Experimental design

Why did you use naïve horses and double bridles?

Designing a single study so that it is as informative as possible to as many stakeholders as possible is always a challenge. As with all good science, one ends up with more questions than answers. So, the need for further work to address many other variables is clear. These variables and our bid to control at least some of them are considered below.

Our decision to use naïve horses was driven by our need to abide by the principles of the 3R’s approach to animal experimentation (Replacement, Reduction and Refinement) and reduced variability by using naïve horses. This reflects the need to avoid habituation effects because, despite being denied so many normal behaviours, horses may habituate to wearing a tight noseband (although this has yet to be shown).

Arguably, the loss of ability to swallow, lick and chew should probably be broken down into three further variables, so that we can report on which of these comfort behaviours horses are most motivated to perform. However, designing a means of denying horses each of these separately is extremely challenging. There seems no likelihood of the horses wearing a tight noseband being able to retain these behaviours since the tightened nosebands are designed, at least in part, to eliminate jaw movement. Of course, a further variable would be duration of treatment since in this study we applied the treatment for only ten minutes and acknowledge that most horses undergo the restrictions for far longer under normal and competition conditions.

Then, one must add variable rein, poll and curb chain tensions, that were entirely absent in the current study. The stress response reported for the tight nosebands is clearly linked to the inhibition of oral behaviors but we cannot unpick whether it comes from the pressure, the deprivation or both. So, with so many variables to consider, before we can say that we have the exact data to comment on what can be achieved through best practice in the use of restrictive nosebands in elite dressage, we’d need much more time and resources.

It is not clear how long trainers generally wait before introducing the second bit and the tight nosebands nor how long horses take to become less reactive and whether they recover the ability to swallow. On face value, the prospect of horses recovering the ability to move their tongues when nosebands are compressing them appears highly unlikely, given the need to move the tongue freely in normal deglutition. The question is how long horses can cope with being unable to swallow.

The current experiment shows changes in physiology and behavior when the noseband is tightened and the bridle is removed. Each of the treatment levels were exposed to the same regime, meaning that bridle-only effects would be clear.

There is certainly a bridle-effect, as we discuss in the paper. When one examines the swallowing in Figure 6. All of the groups increase swallowing when the bridle is removed. However, the blue, Unfastened Noseband, group appears to have the least obvious response in that the line is relatively straight. The Conventionally Fastened (CAUN) group also swallow more during recovery but they appear to swallow just as much as the UN group during the treatment phase. It is clear that both the HCAUN treatment and the tight noseband treatment (NAUN) both reduce swallowing greatly in comparison to the two other treatments. Horses in both these groups swallow less during treatment than baseline, unlike the UN and CAUN groups. When the bridle is removed both groups increase their swallowing significantly above the baseline readings, indicating a post-inhibitory rebound effect.

This graph may demonstrate a ‘bridle only’ effect. When one examines the UN and the CAUN treatments, swallowing increased during the 10min treatment phase for both groups (significantly for the CAUN group), possibly indicating the bits increase the desire to swallow, and again when the bridle was removed – indicating a slight, but not significant, increase in swallowing as a response to the bits being removed. Whereas both the HCAUN and NAUN decreased swallowing during the 10min treatment phase (the NAUN group coming close to zero swallowing) and were the only groups to show a significant increase in swallowing between the treatment and recovery sessions.

If we take one of the physiological measurements: heart rate, shown in Figure 2, we see a similar effect. Baseline readings were taken with the bridle on and recovery readings with the bridle off. Any bridle-only effects would show up here. If the two bits were causing a physiological effect, then we should see the heart rates decrease during the recovery phase but this was not the case.

Fundamentally, of course, we are looking here for changes in the readings. As you can see, each treatment resulted in a relatively flat graph line, apart from the NAUN treatment. The NAUN treatment began at a normal baseline reading and went significantly higher during the treatment phase. The NAUN heart rate did not return to normal during recovery, when the bridle was removed, but it may have required more than 10 minutes for that to occur. A similar peak and recovery are evident in the graph on eye temperature.

We note the need to monitor the horses for longer in any further studies, if indeed the animal ethics committee were to approve them and if they were to be funded.

This graph shows that the introduction of the two bits to naïve horses, in conjunction with a correctly fitted noseband, is not detrimental. According to these physiological data, the only problem arises when the noseband is tightened so that there is no space underneath it.

Should the experiment have included measurements of pressure?

It is true that tightness and pressure are not the same thing. In a pilot study, Casey et al (2013) showed how the CAUN treatments reported here relate to pressures under the noseband. For what it is worth, the reported pressures ranged from 200 to 400 mm Hg;
pressures that, in humans, are associated with nerve damage and other complications.

Using a practical measure of area under the noseband was the most important consideration for the current study. Our decision to use the ISES taper gauge was predicated on it being available to riders and it having been used in previous studies (McGreevy et al 2012) and a soon-to be released audit of noseband use at FEI events. Unfortunately, at this stage, we do not have the technology available to test the resulting pressure under the noseband. However, we are aware of some formative work in the domain and the prospect of a patented probe for measuring noseband tightness. For the purposes of the current study, we simply used the noseband taper gauge to measure the space under the noseband. It would seem difficult to argue that there is not relentless pressure when there is no space under the noseband, especially when these nosebands are often tightened beyond that point, i.e. cranked.


Results

 Are 12 horses enough for the study to be valid?

The statistical tests show that the number of horses we studied was ample. Each horse acted as its own control. The values we list as significant are at the level of p <0.05. In essence, this means that the chances of the result being a fluke are less than 5%.

Are outliers factored into the analysis?

The statistical tests factored-in the effects of any outliers.

Is it possible to get more data?  

The raw data are available online.

Where are the standard deviations?

Standard deviations appear as s.e.d. values into each of the tables containing mean data.

Why do the eye temperature and cardiac responses not correlate?

We only found that the eye temperature and heart rate did not correlate on a minute-by-minute basis. The importance of eye temperature in animal welfare science is growing because eye temperature often increases during aversive procedures. However, absolute synchrony between eye temperature and other physiological stress responses has not been reported, to our knowledge. In the current study, eye temperatures did correlate with the intensity of treatment levels, see Figures 2 and 4.

Looking at both eye temperature and heart rate, we can see that only the NAUN treatment had a significant effect. It does appear that the heart rate is slower to return to normal than the eye temperature after a period of stress and the exact timing of this merits further investigation. Both measures revealed a significant stress response during the NAUN treatment and during none of the other treatments.

The scale of the increase in eye temperature is consistent with events such as jugular catheterization of dairy cows (Pierard et al., 2015) and horses subjected to the acute stress effects of showjumping (Valera et al., 2007).

In Fig 4, the change in eye temperature for NAUN appears only just above that for UN.  Does this mean that CAUN and HCAUN were less stressful than UN?


The change in eye temperature is the important effect to observe here (see Figure 4). While the change in NAUN appears to be only just above UN, that is not the case. The only significant change for eye temperature emerged in the NAUN group, where eye temperature remained significantly high throughout treatment and returned to normal during recovery. No other group showed a significant change in eye temperature during the 3 phases of the experiment. The same holds for heart rate and heart rate variability, although the heart rate parameters failed to return all the way back to normal during the recovery phase, perhaps indicating the need for a longer recovery period.

Interpretation

What does using naïve horses tell us about the welfare of horses that may have habituated to the level of tightness over a prolonged period? 

We acknowledge, in the paper itself, the need to understand our decision to use naïve horses and the need to assess the reactions of more seasoned horses in further studies. Knowing how long they’d be subjected to the practice would be an important variable but also knowing how much rein and curb chain tension they’d been exposed to at the same time would present a challenge.

That said, it is important to note that the horses in the current study were not having the two bits activated with rein tension. Instead they were simply standing with the bits without any reins attached. They were all riding horses and accustomed to being ridden in simple snaffle bits. As noted above, no rein tension was applied to the bits, an area yet to be investigated, but this additional factor is one that young dressage horses will experience.

There is an untested assumption that horses habituate to tight nosebands over time and, from that, there is the suggestion that gradually tightening them up over a prolonged period will negate any stress that appears in naïve horses. This theory must be tested before we can apply relentless pressure and deny horses yawning, chewing, licking and swallowing with confidence that it is ethical. Ethical equitation requires that we eliminate as many stressors as possible before justifying those that remain. Since we know that horses can perform grand prix tests without tight nosebands, any justification for denying horses jaw movement is elusive.

There is no evidence from the current study that horses habituate to the NAUN treatment but it would certainly be an interesting experiment to conduct, if funding emerged. However, if horses do habituate then using experienced dressage horses would present a sizeable challenge as some may have habituated more than others, so we would have to counter a great deal of variation. The only way to do the test would be to determine the point at which horses habituate by testing them in similar circumstances to the experiment we did here – starting with a cohort of naïve horses and testing over time to see at what point they stopped showing the significant physiological and behavioural responses. Even though, in the real world, nosebands are tightened to a restrictive extent every day, it is unlikely that one would secure ethics approval for such a study.

Habituation is a convenient term but it seems that we often assume horses habituate to procedures without actually testing this (Hall et al., 1997). When given the choice, young horses were found to avoid bit pressure; not habituate to it as was thought (Christensen, 2011). Some researchers question whether habituation is the correct term when the horse is forced to put up with an aversive event or whether the correct term for the change in the horse’s behaviour is learned helplessness.


How does the way you report physiological stress relate to human stress?


This is a great question! It is beguiling to imagine that, one day, animal welfare science will allow us relate a horse’s stress response to our own experience and affective state but that outcome is unlikely to be achieved and may not even be terribly useful. Relating our experience to those of other humans is problematic. I don’t know how being miffed differs from being suicidal for you and so must moderate my estimation of these affective states. Relating our experience to those of other non-human animals is even more problematic. We don’t know what is like to be an animal that chews for up to 16 hours a day, let alone know how miffed such an animal would be to be denied the opportunity to chew for ten minutes. In addition, in contrast to human subjects, we cannot tell the horses that we study that they will feel this pressure for only ten minutes.

The reduced swallowing noted in the current study may relate to restricted jaw movement and compression of the tongue that may immobilise the tongue. Again, relating these processes to a human affective state is questionable. If we had suggested that people try this at home by using two strips of metal across their tongue and clenching their teeth for ten minutes, we’d have been laughed out of the journal’s editorial office.

Nevertheless, other studies have revealed how much tongue compression occurs with a single bit even without tight nosebands and before rein tension (Engelke and Gasse, 2003). When one adds noseband tightening and two bits, tongue compression is likely to increase with the tongue being driven ventrally against the bony tissues of the mandible. Of course, two bits must take-up more compressible virtual space in the mouth than one bit and we should not be surprised to see less swallowing, presumably because this compression effectively immobilizes the tongue. Fluoroscopy (radiographic videoing) studies would be needed to reveal exactly why horses swallow less when subjected to this condition. Again, it is worth noting here that there was no rein tension applied to the bits in the current study, something that would greatly increase tongue compression when applied in the real world.

As always, experiment situations differ from the real world because we are attempting to isolate the effect of a particular event or stimulus; in the case of the current study noseband tightness. The NAUN treatment produced a significant stress response in isolation, meaning it was not a result of the bridle, new environment, two bits or standing still for 30mins (3 x 10m minute phases: baseline, treatment and recovery). The important point here is that the response was significant. When the horse is then taken-out into the real world, many other factors will come into play that may further stress the horse.

Knowing that the NAUN treatment significantly stresses the horse in isolation is objective and the only way to be objective. The level of any distress must be considered in context in that stress responses seem significantly greater in the NAUN treatment than when the noseband is conventionally tightened. At what point that drives a given horse into a depressed state or into learned helplessness is, of course, still unknown.

Paul McGreevy and Kate Fenner


References

Casey, V., McGreevy, P.D., O’Muiris, E., Doherty, O., 2013. A preliminary report on estimating the pressures exerted by a crank noseband in the horse. Journal of Veterinary Behavior: Clinical Applications and Research. 8, 479-484.

Christensen, J. W., Zharkikh, T. L., Antoine, A., Malmkvist, J., 2011. Rein tension acceptance in young horses in a voluntary test situation. Equine Vet. J. 43, 223-228.

Engelke, E., Gasse, H. 2003. An anatomical study of the rostral part of the equine oral cavity with respect to position and size of a snaffle bit. Equine Veterinary Education. 15 (3) 158–163.

Hall, C.A., Goodwin, D., Heleski, C., Randle, H., Waran, N., 2007. Is there evidence of ‘Learned Helplessness’ in horses? Proceedings of the 3rd International Equitation Science Symposium, Michigan. Eds: D. Goodwin, C. Heleski, P. McGreevy, A. McLean, H. Randle, C. Skelly, M. van Dierendonck and N. Waran. MSU, Michigan, USA.

McGreevy, P., Warren-Smith, A., Guisard, Y. 2012. The effect of double bridles and jaw-clamping crank nosebands on facial cutaneous and ocular temperature in horses. Journal of Veterinary Behavior: Clinical Applications and Research. 7, 142-148.

Pierard M, Hall C, König von Borstel U, Hawson LA, McLean A, Averis A, et al. Evolving protocols for research in equitation science. Jorunal of Veterinary Behaviour: Clinical Applications nad Research. 2015;10(3):255-66.

Valera M, Bartolome E, Jose Sanchez M, Molina A, Cook N, Schaefer A. Changes in Eye Temperature and Stress Assessment in Horses During Show Jumping Competitions. Journal of Equine Veterinary Science. 2012;32(12):827-30.