A survey of diseases of working farm dogs in New Zealand

Abstract

AIMS: To record the incidence of specific diseases affecting working farm dogs; identify any apparent breed predispositions; determine the most important causes of loss from death, euthanasia, or retirement of the dog; and identify potential interventions that could reduce the incidence of disease.

METHODS: A national questionnaire-based cross-sectional study was conducted to identify the incidence of specific diseases, and circumstances of trauma affecting working farm dogs that presented to rural veterinary practices over a 12-month period, between April 2008 and April 2009. The practices were convenience-sampled on the basis of geography and size.

RESULTS: Sixty-six practices were approached, and veterinarians from 30 practices agreed to participate, with representation from all major rural regions of New Zealand. Data were provided on 2,214 presentations, excluding revisits, of which 51% were for Huntaways and 39% Heading dogs. Trauma was identified as a cause of injury in 848 (38%) visits. Huntaways were apparently over-represented in cases of constipation, gastric dilatation-volvulus (GDV), theriogenological problems, laryngitis, hip dysplasia, and degenerative lumbosacral disease. In contrast, Heading dogs were over-represented among cases of multiple ligamentous injury of the stifle, disruption of the gastrocnemius or Achilles tendon, tarsal injuries, and hip luxation. Traumatic injury involved injury by stock (20%), automotive incidents (19%), transit across fence lines (16%), and dog bites (12%). Loss occurred following 10% of visits, of which trauma was known to be involved in 32%. The most important nontraumatic causes of loss were GDV, degenerative joint disease, mammary neoplasia and diseases involving the female reproductive tract, cardiac disease, and poisoning.

CLINICAL RELEVANCE: Several important diseases appear amenable to intervention through nutrition, neutering, or behavioural modification, including constipation, GDV, theriogenological problems, dog-bite injuries, and laryngitis. Altering modes of transit across fence lines has the potential to reduce large numbers of serious orthopaedic injuries to working farm dogs.

KEY WORDS:

• Working farm dog
• Huntaway
• Heading dog
• disease survey

GDV	=	Gastric dilatation-volvulus

Introduction

It has been estimated that as many as 200,000 dogs currently perform work on farms across New Zealand, although no accurate census data have been collected (Dalton 2008). As some aspects of pastoral farming would be virtually impossible without working dogs, they are of paramount importance to the industry. Despite that, little research has been performed on any aspect of working dogs and their working conditions, with the exception of a few small surveys. Working farm dogs in New Zealand are dominated by two breeds, the Huntaway and the Heading dog. The Huntaway, averaging around 28 kg, is primarily used for driving a flock or herd from behind, and is required to bark frequently. The Heading dog, averaging around 19 kg, is used in a fashion similar to a Border Collie, namely to head off a flock or herd and control moving animals from the sides and front.

Specific disease surveys conducted by veterinary practitioners have previously suggested that certain diseases may be common. Hip dysplasia was found to be present in 22/93 Huntaways and 3/48 Heading dogs that were otherwise apparently healthy (Hughes 2001). Likewise, a survey of adult working dogs at national dogtrial meetings found retinal lesions in 178/371 (48%) dogs when examined ophthalmoscopically (Hughes and Joyce 1981). A survey of faecal samples from working dogs in the North Island identified 307/481 (63%) infected with coccidia, and that infection was much more common in rural than urban dogs (McKenna and Charleston 1980). Exposure to Neospora caninum was very common among working farm dogs, with 75% of 161 dairy farm dogs and 97% of 154 sheep/beef-farm dogs found to be seropositive (Antony and Williamson 2003). The seroprevalence of Leptospira spp. in dogs in New Zealand has also been found to be relatively high, although no comparison between urban and rural dogs has been made (O'Keefe et al. 2002). Reports of sporadic diseases in working dogs in New Zealand have included dilated cardiomyopathy, mucopolysaccharidosis, progressive myelopathy and neuropathy, and a photosensitivity in Collie working dogs (Jolly et al. 2000a b; Munday et al. 2006).

There has been no survey of general diseases of working farm dogs to determine relative frequencies, and it is likely that the literature from New Zealand is biased towards the unusual, and does not describe the common. At present, veterinarians and farmers alike are left to draw on personal experience, anecdotes, or professional opinion as to what is common, or important (Cave 2008). However, there is no general quantification of causes of loss from death or retirement in working dogs, or published incidences or causes of traumatic diseases, despite trauma being suspected as a leading cause of disease and loss.

The principle aim of this survey was to record the number of specific diseases affecting working farm dogs that were presented to participating veterinary clinics over a 12-month period, as a broad indicator of the incidence of those diseases during that period. Subsequent aims were to identify any apparent breed predispositions, determine the most important causes of loss, and to identify potential interventions that could reduce the incidence of disease.

Materials and methods

Recruitment of practices

There are currently 481 veterinary practices registered with the New Zealand Veterinary Council, although a large number of those are urban or strictly equine or production animal and do not attend to working farm dogs. The register does not distinguish between urban practices that attend to working farm dogs, and practices listed as ‘equine’ for instance, that do. Sixty-six rural veterinary practices throughout New Zealand were conveniencesampled from all those registered with the New Zealand Veterinary Council, using the following criteria. Participating practices were approached on the basis of geographical location within the known large rural regions of New Zealand (and excluding the main urban centres of Auckland, Wellington City, and Christchurch), and the size of the practice (more than two veterinarians). Selected practices were initially contacted by phone, and invited to participate in the study. Those practices that agreed to participate were enrolled during March 2008, and collection of data commenced during April 2008, and continued until 30 April 2009. The primary constraint to participation was the willingness of the practice, so convenience sampling was chosen over random sampling to most easily achieve a significant number of participating practices; no incentive was given to participating practices.

Collection of data

One questionnaire was to be completed for every working farm dog that was presented to the clinic during the period of study. The questionnaire was a single page, with three sections, covering basic details of the patient, descriptions of the farm, and details of the visit to the clinic (Supplementary Table 1 ¹ ). A pilot questionnaire was assessed by three local participating veterinary practices, and the format and some questions were modified in response to comments. Veterinarians were asked to be as specific as possible when making the diagnosis and defining the affected anatomical region. The questionnaire was to be completed by either the attending veterinarian or assisting veterinary support staff, and posted to Massey University, Palmerston North. The conduct of this study was deemed a low risk by the Massey University Human Ethics Committee, Palmerston North, New Zealand.

Management of data

Data from returned questionnaires were entered into a spreadsheet, and double-checked by two of the authors. Diagnoses/presenting problems were coded using a six-digit alphabetised code, denoting body system, region, and exact condition. Up to two different diagnoses were entered for each visit, and when more than two diagnoses were reported one was selected as the primary presenting problem and another as the most serious secondary problem.

Statistical analysis

Continuous data that were non-normally distributed were summarised using minimum, maximum and percentiles, while normally distributed data were summarised using mean and SD, using Minitab software v14 (Minitab Inc, State College PA, USA). Nominal data are presented as counts or percentages where appropriate. Dogs that were re-presented for the same condition more than once during the period of the study were defined as ‘true revisits’ for subsequent visits. Dogs that were described as ‘revisits’ by the clinic, but which had not previously had a visit recorded in the survey, were included among the original visits.

The number of original visits, excluding true revisits, was stratified by region and month. The results for region are presented as counts, and the percentage of all visits; the results by month are presented as counts.

For selected diseases, e.g. theriogenological, the total number of cases, and the number and percentage of cases that were Heading or Hunting dogs were determined. The median, minimum and maximum age for dogs presented with these selected reproductive diseases were also determined.

Visits were classified as traumatic or non-traumatic. In cases where the initiating circumstance was unknown, the visit was classified as unknown or, for diseases known to occur either as a result of trauma or in the absence of trauma, e.g. degenerative arthritis, the visit was defined as non-traumatic. For visits, the number of injuries was determined stratified by the location on the farm where the injury occurred, and the anatomical location of the injury. Results are presented as counts and percentages of total number of traumatic injuries.

Each visit was classified as either resulting in loss or not resulting in loss. A visit was classified as resulting in loss if it resulted in death, euthanasia, or retirement of the dog. The common nontraumatic causes of loss were determined stratified by body system and category of complaint. Results of non-traumatic loss are presented as counts and percentages of the total number of non-traumatic causes of loss. Similarly, the number of traumatic injuries resulting in loss were determined stratified by cause, and results are presented as counts and percentages of the total number of traumatic causes of loss.

Results

Participation

Of the 66 practices approached, veterinarians from 30 practices (45%) agreed to participate, with representation of all major rural regions of New Zealand (Table 1). At worst, this sample of 30/481 (6%) of all veterinary practices in New Zealand represents a much greater proportion of rural practices. After discarding incomplete questionnaires, data were available for 2,372 veterinary visits, which, after excluding true revisit consultations, included 2,214 original visits. Throughout the duration of the study, reported visits declined noticeably, stabilising between 100 and 150 visits per month for the last 5 months (Figure 1). Of the 2,214 original presentations, 1,024 were for non-traumatic diseases, 848 were for traumatic diseases, and the remaining visits were either unknown if trauma was involved, or were vaccinations.

Table 1. Geographic distribution of 30 veterinary practices, and the presentations (visits) of working farm dogs throughout New Zealand, from April 2008 to April 2009.

Region	No. of practices	Visits
		n	%
East Coast	2	677	28.5
Manawatu-Wanganui	5	440	18.5
Canterbury	4	343	14.5
Southland	2	325	13.7
Hawke's Bay	3	195	8.2
Marlborough	2	112	4.7
Central Plateau	2	77	3.3
Taranaki	1	63	2.7
Wairarapa	2	55	2.3
Otago	1	38	1.6
Auckland	2	22	0.9
Waikato	2	13	0.6
Northland	1	10	0.4
West Coast	1	2	0.1
Total	30	2,372	100

Figure 1. Number of reported visits of working farm dogs presented to 30 veterinary practices throughout New Zealand, from April 2008 to April 2009, and stratified by month.

Farms

Specific data relating to farms were not recorded to uniquely identify each property, and thus several dogs could have been presented from a single property. Of the 2,372 visits, 2,128 were for dogs from sheep-and-beef farms, 111 were from dairy farms, 55 from sheep-only farms, five from beef-only farms, and 73 from farms that were classified as other, or undefined.

Dogs

Of the 2,214 original presentations, complete signalment of the dog was recorded for 2,198 visits, which included 51% Huntaways, 39% Heading dogs, 4% Beardies, 3% Crossbreeds, <1% each of Blue Heelers and Smithfields, whilst the remaining dogs were ‘other’ (2%). Of the 2,198 dogs, 54% were entire males, 41% entire females, 2% castrated males, and 3% spayed females. There were no significant differences in sex distribution between the six main breed categories. A total of 370 different diagnoses were recorded over all 2,214 original presentations, and with few exceptions only diagnoses recorded on =15 occasions are presented here. However, in many cases the forms were completed at the beginning of the visit, therefore the information provided described the presenting clinical problems rather than the final diagnoses.

Non-traumatic diseases

After excluding revisits, 1,024 visits were described as being unrelated to trauma. Of those, 200 (9% of all original visits) were related to the gastrointestinal tract. Constipation was the single most common diagnosis (n=51), affecting 31 entire males, 16 entire females, three neutered males, and one of unknown sex; Huntaways were over-represented (n=43) compared with Heading dogs (n=4). GDV was the next most common diagnosis (n=36, comprising 33 Huntaways, two Heading dogs, and one Beardie; mean age 6.1 (SD 2.6) years). Acute onset of vomiting and/or diarrhoea was the reason for presentation of 32 dogs in which a specific diagnosis was not made. Of those, 10 were described as being severe, with dysentery, haematemesis, or fever. Parvoviral enteritis was suspected in a further six cases. Foreign bodies were removed from the stomach of five dogs, and the small intestine of nine dogs. The foreign bodies were identified as ear tags (n=6), bones (n=2), corn cobs (n=2), peach stone (n=1), stick (n=1), and two were unidentified. Intestinal parasitism was only reported in two cases. Dental or oral-cavity disease was reported for 39 visits, which consisted mostly of tooth-root abscesses and fractured teeth requiring extraction (n=23).

Non-traumatic skin disease was a reason for 149 visits, of which 96 were for abscessation or cellulitis. In only four cases was a foreign body identified (all grass seeds), whilst the remaining cases were of unknown or unreported cause. Abscessation or cellulitis was confined to the abdomen (n=20), head (n=16), foot (n=13), leg (n=18); and neck, thorax or other (n=29). Cutaneous parasites were identified at 23 visits (n=11 Demodex spp., n=6 fleas, n=4 Otodectes spp., and n=2 Sarcoptes spp.). Non-parasitic pruritic skin disease was reported at 54 visits, and those included otitis externa (n=30), pododermatitis (n=6), suspected atopic dermatitis (n=9), and generalised pruritis (n=9).

Problems relating to the reproductive tract were responsible for 197 visits, including 108 Huntaways and 63 Heading dogs (Table 2). Huntaways appeared over-represented for pyometra/endometritis, vaginal prolapse, and vaginal hyperplasia. For mismating, spaying was performed on eight dogs, medical therapy was given to 28 dogs, and the treatment for the remaining five was unrecorded. Benign prostatic hyperplasia was diagnosed in 15 dogs, and was recognised as a cause for constipation in six dogs. In contrast, prostatitis was only diagnosed in two dogs.

Poisonings were identified or suspected at 36 visits, and of those identified included anticoagulant (n=26), metaldehyde (n=3), and one each of ivermectin, organophosphate, rumensin, and paraquat. Seizures unrelated to known toxicity were a cause for presentation on eight occasions. Systemic infectious diseases included leptospirosis (n=5), and tetanus (n=4). Of the five dogs suspected to be infected with Leptospira spp. all were in the central to lower North Island (Taranaki, Hastings, Manawatu, and Central Plateau).

Respiratory disease was a relatively uncommon cause for presentation, and included predominantly laryngeal disease, diagnosed as laryngitis in 18 cases, and reported to be associated with dysphonia in nine, and laryngeal paralysis in two dogs. All cases of laryngeal disease were seen in Huntaways, and one Beardie (dysphonia). Dilated cardiomyopathy was only diagnosed twice, both in entire male Huntaways. Congestive heart failure was seen in a further eight dogs, which was associated with fever in two cases. All six of the non-febrile cases of congestive heart failure were Huntaways, all but one of which were entire males.

Table 2. Median age at presentation, and incidence of the most common reproductive conditions in Heading dogs and Huntaways presented to 30 veterinary practices in New Zealand, from April 2008 to April 2009. Percentages refer to the proportion of all dogs presenting with that condition.

Condition	n	Age (years)		Breed
				Heading dog		Huntaway
		Median	Range	n	%	n	%
Pyometra/endometritis	17	8.0	2.0–12.8	5	29	11	65
Mammary neoplasia	27	8.5	1.3–14.0	7	26	15	56
Vaginal prolapse	18	2.0	0.6–9.0	0	0	18	100
Vaginal hyperplasia	7	2.0	1.5–3.3	0	0	7	100
Mismating	41	5	1.0–11	19	46	14	34
Dystocia	15	6	2.0–11	7	47	8	53

The most common non-traumatic musculoskeletal condition was degenerative arthritis, which was reported for 129 visits. It was localised to the coxofemoral joint (n=23), carpus (n=18), elbow (n=7), and hock (n=3), whilst the remainder were of an unspecified location. Other specific diseases were defined as lumbosacral disease (n=38), toe infections (n=18), and hip dysplasia with or without degenerative arthritis (n=23). Huntaways were over-represented in cases of hip dysplasia (n=16) compared with Heading dogs (n=5), and in cases of lumbosacral disease (n=27, compared with nine Heading dogs). Ataxia was reported in four cases; one resolved, one was associated with fever, and in the other two, intervetebral disc disease was suspected.

Traumatic diseases

A total of 848 visits were described as being the result of trauma, excluding revisits. Heading dogs were involved in 440 and Huntaways in 367 traumatic incidents. The general body regions injured, and the location on the farm where the injury occurred, are presented in Table 3. For the six geographic regions from which more than 100 visits were submitted, trauma was related to between 25 and 46% of all original visits.

Of the 260 cases of foot trauma, foreign bodies were responsible in 43 cases, of which 19 were interdigital. Other soft-tissue foot injuries included contusion of the skin or pad (n=41), metacarpal or metatarsal fractures (n=30), phalangeal fractures (n=15), and unspecified trauma to the toe (n=18). Toe infections secondary to trauma were diagnosed 50 times, and amputation of the toe was performed on 34 occasions.

Cruciate-ligament injuries were seen on 65 occasions, 35 in Heading dogs and 25 in Huntaways. Of those 65, multiple injuries to the stifle ligament (cranial cruciate plus one or two collateral ligaments) were seen in 18 dogs, of which 13 were Heading dogs and three were Huntaways, and 12 of which occurred during transit of a fence. Incidents leading to injury of the cruciate ligament were associated with fences in 20 (31%), paddocks in 17 (26%), stock yards in five (8%), and on roads in two (3%) cases, and was unknown or unreported in 21 cases. Thus, when the location was known fences were associated with 45% of injuries of the cruciate ligament, and 39% occurred while dogs were in paddocks. Of the fence-associated injuries, nine were Heading dogs and 10 were Huntaways. Disruption of the gastrocnemius or Achilles tendon was reported in 22 cases, of which 16 were Heading dogs and six were Huntaways. Tarsal injuries were recorded in 32 cases, of which seven were fractures, and 25 were luxations of the tarsus or tarsometatarsus. Of those 32 tarsal injuries, 23 were in Heading dogs, whilst eight were in Huntaways. Hip luxation was diagnosed on 27 occasions, 17 times in Heading dogs and seven times in Huntaways.

Table 3. Summary results for working farm dogs presented with traumatic injuries at 30 veterinary practices in New Zealand, from April 2008 to April 2009, showing the location on the farm where the injury occurred, and the anatomical location.

Category	Location	Injuries
		n	%
Location on farm (n=553)	Paddock	289	52.2
	Fence	131	23.7
	Yards	72	13.0
	Road	36	6.5
	Race	23	4.2
	Shed	2	0.4
Anatomical location (n=655)	Foot	260	39.7
	Stifle	100	15.3
	Tarsus	68	10.4
	Head	58	8.8
	Abdomen	54	8.2
	Hip	31	4.7
	Thorax	26	4.0
	Carpus	22	3.4
	Elbow	19	2.9
	Spinal	12	1.8
	Shoulder	5	0.8

General causes of trauma, and the outcome of the traumatic incident are presented in Table 4. Trauma caused by livestock involved Heading dogs for 91 and Huntaways for 66 visits, and led to a wide variety of injuries. Of the incidents of automotive trauma, 74 (47%) were from vehicles hitting or running over dogs, 41 (26%) from dogs falling off vehicles, 29 (18%) from dogs jumping off vehicles, and 13 (8%) were unknown. Thirtyeight of these incidents involved Heading dogs and 24 involved Huntaways, and a wide range of injuries ensued, with cutaneous trauma being the most common (n=19). Fences were associated with injury to Heading dogs in 67 and Huntaways in 61 cases. Injury was confined to the skin in 75 cases, whilst 65 cases involved musculoskeletal injury. The most common musculoskeletal injury associated with fences was injury of the cranial cruciate ligament in 20 cases, of which 11 also included injury to at least one collateral ligament. Tarsal fracture or fracture-luxation occurred in nine cases, and femoral fracture occurred in five cases involving fences.

Table 4. Traumatic incidents occurring to working farm dogs presented to 30 veterinary clinics in New Zealand, from April 2008 to April 2009, stratified by general cause. A visit was classified as resulting in loss if it resulted in death, euthanasia, or retirement of the dog.

Category	Total visits		Visits resulting in loss
	n	%	n	%
Unknown	260	30.7	1	2
Injured by stock	170	20.0	21	41
Automotive	159	18.7	24	47
Fence	140	16.5	0	0
Dog bite/fight	99	11.7	0	0
Fall (no vehicle)	16	1.9	4	8
Gunshot	4	0.5	1	2
Total	848	100	51	100

Bite injuries included 43 Heading dogs and 41 Huntaways, and comprised 74 (75%) entire males, 18 (18%) entire females, four (4%) spayed females, and two (2%) castrated males, with one undefined.

Outcome

Of the 2,372 visits, 1,958 (82.5%) resulted in the dog returning to work. The dog was rested following 206 (8.7%) visits, and the fate of the dog was unrecorded or unknown for 306 (12.9%) visits. Loss was reported following 270 (11.4%) of all visits; it resulted in death or euthanasia in 189 of those visits (8.0% of total visits). Trauma was involved in 86 (31.9%) cases of loss, and the prominent causes were injury by stock (21/86; 24% of traumatic loss), being hit by a vehicle (13/86; 15%), falling from a vehicle (5/86; 6%), jumping off a vehicle (2/86; 2%), gunshot (1; 1%) and unknown (1; 1%). The most common non-traumatic causes of loss are presented in Table 5.

Discussion

Described here is the first large-scale investigation of diseases affecting working farm dogs in New Zealand. Data were successfully collected from 30 veterinary practices over a 12-month period. Practices were sampled based on two criteria, viz location and size. Location of the practice was important because we wanted to ensure that the catchment area included a population of working dogs. Although there are not yet accessible data on the true number or geographical distribution of working farm dogs in New Zealand, the geographical reach of this survey is thought to have reasonably represented that distribution. Regions that were probably under-represented include the Waikato, Northland, and Wairarapa. The willingness of the practices was also considered important because collection of data required a high level of compliance; reluctance to provide the information could have compromised the quality of the data.

Sampling veterinary practices using non-probability methods was associated with a risk of bias (Thrusfield 2007). Due to limited availability of information about veterinary practices in New Zealand it was not possible to compare those that participated in the study with either those that were selected and did not participate, or the general population. Therefore the bias created by the sampling method cannot be quantified. Despite this we believe the data make a significant contribution to knowledge because until now farmers and veterinarians have relied on anecdote, dogma, and experience to decide which are the numerically important diseases, and which are the most important causes of loss. Many of the data could be considered intuitive or expected, however several results were discordant with our expectations.

Table 5. The most common non-traumatic causes of loss to working farm dogs presented to 30 veterinary clinics in New Zealand, from April 2008 to April 2009. A visit was classified as resulting in loss if it resulted in death, euthanasia, or retirement of the dog. Percentage refers to the proportion of all cases of loss.

System	Disease	n	%
Gastrointestinal disease	Gastric dilatation-volvulus	11	4
	Gastric ulceration/gastritis	3	1
	Constipation	5	2
	All causes	36	13
Degenerative joint disease	Hip	9	3
	Carpus/elbow	6	2
	Other	3	1
	All causes	18	7
Reproductive tract disease	Pyometra/metritis	3	1
	Mammary neoplasia	6	2
	Dystocia	3	1
	Mismating	3	1
	All causes	15	6
Cardiac disease	All causes	8	3
Poisoning	All causes	7	2

Submission of completed survey forms declined steadily throughout the year. Possible reasons for this decline include waning interest, changing workloads across the year, or it could have been a true reflection of the clinic admissions during the year. The smallest number of submissions in a month was 100, which occurred during December. Thus, even during the leanest month a large number of submissions continued to be made.

It is conceivable that different work demands, stock behaviour, and terrain would produce different incidences of disease, especially for traumatic injuries. However, when geographical differences were assessed there were no significant differences. Care should be taken not to interpret this as evidence of no geographical differences, as the survey captured no information about the size of the underlying population. Furthermore, the low number of dairy farms represented here prevented any meaningful comparisons between the diseases seen in dogs from different farming types, e.g. dairy vs sheep and beef.

Huntaways and Heading dogs remain the dominant breeds in rural New Zealand, and were fairly equally represented, making apparent differences in disease incidences between the breeds more noticeable. The survey was constructed to contain only six choices of breed, and an ‘other’ category, so more refined definitions for varieties of Heading dogs, Huntaways and Beardies were not provided. Future studies could evaluate if varieties of breed have different predispositions for diseases, which could assist empirical breeding decisions for heritable diseases such as hip dysplasia.

Non-traumatic disease was both a more common reason for presentation and a more common cause of loss than was expected. Diseases of the gastrointestinal tract accounted for 9% of all original presentations, and constipation was responsible for a quarter of those. Variables that are likely to contribute to the large incidence of constipation include ingestion of bone or hair, dehydration, insufficient opportunity for frequent defaecation, and variable feeding frequencies. The apparent over-representation of Huntaways suggests a breed difference in colonic function, as the other variables are unlikely to be significantly different between the two main breeds. If that is the case, it suggests potential for breed-specific nutrition. GDV remains an important cause of suffering, loss, and expense in Huntaways but predictably not in Heading dogs. Previous studies have identified rapid eating of large volumes, elevated feeding and increasing atmospheric pressure in the previous 2 days in the development of GDV (Theyse et al. 1998; Glickman et al. 2000; Levine and Moore 2009). It is likely that genetics and anatomy, feeding practices, exercise demands, and housing all differ between Huntaways and the dogs described in previous studies of GDV. There is, therefore, a serious need for determining which risk factors for the development of GDV apply to working farm dogs in New Zealand.

Acute gastroenteritis with evidence of severe mucosal damage (haematemesis, fever, dysentery) was present in at least 16 cases. Although parvoviral enteritis was only reported as suspected to be involved in six cases, it is possible that parvovirus remains an important if sporadic cause of enteritis in working dogs. Other infectious causes could include Salmonella spp., Campylobacter spp., Clostridium perfringens, coronavirus, and enterotoxagenic Escherichia coli (Morse and Duncan 1975; Brown et al. 1999; Cave et al. 2002). Non-infectious causes are likely to include toxicity, such as Amanita mushrooms (Amanita phalloides), and heavy metals. The nature of disease caused by leptospirosis infections in dogs in New Zealand has not been well-described, and could potentially be responsible for some cases treated empirically as ‘parvoviral enteritis’. In the lower North Island, 14% of 466 dogs were found to be seropositive to leptospiral serovars (O'Keefe et al. 2002). In the current survey, only five dogs were suspected of being infected with Leptospira spp. Renal disease was a rare cause for presentation, and acute hepatopathies were also surprisingly uncommon. Thus, despite the relatively high seroprevalance of Leptospira spp. in dogs in New Zealand, diseases classically associated with infection were rarely seen in those dogs predicted to be most at risk, i.e. dogs with a high chance of exposure to carrier species such as rats, cattle, and pigs. The efficacy of vaccination could be responsible, but atypical disease, or the dominance of low- or non-pathogenic serovars could equally explain these findings.

Allergic or pruritic skin disease was reported much less frequently than expected. The inclusion of German Shepherd dogs and Border Collies in the heritage of Huntaways and Heading dogs is predicted to make them predisposed to atopic dermatitis (Scott and Paradis 1990). Of the reported cases, otitis externa dominated but pododermatitis was not commonly reported. It is possible that this survey demonstrated the true unimportance of these diseases in working dogs in New Zealand, or it could simply be the result of under-reporting, or the low priority placed on such diseases by farmers. Otitis externa and pododermatitis can seriously affect the working performance of a dog, whilst pruritis in other areas may not. Equally, the finding of localised or generalised skin disease in dogs during presentation for other more serious diseases may not have been recorded by attending veterinarians. It is therefore hard to ascribe the relative importance of allergic skin disease in working farm dogs, although the small number that presented for pododermatitis suggests that component is minor, since serious cases would likely impair working, and instigate presentation.

The overwhelming majority of working farm dogs in New Zealand are sexually intact. Breeding of these dogs tends to be delayed until well into adulthood, once the performance of the dog has been established. If a decision to breed is not made until a dog is 4–5 years old, and there is a strong interest in producing several litters, the dog may be kept entire until relatively late into its life. It is not surprising then that theriogenological problems were common. Mismating was the most common category, and its frequency suggests there is need for better management or education to reduce this largely preventable occurrence. Huntaways were apparently over-represented in mammary neoplasia, pyometra, and what was described as vaginal prolapse or vaginal hyperplasia. The median age of dogs diagnosed with vaginal prolapse was 2 years, which is in contrast to current understanding of the condition as one of older dogs following parturition (Soderberg 1986; Ragni 2006; Alan et al. 2007). It is possible that the Huntaway is highly unusual in its susceptibility to non-parturient vaginal prolapse. However, it is suspected that the description of vaginal prolapse may have been misapplied to cases of vaginal hyperplasia, which is usually seen at the time of the first or second oestrus (Soderberg 1986). None-the-less, the Huntaway bitch appears more susceptible to mammary neoplasia, endometritis, and vaginal hyperplasia/prolapse than the Heading dog.

Anticoagulant toxicity was responsible for the majority of poisonings, whereas other causes were sporadic. Although organophosphate and paraquat toxicity would have once been more common, they are now rare. Toxicity from abermectins was expected to be more common because of the ready access, use of products off-label, and potential for the multi-drug resistance-1 gene deletion mutation associated with ivermectin sensitivity in Collies and their descendants (Neff et al. 2004). However, either abermectin toxicity does not occur, is rapidly fatal, or not recognised as such.

Despite the athletic demands, exposure to inhaled aero-allergens, and likelihood of suboptimal nutrition and stress in working farm dogs in New Zealand, disease of the lower airway and lung parenchyma was uncommon. Laryngitis and dysphonia were relatively common, and seen almost exclusively in Huntaways. Possible causes for laryngitis in dogs include hypersensitivity, bacterial infection, viral infection, gastro-oesophageal reflux disease, and neoplasia, whilst dysphonia without inflammation can result from laryngeal hemiplegia (Yamamoto 1966; Wheeldon et al. 1982; Venker-van Haagen 1992; Jackson-Menaldi et al. 1999). However, excessive barking (hyperphonation) can lead to vocal damage and temporary vocal loss. Thus, the propensity for Huntaways to experience vocal loss may be associated with specific disease, such as laryngeal hemiplegia, but is more likely to be associated with overuse, so is preventable through appropriate management.

Non-traumatic musculoskeletal disease was dominated by degenerative arthritis, many cases of which could probably be of traumatic origin. Hip dysplasia and lumbosacral disease were apparently more common in Huntaways, consistent with previous suggestions (Hughes 2001). Of the non-traumatic diseases of working farm dogs in New Zealand reported sporadically, such as mucopolysaccharidosis, progressive myelopathy, and Collie photosensitivity, only cardiomyopathy was reported during the period of this study.

It had been assumed that traumatic injuries would be responsible for the large majority of cases of loss. However, only 32% of visits culminating in loss were known to be the result of trauma. Although many more were probably traumatic in origin, the significance of non-traumatic disease was surprising. None-the-less, trauma was still a prominent cause of disease in this population.

Foot trauma was by far the greatest single category of trauma, and was dominated by cutaneous injury, with or without secondary infection. More serious was injury to stifle ligaments, which appeared slightly more common in Heading dogs than Huntaways. Of even greater note was the apparent over-representation of Heading dogs amongst dogs with multiple injuries to stifle ligaments. This disparity suggests either different dog or work behavioural risks, or both. Risks to a dog for cruciate disease include angulation of the tibial plateau, a narrow tibial tuberosity, breed, and large bodyweight (Duval et al. 1999; Witsberger et al. 2008; Buote et al. 2009; Inauen et al. 2009). Work behaviour is likely to be different between breeds with regard to stifle injuries because of the over-representation of Heading dogs in the category of multiple injuries to ligaments. Such cases were presumably the result of extreme pathological forces applied to the stifle, whereas a single injury is likely to involve conformational differences (Bruce 1998; Schwandt et al. 2006). The majority (67%) of multiple injuries to stifle ligaments occurred during transit of a fence, and all nine Heading dogs suffering this injury during such transit incurred multiple injuries to the ligaments. Despite that, all fence-related injuries were evenly distributed between the two breeds. This finding does not, however, allow discrimination between behaviour, e.g. style of jumping or speed, and anatomical or physiological differences. Heading dogs were again apparently over-represented in injuries to the Achilles mechanism or tarsal joints, fractures and fracture-luxations of the tarsus, and hip luxations. It is suggested that this propensity for serious musculoskeletal injuries in Heading dogs is a consequence of their faster working behaviour and their increased running requirements during heading activities. It is important that the true risk factors for these types of injuries be established because these are serious injuries that cause significant suffering and expense for the farmer. If dog factors such as anatomical or physiological limitations are key, then selective breeding can be directed towards lower-risk dogs. If working behaviours are key, then alterations to the way the dogs are worked could prevent injury.

There was no apparent breed bias to dog bites, but entire males were the recipients in the overwhelming majority. It is a matter of opinion as to how many bite wounds are inevitable, or acceptable. However, proper socialisation, housing, training, and control will influence the incidence. As these are aspects of the dog's use that can be improved, they should be, where possible. In clinical practice, judging the efficacy of those management variables in a farming client is practically and professionally difficult. The veterinary profession should decide what role we could or should play in improving behavioural management of dogs.

The paddock was the location on the farm where the greatest number of traumatic events (52%) was known to have occurred, whilst the yards, often presented as an area of greatest risk, was where only 13% of events occurred. Movement through fences is clearly an important risk for injury, often serious injury. This was an important, if predictable, finding because it is an area that is amenable to intervention. This survey did not obtain sufficiently detailed information about the nature of the injuries associated with fences, and further work is warranted to elucidate the risk factors. Injury during transit of fence lines can occur when the dog crawls beneath a barbed-wire strand, jumps between strands, or jumps over the fence. Modifications to fences to provide safe ‘transit points’, training dogs to cross fences in the safest manner, and encouraging movement through gates when possible all appear intuitively sensible recommendations. However, until there is clear evidence as to the most important risks associated with transit, recommendations have the potential to be ineffective, impractical, or even detrimental.

It is notable that the incidence of several important diseases could be reduced by improved nutrition, neutering, behavioural modification or altered training . Improvement in feeding regimes is likely to reduce the incidence of constipation, GDV, and might reduce the occurrence of acute gastroenteritis. Nutritional influences on soft-tissue and skeletal injuries have been established in dogs during training, and determining what constitutes an ideal diet for working farm dogs could improve athletic performance and reduce the incidence of serious injuries (Reynolds et al. 1999). Neutering would be predicted to have a significant effect on several important diseases in this population, including mammary neoplasia, pyometra, and mismating. The over-representation of entire males in dog-bite injuries suggests that, too, might be amenable to neutering. As commented above, however, it is unlikely that recommendations to neuter dogs early will be adhered to because of the widespread desire to breed from proven dogs. Behavioural modification or improved early training has the potential to reduce aggression, and could reduce unnecessary barking that can contribute to dysphonia in Huntaways.

There are important limitations regarding this study that should be kept in mind. Although large, this dataset represents a convenience sample, rather than a truly random sample. Care should be taken to define the denominator for each dog or disease variable considered. It is argued that the relative disease frequencies presented here are close to those in the entire population of working farm dogs, however due to the presentation of some dogs more than twice, and the potentially biased submission of cases, the denominator should remain ‘visits’, rather than ‘dogs’. This survey was designed to describe the most common reasons for working farm dogs presenting to veterinarians, so no attempt was made to exclude dogs if they were presented on more than one occasion for different diseases. Additionally, there are several real and potential biases in this dataset. Selecting veterinary visits rather than truly surveying dogs biases away from those dogs not presented to a veterinarian for disease, or the owner's reasons. Participating veterinarians were asked to submit forms for every dog presented that met the inclusion criteria. However, there was likely to be a bias towards more interesting, unusual, or serious diseases, and away from more mundane diseases such as chronic dermatopathies. That may explain the relative unimportance of non-parasitic pruritic skin disease in this population. Submissions reduced during the study, which is expected in a long-duration study such as this. Seasonal changes in relative disease incidences were not able to be truly assessed because we only included one calendar year, and there was a large variation in the number of submissions each month.

Despite the limitations, this study is the first to record a large number of diseases of working farm dogs in New Zealand. Nontraumatic disease was more common than expected, but trauma remains an important challenge to a dog's welfare. Notable breed differences were seen between the two dominant breeds, the Huntaway and Heading dog. Huntaways were apparently over-represented among cases of constipation, GDV, pyometra, mammary neoplasia, laryngeal disease causing dysphonia, hip dysplasia, and degenerative lumbosacral disease. Heading dogs were apparently over-represented among cases of multiple injuries to stifle ligaments (not associated with fences), tears of the Achilles or gastrocnemius apparatus, tarsal-tarsometatarsal fracture- luxations, hip luxation, and stock-associated injuries. Some of these breed differences may be due to differences in working behaviour, but anatomical or physiological variables are also likely to be involved. More focussed research is clearly indicated in this understudied population, aimed at prevention of traumatic injuries related to transit of fences, dietary influences on constipation, risk factors for GDV, the effect of nutrition on injury and performance, heritability of the most important diseases, and the efficacy of training or behavioural manipulation on diseases such as laryngitis, and dog-bite injuries. Clearly, there is much to be done to improve the health and welfare of this vital component of our agricultural industry.

Acknowledgements

We would like to sincerely thank all the veterinarians and veterinary support staff who participated in this survey. This study was funded internally.

Notes

http://www.sciquest.org.nz/node/60926