Reasons for performing study: There has been growing interest in training techniques with respect to the head and neck position (HNP) of the equine athlete. Little is known about the influence of HNP on neuromuscular transmission in neck muscles.
Objective: To test the hypothesis that different HNPs have effect on single fibre (SF), quantitative electromyographic (QEMG) examination and muscle enzyme activity directly after moderate exercise.
Methods: Seven Warmblood horses were studied using a standard exercise protocol in 5 HNPs: HNP1: unrestrained; HNP2: neck raised; bridge of nose around the vertical; HNP4: neck lowered and considerably flexed, bridge of nose pointing towards the chest; HNP5: neck raised and considerably extended; bridge of nose in front of the vertical; HNP7: neck lowered and flexed; bridge of nose pointing towards the carpus. Mean consecutive difference (MCD) of single muscle fibre potentials and motor unit action potential (MUP) variables (amplitude, duration, area, turns and phases) were recorded in each fixed position directly after exercise at rest using commercial EMG equipment. Muscle enzyme activity was measured before and 4, 6 and 24 h after exercise.
Results: Mean consecutive difference in all HNPs was higher than in HNP1 (22 µs, P<0.001) of which HNP4 was highest with 39 µs compared to 30 µs in HNP2 (P = 0.04); MCD in HNP 5,7 was with 25 µs lower than in HNP 2 and 4 (P<0.001). Odds ratio for MCD suggestive for conduction delay or block was 13.6 in HNP4 compared to HNP1 (P<0.001). Motion unit action potential variables followed the same pattern as MCD. Lactate dehydrogenase (LDH) activity increased in HNP4 at 4 h (P = 0.014), 6 h (P = 0.017) and 24 h (P = 0.038) post exercise and in HNP5 and HNP7 at 4 h (P = 0.037; 0.029).
Conclusions and clinical relevance: HNP4 in particular leads to a higher rise in LDH activity, MCD and MUP variables, indicating that HNPs have effect on variables characterising neuromuscular functionality.
Various methods are used in the training of dressage and jumping horses to achieve a well balanced horse able to show its qualities in the jumping or dressage arena. Among riders, it is generally believed that altering the head and neck position (HNP) is an important aid in achieving this goal (Weishaupt et al. 2006). Concern has been raised in the public opinion regarding the effect of a deeply flexed HNP on the welfare of horses. This position has been named ‘rollkur’ (Meyer 1992), ‘low, deep and round’ (Janssen 2003) or ‘hyperflexion’ (Jeffcott et al. 2006). A change in HNP does influence back kinematics and loading pattern of the locomotor apparatus in the unridden as well as in the ridden high-level dressage horse (Gómez Álvarez et al. 2006, Weishaupt et al. 2006; Rhodin et al. 2009; Waldern et al. 2009). These studies discuss the effects of this training method without resulting in a clear-cut conclusion. Although the FEI recognised that training with an extreme flexed HNP may be a risk for the welfare of the horse (Jeffcott et al. 2006), it remains uncertain on what topic of welfare or to what extent this training technique could possibly have an effect on the well-being of the equine performance individual. This uncertainty validates more research on this topic.
Single fibre EMG allows precise study of the microphysiology of the motor unit. The physiological parameters that can be quantified include impulse transmission along the intramuscular axon collaterals, pre- and post synaptic events at the neuromuscular junction, and muscle fibre membrane properties (Trontelj and Stålberg 1995). In the normal muscle, a large excess of acetylcholine and an overload of acetylcholine receptors needed for the generation of an endplate potential are present in the endplate zone in order to serve as a built-in safety feature. Single fibre EMG allows evaluation of a small portion of the muscle and of a portion of the nerves across a suspected area, making it suitable to detect subtle changes that conventional conducting studies cannot detect (Padua et al. 2001). The latency from the stimulus to the response varies and is called ‘jitter’. It becomes increased whenever the ratio between the action potential threshold and end plate potential becomes increased. In diseases of abnormal neuromuscular transmission this jitter becomes increased and is expressed as the mean value of consecutive differences of the successive interpotential intervals calculated from the formula (IPI1 − IPI2) + (IPI2 − PIP3) + . . . + IPIn-1 − IPIn) divided by (n − 1). IPI1 is the stimulus response interval in case of stimulated single fibre EMG (Sanders and Stålberg 1996) (Fig 1).
In the present study, a combination of modern neurophysiological diagnostic techniques was used to test the hypothesis that alterations of HNPs influencing intervertebral foramina dimensions in the equine cervical spine (Sleutjens et al. 2010), in healthy, base-level trained Warmblood riding horses directly following moderate lungeing exercise, will alter neuromuscular functionality.
Materials and methods
Seven healthy, base-level trained Royal Dutch Sport horses (5 mares and 2 geldings; mean age 10.3 ± 3.6 years; mean height at withers 161.2 ± 1.4 cm; mean weight 531 ± 47.3 kg) with no disease history participated in the study. No abnormalities were found on full clinical examination including neurological examination. Radiographic and ultrasound examination of their cervical spinal column showed no abnormalities. To accustom the horses to the experimental set-up, they were trained in the different HNPs on the lunge for at least 3 weeks according to the experimental design below. This study was approved by the Committee on Animal Welfare of Utrecht University.
Each horse performed a standardised exercise test on the lunge in the anticlockwise direction; warming-up (1 min walk, 3 min trot, 1 min canter, unrestrained) followed by trot1 (10 min, heart rate [HR] 101 ± 8 beats/min), canter (4 min, HR 128 ± 11 beats/min), trot2 (5 min, HR 104 ± 8 beats/min), walk (5 min, HR 73 ± 8 beats/min) and cooling-down (5 min walk, unrestrained). After the warming-up, the predetermined HNP was accomplished using side reins. The 5 standard, predetermined HNPs, of which HNP1, HNP2, HNP4 and HNP5 were positioned comparable to those used in an earlier experiment (Gómez Álvarez et al. 2006; Weishaupt et al. 2006; Rhodin et al. 2009; Waldern et al. 2009) (Fig 2). In addition, they were discussed with an international dressage team and a researcher of the formerly mentioned research group to check for realistic and correct interpretation. HNP7 was included because 2 interpretations of the rollkur training position were found to exist among riders. The correct HNPs were determined based on the study of Elgersma et al. (2010) (Table 1). SF- and QEMG were performed in the same session with the horse standing still in the fixed HNP directly after the exercise test before the cooling down. Pre- and post exercise HNP1 were measured in all horses at the start of the measuring period, after acclimatisation of the horses to the training, the other HNPs were measured post exercise in a random order. There were at least 2 days between each measurement with the exception of HNP1.
Table 1. Mean ± s.d. angles and distances defining the head and neck positions (HNPs) in the group of horses (n = 7)
Angle 1: T6-wing of atlas with the horizontal. Angle 2: T6-wing of atlas-lower part of the facial crest. Angle 3: Wing of atlas-lower part of the facial crest with the vertical. Angle 4: Bridge of nose-vertical. Distance A: The horizontal distance between the vertical lines passing through the lower part of the facial crest and the supraglenoid tubercle of the scapula. Distance B: The vertical line between the horizontal lines passing through the lower part of the facial crest and the lateral styloid process of the radius (for illustration see Fig 2).
−16.2 ± 5.4
131.9 ± 5.4
25.7 ± 4.5
11.9 ± 3.5
73.5 ± 9.1
60.2 ± 12.8
3.60 ± 5.4
87.2 ± 5.0
0.9 ± 6.6
−12.8 ± 6.7
43.9 ± 4.4
89.8 ± 7.9
−33.3 ± 3.2
84.4 ± 3.4
−38.9 ± 2.8
−51.7 ± 4.9
34.5 ± 3.3
46.7 ± 5.6
34.8 ± 5.2
119.2 ± 9.9
63.9 ± 9.4
47.2 ± 9.2
45.5 ± 5.4
144.8 ± 10.4
−41.5 ± 4.6
103.6 ± 6.9
−27.8 ± 4.0
−41.0 ± 5.9
37.2 ± 4.4
18.6 ± 8.7
Rectal temperature was measured by a commercial digital thermometer (Microlife, MT 1831)1, before exercise, at the end of exercise and at the end of the EMG measurements.
Muscle enzyme activity
Muscle enzyme activities creatine kinase (CK), lactate dehydrogenase (LDH) and aspartate aminotransferase (ASAT) were measured using a Synchron CX52 from 4 ml heparinised blood taken form the jugular vein before exercise and 4, 6 and 24 h post exercise. Heparin plasma was separated directly after sampling and stored at -20°C overnight until analysed.
Selection of the muscle
For both QEMG and stimulated single fibre EMG, the muscle selection and location to be measured were based on 3 other studies (Wijnberg et al. 2009a,b; Sleutjens et al. 2010). A study preparing for this study determined that the largest significant effects of flexion and extension on intervertebral foramina height and length occurred in the segment C6–C7 (Sleutjens et al. 2010). In combination with the knowledge that cervical pathology especially in this segment is a common cause of neurological problems in equine athletes (Moore et al. 1992; Levine et al. 2008), this segment was chosen as study object. The segments in between 2 vertebrae were identified in vivo by manual palpation of the transverse processes based on studies of Wijnberg et al. (2009a,b).
QEMG of the serratus ventralis muscle
Details on definition, materials and methods of EMG examination can be found in previous publications (Wijnberg et al. 2002b,c,d). In brief, EMG signals were recorded using a portable apparatus (Viking Quest EMG system; software version 11.0)3 connected to a portable computer to record and store signals and disposable concentric needle electrodes (26 gauge concentric EMG needle, length 50 mm; diameter 0.45 mm; sampling area 0.07 mm)4. A surgical pad attached to the horse with a girdle placed directly behind the withers and connected to the preamplifier served as the ground electrode. The area was swabbed with alcohol and the pad covered with electrode gel for optimal contact. Frequency bandwidth was 5–10 kHz. Sweep speed was 10–20 ms/division. The sensitivity of the oscilloscope was set at 50–100 µV for spontaneous activity and at 100–500 µV/division for motor unit action potential (motor unit action potentials) recording depending on the size of obtained motor unit action potentials.
EMG recordings of motor unit action potentials
Insertional activity, pathological spontaneous activity (PSA) and motor unit action potentials were recorded per segment of the cervical part of the left serratus ventralis muscle 5–10 cm rostral of the cranial border of the subclavian muscle (Fig 3). At least 3 insertions and 3 directions per insertion were made per session in this area. The needle was redirected several times and, by selecting sharp sounding motor unit action potentials while the needle was withdrawn with 3 mm increments, sampling was performed throughout the indicated muscle area maximally 3–4 cm deep (Wijnberg et al. 2002b,c,d, 2009a,b). Motor unit action potentials were selected partly in an automatic way, using a trigger line that selects identical motor unit action potentials above chosen amplitude. The 20–30 motor unit action potentials per horse per HNP included for analysis had a maximal rise time of 0.8 ms and were identically firing at least 4 times. The EMG equipment automatically provided individual motor unit action potentials amplitude, duration and number of phases including calculation of mean. The automatic motor unit action potentials selection was manually filtered from nonidentical motor unit action potentials within a recording and the end marker of duration was manually corrected if necessary (Sonoo and Stålberg 1993; Kimura 2001a). Number of turns and the ratio of motor unit action potentials area : amplitude were calculated manually.
Stimulated single fibre EMG
The same portable EMG apparatus as mentioned above was used. A single fibre EMG needle electrode (40 mm long, 0.45 mm diameter)4 was used to record the single fibre muscle action potentials. Filter settings were set at 10 kHz for the high frequency filter and at 500 Hz for the low frequency filter. Sensitivity was 0.2–1 mV/cm, sweep speed was 1 ms/cm. Two monopolar Teflon coated needle electrodes (length 37 mm, diameter 0.46 mm, sampling area 0.34 mm2)4 were placed i.m. 3 cm apart in the cervical part of the left serratus ventralis muscle, 5–10 cm rostral of the cranial border of the subclavian muscle (Fig 3) around 3 cm deep and served as stimulating electrodes. Stimulation frequency was set at 1 Hz for locating the action potential, for capture this was 3–5 Hz (stimulus duration was 0.05 ms) (Kimura 2001b). Motor point location was identified by subtle movements of the cathode of the monopolar needle in combination with adjusting the stimulating current until slight muscle twitching was induced. The SF needle electrode was inserted into the twitching portion of the muscle belly, 1–2 cm distant to the stimulation electrodes. Stimulating current was increased to supra-maximal stimulation until maximum amplitude was generated in combination with no latency reduction from the same muscle fibre. For each HNP at least 30 muscle fibres were studied, in which 50–80 consecutive single muscle fibre action potentials were recorded per muscle fibre if the amplitude was at least 300 µV and rise-time shorter than 300 µs (Kimura 2001b). Latencies of the evoked SF action potentials were measured from the stimulus artefact to the quick rising phase on the positive-negative aspects of the waveform. Mean consecutive difference (MCD) was calculated automatically by the incorporated software in the single fibre EMG program of the EMG machine for each muscle fibre. Mean MCD (Fig 1) was calculated by the EMG software based on 50–80 consecutive single muscle fibre action potentials per muscle fibre studied (n≥30 per horse per HNP).
Post exercise data of HNP1 were used as reference. Differences between HNPs for muscle enzyme activity, motor unit action potentials variables and MCD were tested using a linear mixed model, with head and neck position as fixed effect, random intercept and horse as experimental unit using SPSS version 15.05. Differences in muscle enzyme activity, motor unit action potentials variables and MCD values were transformed to natural logarithms (ln) in order to obtain normal distribution and enable statistical analysis. Geometric mean (gmean) was calculated from back transformation of mean ln values. Odds ratio for risk on pathological range MCD was calculated and tested in a general linear model with software analysis packet R version 2.8.16. Paired t tests were performed to calculate rectal temperature differences before and after exercise. Experiment wise, significant levels were set to P<0.05 after post hoc Bonferroni correction.
Before exercise the mean rectal temperature was mean ± s.d. 37.4 ± 0.16°C. After exercise this increased to 37.8 ± 0.38°C (P<0.001). At the end of the study temperature was not significantly altered (37.8 ± 0.27°C). Rectal temperature was 38.0 ± 0.2°C in HNP1, 37.7 ± 0.2°C in HNP2, 37.8 ± 0.3°C in HNP4 and 37.8 ± 0.2°C in HNP5 and HNP7. Rectal temperature in HNP2 was significant lower (P = 0.033) compared to HNP1.
Muscle enzyme activity
Of the 3 muscle enzyme activities measured, only the LDH showed significant elevations. The largest increase was present at 4 h after exercise (Fig 4) with a significant increase in HNP4 (P = 0.014), HNP5 (P = 0.037) and HNP7 (P = 0.029) compared to the control HNP1. The increase from mean ± s.d. 301 ± 56 u/l to 337 ± 57 u/l was highest in HNP4, and with 302 ± 37u/l to 324 ± 58 u/l SD lowest in HNP5. HNP4 showed a significant increase at all measured time points (from 301 ± 56 u/l to 337 ± 57 u/l at 4 h, P = 0.014; to 332 ± 54 u/l at 6 h, P = 0.017; and to 304 ± 55 u/l at 24 h, P = 0.038). CK activity did increased by >2-fold after exercise in HNP4, HNP5 and HNP7. However, in individual horses these increases were highest in HNP4 (5–17 times value pre-exercise) and lowest in HNP7 (6–9 times pre-exercise value) (Table 2).
Table 2. Illustration of highest creatine kinase (CK) activity (u/l) values pre-exercise (0), 4, 6 and 24 h after exercise in each HNP
It can be seen that the increase in CK activity was highest in HNP4 (17 × pre-exercise value) and lowest in HNP1 (1.4 × pre-exercise value). Lab reference CK activity <200 u/l.
Single fibre EMG
Mean motor unit action potentials variables are presented in Table 3. In HNP1, no significant difference was present before and after exercise. Pathological spontaneous activity as defined in a former study (Wijnberg et al. 2002d) was recorded in HNP4 in 2 horses. The ratio mean area : mean amplitude was >1 in HNP4 and HNP7 (1.11 and 1.10, respectively), whereas in this was 1 in HNP1 and HNP2, and <1 for HNP5. HNP 2 showed no significant differences compared to HNP1. HNP4 showed a higher number of turns (P = 0.02). HNP5 resulted in a higher motor unit action potentials amplitude (P < 0.001); and HNP7 in a longer motor unit action potentials duration (P = 0.008), higher number of turns (P = 0.02) and motor unit action potentials area (P = 0.04).
Table 3. Description of mean and back transformed gmean motor unit action potential variables in 7 horses in 5 head and neck positions
LB, lower boundary; UB, upper boundary; MUP, motor unit action potential; HNP, head and neck position; gmean, geometric mean derived from back transformation of Ln values. ** P = 0001* P = 0.05 after post hoc Bonferonni correction compared with HNP1 (control position) based on gmean statistics.
Mean consecutive difference is presented in Figure 5 and Table 4; it was significantly higher pre-exercise (26 µs) compared to post exercise (22 µs) in HNP1 (P<0.001). Of all HNPs, MCD was highest in HNP4 (P<0.001). HNP2, HNP5 and HNP7 also induced a higher MCD (25 μs) than HNP1 (P<0.001). If human references are applied (Kimura 2001b) the odds ratio for a MCD in the pathological range (≥55 µs) or a conduction block (≥80 µs) was for HNP4/HNP1 13.6 (P<0.001).
Table 4. Description of gmean consecutive difference (µs) in 7 horses in 5 head and neck positions
Gmean, geometric mean derived from back transformation of Ln values; LB, lower boundary; UB, upper boundary; MCD, mean consecutive difference, a method to express jitter between consecutive muscle fibre motor unit action potentials; HNP, head and neck position. ** P = 0.001* P = 0.05 after post hoc Bonferonni correction compared with HNP1 (control position) based on gmean statistics.
In the current study, various effects of the different HNPs were determined. The public concern expressed in discussions in global media on training in bended neck positions that has been ongoing for some years necessitates performing scientific studies in this topic to address the concerns on horse training methods objectively. Most effects were present in HNP4, but none of the HNPs tested in comparison with the neutral HNP1 were without effect. The increases in LDH activity, motor unit action potentials amplitude, duration, area and number of turns and MCD induced by HNPs other than the control HNP1, indicate that changes in HNP alter neuromuscular functionality, especially synchronicity expressed by the increased jitter.
Lungeing a horse in a fixed bended neck position for a relatively short period appears to be a common training method for warm blood sport Horses amongst horse owners training at both low and high level. The same experimental set-up with horses ridden by their own riders would induce a human factor that would result in a (daily) variability that was considered larger than that in a lungeing situation with the same person. Therefore, the hypothesis tested was that lungeing at moderate exercise in different HNPs would influence variables related to neuromuscular functionality.
The significant increase in LDH activity indicated that leakage from muscle fibres was present in HNP4, 5 and 7. In addition, CK activity increased ≥2–17 times in individual horses in HNP4, HNP5 and HNP7 whilst increases ≥50–250% are considered pathological (Harris 1998; Ludvikova et al. 2008). This is in agreement with former studies concluding that of CK, ASAT and LDH activities, LDH activity appeared to be the most sensitive marker of increased workload (Krzywanek et al. 1996; Wijnberg et al. 2008) or myopathy (Wijnberg et al. 2003a). In HNP4 this effect was the largest and longest measurable, suggesting a larger workload in this position. Whether this increase suggests higher workload (Krzywanek et al. 1996) on neck muscles or on muscles involved in the altered gait (Gómez Álvarez et al. 2006; Rhodin et al. 2009) cannot be concluded from this study.
The increased motor unit action potentials amplitude in HNP5 suggests recruitment of larger motor units indicative of increased force support, the idea of increased workload (Weishaupt et al. 2006; Waldern et al. 2009). However, an increase in ratio area : amplitude is seen in human neuropathies whereas in myopathy a decrease is expected (Nandedkar et al. 1988). HNP4 and HNP7 resulted in a ratio >1; in the other HNPs this was ≤1. The increases in number of turns in HNP4 and HNP7 and duration in HNP7, suggest a less synchronous arrival of the individual motor unit action potentials of which the recorded motor unit action potential is composed (Gabriel et al. 2006; Wijnberg et al. 2008), which is in agreement with the results of the single fibre EMG. An increase in MCD measured by stimulated single fibre EMG in man and dogs is considered a delay in neuromuscular transmission (Sanders 2002; Añor et al. 2003). The latency from the stimulus to the response varies when repetitive nerve stimulation induces muscle action potentials. This variation is called the jitter and is measured with single fibre EMG and expressed as MCD. Stimulated single fibre EMG has the advantage enabling testing in uncooperative patients (Kimura 2001b). Increased jitter is regarded as a subclinical sign of impaired neuromuscular transmission in man (Stålberg and Trontelj 1997), whilst, in case of further impairment, intermittent or persistent blocking will occur.
When blocking occurs in many endplates in a muscle, this results in clinical weakness in man (Sanders and Stålberg 1996). Many diseases, both myogenic and neurogenic can cause an abnormal neuromuscular transmission and thus induce increased jitter (Sanders and Stålberg 1996; Padua et al. 2007). Montagna et al. (2001) have studied nerve conduction velocities and neuromuscular transmission using repetitive nerve stimulation; however, single fibre EMG studies of neuromuscular transmission by axonal microstimulation have not been reported in equine literature to the authors' knowledge and reference values are not available. For that reason in all 7 healthy horses the neutral HNP1 served as their own control. A human study is considered abnormal if the mean or median jitter exceeds the upper limit for the muscle, or >10% of pairs or endplates have increased jitter (Sanders and Stålberg 1996). If human references are applied, in which MCD of >55 µs suggest pathological conduction and >80 µs conduction block, in HNP4 the odds ratio was 13.6 times higher to fall into this category, even though the temperature had increased after exercise (e.g. higher temperatures are negatively correlated to nerve conduction speed) indicating that the neuromuscular transmission seems to be delayed. The increase in number of turns of the motor unit action potentials in HNP4 and HNP7, and increase in motor unit action potentials duration despite the increase in temperature is in agreement with this finding. Although the rectal temperature in HNP2 decreased slightly but significant, it stayed in between the reference values and therefore has no physiological relevance. Upper normal limits of MCD of voluntary muscle activation depend on the muscle examined (29–49 µs) and increase with age (Sanders and Stålberg 1996). Newer human studies indicate an even lower normal value of 36 µs and single fibre EMG proved to be useful in diagnosing early, mild or partial myelin damage (Padua et al. 2007). No further data exist on MCD in horses, which leaves the suggestion open as to what extent the increased jitter found in this study impairs the horse functionality. At which level the impairment at the endplate is originating, nerve, muscle or synapse, is not clarified nor examined in this study. Alternative explanations such as that repeated needle insertions would have damaged or altered muscle fibre and thereby altered the enzyme activity or EMG variables, or that fatigue or increased muscle force explain the results shown above are unlikely. A study of Steiss and Forsyth (1984) failed to show significant increase in CK activity after EMG in horses. Increased muscle tone would lead to increased amplitude, faster firing rate and eventually recruitment of larger motor units. In the case of muscle fatigue, motor unit action potentials amplitude reduces and the number of spikes decreases, reflecting drop out of motor units, decreased firing rate and decreased central motor drive resulting in less maximal activation (Kimura 2001c).
Two interpretations of ‘rollkur’ positions (Meyer 1992; Janssen 2003; Jeffcott et al. 2006) namely HNP4 and 7 were tested in this study, of which the HNP7 induced overall smaller changes than HNP4. This might suggest that if a deeply flexed HNP in a fixed position is applied for lunge training in average riding horses trained at this level, HNP7 reflects the neutral HNP1 more than HNP4. However, also the elevated position (HNP5) as well as the required position for competing (HNP2) are not without effect. The results of this study encourage further research on this topic expanding the knowledge on single fibre EMG in horses and applying this to highly trained equine athletes also in a ridden situation.
R. Schoobaar, Viasys Health Care, Houten, the Netherlands for his enthusiastic technical support. Animal Caretakers, Faculty of Veterinary Medicine, Utrecht University, Utrecht, The Netherlands for their professional and kind assistance.