Effects of cattle breed and production system on veterinary diagnoses and administrated veterinary medicine in alpine dairy farms

Abstract

The aim of the following study was to investigate the effect of dairy production systems, including housing, farm management, pasture access and concentrate supplementation on veterinary diagnosed diseases as well as on administrated drugs in local dual-purpose Alpine Grey (AG) cattle and specialised Brown Swiss (BS) cattle in alpine dairy systems. In the final dataset 916 veterinary diagnosis from 524 cows farmed in 6 low-concentrate Alpine Grey farms (LAG), 9 low-concentrate Brown Swiss farms, (LBS) 11 high-concentrate Alpine Grey farms (HAG) and 15 high-concentrate Brown Swiss farms (HBS) were available for the statistical analysis. The most diagnoses referred to udder health (mastitis, dry-cow therapy) followed by parasite infections and reproduction disorders. In detail, BS had a higher occurrence for mammary infections than AG while HBS farms showed a significant higher incidence for udder diseases compared to all other farm categories. Pasture access had in tendency a beneficial effect on udder health but was detrimental for cows fertility. For latter, special attention should be put on providing sufficient energy supply to grazing animals for avoiding severe negative energy balances and resulting metabolic disorders which reduce fertility. Finally, to reduce the use of antimicrobials classified as highly critical by the WHO in terms of resistance, selective therapies, particularly for mammary infections and dry-cow therapies as well as the inclusion of preventive strategies in every days farm management should be encouraged.

Highlights

• Pasture access has a favourable effect on udder health

• Selective veterinary treatments for udder infections and dry-off practice should be promoted

• Brown Swiss cows are more susceptible to mammary infections than Alpine Grey cows

Keywords:

• Dairy systems
• animal health
• pasture
• veterinary diagnose

Introduction

Animal health and welfare are crucial for the economic success of a farm (Pfeiffer et al. 2015). In fact, health and fertility disorders were stated to have a negative effect on longevity and reproductive ability of dairy cows and create next to veterinary treatment costs a series of follow-up costs caused by premature culling and higher restocking rates, reduced milk production and prolonged calving intervals (Heikkilä et al. 2012; Kofler 2012; Martin et al. 2018; Mostert et al. 2018). Alpine dairy farms are very different production systems in sense of size, husbandry and feeding systems and overall productivity as those present in the lowland areas of Europe and North America (Firth et al. 2019). Consequently, production costs per unit of output are higher and remuneration per working hour is lower (Poulopoulou et al. 2018; Kühl et al. 2020). Thus, a cost-effective production is of high importance in order to maintain farm profitability in such area. Next to feeding costs veterinary treatment costs were shown to have a relevant impact on total variable costs of alpine farms (Kühl et al. 2020). Accordingly, it is of farmers’ and other stakeholders’ interest to improve animal health and welfare in order to reduce veterinary costs as well as meet the increasing consumers’ demand for animal welfare and food safety (Egger-Danner et al. 2015). Latter is of high importance because of the rising concern of antimicrobial resistances (AMR) due to the overuse of partly poor-quality veterinary medicine for therapeutic use in livestock production systems (Clifford et al. 2018).

In Italy the registration of veterinary treatments in form of an electronic registration has become mandatory by law in 2019 (Law of the 20th November 2017 Nr 167 Art 3). Thereby, veterinarians have to register the prescribed drug and the general group of the observed diagnosis according to a specific diagnose key in the electronic recording system, which was established and is managed the by the Italian Ministry of Health. Latter should enable the traceability of veterinary medicinal products and medicated feed and therefore promote a more diligent application of antimicrobial substances in livestock production. However, the electronic recording system provides only information about applicated drugs but does not display information neither regarding farm system nor about farm management, which are known to have relevant impact on animal health and welfare (Ivemeyer et al. 2012; Åkerfeldt et al. 2021).

Therefore, information regarding veterinary diagnoses and administrated drugs in combination with data on management practice are required to gain better insights for preventing diseases and consequently reduce the use of veterinary medicine and the associated risk for AMR (Landers et al. 2012; Bengtsson and Greko 2014; De Monte et al. 2020). The objective of the following study was to investigate the effect of housing systems, farm management, pasture access and concentrate supplementation on veterinary diagnoses as well as on administrated drugs in local dual-purpose Alpine Grey (AG) cattle or specialised Brown Swiss (BS) cattle, both frequently farmed in low-input and high-input alpine dairy systems (Kühl et al. 2020; Zanon et al. 2020). Furthermore, a risk factor analysis was conducted in order to determine possible risks for veterinary diagnoses in analysed dairy farms.

Material and methods

Study area

The province of South Tyrol is located in the North-Eastern part of Italy on the border to Austria and Switzerland. Livestock production is characterised by small-scale farms, mostly run on a family basis. Approximately 72,000 hectares of field are used for feed production and 128,000 cattle are kept in the region of which 52% are dairy cows (Agrar- und Forstbericht 2019). The average South Tyrolean dairy farm is keeping 15 dairy cows and produces approximately 88,500 kg milk per year (Tätigkeitsbericht Sennereiverband 2019).

Data collection and editing

In a first step, 64 dairy farms were pre-selected considering results of the sustainability report of the dairy sector in the province of South Tyrol (Sennereiverband 2017). Possible trial farms had to keep either local dual-purpose Alpine Grey cattle or specialised Brown Swiss cattle, feed either high or low amounts of concentrates in the feed ratio, possibly using pasture during the vegetation period, keeping the cows in either tie-stall or free-stalls, to be organised on a part-time (commonly practiced in alpine dairy production systems of South Tyrol) or full-time basis. The participation was on a voluntary basis. In order to consider the heterogenous degrees of production intensity in alpine dairy-production systems (Kühl et al. 2020) farms with different production intensity level in terms of concentrate use in the feed ration were considered and classified as follows (Kühl et al. 2020): 1) low-input AG (LAG) ≤ 3.5 kg concentrates/day and cow; 2) low-input BS (LBS) ≤ 4.5 kg concentrates/day and cow; 3) high-input AG (HAG) ≥ 6.0 kg concentrates/day and cow; 4) high-input BS (HBS) ≥ 7.5 kg concentrates/day and cow. Different values for classifying systems in low and high input for AG and BS are because for AG generally lower amounts of concentrates are required in the feed ratio to fulfil the energetic requirements compared to the high yielding BS (Horn et al. 2014). Crude protein in concentrate fodder varied between 16% and 21%. Following the classification criteria according to Kühl et al. (2020) 41 focus farms remained for the final investigation.

Direct health data per single animal regardless of sex and age were collected during farm visits considering the farm drug registry of respective farms. The observation period ranged from October 2014 to October 2017. Each event of the same disease was considered a new case on the same animal if it occurred at least one week after the previous event (Pérez-Cabal and Charfeddine 2013). For the evaluation of the drugs administrations the names of active substances of respective medical were considered. The active substances were identified individually by searching for the commercial name of each drug administered in the Dictionary of Veterinary Medicines and Animal Health Products XII edition 2017. Data were evaluated on the base of the classification of the critically important antimicrobials list of the World Health Organisation (WHO – CIA List; WHO 2019). Also, both painkiller and anti-inflammatory drugs were evaluated based on the active substance (e.g. NSAIDs, glucocorticoids). With regard to the drugs used on the various farms, it was decided to only describe the frequency of the active ingredients administered, as data on the dose administered per treatment were not available for all farms. According to the European Medicines Agency (EMA 2015, 2018), in order to assess quantitatively the antibiotic use, it is necessary to have data available on the daily amount administered per animal and the duration of treatment, in addition to the concentration of the antimicrobial contained in that particular commercial formulation and live weight of the animals (Defined daily dose- DDDvet and Defined Course dose-DCDvet).

After editing, a total of 916 diagnosis from 524 cows farmed in 6 LAG, 9 LBS, 11 HAG and 15 HBS farms were available for the statistical analysis.

Statistical analysis

Statistical analysis was performed with SAS software v. 9.4 (SAS Institute Inc., Cary, NC). Incidence of a diagnosis was defined as the proportion (%) of new cases of the disease on the total number of animals present in the barn over the studied period ((number of new cases/total animals at risk) × 100). Statistical analysis was only applied to veterinary diagnoses with an overall prevalence of 4% in the original dataset. Since the Kolmogorov-Smirnov and Shapiro-Wilks test for normality was significant (p<.05) for farm incidences and frequency of administrated substances, non-parametric test (Kruskal–Wallis) was applied. Significance level was set at p < .05. p-values were adjusted considering the Bonferroni correction method. Moreover, a multinomial logistic regression was performed, using the proc glimmix procedure to test potential risk factors associated with veterinary diagnosed diseases. The farm was included as a random effect for considering repeated health problems and related housing and management factors for multifactorial diseases. Each veterinary diagnosed disease was expressed as a binary variable (YES | NO) at the level of the individual dairy cow. Resource based measures and animal-based parameters depicted in Table 1 were classified as potential risk factors and were defined as binomial parameters. Descriptive statistics were performed to determine logical thresholds (Table 1). Farm was considered as random effect. For inclusion and exclusion of the potential risk factor a stepwise forward and backward selection was conducted. Only risk factors with a p-value < .05 were maintained in the final model. Odds ratios (OD) with 95% confidence intervals were calculated for every maintained risk factor in the final model. The goodness of fit of the models for respective diagnose was evaluated by drawing ROC curves and in determining the area under the curves (AUC). The AUC for the model describing udder diseases was 0,67 (SE:0.02; p < .001), for the model describing reproduction disorders 0.65 (SE:0.04; p < .001), for the model describing dry-off practice 0.67 (SE:0.02; p < .001), for the model describing parasite infections 0.89 (SE:0.01; p < .001) and for the model describing respiratory diseases 0.75 (SE:0.06; p < .001).

Table 1. Classification of resource-based measures and animal-based parameters as potential risk factors.

Parameters	Classes
Pasture access	Yes | no
Pasture period	>45 days | ≤45 days
Farm dimension	>12 cows | ≤12 cows
Husbandry System	Free-stall | Tie-stall
Concentrates level	High | low
Business organisation	Full-time | part-time
Breed	Brown Swiss | Alpine Grey

Results

Farm characteristics

Farm characteristics of the considered trial farms are depicted in Table 2. In terms of metres above sea, level HAG were the highest located farms (p < .05) whereas LBS were the lowest located farms (p < .05). Farms rearing AG cattle and farms feeding lower amounts of concentrates had significant longer pasture periods compared to high concentrate feeding farms and BS farms (Table 2). Further, LAG and HAG farms kept significant more animals (p < .05), in particular more calves and heifers, than HBS and LBS farms. The dominant husbandry system in all farms was tethering (70%), except for LAG, where free-stall and tethering systems were equally found (Figure 1). High concentrated feeding farms were mostly run on a full-time basis (>80%) whereas approximately one third of low concentrated feeding farms were run on a part-time basis (Figure 2).

Figure 1. Frequency of husbandry systems in respective farm category (HBS: High-Input Brown Swiss farms; HAG: High-Input Alpine Grey farms; LBS: Low-Input Brown Swiss farms; LAG: Low-Input Alpine Grey farms).

Figure 2. Business organisation in respective farm category (HBS: High-Input Brown Swiss farms; HAG: High-Input Alpine Grey farms; LBS: Low-Input Brown Swiss farms; LAG: Low-Input Alpine Grey farms).

Table 2. Structural characteristics of investigated farms.

	HBS (n = 15)		HAG (n = 11)		LBS (n = 9)		LAG (n = 6)		Kruskal–Wallis** (df = 3)
	Mean	Std	Mean	Std	Mean	Std	Mean	Std	Chi²	p
Masl, m*	1,134^b	197	1,294^a	206	1,028^c	146	1,160^b	258	188.800	<.001
Pasture days, d	16^d	26	71^b	39	48^c	20	103^a	68	357.000	<.001
Cows, n	12^b	2	12^c	1	11^d	2	13^a	3	68.000	<.001
Calves, n	2^d	2	4^b	2	3^c	2	6^a	4	221.700	<.001
Heifers, n	4^b	3	6^a	3	4^b	1	7^a	4	135.700	<.001
Tot. Animals, n	18^c	3	22^b	4	18^c	2	26^a	10	206.500	<.001

Different superscript letters within a row are significantly different (p < .05) for pairwise comparison of group means using the Wilcoxon rank-sum test; *Metres above sea level; **the Kruskal–Wallis test was used to test for differences among the 4 farm categories.

Veterinary diagnoses

Number of diagnoses and respective frequency of diagnoses are illustrated in Table 3. A total of 916 veterinary diagnoses were recorded in the farms during the study period. The most frequent veterinary diagnoses concerned udder health management with mastitis and dry-off practice. Moreover, diagnoses regarding parasite infection played a relevant role (Table 3). In latter, nearly all referred to prophylactic and metaphylactic treatments against helminths, particularly. Regarding reproduction disorders retained placenta and metritis were the most frequent diagnoses (Table 3). Disbudding and digestive disorders presented a frequency of 4% and respiratory diseases 5% of which crowding disease was the most frequent veterinary diagnose. Veterinary diagnoses with an overall frequency of ≤4% are summarised in Others and concerned locomotive apparatus, urinary apparatus, skin, metabolic status, nervous system and calve diseases (Table 3).

Table 3. Descriptive statistics of veterinary diagnoses.

Veterinary diagnoses	Frequency	n = 916
Udder diseases	38%	352
Mastitis	38%	346
Others	1%	6
Dry-off practice	21%	190
Parasites	11%	103
Helmints	11%	102
Coccidiosis	0%	1
Reproduction disorders	9%	86
Retained placenta	6%	53
Metritis	2%	15
Others	2%	18
Respiratory diseases	5%	45
Crowding disease	3%	29
Pneumonia	2%	16
Digestive dissorders	4%	40
Diarrhea	1%	12
Others	4%	38
Disbudding	4%	39
Others	7%	61

Bold letters indicate Diagnose groups.

Farm incidences

Mean farm values with respective standard deviations are depicted in Table 4. The incidence for udder diagnoses was significantly highest (p < .05) in HBS farms but was not significantly different between the other farm categories. Dry-off practice was not significantly (p < .05) different among farm categories. Similarly, no statistical difference was observed for diagnoses referring to parasite infections (Table 4). The incidence for reproduction diagnoses were higher in BS farms but not significantly different from farms rearing AG cattle. Incidences for respiratory diseases were not significant between farm systems. Finally, incidences for diagnoses concerning digestive apparatus and disbudding were not significantly different between respective farm categories (Table 4).

Table 4. Calculated farm diagnoses incidences for respective farm categories.

	HBS		HAG		LBS		LAG		Kruskall–Wallis* (df = 3)
Diagnoses group	Mean	Std	Mean	Std	Mean	Std	Mean	Std	Chi²	p
Udder, %	27.600^a	12.200	10^b	7.700	10.500^b	5.700	8.800^b	7.900	18.900	<.001
Dry-off, %	10.400	19.300	3.400	10.000	5.600	5.200	17.100	19.300	7.300	.060
Parasites, %	2.300	5.800	8.000	13.600	0.000	0.000	9.900	18.300	3.600	.310
Reproductive app., %	4.600	6.800	3.400	3.200	4.300	3.000	2.400	3.000	1.200	.740
Respiratory system, %	3.400	7.300	0.700	1.300	2.800	5.900	0.200	0.400	3.000	.390
Digestive app., %	1.900	2.500	2.100	3.800	2.800	4.200	0.200	0.600	4.400	.220
Disbudding, %	4.100	8.300	0.000	0.000	0.000	0.000	1.900	4.500	5.000	.170

Different superscript letters within a row are significantly different (p < .05) for pairwise comparison of group means using the Wilcoxon rank-sum test; *the Kruskal–Wallis test was used to test for differences among the 4 farm categories.

LAG: low-concentrate Alpine Grey farms; LBS: low-concentrate Brown Swiss farms; HAG: high-concentrate Alpine Grey farms; HBS: high-concentrate Brown Swiss farms.

Use of active substances

During the observation period 1732 drugs were administrated. Thereby antibiotics represented 78.52% of all treatments (N = 1360), followed by painkillers and anti-inflammatory active substances (NSAIDs, glucocorticoids and analgesics; 7.85%; N = 136) and antiparasitic drugs (5.95%; N = 103). Hormones (1.56%; N = 27) and local anaesthesia (2.54%; N = 44) were less frequent. The most common antiparasitic drugs contained ivermectin (47.57%; N = 49), followed by the combination of ivermectin and clorsulon (40.77%; N = 42). Eprinomectin, which is normally used in lactating animals, was administrated in 11.7% of all treatments. Further, 18.5% of administrated antibiotics belonged to groups classified by the WHO as Highest Priority Critically Important Antimicrobials (N = 252), 31.6% belonged to groups classified by the WHO as High Priority Critically Important (N = 430) and the remaining 49.9% were antimicrobials classified as highly important (N = 678). In general, Highest Priority Critical Important antimicrobials seemed to be used more often in farms rearing BS (43,4%, Table 5). Moreover, 25.71% of the antimicrobials used for udder diseases belonged to the WHO Highest Priority Critically Important antimicrobials (N = 190), 16.33% to WHO High Priority Critically Important antimicrobials (N = 221) and 44.38% to Highly Important Antibiotics (N = 328). Highest Priority Critically Important antibiotics were used more carefully for dry-off and reproductive disorders (7.81% and 7.31%, respectively). For dry-off treatments WHO High Priority Critically Important antimicrobials and Highly Important Antibiotics were used at a similar percentage (46.90% and 45.27%, respectively). Finally, disorders affecting the reproductive apparatus were treated mainly with less significant antimicrobials in terms of AMR (67.47%).

Table 5. Descriptive statistics of antimicrobials used in respective farm categories.

WHO-class	HBS (n = 673)		HAG (n = 272)		LBS (n = 189)		LAG (n = 226)
	n	%	n	%	n	%	n	%
Highest priority critically important	157	23.3	38	14.0	38	20.1	19	8.4
High priority critically important	210	31.2	83	30.5	56	29.6	81	35.8
Highly important	306	45.5	151	55.5	95	50.3	126	55.8

LAG: low-concentrate Alpine Grey farms; LBS: low-concentrate Brown Swiss farms; HAG: high-concentrate Alpine Grey farms; HBS: high-concentrate Brown Swiss farms.

Risk factor analysis

Resourced based measures and animal-based parameters influenced the risk for veterinary diseases diagnosed by veterinarians (Table 6). For the veterinary diagnoses referring to digestive apparatus and disbudding no significant risk factor was determined. The risk for udder diseases was higher (OR 2.795) in BS and in farms keeping more than 12 cows. In contrast, the occurrence of udder diseases was lower when cows had access to pasture or longer pasture periods and were fed with lower amounts of concentrates (Table 6). Further, reproduction disorders were more likely to occur when animals had access to pasture or had longer pasture periods. However, the risk for reproduction disorders was lower in farms organised on a full-time basis. Dry-off practice was more common in full-time farms keeping AG in tie-stalls and feeding low amounts of concentrates whereas farms using pasture and long pasture periods showed a lower occurrence for dry-off practice (Table 6). Furthermore, the risk for parasite infections was lower in BS and in big farms feeding low concentrate levels and keeping the animals in tie-stalls. Finally, respiratory diseases were more likely to occur in BS cattle (Table 6).

Table 6. Risk factor analysis for veterinary diagnoses, considering cattle breed, concentrate feeding level, pasture access, pasture period, farm dimension, business organisation and husbandry system as risk factors.

Risk factor^a	p-value	Level 1	Level 2	Odds ratio^b	df	Confidence interval
Udder diseases
Breed	<.0001	BS	AG	2.7650	775	1.9200–3.9800
Level	.0085	Low	High	0.5990	775	0.4100–0.8800
Pasture access	.0008	Yes	No	0.3460	775	0.1900–0.6400
Pasture period, days	.0202	>45	≤45	0.4870	775	0.2700–0.8900
Farm dimension, n	<.0001	>12 cows	≤12 cows	1.9930	775	1.4200–2.8000
Reproductive disorders
Pasture access	<.0001	Yes	No	6.9320	777	3.3700–14.2700
Pasture period, days	<.0001	>45	≤45	4.1140	777	2.1500–7.8600
Business organisation	.0001	Full-time	Part-time	0.3390	777	0.2000–0.5900
Dry-off practices
Breed	.0075	BS	AG	0.4870	772	0.2900–0.8300
Level	<.0001	Low	High	3.8410	772	2.2900–6.4500
Husbandry system	.0091	Free-stall	Tethering	0.4070	772	0.2100–0.8000
Pasture access	.0026	Yes	No	0.1340	772	0.0400–0.4900
Pasture period, days	.0012	>45	≤45	0.1190	772	0.0300–0.4300
Business organisation	<.0001	Full-time	Part-time	3.7540	772	2.1600–6.5300
Parasites infections
Breed	<.0001	BS	AG	0.0480	776	0.0200–0.1000
Level	.0040	Low	High	0.4250	776	0.2400–0.7600
Husbandry system	.0130	Free-stall	Tie-stall	2.0930	776	1.1700–3.7500
Farm dimension, n	<.0001	>12 cows	≤12 cows	0.0400	776	0.0100–0.1700
Respiratory diseases
Breed	.0204	BS	AG	3.6630	779	1.2200–10.9700

^aAll risk factors for the final model were significant (p < .05).

^bOdds Ratio is given for the pairwise comparison between risk factor levels. Level 2 is the reference level. When Odds Ratio <1 the risk is decreased in comparison with the reference level. When Odds Ratio >1 the risk is increased in comparison with the reference level.

AG: Alpine Grey; BS: cattle and specialised Brown Swiss.

Discussion

To the best of our knowledge, this is the first study combining management data and veterinary diagnosed diseases with information about administrated drugs. The aim of the following study was to investigate the effects of cattle breed, housing system and farm management on veterinary diagnosed diseases and on the application of veterinary administrated medicals.

The high frequency of udder diseases observed in the following study (Table 3), of which the lions share is related to mastitis, is in line with results reported in literature (Egger-Danner et al. 2012; Ribeiro et al. 2013; De Monte et al. 2020). The significant higher incidence for udder diseases in HBS (Table 4) might be related to both, cattle breed and intensive farm management. In fact, it is well known that mastitis is a multifactorial disease (De Vliegher et al. 2018) for which environmental, management and animal related factors all play an important role. In our case the higher incidences for udder diseases observed in HBS might be explained by unfavourable genetic association between udder health traits and milk yield traits (Heringstad et al. 2000; Koivula et al. 2005). Latter is further illustrated by the higher risk for udder disease in BS compared to AG observed in our study (Table 6). Moreover, pasture use and length of pasture period were shown to reduce the risk of mammary infections (Table 6) which is in line with results published in White et al. (2002), Fontaneli et al. (2005) and Firth et al. (2019), where pastured cows showed lower incidence of subclinical and clinical mastitis than confined cows. The lower risk associated with udder infections in pasture-based systems in our case might be related to better hygiene conditions on pasture compared to confined systems (Gösling et al. 2019). Further, an additional factor which might have favoured this outcome could be the fact that the prevalent housing systems in our trial were tie-stalls which were stated to have a higher incidence rate of clinical mastitis than free-stalls (Riekerink et al. 2008). The higher occurrence of reproductive disorders in farms with pasture access and longer pasture periods could be related to an insufficient energy supply in pasture fed cows compared to confined cows. Indeed in literature a lower BCS score was observed in pasture feeding systems compared to confined systems which might indicate a more severe negative energy balance resulting in a higher body fat mobilisation in pastured cows (Washburn et al. 2002; Hofstetter et al. 2014; Zendri et al. 2016). The resulting higher concentration of NEFA in blood was shown to negatively affect immune cell function and consequently increase the susceptibility for infectious diseases like metritis and subclinical endometritis (Ribeiro et al. 2013; Esposito et al. 2014). In addition, the risk for reproductive disorders was lower in full-time farms as farmers might have more time for animal control and observation than part-time farmers, who pursue additional employment outside the farm. Further, the higher occurrence of dry-off practice in low-input farms and AG cows could be related to a seasonal production pattern commonly used in pasture-based low-input systems. Finally, the risk for respiratory diseases was higher in BS than in AG which might be related to the higher adaptability of AG to local farming condition (Mattiello et al. 2011).

Antibiotic administration

In Austria, Firth et al. (2017), reported that antimicrobials were mainly administrated for treating udder diseases in 92% of the farms observed by which 79% used from WHO classified Highest Priority Critically Important antimicrobials (primary 3rd generation cephalosporin and cefoperazone) and High Priority Critically Important antimicrobials like beta-lactams. Similarly, Obritzhauser et al. (2016) described a wide use of Highest Priority Critically Important antimicrobials, corresponding to 24.6% of the total amount of antimicrobial applicated in Austria. In our study, the overall frequency of Highest Priority Critically Important antimicrobials applied was lower (18.52%), however, it should be noted that for udder diseases the use of Highest Priority Critically Important antimicrobials was with 14% of the total antimicrobials administrated in the present study relatively high. Therefore, in future treatments bacteriological test should be performed before actual treatment to identify the specific pathogen in order to formulate a respective treatment protocol which would enable a more targeted therapy. Latter could contribute to reducing the application of critical antimicrobial substances risky for causing resistance without minimising the success of the clinical outcome (Vasquez et al. 2017). Moreover, the use of selective dry cow therapy systems should be promoted (Cameron et al. 2015; Krömker and Leimbach 2017; Vanhoudt et al. 2018) as the use of antimicrobials for dry-off in our study was next to mastitis treatment the most common treatment-related to udder health management (Ferroni et al. 2020). In fact, according to the Italian guidelines on the use of antibiotics to contain the development of AMR bacteria that will be in force in 2022 (REGULATION (EU) 2019/6 OF THE EUROPEAN PARLIAMENT AND OF THE COUNCIL), treatments should be carried out on the basis of a specific diagnosis and CIA Highest Priority Critically Important antimicrobials may only be administered on the basis of an antibiogram which indicates that type of active ingredient as first choice antimicrobials or in the absence of other effective active ingredients. Moreover, some very important antimicrobials necessary for treating human infection might no longer be allowed for veterinary use. For this reason, it can be assumed that in the next future the types and methods of antibiotic administration will vary considerably, as will the veterinarians work in the study area. Overall, regarding udder health, special attention should be put on preventive strategies like hygiene, milking management, milking technique, feeding of Vitamin E and Selen or genetic selection for resistance for reducing the application of antibiotics (Ruegg 2017). Finally, the frequency of antibiotics applied for reproduction disorders (9%) was similar to that reported in Ferroni et al. (2020).

Conclusion

In summary, the most common veterinary treatments in our study referred to the udder, where BS had a higher occurrence for mammary infections than AG and HBS farms showed a significantly higher incidence for udder disease compared to all other farm categories. Pasture access revealed to be favourable for udder health but problematic in terms of reproduction disorders as special attention should be put on providing sufficient energy supply to the grazing animals in order to avoid severe negative energy balances resulting in decreased fertility. Further, selective therapies for mammary infections and dry-cow therapies as well as the implementation of preventive strategies in respective farm management should be promoted to reduce the risk for AMR.

Ethical approval

The experimental and notification procedures were carried out in compliance with Directive 86/609/EEC.

Author contributions

TZ analysed the data, drafted the manuscript, prepared the final report; EDM collected the field data; analysed the drug data and general diagnoses incidences, assisted in drafting the manuscript; MG designed the project, supervised all the data collection and data analysis. All authors read and approved the final manuscript.

Acknowledgements

We would like to thank the farmers for their valuable contribution to this project. The project was carried out within the framework of the Mountain Farming Action Plan of the province of South Tyrol (Aktionsplan Berglandwirtschaft).

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The data that support the findings of this study are available on request from the corresponding author, Dr. Thomas Zanon. The data are not publicly available due to privacy.

Additional information

Funding

This work was supported by the Open Access Publishing Fund of the Free University of Bozen-Bolzano.