Relationship between saddle pressure measurements and clinical signs of saddle soreness at the withers




Reasons for performing the study: Similar to human decubitus ulcers, local high pressure points from ill-fitting saddles induce perfusion disturbances of different degrees resulting in tissue hypoxia and alteration in sweat production.

Objective: To relate the different clinical manifestations of saddle sores to the magnitude of saddle pressures at the location of the withers.

Methods: Sixteen horses with dry spots after exercise (Group A) and 7 cases presented with acute clinical signs of saddle pressure in the withers area (Group B) were compared with a control group of 16 sound horses with well fitting saddles (Group C). All horses underwent a saddle pressure measurement at walk, trot and canter. Mean and maximal pressures in the area of interest were compared between groups within each gait.

Results: Mean pressures differed significantly between groups in all 3 gaits. Maximal pressure differed between groups at trot; at walk and canter, however, the only significant difference was between Group C and Groups A and B, respectively, (P>0.05). Mean and maximal pressures at walk in Group A were 15.3 and 30.6 kPa, in Group B 24.0 and 38.9 kPa and in Group C 7.8 and 13.4 kPa, respectively; at trot in Group A 18.1 and 43.4 kPa, in Group B 29.7 and 53.3 kPa and in Group C 9.8 and 21.0 kPa, respectively; and at canter in Group A 21.4 and 48.9 kPa, in Group B 28.6 and 56.0 kPa and in Group C 10.9 and 24.7 kPa, respectively.

Conclusion: The study shows that there is a distinguishable difference between the 3 groups regarding the mean pressure value, in all gaits.


The discussion about the definition of a fitting saddle is still controversial. However, there is a general consensus that a saddle should neither traumatise the horse's skin nor injure the underlying muscular or neural tissues. Since it became possible to measure saddle pressure, several studies have tried to define the upper limit of tolerated pressure. Earlier investigations related saddle pressure to the occurrence of back pain (Werner et al. 2002; Nyikos et al. 2005) or to the fit of a saddle (Harman 1994). These studies gave a good overview on what was to be expected when dealing with badly fitting saddles and demonstrated how diverse saddle problems and their potentially negative influence on the horse's back can be. The most frequently encountered problems are: bridging, ill-fitting headplates and incorrect stuffing of the panels. The majority of cases presented at the Equine Hospital of the University of Zurich for poor performance related to a saddle problem showed different degrees of saddle soreness in the area of the withers underlying the headplate of the saddle (Fig 1).

Figure 1.

Area of interest: horse with dry spot at the withers after intense exercise included in Group B.

Several models of human decubitus ulcer show that the condition arises from local ischaemia resulting from capillary closure pressure being exceeded. These alterations start at the bone-muscle transition, as the muscle tissue, due to its high perfusion, shows the lowest tolerance to pressure strain (Daniel et al. 1981; Le et al. 1984). Persistent malnutrition of the tissue leads to muscle atrophy and necrosis along with local inflammation and swelling. Being the most resistant, the skin is generally the last tissue to show macroscopic damage (Nola and Vistnes 1980) and, when it does, the underlying muscles are already traumatised. Acute histological signs of pressure-related muscle injury are loss of cross striation due to damaged cell membranes and nuclei (Bouten et al. 2003; Linder-Ganz et al. 2006).

In horses, focal back muscle soreness is often accompanied by dry spots in the saddle area. Sweat glands are embedded in a dense network of capillaries. Due to high loading, local ischaemia results in dwindling sweat production (Ferguson-Pell et al. 1988). Therefore, the associated symptom of dry coat spots might be used together with the more obvious signs of a saddle sore as a semi-quantitative estimate of pressure load and can help to determine the upper limit of acceptable pressure underneath the saddle.

In contrast to the occurrence of human pressure ulcers, which mainly occur in patients confined to a wheelchair or bed, where the tissue suffers from constant loading, saddle pressure in horses is only a transient load and the horses are in motion. Therefore human pressure models cannot be directly applied to horses.

The aim of this retrospective study was to compare the saddle pressure of patients with dry spots and of patients suffering from distinct saddle sores in the area of the withers (Fig 1) with the pressure at the same location of sound horses with well fitting saddles. We hypothesised that a threshold of tolerated pressure would exist for both pathologies.

Materials and methods


All horses included in this study were presented at the Equine Hospital between 2006 and 2009 with saddle problems. As the area of interest was restricted to the region underneath the headplate of the saddle at the horses' withers, only horses with clinical signs of saddle soreness in this area (Fig 1) were included. Horses were divided into 2 groups according to the severity of clinical signs.

Group A included 16 horses (6 mares, one stallion, 9 geldings) aged 7–17 years (mean 9.1 years). All the horses showed an area at the withers that stayed dry underneath the saddle even after intense exercise (Fig 1) in addition they all showed increased muscle soreness in this area.

Group B consisted of 7 horses (4 geldings, 3 mares) aged 6–17 years (mean 10.4 years). This group showed clinical signs of saddle sores in the same area (Fig 1) as Group A showed dry spots. The area was swollen, warm and very painful to palpation. The skin was traumatised to various degrees.

These 2 groups were compared to a control group (Group C) of 16 healthy horses (4 mares, 2 stallions, 10 geldings) without any signs of saddle or back problems aged 5–18 years (mean 11.3 years). Each of the 3 groups was composed of a representative selection of breeds in Switzerland used as riding horses; ponies were excluded.

The saddles were all English saddles of different brands. All horses were privately owned leisure horses, ridden daily for 1–2 h and regularly schooled in dressage.


All horses underwent an orthopaedic examination with emphasis on the back, carried out by the same experienced veterinarian (K.v.P.). The saddle area was palpated carefully and the sore areas recorded. The saddle fit was checked manually prior to the saddle pressure measurements. Subsequently, a pressure sensitive mat (Pliance)1 was placed on the horse's back and reset to zero before placing the saddle and tightening the girth. Measurements were taken when the horse was moving in a relaxed way along the long side of the riding arena at walk, rising trot and canter at a straight line on both hands. All horses where ridden by their usual rider.

The calibration of the pressure sensitive mat was rechecked and recalibrated every 3 weeks, as recommended by the manufacturer.


The count and number of sensors (area = 9.37 cm2) overlying the area of interest at the withers (saddle sores or dry areas, respectively) were recorded before removing the saddle pressure-measuring mat. In general, 4 sensors were overlying the area of interest on each side of the horse (mean ± s.d.: Group A: 4.16 ± 0.6; Group B: 3.9 ± 0.59). In Group C, 4 sensors on each side in the same area (Fig 1) were chosen as points of comparison. Since the symptoms were always bilateral, left and right values were pooled. From each measurement 15 stride cycles were evaluated using Novel Software1 to analyse: 1) Pmean, mean pressure of the chosen sensors averaged over the entire measurement period; and 2) Pmax, maximum pressure attained during the entire measurement period. Values were calculated separately for each saddle mat half and pooled before calculating the group means for each gait.


Data were analysed separately for each gait with one-way ANOVA or the nonparametric equivalent (Kruskal-Wallis ANOVA on Ranks) depending on the preceding normality test (Kolmogorov-Smirnov). Succeeding pair-wise comparison tests (Holm-Sidak or Dunn's method) enabled identification of group differences. Statistical analyses were done with SigmaStat 3.52 and the overall significance level was set to P = 0.05.


Out of the 16 horses in Group A, only 14 could be cantered. Two horses were uncomfortable at canter and would buck, so that no correct measurement would have been achieved. In Group B one horse would neither trot nor canter and one horse resisted cantering, both due to the severe pain from saddle sores. For welfare and safety reasons the horses were not forced to do so.

Pressure values increased consistently and in each gait from Group C to Group A to Group B.

Pressure results are presented in Figures 2a and b: In all gaits Pmean (Fig 2a) differed between groups. Pmax (Fig 2b) differed between groups at the trot. At walk and canter, Pmax of Groups A and B differed only from the healthy control Group C.

Figure 2.

a) Mean pressure (kPa) of the area of interest at the withers: All groups differed significantly from each other in all the gaits. b) Maximal pressure (kPa) of the area of interest at the withers: Only at trot did the 3 groups differ from each other, whereas at walk and canter the difference was significant between the Group C (control group) and Group A (horses with dry spots at the withers) and Group B (horses with saddle pressure), respectively. Different shades indicate significant difference between the groups within the gait.


This study provides clear evidence that a commonly tolerated pressure threshold exists, above which macroscopic tissue damage occurs.

As seen in other studies, Pmean is a more reliable parameter than Pmax because of its higher repeatability (Bojer et al. 2001; de Cocq et al. 2006). Hence it was not surprising that Pmean could better distinguish between groups in all gaits. Pmax could distinguish between the 3 groups only at trot, which is the most regular gait and therefore produces very reproducible results. Peham et al. (2004) reported reduced gait regularity in horses ridden with a badly fitting saddle. This has to be assumed for the horse of the Groups A and B, especially when cantering. In order to stay in balance the rider has to compensate for this irregularity, which could partly explain the large standard deviation for the Pmax values in Groups A and B at canter. Nevertheless, Group A and B differed from Group C in all the measurements and made it possible to clearly distinguish between painful and tolerable pressure.

Unfortunately, there was a differing number of cases in Group A and B. Horses with obvious saddle sores are presented less frequently at the hospital than horses with discomfort in the back, because the pathology is quite evident and does normally not require any further investigation. On the other hand, the characteristic dry spots are very often overlooked, although they are regularly accompanied by focal pain and muscle atrophy. These dry spots appear rather unspectacular and are therefore not taken seriously by the rider mostly because of unawareness of their significance. If a patient is presented with a history of poor performance and shows these symptoms, the saddle has to be considered as a possible cause.

The breed composition of the 3 groups was a mixture between Warmbloods, Thoroughbreds and light draught horses. One would expect that Thoroughbreds would be more sensitive to pressure because of their slightly thinner skin (1.094 mm) in comparison to non-Thoroughbred horses (1.271 mm) (Sneddon et al. 2008). However, despite these differences there was no evidence of any breed-specific tendencies in the tolerated pressure threshold and therefore we presume that these pressure thresholds are applicable for the average riding horse. Ponies have not been presented to our service with saddle problems up to date. It is not clear whether this is due to their being preferentially ridden by lightweight riders (children and small adults) and therefore not exposed to too high pressure strains, or if they are less prone to saddle sores due to their different conformation.

Human patients with prolonged exposure to pressure above 4.26 kPa (Chang and Seireg 1999), just above the capillary closure pressure, are considered at risk for pressure ulcer formation. In this study, the critical pressure of 15.3 kPa at walk is >3 times higher; the healthy control group shows (Pmean walk: 7.8 kPa) that the horses tolerate much higher pressure than human patients. Although the mechanisms of developing pressure ulcers and saddle sores are assumed to be similar, a possible explanation for the horse's higher tolerance is that human pressure ulcers occur when the patient is permanently confined to a bed or to a wheelchair, whereas in horses, the pressure load of the saddle and rider is only transient and during motion (Herrman et al. 1999).

Harman (1994) correlated the level of saddle pressures with the degree of saddle fit. Well to excellently fitting saddles did not exceed values over 23.3 kPa. Unfortunately, it is not clear whether the pressure values referred to are mean or maximal pressure values. Furthermore, it is not stated if the analysis took into account the whole saddle area or only specific areas of interest. In the present study, Pmax, values of up to 24 kPa were measured in the healthy control group.

Werner et al. (2002) attempted to define an upper tolerated pressure value in relation to back pain symptoms. Mean pressure values of >15 kPa and maximal pressure values of >35 kPa measured at the sitting trot correlated with back pain. Compared to the present results, the mean pressure values were clearly lower. This could be due to the fact that in Werner et al. (2002) the whole saddle area was taken into account and not a selected area of interest (Fig 1) as in the study reported here. In general, much lower pressure values occur in the peripheral region of the saddle area, which consequently lowers the average pressure values. The greatest difficulty of such an approach is to find a group of horses with homogenous symptoms and/or defined origin of back pain (Jeffcott 1979). A later study (Nyikos et al. 2005) divided the saddle area into thirds and found an increasing sensitivity from front to hind but also only included measurements at sitting trot. Reported mean pressure exceeding 13.2 kPa and maximal pressure exceeding 31.5 kPa in the front third of the saddle area were associated with back pain. Those mean pressures were lower than in the present study by 5 kPa and peak values by 5–10 kPa. The difference between the studies was, however, that Nyikos et al. (2005) noticed no dry spots or skin damages. As the standard deviations of Pmean and Ppeak in our Group A at trot was 8.8 kPa and 3.4 kPa, respectively, the values observed by Nykios et al. (2005) would fit in the lower, but already damaging, part of the range.

Meschan et al. (2007) compared the force between 3 saddles. The saddle with the lowest overall force served as a reference. The saddle area was divided into transverse and longitudinal thirds, resulting in an area 2–10 times larger than our area of interest.

The mean pressure of the healthy horses (Group C) at walk was 7.8 ± 1.7 kPa, which can best be compared with the cranial third of the saddle area in the study by Meschan et al. (2002), which lay between 7.4 and 7.6 kPa for the saddle with the lowest overall force (i.e. the best fitting saddle).

At trot, the values in the study by Meschan et al. (2007) are around 14 kPa in the lowest overall force saddles, whereas in our study the mean pressure was 9.8 kPa in the area of interest (Fig 1). This might be because the tree size with the lowest overall force was not sufficiently well fitting, leading to higher pressure in the more dynamic gaits. In our case, all horses in the control group had individually fitted saddles, and therefore showed a lower pressure. Further differences between the studies were found regarding the zero baseline, which was before saddling and tightening the girth in the present study, whereas Meschan et al. (2007) zeroed the system after saddling and tightening the girth, which could alter the measurement behaviour of the system.

Beside the normal force, shear forces also enhance the development of pressure ulcers (Chang and Seireg 1999; Linder-Ganz and Gefen 2007). It must be assumed that the area underneath the headplate is the area with the most shear force, since there is significant muscle movement due to the pro- and retraction of the front legs (von Peinen et al. 2009). Furthermore, the stirrup bars overlie this region and directly transmit the weight of the rider to the withers area, especially in the situation where the rider stands in the stirrups. If the saddle construction fails to distribute and/or absorb this pressure, a local peak pressure point occurs. Therefore, it has to be emphasised that, due to the exposure to pressure and shear forces and because of the anatomy of the horse, the withers area underneath the headplate is probably the key area to be considered when fitting a saddle.


These values of tolerated, but harmful saddle pressure can be used as a guideline for fitting saddles in practice. Since the horses change in body condition and the saddle panels alter their shape over time due to compression, the riders have to be made aware of the importance of regularly checking their horse's back after riding for dry spots, as an indicator of an ill-fitting saddle.


The study was supported by the ‘Stiftung Forschung für das Pferd’, Zurich, Switzerland. The authors would like to thank the saddlers Clemens Santschi, Urban Truninger and Peter Menet for their constructive inputs and Dr Isabel Imboden for corrections.

Conflicts of interest

The authors have declared no potential conflicts.

Manufacturers' addresses

1 Novel GmbH, Munich, Germany.

2 SPSS Inc, Chicago, USA.