking remains limited, and functional assays are not always performed on the finished food product. In meat and seafood systems, microbial-derived preparations can improve microbiological stability, but antimicrobial efficacy is concentration- and matrix-dependent, and higher concentrations may adversely affect texture. Moreover, many studies use cell-free supernatants or isolated antimicrobial products that do not independently meet the strict ISAPP postbiotic definition. Infant formulas provide comparatively extensive clinical evidence, but most interventions are multicomponent formulations, preventing attribution of outcomes to the non-viable microbial fraction alone. Therefore, no single food matrix can currently be considered universally optimal for incorporating non-viable microbial or fermentation-derived preparations. The principal industrial trade-off is balancing biological activity with process tolerance, sensory compatibility, compositional reproducibility, storage stability, regulatory requirements, and the strength of preparation-specific evidence.

5.2.8 Market trends and commercial products

Commercial interest in ingredients marketed as postbiotics is increasing, although market-size estimates vary because definitions and product categories differ among commercial analyses. Accordingly, such figures should be interpreted as industry estimates rather than scientific measures of adoption. Peer-reviewed literature documents several commercially available non-viable microbial preparations, including EpiCor®, HT-BPL1, LAC-Shield®, and heat-killed L-137 (Liang and Xing, 2023). Manufacturer data further indicate broad incorporation of LAC-Shield® into finished products, but commercial availability should not be equated with consumer familiarity. A 2026 survey of 346 university students in Türkiye found low awareness and limited supplement use, illustrating that consumer knowledge may lag behind product development (Kahve and Yıldız, 2026). Overall, market expansion reinforces the need for precise terminology, transparent product characterization, and evidence-based claims. The principal clinical and food-industry applications are summarized in Figure 3.

6 Discussion: challenges and translational priorities

Translation of paraprobiotics and postbiotics is constrained by inconsistent terminology, heterogeneous manufacturing and characterization, limited preparation-specific clinical evidence, and fragmented regulatory pathways. These issues complicate comparison across studies and support a product-specific rather than class-wide approach to efficacy, safety, and technological performance. A central conceptual challenge is the inconsistent use of the term “postbiotic” across the literature. Earlier and some contemporary studies apply the term to isolated metabolites, cell-free supernatants, bacteriocins, or extracellular vesicles, whereas the current ISAPP definition requires a preparation containing inanimate microorganisms and/or their components together with a demonstrated health benefit. This discrepancy complicates comparison among studies, regulatory classification, and interpretation of efficacy. Future research should therefore report the composition and processing history of each preparation rather than relying on broad biotic labels alone.

6.1 Standardization and regulatory challenges

The regulatory landscape for paraprobiotics and postbiotics remains fragmented because these preparations do not currently constitute a dedicated and harmonized regulatory category across major jurisdictions. The 2021 ISAPP consensus defined a postbiotic as a “preparation of inanimate microorganisms and/or their components that confers a health benefit on the host”, requiring a demonstrated health benefit rather than microbial origin alone. Nevertheless, this scientific definition has not been translated into a single internationally harmonized regulatory pathway, and commercial products are generally assessed according to their composition, intended use, safety, and claims under existing food, food-supplement, or medicinal-product legislation (Salminen et al., 2021; Guglielmetti et al., 2025).

In the European Union, nutrition and health claims made on foods are governed by Regulation (EC) No 1924/2006. The European Food Safety Authority (EFSA) evaluates the scientific substantiation of proposed health claims, whereas authorization decisions are taken within the European Union regulatory framework and authorized and non-authorized claims are recorded in the EU Register of Health Claims. Current EFSA guidance requires adequate characterization of the food or constituent, demonstration that the claimed effect is relevant to human health, human efficacy studies supporting the claim, and reports of biological plausibility. Therefore, a preparation described commercially as a postbiotic does not obtain a health claim solely based on being non-viable or fermentation-derived (European Parliament and Council, 2006; EFSA, 2026).

In the United States, there is likewise no dedicated regulatory category for postbiotics. A microbial-derived ingredient intended for conventional food use must comply with the applicable food-ingredient framework; substances intentionally added to food generally require food-additive authorization unless their intended use is generally recognized as safe (GRAS). For dietary supplements containing a new dietary ingredient, a premarket New Dietary Ingredient Notification may be required at least 75 days before marketing. Dietary supplements may bear substantiated structure/function claims, but the manufacturer must notify the FDA (U.S. Food and Drug Administration, 2026a) within 30 days after first marketing the product with the claim, include the statutory disclaimer, and avoid claims that diagnose, treat, cure, or prevent disease (FDA, 2026a, FDA, 2026b).

Regulatory pathways also differ in Asia. In Japan, Foods for Specified Health Uses (FOSHU) require government evaluation of product safety and efficacy before authorization, whereas Foods with Function Claims (FFC) may carry reports -based function claims under the responsibility of the food business operator after the required safety and efficacy information has been submitted to the Consumer Affairs Agency before marketing. Unlike FOSHU, FFC products are not individually pre-approved for efficacy by the government. In China, safety review of new food ingredients falls under the National Health Commission, whereas health-food registration and filing are administered through the State Administration for Market Regulation. (Consumer Affairs Agency of Japan, 2026; National Health Commission of China; SAMR-State Administration for Market Regulation).

Standardization remains equally important at the scientific and manufacturing levels. A 2026 review identified 101 published clinical studies involving preparations described as postbiotics; 69 studies used heat treatment, but reported inactivation temperatures ranged from 70 °C to 140 °C and treatment durations from approximately 30 s to 10 h, while several studies provided incomplete information on the inactivation procedure. The authors therefore recommended standardized reporting of microbial taxonomy, inactivation conditions, administered quantity, product characterization, and quality control, including verification of residual viable cells and definition of an acceptable viability limit (Anjum and Ouwehand, 2026).

This meaningful standardization should include identification of the source microorganism to an appropriate taxonomic level, documentation and validation of the inactivation procedure, verification of residual viability, quantitative characterization of relevant cellular or metabolic markers, assessment of biological activity, contaminant and safety testing, and stability studies under the intended processing and storage conditions. These requirements are necessary to improve reproducibility, batch-to-batch consistency, comparison among studies, and regulatory substantiation of product-specific health claims (Guglielmetti et al., 2025; Anjum and Ouwehand, 2026).

The principal regulatory gap is therefore not the absence of food regulation, but the absence of a dedicated, internationally harmonized category for paraprobiotic and postbiotic preparations. This gap creates practical uncertainty regarding terminology, minimum compositional characterization, acceptable evidence of efficacy, dose expression, stability requirements, and the claims that may accompany a commercial product. A harmonized framework should establish, at minimum, requirements for taxonomic identification of the source microorganism, validated inactivation and residual-viability testing, characterization of relevant cellular and molecular components, preparation-specific potency markers, contaminant and safety assessment, storage stability, intended use, and evidence supporting any health claims. Until such harmonization is achieved, regulatory assessment will remain jurisdiction- and product-specific, and manufacturers should avoid using the term “postbiotic” as a substitute for the compositional and efficacy evidence required under the applicable regulatory framework.

6.2 Production methods and biological activity

The production method is a major determinant of the physicochemical and biological characteristics of non-viable microbial preparations. Available inactivation approaches include conventional heating, ultrasound, high hydrostatic pressure, irradiation, changes in pH, supercritical carbon dioxide, and other physical or chemical treatments. Despite this, no single method can be considered universally superior because microbial species and strains respond differently to processing conditions.

Direct observations for this strain-dependent response were provided by Almada et al., who compared six inactivation approaches—heat, ultrasound, low pH, high pH, gamma irradiation, and supercritical CO₂ for Lactobacillus acidophilus, Lacticaseibacillus casei, and Bifidobacterium animalis. Flow cytometry and electron microscopy demonstrated significant differences in membrane integrity, enzymatic activity, and cellular morphology according to both strain and treatment. Following ultrasound treatment, L. casei retained 86.2% enzymatically active cells, while only 4.6% showed membrane damage. In contrast, ultrasound markedly altered the morphology of L. acidophilus and B. animalis and compromised enzymatic activity in more than 80% of their cells. These results demonstrate that the same inactivation technology can produce substantially different preparations depending on the microbial strain (Almada et al., 2021).

Processing intensity can also modify biological activity after successful microbial inactivation. Sun et al. exposed Lacticaseibacillus paracasei ET-22 and Bifidobacterium animalis subsp. lactis BL-99 to 70, 80, 90, 100, or 121 °C for 10 min. All treatments effectively eliminated cultivability, but increasing temperature progressively disrupted cell morphology and promoted leakage of intracellular material. Antioxidant and anti-inflammatory activities were largely preserved after treatments below 100 °C, whereas treatment at 121 °C significantly reduced antioxidant activity and substantially altered the chemical profiles of the resulting preparations. For BL-99, for example, DPPH (2,2-diphenyl-1-picrylhydrazyl) radical-scavenging activity decreased to 23.45% after treatment at 121 °C. Metabolomic analysis further showed that several compounds abundant after treatment at 80 or 100 °C were strongly reduced or no longer detected following heating at 121 °C (Sun et al., 2023).

Biological effects are also strongly strain-dependent even when the same general inactivation strategy is used. Noh et al. compared heat-killed Lactiplantibacillus plantarum KC3, L. plantarum CKDB008, and Limosilactobacillus fermentum SRK414 in RAW264.7 macrophages. Heat-killed L. plantarum CKDB008 produced the strongest response and increased macrophage phagocytic activity to 119% of the zymosan-treated control at 1 × 10⁸ cell equivalents/mL. The same preparation increased NO and cytokine production in unstimulated macrophages but suppressed NO, TNF-α, IL-6, and IL-1β (interleukin-1 beta) production in LPS-stimulated cells, illustrating that biological responses depend not only on strain and preparation but also on the physiological context of the target cells (Noh et al., 2022). The available evidence does not support ranking any single inactivation method as universally superior for preserving biological activity. Instead, functional retention appears to depend on the microbial strain, processing intensity, target molecular structure, and biological endpoint. Mild or intermediate processing may preserve selected cell-surface structures or metabolites in some preparations, whereas more severe treatment can disrupt membranes, denature proteins, promote intracellular leakage, or modify metabolite profiles. For example, Sun et al. (2023) reported that antioxidant and anti-inflammatory activities of heat-treated preparations were largely retained below 100 °C, whereas treatment at 121 °C markedly reduced antioxidant activity and altered chemical profiles. Almada et al. (2021) demonstrated that even the same inactivation method can have markedly different effects on membrane integrity and enzymatic activity among microbial strains. Noh et al. (2022) showed that immunomodulatory responses differed among heat-killed strains and according to the inflammatory state of the target cells. Therefore, preservation of antioxidant, immunomodulatory, antimicrobial, or barrier-related activity should be demonstrated experimentally for each strain–process–matrix combination rather than inferred from the inactivation technology alone. Recent evidence further emphasizes that food processing and matrix interactions can alter the integrity, bioaccessibility, and functional expression of microbially derived effectors (Zhao et al., 2026). Because the biological consequences of inactivation depend on the microbial strain, processing intensity, and functional endpoint, Table 4 summarizes representative evidence comparing different inactivation approaches and their effects on structural and biological properties.

TABLE 4. Representative evidence on the relationship between microbial inactivation conditions and retained bioactivity.

6.3 Industrial scalability and economic considerations

Direct head-to-head techno-economic studies comparing live probiotics with non-viable microbial preparations remain scarce; consequently, a universal cost advantage cannot yet be assigned to non-viable microbial products. Manufacturing live probiotics may require viability-preserving downstream processing, stabilization, formulation overages, and controlled storage or distribution, depending on the strain and product. Non-viable preparations may reduce some of these viability-related constraints but introduce other cost drivers, including validated inactivation, confirmation of residual non-viability, compositional characterization, potency testing, and, for some products, downstream separation or purification. Process economics therefore depend on the fermentation substrate, microbial strain, production scale, inactivation technology, degree of purification, target dose, packaging, quality-control requirements, and regulatory pathway rather than on biotic classification alone.

Scale-up studies nevertheless demonstrate increasing technical feasibility. Kongsinkaew et al. (2024) transferred paraprobiotic soy milk production from flask culture to a stirred-tank bioreactor using fermentation followed by in situ thermal inactivation, illustrating the potential to integrate biomass production and inactivation within a scalable process. A separate 3 L bioreactor study by Hernández et al. (2025) reported a preliminary approximate 25% reduction in production cost per milligram of a protein-rich, microbially derived preparation relative to flask cultivation, attributed mainly to reduced manual labor, increased batch volume, and lower contamination risk. However, this estimate is product- and process-specific and does not constitute a general probiotic-versus-postbiotic cost comparison. Industrial translation will additionally require control of heat and mass transfer, oxygen availability, batch-to-batch consistency, bioactive yield, downstream recovery, and end-of-shelf-life potency. Formal techno-economic analyses at pilot and industrial scale therefore remain an important research priority.

6.4 Clinical validation and evidence gaps

Clinical translation remains one of the principal challenges in postbiotic research. According to the ISAPP consensus, a postbiotic is a preparation of inanimate microorganisms and/or their components that confers a demonstrated health benefit on the host; therefore, microbial metabolites or inactivated cells cannot be classified as postbiotics solely based on their origin or absence of viability (Salminen et al., 2021).

The clinical evidence base is expanding but remains heterogeneous. A comprehensive review published in 2026 identified 101 clinical studies involving preparations described as postbiotics. Most evaluated preparations derived from single microbial strains, whereas only 12 studies evaluated mixed-strain preparations or heat-treated yogurt. Heat treatment was the most frequently reported inactivation approach, being used in 69 studies, but substantial differences in processing conditions, dose reporting, taxonomy, and quality-control procedures limited direct comparisons among trials (Anjum and Ouwehand, 2026).

A search of the World Health Organization (WHO) International Clinical Trials Registry Platform performed up to 12 February 2026 identified 76 registered postbiotic trials. The largest groups targeted metabolic outcomes (18 registrations), oral health (13), digestive health (12), cognitive outcomes (8), and immune outcomes (7), whereas only two registrations listed safety as the primary outcome. These data demonstrate substantial diversification of clinical research but also underline the need for dedicated safety assessment and standardized trial design (Anjum and Ouwehand, 2026).

Even in indications with multiple randomized trials, the certainty of results may remain limited. A 2026 systematic review and meta-analysis of postbiotic interventions for IBS identified only four randomized controlled trials involving 1,062 participants. Postbiotic treatment was associated with improvements in IBS symptom severity and abdominal pain compared with placebo, while treatment-emergent adverse-event rates did not differ significantly between groups. On the other hand, heterogeneity was moderate to substantial, and the overall certainty of documentation was rated as low, leading the authors to call for larger and more standardized randomized (Ayad et al., 2026).

Major limitations therefore include heterogeneity in microbial strains and preparations, inconsistent definitions and characterization, variable inactivation technologies, non-standardized dose units, short intervention periods, limited long-term follow-up, and insufficient evidence in vulnerable populations. Differences in target populations and baseline clinical or microbiome characteristics further complicate comparisons between studies. Future clinical trials should use clearly characterized and reproducibly manufactured preparations, prespecified dose-response designs, appropriate placebo controls, clinically relevant primary outcomes, and systematic safety monitoring. Importantly, efficacy and safety conclusions should remain preparation- and indication-specific rather than being generalized to paraprobiotics or postbiotics as a whole (Anjum and Ouwehand, 2026; Guglielmetti et al., 2025).

7 Future research directions

Future development will depend on molecular characterization, controlled biomanufacturing, targeted delivery, standardized vesicle purification, and biomarker-guided clinical validation (Figure 4). Future research on paraprobiotics and postbiotics should move from broad functional claims toward the molecular characterization of defined preparations, reproducible manufacturing, targeted delivery, and biomarker-supported clinical validation. Integration of genomics, proteomics, metabolomics, and computational approaches is particularly relevant because the biological activity of these preparations depends on complex and strain-specific combinations of cellular structures, proteins, metabolites, and extracellular vesicles. Multi-omics approaches provide a basis for defining the molecular fingerprint of individual microbial preparations. Multi-omics approaches can define preparation-specific molecular fingerprints. Odorskaya et al. combined genomic, metabolomic, and proteomic analyses of Limosilactobacillus fermentum U-21 and identified distinct metabolite and protein profiles in culture fluid and extracellular vesicles, together with genes associated with antioxidant potential. This integrated approach illustrates how candidate functional components can be linked to a defined microbial preparation rather than inferred from species identity alone (Odorskaya et al., 2024).

The need for strain-specific molecular characterization is further supported by a recent untargeted metabolomics comparison of cell-free preparations obtained from 14 Bifidobacterium and Lactobacillus strains. Liu et al. identified 3,333 metabolites, of which 1,262 were shared across all preparations; principal-component analysis nevertheless revealed clear differences between bacterial groups and individual strains. These findings demonstrate that preparations derived from different microorganisms may have substantially different chemical compositions even when they are grouped under the same general biotic terminology (Liu et al., 2026).

Targeted delivery may be particularly useful for microbial-derived metabolites with unfavorable organoleptic or pharmacokinetic properties. Carvalho et al. developed PLGA nano-in-microparticles for colonic delivery of butyrate and propionate and demonstrated reduced premature release under acidic conditions, supporting controlled intestinal delivery (Carvalho et al., 2025). These purified SCFAs should nevertheless be considered microbial-derived metabolites rather than postbiotics under the strict ISAPP definition.

Metabolic and genome engineering may improve the controlled production of defined microbial metabolites and components, although this strategy should be regarded as an upstream biomanufacturing approach rather than evidence of postbiotic efficacy. Bai and Mansell engineered Escherichia coli Nissle 1917 to produce butyrate from chromosomally integrated pathways, achieving concentrations of approximately 1 g/L, corresponding to 11 mM, and developed a fluorescent biosensor for real-time monitoring of intracellular butyrate production (Bai and Mansell, 2020). More broadly, Wilding-Steele et al. developed a CRISPR/Cas9 (clustered regularly interspaced short palindromic repeats/CRISPR-associated protein 9) platform for Clostridium acetobutylicum that achieved 100% editing efficiency for upp and ldhA deletion and 93% efficiency for deletion of the ptb-buk operon. These studies demonstrate the technical feasibility of precise strain engineering, but they do not establish that genetically engineered microorganisms automatically generate standardized or clinically effective postbiotic preparations (Wilding-Steele et al., 2021).

Extracellular vesicle biomanufacturing is another emerging area requiring improved purification and quality control. Pirolli et al. combined tangential-flow filtration with high-performance anion-exchange chromatography to purify bEVs produced by probiotic Escherichia coli and LAB. Size-based filtration alone co-isolated protein contaminants, whereas the combined size- and charge-based procedure substantially improved vesicle purity and enhanced the anti-inflammatory activity of the purified lactic-acid-bacterial vesicle preparations. Because both technologies are scalable, this approach provides a practical model for future manufacturing of defined bacterial extracellular-vesicle products (Pirolli et al., 2023).

Integration of clinical phenotyping with microbiome, metabolomic, and machine-learning analyses may help identify responders and mechanistic biomarkers. A 2026 randomized, double-blind, placebo-controlled trial involving 149 patients with moderate-to-severe cancer-related pain evaluated JK5G, a preparation containing inactivated LAB and microbial-derived components, as an adjunct to standardized analgesic treatment. The investigators integrated microbiome and metabolomic data and compared support-vector machine, random-forest, LASSO (least absolute shrinkage and selection operator), and logistic-regression models. In the independent test subsets, the corresponding microbiome-based models showed predictive performance across several machine-learning approaches, with LASSO providing the highest reported area under the curve. Five microbial features were consistently selected across all four approaches: Akkermansia muciniphila, Bifidobacterium, Escherichia-Shigella, Blautia, and Streptococcus. Although these biomarkers require external validation and the trial was conducted at a single center, the study provides a concrete example of how multi-omics and machine learning can be incorporated into clinical postbiotic research (Chen et al., 2026).

Future studies should therefore prioritize product-specific molecular fingerprints, validated potency markers, dose–response relationships, scalable purification methods, and prospective biomarker validation. Organoid, organ-on-chip, and gnotobiotic models may help establish causal links between individual microbial components and host responses before clinical translation. In parallel, machine-learning models should be trained and externally validated using sufficiently large, independent cohorts rather than being used only for exploratory classification within individual studies. These developments should ultimately enable a transition from broadly defined paraprobiotic and postbiotic preparations toward compositionally defined and indication-specific products. Non-viable microbial preparations should not be regarded as functionally equivalent substitutes for viable probiotics. Although some mechanisms may overlap, loss of viability eliminates microbial growth and ongoing metabolic activity and may alter the abundance, accessibility, or structure of biologically active components. Inactivation can preserve or expose selected cell-associated molecules and may provide technological advantages when maintaining viability is undesirable or impractical. Efficacy demonstrated for a viable probiotic strain cannot be assumed to persist after inactivation, and effects observed for an inactivated preparation cannot be extrapolated to its viable counterpart. Probiotics, paraprobiotics, and postbiotics should therefore be viewed as complementary, preparation-specific strategies whose suitability depends on the intended biological effect, target population, food matrix, processing conditions, and available evidence.

8 Limitations of this review

This narrative review is limited by substantial heterogeneity in terminology, microbial strains, inactivation procedures, preparation composition, dose reporting, models, populations, and outcomes. Some studies predate the current ISAPP definition, while others evaluated viable probiotic comparators or multicomponent formulations that prevent attribution of effects to a single non-viable component. The limited number of preparation-specific randomized trials and potential publication bias further restrict broad conclusions. Evidence should therefore be interpreted at the level of the specific preparation, indication, and experimental context.

9 Conclusion

Paraprobiotics and postbiotics have broadened the concept of microbiome-related functional ingredients by demonstrating that selected biological effects associated with microorganisms do not necessarily require microbial viability. Non-viable microbial cells and their structural components can interact with host pattern-recognition receptors and modulate inflammatory and immune-signaling pathways, while microbial-derived metabolites such as SCFAs contribute to epithelial barrier regulation and inter-organ communication. Conversely, these mechanisms are highly dependent on the microbial strain, preparation method, molecular composition, dose, and experimental context, and should not be generalized to all paraprobiotic or postbiotic preparations.

Clinical translation remains uneven. The strongest evidence currently supports specific, well-characterized preparations in selected indications, such as heat-inactivated Bifidobacterium bifidum MIMBb75 in IBS. By contrast, evidence for IBD, metabolic disorders, infection prevention, neurological outcomes, and other systemic applications remain heterogeneous and is frequently derived from preclinical models, viable probiotic comparators, or multicomponent formulations in which the independent contribution of the non-viable microbial component cannot be established. Similarly, non-viability eliminates the capacity for microbial replication but should not be interpreted as proof of complete safety; compositional, toxicological, and clinical evaluation remains necessary for each preparation and intended population.

From a food-technology perspective, non-viable microbial preparations offer important formulation opportunities because their intended activity does not depend on maintaining viable cell counts. Studies in dairy, bakery, meat, seafood, and infant-nutrition systems demonstrate that these preparations can influence product stability, texture, microbial quality, sensory acceptance, and selected functional properties. Nevertheless, their performance is strongly matrix- and process-dependent, and thermal resistance, storage stability, sensory compatibility, and biological activity must be established experimentally for each product rather than assumed as universal characteristics.

Progress toward wider clinical and commercial application will therefore depend on improved standardization. Critical priorities include precise taxonomic identification, reproducible inactivation procedures, verification of residual viability, molecular and compositional characterization, validated potency markers, standardized dose reporting, stability assessment, and adequately powered randomized clinical trials. Harmonization of terminology and regulatory approaches will also be essential to distinguish scientifically defined postbiotics from purified microbial metabolites, cell-free products, and other microbial-derived ingredients that do not necessarily meet the ISAPP definition.

Future development should increasingly integrate multi-omics characterization, scalable purification, targeted delivery, metabolic engineering, and biomarker-supported clinical research. These approaches may enable a transition from broadly defined microbial preparations toward compositionally defined, reproducible, and indication-specific products. Rather than being regarded as universal replacements for probiotics, paraprobiotics and postbiotics should therefore be considered complementary microbial-derived strategies whose clinical and technological value must be demonstrated on a product-specific basis.

Generative AI statement

The author(s) declared that generative AI was used in the creation of this manuscript. During the preparation of this manuscript, the authors used Grammarly, InstaText, and ChatGPT for language editing and structural refinement. Canva and OpenAI image-generation tools were also used to generate and/or edit graphical elements in Figures 1–4. All AI-assisted textual and visual content was critically reviewed and verified by the authors to ensure scientific accuracy, citation integrity, and consistency throughout the manuscript. The authors take full responsibility for the final content of the manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Shashi Kant Bhatia, Kookmin University, Republic of Korea

Reviewed by: Junxin Zhao, Xuchang University, China

Dmitrii Khrundin, Kazan National Research Technological University, Russia

Footnotes

bEVs, bacterial extracellular vesicles; CFU, colony-forming units; DSS, dextran sulfate sodium; EPS, exopolysaccharides; IBS, irritable bowel syndrome; IL-8, interleukin-8; ISAPP, International Scientific Association for Probiotics and Prebiotics; LAB, lactic acid bacteria; lcFOS, long-chain fructo-oligosaccharides; LPS, lipopolysaccharide; NF-κB, nuclear factor kappa B; SCFAs, short-chain fatty acids; scGOS, short-chain galacto-oligosaccharides; TLR2, Toll-like receptor 2; TNF-α, tumor necrosis factor alpha.