A Systematic Review of the Biomechanical Effects of Harness and Head-Collar use in Dogs

The number of dogs in the UK is on the rise, as are canine sports involving the use of a harness to allow the dog to pull against an interface in the same way as a husky might pull a sled. Service dogs and those involved in essential work commonly wear a harness throughout their working lives, yet little is understood regarding the biomechanical impact of their use. This systematic review was conducted to review reported evidence of the biomechanical effects of harness and head collar (Halti) use in dogs. Searches were applied covering 1910 to 2018 on the following databases: PubMed, Web of Science and Writtle Discovery. Three publications were identified as suitable which were then critically evaluated using predefined criteria and ARRIVE based guidelines for bias assessment. Only one was considered to provide the most reliable data regarding the influence of harnesses on gait, whilst the remainder were considered to suffer a variety of issues including poor sample size, repeatability and study execution. The most appropriate study found that wearing a chest strap harness reduced shoulder extension in both walk and trot by up to 8° of movement, whilst a Y-shaped harness commonly marketed as non-restrictive reduced shoulder extension by up to 10° of movement, suggesting that the use of harness type restraints can affect canine gait, whereas no studies were found relating to the biomechanical effects of head-collar usage.

Non funded .  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60  61  62  63  64  65  66  67  68  69  70  71  72 Introduction 92 The canine population in the UK is currently estimated to be in excess of 9 million, whilst owner 93 expenditure is in excess of £10 million per annum [1]. A fundamental requirement of dog ownership is control 94 outside of the home, and owners spend even more time and money on puppy classes, obedience training 95 and behaviourists in the hope of having a sociable and obedient pet, yet nearly a quarter of dogs given up to 96 the Dogs Trust are there because of behavioural issues, such as a lack of control or aggression towards other 97 dogs and/or humans [2].

98
A common solution for owners when faced with an unruly dog is the use of a restraint such as a harness or 99 head collar (commonly known as a Halti), with manufacturers routinely advertising them on the basis of how 100 they can benefit the owner, using product names such as Non-Pull™ and Easy walk™. Training a dog is vital 101 in their early years and the foundation of correct behaviour [3] and harnesses are often used during the 102 training period or as a training aid. It is surmised therefore that an owner is more likely to use these types of 103 restraint when an animal is younger and relatively unruly, which raises questions regards their suitability and 104 possible impact on a developing musculoskeletal system and its associated growth plates.

105
Canine sports such as Canicross (also known as Cani-fit) and Bikejoring are also growing in popularity 106 in the UK, and these sports use harness systems to allow an animal to pull against an interface in much the 107 same way as a husky may pull a sled, utilising the canines instinct to pull against pressure [3]. Harness systems 108 of varying designs are also worn by all manner of service dogs, from guide dogs to search dogs and those 109 involved with armed forces and policing.

110
It is clearly appropriate that a dog is under control at all times, for its own safety and the safety or 111 others, yet there is very little discussion around the welfare consequences of using restraint devices, or 112 whether they may prevent walking at the most natural, biomechanically efficient gait. As such they may have 113 the potential to impact the dogs long term health and potentially compromise welfare.

114
If this proves to be true then the resultant costs may far out way any initial training expenditure needed to 115 negate the need for restraint devices -the cost of veterinary care continues to rise, with insurers paying out 116 on average £2 million per day for pet claims, an increase of nearly 56% in the last eight years [4]. Intervention (harness OR restraint* OR "head collar" OR "head-collar" OR halti OR "no pull" OR "no-pull" OR "non-pull" OR "gentle leader" OR "julius-k9" OR dogmatic OR ruffwear) OR "vest harness")

162
The full text of any remaining papers was then used to confirm suitability. Bibliographies of the 163 remaining papers were also used to identify any studies that were not located within the electronic search 164 A standardised model of data collection was then used as set out within PRISMA guidelines [16] to extract 165 key information from each of the included studies. definition, statistical methods, baseline data, numbers analysed, outcomes and estimation and adverse 176 events. All domains were then scored as either 1) low risk of bias 2) unclear risk of bias or 3) high risk of 177 bias and results were collated using excel to produce a graph which would indicate the total risk of bias for 178 the pool of papers as a whole.

179
In addition, papers included in the review were checked for evidence of conflicts of interest such as funding 180 from organisations that may gain from specific research results.

182
Results of the search and subsequent exclusions can be seen in (Fig 1) whilst the results extracted from 183 each study can be seen in table 4.

184
The three papers identified as suitable for review are as follows;    Y-shaped harness is designed to stop the animal pulling, so addition of weight was contrary to the design.
Skin displacement over joints during locomotion will have affected accuracy of results.
Treadmill can affect gait pattern. shoulder extension at both walk and trot, however the non-restrictive (Y-shaped) harness actually 200 decreased shoulder extension more than the chest harness, by an additional 2.56 0 reduction in extension at 201 walk and an additional 4.82 0 in trot. Full results are shown in table 5 and illustrated in (Fig 2).

210
There was insignificant loading of the spine from all three types of harnesses studied, as well as variable 211 loading of the shoulders as seen in (Fig 4). Data from the study can be seen in table 6.

224
The individual results of the ARRIVE Bias assessment of included studies are shown in (Fig 4). No 225 conflicts of interest were Identified. Two papers failed to report details of animal housing and husbandry including procedures to monitor test subject's welfare during the study, whilst the remaining publication 227 partially disclosed husbandry only. Two papers failed to fully discuss how treatments were allocated to each 228 test subject, although as they were cohort studies no randomisation was expected. Two papers also did not  interestingly this addition reduced shoulder extension even further. This was not consistent with both types 248 of harness, indicating that the shape of the harness could be a contributing factor as opposed to the load 249 pulling the limb caudally. It may also be that the dog shifts its centre of mass cranially to allow it to pull more 250 effectively, which is especially pertinent where canine sports such as canicross are concerned as the animal 251 is expected to be able to manoeuvre at speed, with a harness that is padded enough so as not to cause injury, 252 but thin enough to allow the limbs to move freely. The addition of 5 kg of weight is relatively light when compared to the potential forces caused by a runner attached via a bungee lead, especially if the lead is at 254 the end of its stretch capacity. One unexpected result is that a guide dog harness did not create pressure on 255 the dorsal spine, but this may be due to the handler needing to maintain contact by lifting the harness slightly  spinal movement it did conclude that a harness will impact lateral movement of the spine which adds an 275 additional dimension -a harness will need to allow for adequate flexion and extension, lateral bending, and 276 axial rotation of the spine, all of which will alter through changes in head and neck position at different gaits.

277
It would therefore seem logical that the larger or wider the harness, the more these will be impaired. As has 278 been noted skin displacement over anatomical landmarks during locomotion can lead to incorrect data 279 collection [22] so this would also need to be addressed in futures studies. No studies to date have explored any impact on gait when using a head collar and leash, which would be necessary if it is to be compared to 281 the suitability of harnesses.

282
Risk of bias was low, but none of the studies adequately discussed the housing and husbandry of the 283 test subjects, and almost all did not fully examine or record baseline data prior to any intervention which 284 again limits the validity of the results. policing and security services as mentioned above. This would also reduce the overall number of breeds that 305 would need to be studied initially as well as having the greatest potential impact. What is clear is that future studies will need to be of a sufficiently robust nature to be able to provide appropriate data, which has been 307 lacking in some of the research so far.

308
The only clinically relevant data that can be taken from this review is that shoulder extension is limited 309 by two of the most common types of harness. At present the use of relatively low-cost technology to assess 310 gait is still underutilised in veterinary practice, but what is clear is that quantitative analysis is the most 311 effective way of detecting biomechanical abnormalities as well as the underlying reasons. adopt an abnormal gait, as well as canine athletes who may be subject to the same restrictions but also be 328 expected to work at their maximum capacities.