Muscle responses during radial nerve-biased upper limb neurodynamic test in asymptomatic individuals: a cross-sectional study

ABSTRACT

This laboratory cross-sectional study aimed at explore the muscle response (MR) of the upper trapezius, infraspinatus, biceps brachii and extensor carpi radialis brevis (ECRB) during the radial nerve-biased upper limb neurodynamic test (RN-ULNT) in healthy participants. Myoelectric activity was stage-by-stage recorded during two sequencing variants of the RN-ULNT: S1, in which elbow extension was the last movement; and S2, in which wrist flexion was the last movement. Final elbow and wrist joint angle and sensory response (SR) in five zones (Z1-Z5) were also registered. MR was qualitatively categorized as ‘absent’ (No-MR), ‘true’ (TMR) or ‘uneven’ (UMR). In both sequences, significant increases in muscle activity occurred mostly during shoulder abduction and elbow extension (p ≤ 0.009). Also, elbow extension but not wrist flexion increased the activity of the ECRB muscle (p ≤ 0.009). S2 showed significantly higher upper trapezius (p = 0.04) and biceps brachii (p = 0.036) muscle activity during wrist flexion, and higher report of SR in Z1 and Z4 (p < 0.001) compared to S1. Only the ECRB muscle showed significant differences in the MR type between S1 and S2 (TMR, p = 0.016; UMR, = 0,012). Our results may be useful in the assessment of upper limb musculoskeletal disorders.

Background

Applying stress through the use of upper limb neurodynamic tests (ULNTs) has become an essential method for assessing the sensitivity of peripheral nerves of the upper extremity to mechanical stimulation in physical therapy practice [1]. ULNTs involve an ordered sequence of head and neck, shoulder, elbow, forearm, wrist and hand positioning and movements aiming at exerting tension on a specific nerve trunk. Among physical responses to nerve stress, the muscle response (MR) represents a neuroprotective mechanism aiming at preventing an excessive lengthening of the neural bed and provides the clinician with indirect information regarding the link between nerve trunk irritability and muscle irritability, and its relation to signs and symptoms [2]. On the other hand, the sensory response (SR) involves descriptors such as stretching, burning, pricking or other sensations that may arise during nerve testing [3]. By far, the median nerve-biased ULNT has garnered the most attention in clinical and laboratory studies and, to date, MR during neurodynamic testing of the upper limb has been only explored during this test. Increased activity of the upper trapezius muscle and positioning-dependent restrictions in the cervical spine and elbow mobility has been demonstrated during median nerve testing in healthy individuals and patients with upper limb disorders [4–9].

The median nerve-biased ULNT has been recommended for the assessment of patients with carpal tunnel syndrome and cervical radiculopathy [10]. However, in patients with painful conditions that are anatomically and functionally related to the radial nerve (RN), such as lateral epicondylalgia [11], de Quervain’s tenosynovitis [12], and thumb osteoarthritis [13], the RN-biased ULNT (RN-ULNT) may be preferred in order to better assess if the patient’s symptoms are related to peripheral nerve irritability. While previous clinical studies support the use of RN-ULNT in painful conditions of the upper limb [14–16], the MR during the test has not been reported yet. Since upper limb positioning in both tests is different, it logically follows that the MR would be different as well. Unlike the median nerve-biased ULNT, during the RN-ULNT the shoulder is internally rotated, which has the potential to induce an MR from the infraspinatus muscle. Likewise, while for the median nerve testing the wrist undergoes extension, the RN-biased ULNT involves wrist flexion, which has the potential to activate the extensor carpi radialis brevis (ECRB) muscle [17]. Regarding the elbow, in both maneuvers the joint undergoes passive extension, which has been proven to induce an MR from the biceps brachii muscle [7]. Nevertheless, a higher response from this muscle could be expected during the RN-ULNT, since this test also involves pronation of the forearm.

Another important issue when applying ULNT is sequencing. The standard description of the RN-ULNT involves leaving shoulder abduction to the last stage. Although a standardized order is recommended in order to allow comparison between studies [10], the order of movements may be adjusted in clinical practice according to the patient’s presentation [18]. For example, in a patient with lateral elbow pain, the RN-biased ULNT may be performed by extending the elbow at the end of the test, while in a patient with dorsal wrist pain the clinician may leave wrist flexion for the last stage. Nee et al. found no differences in the final nerve position or strain at the forearm when three sequences of the median nerve-biased ULNT were compared in cadavers [18]. On the contrary, differences in the SR have been reported. Zorn et al. informed higher SR in the distal part of the upper limb when the median nerve-biased ULNT was sequenced from distal to proximal [19]. However, when the nearest joint to the affected segment of the nerve is moved first, afferent sensations are likely to be predominantly originated from joint tissues and related myofascial structures, and not from the nerve itself, since at the first stage the rest of the nerve is slack. On the contrary, if a progressive proximal-to-distal strain is induced, reaching a longitudinal ‘pre-tension’ state in the neural bed and removing the slack, then is possible that the last movement would place the distal part of the nerve under tension before a significant elongation is achieved in joint and myofascial tissues [10]. Thus, altering the neurodynamic sequencing during the RN-ULNT may have also the potential to influence both MR and SR, although this has been never studied. Elucidating normal responses RN-ULNT using different sequencings may contribute to improving the interpretation of testing outcomes in the clinical context.

Thus, the aim of the present study was to measure the normal MR and its relation with SR during two sequencings of the RN-ULNT in asymptomatic individuals. The working hypotheses were that: 1) the RN-ULNT evokes an MR from the upper trapezius, infraspinatus, biceps brachii, and ECRB muscles; and 2) the sequencing variants used in this study have an influence only over the activity of biceps brachii and ECRB muscles.

Materials and methods

Study design

A cross-sectional study was conducted at the University Center of Assistance, Teaching and Research (CUADI) of the University of Gran Rosario (UGR), following the STROBE guidelines. Thirty healthy male and female individuals (15 male, 15 female) aged between 20 and 50 years were recruited by convenience sampling from August 2019 to November 2020. After explaining the overall aim of the study and collecting general information, subjects were asked about their job and sports participation and to complete the Revised Upper Extremity Work Demands (R-UEWD) scale [20] to assess the exposure to mechanical overuse at work. Table 1 shows the demographic characteristics of the sample. The exclusion criteria were: having been tested with ULNTs before, previous surgery in the neck and/or upper limbs, pain in the upper quadrants lasting more than 3 weeks in the previous 6 months, and preexisting mobility deficits in the scapular girdle, shoulder, elbow, wrist, and fingers. The ethical principles of the Declaration of Helsinki were followed and participants signed the informed consent. The study was approved by the Comité Provincial de Bioética (number 789).

Table 1.

Demographic characteristics of the sample.

	Sample (n = 30)	Men (n = 15)	Women (n = 15)
Age – years, SD (Min, Max)	31.57 ± 5.54 (21. 41)	33 ± 5.79 (22. 41)	30.3 ± 5.7 (21. 39)
Height – cm, SD	171.17 ± 9.59	178.4 ± 5.95	163.87 ± 6.42†
Weight – kg, SD	72.4 ± 16	76.83 ± 12.83	60.87 ± 8.99†
UEWD-R – mean, SD	9.43 ± 2.54	9.33 ± 3.17	9.53 ± 1.8

† Statistically significant (p < 0.05)

Procedures

Equipment set-up

Participants were assessed in a single session. Surface electromyography (sEMG) assessment was performed using an acquisition module with four analog channels (Miotec™, Biomedical Equipments, Porto Alegre, Brazil). The conversion from analog to digital signals was performed by an A/D board with 14-bit resolution input range, sampling frequency of 2 kHz, common rejection module greater than 100 dB, signal–noise ratio less than 03 μV Root Mean Square and impedance of 109 Ω. The sEMG signals were recorded through surface Ag/AgCl electrodes (Meditrace™, Canada) with a center-to-center distance of 20 millimeters, in a parallel orientation to the underlying muscle fibers. Preparation of the skin was performed following standardized guidelines. Table 2 resumes the location of electrodes and the rationale for the selection of each muscle. A ground electrode was placed in the lateral epicondyle of the contralateral arm. Appropriate muscle activity was confirmed by active contraction against resistance [21]. Testing was videotaped to allow synchronization between movement and sEMG data. A standard goniometer was used for the measurement of the final elbow and wrist flexion angle (EFA and WFA, respectively). A body chart of the upper quadrant was used for the registration of the SR (see Data processing).

Table 2.

Rationale for selection and electrodes location for each muscle.

Muscle	Rationale	Elecrodes location
UT	Protects the brachial plexus by elevating the scapular girdle	Halfway between C7 acromion process and acromion
IS	This muscle may protect the RN against stretch during shoulder internal rotation	Two fingers below the spine of the scapula over the muscle belly
BB	This muscle may protect the RN against stretch during elbow extension and forearm pronation	Over the muscle belly, approximately halfway of the arm
ECRB	The ECRB has a role in wrist joint stability, and it may prevent an excessive RN stretching during wrist flexion	Lateral to the brachioradialis, at the proximal third of the forearm palpating the muscle belly in line with the lateral epicondyle

UT = upper trapezius muscle; IS = infraspinatus muscle; BB = biceps brachii muscle; ECRB = extensor carpi radialis brevis muscle; RN = radial nerve.

Testing protocol

Participants were asked to lie in supine position having the contralateral arm relaxed at the side of the body. The legs were placed on a wedge-shaped pillow in order to obtain a neutral tension position in the lumbar and lumbosacral plexus. The dominant arm was assessed in all participants. Participants were asked to keep the head in neutral position. Next, participants were explained that the arm would be passively, sequentially and gently moved to progressively lengthen the RN. The use of words with a negative or warning connotation such as ‘tension’ and ‘pain’ was avoided. For familiarization purposes, a quick demonstration was performed using the ipsilateral lower limb as if the sciatic nerve was to be tested.

The RN-ULNT was performed manually, in order to mimic the use of the test in clinical practice more closely. Two sequences of the test (S1 and S2) were studied, altering the order of joint movements in the distal part of the arm [10]. Both sequences included six stages, each lasting 10 seconds. Three assessors were involved in the testing protocol, one of them being exclusively responsible for the stage 1 of the test, i.e. full scapular girdle depression, since this component of the test is critical to prevent looseness in the brachial plexus during ULNTs [9]. The second assessor was responsible for the movement of the arm, which was induced at a low pace and without a formal measurement of movement speed. Starting with the elbow of the tested side flexed at 90°, the following stages of the test were [22,23], for S1, 2) shoulder abduction to near 90°, 3) full shoulder internal rotation, 4) full forearm pronation, 5) full wrist flexion with fingers flexed, and, finally, 6) elbow extension from 90° flexion to 0°. This sequence sought to add tension on the RN both proximally and distally to the elbow, leaving some degree of nerve slack in order to assess how elbow extension affects MR and SR [23]. For S2, the sequence was the same from stages 1 to 4, but elbow extension (stage 5) preceded the wrist flexion (stage 6). During S2, it was not pretended to achieve a full elbow extension. Instead, participants were asked to notify the assessor when the first stretching, non-painful sensation was perceived as elbow extension progressed. At that moment, the movement was stopped in order to reach a ‘pre-tension’ state and prevent an excessive strain in the RN before wrist flexion. In both sequences, movement was stopped in the last stage when the participant perceived the SR and the third assessor measured the final EFA (S1) and EFA (S2), and registered the SR. Figure 1 shows the initial and final positioning for the RN-ULNT.

Figure 1.

Positioning for the RN-ULNT. (a) Initial position. (b) Final position.

Data processing

Electromyographic signals were analyzed using the Miotec™ Suite 1.0 Software (Miotool 400, Porto Alegre, Brazil). Raw signals were rectified using root mean square conversion. Mean muscle activity in µV was recorded from the movement onset to the end of each stage. Also, since a trend to a homogeneous MR pattern was not observed among participants, an analysis of the behavior of each muscle was conducted considering the moment of muscle onset. Muscle onset was determined using the threshold of three standard deviations (SD) above the mean activity during the resting stage [9]. In order to describe each muscle’s behavior, MR was categorized as either ‘absent’, ‘true’ or ‘uneven’ (Table 3). In the case of a true MR (TMR), the stage in which onset occurred was also registered. EFA and WFA were measured in degrees. The body chart of the upper quadrant used for registering the SR included five zones (Z1-Z5; Figure 2) [15].

Table 3.

Definition of MR types.

No-MR	Absence of MR, i.e. no sustained response is observed through the test. This category included any transient increase in muscle activity exceeding the threshold, but not lasting until the end of the stage in which it occurred
TMR	An increase in muscle activity above resting levels -i.e. onset- in any stage lasting to the end of the test, without dropping back below the threshold
UMR	Muscle onset lasting until the end of the stage, but showing thereafter a decrease in activity below the threshold in any point of the rest of the test. Type I: After falling below the threshold, a new onset occurs lasting to the end of the test. Type II: No new onset is observed

No-MR = No muscle response; TMR = True muscle response; UMR = Uneven muscle response

Figure 2.

Body chart of the upper quadrant for the registration of the SR. The zones (Z) were 1) neck and shoulder, 2) lateral part of the elbow and proximal forearm, 3) posterior aspect of the forearm, 4) posterior aspect of the wrist, and 5) any other part of the arm.

Statistical analyses

Since this was an observational study, no formal calculation of the sample size was performed [3], and the final sample was determined based on previous studies assessing MR in asymptomatic individuals [3,4,6,7,9]. All statistical analyses were performed using SPSS software 20.0 (IBM Corporation, Somers, NY, USA). The Kolmogorov–Smirnov test showed that mean MR values had a non-parametric distribution. Wilcoxon signed rank tests were used for both within-sequence (comparison between each stage and the following one) and between-sequence (differences in MR per stage) analysis. The U-Mann Whitney test was used for between-sex comparisons in the MR at the final stage of S1 and S2. The correlation between muscle activity and joint angle (EFA and WFA) was assessed using the Spearman test. Descriptive data for categorical variables such as MR type and SR was expressed as number and percentage. The McNemar test was used to assess differences between sequences in the frequency of SR and MR type. The binomial test was used to compare the distribution of the MR type for each muscle within a sequence. A p-value less than 0.05 was considered statistically significant.

Results

Muscle activity

Figure 3 shows the activity of each muscle during S1 and S2.

Figure 3.

From top to bottom, muscle response for the upper trapezius (UT), infraspinatus (IS), biceps brachii (BB) and extensor carpi radialis brevis (ECRB) muscles during S1 and S2. SGD = scapular girdle depression; SA = shoulder abduction; SIR = shoulder internal rotation; FP = forearm pronation; WF = wrist flexion; EE = elbow extension. (*) indicates a statistically significant change in muscle activity compared with the previous stage. (ᶲ) indicates a statistically significant higher muscle activity compared to the other sequence.

Within-sequence analyses

In both sequences, shoulder abduction provoked a statistically significant increase in muscle activity for the upper trapezius, infraspinatus and biceps brachii muscles (p ≤ 0.001), while shoulder internal rotation consistently increased the activity of biceps brachii and ECRB (p < 0.001). Curiously, forearm pronation produced a decrease in biceps brachii muscle activity which was only statistically significant in S2 (p = 0.008). Elbow extension significantly increased the activity of all muscles in S1 and S2 (p ≤ 0.009), except for the infraspinatus muscle in S1. In S1 and S2, elbow extension but not wrist flexion induced a significant increase in the myoelectric activity both in biceps brachii and ECRB muscles (p ≤ 0.009).

Between-sequence analyses

In between-sequence analyses, Wilcoxon signed rank tests revealed statistically significant differences in upper trapezius muscle activity during shoulder abduction (S1< S2; p = 0.039), and in infraspinatus muscle activity during forearm pronation (S1> S2; p = 0.01). No difference in the activity of any muscle was observed during shoulder internal rotation. In S2, wrist flexion induced a significantly higher activity in upper trapezius (p = 0.04) and biceps brachii (p = 0.036) muscles compared with S1, while no difference was observed during elbow extension.

Between-sex comparison in mean muscle activity at last stage revealed no significant differences, except for the upper trapezius muscle activity in S2 in which women showed a slightly higher activity (p = 0.40).

Correlation between EFA and WFA with MR

In both sequences, Spearman correlation analyses revealed no significant correlation between mean muscle activity and final joint angle, except for the ECRB activity and EFA which showed a weak positive correlation in S1 (p = 0.02; r = 0.42).

Sensory response

The dorsal aspect of the forearm (Z3) was the most common region of SR (63.3% in S1, 70% in S2), followed by Z5 (36.7% in both S1 and S2), Z2 (30% in S1, 36.7% in S2), Z4 (6.7% in S1, 10% in S2) and Z1 (3.3% in S1, 6.7% in S2). McNemar tests revealed that there was a statistically significant higher frequency of SR in Z1 and Z4 during S2, in comparison with S1 (p < 0.001).

Type of MR

Table 4 shows the distribution of TMR and UMR type I and II for each muscle during S1 and S2, respectively (No-MR cases are not included). Only the ECRB muscle showed statistically significant differences in the frequency of the MR types between both sequences (TMR, p = 0.016; UMR, = 0.012).

Table 4.

Frequency of each type of MR in S1 and S2 (No-MR not included).

MR type	S1	S2	S1	S2	S1	S2	S1	S2
TMR
SGD	17.9	3.7	8.7	0	0	0	0	0
SA	67.9	81.5	43.5	58.3	22.2	11.5	0	8.7
SIR	3.6	7.4	4.3	25	25.9	23.1	20.8	8.7
FP	7.1	0	8.7	0	3.7	0	4.2	21.7
EE	0	3.7	0	4.2	0	30.8	12.5	17.4
WF	0	0	8.7	4.2	14.8	15.4	0	13
Subtotal	96.4	96.3	73.9	91.7	66.7	80.8	37.5	69.6†
UMR
Type I	3.6	3.7	26.1	8.3	29.6	19.2	50	30.4
Type II	0	0	0	0	3.7	0	12.5	0
Subtotal	3.6	3.7	26.1	8.3	33.3	19.2	62.5	30.4†
TotalTMR + UMR	28	27	23	24	27	26	24	23

Values expressed as number (percentage). MR = muscle response; TMR = true muscle response; UMR = uneven muscle response; SGD = scapular girdle depression; SA = shoulder abduction; SIR = shoulder internal rotation; FP = forearm pronation; WF = wrist flexion; EE = elbow extension; UT = upper trapezius; IS = infraspinatus; BB = biceps brachii muscle; ECRB = extensor carpi radialis brevis

† Statistical significant difference for the same muscle between S1 and S2 (p < 0.05)

* Statistical significant difference for the same muscle between TMR and UMR (p < 0.05)

Discussion

The aim of the present study was to describe the MR during the RN-ULNT. The results showed that, compared with the activity levels at rest, upper trapezius, infraspinatus, biceps brachii and ECRB muscle activity was significantly higher during the test, thus confirming our first hypothesis and suggesting that these muscles may activate to prevent excessive RN lengthening. The second hypothesis was partially confirmed, since biceps brachii and upper trapezius, but not ECRB muscle activity was affected by the change in the sequencing.

Previous studies assessing MR during median nerve-biased ULNT suggested that upper trapezius muscle acts protectively against brachial plexus tensioning [4,6,7,9]. The upper trapezius muscle is lengthened when the distance between the cervical spine and the acromion increases during scapular girdle depression. However, as reported by Balster and Jull [4], in our study scapular girdle depression by itself did not significantly increase muscle activity above resting values, probably because the rest of the system is still in slack. Adding shoulder abduction significantly raised the activity of upper trapezius, infraspinatus and biceps brachii muscles. Shoulder abduction increases the distance between the midpoint of the humeral diaphysis and the coracoid process, lengthening the neural bed and thus increasing brachial plexus tension [15]. On the other hand, the activation of infraspinatus and biceps brachii muscles during this stage was unexpected. It is possible that these muscles activate in order to pull the humeral shaft proximally, although no previous study support this notion. In addition, unlike what was expected, a trend toward a decrease in upper trapezius and -especially- biceps brachii muscle activity was observed during forearm pronation. It has been reported that the radius migrates proximally during pronation, which may contribute to decrease biceps brachii muscle activity [24–26]. On the other hand, forearm pronation showed a trend to a non-significant increase in ECRB activity which might be due to a combination of local tension and pressure upon the RN [15,27]. Thus, it is possible that pronation reduces muscle activity proximally to the elbow, but simultaneously increases RN stress and muscle activity at the proximal forearm.

Between-sequences analyses showed that the muscle activity pattern of the upper trapezius and biceps brachii varied similarly. In both muscles, a decrease in activity – which was significant only for the upper trapezius – was observed when wrist flexion preceded elbow extension (i.e. in S1), while activity was increased when elbow extension was introduced before wrist flexion (S2). Anatomic studies showed that elbow flexion decrease RN tension at the elbow, and that RN slides proximally if shoulder abduction to 110° is added [28,29]. On the other hand, wrist flexion with the elbow flexed to 90° elicits distal excursion and compression of the RN [29], and elbow extension plus wrist flexion causes distal excursion and increased strain on the RN [29]. Thus, it is possible that in S1, before elbow extension, proximal muscle activity would be more directed to protect the brachial plexus, while distal – i.e. ECRB – muscle activity is likely to be protective of the RN. Then, when elbow extension is added, proximal muscle activity may again be increased in order to protect the RN. On the other hand, in S2, elbow extension significantly increased the activity of all muscles, and final wrist flexion contributed to an additional increase in muscle activity probably because the RN was already under strain [30].Therefore, wrist flexion may be more provocative if added at the end of the test. This might explain the higher activity of upper trapezius and biceps brachii muscles and the higher report of SR in Z1 and Z4 during S2, compared to S1.

To the best of our knowledge, no previous study has performed a distinction of MR types based on muscle onset such as in the present study. The proposed categorization had the intention of establishing a qualitative distinction between the assessed muscles regarding their behavior during RN-ULNT. Previous studies showed a higher rate of upper trapezius response compared with more distal muscles [6]. Legakis and Boyd reported a response rate of 68.4% for the upper trapezius and 52.6% for the biceps brachii using the median nerve-biased ULNT [9]. The higher rate of response in our study may be explained by the method of scapular stabilization. In our study, one of the assessors was exclusively responsible for scapular girdle depression, while in the study from Legakis and Boyd the same tester blocked the scapula with the forearm. This might have resulted in some loss of scapular stability at the end of the test. On the other hand, the ECRB showed the greatest variability in the MR type between sequences. In S1, only 9 participants (37.5%) had a TMR. It is possible that, after ECRB activation during shoulder internal rotation and the elbow still flexed to 90°, changing the grasp from the distal forearm to the hand for inducing wrist flexion may have deactivated the muscle in some participants. Reactivation was likely to occur at the end as a result of elbow extension. On the contrary, during S2, it is possible that ECRB remained active before wrist flexion due to increased stress of the RN because of elbow extension, and that sustained ECRB activity during wrist flexion was most probably the result of RN loading, likely explaining the statistically significant higher rate of TMR (69.6%). Unfortunately, since the stage of muscle deactivation and reactivation was not registered in cases of UMR, it is not possible to conclude whether these differences are due to the change in the sequencing.

This study has some limitations. We believe that one of the most important is that only healthy individuals were included. In some fields such as neurophysiology and biomechanics, studies in non-symptomatic participants serve as first step in the exploration of biological phenomenon when no previous data are available. Therefore, our results may serve as reference for future studies including patients with upper quadrant pain. Also, the order of S1 and S2 was not randomized. It is reasonable that this might have resulted in a less overall mean muscle activity during S2 compared to S1, because of the effect of familiarization with the test. However, judging by between-sequence analysis, no significant differences existed. Another important limitation was that we did not assess a muscle not expected to respond to RN testing (e.g. triceps brachii), which may have resulted in a confirmation bias. Also, not using instrumental methods for stabilizing the upper limb resulted in some unexpected responses. Thus, the results of the present study may not be directly and entirely comparable with previous electromyographic studies using electromechanical devices for moving the arm. As we stated, our aim was to more closely reproduce the use of the test in clinical practice. Finally, disparities in ECRB muscle activity may be due, at least in part, to cross-talk and changes in the location of the electrodes as a result of arm movement.

Conclusion

In this sample of healthy participants, our results showed that upper trapezius, infraspinatus, biceps brachii and ECRB muscles were responsive to the RN-ULNT, especially during shoulder abduction and elbow extension. In addition, changing the sequencing affected both MR and SR. Our results may be considered by clinicians assessing the mechanosensitivity of the RN and the potential involvement of upper trapezius, infraspinatus, biceps brachii and ECRB in patients with painful conditions of the upper quadrant.

Funding Statement

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Disclosure statement

No potential conflict of interest was reported by the author(s).