The International Scientific Association of Probiotics and Prebiotics (ISAPP) consensus statement on the definition and scope of postbiotics

In 2019, the International Scientific Association for Probiotics and Prebiotics (ISAPP) convened a panel of experts specializing in nutrition, microbial physiology, gastroenterology, paediatrics, food science and microbiology to review the definition and scope of postbiotics. The term ‘postbiotics’ is increasingly found in the scientific literature and on commercial products, yet is inconsistently used and lacks a clear definition. The purpose of this panel was to consider the scientific, commercial and regulatory parameters encompassing this emerging term, propose a useful definition and thereby establish a foundation for future developments. The panel defined a postbiotic as a “preparation of inanimate microorganisms and/or their components that confers a health benefit on the host”. Effective postbiotics must contain inactivated microbial cells or cell components, with or without metabolites, that contribute to observed health benefits. The panel also discussed existing evidence of health-promoting effects of postbiotics, potential mechanisms of action, levels of evidence required to meet the stated definition, safety and implications for stakeholders. The panel determined that a definition of postbiotics is useful so that scientists, clinical triallists, industry, regulators and consumers have common ground for future activity in this area. A generally accepted definition will hopefully lead to regulatory clarity and promote innovation and the development of new postbiotic products.

The past few decades have demonstrated unequivo cally the importance of the human microbiota to both short term and long term human health. Early pro gramming of the microbiota and immune system dur ing pregnancy, delivery, breastfeeding and weaning is important and determines adult immune function, microbiome and overall health 1 . We have also seen rapid growth in the number of products that claim to affect the functions and composition of the microbiota at different body sites to benefit human health.
Improving human health through modulation of microbial interactions during all phases of life is an evolving concept that is increasingly important for consumers, food manufacturers, health care profes sionals and regulators. Microbiota modulating die tary interventions include many fermented foods and fibre rich dietary regimens, as well as probiotics, prebi otics and synbiotics, some of which are available as drugs and medical devices, as well as foods 2 . The rich, diverse microbial ecosystems and immune cells inhabiting all mucosal and cutaneous surfaces provide targets for intervention, with the goals of reducing the risk of dis eases and improving health status 2 . Consensus defini tions of probiotics, prebiotics and synbiotics have been published previously. Probiotics are "live microorgan isms that, when administered in adequate amounts, con fer a health benefit on the host" 3 , whereas a prebiotic is a "substrate that is selectively utilized by host microorgan isms conferring a health benefit" 4 . A synbiotic, initially conceived as a combination of both probiotics and preb iotics, has now been defined as "a mixture comprising live microorganisms and substrate(s) selectively utilized by host microorganisms that confers a health benefit on the host" 5 . The concept of postbiotics is related to this family of terms and is emerging as an important microorganism derived tool to promote health.
Probiotics are by definition alive and required to have an efficacious amount of viable bacteria at the The International Scientific Association of Probiotics and Prebiotics (ISAPP) consensus statement on the definition and scope of postbiotics time of administration to the host, but most probiotic preparations, especially at the end of shelf life, will also include potentially large numbers of dead and injured microorganisms 6,7 . The potential influence of non viable bacterial cells and their components on probiotic functionality has had little attention.
Fermented foods might also contain a substantial number of non viable microbial cells, particularly after prolonged storage or after processing, such as pasteur ization (for example, soy sauce) or baking (for exam ple, sourdough bread). Food fermentation has a major influence on the physical properties and potential health effects of many foods, especially milk and plant based foods 8 . Many fermentations are mediated by lactic acid bacteria, which can produce a range of cellular structures and metabolites that have been associated with human health, including various cell surface components, lactic acid, short chain fatty acids (SCFAs) and bioactive pep tides among other metabolites 9 . These effector molecules of fermented food microorganisms are thought to be sim ilar to those produced by probiotics, but this link has not been conclusively established. In parallel, bacterial lysates of common bacterial respiratory pathogens have been used for decades to prevent paediatric respiratory diseases by postulated general immune stimulating mechanisms that are not yet well understood 10 . The possibility that non viable microorganisms, their components and their end products play a part in the health benefits of such products is the rationale underlying the need for accurate terminology. We consider that a common understand ing of the emerging concept of postbiotics, including a consensus definition, would benefit all stakeholders and facilitate developments of this field. Herein, we address several aspects pertaining to postbiotics, including pro cessing factors important in their creation, proper char acterization, mechanistic rationale on how they work to improve both intestinal and systemic health, safety and current regulatory frameworks. Key conclusions from this consensus panel are provided in Box 1.
Methods ISAPP, a non profit collaboration of scientists dedicated to advancing the science of probiotics and prebiotics, convened an expert panel of basic and clinical scien tists to address the emerging concept of postbiotics in December 2019. ISAPP activities are determined by a volunteer academic board that functions independently of industry supporters of the organization. The panel comprised experts in probiotics and postbiotics, adult and paediatric gastroenterology, paediatrics, metabolo mics, regulatory affairs, microbiology, functional genom ics, cellular physiology of probiotics and host interactions and/or immunology. Prior to the meeting, panellists agreed on the relevant questions. During the meeting, panellists presented perspectives and evidence, debated the proposed questions and reached consensus. After the meeting, individual panellists wrote sections of this paper and the major contributions were as follows: S.S., regu latory aspects and background; H.S., paediatric health, nutrition and systematic reviews; R.S., paediat rics and evidence based recommendations; A.E.

Proposed definition of postbiotic
The term postbiotic was chosen by the panel as a com posite of 'biotic' , defined as "relating to or resulting from living organisms", and 'post' , a prefix meaning 'after' . Together these terms suggest 'after life'; that is, non living organisms. The concept that non living microorganisms could promote or preserve health is not new, and sev eral terms have been used to describe such substances, although during the past decade, postbiotic has been used most often (Figs 1,2). Other related terms have also been used, including 'paraprobiotics' [11][12][13][14] , 'parapsychobi otics' 15 , 'ghost probiotics' 16 , 'metabiotics' 17,18 , 'tyndallized probiotics' 19,20 and 'bacterial lysates' 21 . However, the field would benefit from coalescing around the use of a single, well defined and understood term rather than the use of disparate terms for similar concepts. We suggest that the term 'postbiotic' be used when applicable.
We propose that a postbiotic is a "preparation of inanimate microorganisms and/or their components that confers a health benefit on the host". Alternative definitions of this word have been proposed (TaBle 1), but we believe this consensus definition best fits the understanding of this concept. This wording was cho sen following substantial debate and consensus build ing. We chose to use 'inanimate' , meaning lifeless, rather than 'inactive' as this latter term might suggest an inert 0123456789();: material. 'Inanimate' simply captures the fact that live microorganisms were present but have now been killed, without implying a loss of function. 'Preparations' was chosen to reflect the likelihood that a specific formu lation of microbial biomass, the matrices and/or inac tivation methods have a role in any beneficial effects. The term 'postbiotic' would, therefore, be reserved for specific preparations, which would include descrip tions of the microorganisms, the matrix and the inac tivation method that had collectively contributed to a demonstrated health benefit. The word 'components' was included because intact microorganisms might not be required for health effects, and any effects might be mediated by microbial cell components, includ ing pili, cell wall components or other structures. The presence of microbial metabolites or end prod ucts of growth on the specified matrix produced dur ing growth and/or fermentation is also anticipated in some postbiotic preparations, although the definition would not include substantially purified metabolites in the absence of cellular biomass. Such purified mol ecules should instead be named using existing, clear chemical nomenclature, for example, butyric acid or lactic acid. Vaccines, substantially purified compo nents and products (for example, proteins, peptides, exopolysaccharides, SCFAs, filtrates without cell com ponents and chemically synthesized compounds), and biological entities such as viruses (including bacterio phages) would not qualify as postbiotics in their own right, although some might be present in postbiotic preparations. To qualify as a postbiotic, the microbial composition prior to inactivation must be characterized, and so preparations derived from undefined microor ganisms are not included in the definition. For example, many traditional fermented foods are made through the action of undefined, mixed cultures, and such a product could not be used for the preparation of a postbiotic. However, postbiotics could be derived from fermented products made using defined microorganisms. The criteria for a preparation to qualify as a postbiotic are shown in Box 2. Many existing postbiotics include inanimate strains belonging to established probiotic taxa within some genera of the family Lactobacillaceae (now comprising 31 genera 22 ) or the genus Bifidobacterium [23][24][25] . However, a microbial strain or consortium does not have to qual ify as a probiotic (while living) for the inactivated ver sion to be accepted as a postbiotic. Specific strains of Akkermansia muciniphila, Faecalibacterium prausnitzii, Bacteroides xylanisolvens, Bacteroides uniformis, Eubacterium hallii, Clostridium cluster IV and XIVa, Apilactobacillus kunkeei and the fungus Saccharomyces boulardii have all been investigated for potential ben eficial effects in an inanimate form and would fit the definition of postbiotic should a health benefit be demonstrated [26][27][28][29][30] . Many bacterial lysates have been used for medical purposes, but there is a clear need for more robust clinical trials. For example, a report by the European Medicines Agency (EMA) describes the assessment of eight different lysates developed for respiratory conditions 31 . The report provides a review of the results of clinical studies, data on adverse effects reported with these medicines, and advice from an expert group on infectious diseases and considers the benefit-risk balance of bacterial lysate based products. Based on this review, EMA recommended that bacterial lysate medicines authorized for respiratory conditions should only be used for the prevention of recurrent respiratory infections and not for treatment or pneu monia. The companies must also provide further data on safety and effectiveness from new clinical studies by 2026. A commercial oral postbiotic developed to pro tect against a variety of respiratory pathogens through boosting immune function illustrates the possible microbiological complexity of postbiotic design 32 . For this preparation, 21 different bacterial strains are grown in individual batches, heat inactivated once they reach a critical mass, harvested, and then subjected to alka line lysis and further purification steps 33 34 and, indeed, efficacy in reduc ing the frequency of acute respiratory infections among those prone to recurrent respiratory infections has been demonstrated in clinical trials 34,35 . In addition, polyvalent bacterial lysates derived from the mechan ical lysis of strains commonly involved in respiratory infections such as otitis media, pharyngitis, sinusitis and sometimes pneumonia induced the maturation of dendritic cells, recruit B and T lymphocytes, increase the number of circulating natural killer cells in treated patients when compared with age matched controls 30 and induced the secretion of specific IgA [36][37][38] in a group of ten healthy volunteers, suggesting some potential in the treatment of chronic pulmonary conditions such as chronic obstructive pulmonary disease. Unfortunately, a large randomized placebo controlled clinical trial with the related lysate in 288 patients (142 in the pla cebo group and 146 in the treatment group) failed to

Box 1 | Main conclusions of the consensus panel regarding postbiotics
• A postbiotic is defined as a "preparation of inanimate microorganisms and/or their components that confers a health benefit on the host". • postbiotics are deliberately inactivated microbial cells with or without metabolites or cell components that contribute to demonstrated health benefits. • purified microbial metabolites and vaccines are not postbiotics.
• A postbiotic does not have to be derived from a probiotic for the inactivated version to be accepted as a postbiotic. • the beneficial effects of a postbiotic on health must be confirmed in the target host (species and subpopulation). • the host can include humans, companion animals, livestock and other targets. meet its primary end point -a reduction in exacer bations of chronic obstructive pulmonary disease 39 . Also, some spirulina formulations could qualify as postbiotics 40 , but only if the processing and microor ganism used (often species Arthrospira platensis) is well described and the health benefit well documented in robust clinical trials.

Drivers of the postbiotic concept Stability
One important factor driving interest in postbiotics is their inherent stability, both during industrial processes and storage. Maintaining stability of live microorganisms is a technological challenge as many probiotic organisms are sensitive to oxygen and heat, but products with a long shelf life can be readily achieved for inanimate micro organisms. Postbiotics might also be more suited than probiotics to geographical regions that do not have reli able cold chains or whose ambient temperature causes problems for storage of live microorganisms.
For the majority of products with a long shelf life, probiotic die off is inevitable during storage. Because the rate of death during storage depends on the physiological characteristics of the probiotic strain and the conditions of storage (time, temperature, water activity, oxygen lev els, and others), it is difficult to generalize about the level of dead cells contained across probiotic products at the end of their shelf life. Responsible probiotic manufac turers often formulate their products with substantial overages to ensure that the labelled count of viable cells is met at the end of its shelf life. Even if such overages are not used, the live to dead ratio of a probiotic product can change substantially over the course of its shelf life 33 . Currently, probiotic product descriptions focus only on the viable cells in the product. This aspect raises some important questions. Is the efficacy of the product at the time of manufacture equivalent to the product at the end of the shelf life? What is the contribution of inanimate microorganisms to efficacy? These questions are espe cially important if the product is undergoing testing in a clinical evaluation. Although not common in the past, it seems important that going forward, quantifying the live and inactivated components of a probiotic product should be conducted over the course of an efficacy trial. Lastly, the safety of the probiotic must be assessed for the actual formulation amount, including overages. All of these concerns related to probiotic viability do not apply to postbiotics, which are likely be extremely stable for several years at room temperature and would be based on a fixed level of a viable microorganisms at the time of manufacture.

Intellectual property protection
Another possible advantage of products devoid of live microorganisms is that the microorganisms from which the postbiotic is derived cannot be isolated from the commercial product, thereby enabling product developers to maintain ownership of their ingredients. However, the ability of researchers to reproduce findings is imperative for progress in this developing field and so we encourage researchers to make available the viable progenitor strains for research purposes, for instance, by depositing them in a public culture collection. The neg ligible level of viable microorganisms could also be an advantage in the development of postbiotics that might include genetically modified microorganisms, for which dissemination into the environment might be hazardous. Finally, if a postbiotic was derived from a microorganism from a country/region protected by the Nagoya Protocol (an international agreement that promotes sharing of benefits arising from biological resources in a fair and equitable way), the country of origin would be able to retain control of the microorganism.

Regulatory considerations
To our knowledge, no regulators have advanced a postbiotic definition or framework specific to postbiotic containing foods or food supplements. Some regulatory requirements have been advanced for postbi otic formulations whose intended use is directed towards medical or pharmaceutical applications 31 .
In Japan, postbiotics (termed 'biogenics' by Mitsuoka in 1998 (reF. 41 )) have been available for more than 100 years. Most of these products contain inanimate forms of lactic acid bacteria or bifidobacteria and are used in an assortment of food products, including juices, ice creams, popcorn, potato chips, natto (fermented soybeans), instant type miso soup (traditional Japanese soup), supplements, tablets, pancake powder and many more. Most of these products are not associated with any health claims, but three products (two fermented milk type drinks and a tablet) display health claims based on a regulation of Foods with Function Claims (FFC) 15,[42][43][44] . The ingredient statements on such products might include, for example, lactobacilli, but they do not always state that the microorganisms added are non viable. This type of labelling could mislead consumers concerning the content of the products. The data that support the plots within Fig. 1 are available from the authors upon reasonable request.
Three regulatory approaches are possible for mak ing health claims on foods in Japan: Food for Specified Health Uses (FOSHU), Foods with Nutrient Function Claims (FNFC) and FFC 45 . However, the FNFC is likely not applicable to postbiotics, leaving two possible routes . To date, no postbiotic food products have health claims based on FOSHU status but a few indicate health claims based on FFC are reported in the data base of the Consumer Affairs Agency of Japan. Applications for FOSHU are reviewed and permitted by the Consumer Affairs Agency of the Government of Japan. Functional analyses and safety assessments of final products are essentially based on human studies. A permission seal from the authority appears on approved products. For FFC, scientific evidence is required from a systematic review of functional components or the prod uct's own clinical studies for applications. A history of the safe consumption of the species or scientific princi ples can be used to establish safety. A permission seal is not available for FFC 45 .
Postbiotics have had a long presence in Europe. Several postbiotics have been marketed or regulated as immune stimulating agents 46 . However, in the European Union, no specific regulation covers probiotics, pre biotics, synbiotics or postbiotics. However, as we pro pose that their definition requires a health benefit, we expect that the use of any of these terms on a food or food supplement would require health claim approval. With regard to safety assessment in Europe, the European Food Safety Authority (EFSA) develops regularly updated lists of microorganisms that meet criteria for presumptive safety for use in foods. This process, called Qualitative Presumption of Safety (QPS), would apply to live microorganisms (including bacteria and yeast) used as progenitor microorganisms for postbiotics. Microorganisms not found on the list require a sys tematic novel food application and approval in Europe before they can be used for postbiotic devel opment for foods or feeds. An example of a safety assessment of a potential postbiotic includes B. xylanisolvens for food 47 , which has undergone safety evaluations con ducted on heat treated or inactivated bacteria. For postbiotics formulated in medical products, the EMA (Directive 2004/27/EC) 48 ) is in charge of both evaluation and supervision. For pharmaceutical preparations and medicinal products, the European Pharmacopoeia has clear criteria, which stipulate maximum allowed levels of live microorganisms 49 . Such criteria should be easily met by postbiotic products. The new EU Regulation 2017/745 (reF. 50 ) for medical devices also has a specific paragraph positioning 'living organisms' out of the scope of the regulation but postbiotics do not seem to be out of scope.
In South America, Brazil has been the most active country in addressing probiotics and incorporating them in their regulations, publishing the first guide lines for their evaluation in 1999. Argentina did the same in 2011 and Chile in 2017 (reF. 51 ). However, Brazil still takes the lead by updating their guidelines as they deem necessary according to the advancing knowledge on probiotics. The fact that Brazil was the first country/ territory to address probiotic regulations, which have been updated several times over the past 20 years, could suggest that it might be the first in the region to incorporate postbiotics. In Argentina, the Argentinian Food Code incorpo rated the concepts of probiotics and prebiotics in 2011 under Articles 1389 and 1390, respectively. However, the topic of postbiotics has not yet been addressed, even though in 2019 an international company launched an infant formula with 30% of its composition being derived from spray dry inactivated milk fermented with Streptococcus thermophilus and a Bifidobacterium strain, and the product was labelled 'with postbiotics' . As in most cases, food development precedes regulation and, for regulation, a clear and well accepted definition of postbiotics is needed.
In the USA, the Food and Drug Administration (FDA) has not specifically addressed postbiotics. A search shows no mention of the term 'postbiotic' on the FDA website. As postbiotics can be developed under different regulatory categories 52 , the FDA will probably approach postbiotics based on the regulations that pertain to the specific regulatory category chosen for a product under development. The product's intended use, safety and efficacy will need to meet the standard for the applica ble regulatory category. Thus, for example, if a postbiotic is to be used as a food ingredient, it will either need to undergo premarket approval as a food additive or need to be evaluated by experts to determine whether it is generally recognized as safe. Any health benefit claims made would need to be approved by the FDA either as a health claim, which identifies a food as able to reduce the risk of disease, or as a non approved general function claim, which identifies a food as influencing the normal structure or function of the human body. Other regu latory categories that postbiotics could potentially fall under include drugs, medical devices or subcategories of foods, such as dietary supplements, infant formulas, foods for special dietary use or medical foods.

Safety
Postbiotics could reasonably be expected to have a better safety profile than probiotics, because the microorgan isms they contain have lost the capacity to replicate and therefore cannot cause bacteraemia or fungaemia, risks that are associated with probiotic administration (albeit extremely rare) 53 . However, postbiotics cannot be pre sumed to be safe solely based on the safety profile of the progenitor microorganism. For example, lipopoly saccharides from Gram negative bacteria can induce sep sis and toxic shock, especially when endotoxin A, which is normally embedded in the outer membrane in living bacteria, is released from dead bacteria 54 . An assessment of safety for the intended use for any postbiotic is needed prior to use. Postbiotics derived from food grade micro organisms or species in the continuously updated EFSA QPS lists might have an easier path to approval. Postbiotics are inanimate by definition, and unless they are rapidly killed under the conditions used to make a product (for example, a strict anaerobe might not survive exposure to atmospheric conditions), they will require an inactivation step. A number of options are available to achieve this objective, and while this sec tion lists some of the likely options, it is not an exhaus tive list of available treatments that could inactivate microorganisms.

Inactivation
Thermal processing is likely to be used in many instances to inactivate microorganisms, as there is a long history of thermal processing in the food industry. Traditional thermal processing (pasteurization, tyndallization, autoclaving) is widely used to confer enzymatic and microbiological stability on food systems. However, the temperature and length of time of heating affect nutri tional value, sensory characteristics and flavour 55 . As a result, thermal processing might not always be optimal when generating a postbiotic preparation intended to be used as a food supplement or as a food.
Other processing technologies can provide useful alternatives to thermal sterilization or pasteurization 56 . Most of the technological knowledge concerning the non thermal inactivation of microorganisms in foods was developed for the inactivation of food borne microbial pathogens or spoilage microorganisms, but these technologies could be used equally well for the production of postbiotics. Non thermal inactivation techniques were designed to obtain safe and stable foods with preserved overall quality and value while maintaining their sensory characteristics close to those of their fresh equivalents. In this context, technologies such as electric field, ultrasonication, high pressure, X rays, ionizing radiation, high voltage electrical dis charge, pulsed light, magnetic field heating, moder ate magnetic field 55 and plasma technology 57 could all potentially be applied to inactivate microorganisms and generate postbiotics. Spray drying is a method of producing a dry powder from a liquid or slurry by rapidly drying with a hot gas. Spray drying has been proposed as a low cost alterna tive to freeze drying to develop dehydrated but viable microbial cultures 58 , and could be used with higher inlet and/or outlet temperatures to achieve microbial inacti vation. Spray dried infant formulas fermented with lac tic acid bacteria and bifidobacteria, but not containing substantial amounts of viable bacteria in the final prod uct, are widely available in many countries 59 . They can therefore be labelled as including postbiotics if they are in agreement with our proposed definition and criteria.
Other drying techniques, such as vacuum and fluid ized bed drying, have been shown to stress microorgan isms and decrease their viability 60 and could potentially be used under harsher operative conditions to com pletely inactivate cultures. Even more effective micro bial inactivation might be achievable by the combined or successive application of these milder technologies, applied either independently or in tandem with other stresses, such as mild temperature 61 .
In addition to the level of microbial inactivation achieved, the functionality of a postbiotic might be affected by the means of production. For instance, it has been shown that different heat treatments applied to the development of dehydrated probiotics (air dry ing, freeze drying and spray drying) can strongly affect both the viability and immunomodulatory properties of probiotic strains, and thus we can surmise that such treatments could also affect postbiotic properties 62 . Non thermal treatments, such as high pressure, have also been reported to modify the in vivo host response to lactobacilli 63 . Figure 3 shows cells of Lacticaseibacillus rhamnosus GG (formerly known as Lactobacillus rhamnosus) before and after spray drying, which resulted in a mixture of live, fully piliated cells and inactivated cells lacking pili surface appendages. Pili are cell surface structures known to mediate bacterial-host immune interactions. For example, loss of pili has been linked to increased induction of pro inflammatory markers such as IL8 and less stimulation of cell proliferation and protection against radiologically inflicted intestinal injury in Caco2 intestinal epithelial cells 64 .
We can learn much about the likely extent of micro bial inactivation that can be achieved by thermal and non thermal processing from studies conducted on food borne pathogens. When heat is used, complete inactivation can be proportional to the level of heat and time of exposure, whereas in non thermal food pro cessing complete inactivation might not always occur in a linear fashion 65,66 . The extent of microbial inactiva tion depends on multiple factors related to the cell type (prokaryotes versus eukaryotes, Gram positive versus Gram negative bacteria, vegetative cells versus spores, cocci versus rod shaped microorganisms), the processing conditions and the composition of the matrix 67 .

Parameters for inactivation
Most postbiotics will contain no viable cells but some survivors might persist depending on the inactivation conditions 47 . Different inactivation technologies (heat, high pressure, exposure time to oxygen for strict anaer obic microorganisms) and procedures could be expected to result in different numbers of remaining viable cells of the progenitor microorganisms, although such com parisons have not yet been published. At the same time, extreme inactivation conditions designed to achieve complete inactivation might negatively influence the nutritional, physical, rheological or sensorial properties of the material. Thus, the inactivation method chosen could result in some residual, live microorganisms. Our intention is not to disqualify such products from our postbiotic definition. Although we do not require that a postbiotic be microbiologically sterile, there must be intentional and deliberate processing designed to inactivate the microbial progenitor strain. Here we do not suggest a precise limit on allowable live microor ganisms remaining after postbiotic preparation as this is more appropriately a matter for regulators, as can be found in an EFSA assessment of B. xylanisolvens 47 .

Quantification
Suitable methods must be available to describe the com position of and to quantify a postbiotic product. These methods must be available for clear product description to facilitate duplicative research as well as for quality control at the production site. Flow cytometry is emerg ing as an alternative to plate counting for microbial detection and enumeration 68 . In addition to being faster, it has the advantage of being able to separate a microbial population into live, damaged and dead cells. Results are expressed as total fluorescent units and active fluores cent units (AFUs). In flow cytometry, cells pass through a narrow aperture and they are analysed individually by a laser. A limitation of this counting method is that the correlation between AFUs and colony forming units (CFUs) is not established, especially when applied to inactivation treatments that might produce several large fragments from a single cell (Fig. 3). Potentially, one cell rendering several fragments could be counted as several AFUs. In cases in which an AFU to CFU ratio of 1:1 is not expected owing to the disintegration of the microbial cell after an inactivation treatment has been Box 2 | Criteria for a preparation to qualify as a postbiotic • molecular characterization of the progenitor microorganisms (for example, fully annotated genome sequence) to enable accurate identification and screen for potential genes of safety concern • Detailed description of the inactivation procedure and the matrix • Confirmation that inactivation has occurred • evidence of a health benefit in the host from a controlled, high-quality trial applied, cell counts before inactivation might be a useful method to report the concentration of the postbiotic in the final product. Alternative analytical methods to ana lyse and quantify microbial biomass include proteom ics and enzyme linked immunosorbent assay based approaches 69 , real time PCR 70 , flow cytometry 68 , drop let digital PCR 71,72 , NMR 73 , atomic force spectroscopy 74 , scanning electron microscopy 75 and Fourier transform infrared spectroscopy 76 , but they are not yet commonly used by industry.
Freshly grown microbial cultures displaying high levels of viable cells can sometimes contain a higher number of non viable cells, even in the absence of any inactivation step 77 . The level of inactive cells will depend on the conditions of the biomass production process, such as the growth phase at harvesting, medium compo sition or the pH profile throughout fermentation. Thus, because postbiotics will be derived from both active and inactive cells, CFU counts prior to inactivation might not prove an effective means of defining the cell biomass of a postbiotic product. Because CFUs before the inac tivation process could underestimate the true biomass, flow cytometry might be a more suitable method.
It is also possible that intact inactivated cells could interact differently with the immune system when com pared with their cell wall and cell membrane fragments, because of the different conformation and avidity of the immune interaction molecules 6 . In this scenario, the type of technology used to inactivate cells (regardless of whether intact cells or cell fragments are generated) might result in products with different functionality compared with the progenitor microbial product. For this reason, it is important that each postbiotic prepara tion is consistently produced using the same technologi cal process as the one used in the study in which a health benefit was demonstrated. If the process is altered, it is important to ensure the resulting product will produce the expected health effect.

Biomolecules mediating health effects
The ability of a postbiotic, which can be a heterogene ous mixture of components, to mediate a health effect in the target host might be driven by many different mechanisms. In some cases, these mechanisms could be similar to those known for probiotics 3,78 . Such mech anisms might act independently or in combination. Understanding the major effector molecules involved in eliciting such beneficial effects is important information to ensure that a commercial postbiotic product retains the attributes necessary for efficacy. Because postbiotics are inanimate, these bioactive molecules must be synthe sized by the progenitor microorganisms prior to inacti vation, and in sufficient amounts to induce a beneficial effect. Here, we review possible mechanisms that could drive postbiotic efficacy. Overall, five main modes of action are considered, as depicted in Fig. 4.

Beneficial modulation of microbiota
Although effects of postbiotics on the microbiota might be temporary, they could still have an important mech anistic role. Molecules present in postbiotics, such as lactic acid 79 and bacteriocins 80 , can have direct antimi crobial activity according to in vivo studies. Postbiotics could also modulate the microbiota indirectly, for exam ple by carrying quorum sensing and quorum quenching molecules 81 or by carrying lactic acid that can be con sumed by some members of the microbiota resulting in SCFAs and butyrate, which have a beneficial function 82 . Postbiotics can also compete with resident microorgan isms for adhesion sites if the postbiotics provide adhes ins (such as fimbriae 83 and lectins 84 ) that remain intact after processing.

Enhancing epithelial barrier function
Activities that enhance epithelial barrier function can be mediated by secreted proteins, such as the major secreted proteins Msp1/p75 and Msp1/p40 (reF. 85 ) or the protein HM0539 (reF. 86   form showing that processing steps to obtain postbiotics can have a major effect on the physical and functional properties of the bacteria, even if the overall biomass and rod shape is preserved. Inactivation was performed in this case by spray drying that resulted in a mixture of live, full piliated cells and inactivated cells lacking pili surface appendages (as described in Kiekens et al. 75 ). The bacteria were spotted on a goldcoated membrane, which is especially visible after processing. Adapted with permission from reF. 75  from Bifidobacterium species, can promote barrier func tion by reducing inflammation via yet to be defined signalling mechanisms 87 . Increasing evidence shows that certain Bifidobacterium species induce signalling pathways, such as MAPK and AKT, that promote tight junction functioning via autophagy and calcium signal ling pathways 88 . SCFAs present in a postbiotic prepa ration have the potential to modify epithelial barrier function and protect against lipopolysaccharide induced disruption, if present at sufficient levels 89 . For example, acetate (0.5 mM), propionate (0.01 mM) and butyrate (0.01 mM), alone or in combination, were shown to increase transepithelial resistance and stimulate the for mation of tight junction in Caco2 intestinal epithelial cells in vitro 89 . In another study, butyrate was demon strated to alter the permeability of tight junctions via lipoxygenase activation through histone acetylation in Caco2 cell lines 90 .

Modulation of immune responses
Immune modulatory activities at both local and systemic levels are generally exerted by microorganism associated molecular patterns interacting with specific pat tern recognition receptors of immune cells, such as Toll like receptors (TLRs), nucleotide binding oligomer ization domain (NOD) receptors and C type lectins, resulting in the expression of various cytokine and immune modulators 91   . Conceptually, the activity of effector molecules could be better retained if the cellular structure of the postbiotics is preserved, for example, through increased avidity in interactions with immune receptors or through increasing the residence time of the active molecules inside the host. The cell wall protects against rapid degradation by digestive enzymes and immune attack inside the host. This aspect is similar to the situation with vaccines, which also function best if cellular structure is preserved, but with the most toxic and/or pathogenic parts being inactivated or deleted. BSH, bile salt hydrolase; EPS, exopolysaccharide; MAMP, microbe-associated molecular pattern; PRR, patternrecognition receptor; SCFAs, short-chain fatty acids; TCR, T cell receptor; T H cell, T helper cell; T reg cell, regulatory T cell. postbiotics derived from Gram negative bacteria, such as Escherichia coli Nissle, mostly interacting with TLR4 and sometimes TLR2 (reF. 96 ); β glucans in yeast, such as Saccharomyces cerevisiae, interacting with TLR2 and lectin immune receptors 97 ; and lipoproteins mostly interacting via TLR2 (reF. 98 ). These microbe associated molecular patterns could also be present in postbiotics if not destroyed or modified by the inactivation pro cess. Some of the immunostimulatory bacterial lysate mixtures mentioned earlier contain lysates from both Gram positive and Gram negative bacteria and have been shown to interact with TLR4 and TLR2 (reF. 99 ). In addition, metabolites, such as lactic acid, have been reported to mediate immune effects through, for example, the GPR31 dependent dendrite protrusion of intestinal CX3CR1 + cells 100 . Similarly, indole deriv atives of tryptophan generated by Limosilactobacillus reuteri (formerly Lactobacillus reuteri) can activate the aryl hydrocarbon receptor in CD4 + T cells in the mouse gut, inducing differentiation into CD4 + CD8αα + double positive intraepithelial lymphocytes 101 . However, it is unknown whether indole derivatives are stably contained in postbiotic formulations. Other immuno modulatory microbial metabolites that could be present in postbiotics, based on molecular research in related microorganisms, include histamine 102 , branched chain fatty acids and SCFAs 103 , which have been shown to influence a number of immune responses, including suppression of NF κB.

Modulation of systemic metabolism
Effects on systemic metabolic responses can be directly mediated by the metabolites or enzymes inside and on the surface of the inactivated microorganisms in the postbiotics. One example is bile salt hydrolase (BSH). This microbial enzyme is responsible for the deconjuga tion of bile acids that enables further microbial biotrans formation to occur, diversifying the overall circulating bile acid pool 104 . Bile acids can modulate the commu nity structure of the microbiota generally and interact with various host receptors, with downstream effects on a range of host metabolic processes, including glucose, lipid, xenobiotic and energy metabolism 104 . BSH is pre dominantly expressed in the cytoplasm of microorgan isms, but extracellular forms have also been observed, and its activity has been demonstrated in the filtered supernatant of the probiotic Lactobacillus johnsonii 105 . Interestingly, a loss of gut microbiota derived BSH pre disposes individuals to recurrent Clostridioides difficile infection, but restoration of this activity through faecal microbiota transplantation has been shown to assist in treating this infection, which was demonstrated in a study analysing stool samples from 26 patients and their 17 donors, followed by validation in a mouse model 106 . Another example is succinate, a bacterial intermediate of carbohydrate fermentation. Succinate is a substrate for intestinal gluconeogenesis that has been found to improve glycaemic control in mice 107 . Other known modulators of host metabolism include microbial derived vitamins and SCFAs. Propionate can improve insulin sensitivity and glucose tolerance and modify lipid metabolism 108 , whereas butyrate can upregulate the antioxidant glutathione and can affect oxidative stress beneficially in the colon of healthy humans 109 .

Signalling via the nervous system
Microorganisms can produce various neuroactive com pounds that can act on both the enteric and central nervous systems with the potential to modulate behav iour and cognitive function in animals and humans 110 . This includes several neurotransmitters such as seroto nin, dopamine, acetylcholine and GABA, and various compounds that can bind to receptors expressed in the brain (for example, indoles and bile acids). Microbial enzymes can also metabolize dietary precursors for host neurotransmitter synthesis (for example, tryp tophan (for serotonin) and tyrosine (for dopamine)), reducing their bioavailability 111 . In addition, microbial metabolites, such as SCFAs, if present in a sufficient quantity in the postbiotic preparation, could stimulate enterochromaffin cells to produce serotonin, which can subsequently enter the bloodstream 112 . A study in mice and human enteroids using live and heat killed Bifidobacterium dentium has highlighted that viability is crucial for serotonin induction by this microorganism 113 , so that it remains to be investigated whether postbiotic preparations other than heat killed preparations could exert this effect. Moreover, SCFAs have been shown in human intervention studies to be able to modify feeding behaviours through the promotion of satiety by stim ulating the release of anorexigenic hormones, such as glucagon like peptide 1 and peptide YY 114,115 . In mice, gut derived acetate has also been shown to enter the brain and regulate appetite through a central metabolic mechanism 116 . Bacterially synthesized vitamins, such as B vitamins (riboflavin, folate and cobalamin), can also be present in probiotics 117 and probably also retained in postbiotics. B vitamins have important beneficial roles in central nervous system function 118 . However, how much of these neuroactive molecules are retained in postbiotics is not well documented at present.
We hypothesize that, as a general rule, the activity of effector molecules is increased if the cellular structure of the postbiotics is preserved, for example, through increased avidity in interactions with immune receptors or through increasing the residence time of the active molecules inside the host, because the cell wall protects against rapid degradation by digestive enzymes and immune attack inside the host, but further experimental proof is needed. This situation is similar to that with vac cines, which also function best if the cellular structure is preserved, but with the most toxic or pathogenic parts inactivated or deleted 119 . However, it cannot be ruled out that the activity and bio availability of effector metabo lites such as amino acid derivatives and SCFAs might be increased when the cellular structure is degraded owing to the molecules becoming more exposed and available.
Health benefits of postbiotics Postbiotics in general have been studied in the pre ventative and treatment contexts. Most of the research cited is in the medical field for therapeutic applications, but postbiotics could also have nutritional benefits. The following discussion focuses on preclinical studies and postbiotic mediated benefits in adults and paediatric populations.

Animal studies
The possibilities for postbiotics as clinical interventions have been well illustrated in the laboratory. Observations in animal models have, for some time, demonstrated biological activity of inanimate bacteria, which offer considerable formulation, safety and regulatory advan tages over their 'live' counterparts. An example is a postbiotic derived from Limosilactobacillus fermentum and Lactobacillus delbrueckii that influenced behav iour in a mouse model. The fermentate was subjected to a high temperature treatment to achieve microbial inactivation 120 . The postbiotic fed animals demonstrated increased sociability and lower baseline corticosterone levels (stress hormone) and had subtle but statistically significant changes in the composition of their gut microbiota when compared with controls receiving a standard rodent chow. The study found that less abun dant taxa were most affected. The same research group went on to use the same postbiotic in a mouse model of Citrobacter induced colitis, which is characterized by a shortening of the small intestine and an increase in colon crypt depth 121 . The postbiotic did not prevent Citrobacter infection, but postbiotic fed mice had a longer small intestine and reduced colon crypt depth compared with control animals that received standard mouse chow alone.

Postbiotics in adults
Available evidence. For evidence on the health benefits of postbiotics in adults, the Cochrane Central Regis ter of Controlled Trials and MEDLINE databases were searched for randomized controlled trials (RCTs), cohort studies, or their meta analyses that compared postbiotics with placebos or no therapy. Data from human studies are limited but efficacy for orally administered, inacti vated lactic acid bacteria has been demonstrated in the eradication of Helicobacter pylori infection 122 , reduction of symptoms in patients with irritable bowel syndrome (IBS) 25,123 and chronic unexplained diarrhoea 124 , and in the abrogation of the negative effects of stress 15,125 . In a randomized, double blind, placebo controlled trial in 443 individuals with IBS involving orally administered, heat inactivated Bifidobacterium bifidum MIMBb75, the postbiotic substantially alleviated symptoms asso ciated with IBS, such as abdominal pain or discomfort, abdominal bloating and abnormal bowel habits 25 .
No benefits were seen in terms of modulating gut barrier function in 25 patients with increased permeabil ity secondary to obstructive jaundice treated with inac tivated Lactiplantibacillus plantarum (formerly known as Lactobacillus plantarum) 126 . Other inactivated strains, such as Bacillus coagulans (effect on responses to vig orous exercise among soldiers undergoing self defence training) 127 , Mycobacterium manresensis (in tuber culosis) 128 , Mycobacterium phlei (in asthma) 129 and H. influenzae (in severe chronic obstructive pulmonary disease) 130 have also been studied in humans. As is the case with this entire category, data from human studies are limited, are of variable quality and have resulted in varying clinical impacts. Mycobacterium vaccae has attracted considerable attention because of the immuno regulatory and anti inflammatory properties of the heat killed microorganism, as demonstrated in the central nervous system, for example 131 . Others are also deve loping topical products with lysates of the probiotic L. rhamnosus GG for skin applications 132 . A preparation incorporating autologous platelet rich plasma, biomi metic peptides, postbiotics (plantaricin A, A. kunkeei bee bread) and Tropaeolum majus flower, leaf or stem extract, was shown to be superior to placebo in the treatment of alopecia areata in 160 patients 133 . These preparations could therefore be termed skin postbiotics according to the new consensus definition. Further examples of post biotics being used for therapeutic purposes in humans are delineated in TaBle 2.
Potent examples of the power and clinical impor tance of substances produced by microorganisms are numerous. Perhaps the most important examples are antibiotics, the first of which, penicillin, came from the mould, Penicillium notatum. A truly game changing immunosuppressant ciclosporin was derived from the fungus Tolypocladium inflatum. A variety of other anti bacterial molecules have been isolated from gut and other microbiota, including topically applied bacterioc ins such as nisin 134 and ESL5, a bacteriocin isolated from Enterococcus faecalis SL5 (reF. 135 ). Topical application of these substances circumvented challenges faced by an orally administered bacteriocin in the treatment of mas titis (n = 8) and acne vulgaris (n = 70), respectively. Given the increasing concerns presented by antibiotic resistant strains of a variety of human pathogens, the exploration of the microbiota for novel antimicrobials assumes great urgency. Such substances in a purified form fall outside the scope of postbiotics as defined herein, but they could contribute to functionality of preparations of inactivated microorganisms.
Clinical use. Clinical use of postbiotics has been limited by issues of delivery and formulation, but these issues are being addressed 136 and one looks forward to the realization in the clinic of the promise that basic science has shown. One group of products of microbiota-diet interactions, SCFAs, has been subjected to clinical trials in humans with some encouraging results. Butyrate ene mas have been used in clinical trials to treat ulcerative colitis (some cohort trials and some open label studies; the number of participants in individual studies ranged from 10 to 47) 137-142 and, to a limited extent, radiation proctosigmoiditis (RCTs; the number of participants ranged from 15 to 166) 143-146 and visceral hypersensi tivity (RCT in 11 healthy volunteers) 147 . SCFA enemas have become standard therapy for diversion colitis [148][149][150] . However, SCFAs used as purified substances, and not as a component of an inactivated microbial preparation, would not be considered postbiotics.  ranging from inflammatory bowel disease to radiation induced mucositis and food allergy [151][152][153][154][155][156][157] . Some tanta lizing hints of clinical efficacy have been generated for GMOs 154,157 , but regulatory challenges, as well as the court of public opinion in some regions of the world, have hampered progress in this area. Furthermore, the clinical use of preparations of inactivated GMOs as post biotics has -to the best of our knowledge -not yet been published, although such preparations are proba bly in development 158 . For feed applications in animals, some products are marketed in Europe 159 . For example, PL73 (LM) is a dried, heat inactivated bacterial bio mass used as a feed material produced from an E. coli K12 strain, which was genetically modified to over produce lysine. As mentioned earlier, we have consid ered vaccines, including from GMOs, outside the scope of the postbiotic definition, because they already have a dedicated term.

Summary.
It is clear that several clinical indications could benefit from the availability of effective post biotics, including: new antimicrobials; targeted anti inflammatory and immunoregulatory agents; novel signalling molecules that affect gut pain, sensation, secretion and motility; and agents that enhance vacci nation efficacy or modulate immune responses or that exert beneficial metabolic effects via interactions with dietary components. All could have a valuable role in clinical medicine. High quality randomized placebo controlled (or alternately, active agent controlled) trials will provide the ultimate proof.

Postbiotics in infants and children
For evidence on the health benefits of postbiotics in children, the Cochrane Central Register of Controlled Trials and MEDLINE databases were searched for RCTs or their meta analyses that compared postbiotics with placebos or no therapy (TaBle 3).

Fermented formulas.
Fermented formulas are those that are fermented with certain lactic acid bacteria during the production process and that do not contain substantial amounts of viable bacteria in the final product. Exact acceptable levels of live microorganisms have not been established by regulatory authorities. To the extent that the microorganisms used to ferment these formulas are characterized adequately, these products would fall under the postbiotic definition. Infant formulas serve as the sole nutrition source for infants who are not being breast fed. Thus, infant formulas are heavily regulated worldwide for their nutrient content as well as any added ingredients.
In 2007, the European Society for Paediatric Gastro enterology, Hepatology and Nutrition (ESPGHAN) Committee on Nutrition reviewed the evidence on fermented infant formulas. Based on two RCTs, the Committee concluded that the available data do not allow general conclusions to be drawn on the effects of fermented formulas in infants 160 . Updated data on fermented formulas can be found in TaBle 3    Formulas for pre term infants are not covered by the Codex Alimentarius, and this issue will eventually pose a challenge to the use of fermented formulas in this age category. Data on the use of fermented formula in pre term infants are limited to one RCT, which evaluated the effect of a formula fermented by Bifidobacterium breve and S. thermophilus in a total of 58 infants (gestational age 30-35 weeks) 161 . There was a reduced incidence of abdominal distension in infants fed fermented preterm formula compared with those fed standard preterm for mula, as well as statistically significantly lower faecal calprotectin levels in the former group (P = 0.001). 162 of four RCTs of varied methodological quality, involv ing 304 children aged 1-48 months, showed that heat inactivated Lactobacillus acidophilus LB reduced the duration of diarrhoea in hospitalized, but not outpa tient, children compared with a placebo. The chance of a cure on day 3 was similar in both groups, but L. acidophilus LB increased the chance of a cure on day 4 of the intervention. One trial investigated the effect of heat inactivated L. rhamnosus GG compared with via ble L. rhamnosus GG in children with acute rotavirus diarrhoea. Clinical recovery from rotavirus diarrhoea was similar in both groups 163 . A recent review covers the mechanisms as suggested by several in vitro studies 164 .

Prevention of common infectious diseases.
Data on pre venting common infectious disease are inconsistent [165][166][167][168] , However, limited results pooled from two RCTs (n = 537) carried out in healthy children aged 12-48 months attending day care or preschool for at least 5 days a week suggest that heat inactivated Lacticaseibacillus paracasei CBA L74 (formerly known as Lactobacillus paracasei) might reduce the risk of diarrhoea 165,168 , pharyngitis 165,168 , laryngitis 165,168 and otitis media 165 . By contrast, one trial 167 investigated the effect of micronutrients (including zinc) with or without heat inactivated L. acidophilus com pared with a placebo in infants aged 6-12 months at high risk of diarrhoea related mortality (defined as at least one episode of diarrhoea in the preceding 2 weeks). The prevalence of diarrhoea was 26% in the group receiv ing micronutrient with L. acidophilus, 15% in the group receiving micronutrient and 26% in the group receiving placebo. There was no statistically significant difference between the micronutrient with L. acidophilus and pla cebo groups. The authors concluded that the addition of heat inactivated L. acidophilus had a negative effect in these children.
Cow's milk allergy management. Kirjavainen et al. 169 evaluated the effects of an extensively hydrolysed whey formula (EHWF) supplemented with live or killed L. rhamnosus GG compared with the effects of non supplemented EHWF in 35 infants (mean age 5.5 months) with atopic eczema and cow's milk allergy 170,171 . The authors reported statistically significant reductions in the Scoring Atopic Dermatitis scores in the EHWF group, EHWF/viable L. rhamnosus GG group and the EHWF/heat inactivated L. rhamnosus GG group (base line versus end of a 1 month intervention). No adverse events in the EHWF group and the EHWF/viable L. rhamnosus GG group were reported. However, com pared with these two groups, the administration of the EHWF/heat inactivated L. rhamnosus GG resulted in a significantly higher risk of diarrhoea (P = 0.05).

Non-clinical outcomes.
A number of studies evaluated additional non clinical effects 163,[172][173][174][175] . For example, the use of fermented formula was found to reduce faecal pH values. However, whether the faecal pH reduction per se is of benefit is not well established. The same applies to other stool parameters, such as faecal IgA levels and bifidobacteria levels.
Summary. The effects of postbiotic supplementation have been studied mainly for fermented infant formulas and bacterial lysates. Overall, there is only limited evi dence to suggest that these products provide a health benefit compared with non postbiotic containing for mulas in the paediatric setting. The safety and poten tial harms of postbiotic interventions remain poorly explored and understood. Further multicentre studies are necessary to determine the effects and safety of different postbiotics.

Conclusions
This panel was conceived in response to the rise of the term 'postbiotics' both in the scientific literature and in relation to commercial products, as well as to the con comitant lack of clarity regarding the appropriate use of the term. The panel was interested in defining use ful, science based parameters for this emerging term. By providing a definition for the term, we hope that all stakeholders will use the term appropriately, thereby assuring a common foundation for developments in the field. If this can be achieved, it will enable scientists    and intellectual property lawyers to track publications on postbiotics easily. It will provide a common under standing of the term for researchers, industry, regulators and consumers. Responsible use of the term 'postbiotic' on a product label will compel manufacturers to meet the minimum criteria imposed by this definition, includ ing availability of controlled studies in the target host demonstrating a health benefit.
We have also clarified how postbiotics differ from other related substances, including probiotics, prebiot ics and synbiotics. The conflation of these terms leads to confusion. Furthermore, we have called out issues that should be considered when investigating postbiot ics, such as the starting material, the means of inacti vation and assurance of safety. Careful control of these parameters is important for reliable and repeatable research.

Data availability
The PubMed search data that support the plots within this paper are available from the authors upon reasonable request.