Effect of ‘chlorine-free’ cleaning of milking equipment on the microbiological quality and chlorine-related residues in bulk tank milk
Abstract
The influence of new chlorine-free milking equipment cleaning protocols was compared with that of traditional chlorine-based protocols from the perspective of the microbiological quality and chlorine-related residues in bovine bulk tank milk. Commercial dairy farms using different cleaning protocols were identified, and bulk tank milk from these farms was sampled and tested on three occasions during the milk production season for microbial load and the chlorine-related residues trichloromethane and chlorate. Total bacterial counts and residue levels were lower with chlorine-free than with chlorine-based protocols, demonstrating the new chlorine-free cleaning protocols had a positive impact on milk quality when implemented on commercial farms.
Introduction
Bovine milk is used to produce a wide range of dairy products. Bacterial numbers in milk represent a significant parameter in terms of achieving specifications for dairy products as determined by regulatory authorities and international markets. Production of raw milk under appropriate hygienic conditions is critical to control bacterial numbers. Several studies have focused on quantifying and identifying bacterial types in raw milk on-farm and their effect on dairy products (Barbano et al. 2006; Murphy et al. 2016). Inadequate cleaning protocols at farm level can affect the number of bacteria in milk (Kelly et al. 2009; O’Connell et al. 2015) including the following: mesophilic, psychrotrophic, lipolytic, proteolytic, thermoduric and thermophilic bacteria (Paludetti et al. 2019). Some spore-forming bacteria such as Bacillus cereus, Paenibacillus and Sporosarcina can also enter milk at farm level (Huck et al. 2008). These spore-forming bacteria can survive thermal protocols during dairy processing. Due to their ability to withstand pasteurisation, thermoduric bacteria can limit the shelf life of pasteurised milk (Te Giffel et al. 1997). The microbiota in unpasteurised milk can be highly diverse and differ according to season. This diversity can be attributed to farm factors, such as bedding, feed, milking equipment and milk storage (Elmoslemany et al. 2010; Kable et al. 2016). The study by Paludetti et al. (2019) demonstrated that better microbiological quality of mid-lactational milk resulted in the production of milk powder with lower bacterial counts in contrast to powder produced during late lactation with milk of inferior microbial quality. The effect of season and/or stage of lactation of milk production had an influence on the abundance of different bacterial types in milk. Milk residues on plant surfaces may not become apparent for a number of months after cleaning protocols are put in place on-farm (Gleeson et al. 2013). Thus, any on-farm evaluation of cleaning protocols should be conducted across seasons to capture both the impact of the detergent contact time and contribution of season to microbial diversity.
Chemical residues in raw milk can also impact upon food safety (McCarthy et al. 2018). There are increased food safety concerns regarding the use of chlorine for cleaning milking equipment due to the potential occurrence of residues associated with chlorine, such as trichloromethane (TCM) and chlorate in raw milk and dairy products. Should chlorine come into contact with milk, it can result in the formation of TCM (Tiefel and Guthy, 1997). If milk is contaminated with high levels of TCM, chlorinated hydrocarbons accumulate in the fat-rich fractions, and as a result, products such as butter and cream could contain high concentrations of TCM (Hubbert et al. 1996). A limit of 0.00124 mg/kg for TCM in bulk tank milk was introduced (01/01/20) by Ornua, an Irish agri-food co-operative, which markets and sells dairy products on behalf of Irish dairy processors and Irish dairy farmers. In 2020, the EU commission imposed a maximium residue limit (MRL) of 0.10 mg/kg for chlorate in milk (EU Regulation 749, 2020). Chlorate enters raw milk as a disinfectant by-product, through contact of milk with chlorinated water or as a residue from cleaning chemicals present on equipment surfaces. As in the case of TCM, this contamination can occur at the producer stage due to on-farm practices or/and at processor level. The Board of Ornua passed a resolution to remove all chlorine-based detergents from both farms and processing plants in Ireland from January 2021. It was considered that the removal of chlorine as a cleaning agent from cleaning routines would significantly reduce the risk of these residues in milk and make it easier to achieve dairy product specifications. However, little knowledge is available on the impact of removing chlorine-based detergents (sodium hydroxide combined with sodium hypochlorite) and using alternative chlorine-free (CF) products (sodium hydroxide) on the microbiological quality of bulk tank milk. Internationally, CF cleaning is used with robotic milking systems, where three cleaning operations are undertaken within a 24-h period using hot water. Chlorine-free cleaning of equipment in traditional milking systems has been demonstrated to be effective in the short term (3-month test period) and within a research environment where detergent usage rates, water temperature and rinsing protocols are closely monitored (Gleeson et al. 2013). However, the adoption of new Teagasc-recommended CF protocols (https://www.teagasc.ie/media/website/animals/dairy/research-farms/Chlorine-free-wash-routines_2020.pdf) on commercial farms may have some associated microbiological risk in the long term, particularly if some specific parameters are not adhered to. For example, re-calibration of the automatic detergent dosing systems for both the milking machine and bulk milk tank is necessary to ensure correct uptake rates of CF products as these products generally have a higher caustic content than those used in conjunction with chlorine (Gleeson 2018). Furthermore, there is an increased requirement for more frequent hot washes and the use of alternative disinfectants, such as peracetic acid, which has similar antimicrobial properties to sodium hypochlorite and is effective against a broad spectrum of bacteria, yeasts, moulds and viruses (Tamine 2009). Thus, the objective of this study was to measure the impact of CF cleaning protocols on the microbiological quality and chlorine-related residues in bulk tank milk, on commercial dairy farms across production seasons.
Materials and Methods
Sample selection
The effect of CF cleaning of milking equipment was examined on commercial dairy farms during the milk production season of 2019, from both a microbiological and a chlorine-related residue perspective. Fifty-nine spring-calving herds producing manufacturing milk and using different cleaning protocols were identified by four milk processors to take part in this study. Chlorine-based cleaning protocols involve the use of a product (liquid or powder) containing a combination of sodium hydroxide and sodium hypochlorite and are used in conjunction with acid descale products (generally phosphoric/nitric). Chlorine-free cleaning of milking equipment involves the use of sodium hydroxide (liquid or powder) as the main cleaning agent that is used in conjunction with the same types of acid-based products as previously used for chlorine-based (CB) cleaning. Chlorine-free cleaning also requires more hot washes per week at higher temperatures (https://www.teagasc.ie/media/website/animals/dairy/Chlorine-free-wash-Milk-quality-workshop2020NA.pdf). Therefore, farms were selected for inclusion in the study based on a number of factors including the following: cleaning protocol used; cleaning chemicals used; willingness to take part in the study; and location of farm to facilitate milk sample collection and operate a spring-calving herd. Twenty-eight farms that used CF cleaning products for both the milking machine and bulk milk tank, which were previously identified as good operators achieving good quality milk, were selected for inclusion in the study. Twenty-one farms, which were also identified as good operators using CB cleaning and which were in close proximity to the CF farms, were also selected for the study. A third group of milk suppliers (n = 10) used CF products for cleaning bulk milk tank only (BTCF) and continued to use CB products for milking machine cleaning. During the study period, eight of the farms chosen as CB farms have changed to CF products for cleaning either the bulk tank only or milking machine, resulting in 13 farms in the CB group and leaving a total of 51 farms available for analysis.
The cleaning protocols used were not influenced prior to or during the study as the objective was to capture the impact of the adoption of CF on commercial farms. Milk sampling was undertaken on three test dates (24/4/19, D1; 26/08/19, D2; and 13/11/19, D3). All milk samples (400 mL) were collected on the same morning directly from the bulk milk tanks by the milk quality advisor associated with each farm. Samples were held in cooler boxes and delivered to the Milk Quality Laboratory at Teagasc, Moorepark, Fermoy, Co Cork, Ireland, and microbiological analysis commenced within 12 h of sample collection. A questionnaire survey was conducted by the milk quality advisor with the milk supplier at each sample collection test day to confirm that cleaning protocols remained as defined during the trial period. The questionnaire also captured information on the size of milking plants, cleaning products used, water sources, rinse water levels and frequency of hot water use (Table 1).
| Cleaning protocol | CB | CF | BTCF |
|---|---|---|---|
| No. of farms | 13 | 28 | 10 |
| Average no. of milking units | 17 (9-26) | 16 (6-24) | 17 (8-60) |
| Farm water supply – own well, % | 67 | 79 | 93 |
| Automatic plant washer, % | 33 | 62 | 53 |
| Powder detergent used, % | 20 | 24 | 13 |
| Hot washes ≥5 weekly, % | 47 | 55 | 53 |
| Estimated water temperature, 0 C | 73 | 70 | 69 |
| Peracetic acid – used weekly, % | 40 | 55 | 4 |
| Descale wash > once weekly, % | 30 | 34 | 7 |
| Rinse water levels (litres/unit) | 9.6 | 12.5 | 9.2 |
Microbiological analysis
All milk samples were tested in duplicate for a range of bacteria types. All the microbiological analyses were performed according to the Standard Methods for the Examination of Dairy Products (Wehr and Frank, 2004). Total (TBC), psychrotrophic (PBC), thermoduric (TDBC) and thermophilic (TPBC) bacterial counts were measured using Petrifilm aerobic count plates (ready-to-use media) (3M, Technopath, Tipperary, Ireland), in accordance with the procedures described by Laird et al. (2004). Samples for TDBC were pasteurised at 63°C for 35 min, including time to allow samples to reach the required temperature (Frank and Yousef, 2004); afterwards, the samples were cooled to 10°C using iced water before testing. Samples tested for TBC and TDBC were incubated for 48 h at 32°C (Laird et al. 2004; Pantoja et al. 2009; O’Connell et al. 2016), while samples tested for TPBC were incubated for 48 h at 55°C. The Petrifilms corresponding to the PBC test were incubated for 10 days at 7 ± 1°C (Frank and Yousef, 2004). Regarding PBC, other studies have used Petrifilm incubated at 7°C (Ramsahoi et al. 2011). Presumptive Bacillus cereus group counts (BAC) were performed using Bacara agar (bioMerieux, Hampshire, UK), with plates incubated at 32°C for 24 h; orange colonies with an opaque halo were considered as presumptive B. cereus colonies (ISO 16140:2016). The enumeration of enterococci in milk samples was undertaken following the standard method (BS 4285: 1985). The pour plate method was used in duplicate, and samples were incubated at 37°C for 48 h. Colonies surrounded by a dark halo were counted as enterococci.
Residue detection in milk samples
Quantification of trichloromethane
Trichloromethane was quantified in the milk samples using static headspace gas chromatography (HS-GC) with electron capture detector (ECD) and fitted with a low thermal mass system (LTM) (Agilent 7890A; Agilent Technologies, Santa Clara, CA, USA). The TCM detection limit was 0.0001 mg/kg. The methodology applied was an adaption of the procedure of Resch & Guthy (1999). This analysis was performed in the Milk Quality Laboratory in Teagasc Moorepark.
Quantification of chlorate
The quantification of chlorate and perchlorate (PCHLO) was performed by high-performance liquid chromatography coupled with tandem mass spectrometry (LC/ MS-MS) with electrospray ionisation (ESI) in negative mode at the Teagasc Residue Laboratory, Ashtown, Dublin, Ireland (Danaher et al. 2021). The detection limit of chlorate and PCHLO in milk was 0.0010 mg/kg.
Statistical analysis
Analysis of the measurements was conducted using SAS 9.3. (SAS Institute Inc., Cary NC USA, 2016) A general linear model accounting for the fixed effects of protocol, test day and their interaction was fitted using the mixed procedure. Model: where represents the bacterial and residue measures from each cleaning protocol i (n = 3) and from each test day j (n = 3), fitted as fixed effects; and is the residual term. Interaction between test days for microbiological parameters and main effect (cleaning protocol) means was compared using a Tukey adjustment to P-values to account for multiplicity. Residual plots were used to ensure that all assumptions of the analysis were met. Log transformation was used as appropriate to ensure that distributional assumptions were satisfied and back-transformation of means from the analysis was used as required. Results were deemed significant when α level was below 0.05, and a significant tendency was considered when α level was between 0.05 and 0.1. Other farm management factors were not included in the analysis due to insufficient farm numbers in the sample size.
Results
The farms were chosen for this study based on having a CF cleaning protocol already in place for both the milking machine and bulk tank or the bulk tank only or using traditional CB cleaning protocols. The results of the questionnaire established the differences in how cleaning protocols were implemented on farms (Table 1). The average no. of milking units (n = 17) on farms did not differ between cleaning protocols. A higher percentage of CF farms (62%) had automatic equipment for cleaning the milking machine as compared to that of CB farms (33%). A similar amount of CF and CB farms used powder or liquid detergent products. The number of hot washes per week was highest with CF (55%) than with CB farms (47%), with no differences in the water temperature used. The number of additional descale washes applied (>1) was similar for CB and CF farms, with a much lower level observed with BTCF farms. The usage of peracetic acid as a disinfectant in the rinse water was highest with CF farms (55%) than with BTCF (4) and CB (40%) farms. Rinse water volumes were largest for CF protocols (12.5l/unit) than for CB (9.6l/unit) and BTCF (9.2 L/unit) protocols.
Microbiology
The bacterial counts of bulk milk tank samples obtained from the three milking equipment cleaning protocols are shown in Table 2. Total bacterial counts were significantly lower with CF protocols (3168 cfu/mL) than with CB protocols (12 454 cfu/mL) (P < 0.001) with no difference with the BTCF protocol (6091 cfu/mL). Similarly, PBC and TPBC counts were lower for CF than for CB protocols (P < 0.001) and B. cereus tended to be lower for CF and BTCF than for the CB protocol (P < 0.07). There were no differences between all three cleaning protocols for TDBC and enterococcal counts in milk samples.
| Bacterial counts | Treatments | Significance* | ||||
|---|---|---|---|---|---|---|
| CB | CF | BTCF | Treatment | Test day | Treat × day | |
| Total |
12,454a (8,307-18,672) |
3,168b (2,406-4,172) |
6,091ab (3,874-9,580) |
<0.001 | <0.05 | NS |
| Psychrotrophic |
2,442a (1,560-3,822) |
838b (620-1,134) |
1,291ab (783-2,131) |
<0.001 | <0.001 | NS |
| Thermophilic |
50a (9-292) |
1b (0.4-3.9) |
15ab (2-107) |
<0.001 | <0.001 | NS |
| Thermoduric |
92 (29-290) |
43 (20-92) |
81 (23-292) |
NS | <0.001 | NS |
|
Enterococci |
68 (24-196) |
147 (72-300) |
48 (15-157) |
NS | <0.001 | 0.03 |
| Bacillus cereus group |
0.022a (0-0.01) |
0.001b (0.02-0.19) |
0.002b (0-0.03) |
0.07 | <0.001 | NS |
- * Significance levels represent differences in log measures between treatments where values on the same line have different superscripts.
However, significant differences in bacterial counts were observed across test days (Table 2). Total bacterial counts did not differ between D1 and D3 but were significantly lower at D2 (P < 0.004). Psychrotrophic counts were higher on D1 than on D2 or D3 (P < 0.001). Thermophilic counts were higher on D3 than on D1 or D2 (P < 0.001), and TDBC was lower on D2 than on D1 and D3 (P < 0.001). Enterococcal counts were lowest on D3 than on D1 and D2 (P < 0.001).
There was no protocol-by-test day interaction for TBC, PBC, TPBC, TDBC and B. cereus.
However, there was a protocol-by-test day interaction for enterococci, with lower counts observed for the CB protocol on D3 than on D1 and D2 (P < 0.05). The CB protocol also had lower counts on D3 than on the CF protocol (P < 0.05).
Chlorine residues
When cleaning protocols were compared, the CB protocol had significantly higher mean TCM levels (0.0013 mg/kg) than CF (0.0004 mg/kg) and BTCF (0.0005 mg/kg) (P < 0.001). Levels of TCM were significantly higher for CB than for CF and BTCF on D1 (P < 0.03) and D2 (P < 0.001) with no difference between protocols on D3 (Table 3).
| CB | CF | BTCF | Significance* | |
|---|---|---|---|---|
| Sample no. | 13 | 28 | 10 | |
| Treatment |
0.0013b (0.0010-0.0016) |
0.0004a (0.0003-0.0005) |
0.0005a (0.0003-0.0006) |
0.001 |
| Test day | 1 | 2 | 3 | |
|
0.0006ab (0.0005-0.0008) |
0.0005b (0.0004-0.0006) |
0.0008a (0.0006-0.001) |
0.03 | |
| Treatment × test day | ||||
| D1 |
0.0011a (0.0007-0.0018) |
0.0004b (0.0003-0.0006) |
0.0005b (0.0003-0.0009) |
0.03 |
| D2 |
0.0017a (0.0011-0.0027) |
0.0002b (0.0002-0.0003) |
0.0003b (0.0002-0.0005) |
0.001 |
| D3 |
0.0011 (0.0006-0.0017) |
0.0006 (0.0004-0.0009) |
0.0007 (0.0004-0.0012) |
NS |
- * Significance levels represent differences in log measures where values on the same line have different superscripts.
The average chlorate levels in milk samples where chlorate was detected were 0.005, 0.0098 and 0.110 mg/kg for CF, BTCF and CB, respectively. The percentage of milk samples with chlorate detected (≥0.001 mg/mg) for the three wash protocols were 23, 6 and 11% for CB, CF and BTCF, respectively. The number of samples greater than the legal limit of 0.01mg/kg was highest for CB and lowest for CF (P < 0.01). The percentage of milk samples with chlorate detected (≥0.001 mg/mg) across test days were 3, 9 and 29% for D1, D2 and D3, respectively. The highest per cent of samples with chlorate detected was with CB protocol on D3, which also had a significantly higher number of samples with chlorate detected than on D1 (P < 0.04). Perchlorate was not detected in any milk samples.
Discussion
There were no negative microbiological results observed with CF for any of the parameters measured. In fact, the results for TBC, PBC, TPBC and B. cereus were significantly better than those observed for the CB-based protocol. Achieving minimum TPBC and Bacillus in raw milk is a key measure for the production of dairy products in particular infant milk formula (Haughton et al. 2010), so maintaining low levels is critical with any cleaning routine. While some differences were observed for these parameters, they were not considered biologically important.
Chlorine-free cleaning of milking equipment has previously been shown to be effective in maintaining low bacterial levels in bulk tank milk under experimental conditions and over a 9-week test period, on individual research farms (Gleeson et al. 2013). However, this is the first such study to monitor chlorine-free cleaning of milking equipment on commercial dairy farms and for an extended test period of 8 months. There are five chlorine-free cleaning protocols recommended by Teagasc for cleaning milking machines and three protocols for bulk tank cleaning (https://www.teagasc.ie/media/website/animals/dairy/research-farms/Chlorine-free-wash-routines_2020.pdf). The adoption of CF cleaning protocols for cleaning milking equipment requires some changes in cleaning steps, including an increase in the number of hot washes per week, and an increased use of peracetic acid and acid descalers (phosphoric/nitric acid). Farms were not required to have implemented all steps before being included in the study as long as they adhered to the inclusion or not of CF or chlorine-based products. In this study, CF farms used higher rinse water levels and a higher number of hot washes per week, and an increased amount of peracetic acid, than CB farms. However, some of the cleaning steps recommended for CF cleaning were only partially implemented on individual farms. For example, CF protocols require a minimum of two and up to seven descale washes per week depending on the CF protocol chosen. However, only 34% of farms used >1 descale washes per week. Water temperatures of 75/80°C are recommended for CF, and the average temperatures applied on farms were lower at 70°C. Information on wash water temperature was established from discussions with the farmer as it was not possible to be present to measure exact water temperatures on each farm during actual cleaning. While rinse water levels were higher at 12.5 L/unit for CF protocols, these levels are still less than that recommended (14 L/unit). Considering that not all criteria were reached with regard to the best Teagasc advice on CF cleaning, good microbiological results were achieved with CF cleaning and this indicates the potential to achieve even better quality results if cleaning guidelines were fully followed. However, it is acknowledged that more progressive farmers would be among the first to move to CF cleaning and would have received some guidance from their respective milk quality advisors on moving to CF within a few months of the trial start date. This may partially account for the positive microbiological results obtained as compared to farms on CB systems that may not have received advice on cleaning protocols for some time prior to initiating the trial.
Higher bacterial levels were observed for PBC and TPBC and tended to be higher on D3 for TBC than on D1, with enterococcal counts also higher for the CF protocol on D3 than for the CB cleaning protocol. Higher bacterial counts would be as expected on D3 as late-lactational milk (from spring-calving herds) (Paludetti et al. 2019a). Apart from the treatment-by-day interaction observed for enterococcal counts, hygiene standards were not compromised over time with CF. The microbiota of farm milk can differ across farms due to farm factors other than the milking equipment cleaning protocol applied, such as the environment and teat preparation procedures (Elmoslemany et al. 2010). The authors acknowledge that the outcomes of this study could be dependent on other management factors; however, it was not possible to compare individual factors due to the sample size.
Concentrations of TCM were monitored in bulk tank milk at the three test days during the production season. There are no European residue limits for TCM in raw milk or dairy products. However, Irish dairy processors introduced a limit of 0.00124 mg/kg for milk in 2020 destined for the production of lactic butter, which should have less than 0.030 mg/kg of TCM, as required by the export market (Ryan et al. 2013). While there were no significant differences in TCM levels between sampling test days, CB farms had significantly higher TCM levels than CF and BTCF farms as would be expected on two of the three test days. Increased TCM levels in milk are associated with the use of detergent steriliser cleaning products (Ryan et al. 2012). Five of the CF farms had TCM levels above or equal to the specified limit (0.00124 mg/kg) on D3. It was not established whether chlorinated water or the teat disinfectant products used on farms contributed to these levels, and this requires further investigation. Three BTCF milk samples were above or equal to the limit on D3. While the bulk tank was cleaned on BTCF farms with a chlorine-free product, the higher residue levels observed in milk samples could be associated with the use of detergent steriliser products for cleaning the milking machine.
Concentrations of chlorate were monitored in bulk tank milk at three test days during the production season. In Europe, a default threshold limit of 0.0100 mg/kg for chlorate and perchlorate is applied for raw milk (EU Regulation 396, 2005). For D1 and D2, the mean chlorate concentration in the CB, CF and BTCF farms (in which chlorate was detected) was lower than that limit; however, on D3 (November), the mean chlorate concentration was higher than the European limit and higher than the levels recorded on D1. The higher concentrations of chlorate and the higher number of milk samples with chlorate detected in late lactation were similar to those observed by Paludetti et al. (2019b). It was concluded that an increase in chlorate levels in late lactation may be related to lower milk volumes in bulk milk tanks due to the stage of lactation and may also be due to the length of time cleaning products have been stored on farms as chlorate levels increase with storage time (McCarthy et al. 2018). The highest concentrations of chlorate were recorded with CB farms as would be expected. Three of the ten CB farms had milk samples over the legal limit of 0.01 mg/kg, whereas there was no CF sample over the limit on D3.
However, chlorate was detected in some milk samples on CF farms and the reasons for this were not established in this study. However, potential sources of this chlorate could include a contribution from chlorinated water supplies (McCarthy et al. 2018), chlorine dioxide-based teat disinfectants (Fitzpatrick and Gleeson, 2019) and a low level of chlorate formed in CF products during the manufacture (Gleeson, 2018). Even though chlorate was detected in CF farm bulk milk tanks, these levels were much lower than those observed on CB farms and below the threshold limit.
Conclusions
Farms using wash protocols that excluded chlorine had milk of a higher microbiological quality than farms using traditional chlorine-based cleaning; this may be due to these farms being more focused on the new protocols having received specific guidance from milk quality advisors. There was no deterioration in the milk quality parameters measured across test dates for the CF protocol, indicating no negative impact of the prolonged use of CF products. However, milking equipment was not inspected for visual cleanliness and this should be considered in future studies on commercial farms. Considering that some farms were not following all steps of the CF cleaning guidelines indicates the potential to achieve even better microbiological results when protocols are fully implemented. An increase in the percentage of milk samples with TCM and chlorate residues was observed on the third test day, which coincides with the end of lactation. This increase may be related to milk volume reduction and the use of chlorine products that have been stored on farm for a long period. An explanation as to why some milk samples from farms using chlorine-free products had detected levels of TCM residues requires further investigation. The development and evaluation of chlorine-free cleaning protocols is of national importance and will be critical to the industry in achieving future chlorine-related residue and microbiological targets in dairy products.
Acknowledgements
The authors would like to acknowledge the four milk processors who submitted farms for this study and the milk advisors who obtained samples and recorded questionnaire data. The authors would also like to acknowledge Dr Martin Danaher for the quantification of chlorate, Aoife McDonald for the quantification of trichloromethane and Dr Jim Grant for statistical analysis of the data. The authors acknowledge funding from the Dairy Research Levy Fund administered by Dairy Research Ireland to undertake this research (project MKLS1163). The authors declare they have no competing interests. Open access funding enabled and organized by IRel.
Conflict of Interest
All the authors report no conflicts of interest and are solely responsible for the content and writing of this article.
Author Contributions
David Gleeson: Conceptualization; Data curation; Funding acquisition; Methodology; Project administration; Supervision; Writing – original draft; Writing – review & editing. Lizandra Paludetti: Formal analysis; Methodology; Writing – review & editing. Bernadette O'Brien: Writing – review & editing. Tom Beresford: Writing – review & editing.
Open Research
Data Availability Statement
Research data are not shared.
