Disinfectant testing for veterinary and agricultural applications: A review

Disinfectants for veterinary and livestock use, plus skin antiseptics, are critical elements for the control of infectious agents, including zoonotic and antimicrobial‐resistant micro‐organisms, in managed animal species. Such agents impact animal welfare, economic performance and human health. Testing of disinfectants is needed for safety, efficacy and quality control. The present review examines the principal types of test (carrier, suspension, surface and field) that have been developed or attempted, plus the features inherent in the respective tests, particularly with respect to variability. Elements of testing that have to be controlled, or which can be manipulated, are discussed in the context of real‐world scenarios and anticipated applications. Current national and international testing regimes are considered, with an emphasis on the UK, continental Europe and North America, and with further detail provided in the Supporting Information . Challenges to disinfectant efficacy include: the nature of the biological targets (bacteria, fungi, yeasts, spores, viruses and prions), the need for economical and safe working concentrations, the physical and chemical nature of contaminated surfaces, constraints on contact times and temperatures, the presence of organic soil and other barrier or neutralising substances (including biofilms), and thoroughness of pre‐cleaning and disinfectant application. The principal challenges with veterinary disinfectant testing are the control of test variability, and relating test results to likely performance in variable field conditions. Despite some ambitions to develop standardised field tests for disinfectants, aside from skin antiseptic trials the myriad problems such tests pose with respect to cost, reproducibility and generalisability remain intractable.

• Many real-world factors modulate disinfectant action, and a variety of basic and application-focussed tests are needed to address these.
• There has been some progress in development of international testing standards, but standardised field tests currently remain unachievable.

| Biocides: definitions, use in veterinary settings and the need for testing
The control of micro-organisms affecting production and companion animals, and colonising their environments, is dependent on effective use of biocides. Such organisms include pathogens and antimicrobial-resistant organisms that can be transmitted to humans, either via the food chain or directly, such as Salmonella, Campylobacter and methicillin/meticillin-resistant Staphylococcus aureus. The present review addresses the principles and challenges of disinfectant testing and examines contemporary testing regimes for biocide efficacy in veterinary, farming and related settings.
Disinfectants are applied to inanimate surfaces to rapidly kill or inactivate micro-organisms and sometimes spores. Antiseptics have similar activity but are used on living tissues, and are considered here insofar as they relate to skin hygiene. Disinfectants generally disrupt membrane function and/or interfere with nucleic acids or cytoplasmic components of biological targets, and typically involve multiple target sites (Maillard, 2002;McDonnell & Russell, 1999;Morente et al., 2013).
Beyond basic testing of the biocidal activity of disinfectant agents, appropriate 'in-use' working concentrations and exposure times need to be established, for general and special applications and for operator and environmental safety considerations. In principle, test results should be repeatable within the same laboratory and reproducible in different laboratories. They should also yield results (pass/fail or quantitative values) that relate to anticipated applications. Organic soil is almost universal in livestock environments and such material is strongly associated with attenuation of disinfectant efficacy (Brouillaud-Delattre et al., 1994;Kirk et al., 2003;Luyckx et al., 2017;Steinmann et al., 1995;Wales et al., 2006), but it has markedly differing effects between microbicidal agents (Bessems, 1998).

| Factors affecting disinfectant efficacy in the field
Biofilms comprise extracellular polysaccharide and protein matrices constructed by mono-or multi-species bacterial and/or yeast communities on surfaces and at fluid/gas interfaces. Biofilm matrix presents a diffusion barrier and a neutralising environment for some biocides. This protective effect varies with: substrate surface, age of biofilm, hydration, fluid shear conditions during biofilm formation, and the mix of microbe species in the biofilm (Akinbobola et al., 2017;Buckingham-Meyer et al., 2007;Das et al., 1998;Harding et al., 2014;Howard et al., 2015;Papaioannou et al., 2018). Biofilm-embedded organisms often show less biocide susceptibility than counterparts dried-on to surfaces (Buckingham-Meyer et al., 2007;Maris, 1992;Otter et al., 2015), and they are markedly less susceptible than suspended cells (Cabeça et al., 2012;Harding et al., 2014).
In field situations, gross organic soil, biofilm and post-cleaning residual hydrophobic organic material on surfaces can protect persistent resident pathogens despite regular cleaning and disinfection (C&D), for example in the case of Salmonella in hatcheries. The selection and application of pre-disinfection cleaner can substantially affect disinfectant performance in the face of organic soil and/or biofilm. Cleaning agents (usually detergents) may also show substantial microbicidal activity in their own right (Hancox et al., 2013) and may variously interfere or synergise with the action of a subsequentlyapplied disinfectant .
A post-disinfection drying downtime of one or more days may add substantially to the observed microbicidal effect of C&D procedures, at least for materials where effective drying occurs (Connor et al., 2017;Hancox et al., 2013;Martelli et al., 2017). Furthermore, drying between applications of cleaner and of disinfectant will allow porous surfaces to absorb more disinfectant in the latter stage (Böhm, 1998).

| Types of disinfectant tests
Approaches to disinfectant testing have been reviewed by Reybrouck (1998). Biological targets are usually defined bacteria, yeasts, spores, viruses, or (latterly) prions, or sometimes mixtures of these. The principal types of test (carrier, suspension, capacity, surface and field) are summarised in Table 1. Simple minimum inhibitory or minimum bactericidal concentration (MIC or MBC) multi-well suspension tests can be used to screen for activity (Bloomfield et al., 1994;Roesler et al., 2007) and to avoid approving an in-use concentration that is below the disinfectant's MIC, which can happen with short, 'clean' tests with low nutrient availability and minimal organism adaptation time (DVG, 2017a). Performance can be further assessed using suspension tests, typically at anticipated 'in-use' concentrations, with and without interfering substances added to the solution to mimic organic soil.
Surface tests introduce additional variables, for example: surface characteristics, attachment and changes in viability of target organisms before disinfection, and variation in disinfectant concentration according to local surface features or evaporation. This allows some additional examination of field-type conditions. Surfaces may be inert (such as steel, glass or concrete) or biological, which includes skin or substitutes thereof (important for hand hygiene) or a complex matrix in the case of biofilm. Biofilm is potentially of great significance in infrequently-cleaned and disinfected environments such as animal housing.
Field tests typically have been reported in livestock and human hospital environments. Their value potentially lies in comparative evaluation of differing protocols and agents in similar environments, and in identifying scenarios (environment, surface or equipment) that pose particularly stiff challenges. Furthermore, such tests may identify common pitfalls and practical limits to what can be achieved with certain approaches (Decun et al., 2009;Roesler et al., 2007). However, many factors militate against standardisation in agricultural settings. These include: local water quality, prevailing temperature and relative humidity, varied materials and configurations (structural, flooring and equipment), damaged surfaces, the practicability of comprehensive cleaning, presence of biofilms, the density of bacterial communities, prevailing bacteria and/or viruses, plus drainage and airflow affecting pooling and drying (Burbarelli et al., 2015;Connor et al., 2017;van Klingeren et al., 1998;Mannion et al., 2007;Mueller-Doblies et al., 2010;Pintaric et al., 2011;Tamási, 1995). An alternative approach, of using field-contaminated materials in laboratory testing (Furuta & Yoshizawa, 1997;Huneau-Salaün et al., 2010) has in some cases demonstrated a considerable blunting of biocidal performance compared with standardised test results (Furuta & Yoshizawa, 1997).  (Bloomfield et al., 1994;Cabeça et al., 2012). For some specific applications, other organisms may be chosen for relevance or safety ( Table 2). Details of commonlyemployed yeasts and fungi are also in Table 2.

| K E Y FAC TOR S IN THE DE S I G N AND CONDUC T OF TE S TS
There can be marked variation in the susceptibility of bacterial species, and of strains within a species, to disinfectants under test conditions (Sander et al., 2002). Furthermore, the modulating effects of other test variables on disinfectant activity can vary markedly according to the bacterial target. Examples include the interaction observed between target, water hardness and activity of quaternary ammonium compounds (QAC) in one study (Bessems, 1998), and the effect of organic soil on susceptibility of Salmonella Typhimurium or S. aureus to disinfectants (Stringfellow et al., 2009).

| Viruses
Test viruses ideally should be readily propagated to high titres, be safe to handle, and be robust in the face of disinfectants and (for surface tests) of drying (Krug et al., 2011;Rabenau et al., 2014).
There is much variation in these qualities, with non-enveloped viruses generally being more robust to adverse environmental and chemical conditions, but loss of activity upon drying is still commonly substantial (Krug et al., 2011;Schürmann & Eggers, 1983).
The complexity of propagation, recovery and enumeration of viable virus adds to the challenge of reproducibility (Rabenau et al., 2014).   (Reybrouck, 2007). More susceptible to issues with repeatability than suspension tests Field Attempts to evaluate efficacy in real-world environments with protocols that follow recommended or observed practice. Wide range of possible scenarios. Also testing of field-contaminated materials in laboratory (Furuta & Yoshizawa, 1997;Huneau-Salaün et al., 2010). Targets for recovery may be pathogens (e.g. Salmonella, Staphylococcus aureus), or endogenous or applied indicator organisms (e.g. Enterobacteriaceae, Serratia marcescens). Sampling by swabs, rinses, impression plates, pieces of material, etc.
Often comparative culture of treated and untreated (control) areas. May use molecular biological tools to detect viable target, particularly for viruses. Reduction against control or to a predefined surface density are typical measures Generally, standardisation not currently achievable due to the number of variables and resources needed for repetitive trials . Tight control of variables, even if possible, would limit the generalisability of results in agricultural settings. Hand sanitiser tests may be an exception (Cremieux et al., 1989)

| Preparation of the biological target
The physiological and physical state of the biological target will affect test performance as will any associated material, for example culture medium or biofilm. As previously mentioned, stock strains also appear to alter in susceptibility over time (Bloomfield et al., 1994;van Klingeren, 1995). For suspension tests with bacteria, the use of stationary-phase cultures sedimented and re-suspended in defined media provides targets in reproducible numbers and in a state of limiting nutrition and physiological stress, such as is likely to be encountered in real-world applications.
Organisms dried on to surfaces may show increased resistance to disinfectant action, but they also spontaneously lose viability over a matter of hours, necessitating tight control over the timing of preparation and test procedures (Pines et al., 2013). In addition, the bacterial species and density of inoculum will affect surface distribution patterns and consistency of results, and mechanical spreading tends to reduce the viability of the inoculum (Bloomfield et al., 1994).

| Qualities of diluting water
Water hardness (dissolved calcium and magnesium minerals) can have a substantial effect on interactions between certain combinations of bacteria and biocide (Bessems, 1998), and standardised testing usually specifies deionised/distilled water (EPA, 2014), or water of a defined hardness (APHA, 2016; Bloomfield et al., 1993;CEN, n.d.). Water supplies in the field can have effects on disinfectant efficacy that are either understood, such as the effect of pH on free chlorine (Holah, 1995), or which may be unpredictable and not correlated with usually-measured physical and chemical properties (Davison et al., 1996;Wales et al., 2013).

| Surface qualities
Surface variables include: roughness, porosity, chemical nature and qualities such as hydrophobicity. Such factors need to be controlled, but they do allow testing to address certain real-world scenarios, for example polyethylene drinking water systems or concrete, metal and wood structural material in livestock accommodation (Hancox et al., 2013). Highly standardised surface tests tend to use readilyavailable and non-reactive materials, typically stainless steel or glass, with a defined surface roughness (van Klingeren et al., 1998;Pines et al., 2013).
Results of laboratory studies (Marin et al., 2009;Yilmaz & Kaleta, 2003) and field trials (Connor et al., 2017;Luyckx et al., 2015;Martelli et al., 2017) emphasise the tendency for rough and damaged surfaces to reduce the efficacy of disinfectants. In addition, porous surfaces present a large surface area and possibly enhanced attachment opportunities for target organisms. They too are associated with reduced bactericidal and virucidal effect of disinfectants (Gamal et al., 2014;Jang et al., 2014;Lyutskanov et al., 2010).
Skin is a complex surface type, with studies in this area commonly comparing contamination and recovery techniques on volunteers' hands (Cremieux et al., 1989;Macinga et al., 2011). One ASTM International test (ASTM E2897) employs prepared and sterilised pig skin as the surface material, whereas an artificial skin substitute is used elsewhere (prEN 17422, Table S3).

| Mode of disinfectant application
In standardised surface tests a measured volume of disinfectant is placed on a surface, or applied by using a prescribed wiping or spraying pattern (Bloomfield et al., 1993;EPA, 2014), whereas standard spray-disinfection procedures are usually employed in field tests (Connor et al., 2017;Fankem et al., 2009). With small disinfectant volumes evaporative effects during the tests may be of significance, and larger volumes can enhance the observed biocidal effects (van Klingeren et al., 1998). Therefore appropriate and consistent volumes need to be employed, and in standardised conditions of temperature and humidity.
There have been recent developments of airborne disinfection systems, using either aerosols (fogging or ionised 'dry mist' systems) or disinfectant vapour, commonly intended for disinfection of complex environments such as hospital patient rooms and animal houses (Barbut et al., 2009;Luyckx et al., 2017;Schneider, 2013

| Disinfectant exposure: concentration, time and temperature
The theoretically predictable relationships between killing/inactivation and: exposure time, disinfectant concentration and temperature become complex outside of tightly-controlled suspension tests (van Klingeren, 1995;Mazzola et al., 2009), and even these show marked differences in observed parameters between differing combinations of agent and target (Bessems, 1998). Furthermore, variability in the results of disinfectant tests is likely to be higher when the concentration of active disinfectant experienced by the target is at or near 'borderline' biocidal conditions (van Klingeren et al., 1998).
Whilst 'in use' concentrations should be well within the highlyeffective killing range, the addition of 'real-world' factors to a test (for example, a porous surface, organic soil or biofilm) may consequently reveal much less efficacy, or more variability, than anticipated (Furuta & Yoshizawa, 1997). Microbicide classes vary substantially in their dilution coefficients. Thus, the activity of phenolic and cresol-based agents reduces rapidly with decreasing concentration (high dilution coefficient), whereas QAC have a low dilution coefficient and therefore they may show a wider effective range of concentrations, performing better in tests that tend towards 'borderline' conditions for other microbicide classes Russell et al., 1979).
Exposure time is a factor for which testing regimes take their cues from real-world applications, and between five and 30 min have been popular intervals in experimental and standardised tests. Five minutes has been considered realistic for water-based non-foaming agents on non-porous vertical surfaces (Holah, 1995), although there may not be good field data to substantiate this. Where interfering organic material is present, extended exposure times appear not to be associated with greater biocidal effect (Berchieri & Barrow, 1996), whereas increased disinfectant concentration can overcome such a limit (Furuta & Yoshizawa, 1997;Gosling et al., 2016).
Very short contact times, such as those encountered in agricultural and veterinary practice when using boot dips, are generally not covered in standard tests and they pose challenges for laboratory testing. Nevertheless this information may be highly relevant for day-to-day disinfectant use on farms where high numbers of organisms are present, for example with Campylobacter on poultry farms.
Ambient temperature may correlate well with effect, even for surface and field tests (Böhm, 1998), but field temperatures (particularly in colder climates) may depart markedly from those usually employed in the laboratory, making specific testing at such temperatures potentially more informative and reliable than extrapolating from 'standard' temperature results (Dee et al., 2005).

| Neutralisation of disinfectant
Neutralising disinfectant action provides a defined end to biocidal activity, i.e. standardisation of single or multiple sampling points.
Testing of neutraliser methods should be performed using the relevant combination of target and microbicide(s) (MacKinnon, 1974).
Additional benefits, of protecting the detection system (usually bacterial or cell culture) and allowing damaged yet viable targets to recover and be detected, have prompted recommendations that control procedures should include testing neutralising agents for inhibitory effects upon the viability of disinfectant-exposed targets and the systems used to detect them (MacKinnon, 1974;Russell et al., 1979).
Dilution is a simple neutralisation method, either using conventional dilution or membrane filtration plus washing (Russell et al., 1979). Some disinfectants that associate strongly with bacterial cell walls (QAC, phenolic agents), or which have low dilution coefficients (QAC) may resist this mode of neutralisation (MacKinnon, 1974;Roesler et al., 2007). Neutralisation by addition of chemical agents, is a well-established alternative, with agents having been established empirically for the most part. They include: non-ionic surfactants, lecithin, thiosulphate compounds, various culture media, histidine or histamine (for formaldehyde), catalase (for hydrogen peroxide), and pH-modifying agents for organic acids (ASTM, 2013a;Carrique-Mas et al., 2007;MacKinnon, 1974;Neuberger, 1944;Russell et al., 1979). Chemical neutralisers have been used in field investigations (Luyckx et al., 2017;Lyutskanov et al., 2010), where a 'disinfect plus dry downtime' cycle may be appropriate before applying neutraliser.

| Recovery and detection, with or without enumeration, of viable biological target
Laboratory test results may be expressed as pass/fail against a reduction target or as a (semi-) quantitative reduction, against appropriate controls. For surface-mounted targets, physical removal via methods such as scraping, agitation with glass beads and/or sonication, is usually needed for enumeration by culture. Any variability in outcomes associated with these removal methods remains unquantified. Methods for enumeration in situ of surface-adhered viable bacteria rely on microscopy or spectrophotometry (Günther et al., 2017;Köse & Yapar, 2017;Smith & Hunter, 2008;Yu et al., 1993). These techniques have not been formally compared with culture techniques in equivalent settings, and as yet they have not been adopted for standardised testing.
In the field, agar contact plates are convenient and of suitable sensitivity for aerobic counts after disinfection, but sampling for specific organisms (such E. coli or enterococci) benefits from a more sensitive technique such as wiping with neutraliser-moistened swabs (Luyckx et al., 2015). For the assessment of hand sanitisers, recovery has been performed by sampling defined areas or by complete immersion of the test hand in neutraliser/recovery medium, a technique that recovers a large proportion of bacteria or viruses (Cremieux et al., 1989;Macinga et al., 2011;Steinmann et al., 1995), and which has been adopted in many standards as the glove or 'glove juice' method (ASTM, 2013b, 2017a).
Assessing the activity of disinfectants against bacterial or fungal spores has been reviewed recently by Leggett et al. (2016).
Complications arise in distinguishing between lethal and static effects, between activity against spores versus against derived vegetative cells, and in recognising sub-lethal damage that can be bypassed by treatment with lysozyme or, plausibly, following ingestion by an animal. It is evident that that there are significant knowledge gaps in this area.
For virus recovery, cell culture remains the principal system, but inoculation of susceptible species with recovered material has also been used to determine if an infective dose remains after disinfection (Baker et al., 2018). For prions, detecting residual infectivity has potentially become simpler with the development of in vitro cell culture and biochemical assay techniques that correlate with activity in whole-animal studies (Wagenführ & Beekes, 2012).

| Organic soil and biofilm
Typically serum, serum albumin or yeast are used as interfering or-

| Cleaning and drying
There is currently a lack of verified methods for assessing cleaning before disinfection. Bacterial spores can be used as markers if the disinfection step in a laboratory test is not expected to be sporicidal (Brouillaud-Delattre et al., 1994). As discussed in the introduction, a drying stage may contribute significantly to the microbicidal effect of C&D. Therefore, depending on anticipated conditions of disinfectant use, tests that include drying steps after application of cleaning and/or biocidal agents, may more closely indicate the efficacy of disinfection procedures, compared with simpler tests that aim to quantify the effect of the disinfection agents alone.

| Special applications
Beyond the disinfection of environmental surfaces in livestock accommodation and veterinary clinical environments, special applications include hand and skin sanitisation (discussed elsewhere) and chemical sterilisation (to include sporicidal activity) of equipment such as endoscopes. C&D of animal transporters is one further special application of particular relevance. Challenges for transporters include: a major risk for the spread of pathogenic or antimicrobialresistant organisms between groups of animals, a number of surfaces with varying texture and finish, use in commercial situations demanding a fast turnaround, and C&D sometimes in very low ambient temperatures (Baker et al., 2018;Böhm, 1998;Dee et al., 2005).
Given their particular circumstances, their relatively small size and the limited range of materials and structures encountered, this is one area where cost-effective and informative field tests might realistically be developed.

| Consistency and interpretation of tests
Repeatability and reproducibility are maximised in tests where there are a few, inherently-controllable variables and where the measured outcome is not very sensitive to changes in those variables. For example, a test of stationary-phase planktonic bacteria, with or without well-defined uniformly distributed organic matter (such as horse serum or yeast extract), fully-mixed with disinfectant and with a readily-achievable pass/fail criterion (Bloomfield et al., 1993). If undissolved 'soil' is introduced, for example faeces or chicken litter, then liquid/solid interfaces are introduced, plus sometimes other organisms, and outcomes are more variable (Berchieri & Barrow, 1996;McLaren et al., 2011;Stringfellow et al., 2009).
For surface tests, the interface and small disinfectant volume introduce variables not present in suspension tests, and there is not a consistent or predictable relationship between performance in a suspension and a surface test (van Klingeren, 1995). Nonetheless, the repeatability of a highly-standardised surface test can in practice be similar to a suspension test, but in both cases variation in the biological target inoculum can be a substantial constraint on both reproducibility and repeatability (Bloomfield et al., 1994). In general, results become less repeatable or reproducible as more variables are introduced into a test, and the more a test incorporates 'borderline' conditions that will demonstrate substantial changes in outcome when conditions alter (Møretrø et al., 2009).
The need to consider what counts as non-significant variation in test outcomes is illustrated by intra-and inter-laboratory trials of microbicidal effect (ME, i.e. reduction in viable bacterial counts).
Two studies of repeatability and/or reproducibility (the latter using three laboratories) for stainless steel surface tests reported that non-significant variation in ME exceeded 1.6 log 10 units (Bloomfield et al., 1994;van Klingeren, 1995). Similarly, in a seven-laboratory ring trial of steel surface tests using local stocks of defined target bacteria strains, the range of ME values for a given test (same agent, concentration and target) generally exceeded two (and often four) log 10 units (van Klingeren et al., 1998).
Surface tests are a versatile tool for assessing disinfection processes in highly standardised model scenarios, but there is no ready translation of results to outcomes in the field. In farm situations, target end-points of routine C&D are either undefined or empirical, for example up to one colony-forming unit of aerobic bacteria per square centimetre (Tamási, 1995), and it is not clear whether present testing methodology allows a pass/fail criterion to be set that has a predictable relationship with the likelihood of pathogen elimination or a particular end-point in the field. There are also factors in the field which can be intractable even with optimised C&D technique. For veterinary and agricultural applications these include: substantial residual soiling of certain materials and structures (for example, belts conveying eggs or droppings), short persistence time on non-porous vertical surfaces or the undersides of equipment/fittings, and low ambient temperatures (Böhm, 1998;Huneau-Salaün et al., 2010;Mueller-Doblies et al., 2010;Tamási, 1995).

| REG UL ATORY AND TE S TING REG IME S FOR DIS INFEC TANT APPROVAL
Requirements for tests to be consistent (using few and highlycontrollable variables) and to give a firm indication of performance in the field are to some extent irreconcilable. Using current methodologies the various aims of disinfectant testing require that more than one type of test be performed, and current testing schemes recognise this. There follows a summary of notable examples for which details are readily available.

| International testing authorities
The principal standards bodies that develop disinfectant tests and have international stances are: AOAC International, ASTM International, the European Committee for Standardization (CEN), and the OECD. The International Organization for Standardization (ISO) provides a classification framework (ISO, n.d.). AOAC and ASTM standards have been adopted as the basis for many of the USA-approved tests, as discussed below. The USA and OECD used international ring trials of candidate methods to develop a draft guidance document on quantitative methods for testing microbicides used on hard non-porous surfaces (OECD, 2013; Table S5).

| European Union
A technical committee (TC 216) of the CEN was established in 1989 to develop a unified European disinfectant testing regime for food hygiene, medicine, veterinary practice and agriculture (Bloomfield et al., 1993). The testing structure adopted includes basic (Phase 1) quantitative suspension tests to establish biocidal activity and  Tables S1 and S3. An ambition to develop 'Phase 3' field performance tests has encountered severe problems with standardisation (Bessems, 1998;van Klingeren et al., 1998), as discussed earlier.

The European Union Biocidal Product Regulations (BPR)
528/2012 governs safety and efficacy approval for biocide active substances and compound products, and the CEN/TC 216 testing scheme provides a route to compliance with the BPR efficacy requirements for member states. However, other tests are acceptable for BPR compliance, such as variants used within the multistage licensing regime of Germany.

| Germany
The German Veterinary Society (Deutsche Veterinärmedizinische Gelleschaft; DVG) examines and approves products for animal husbandry, the food sector, veterinary practice and animal shelters (DVG, n.d.). A particular feature of the German approach is the use of independent expert reviewers approved by the Committee for disinfection that supervises the process for each product. Reviewers oversee testing by laboratories (which are required to participate regularly in ring trials) and provide reports. Four principles are advanced as the basis of DVG product testing (DVG, 2019a): • Independence, via the use of approved experts • Repeatability, via two independent experts and at least two independent test repetitions • Validity, via further review by the Committee • Relevance, with reference to anticipated applications and to managing the risk of selecting for reduced susceptibility by establish- ing MIC values for all tested organisms Guidelines for disinfectant testing are published by DVG (2019b), and these vary by sector. In general, an initial testing report is required, including results from EN-compliant quantitative suspension and surface tests (DVG, 2015). This is reviewed by the Committee,

Livestock sector
Requirements for bactericide testing and approval within the German livestock sector are outlined in the Supporting Information (Table S11 and Appendix S1), and summarised on the DVG website (DVG, 2018a). Appendix S2 in the Supporting Information gives a model testing regime for approval of a livestock bactericide (DVG, 2017b(DVG, , 2018b.

| United Kingdom
Currently the UK adheres to the BPR, and the Department for  (Table 3 and    National testing and approval regimes still vary substantially.

| D ISCUSS I ON
Within Europe, the CEN/TC 216 process has focussed on a single, structured testing regime with the flexibility to incorporate application-specific tests, including tests for surgical hand antisepsis and for fogging and misting devices. In the USA, the approach appears to have been more eclectic, selecting and adapting (where necessary) useful tests from a number of sources. This includes, importantly, tests for surface-biofilmed organisms. The approval process for German livestock products stands out in its provision for independent repeat testing and review.
The advent of biofilm surface tests is an important step, particularly for applications in environments that are habitually moist and not intensively cleaned and disinfected, such as livestock drinker systems. Hand sanitiser and antiseptic tests have to contend with variation between regions (particularly fingernails), and variation between volunteers in respect of, for example, skin and nail surface, resident microorganisms and prior hygiene routines.
There are starting to be test protocols involving ex vivo animal skin or artificial skin, which may yield less variability or at least provide the opportunity to increase repetitions to demonstrate consistency.
The lack of a satisfactory 'one-size-fits-all' test has led authorities to consider specific variants of standardised tests for specific applications. This allows test conditions and pass/fail criteria to be set that either do not have to account for a 'worst case' disinfection scenario, or which address a particular biological target of concern.
One such emerging concern, likely requiring higher concentrations or application rates of disinfectants, is eliminating persistent contamination of animal housing by commensal bacteria that carry multiple drug resistance.

ACK N OWLED G EM ENTS
The authors thank Emma West, for supplying details of the Defra disinfectant scheme.

CO N FLI C T O F I NTE R E S T
The authors declare that there are no conflicts of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
Data sharing not applicable to this article as no datasets were generated or analysed during the current study.