Article Text

Download PDFPDF
The Elbow-EpiTrainer: a method of delivering graded resistance to the extensor carpi radialis brevis. Effectiveness of a prototype device in a healthy population
  1. Rishi Navsaria1,
  2. Dionne M Ryder2,
  3. Jeremy S Lewis2,
  4. Caroline M Alexander1
  1. 1Department of Physiotherapy, Imperial College Healthcare NHS Trust, London, UK
  2. 2Department of Allied Health Professions and Midwifery, University of Hertfordshire, Hertfordshire, UK
  1. Correspondence to Dr Jeremy Lewis, Department of Allied Health Professions and Midwifery, School of Health and Social Work, Wright Building, College Lane Campus, University of Hertfordshire, Hatfield, Hertfordshire AL10 9AB, UK; j.lewis8{at}


Background Tennis elbow or lateral epicondylopathy (LE) is experienced as the lateral elbow has a reported prevalence of 1.3%, with symptoms lasting up to 18 months. LE is most commonly attributed to tendinopathy involving the extensor carpi radialis brevis (ECRB) tendon. The aim of tendinopathy management is to alleviate symptoms and restore function that initially involves relative rest followed by progressive therapeutic exercise.

Objective To assess the effectiveness of two prototype exercises using commonly available clinical equipment to progressively increase resistance and activity of the ECRB.

Method Eighteen healthy participants undertook two exercise progressions. Surface electromyography was used to record ECRB activity during the two progressions, involving eccentric exercises of the wrist extensors and elbow pronation exercises using a prototype device. The two progressions were assessed for their linearity of progression using repeated ANOVA and linear regression analysis. Five participants repeated the study to assess reliability.

Results The exercise progressions led to an increase in ECRB electromyographic (EMG) activity (p<0.001). A select progression of exercises combining the two protocols increased EMG activity in a linear fashion (p<0.001). The ICC values indicated good reliability (ICC>0.7) between the first and second tests for five participants.

Conclusions Manipulation of resistance and leverage with the prototype exercises was effective in creating significant increases of ECRB normalised EMG activity in a linear manner that may, with future research, become useful to clinicians treating LE. In addition, between trial reliability for the device to generate a consistent load was acceptable.

  • Eccentric exercise
  • Exercise
  • Elbow injuries
  • Tendons

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.



Tennis elbow or lateral epicondylopathy (LE) is experienced as lateral elbow pain with a reported prevalence of 1.3%,1 and presents more commonly in the fourth and fifth decades.1 ,2 General practice data suggest that the annual incidence of lateral elbow pain is 4–7/1000 people.25

An episode of LE may last between 6 and 18 months.2 ,6 It is associated with considerable morbidity afflicting elite athletes, recreational athletes as well as non-sporting populations. In a study of 532 tennis players, 211 [39.7%] reported previous or current LE, with 24% reporting recurrent symptoms.7 LE is associated with a substantial economic and social burden, and may threaten participation in sport and adversely affect routine functional activities.

The typical symptoms of lateral elbow pain and weakness of the wrist extension with elbow supination often occurs if the tendinopathy specifically involves the extensor carpi radialis brevis (ECRB) tendon.8 ,9 Treatment for this condition includes non-surgical and surgical options, with surgery generally being considered only when non-surgical treatment has failed to achieve the desired outcomes. Non-surgical options include: rest, analgesics and non-steroidal anti-inflammatory medications, corticosteroid injections, taping, bracing, exercise and physical therapy.8 ,1012 Although the short-term effect of corticosteroids appears beneficial, the results 1 year postinjection suggest a poorer response than no injection and this practice may need to be reconsidered.8 ,13 ,14 A short-term benefit has also been reported with bracing and physical therapy mobilisations.5 ,8 ,15 Based on a continuum model for treating tendinopathy,16 ,17 a graduated exercise progression appears to be effective in pain control and restoration of function. With respect to LE, rehabilitation programmes focus on training the wrist extensors and elbow supinators.18 ,19 Although resistance exercises have been suggested, until now the electromyographic (EMG) response of a system to deliver an easily accessible (in clinic and at home) graduated resistance exercise progression has not been tested.

The aim of this investigation is to assess the effectiveness of a prototype device developed by one of the authors (JL) to deliver a graduated exercise progression for the wrist extensors and elbow supinators.



The study was conducted between November 2011 and January 2012, and took place in the Physiotherapy Department of Charing Cross Hospital, London, UK.


A convenience sample of 18 participants took part in the study. As this exercise protocol has not been studied in the past, the sample size was estimated using available data where integrated EMG from wrist extensors was normalised to maximum voluntary contraction.20 With a power of 0.8 and a significance set at 0.05, using ANOVA across 16 exercise progressions, a meaningful difference between the levels set at 5% and SD of 3%, a sample size of 15 is required.

Inclusion criteria were (1) aged 16–65 years, (2) no cervical or upper limb symptoms in the past 3 months, (3) no history of LE, (4) no forearm fractures within the previous 12 months and (5) competency in written and spoken English. Exclusion criteria included (1) a Beighton score of four or more, (2) any systemic inflammatory condition, (3) a history of carpal tunnel syndrome and (4) wrist and hand deformities (that would prevent holding the device). A Beighton score of four and above is a major criteria factor in diagnosing joint hypermobility syndrome.21 As joint hypermobility may influence joint movement and control,22 participants with a score of four or greater than four were excluded in case this introduced a confounding variable.

Study design

The study design involved a descriptive study of two exercises. All participants provided written informed consent prior to study inclusion (Research Ethics Committee number 09/H0707/3). All participants were then assessed for dominant handedness using the Edinburgh Handedness Inventory.23

The study incorporated two elements: (1) measuring EMG activity during the movement of the wrist from a position of wrist extension to wrist flexion and (2) measuring EMG activity during the movement of elbow supination to pronation. The order of these elements was randomised. In addition, a sample of convenience of five participants was recruited to repeat these two elements on a second occasion. They were tested at the same time of day but between 5 and 11 days after initial testing. This variation in time was to fit with the participant's availability.

Participant start position

Participants were asked to stand upright against a wall, with their dominant arm supported by an adjustable height table, with their elbow flexed to 90° and their shoulder resting against a wall in 0° of flexion, abduction and rotation.2426 The height of the supporting table was adjusted to support the dominant arm and to maintain the shoulder girdle at approximately the same level as the non-dominant side. The dominant wrist was free to fully flex over the end of the table (figure 1) when in the pronated position.

Figure 1

The prototype device. The length of the device can be adjusted telescopically (eg, lengths A and B here) using numbers that were sequentially assigned and evenly spaced to the holes (provided by the manufacturers) along the crutch. Additionally, the mass of the device was altered by placing 0, 0.5, 1 or 1.5 kg on the end of the crutch. Crutches A and B have the 1.5 kg weight placed upon the crutch. (1) Illustrates that a grip made of rubber was fashioned along the shaft of the crutch just below the original handle. The participant is able to hold the crutch along the shaft, which allows the crutch to be turned in the direction of elbow supination and pronation. (2) The participant holds the handle of the crutch with the wrist in extension, ready to lower the crutch into a flexed position.


Surface Ag-AgCl electrodes with a pick-up area measuring 6 mm×4 mm (Ambu N-10-A; Ballerup Denmark) were used to record EMG activity from the extensor muscles of the forearm. They were placed with the recording area 2 cm apart and positioned in a standardised way,27 with the midpoint of the two electrodes placed 47.6% of the distance on a line between the midpoint of the supracondylar ridge and the radial styloid.28 The aim of this position was to bias the recording such that it measured EMG activity from ECRB.28 EMG was amplified and filtered with a bandwidth of 30 Hz to 6 kHz (Digitimer; Herts UK). The data were collected and rectified at a sampling frequency of 4 kHz using an analogue to digital converter (1401, Cambridge Electronic Design; Cambridge UK) and stored using Spike 2 software (Cambridge Electronic Design; Cambridge UK). During the data collection, the researcher observed the movement of the participant and digitally marked a time point at the onset and offset of movement. Two cursors were subsequently positioned at the onset and offset of the movement according to the digital marks. To record the amplitude of the activity during the exercise, the root mean square (RMS) of the rectified EMG was measured between these cursors. To record the maximum activity, the maximum amplitude of rectified EMG was recorded for each of three contractions and an average calculated. EMG activity was normalised to the level of EMG activity achieved at maximum voluntary contraction. Therefore, each participant was asked to clench their hand into a fist and hold the wrist in neutral and pronation in order to activate ECRB.29 The participant was then asked to maximally contract their wrist extensors by pushing up against the investigator's resistance and the maximum EMG voltage was recorded.

Exercise resistance

A modified Coopers elbow crutch was developed for this investigation as a prototype device. The rationale for using this device was that the telescopic lower section, which is used in clinical practice to adjust the length of the crutch for an individual to facilitate gait, could also be used to incrementally increase and decrease the load on the muscle-tendon unit.

With the device being held at the original handle, a load could be applied to the muscle-tendon unit while the wrist was moved from extension to flexion (figure 1). This load could be increased as the telescopic lower section of the device was incrementally extended or a mass was applied to the end of the device. Additionally, a grip made of rubber was fashioned along the shaft of the device just below the original handle. This allowed the participants to hold the device along the shaft (figure 1). This allowed the device to be turned in the direction of elbow supination and pronation. Again, increasing the length of the telescopic lower section of the crutch or the addition of a mass at the end of the device would translate as increased load through the muscle-tendon units. However, the objective of this study was to determine if the prototype device would incrementally increase the EMG activity of ECRB as the resistance was varied.

The resistance from the exercise prototype was changed using two variables. First the length of the crutch was varied (figure 1) by using the different predefined holes (provided by the manufacturer) that change the crutch length. These positions were sequentially assigned a number and evenly spaced to the holes along the crutch. In this prototype device, numbers were placed at every third hole. Four numbers (1–4) were used and the largest number (4) was assigned to the longest position of the extendable arm. Second, at each of these predetermined lengths, no weight, 0.5, 1 or 1.5 kg weight was placed on the end of the crutch in a predefined position (figure 1). This gave 16 different exercises of different torques, that is four different device lengths with four different weights. Such methods have advantages and disadvantages. The advantages are that the exercise may be easily replicated and allow for incremental progressions. Additionally, using this pragmatic approach, this exercise could be progressed when tested in people with symptoms, and so may be appropriate as a home exercise for those suffering from LE. To this end, no fixation was applied to the participant as that would not be possible to easily recreate as a home exercise. The lack of fixation of the arm might lead to additional muscle activity as well as upper and lower arm movements as the resistance changes. An additional disadvantage of this method is that gravitational torque will vary depending upon the angle of the crutch and when adding additional weights.

Wrist extensor eccentric resistance

Participants were asked to grip the manufacturer's handle of the crutch in their dominant hand so that the length of the crutch was positioned along the lateral aspect of the hand. It was held with the wrist in full extension and slowly lowered to full wrist flexion. To standardise the time taken to complete the movement, participants were instructed to keep time with a computer-driven metronome which counted up from 0 to 5 s. Sixteen different sets of exercises were performed, giving the participants as much time as they needed to recover between each exercise (generally 1–2 min). The level of resistance was varied by changing the length of the crutch using the four predefined lengths and by adding the weights to the predefined position. Thus, the conditions used were: the gravitational moment of the crutch, and the gravitational moment of the crutch plus three additional masses of 0.5, 1.0 and 1.5 kg at each extension position of the crutch. Each mass was applied and tested separately. The order of variations in the crutch length and mass applied to the end of the crutch was randomised.

Elbow supination to pronation exercise progression

The participants’ start positions were as described previously. However, for this exercise, participants gripped the crutch in their dominant hand with the wrist in full supination and the crutch positioned at right angles to their forearm, holding the crutch on the fabricated rubber grip placed just inferior to the manufacturer's crutch handle (figure 1). The participant then actively rotated the crutch to full wrist pronation in time with the metronome counting from 0 to 5 s. The participant then repeated this process for each of the four positions of the extendable arm and the four weight levels at each of these positions.

Statistical analysis

A repeated-measures ANOVA was undertaken using SPSS V.19 (IBM, UK) to determine if the surface EMG activity as a percentage of MVC differed across the graded exercise gravitational moments.

Although unproven, clinically a graduated and linear increase in activity of the muscle is thought to be beneficial in treating tendinopathies. As such, regression analysis was used to explore the change between one level of exercise and the next in sequence. These changes were tested against relationship models including linear models to explore the best curve fit. Statistical significance was accepted at p≤0.05.

Intraclass correlation was used to indicate the reliability of the normalised electromyography (NEMG) from five participants’ data in all 16 conditions with Bland and Altman graphs created to interpret the reliability of the data.

In order to investigate whether a combination of the two exercise protocols led to an incremental change in NEMG, the data from the two exercise protocols were then plotted on one graph (figure 2). To explore how the amplitude of the NEMG varied with each progression, combinations of the two exercise protocols were then tested against relationship models including linear models to explore the best curve fit.

Figure 2

The surface electromyographic (EMG) activity of Extensor Carpi Radialis Brevis (ECRB) during two graded exercise progressions. The EMG activity is a percentage of EMG at maximum voluntary contraction. The error bars represent SD. The error bar has been placed in one direction for ease of visualisation. The exercise number represents an increase in crutch length (L1, L2, L3 and L4) with increasing mass (kg) placed on the end of the crutch (W0, W0.5, W1 and W1.5). The different lengths of the crutch were sequentially assigned and evenly spaced to the holes (provided by the manufacturers) along the crutch. The weights were placed in a predetermined position on the end of the crutch. The unfilled squares represent EMG activity of ECRB during eccentric exercise of the wrist extensors. The diamonds represent EMG activity of ECRB during pronation from a supinated position.


Eighteen participants (10 male) aged 28 years (SD=4.4; range 23–41 years) completed the study and five participants repeated the study to provide reliability data. No adverse effects were reported.

Both of the exercise progressions demonstrated an increase in percentage EMG activity (figure 2; eccentric wrist extensor exercise F ratio 129.2, df 15, p=0.000; supination to pronation exercise F ratio 103.0, df 15, p=0.000). Neither programme was found to produce a linear or incremental relationship between the EMG activity and the sequential load. However, in post hoc analysis, a selected combination of both programmes was found to have a statistically linear relationship (p<0.001). Here, the exercise programme begins with the eccentric exercise of the wrist extensors using the crutch at length 1 through length 4 without a weight (exercise L1W0–L4W0, figure 2). The programme then changes to the supination to pronation protocol at length 1 through length 4 with a 0.5 kg weight (exercise L1W0.5–L4W0.5, figure 2). It progresses to 1 kg weight using length 1 through length 3 (exercise L1W1–L3W1, figure 2) before reverting to the wrist extensor exercise with the crutch at length 1 and the addition of a 0.5 kg weight (exercise L1W0.5, figure 2). The programme then continues using the wrist extensor protocol of incrementally increasing the length before increasing weight.

Reliability of the study

The mean ICC value (model 2 individual) of RMS of NEMG of the five participants was 0.78 (95% CL 0.67–0.87). This indicates moderate agreement between the data. Note that the Bland and Altman plot indicates that a minor bias across the testing and meaningful differences in NEMG in these participants would likely be greater than 10% MVC. The Bland and Altman plot is presented in figure 3.

Figure 3

A Bland and Altman plot. The triangle represents two SDs above the mean, X represents two SDs below the mean. The square represents the mean. The data represent five participants repeating the two exercise progressions.


Many studies advocate the use of exercises in the treatment of LE.8 ,26 ,30 However, we were unable to identify any research that has investigated if a purpose-built device designed to incrementally increase resistance on the ECRB muscle-tendon unit produces a graded exercise programme. Among a number of conclusions, a recent systematic review of resistance exercises in the treatment of lateral epicondylosis reported that optimum dose and exercise progression had limited supporting evidence.31 This study is the first to identify a systematic and linear graded exercise for ECRB. The relevance of this finding would need to be tested in the treatment of people with LE.

It is interesting to note that the participants did not achieve more than an average of 30% of the EMG activity at maximum voluntary contraction (figure 2). This maximum was collected isometrically at a mid-point in range; however, the data were collected during isotonic eccentric contractions particularly during the extension to flexion exercise. EMG activity during an eccentric contraction is lower than during other types of muscle contractions.32 Nevertheless, given the difficulty that some participants had in completing the exercise protocol, it might seem surprising that a greater level of EMG activity was not achieved. One reason might be that even with regular breaks, fatigue could have impacted on the participants’ ability to complete the tasks. Another may be activity in other muscles that was not detected in this investigation.

It is important to note that EMG activity and not the load going through the tendon or the force generated by the muscle was measured here. It is unclear if there is a linear relationship between EMG activity and such a load through the tendon or force generated by the muscle. Indeed, it is clear that the relationship between force and EMG is complex,33 ,34 and can change even during the time of a contraction.35 However, it seems reasonable to suggest that this programme may represent a systematic increase in load. Importantly, the reliability data suggest that the outcomes of the study are likely to be reproducible.

Limitations of the study

Research using surface EMG is not without limitations. These include recording EMG from other muscles as well as ECRB.36 However, this is unlikely to confound the result presented here as the electrode position remained constant throughout the recording procedure and the recording of crosstalk is likely to have been stable throughout and would most likely be limited to agonists.

This study was conducted on healthy participants, and further studies recruiting people suffering with LE is required to determine whether the same graded increase in the EMG activity is found in the presence of a pathological tendon. Although it is not our intention to generalise our findings to people with LE, the participants in the study do not represent the entire spectrum of ages of people who suffer from LE. The influence of perceived exertion and fatigue was not tested in the current study. This may be a relevant factor to investigate in future research.

Finally, more robust conclusions may be derived with a larger study sample.


Graduated loading is considered to be the keystone for the management of overuse tendinopathies. This study investigated the EMG activity of the ECRB muscle in people without symptoms using a novel purpose-built device. If combinations of the two exercise progressions are used in a particular sequence, they can produce a linear increase in the surface EMG of ECRB. In addition, the results suggest that the technique is reliable. At present, we are unable to determine the response in people with symptoms or its effectiveness in clinical practice.

What are the new findings?

  • A combination of eccentric wrist extension and pronation exercises produces a graded increase in surface electromyographic activity in the healthy extensor carpi radialis brevis muscle.

  • The proposed graded exercise progression was fairly reliable in a small sample of uninjured participants.

How might it impact clinical practice in the near future?

Clinical studies are currently being conducted to determine if this approach is effective in the treatment of lateral epicondylopathy.


The authors wish to acknowledge the contribution of Professor Jill Cook (Monash University, Victoria, Australia), who recommended a method to progressively increase the resistance and activity of ECRB using supination and pronation at the elbow.



  • Contributors JL conceived the study, contributed to the overall methods, analysis of the data and writing of the manuscript. CA developed the EMG protocols, was involved in the collection of data and analysis and writing of the manuscript. DR contributed to the overall methods, was involved in the analysis of the data and writing of the manuscript. RN contributed to the methods, and was involved in the collection of data and analysis and writing of the manuscript.

  • Competing interests None.

  • Ethics approval Research Ethics Committee number 09/H0707/3.

  • Provenance and peer review Not commissioned; externally peer reviewed.