The cause for the high prevalence of cefotaximase-producing Escherichia coli reported in dairy calves is unknown but may be partly due to the selective pressure of antimicrobial residues in waste milk (milk unfit for human consumption) fed to the calves. Antimicrobial use and waste milk feeding practices were investigated in 557 dairy farms in 2010/2011 that responded to a randomised stratified postal survey. The mean number of cases of mastitis per herd in the previous year was 47, and 93 per cent of respondents used antibiotic intra-mammary tubes to treat mastitis. The most frequently used lactating cow antibiotic tubes contained dihydrostreptomycin, neomycin, novobiocin, and procaine penicillin (37 per cent), and cefquinome (29 per cent). Ninety-six per cent of respondents used antibiotic tubes at the cessation of lactation (‘drying off’). The most frequently used dry cow antibiotic tube (43 per cent) contained cefalonium. Frequently used injectable antibiotics included tylosin (27 per cent), dihydrostreptomycin and procaine penicillin (20 per cent) and ceftiofur (13 per cent). Eighty-three per cent of respondents (413) fed waste milk to calves. Of these 413, 87 per cent fed waste milk from cows with mastitis, and only one-third discarded the first milk after antibiotic treatment. This survey has shown that on more than 90 per cent of the farms that feed waste milk to calves, waste milk can contain milk from cows undergoing antibiotic treatment. On some farms, this includes treatment with third- and fourth-generation cephalosporins. Further work is underway to investigate the presence of these antimicrobials in waste milk.
Statistics from Altmetric.com
In recent years, bacteria, particularly Escherichia coli, have emerged with resistance to extended-spectrum cephalosporin antibiotics (Canton and Coque 2006, Nicolas-Chanoine and others 2008). These antibiotics are widely used as first-line treatments in human medicine, and the development and dissemination of bacteria resistant to these medicines is a serious medical problem. Resistance is mainly conferred by the production of Extended-Spectrum β-Lactamases (ESBLs), such as cefotaximases (CTX-M), which are able to degrade and inactivate a wide range of cephalosporin antibiotics (Bonnet 2004). A recent investigation of the epidemiology of CTX-M ESBL E coli on a commercial dairy farm in the UK found that 82.8 per cent of calves examined tested positive for these bacteria at day 1 of age (Watson and others 2012). The use of cephalosporins in dairy calves is usually limited to the injection of individual animals, and this is not common practice in that age group of cattle (Hornish and Kotarski 2002). However, most ESBL-producing bacteria isolated and identified in studies of cattle in England and Wales have been enteric E coli from dairy calves (Teale and others 2005, Liebana and others 2006, Watson and others 2012). As such, an alternative route of exposure could be the feeding of waste milk containing antibiotics to calves, which may select ESBL-producing bacteria in the calves' intestine, and may also contain resistant organisms. Previous studies have investigated the relationship between feeding waste milk with antibiotic residues and antibiotic resistance in enteric bacteria of calves, although none has specifically investigated ESBL resistance. Increased resistance to streptomycin among faecal E coli isolates from calves was demonstrated when calves were fed waste milk that had fermented, although a second trial did not show the same effect (Wray and others 1990). Also, increased resistance of gut bacteria to penicillin was demonstrated when calves were fed milk containing penicillin (Langford and others 2003).
Waste milk is that which is unfit for human consumption. Waste milk may comprise colostrum, milk from cows with mastitis, or milk from cows treated with antibiotic or non-antibiotic medicines that may leave a residue in the milk. A wide range of antimicrobials is used for the prophylaxis and therapy of mastitis including the third- and fourth-generation cephalosporin antibiotics. These are available as milking cow intra-mammary tubes for treatment of mastitis during lactation, as dry cow intra-mammary tubes to prevent any infection developing during the non-lactating (dry) period and clear residual infection, and also as injectable preparations (NOAH 2011). Antibiotics are commonly excreted in milk after intra-mammary or parenteral treatment, and withholding times for milk for human consumption are provided for veterinary medicinal products (NOAH 2011).
There is little current information on how widespread different waste milk feeding practices are in the UK dairy industry. Feeding practices are likely to vary from farm to farm, for example, in terms of the composition of waste milk fed to calves, how often it is fed, and for how long, how much is fed and how it is stored. In 1993, feeding waste milk containing antibiotics to calves was considered widespread in the UK and USA (Goodger and others 1993). This study examines the current situation relating to use of antibiotic in dairy cows, which might lead to residues in colostrum and waste milk, and practices associated with feeding these to calves.
Materials and methods
The questionnaire was designed to collect information on farm practices relating to different therapeutic regimens and variations in usage of veterinary medicinal products. Questions about the occurrence of mastitis and lactating/dry cow treatments were included. Among these questions, respondents were asked to identify the three lactating cow antibiotic tubes and three dry cow antibiotic tubes they most commonly used, and rank them as first, second and third in frequency used. They were also asked to name the two injectable antibiotics they use most frequently in their dairy cows, and the proportion of animals treated in the previous 12 months.
The questionnaire also investigated practices relating to the feeding of waste milk to calves, including any storage, fermentation or pasteurisation and the disposal of waste milk. Basic farm details were also collected, such as the number of milking cows and husbandry practices.
Selection of farms
At least 385 responses were required to estimate the proportion of farms carrying out any given practice with an accuracy of 5 per cent and 95 per cent confidence. A sampling frame of dairy farms in England and Wales was constructed from the Cattle Tracing System (CTS) (Defra 2012) in which only farms with more than 50 animals of dairy breed were eligible for inclusion. The number of dairy farms in England and Wales at the time of the survey was estimated at 12,590 based on herds with a registered County Parish Herd (CPH) number. This encompassed approximately 2.9 million animals. The sampling frame was stratified by herd size, and the five strata used were herd sizes of 50–99, 100–149, 150–249, 250–399 and ≥400. The survey took place between September and November 2010. Initially 1000 farms randomly selected from the sampling frame were contacted, but this did not result in sufficient responses, so a further 1003 farms were selected. Each farm received a postal questionnaire and covering letter explaining the purpose of and background to the study, with a prepaid envelope to return their response. As the questionnaires were not anonymous, information on data protection was provided to farms within the questionnaire. Each farm was assigned a unique farm ID to anonymise the data for the analysis. A small inconvenience payment was offered for completion and return of the survey.
Analysis of responses
An MS Access database was constructed to hold the data obtained in the completed questionnaires. Data were extracted from the database and analysed using STATA (StataCorp. 2007: release 10) and MS Excel. Respondents with less than 50 adult milking cows were excluded from the analysis to ensure only commercial units were investigated. Statistical testing was performed using the Wilcoxon rank-sum test and the chi-squared test with Yates correction for two categories of data. Additional information on non-respondents was extracted from the CTS database and also analysed using STATA and MS Excel.
Respondents versus non-respondents
Of the 2003 farms that were contacted, 557 (28 per cent) successfully completed their questionnaires, 34 (1.7 per cent) did not wish to participate in the study, were no longer a dairy farm, or had less than 50 milking cows, and 19 (0.95 per cent) responded after the deadline, so were not included in the analysis. The remaining 1393 farmers did not respond. There was no difference in response rate between the first and second recruitment phases (30.2 per cent and 30.7 per cent, respectively). The low overall response rate of 28 per cent (30.4 per cent including those not analysed) could have resulted in selection bias, so a comparison was made of responders and non-responders. Dairy herd size and farm location were both investigated as possible factors influencing response.
The farms randomly selected to receive questionnaires were stratified to reflect the size of dairy herds across England and Wales (range = 1854, median = 132), and the distribution of the respondents' and non-respondents' populations reflected this.
The median value for the herd size based on data from the CTS database was 136 for respondents and 133 for non-respondents. The similarity between these datasets was tested using the Wilcoxon rank-sum test where P = 0.410, indicating no significant difference in the herd sizes of the two groups.
Farm location and the local density of cattle farming were both investigated as possible factors influencing response. Similar distributions of respondents and non-respondents were observed both geographically when locations were mapped by postcode, and in relation to the density of cattle holdings as described by the number of holdings per 100 km2 (data not shown to preserve anonymity). This suggests that the non-response level was not influenced by holding location or the intensity of cattle farming in the region.
General farm characteristics
Of the 557 farms that responded to the survey, 514 (92 per cent) answered yes to both keeping dairy cows and rearing calves to weaning, and so could complete the questionnaire in full (Table 1). Of these, 83% (95% CI 79% to 86%) did feed waste milk to calves. Respondents were located in 48 different counties across England and Wales with the highest proportion being in Devon (11 per cent) and Cumbria (8 per cent). This correlates well with the proportion of all British cattle in those counties in the CTS (10.5 per cent and 8.5 per cent respectively) (Defra 2008).
Twenty-eight respondents were organic farmers, and of these, 26 reported feeding waste milk to calves. This was 93% of organic farms (95% CI 76% to 99%) compared with 83% of non-organic farms (95% CI 79% to 86%). The average responses from these farms were similar to the average responses from all respondents, with the same antibiotic tubes and injectable antibiotics frequently used. Cefquinome, a fourth-generation cephalosporin, was the active ingredient in the second most frequently used milking cow tubes among organic farms (n = 5, 19 per cent), and also on non-organic farms (n = 147, 29 per cent). Nineteen of the organic farms used dry cow antibiotic tubes, but only 5 (18 per cent) used them in all cows compared with 85 per cent of non-organic farms. This shows that widespread usage of dry cow tubes in the sample was less common on organic farms compared with non-organic farms. The use of antibiotics on organic farms will be considered further on in the discussion.
The herd size (adult milking cows) reported by all respondents, ranged from 1 to 1000 with a mean herd size of 156 (95% CI 147 to 165; median = 130). The official data held within the CTS database does not specify adult milking herd size, only an estimation of total herd size (all stock). Therefore, in order to compare the sample population with the underlying population, the corresponding total herd size, as specified in the CTS database was considered. With this data, the sample population ranged from 52 to 1952, with an estimated mean total herd size of 297 (95% CI 279 to 316; median = 245). The distribution of total herd sizes in the underlying population was similarly skewed, ranging from 1 to 3042, with an estimated mean total herd size of 232 (95% CI 228 to 235; median = 188). The slightly higher mean total herd size observed in the sample population, compared with the underlying population, is not surprising given that farms with less than 50 animals were excluded from the sampling frame. The number of unweaned calves on responding farms ranged from 0 to 200, with the average being 23 calves (95% CI 21 to 25). The majority of respondents (75%; 95% CI 70% to 78%) calved all year round. Of those that calved seasonally, autumn was the most common season for calving (12%; 95% CI 10% to 15%) followed by spring (8%; 95% CI 6% to 11%). Only five respondents calved mainly in winter (1%; 95% CI 0.3% to 2%).
Unsurprisingly, dairy heifers were the most common type of calf retained on the farms until weaning (n = 493 respondents) when compared with bull calves (n = 177 (dairy breed), n = 208 (cross-breed)) and cross-breed heifers (n = 222).
The number of cases of mastitis treated during lactation in the preceding year ranged from 1 to 500, with a mean of 47 (95% CI 42 to 52) (Table 2). This did vary by herd size, but the proportion of cases remained similar with an increase in herd size. The average number of cases of mastitis treated was 20 (95% CI 17 to 23) on farms with less than 100 milking cows, 30 (95% CI 26 to 34) on those with 100–149 milking cows, 57 (95% CI 45 to 69) on farms with 150–199 milking cows, 67 (95% CI 57 to 76) on those with 200–249 milking cows, and 105 (95% CI 83 to 127) on farms with more than 250 milking cows (data not shown).
Therapeutic treatment for mastitis
Ninety-three per cent of respondents used antibiotic intra-mammary tubes to treat mastitis cases (95% CI 91% to 95%). The majority of respondents (81 per cent) used their first-choice lactating cow antibiotic tube to treat between 61 per cent and 100 per cent of mastitis cases (Table 2). Of those ranked as first choice, the most frequently used lactating cow antibiotic tube contained the active ingredients dihydrostreptomycin, neomycin, novobiocin and procaine penicillin (37 per cent). Cefquinome, a fourth-generation cephalosporin, was the active ingredient in the second most frequently used tubes (29 per cent) (Table 3). Fewer respondents stated a second-choice tube (293 respondents v 506). Where a response was given, tubes containing cefquinome were most frequently used (25 per cent) followed by tubes containing dihydrostreptomycin, neomycin, novobiocin and procaine penicillin (21 per cent). Even fewer respondents (130) stated a third-choice tube, where tubes containing amoxicillin and clavulanic acid were most frequently used.
Prophylactic treatment for mastitis
The proportion of respondents using dry cow antibiotic tubes at the end of lactation (drying off) was 96% (95% CI 94% to 98%). Of these, 433 farms (85%; 95% CI 82% to 88%) used dry cow tubes in all cows, 45 (9%; 95% CI 7% to 12%) used dry cow tubes only in cows with mastitis problems in their last lactation, and 29 (6%; 95% CI 4% to 8%) applied other criteria, such as cows with high somatic cell counts and cows with longer drying periods. The majority of respondents (74 per cent) reported that they used their first-choice dry cow antibiotic tubes in 81–100 per cent of cows. Of the tubes ranked first, the most frequently used dry cow antibiotic tube (43 per cent) contained cefalonium, a first-generation cephalosporin (Table 3). As with the lactating cow tubes, fewer responses were given when farmers were asked for their second-choice tubes (156 v 488). Of these, dry cow tubes containing cefalonium were again most frequently used (28 per cent), followed by tubes containing framycetin, penethamate hydriodide and procaine penicillin (26 per cent). Only 29 respondents identified a third-choice dry cow tube, and in this case, cloxacillin benzathine was the most frequently used (45 per cent).
On average, respondents used the two most frequently used injectable antibiotics in 17 per cent of their cows. The most commonly used injectable antibiotics were tylosin (27 per cent of farms) dihydrostreptomycin and procaine penicillin (20 per cent of farms) and ceftiofur, a third-generation cephalosporin (13 per cent of farms) (Table 4).
This survey revealed that on the majority of farms, dairy heifer calves are left to suckle from their dams for up to 24 hours after birth (Table 5), although around 15 per cent of farms did remove calves immediately. The same practice was reported for non-dairy heifer and bull calves. Respondents were asked what types of colostrum are usually fed to dairy heifer calves after they are removed from their dam. Most commonly, this was reported as colostrum from their own dam followed by colostrum pooled from several dams. Very few farms fed powdered colostrum or a commercial substitute (n = 20) (Table 5). Ninety-four per cent (95% CI 92% to 96%) of respondents fed fresh colostrum to calves, while 11% (95% CI 10% to 16%) fed previously frozen colostrum. Five per cent (95% CI 3% to 7%) of respondents refrigerated or fermented their colostrum, and only one respondent pasteurised colostrum. The age at which feeding of just colostrum to calves stopped ranged from 0–21 days, with a mean of four days (median = 4; sd 1.9).
Of the 83 per cent of respondents feeding waste milk to calves, 87% (95% CI 83% to 90%) reported feeding milk from cows with mastitis (Table 6).
Significantly fewer replacement heifer calves (n = 320) were fed waste milk than other calves (n = 378) (P = 0.05). The proportion of the liquid feed diet that was waste milk did differ slightly between different calf types (Table 6).
Most farms that fed waste milk to calves stated that their waste milk sometimes contained milk or colostrum from freshly calved cows and heifers that had received dry cow antibiotics, or from lactating cows that had received antibiotic therapy (93 per cent and 90 per cent, respectively). Milk from high cell count cows was also reported to be sometimes included in waste milk fed to calves on over 70 per cent of farms (Table 6). The majority of respondents (407/413; 99 per cent) fed fresh waste milk to calves, while 6 per cent let it ferment prior to feeding. Very few respondents refrigerated, froze or pasteurised their waste milk (1 per cent, 1 per cent and 0.5 per cent respectively). Only 30% (95% CI 26% to 35%) discarded the milk from the first milking after antibiotic treatment, and 87 per cent fed milk to calves from cows with mastitis (95% CI 83% to 90%).
The most common reasons reported for feeding waste milk to calves were to save money, and because it was convenient (Table 6). Nine per cent fed waste milk because of difficulties with disposal. Other reasons that were given generally consisted of this being a way of improving calf growth and immunity. Over half (n = 264; 57 per cent) the respondents also fed their calves a commercial milk substitute, and around a third of them fed the calves with milk from a bulk tank (34 per cent). Of those using a commercial milk substitute, 4 per cent reported using one that contained a prescription antibiotic; a further 18 per cent did not know whether their milk substitute contained antibiotics. The age at which milk feeding usually stopped ranged from 1 to 39 weeks with an average of 8.5 weeks (sd 2.6).
The results reported here are from the first survey in England and Wales on antibiotic treatment in dairy cows combined with the feeding of waste milk to calves. The response rate for this survey was 30 per cent which is low but not uncommon for mailed questionnaires, and is in fact more than double the response rate obtained in a similar survey of dairy farms in Germany (Heuwieser and others 2010). Methods of increasing the response rate to postal questionnaires have been reported elsewhere (Edwards and others 2009) and include monetary incentives, recorded delivery, pre-notification and follow-up contact. Monetary incentives were used in this study which were likely to have encouraged the responses; however, further methods could have been employed to improve the response rate. A possible reason for the low response rate could be that some non-dairy farms were contacted due to inaccuracies in the database used for selection. The database relies on updates from local authorities and, so, may not always be current. Despite the low response rate, the sample size requirements were met. No significant difference was observed between the herd size of farmers who responded and those who did not. Additionally, the response rate did not appear to be associated with the location of the farm or the cattle-holding density within the farm's locality. However, it is not known whether antibiotic usage and waste milk feeding practices of respondents in this survey were representative of the general population.
A survey by Bradley and others (2007) suggested that the incidence of clinical mastitis in England and Wales is between 47 and 65 cases per 100 cows per year. In this study, the average number of cases of mastitis treated during lactation was less, at 47 per herd per year which, with an average herd size of 156 cows, would equate to around 30 cases per 100 cows per year.
Eighty-three per cent of responding dairy farmers who kept calves to weaning fed waste milk to calves. This practice has not been surveyed previously in the UK, but the frequency of feeding waste milk to calves reported here is higher than recent reports from the United States Department of Agriculture's (USDA) National Animal Health Monitoring System, where 30.6 per cent and 2.8 per cent of US dairy farms were reported to have fed unpasteurised and pasteurised waste milk to dairy heifer calves, respectively (USDA 2007).
The reason why waste milk was reportedly fed less frequently to dairy heifer calves than to other calves in the current survey was not established, but it is possible that this may be to reduce the risk of infection with Mycobacterium avium paratuberculosis in the replacement heifers (Defra, 2004).
This survey has shown that antibiotic intra-mammary tubes are commonly used in lactating cows in England and Wales, and that cefquinome, a fourth-generation cephalosporin is the first-choice treatment for mastitis on nearly a third of responding dairy farms. Cefquinome has been shown to have a similar selective effect as ceftiofur for ESBL-producing E coli in the intestinal flora of pigs, and a greater selective effect than amoxicillin (Cavaco and others 2008). Cefquinome concentrations of between 10,000 and 27,000 ng ml−1 have been detected in treated quarters, but this fell rapidly below the European Union Maximum Residue Limit 2–3 days after the last intra-mammary dosing (Thal and others 2011). This indicates that discarding the first milking after treatment may reduce exposure of calves to this antibiotic in waste milk. However, this was done on only 30 per cent of dairy farms in this survey.
Antibiotic intra-mammary tubes are also commonly used at drying off, and often in the whole herd as a prophylactic measure. Therefore, on farms where the whole herd is treated, all calves would normally receive colostrum containing antibiotic residues, unless they feed a powdered colostrum substitute, which was rare in this study (approximately 3 per cent). All calves need to ingest colostrum during the first few hours of life to provide protection against infection, and it is beneficial to continue feeding them colostrum, as the antibodies it contains have a local action within the gut to protect against enteric diseases. In this survey, the most commonly used dry cow intra-mammary antibiotic was cefalonium, a first-generation cephalosporin.
The survey has also shown that widespread usage of dry cow tubes is less common on organic farms compared with non-organic farms. The regulations for antibiotic use in dairy cows on organic farms vary according to the organisation providing the organic certification. However, the Soil Association standards state that antibiotics must not be used on a whole herd or flock basis to prevent disease (Soil Association 2012).
Ceftiofur, which is only available as an injectable product, was one of the two most frequently used injectables administered on 13 per cent of responding dairy farms in England and Wales where, on average, it was used in 18 per cent of animals per year. Ceftiofur has a zero milk withdrawal period after injection (NOAH 2011). Milk from ceftiofur-treated cows can be suitable for human consumption dependent on the condition undergoing treatment and whether milk is otherwise wholesome; it therefore may not necessarily be fed to calves as waste milk. Most other injectable antibiotics licensed for dairy cattle, including cefquinome, do have withdrawal periods for milk destined for human consumption, and milk from treated cows will be considered waste milk.
Fermentation of colostrum and waste milk may degrade certain antibiotics, however, the appropriate conditions for fermentation are difficult to achieve in the UK (Wray and others 1990). This survey has shown that only 5 per cent and 6 per cent of responding dairy farms in England and Wales ferment their colostrum or waste milk, respectively. Pasteurisation of waste milk before feeding has been recommended to reduce microbial load. However, this practice is rarely carried out (only one respondent in this survey pasteurised colostrum, and only two pasteurised waste milk). This is similar to findings from studies in the USA and Canada (Kehoe and others 2007, Vasseur and others 2010). The low occurrence of this practice may well be due to the significant investment required to routinely pasteurise small volumes of liquid. Even when used, pasteurisation is unlikely to affect the antimicrobial concentration in milk and may reduce the immunological benefits of colostrum. A 26.2 per cent reduction in immunoglobulin concentrations following pasteurisation of colostrum was observed (Godden and others 2003), but lower temperatures and longer heating times using batch pasteurisers have reduced this problem (Godden and others 2006, McMartin and others 2006). Most respondents to this survey (99 per cent) fed their waste milk to calves fresh; that is, with a minimal period of storage. If the milk is not fed immediately to calves but is left at room temperature until the next feeding, the microbial load will increase substantially (Quigley 1998) exposing calves to greater numbers of bacteria.
More than a third of the farms surveyed in England and Wales leave calves to suckle their own dam for up to 24 hours after birth (39 per cent), while 15 per cent remove calves from their dam straightaway. In Quebec (Canada), 32.5 per cent of farms separated the calf from the cow within two hours of birth, and 73.2 per cent before 12 hours of age (Vasseur and others 2010), and in the USA, 55.9 per cent of farms remove calves immediately after birth (USDA 2007). An American study of dairy farms in Pennsylvania reported that it was common practice to remove dairy heifer calves from their dam straightaway, with 87 per cent of farms feeding colostrum via a bucket or bottle (Kehoe and others 2007). They also reported that on 65 per cent of farms, dairy heifer calves were fed colostrum from their own dam. This is comparable with our findings that on farms where the calf is removed from its mother, 50 per cent of responding farms in England and Wales fed dairy heifer calves colostrum derived from their own dam. Pooling of colostrum, considered to be a risk in the transmission of Mycobacterium avium paratuberculosis infection, occurred on 32 per cent of farms in our study compared with 21 per cent of US farms (USDA 2007). It is possible that removing calves from their dam and feeding them pooled waste milk and colostrum could provide greater exposure to antimicrobials in waste milk. This data provides an insight into current calf management practices, and can be used to inform and structure future investigations into the effects of feeding waste milk on antimicrobial resistance in calves.
Feeding of milk replacer containing antibiotics is currently at a low incidence on responding farms in England and Wales (4.3%; 95% CI 2.2% to 7.5%). This is less than the 57.5 per cent reported in the USA (USDA 2007).
The most common reason for farmers to feed waste milk to calves was to save money. If feeding of waste milk containing antibiotic residues is shown to be associated with the occurrence of ESBL E coli in calves, then there could be economic implications in advising not to feed it. This survey is part of a larger study funded by the UK government, and further work is underway to detect antibiotic residues in waste milk to assess their stability under different storage conditions, and to investigate whether residues of antibiotic detected in waste milk are able to select for ESBL-producing E coli in the calves' gut. The fourth-generation cephalosporin cefquinome has been identified in this survey as the cephalosporin most likely to be present in waste milk fed to calves. This survey provides an overview of current practices on a large sample of English and Welsh dairy farms in relation to antimicrobial usage and the feeding of waste milk, information which has not been available previously.
This work was funded by the UK Government Department for Environment, Food and Rural Affairs as part of project OD2031. The authors are grateful to all the farmers who took the time to participate in this survey.
Provenance: not commissioned; externally peer reviewed
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.