1.2.2. Notification of microorganisms for QPS assessment

A QPS assessment is triggered upon receipt by EFSA of an application dossier seeking a market authorisation of a regulated product that requires a safety assessment of a microbial strain. After the establishment of the first QPS list, no new starter organisms used for food fermentation have been included because they are not subject to a notification to EFSA for market authorisation. Because the QPS status list only considers those microorganisms sent to EFSA in the frame of notifications for market authorisation, it is not exhaustive.

1.2.3. Decision of exclusion from QPS assessment

Some microbial groups are excluded from the QPS assessment based on an ambiguous taxonomic position, lack of a sufficient body of knowledge or the possession of potentially harming traits (e.g. pathogenicity, presence of virulence factors or production of biologically active toxic secondary metabolites), because it is considered unlikely that any TUs within these groups would be granted QPS status in the foreseeable future. Thus, the assessment of members of these biological groups needs to be done at a strain level, on a case‐by‐case basis, by the relevant EFSA Unit:

Filamentous fungi

While knowledge of fungal secondary metabolites has grown substantially, information on their toxic effects on humans and animals is evolving at a much slower rate. Therefore, it was decided that until further notice, filamentous fungi would be excluded from the QPS evaluations.

Bacteriophages

Bacteriophages were not considered appropriate for consideration as QPS organisms because: (i) the lowest level of phylogenetic TU should be the order Caudovirales (which includes 95% of all known phages), which is considered to be too wide; (ii) distinguishing between transducing and non‐transducing phages or whether they carry virulence factor determinants (encoding toxins, adhesins, antibiotic resistance, etc.) would involve thorough analysis of the genomes and has to be done at the individual phage level.

Streptomycetes

Streptomycetes are essentially non‐virulent, with the exception of some plant pathogens such a S. scabies. Genome sequencing has confirmed that all streptomycetes carry gene clusters for the production of secondary metabolites, which include antimicrobial compounds, depressors of the immune system and herbicides (Butaye et al., 2003). Many of these may select for antimicrobial resistance or being toxic. The Panel therefore decided to exclude the genus Streptomyces from future QPS evaluations.

Oomycetes – Pythium oligandrum

Pythium oligandrum is an oomycete used in plant protection products. Several factors lie behind the decision that the oomycetes collectively (Class Oomycota) are not considered eligible for QPS status: (i) They seem to be rarely used as components of food and feed fermentations; (ii) There is lack of experience of actively adding oomycetes in the food and feed chains (except for a few biocontrol agents); (iii) a high proportion of the known oomycete species are pathogens, particularly to plants; (iv) there is little information available about the production of toxins by oomycetes in general (Stam et al., 2013; Berger et al., 2016; Amaro et al., 2017; Kushawa et al., 2017). The Panel therefore decided to exclude oomycetes from future QPS evaluations.

Enterococcus faecium

The inappropriateness of granting a safety status to the species E. faecium has been recognised in several EFSA Opinions (EFSA, 2007, 2008) due to the pathogenic potential of some strains, as defined by their possession of putative and confirmed virulence markers (Freitas et al., 2018).The Panel confirmed the exclusion of this species from future QPS evaluations.

Escherichia coli

Many strains of E. coli are pathogens for humans and animals. In fact, numerous strains of this species are the leading urinary and intestinal bacterial pathogens in developed and underdeveloped countries, respectively. In addition, they are frequently a cause of sepsis and many other systemic infections.

The Panel decided to exclude this species from future QPS evaluations.

2.1.1. Taxonomic identification

The TU for which the QPS status is granted is the species for bacteria, yeast and protists/algae, and the family for viruses. Only unambiguously defined biological TUs are considered for inclusion in the QPS list. Microbial taxonomy is a very dynamic discipline, recently supported mainly by phylogenetic analysis of housekeeping genes and whole genome relatedness (e.g. ANI, phylogenomics). The resulting reclassifications of microorganisms will lead to necessary adaptations in the QPS list, which are updated in the successive QPS Statements.

Bacterial taxonomy

Taxonomic identity is based on the internationally accepted classification, overseen by the International Committee on Systematics of Prokaryotes. The nomenclature of bacteria and the nomenclatural changes as cited in the Approved Lists of Bacterial Names or validly published in the International Journal of Systematic Bacteriology or in the International Journal of Systematic and Evolutionary Microbiology are reported in the website List of Prokaryotic Names with Standing in Nomenclature (LPSN) (Parte, 2018).

Fungal taxonomy

The nomenclature and taxonomy of fungi, including yeasts, is covered by the International Code of Nomenclature for algae, fungi and plants (ICN) (Turland et al., 2018). The most recent authoritative taxonomy of yeasts was published in 2011 (Kurtzman et al., 2011).

Virus taxonomy

The taxonomy and nomenclature of viruses are the responsibility of the International Committee on Taxonomy of Viruses (ICTV, 2018). Updates are made annually, based on proposals by Study Groups and after adoption by the Executive Committee. These updates form the 10^th Report of the ICTV and are available through the ICTV website (https://talk.ictvonline.org/taxonomy/vmr/). The most recent update is from 2018 (ICTV, 2018). Two orders and 20 families of plant viruses have been recognised by the ICTV (ICTV, 2018). Two families (Alphaflexviridae, 49 species and Potyviridae, 160 species) and one insect virus family (Baculoviridae, 55 species) contain viruses notified to EFSA.

A species is the lowest taxon recognised by the ICTV and is based on a consensus sequence of a mixture of genotypes. An isolate or strain is a mixture of genotypes with certain biological characteristics. In the case of baculoviruses, a species is based on a consensus sequence with > 95% sequence homology (Wennmann et al., 2019).

Protists/Algae taxonomy

For protists/algae taxonomy, the Catalogue of Life (https://www.catalogueoflife.org) and the Global Biodiversity Information Facility (https://www.gbif.org) are used as basis for the assessment.

2.1.2. Body of Knowledge

The body of knowledge is one of the pillars of the QPS evaluation and is investigated based on the scientific literature. This includes peer‐reviewed papers published in journals and books that appear in scientific literature databases. To evaluate if the body of knowledge is sufficient to grant a TU, the QPS status several aspects are taken into account, such as the amount of available scientific knowledge indicating a certain degree of exposure of humans and animals through food and feed use.

Aspects on the ecology of the organism are also taken into account. This includes the distribution of the TU in natural environments (e.g. in the gut of humans, wild and farmed animals, and in the plant ecosystem) and their colonisation ability and routes for dispersal. The body of knowledge includes also the history of use of a TU in the agro‐food chain or in other sectors (e.g. biotechnological or medical applications). For this, information on the direct use of viable cells (e.g. as feed additives, food starter cultures, novel foods, probiotic or plant protection products), the use for production purposes (e.g. production of amino acids, biomass, enzymes, vitamins and polysaccharides) or its use in biotechnological or medical applications is examined. When detection in food or feed microbial community is reported, its presence as spontaneous contaminant vs. as main fermentative agent is considered.

2.1.3. Safety concerns in relation to pathogenicity and virulence

TUs assessed for the QPS list should not represent a hazard to human and animal health when used in the food or feed chain.

Relevant information includes case reports of human diseases, particularly infections or human intoxications linked to the TU under assessment. Additional important information is whether any negative impacts are confined to affected persons with conditions favouring opportunistic infections, for example, immunosuppression, and whether transmission occurred through food or other routes (e.g. medical devices). Studies indicating the presence of virulence factors (e.g. toxins and enzymes that may contribute to the pathogenicity of the microorganism) in the TU are also relevant for identification of potential safety concerns.

Several of the QPS‐TUs (e.g. Bifidobacterium species, Lactobacillus and Saccharomyces boulardii cerevisiae) are sporadially reported as causing infections in individuals with recognised predisposing conditions for the acquisition of opportunistic infections e.g. cardiovascular conditions favouring endocarditis, populations in the extreme lower or upper age spectrum or other conditions which can lead to impairment of the immunological system, such as patients submitted to transplants, undergoing cancer therapy, with physical trauma or tissue damage, or HIV patients. Moreover, gastrointestinal tract‐related conditions with mucosal impairment can also be predisposing factors for infections. Previous use of the microorganisms as food supplements for humans was reported in many of these cases. The living microorganism used as a food supplement does not fall under the remit of the QPS assessment. Nevertheless, QPS assessment will continue to take into consideration these reports, extracting relevant information whenever justified.

Assessment of allergenicity to microbial residual components is beyond the QPS assessment remit; nevertheless, if there is science‐based evidence for some microbial species related to well‐defined clinical cases, this is taken into consideration. Although general human safety is part of the evaluation, specific issues connected to exposure of users handling the product (e.g. dermal, inhalation, ingestion) are not addressed.

Reports of infection, intoxication or other diseases caused by the assessed TU on livestock domesticated and wild animals are also a relevant set of information for identifying potential safety concerns. As with safety concerns for humans, whether diseases are acquired through exposure via feed or other routes (e.g. wounds, inhalation) is also relevant information.

2.1.4. Safety for the environment

The assessment of environmental safety considers information on the natural presence of the TU in the microbiota of humans and animals, and the wider environment, and if its use is expected to pose additional risks to these different environments.

3.1.2. Corynebacterium ammoniagenes

This is a new taxonomic unit evaluated from notifications received from October 2016 and now included in the QPS list. The evaluation was published in a recent BIOHAZ Panel Statement (EFSA BIOPHAZ Panel, 2019b).

Corynebacterium ammoniagenes is a species with standing in nomenclature. It was first described by Cooke and Keith (1927) as Brevibacterium ammoniagenes, a urea‐splitting bacterium isolated from the human intestinal tract. It was transferred to the genus Corynebacterium by Collins (1987).

C. ammoniagenes is used for the industrial production of nucleotides, nucleosides and riboflavin (Koizumi et al., 2000; Serrano et al., 2017). C. ammoniagenes derived single‐cell protein can also be used as a non‐conventional protein source in animal diets, such as for broilers (An et al., 2018) and growing pigs (Wang et al., 2013), without any negative effects on blood, bone characteristics or meat quality (An et al., 2018).

No information was found in relation to pathogenicity of the organism, including when it was used as single cell protein for feed (Oliveira et al., 2017).

C. ammoniagenes can be recommended for the QPS list with the qualification ‘for production purposes only’.

3.1.3. Corynebacterium glutamicum

A search for papers potentially relevant for the QPS consideration of Corynebacterium glutamicum provided 246 references. The analysis of their titles left 11 articles. Only one paper reached the final selection phase (full text), but it did not deal with any safety concerns, so the QPS status of C. glutamicum is not changed.

In parallel to the standard procedure for assessing a TU for a possible QPS status and for the maintenance of a QPS status, it was decided to run a complementary reassessment for another specific end use of this TU as the QPS qualification ‘only applies when the species is used for amino acid production’ (EFSA BIOPHAZ Panel, 2019b). The QPS status of C. glutamicum is confirmed with the qualification extended to other production purposes.

3.1.4. Lactobacillus species

Lactobacillus spp.

A search for papers potentially relevant for the QPS consideration of any of the 37 Lactobacillus species included in the list provided 3040 references. The analysis of their titles left 226 articles and of their titles/articles left 65 articles. Revision of their full texts allowed selection of 31 reports that raised safety concerns, all of which described human pathological processes. The claimed aetiological agents comprise L. acidophilus (Cohen et al., 2016; Haghighat and Crum‐Cianflone, 2016; Hubbard et al., 2018), L. animalis (Somayaji et al., 2016), L. casei (Passera et al., 2016; Vanichanan et al., 2016; Pailhories et al., 2017; Stroupe et al., 2017; de Seynes et al., 2018), L. coryniformis (Datta et al., 2017), L. delbrueckii (Chaini et al., 2016; Maillet et al., 2018), L. gasseri (Chaini et al., 2016; Elikowski et al., 2017; Esquibel et al., 2017), L. paracasei (Harding‐Theobald and Maraj, 2018; Kao et al., 2018; Kato et al., 2016; Pararajasingam and Uwagwu, 2017), L. plantarum (Biesiada et al., 2018), L. rhamnosus (Felekos et al., 2016; Molinaro et al., 2016; Aaron et al., 2017; Norena et al., 2017; Boumis et al., 2018; Kane et al., 2018; Koyama et al., 2018; Naqvi et al., 2018; Nayeem et al., 2018; Zeba et al., 2018) and L. salivarius (Garcia Carretero et al., 2018; Wang et al., 2017).

The pathological processes reported included endocarditis (10 cases), bacteraemia/sepsis (8 cases), abdominal, including liver, abscesses (4 cases), pulmonary infection (4 cases), meningoencephalitis, spondylodiscitis and prosthetic joint, urinary tract and genital infections (one case each).

In three reports (Hubbard et al., 2018; Kao et al., 2018; Nayeem et al., 2018), no information was provided on how the causal microorganism was undet, while in several others (Cohen et al., 2016; Haghighat and Crum‐Cianflone, 2016; Elikowski et al., 2017; Biesiada et al., 2018; Harding‐Theobald and Maraj, 2018; de Seynes et al., 2018; Zeba et al., 2018), phenotypical methods for identification were used, which are not considered completely reliable for lactobacilli.

The articles involved single cases of infection in patients that suffered from predisposing illnesses such as metastatic lung (Biesiada et al., 2018) and pancreas (Nayeem et al., 2018) tumours, congenital heart problems (Norena et al., 2017), haemorrhagic telangiectasia (Boumis et al., 2018), alcoholic cirrhosis (Harding‐Theobald and Maraj, 2018), anastomotic leak from bariatric surgery (Garcıa Carretero et al., 2018), chronic obstructive pulmonary disease, caries and uncontrolled diabetes (Hubbard et al., 2018; Pailhoriès et al., 2017) or were immunocompromised due to untreated AIDS and cirrhosis (Haghighat and Crum‐Cianflone, 2016), prematurity (Molinaro et al., 2016) or had received a bone marrow transplant (Koyama et al., 2018).

Based on the available evidence as described above (the safety concerns identified were considered to be linked to severe underlying health conditions or to immunocompromised people or had methodological problems in the identification of the strain), the QPS status of the lactobacilli involved in the reported cases and, by extension, of all others included in the QPS list, is not changed.

Lactobacillus animalis

This is a new taxonomic unit evaluated from notifications received from October 2016 and now included in the QPS list. The full evaluation was published in a previous BIOHAZ Panel Statement (EFSA BIOHAZ Panel, 2017).

Lactobacillus animalis strains have been isolated from the oral cavity and the gastrointestinal tract of animals (Dent and Williams, 1982) and from kimchi, a traditional Korean fermented vegetable dish (Nam et al., 2011). The strain type is L. animalis NCDO 2425. L. animalis is an obligate homofermentative organism that produces mainly L‐(+) lactic acid and is closely related to Lactobacillus acidophilus and L. ruminis. DNA base composition is between 41.3 and 44.4% G + C. A search for the body of knowledge on L. animalis was undertaken in the Web of Science Core collection (search strings in Appendix A) and a total of nine papers were retrieved, with one considered relevant for QPS (Dent and Williams, 1982). Another literature search was performed in PubMed, 289 papers were identified, from which nine were further considered.

L. animalis strain TMW 1.971 has been shown to improve the water holding and gas retention ability of gluten‐free doughs by the production of exopolysaccharides (Rühmkorf et al., 2012, 2013). Effects of L. animalis strain LA4 on the composition and the metabolism of the intestinal microbiota in dogs indicate that it might be considered as a potential probiotic for dogs (Biagi et al., 2007). L. animalis DPC6134 (Hayes et al., 2007) generated peptides with angiotensin‐converting enzyme inhibitory activity from bovine caseinate containing media, with the potential to reduce blood pressure and antihypertensive effects. Bacteriocin production has been characterised for L. animalis strain TSU4 (Sahoo et al., 2015).

The genome sequence of L. animalis P38, isolated from the caecal content of chickens (Rezvani et al., 2016), of L. animalis 381‐IL‐28, a component of a multistrain commercial food biopreservative (Sturino et al., 2014) and of L. animalis KCTC 3501, isolated from kimchi (Nam et al., 2011), have been determined and not found to harbour any genes encoding known virulence factors.

A case of chronic hip prosthetic joint infection caused by L. animalis has been described (Somayaji et al., 2016). This occurred in a 70‐year‐old patient, 5 years after a transient bacteraemia by the same organism as deduced through whole genome sequencing of both causal agents. The patient presented a medical history of type 2 diabetes mellitus and pancreatic cancer.

The species L. animalis is a component of the bacterial communities that colonise the oral cavity and gastrointestinal tract of diverse animal species, and is also commonly used as a starter for fermented vegetables. A single case of human infection by the organism has been reported, but it was linked to life‐compromising predisposing factors. It is therefore concluded that L. animalis does not pose a health risk for the consumer. Consequently, QPS status can be granted for this species.

Lactobacillus parafarraginis

This new taxonomic unit was evaluated from notifications received since October 2016 and is now included in the QPS list. The full evaluation has been published in a previous BIOHAZ Panel Statement (EFSA BIOHAZ Panel, 2020).

Lactobacillus parafarraginis is a valid species name according to the List of Prokaryotic Names with Standing in Nomenclature. It was first described upon isolation from Shochu compost (shochu is a sake‐derived distilled beverage) (Endo and Okada, 2007) and belongs to the L. buchneri group of heterofermentative lactobacilli (Salvetti et al., 2018).

L. parafarraginis is used in the fermentation of food and feed. Consequently, it is frequently consumed by humans and livestock. There are no reports of safety concerns.

L. parafarraginis is recommended for inclusion in the QPS list.

3.1.5. Lactococcus lactis

A search for papers potentially relevant for the QPS consideration of Lactococcus lactis provided 958 references. Analysis of their titles left 29 articles and of their abstracts, 17. Full text revision of these allowed the selection of 10 papers that raised safety concerns.

Of these, two dealt with animal infections; one reported bovine mastitis (Rodrigues et al., 2016) and the other involved farmed Alosa‐Alosa fish (Wünnemann et al., 2018). The remaining papers described pathological processes in humans; four related to endocarditis (Mansour et al., 2016; Chen et al., 2018; Georgountzos et al., 2018; Tato Rodriguez et al., 2018), two to oral lesions (Kabore et al., 2018; Mussano et al., 2018) and the other two to liver abscesses (Fragkiadakis et al., 2017) and cholangitis (Shimizu et al., 2019). In two of the reports (Fragkiadakis et al., 2017; Georgountzos et al., 2018), no indication is provided on how the identification of the microorganism was done, while in the other two (Chen et al., 2018; Kabore et al., 2018), phenotypical methods for identification of the aetiological agent, which are known not to be reliable for L. lactis, were used. Finally, Mansour et al. (2016) report repetitive negative blood cultures and a single PCR‐positive determination, which suggests contamination rather than aetiology. Rodrigues et al. (2016) found a higher proportion of L. lactis in milk from mastitic than from healthy cows which, in itself, does not imply causality. The two oral inflammation articles described polymicrobial infections and concomitant isolation of well‐known pathogens, which make doubtful the L. lactis aetiology of the lesions. Predisposing conditions for opportunistic infection were detected in several of the reports dealing with human infections; these comprise pernicious anaemia, autoimmune atrophic gastritis and severe periodontitis (Fragkiadakis et al., 2017), previous cerebral haemorrhage, coronary heart disease and Alzheimer's disease (Chen et al., 2018), valve replacement and aortocoronary by‐pass (Tato Rodriguez et al., 2018) and insertion of a catheter to allow bile secretion after blocking by a cholangiocarcinoma (Shimizu et al., 2019). Finally, the fish studied by Wünnemann et al. (2018) had been recently captured from the wild and kept in an overpopulated tank under very low oxygen concentrations, conditions described by the authors as very stressful.

Based on the available evidence as described above (all safety concerns identified were considered linked to severe underlying health conditions or to immunocompromised status people or had methodological problems in the identification of the strain), the QPS status of L. lactis is not changed.

3.1.6. Leuconostoc species and Microbacterium imperiale

The search for papers potentially relevant for the QPS consideration of Leuconostoc and Microbacterium imperiale provided 406 references. The analysis of their titles left 42 articles. From these, 19 papers were screened for a possible safety concern. After screening the entire papers, four of them were discarded because they did not deal with safety concerns, did not concern this TU or were not in English. Twelve were analysed in detail for information on the potential safety concern identified.

Leuconostoc species

Twelve articles described safety concerns. Four papers on Leuconostoc mesenteroides dealt with nosocomial infections of patients that suffered predisposing conditions. Franco‐Cendejas et al. (2017) refer to a case of acute infection of a knee prosthesis associated with L. mesenteroides 3 years after surgery. The isolated strain was identified using both phenotypic tests and molecular analyses. The authors proposed that the patient's previous upper respiratory tract infection, which caused hyperpermeability and the subsequent bacterial entrance into the bloodstream, may be the origin of the L. mesenteroides infection. In another case, L. mesenteroides was isolated from the blood of a 50‐day‐old baby hospitalised with diarrhoea, and presenting with catheter‐related septicaemia. In this latter study, only a phenotypic identification procedure was performed (Karbuz et al., 2017). For another paper (Ananieva et al., 2017), the identification was achieved using biochemical methods, and the last article concerned the AMR of a small number of L. mesenteroides strains (Cai et al., 2017).

Two cases of L. pseudomesenteroides were described – one catheter‐related sepsis in which the patient was successfully treated with antibiotic lock therapy (Ho et al., 2016) and another paper on bacteraemia in a patient with acute lymphoblastic leukaemia (Ino et al., 2016) without any indication of the identification procedures.

Another case involved a 44‐year‐old woman with acute myeloid leukaemia under myelosuppression, who had bacteraemia caused by L. lactis (Matsuda et al., 2017).

All safety concerns were considered to be linked to severe underlying health conditions, or patients were immunocompromised, or there were methodological problems in the identification of the strain, and therefore, there is no need to change the QPS recommendation of L. pseudomesenteroides and of other Leuconostoc species included in the QPS list.

Microbacterium imperiale

No articles progressed to the level of screening for potential safety concerns for Microbacterium imperiale, so no new safety concerns were identified. Consequently, the QPS status of M. imperiale is not changed, and nor is the qualification ‘QPS applies for production purposes only’.

3.1.7. Oenococcus oeni and Pasteuria nishizawae

The search for papers potentially relevant for the QPS consideration of Oenococcus oeni and Pasteuria nishizawae provided 221 references. The analysis of their titles left 11 articles for further consideration. One, related to O. oeni, was analysed in full with reference to the potential safety concern identified.

Oenococcus oeni

As described above, one article was analysed in detail, but no potential safety concern was identified. Consequently, the QPS status of O. oeni is not changed.

Pasteuria nishizawae

As described above, no article arrived to the final screening stage, so no new safety concerns were identified. Consequently, the QPS status of P. nishizawae is not changed.

3.1.8. Pediococcus species

The search for papers potentially relevant for the QPS consideration of Pediococcus spp. provided 820 references. Analysis of their titles left 20 articles and of their abstracts, 8. After screening the papers that emerged from the different steps considered, only three papers (Han et al., 2016; Chen et al., 2018; Thumu and Halami, 2019) were analysed in full for potential safety concerns, but they did not relate to food‐borne infections, and/or had methodological problems in the identification of the isolate, so they did not give rise to any new safety concerns. Consequently, the QPS status of Pediococcus spp. is not changed.

3.1.9. Dairy propionic acid bacteria – Propionibacterium species

The search for papers potentially relevant for the QPS consideration of Propionibacterium spp. provided 197 references. The analysis of their titles left three articles. After screening the entire papers, only one paper (Giok, 2016) reached the final selection phase, but it does not refer to a food‐borne disease and had methodological identification problems. Consequently, the QPS status of Propionibacterium spp. is not changed.

3.2.2. Paenibacillus illinoisensis

This is a new taxonomic unit that is now included in the QPS list. The evaluation was published in a previous BIOHAZ Panel Statement (EFSA BIOPHAZ Panel, 2019a).

Paenibacillus illinoisenis, previously known as Bacillus circulans, group 6, was described by Shida et al. (1997). It is a valid species with standing in nomenclature.

P. illinoisensis was isolated from the rhizosphere of soil and characterised for its siderophore‐producing capacity, promoting iron absorption by plants (Liu et al., 2017a,b). Strains of P. illinoisensis were reported to secrete cyclodextrin gluconotransferase (Doukyu et al., 2003; Lee et al., 2013, chitinases (Jung et al., 2008) and enzymes degrading methane (Jhala et al., 2014).

P. illinoisensis can be recommended for QPS with the specific qualifications for production purposes and absence of toxigenic potential.

3.3.2. Komagataeibacter sucrofermentans

This is a new taxonomic unit that is now included in the QPS list. An evaluation was published in a previous BIOHAZ Panel Statement (EFSA BIOPHAZ Panel, 2019a).

The bacterial species K. sucrofermentans (Validation List nr. 149, IJSM 2013, 63, 1‐5) was previously named Acetobacter xylinus subsp. sucrofermentans (Toyosaki et al., 1996) and Gluconacetobacter sucrofermentans (Cleenwerck et al., 2010). The species is clearly described based on a polyphasic approach (Cleenwerck et al., 2010).

K. sucrofermentans strains are characterised by their ability to produce large amounts of cellulose from sucrose in agitated cultures (Cleenwerck et al., 2010). Searching PubMed database for this species delivered 11 hits, all concerning the cellulose production capacity. In Asia, cellulose has traditionally been produced from the fermentation of coconut waste water by K. sucrofermentans, and used in food.

K. sucrofermentans (A. xylinus subsp. sucrofermentans) can be proposed for the QPS list with the qualification ‘QPS applies for production purposes only’.¹⁶

3.4.1. Candida cylindracea

C. cylindracea belongs to the Ogataea clade of the Ascomycetous yeasts (Kurtzman et al., 2011; Daniel et al., 2014). The species was described by Yamada and Machida (1962) and validated by Meyer and Yarrow (1998). No synonym names have been used. Only the anamorphic form is known and described. The type strain for C. cylindracea – CBS 6330 – is also marketed under other designations, e.g. DSMZ 2031 (online) and ATCC 14930 (online). Unfortunately, in the literature on lipase‐producing yeasts, the C. cylindracea type strain has at times been referred to as Candida rugosa (e.g. Benjamin and Pandey (1998); Takaç et al. (2010)). This has caused some confusion since C. cylindracea and C. rugosa are two well‐defined species, not closely related phylogenetically (Kurtzman et al., 2011). It is also unfortunate since C. rugosa is considered an emerging, opportunistic yeast (Miceli et al., 2011). However, identification according to molecular methods can easily separate the two species. It is therefore recommended that the species identity of lipase‐producing strains of Candida is confirmed by using such methods.

No references related to possible concerns for human or animal safety, or other related aspects, were identified, and therefore, the QPS status is not changed. The qualification ‘QPS only applies when the species is used for enzyme production’ is also unchanged.

3.4.2. Debaryomyces hansenii

The anamorph form of D. hansenii is Candida famata. Based on Nguyen et al. (2009), Betancourt et al. (2016) concluded that C. famata should not be considered the anamorphic form of D. hansenii. However, Nguyen et al. proposed that a special riboflavin overproducing form of C. famata, also referred to as Candida famata var. flareri, did not belong to the same species as D. hansenii, while C. famata var. famata should be considered to belong to the same species as D. hansenii. Additionally, Kurtzman et al. (2011) supported the view that C. famata should be considered the anamorph of D. hansenii. Since clinical reports usually do not distinguish between the varieties of C. famata/D. hansenii, it is not possible to say to what extent the flareri variety might be involved. But obviously there is reason to closely follow developments in the taxonomy of D. hansenii and related species.

Seventeen references related to possible safety concerns or relevant aspects of the body of knowledge were identified in the ELS, of which 16 used the name C. famata and one the name D. hansenii.

Most of the studies had some type of methodological problem identified and thus had to be given less weight in the evaluation. Most often there were weaknesses in the methods used for species identification.

Taverna et al. (2019) compared different methods to further identify 38 clinical yeast isolates, tentatively identified as C. famata/D. hansenii by morphological and physiological growth tests methods, from a collection in Argentina. Three different molecular methods and MALDI‐TOF MS were largely convergent and showed that more than half of the 38 isolates actually did not belong to the D. hansenii complex, but to the Candida guilliermondii complex. This suggests that studies relying on morphological and biochemical/physiological identification methods are likely to exaggerate the frequency of D. hansenii in human infections, and reinforces the need to use molecular approaches for correct identification of this species. Karapetsa et al. (2019) claim to be the first to report septic shock due to D. hansenii in an immunocompetent subject, although the patient was characterised as showing ‘immunoparalysis’. The young male had serious injuries after a car accident and was admitted to an intensive care unit. Predisposing conditions included a central venous catheter, recurrent bacterial infections and prolonged use of antibiotics. The patient recovered from the fungal infection after treatment with amphotericin B.

Sonmez and Erbas (2017) identified Candida spp. isolates involved in mycotic mastitis in cattle and investigated susceptibility to antimycotics. Three isolates (12%) were identified as D. hansenii, and all were susceptible to ketoconazole but resistant to fluconazole, miconazole and amphotericin B. However, species identification was only based on physiological growth tests and is therefore uncertain. Lo et al. (2017) found that four isolates of D. hansenii from fresh fruit were susceptible to both the antimycotics tested, fluconazole and triadimenol. Taverna et al. (2019) reported that their eight isolates of D. hansenii all had comparatively low MIC values (i.e. high susceptibility) to all eight tested antimycotics. Espinel‐Ingroff et al. (2019) present MIC distributions and epidemiological cut‐off values for D. hansenii for four triazole antimycotics.

In conclusion, relatively few studies reported isolation of D. hansenii in clinical samples, and in most of them, the species identification was uncertain. No studies reported infection in humans without predisposing factors. In retrospective studies of clinical isolate collections, the prevalence of D. hansenii was generally low compared to other yeasts. No studies reported any increased prevalence of antimycotic resistance. In conclusion, no information was obtained to indicate the need for a change in QPS status.

3.4.3. Hanseniaspora uvarum

The anamorph form of H. uvarum is Kloeckera apiculata. The species name has not been changed since the 2016 QPS opinion.

One reference relating to possible concerns for human or animal safety, or other related aspects, was identified.

Siavoshi et al. (2018) claim that common food‐borne yeasts, among them H. uvarum, can harbour the bacterial pathogen Helicobacter pylori as an intracellular parasite or commensal. Thereby yeasts might possibly function as a vector for transmission of this bacterium. Several earlier studies on possible H. pylori association with yeasts have been published by the same research group, but no studies on this subject by any other laboratory or group could be found. While some studies have reported an association of H. pylori with free‐living amoebae (reviewed by Quaglia and Dambrosio, 2018), more evidence is needed before any firm conclusions can be drawn about the possibility that food‐borne yeasts might function as a vector for H. pylori.

No references indicating any new possible concerns for human or animal safety, or other related aspects, were identified. Therefore, the QPS status of H. uvarum is unchanged.

3.4.4. Kluyveromyces lactis

The anamorph form of K. lactis is Candida spherica. The species name has not been changed since the 2016 Opinion.

No references related to possible concerns for human or animal safety, or other related aspects, were identified. Therefore, its QPS status is not changed. The qualification ‘for production purposes only’ is also unchanged.

3.4.5. Kluyveromyces marxianus

The anamorph form of K. marxianus is Candida kefyr. The species name has not been changed since the 2016 Opinion.

In total, 40 references related to possible safety concerns, or relevant aspects of the body of knowledge, were identified in the ELS. All 40 studies were retrieved when using the anamorph name C. kefyr. The majority of these 40 studies had some type of methodological problem identified and thus had to be given less weight in the evaluation. Most often there were weaknesses in the methods used for identifying the microorganism, or a lack of information regarding the source attribution, or uncertainty about whether or not an infection had actually been diagnosed.

The ability of K. marxianus/C. kefyr to cause opportunistic infections in humans with predisposing disease conditions has received increased attention in recent years. For instance, Charsizadeh et al. (2018a,b,c) reported the isolation of K. marxianus from patients with suspected candidiasis in neonatal and paediatric intensive care units, although the prevalence of K. marxianus compared to other fungal species was only 1–2%. Jahanshiri et al. (2018) reported that K. marxianus constituted 5% of the isolates from cancer patients with oropharyngeal candidiasis in a hospital in Iran. Ghajari et al. (2018) reported that one out of 31 yeast isolates from women with suspected vulvovaginal candidiasis was K. marxianus. This is a low prevalence, it is not entirely clear that this isolate caused an infection, and there is no information regarding whether or not there were any predisposing factors. The review of Benedict et al., 2016 lists an earlier study by Pineda et al. (2012), which reports a case of bloodstream infection in a pregnant woman and her twin fetuses. All three survived the infection after successful antimycotic therapy. The mother became pregnant by in vitro fertilisation and delivered the twins by caesarean section at 29 weeks of gestation. The two bloodstream isolates (from the mother and one of the twins) were identified as two different strains of K. marxianus by ITS sequencing and RAPD. No further details on the identification were reported. The mechanism by which the yeast gained access to the placenta and bloodstreams is not clear, and a link to intake of dairy products containing K. marxianus is only hypothesised.

A few studies reported antimycotic susceptibility of clinical K. marxianus isolates. Nagy et al. (2018) demonstrated differences in susceptibility of planktonic and biofilm cells of K. marxianus to several antimycotics. Farmakiotis and Kontoyiannis (2017) reviewed antimycotic non‐susceptibility in pathogenic and opportunistic yeasts and concluded that there is an increasing trend for resistance in K. marxianus. Another review focused on multidrug resistance in pathogenic Candida spp. (Colombo et al., 2017), but also highlighted K. marxianus as an emerging opportunistic pathogen. Salse et al. (2019) presented epidemiological cut‐off values for K. marxianus that may be useful to identify isolates with potential resistance to antimycotics. Espinel‐Ingroff et al. (2019) present MIC distributions and epidemiological cut‐off values for D. hansenii for four triazole antimycotics. One study reported divergent results of the E‐test and Vitek 2 methods for determining antimycotic susceptibility in K. marxianus (Alfouzan et al., 2017), although there were methodological uncertainties regarding how the selection and identification of the studied isolates were performed. Reales‐Calderon et al. (2016) reviewed information regarding the acquisition and development of resistance to antimycotics in yeasts and other fungi, although no specific information was given regarding K. marxianus.

Karstrup et al. (2017) identified a yeast‐like organism in uterine lavage fluids from three cows with slight signs of inflammation but no mastitis, as K. marxianus. The organism could not be cultured but was identified using PCR and sequencing.

There is no doubt that K. marxianus/C. kefyr can behave as an opportunistic fungus in humans with predisposing factors. Reports where it has been unambiguously shown that food intake of K. marxianus is the cause of infectious disease in otherwise healthy individuals do not exist. Therefore, its QPS status remains unchanged.

3.4.6. Komagataella pastoris

The anamorph of K. pastoris is not described. The previous name of this species is Pichia pastoris. The species name has not been changed since the 2016 QPS opinion.

No references related to possible concerns for human or animal safety, or other related aspects, were identified. Therefore, its QPS status does not change. The qualification ‘QPS only applies when the species is used for enzyme production’ is unchanged.

3.4.7. Komagataella phaffi

This is a new taxonomic unit, evaluated from notifications received since October 2016 and now included in the QPS list. The evaluation was published in a previous BIOHAZ Panel Statement (EFSA BIOHAZ Panel, 2018a).

The anamorph of Komagataella phaffii is not described. K. phaffii is closely related to Komagataella pastoris, a species with a QPS status, from which it was separated (Kurtzman, 2005). The three species of the genus Komagataella, K. pastoris, K. phaffii and K. pseudopastoris show no differences in standard fermentation and growth tests. Consequently, it is recommended that the species be separated based on differences in D1/D2 26S rRNA gene sequences or on differences in restriction patterns of SSU rRNA (Kurtzman et al., 2011).

In total, 24 studies were identified (see Appendix A) and screened, dealing with the properties of the species as a protein expression and model organism. K. phaffii is a sibling species of K. pastoris (Naumov et al., 2013). In the literature, it has been described as being used for the same purpose as K. pastoris; that is for the production of heterologous proteins (Chessa et al., 2017).

There is very little information about the ecology of K. phaffii, but at least some strains have a similar ecology to K. pastoris, since both species have been isolated from sap fluxes in trees (Kurtzman et al., 2011).

There is no information available about any potential safety concerns regarding K. phaffii. However, reports on the safety of K. pastoris as production organism also have relevance for K. phaffii because this was changed, on the basis of taxonomic position, to the species K. phaffii.

The species K. phaffii, a sibling species of K. pastoris, can be recommended for the QPS list with the qualification ‘QPS only when the species is used for enzyme production’.

3.4.8. Lindnera jadinii

The anamorph form of L. jadinii is Candida utilis. Synonyms of this species are Pichia jadinii, Hansenula jadinii and Torulopsis utilis. The species name has not been changed since the 2016 QPS opinion.

Three references related to possible safety concerns or relevant aspects of the body of knowledge were identified in the ELS. All three studies were retrieved when using the anamorph name C. utilis.

Two studies reported L. jadinii in human clinical samples. Yagmur et al. (2016) isolated yeasts from postmortem specimens from 1309 cases with suspected fungal infection in Turkey. L. jadinii was reported in low prevalence (three isolates, =3%). Identification was only by physiological and morphological properties and it is very uncertain whether the putative L. jadinii strains actually caused infection. In a retrospective study by Kim et al. (2016), a relatively high number of ‘Candida’ isolates (304 isolates, =12%) from clinical samples from patients in a hospital in South Korea were L. jadinii. All patients had underlying disease and the most common source of L. jadinii was a urinary catheter.

Watanasrisin et al. (2016) characterised the ABC‐transporters in L. jadinii. These transporters can be involved in the development of resistance to antimycotics, and knowledge about their structure and function can facilitate the development of novel antimycotic substances.

Few studies reported isolation of L. jadinii/C. utilis in clinical samples and no studies reported infection in humans without predisposing factors. Prevalence was generally low compared to other yeasts isolated from collections of clinical samples. No studies reported increased prevalence of antimycotic resistance. In conclusion, no information was retrieved to indicate a change in the QPS status, nor in the qualification ‘QPS only when the species is used for enzyme production’.

3.4.9. Ogataea angusta

The anamorph form of O. angusta is not described. A synonym of this species is Pichia angusta. The species name has not been changed since the 2013 QPS opinion.

No references related to possible concerns for human or animal safety, or other related aspects, were identified. Therefore, its QPS status, and the qualification ‘QPS only when the species is used for enzyme production’ is unchanged.

3.4.10. Saccharomyces cerevisiae/species

The anamorph form of S. cerevisiae is not described. A synonym of this species is Saccharomyces boulardii. The species name has not been changed since the 2016 QPS opinion.

In total, five references were identified in the ELS, after exclusion of the ones from which methodological problems were identified. From these five references, only one described an infection associated with S. cerevisiae, an osteomyelitis after surgical reconstruction following serious physical injury of an arm in an adult when working in a bakery. The infection was cured after antimycotic treatment (Seng et al., 2016, ELS 3245).

Espinel‐Ingroff et al. (2019) presents MIC distributions and epidemiological cut‐off values for S. cerevisiae for four triazole antimycotics. Pérez‐Cantero et al. (2019) evaluated the in vitro activity of nine antifungal compounds against S. cerevisiae and they also studied in vivo efficacy of the three antifungals showing the highest in vitro activity by using a murine model of systemic infection.

These new reports of Saccharomyces cerevisiae did not add any new information that would change the QPS status of this species. These new reports also confirm the previous qualifications, that the consumption of Saccharomyces boulardii (synonym of S. cerevisiae) by patients with fragile health may be considered as the possible origin of the infection, although the use of microorganisms intended to be used as ‘probiotic’ for humans as a health claim does not fall under the remit of the QPS assessment. These new reports also confirm the previous QPS qualifications; the absence of resistance to antimycotics used for medical treatment of yeast infections in cases where viable cells are added to the food or feed chain and inability to grow above 37°C. Therefore, its QPS status, and the qualifications ‘Absence of resistance to antimycotics used for medical treatment of yeast infections in cases where viable cells are added to the food or feed chain’ is not changed. In the case of Saccharomyces cerevisiae, this qualification applies for yeast strains able to grow above 37°C.

3.4.11. Schizosaccharomyces pombe

There are no synonym names in common use for this species, and the name has not been changed since 2016.

No references related to possible concerns for human or animal safety, or other related aspects, were identified. Therefore, its QPS status and qualification are unchanged.

3.4.12. Wickerhamomyces anomalus

The anamorph form of W. anomalus is Candida pelliculosa. Synonyms of this species are Hansenula anomala, Pichia anomala and Saccharomyces anomalus. The species name has not been changed since the 2016 QPS opinion.

Nine references related to possible concerns for human or animal safety, or other related aspects, were identified. Seven references related to possible safety concerns, or relevant aspects of the body of knowledge, were selected for further investigation. These papers reported proper identification of the W. anomalus/C. pelliculosa strain.

The studies reported W. anomalus/C. pelliculosa in human clinical samples with low prevalence in bloodstream infection (Kumar et al., 2017; Liu et al., 2017a,b; Suhr et al., 2017; Tan et al., 2016) or in immunosuppressed patients or in intensive care units (Fernandez‐Ruiz et al., 2017; Jung et al., 2017; Tejan et al., 2017).

W. anomalus was always a minor fraction of the isolates, and there were no indications that exposure may have been through the food‐borne route. No studies reported infection in healthy, non‐hospitalised subjects or signs of increased prevalence of antimycotic resistance. Therefore, its QPS status is not changed. The qualifications ‘QPS only applies when the species is used for enzyme production’ and ‘Absence of resistance to antimycotics used for medical treatment of yeast infections in cases where viable cells are added to the food or feed chain’ are also unchanged.

3.4.13. Yarrowia lipolytica

This is a new taxonomic unit evaluated from notifications received since October 2016, and now included in the QPS list. The evaluation was published in a previous BIOHAZ Panel Statement (EFSA BIOHAZ Panel, 2018b).

Yarrowia lipolytica and Candida lipolytica are the teleomorph and anamorph names of the same species.

The species is widespread in nature, as seen from the sources of isolation. Substrates high in lipids are a common source of this species. Y. lipolytica has several physiological properties of industrial significance. The species is well known for the production of proteases and lipases. It is also a widely reported contaminant in dairy and meat products, and is common on raw poultry products and in traditional sausages (Kurtzman et al., 2011).

Hazen (1995) paid special attention to C. lipolytica, since the species had been implicated as the cause of human infection in one study, and it was therefore considered an ‘emerging pathogen’. Pfaller and Diekema (2004) reported four C. lipolytica isolates among 6,082 Candida spp. isolates (i.e. less than 0.1%) from human bloodstream infections. Shin et al. (2000) reported a temporal outbreak of hospital‐acquired infections with C. lipolytica in five patients during 3 months in a paediatric ward in a hospital in Korea. All patients had suppressed immune systems and all but one had a catheter fitted. All recovered from the infection after chemotherapy. Tumbarello et al. (1996) reported the presence of C. lipolytica in one out of 64 HIV patients with oral candidiasis. A comprehensive review by Groenewald et al. (2014) concluded that all described human infections with this species have occurred in immunocompromised patients with underlying disease, and that the majority were catheter‐related. Antifungal therapy invariably resulted in clearance of the pathogen. In the last 5 years, several reports have confirmed that C. lipolytica can behave as an opportunistic pathogen (e.g.: Trabelsi et al., 2015; Abbes et al., 2017; Boyd et al., 2017).

Based on the available information, Y. lipolytica is a commonly occurring species in many habitats/environments. It may behave as an opportunistic pathogen for immunocompromised patients, especially for those who are using catheters. Yarrowia lipolytica is recommended for the QPS list with the qualification ‘QPS applies for production purpose only’.

3.4.14. Xanthophyllomyces dendrorhous

The anamorph form of X. dendrorhous is Phaffia rhodozyma. The species name has not been changed since the 2016 QPS opinion.

No references related to possible concerns for human or animal safety, or other related aspects, were identified. Therefore, the QPS status is unchanged, as is the qualification.

3.5.2. Euglena gracilis

This new taxonomic unit was evaluated following notifications received since October 2016 and it is now included in the QPS list. The evaluation was published in a recent BIOHAZ Panel Statement (EFSA BIOPHAZ Panel, 2019b).

E. gracilis belongs to the genus Euglena, phototrophic euglenoid flagellates possessing complex chloroplasts. The taxonomy of E. gracilis and closely related species has not been amended so far, based on molecular phylogenetic analyses (Zakryś et al., 2017). The whole genome of the E. gracilis strain Z1 was recently sequenced (Ebenezer et al., 2019).

There are many scientific papers published on this TU. E. gracilis is found in many freshwater habitats, especially in shallow eutrophic ponds. The species is able to synthesise biotechnologically relevant compounds such as polyunsaturated fatty acids, vitamins, b‐glucans and tyrosine, used in cosmetics and food supplements. E. gracilis is also used for bioremediation of heavy metals in contaminated water and as a toxicity bioindicator (Krajčovič et al., 2015).

E. gracilis biomass, generally based on dried cells, is used as a feed additive in aquaculture and in animal feed as well as in human food (Suzuki, 2017). Food products containing E. gracilis are marketed in Japan as cookies, cereal bars and nutritional drinks.

Using dried preparations of non‐viable E. gracilis, no genotoxicity was observed in bacterial reverse mutation and mammalian micronucleus tests. Moreover, subchronic toxicity tests in rats did not show any adverse effect and a no‐observed‐adverse‐effect‐level (NOAEL) of 50,000 ppm was determined (Simon et al., 2016). Literature searches did not provide any evidence of any safety concerns for human or animal health related to any use of E. gracilis.

Euglena gracilis may be recommended for the QPS list with the qualification ‘QPS applies for production purposes only’.