UPDATED The FEI automatically responds to any criticism of hyperflexion of the neck by referring to the workshop held in Lausanne in 2006. But that's the same as an admission to ignorance. All that came out of the workshop was that nobody knows what rollkur does to horses. Four years later, this is still the case.
In the rollkur debate, various references to scientific studies are often used as arguments on both sides. But if proponents and opposers of rollkur all think they have science on their side, then who is right? EPONA.tv has had a rummage through the most frequently referred to studies and the report from the 2006 FEI workshop in Lausanne on which the FEI bases its current acceptance of rollkur. Read on, and get a better perspective of what’s known and – especially – what’s not.
Anyone who has ever been involved in a discussion about rollkur, has either heard or uttered something akin to the following statement: “The FEI looked into hyperflexion, and there’s no evidence it has any long term deleterious effects.”
Firstly, it’s untrue that the FEI has ever looked into whether there are any deleterious effects of rollkur. There was a workshop in Lausanne in 2006, but it didn’t include any presentations by any scientists who have studied the long-term effects of rollkur on the health and wellbeing of horses. What it did include was a presentation by Dr. Emile Welling, a veterinarian employed by the Dutch Equestrian Federation. Dr. Welling showed radiographs of two horses who had been trained to international Grand Prix level using rollkur, and as no radiographic lesions could be seen in either case, the conclusion was that rollkur would not pose any welfare risk if practiced by experienced professionals.
Setting aside the statistical insignificance of a population of two, Tony Tyler from the FEI’s now disbanded welfare sub-committee remarks in Appendix 5 of the workshop report that: “I think it’s fair to assume that a horse winning at this level would not be in pain. There was no radiography of horses that had been schooled using “Rollkur” that were failing to perform well. With almost all types of training, some athletes’ bodies cope well and others injure. I would not be happy with the current evidence to even suggest that it does not damage the horse only that it is more likely that a well prepared supple horse may be able to cope with the demands of Rollkur.”
The FEI’s own welfare expert has a point. In statistics, there is a term called survivor bias. It means that if you subject a group to a treatment and only analyze the group members who are still standing at the end for any ill effects of the treatment, you will get a very different picture than if you were to analyze all individuals belonging to the group in the beginning. For instance, the percentage of 90-year-old smokers who have developed illness from the habit might give a distorted view of whether smoking is bad for you. To be able to say anything about the long term effects of rollkur, scientists (preferably not employed by the Dutch Equestrian Federation) would have to look at a large population of horses which had been subjected to rollkur on a regular basis. And a control group would be needed, consisting of horses of a similar quality, which had never been subjected to any form of unnatural head and neck position during training.
Such a study has never been performed – according to a report on the welfare implications of hyperflexion by Dutch scientist Dr. Kathalijne Visser, commissioned by the Dutch Ministry of Agriculture in 2009, it is due to a lack of funding and suitable horses available to scientists as well as - perhaps more notably: “ethical issues associated with applying aversive procedures in research.” In other words, scientists have trouble finding out whether rollkur is cruel, because it might be too cruel to test it out on living horses.
If there are no studies to directly confirm that horses trained in hyperflexion are at increased risk of developing injuries, there is evidence that overstretching of the horse’s nuchal ligament, as happens during hyperflexion, should be undertaken with caution. In the 1990s. German veterinary pathologist, Horst Weiler, took radiographs (X-rays) of 300 riding horses and found evidence of injury to the poll area in 70 per cent of subjects. “These findings are interpreted as a late morphological correlate of unphysiological strain because of high tension forces, high pressure, position of the head, flexion and/or rotation of the head,” he found. Weiler’s work is barely mentioned in the FEI’s workshop report, and seems to be overlooked on the basis of Welling’s analysis of two Grand Prix horses. It is not known whether Weiler’s 300 horses were all trained in hyperflexion, but, at worst, his data indicates that riding even a bit behind the vertical is detrimental to the horse.
According to Weiler's research, a lot of horses would seem to suffer from what he calls “bizarre bony outgrowths” from the area of the poll where the nuchal ligament attaches, as a physiological response to unnatural strain. No wonder the FEI rules for dressage state that: “The head should remain in a steady position, as a rule slightly in front of the vertical, with a supple poll as the highest point of the neck…”
Anyone who has ever been involved in a discussion about rollkur, has either heard or uttered something akin to the following statement: “They did a study, and it showed that recreational horses are more stressed than rollkur horses.”
It’s true they did a study. “They” in this case are once again the Dutch – more specifically Professor Eric van Breda, an eminent performance exercise physiologist who was hired a few months after the workshop to work with the Dutch dressage team. From the abstract of Professor van Breda’s presentation, which can be found in appendix 4 of the workshop report, the study compared “parameters of stress” in five recreational horses and seven international Grand Prix horses. However, the only parameter of stress, which was measured, was heart rate variability, where the elite horses were found to “show less stress” 30 minutes after their workout, once they had been groomed and put back in their stables. The abstract does not specify how the recreational horses were trained, aside from a general statement that rollkur as a training tool for such horses “is not common.”
Visser notes in her report to the Dutch Ministry of Agriculture that: “Although this study did use HRV (heart rate variability, ed.) as a physiological stress parameter, it did not include behavioural observations. Moreover, in this experiment, HRV was measured 30 – 45 minutes after the training exercise after the horses had been washed and groomed. The HRV reported might have reflected these latter handling procedures instead of (or as well as) the training technique. Another limitation of this experimental design is that the riders were not blind to the treatments. Furthermore, in the paper it is claimed that “pain” was measured, but we are not able to draw the same conclusions from the presented data.”
Professor van Breda seemingly does not share the skepticism of Visser et al regarding his study. In the post Lausanne workshop correspondence between delegates and the FEI, which is included in the final report, he lamentingly states: “The overall impression one might deduct from the report is that hyperflexion of the neck might still pose a threat to the welfare of the elite horses. This is, to my unpretentious opinion, not the outcome of the work presented during the workshop. Rather than to state it is a potential hazard in the wrong hands, it is more appropriate to the outcome of the workshop to state that the technique DOES NOT POSE ANY HEALTH OR WELFARE ISSUE AT THE ELITE LEVEL…”
Further down in his response, van Breda once again resorts to typing in capital letters, when he reminds the FEI that: “In the recommendations and conclusions from the Workshop, there is, to my opinion, again an overemphasis to the POSSIBLE negative side effects of hyperflexion of the neck. This still gives way to those opposing the technique. The peer reviewed science presented during the workshop showed the contrary e.g. there is NO EVIDENCE FOR A POSSIBLE HEALTH HAZARD…”
Professor van Breda’s letter is dated March 1, 2006 – in April of the same year, it was announced that the scientist had been added to the support team of the Dutch A and B teams for Dressage.
Anyone who has ever been involved in a discussion about rollkur, has either heard or uttered something akin to the following statement: “It has been scientifically proven to improve the mobility of the horse’s back.”
At the workshop in Lausanne in 2006, Dr. René van Weeren, professor at the University of Utrecht, presented a paper entitled: “The effect of head and neck position on the thoracolumbar kinematics in the unridden horse.” The data confirms that hyperflexing the horse’s neck will raise up the front portion of the back (the specific point of reference according to the paper is T10 – the 10th thoracic vertebra) – this can come as no surprise to proponents or opponents of rollkur, as all agree that the traction on the nuchal and supraspinous ligament caused by the hyperflexion will hoist up the back, giving the rider a temporary feeling of “throughness” which then disappears as the horse is allowed to raise its head and neck once again. But van Weeren’s data also show that the lumbar spine is hollowed when the horse is in a hyperflexed position.
The following was found to happen in the walk when the horses were walking on a treadmill without rider: "...in HNP4 (hyperflexion, ed.) the spine was more flexed at T10 and more extended at L5, and in HNP6 (traditional forward downward, ed.) the spine was more flexed at T10 but there was no increase in extension of any of the lumbar vertebrae."
In trot, the study does not mention any flexion of the thoracolumbar spine as a result of hyperflexion. On the contrary: "In HNP4 (hyperflexion, ed.) the spine was more extended T13, T17, L1 and L5."
In other words, the artificial or mechanical hoisting up of the front part of the back by way of hyperflexion comes at a cost. And the price to pay is disengagement of the hindlimb – the opposite of the official aim of dressage training: Collection. In trot, it seems the horse drops the portion of the back directly under the rider as well as the portion behind the saddle.
Conversely, by working the horses in a traditional forward downward position on the treadmill, the researchers also obtained flexion of the front part of the back, but without the horses hollowing their lumbar spine. This discovery, which validates traditional dressage theory, seems to have been played down somewhat in the Lausanne workshop report, as was the evidence that the horse will extend (hollow, ed.) the entire length of its back during hyperflexion in trot, which did not happen in traditional long and low with the nose ahead of the vertical.
The finding that hyperflexion of the neck causes extension of the lumbar spine was echoed by Professor Hilary Clayton at two USDF University sessions in 2006. According to a report from the sessions published on the website of Michigan State University's McPhail Equine Performance Center, Dr. Clayton stated that excessive flexion of the neck, tended to "hollow the area behind the saddle a bit."
Still, van Weeren’s conclusion in Lausanne was that these results “lent credibility to the fact that the position (hyperflexion, ed.) could be of value in training…” Even though, according to the published study: “..the results apply to the unridden horse and cannot be directly extrapolated to the ridden situation…”
Professor Heinz Meyer is listed as co-author of the van Weeren research, as well as a paper by M.A. Weishaupt et al. on the effects of head and neck position on the biomechanics of the ridden horse. EPONA.tv met Professor Meyer in Dublin in 2008, and asked him what he thought of the way the results of the research is used as an argument for hyperflexion.
“In my opinion, the interpretations of the research do not represent the effect of hyperflexion,” says Professor Meyer of the research, which he proposed and helped to carry out. He believes that certain data have been selected to support an argument for using hyperflexion, whilst data pointing in the other direction have been largely ignored in the rollkur debate. According to Professor Meyer, there is still disagreement amongst the authors of what the data actually means.
“In Sweden, they did a similar experiment and didn’t find any increase in the mobility of the back,” says Professor Meyer. He doesn’t think it wise to conclude that rollkur has any gymnastic effect, as long as results differ as they do.
In the paper which was presented by René van Weeren in Lausanne, the same Swedish research is briefly mentioned: “In the study of Rhodin et. al (2005), there were no significant changes in the flexion-extension ROM…” The authors speculate that this is due to the horses being different breeds and/or of a lower performance level. But Heinz Meyer is not convinced. “If you take the results of the different studies, okay, you can see that maybe you have a training effect without rider, but if you look at the Swedish research, it is still different. All these questions are unsolved,” he points out.
Indeed, the most interesting thing about the research done in Zürich by Weishaupt et al., which aimed to investigate the effect of head and neck posture on the distribution of weight between the forehand and hind end as well as the purity of the walk and trot, may be that the subjects - rollkur trained elite dressage horses – were unable to produce a regular four beat walk when unrestrained by rein tension. This would beg the question: “Why?” Rather than jump to conclusions as to the wholesomeness of hyperflexion, the authors choose to state that: “…it seems unreasonable to aim for a perfectly regular 4-beat gait under all circumstances as our study suggests that HNPs in which the regular beat was observed all involved certain restrictions to the horse’s range of movement.”
In this context, it should be noted that the treadmill, which the horses were walking on, was on a slower setting during the periods when the horses’ heads and necks were restrained, whereas the horses were walking faster during unrestrained movement. It is possible that the speed of the walk, rather than whether the head and neck were restrained, caused the horses to go lateral. But in any case, these could not have been supple, well trained horses. Not according to the FEI dressage rules, which term this rather commonplace gait abnormality “a serious deterioration of the pace”.
It should also be noted that the seven advanced level horses who were all used to being trained in the hyperflexed position, were ridden by their usual riders in regular snaffles. It was necessary to use draw reins to get the horses into the hyperflexed position for the experiment.
The understanding of dressage theory in the scientific studies which apparently support the theory behind hyperflexion differs from that of the FEI in several cases, which may undermine the usefulness of this research in relation to the correct schooling of dressage horses. The FEI’s understanding of the relationship between the head, neck, trunk and hind quarters is summed up as follows: “The position of the head and neck of a horse at the collected paces is naturally dependent on the stage of training and, to some degree, on its conformation….The arch of the neck is directly related to the degree of collection.”
So the arched neck of a Grand Prix horse is the result of something else going on in the body further back. Not the other way around. At least according to the FEI.
Conversely, the research by Weishaupt et al. refers to: “…collection achieved by the application of side reins,” as if the term collection covers some sort of head and neck set.
According to the paper presented in Lausanne by van Weeren: “It is commonly believed that the higher the HNP (head and neck position, ed.), the more load is shifted to the rear.”
Many horse people would think this a strange statement where dressage and the topic of collection are concerned. Most have learned that the horse will shift its weight back as the hind end becomes more engaged, the haunches flex and the horse “sits”. This is the result of years of gymnastic schooling – not something which any horse can do as long as its neck is in the right position.
The experimental setup where horses’ head and neck positions are manipulated with three different sets of auxiliary reins (regular side reins, side reins to the girth and an over-check) bears little resemblance to actual dressage training as set out by the FEI. So even the head and neck position, which is supposed to emulate “the competition frame” has nothing in common with a collected frame, other than the superficial parameters of head height and poll flexion.
Unsurprisingly, the study done with riders on the horses (instead of side reins, but performing the same function) showed that the riders were unable to get horses to shift any significant proportion of weight back onto the haunches, regardless of where they placed their horse’s heads. The following statement from the FEI rules may shed some light on why that might be. “Collection is developed through the use of half-halts and the use of lateral movements shoulder in, travers, renvers and half pass.” It doesn’t say anything about pulling the head up, down or in, which may be because such actions would do little to collect a horse.
According to Professor Heinz Meyer, himself an accomplished rider and author of numerous articles and books on riding theory, it is problematic when academics without any knowledge of practical riding or theory of dressage attempt to map out the biomechanical function of the horse. “I don’t think the riding aspect is sufficiently dominant in the biomechanical investigations of the horse,” he says. “A lot of the interpretations are too simple. Their way of seeing the horse is too simple. Researchers should know about riding.”
Anyone who has ever been involved in a discussion about rollkur, has either heard or uttered something akin to the following statement: “Even if the horse doesn’t suffer injuries from hyperflexion, it suffers mentally.”
Personally, Heinz Meyer doesn’t think that hyperflexion has anything to do with suppling the horse. “I have not the opinion that it is done for gymnastic effect. It’s mostly done to subordinate and discipline the horse.” During his research for the book, Roll Kur, Heinz Meyer unearthed evidence that hyperflexion has been used for thousands of years as an attempt to control horses with complicated temperaments. Even Sjef Janssen admits that the technique is used for “dominant” horses, which infers horses with a tendency to display violent conflict behaviour. As a horse in rollkur can’t see where it’s going, it is possibly easier to control. And as behavioural zoologist and FEI trainer Dr. Andrew McLean pointed out during the Lausanne workshop: “… cervical flexion as a result of sustained bit pressure has a lot more to do with compliance and pain avoidance than suppleness…”
McLean goes on to explain how using hyperflexion to control conflict behaviour can be a vicious cycle, actually leading to more problems. “It can “deaden the brakes” and that can lead to bolting as a manifestation of habituation, and other unwelcome behaviours. It can compromise welfare since horses can learn that there is nothing they can do to remove the pressure. If the horse’s head is on its chest it has nowhere else to go and may be in a state of “learned helplessness.””
From the Lausanne workshop report, it is evident that: "Delegates were unanimous in their agreement that horses must not be seen to be put under pressure by this or any other training techniques." So rollkur must be obtainable by subtle, sensitive aids, or it's no go. However, top riders are often seen struggling with their horses to maintain the hyperflexed position, and in 2008, the FEI's veterinary committee decided to condemn the technique as a form of mental abuse.
"The FEI condemns hyperflexion in any equestrian sport as an example of mental abuse. The FEI states that it does not support the practice," read the statement in the minutes from the veterinary committee, which was issued a few months prior to the Hong Kong Olympics. Later, the statement disappeared from the document posted on the FEI's website, without further explanation.
(Circumstantial) evidence
Since the workshop, very few studies have been carried out, and the conclusion of Visser’s report for the Dutch Ministry of Agriculture states that “..there is still insufficient scientific evidence to confirm unequivocally whether or not there are welfare issues involved in training techniques using hyperflexion.”
In other words, there is no scientific evidence to show that horses suffer when subjected to rollkur, because the research has not been done. There is also no scientific evidence to prove that horses do not suffer, or that rollkur does any good. On the contrary, the FEI has been presented with evidence that hyperflexion of the neck induces the opposite of “throughness” and collection in the unridden horse, and circumstantial evidence which points in the direction of mental abuse. Last but not least, there is evidence to show that any training method where the horse is behind the vertical, may carry a risk of injuries to the poll area.
At the moment, the FEI has no idea what rollkur feels like for the horse or what the long term effects might be. The federation chooses to believe that it’s okay, which is a far cry from knowing this to be the case. When criticism of the technique is rebuffed by a reference to the workshop held in 2006, it is pure spin. Nobody learned anything in Lausanne, except that more research was needed – research which has not been done, four years later.
So who should get the benefit of the doubt? The riders or the horses? The Code of Conduct of the FEI states very clearly that the welfare of the horse must be paramount at all times, which may be why the federation has chosen to ban such practices as rapping, despite the absence of empirical evidence of long term deleterious effects. The following currently sums up the FEI official position on hyperflexion:
“The FEI does not permit excessive or prolonged Hyperflexion in any equestrian sport, and has a strict stewarding program to protect the performance horse in all disciplines”.
In reality hyperflexion of the neck is allowed without limitations, beause FEI stewards have not been briefed as to the meaning of “prolonged” or “excessive”.
Sources:
Gómez Álvarez, C.B., Rhodin, M., Bobbert, M.F., Meyer, H., Weishaupt, M.A., Johnston, C. and van Weeren, P.R. (2006) The effect of head and neck position on the thoracolumbar kinematics in the unridden horse. Equine Vet J 36, 445-451.
Rhodin, M., Gómez Álvarez, C.B., Byström A., Johnston, C., van Weeren, P.R, Roepstorff, L. and Weishaupt, M. The effect of different head and neck positions on the lumbar back and hind limb kinematics in the ridden horse.
Weiler, Horst. 1997. */Zur Diagnostik der Insertionsdesmopathie des Funiculus nuchae an der Squama occipitalis/*. In: Deutsche Reiterliche Vereinigung (Hrsg.): Diagnostics in horses. Warendorf: FN-Verlag, 106-108.Part II: Pathomorphology.
