Response to comments on: Noseband type and tightness level affect pressure on the horse's face at trot

The use of nosebands remains a contentious issue within the equestrian and scientific community. An evidence-based approach is needed to inform policy and decision-making.¹ Doherty et al.² have raised several questions regarding our recent study,³ which was the first to take a biomechanical approach at quantifying dorsal nasal and ventral mandibular pressures associated with noseband tightness in horses at trot.³ We welcome critical discussion and aim to address the points raised regarding study design, data interpretation, and the application of our findings. While some of the questions posed are reasonable, Doherty et al.² assert that our study compromises equine welfare, a statement that we refute. The Fédération Equestre Internationale (FEI) did not fund either MacKechnie-Guire et al.³ or Clayton et al.'s⁴ work; however, as an indication of the FEI's approach to using evidence-based decision-making, both studies have informed the development and introduction of the noseband measuring tool, which was implemented (1 May 2025) at all international competitions and equestrian sports that are governed by the FEI to prevent horses being competed in tight nosebands.

Doherty et al.² question the inclusion of the drop noseband in our study and data analysis. Compared to the Swedish, flash and cavesson noseband, the magnitude of nasal pressure (kPa) was lower for the drop noseband. Considerable attention has been given to nosebands used in elite dressage, but the vast majority of horses are not at elite level and still require protection and good welfare. Horses training at elite level, training and/or competing in non-elite dressage and other equestrian sports, or being used for pleasure riding do wear a drop noseband,^{5, 6} and we believe that all horses participating in all equestrian sports should be safeguarded from tight nosebands. Therefore, we designed a study that had broader applications than just elite dressage; hence its inclusion. In response to Doherty et al.'s² concerns, Table 1 and Figure 1 display nasal and mandible data and distribution taken from our study,³ with the shaded area representing reprocessed data where the drop noseband data have been removed from the analysis. The median values are slightly increased; however, despite this, no significant differences were found between 2.0 and 1.5 or 1.0 finger equivalent tightness (highlighted in bold in the post hoc comparison column) therefore refuting the suggestion that the inclusion of the drop noseband data skewed the analysis and interpretation.

Doherty et al.² suggest that pressures averaged over the entire pressure mat under-represent the absolute value of the nasal and mandibular pressure. We would agree with this, however, Doherty et al.² have overlooked experimental detail described in our report³ ‘Mean, maximal and minimal pressures (kPa) and total calculated force (N) were summed over all loaded sensors and were determined on a stride-by-stride basis’. Only sensors that were loaded >5 kPa were recorded,³ therefore refuting Doherty et al.'s² suggestion. The noseband encircles the horse's maxilla and mandible, and due to the curvatures of the lateral aspect of the head, the noseband is likely not to be in immediate contact with the lateral aspect of the head,⁷ thereby not loading these pressure sensors (>5 kPa).

Doherty et al.² use the minimal values for the Swedish (also referred to as a crank noseband⁸) noseband when adjusted to 0.0-finger equivalent tightness to suggest that we have overlooked ‘worryingly’ high pressures. Whilst we agree that there is greater mandibular pressure compared to the nasal pressures throughout, the context is important. Doherty et al.² correctly identified that the increase in force (N) between the 2.0 and 1.5 finger-equivalent noseband tightness levels is greater than that observed between the remaining tightness levels. Noseband tightness in our study was standardised using the International Society for Equitation Science (ISES) taper gauge, ensuring consistent application of tightness across all studied nosebands. The observed differences in force are not solely attributable to noseband tightness but are also influenced by the design and construction of the noseband. As described in our study,³ the cavesson noseband consists of a single band of leather that encircles the horse's jaws, whereas the Swedish noseband is constructed using multiple layers of leather, contributing to increased structural rigidity. These differences affect loading of pressure sensors, particularly for the cavesson and flash nosebands. This effect can be visualised in Figure 6 of our study,⁹ where the same tightness level has slightly different sensor loading patterns depending on noseband design. Supporting this, we observed that the force differential between 2.0 and 1.5 finger equivalent tightness for the cavesson (27.8 N) and the upper band of the flash noseband (24.2 N), which are structurally similar except for the flash's additional lower strap, is notably larger than the differential between tightness settings for the Swedish noseband (11.5 N) which had greater contact area due to design (not tightness). This highlights how noseband design, independent of tightness, can influence force distribution and magnitude. Determining the magnitude and duration of pressure at which horses feel discomfort is challenging. Pressure algometry is a repeatable and semi-objective technique to determine the mechanical nociceptive threshold (MNT) at which horses exhibit an avoidance response.¹⁰ The MNT for the temporomandibular joint has been measured as 500–600 kPa.¹¹ In our study,³ when in trot, the highest maximal noseband pressures were recorded for the mandible at 32.9 kPa (0.0 finger-equivalent tightness) when fitted with a Swedish noseband. Whilst MNT values have not been reported for locations beneath the noseband, the values presented (0.0 finger-equivalent tightness) here are 15 times lower than the MNT over the temporomandibular joint.

Doherty et al.² raise concerns that our report³ overlooked the seemingly high pressures associated with the Swedish noseband and recommend that its use should be banned. However, our data³ do not support such a recommendation. Specifically, we found no statistically significant differences in nasal or mandibular pressures when the Swedish noseband was adjusted from 2.0 to 1.5 finger equivalent tightness. Although across conditions, the pressure values for the Swedish noseband were higher than those for the cavesson, these differences did not reach statistical significance. Minimal pressure values for the Swedish noseband were indeed higher than those for the cavesson noseband, but statistically, this was only observed for minimal tightness (0.0 finger equivalent tightness), which is specifically not recommended as a suitable noseband tension for horses in sport. Therefore, the recommendation to ban the Swedish noseband is not substantiated by the data presented in our report.³

Kitchen¹² advocates that for data where skewness and multimodal distribution exist (such as in our pressure data), non-parametric analyses offer a very satisfactory alternative to parametric tests, especially when combined with smaller sample sizes. Therefore, while the assumptions for non-parametric tests are generally weaker than for parametric tests, where the latter's assumptions are not met in a dataset, it is appropriate to use non-parametric analyses. Our data were not normally distributed and the variability present across parameters measured represents a skewed data profile; subsequently, reporting central measures of tendency (median and IQR) and undertaking non-parametric analysis was robust.

Doherty et al.² raise concern that we omitted relevant studies evaluating the impact of nosebands on equine comfort. The reference to the study by Pérez-Manrique et al. (2023) is interesting, but this study did not directly measure noseband tightness or provide any direct evidence of a causal relationship between noseband tightness and bony lesions of the nasal bones or mandibles.¹³ The population represented in the Pérez-Manrique study was of Mexican cavalry horses, which were used for a variety of uses, including ceremonial duties, where the type and fit of tack would be unlikely to reflect that in use for other horse populations.¹³ As such, while informative, its omission does not compromise the scientific integrity or relevance of findings from our report.³

Doherty et al.² appear to place disproportionate emphasis on the use of nosebands exclusively in high-level dressage horses, suggesting that the relevance of our study is limited due to the study's focus on noseband pressures in horses ridden in snaffle bridles. This perspective overlooks the broader scope of noseband use across equestrian disciplines and levels, where snaffle bridles are widely used. It also implies that the welfare of horses that are not elite dressage horses is of less importance, which is not our belief. Understanding the biomechanical implications of noseband tightness in this context remains both relevant and necessary. Crucial insights into pressure distribution and its potential welfare implications should not be dismissed simply because the study population did not represent elite dressage horses ridden in a double bridle.¹⁴

Doherty et al.² reference two studies involving double bridles and appear to consider these citations as offering superior insight and application relative to the findings in our report.³ However, this overlooks key methodological limitations within the cited literature. Specifically, in McGreevy et al.,⁸ not all horses included in the study had prior experience wearing a double bridle, and in Fenner et al.,⁹ none of the horses had been fitted with a double bridle. While both studies contribute to the broader discussion on equine equipment and welfare, the same critique can be applied to the alternative or additional studies cited, particularly when those studies involve horses naïve to such equipment, which is a facet often used to criticise the use of the double bridle.¹⁵

Doherty et al.² suggest that our study³ would have ‘greater value if it included data collection from horses wearing nosebands with greater than 2 cm space allowed under the noseband’. McGreevy et al.⁸ measured the midpoint of the intermediate phalanx of the digitus medius (middle finger) of 10 male and female subjects (>18 years). The mean dimensions were used to design a noseband taper gauge (ISES), which represented the dimensions of two human fingers in a side-by-side orientation. The mean width of the midpoint of the intermediate phalanx of the digitus secundus and digitus medius side-by-side was 3.87 ± 0.09 cm. There were no differences between the subjects' dominant and non-dominant hands, however, male subjects had larger finger dimensions than females.⁸ In our report,³ the ISES noseband tool was used, which indicated two finger (40 mm wide × 16 mm depth) and one (18 mm wide × 11 mm depth) and equivalent tightness levels. Horizontal reference lines were added at 50% of the distance between the 2.0 and 1.0 visual indicators and the end of the tool (0.0 finger tightness) to represent 0.5 (16 mm wide × 10 mm depth) and 1.5 (30 mm wide × 15 mm depth) finger-equivalent tightness.

Doherty et al.² suggest that we should have studied noseband pressures with >2 cm distance beneath the noseband. We are intrigued by this, as based on the dimensions of the ISES tool, this was achieved for the 2.0 and 1.5 finger equivalent tightness settings. If Doherty et al.² are referring to at least 2 cm in distance (height), then this would exclude the ISES tool, as at 2.0 finger equivalent tightness the ISES tool has a depth of 1.6 cm. It is important to consider the experimental approach and findings of Uldahl and Clayton.¹⁶ In that study, a multi-tool was designed where the measurement intervals were converted to linear measurements of <2, 2–3 and >3 cm, which are equivalent to the insertion of less than two, two, or more than two female fingers “placed on top of each other”. McGreevy et al.⁸ designed their tool (ISES measuring tool) with the fingers in a side-by-side orientation, but did report the mean height of the midpoint of the intermediate phalanx of the digitus medius which was 1.59 ± 0.05 cm. The two measuring tools are not comparable, as the ISES tool provides greater width beneath the noseband, compared with the tool used by Uldahl and Clayton,¹⁶ which was far narrower, but provided more height, creating a tenting effect beneath the noseband. We chose to use the ISES tool as it is commercially available, has been used in studies, and is a tool that welfare groups actively promote.

There is concern that we have overlooked a pivotal finding from Uldahl and Clayton.¹⁶ In that study, 9.2% of the 3143 horses examined exhibited oral lesions and/or blood visible at the commissures of the lips. Among horses with nosebands adjusted to less than 2 cm (equivalent to fewer than two fingers (height) in a non-stacked position), the prevalence of oral lesions was 11% (165 of 1529 horses) with a noseband adjusted to <2 cm. Although noseband tightness is a plausible contributing factor to the development of oral lesions, it is important to note that Uldahl and Clayton¹⁶ did not further stratify the <2 cm group (e.g., into 1.5 cm), leaving the possibility that the observed 11% of horses with lesions may be associated with the tightest end of the adjustment range (i.e., 0 cm or less). Interestingly, complete removal of the noseband did not prevent oral lesions, with an absence of an upper noseband being associated with increased risk of oral lesions. The development of oral lesions is complex, and notwithstanding the possible effect that the noseband has, other factors were not controlled for, for example, dental health, type of training programme, or skill level of rider.¹⁶

The request to know the background of the riders studied to make the studies repeatable is an unnecessary distraction, as rider details which complied with ethical approval were included. We recruited high-level dressage horses and riders to reduce the rider variability, as it is known that less experienced riders can ride out of phase with the horse,¹⁷ which may cause inter-stride variation, compromising rider–bridle–horse interaction. We agree that rein tension data would have complemented the study and aided interpretation; unfortunately, due to data capture issues, rein tension data were lost. We have subsequently published studies on bridle/horse interaction where we have included rein tension data to assist in further understanding of horse/rider interaction.¹⁴

We used the ISES taper gauge as our noseband tightness tool and details about how the half measurements were determined on this noseband measuring tool were included in our report.³ We acknowledge that the ISES tool had not been validated under these conditions, however measuring 50% between the 2.0 and 1.0 finger visual indicator should provide the distance of 1.5-finger tightness. Given the relevance of the ISES gauge and its measurement to the area under discussion, and its commercial and professional implications, it is important to point out McGreevy and colleagues informed the design of the ISES noseband taper gauge, which is sold and distributed by ISES (https://www.equitationscience.com/store). We would have preferred randomisation of the order of study of the various nosebands from a scientific perspective. However, as described in our report,³ the order of testing prescribed by the UK Home Office as part of their ethical approval system. Doherty et al.² suggest that ‘it is unlikely that horses can open their mouths and thus manifest pain or discomfort at greater tightness levels’. However, this assertion underscores the importance of clearly defining what constitutes a ‘tight’ noseband. Horses standing square could accept and chew a large treat (4.5 × 1.7 × 1.7 cm) at the same frequency across all noseband conditions, including when the cavesson noseband was adjusted from a 2.0 to a 0.0 finger-equivalent tightness⁴ indicating that horses could open their mouths.

Doherty et al.² suggest that the noseband is acting as a tourniquet when minimum pressures exceed 15 kPa. A band around the skull is not equivalent to a tourniquet applied to a limb to occlude blood flow. Tourniquets are used for emergency or intra-operative occlusion of extremity blood flow, affecting large vessels, with the pressure required to occlude blood flow increasing exponentially with the circumference of the limb. An equine head is not equivalent because there is a small amount of soft tissue overlaying the skull, and the internal contents of the skull are protected from the external pressure.

It is more relevant to discuss the impact of applying focal pressures because studies report specific locations of higher pressures and not circumferential pressures, and these higher pressure locations are not at anatomical locations with large vessels or nerves, unlike a circumferential tourniquet. Our study³ clearly demonstrates that when the noseband is adjusted from 2.0 to 1.5 finger equivalents, the minimum pressures were < 10 kPa for both adjustments. It is plausible that if a noseband were tightened to the extent of visibly compressing superficial soft tissues, it could potentially compress superficial, local small vasculature. However, we did not study such conditions.

We reported that mandibular pressures were higher than dorsal nasal pressures, and suggested that additional padding could be used as a means of redistributing pressure across the mandibles, as has been used for the nasal bones.¹⁸ Regardless of whether padding is present, any correctly fitted noseband must still allow a measuring device to pass beneath it. Under these conditions, a noseband fitted with padding must be of sufficient laxity to allow a measuring tool to pass between the padding and the nasal bones. By necessity, a padded noseband will have to be adjusted to a looser setting to accommodate the combined volume of both the padding and the tool. The assertion that padding is used to facilitate over-tightening is misleading. Indeed, discouraging the use of padding is likely to be deleterious to equine welfare and its removal may inadvertently compromise equine welfare by increasing localised pressure on sensitive anatomical structures.

The Swedish noseband has been previously described as a ‘jaw-clamping crank noseband’.⁸ However, in the study where this term was specifically used, limitations in that study's design made it impossible to determine whether the observed outcomes were solely attributable to the noseband itself or influenced by the concurrent use of a double bridle.⁸ It is important to acknowledge that the Swedish noseband includes design features that may help mitigate pressure. For example, it typically includes additional padding over the nasal bones and mandibles compared to a standard cavesson and incorporates an articulating mechanism that allows the noseband to move in conjunction with the cheekpiece, potentially distributing pressure more evenly, and ensuring the dorsal part of the noseband lies parallel to the horse's nose.¹⁸

While we do not dispute that it is possible to overtighten the Swedish noseband,¹⁹ as it is with other nosebands, our data shows that when Swedish, flash and cavesson nosebands were adjusted to the same tightness levels, pressure differences between the designs were minimal.² In our study, nosebands were specifically fitted to each horse using a qualified bridle fitter, which was considered essential for correct study design and to safeguard the welfare of the horses taking part in the study. Correct bridle fitting and optimising noseband shape and design to each horse is important for optimising welfare, but is not mentioned in most previous studies. It is possible that differences in results between studies might relate to differences in fitting of the noseband and bridle to each horse. The use of the FEI and ISES noseband measuring tools limits over-tightening of the noseband, something that should be consistent across all noseband types.

Further work is underway within our group to study behavioural responses to various noseband types and tightness. However, we hope that this detailed response addresses the questions regarding our publication³ raised by Doherty et al.,² many of which have provided an opportunity to clarify aspects of our study in more detail. We also hope this will contribute meaningfully to the broader discourse on noseband use. We believe our current publications^{3, 4, 7, 14} will serve as a foundation for further research into the biomechanics and welfare implications of noseband use. Again, we must address, in the strongest terms, the suggestion by Doherty et al.² that our study compromised equine welfare, a claim we categorically refute. Notably, the FEI has implemented a significant welfare-driven measure of restricting noseband tightness to no less than 1.5 finger-equivalents. Additionally, the measurement site has been standardised to the nasal bone, rather than the lateral aspect of the head,⁷ and a clear pass/fail protocol has been established (1 May 2025, https://inside.fei.org/system/files/FEI%20Measuring%20Device-General%20Protocol%20with%20Discipline%20Protocols-Clean-17Feb2025-With%20Discipline%20logos.pdf. Accessed June 2025). While critical dialogue is essential in advancing scientific understanding, it is equally important that such critique does not hinder further progress.

R. MacKechnie-Guire: Conceptualization; writing – original draft; writing – review and editing. R. Murray: Writing – review and editing; writing – original draft. J. M. Williams: Writing – original draft; writing – review and editing; methodology. J. Nixon: Writing – review and editing. M. Fisher: Writing – review and editing. D. Fisher: Writing – review and editing. V. Walker: Writing – review and editing. M. Pierard: Writing – review and editing; methodology. H. M. Clayton: Writing – original draft; writing – review and editing.