RESEARCH NOTE

A NEW TOOL TO ASSESS STRETCHING INTENSITY

THE

PERCEPTION

OF

SANDRO R. FREITAS,1 JOA˜O R. VAZ,1 LUIS GOMES,2 RUI SILVESTRE,2 EDGAR HILA´RIO,2 NUNO CORDEIRO,2 FILOMENA CARNIDE,1 PEDRO PEZARAT-CORREIA,1 AND PEDRO MIL-HOMENS1 1

CIPER, Faculty of Human Kinetics, University of Lisbon, Lisbon, Portugal; and 2Superior Health School, Polytechnic Institute of Castelo Branco, Castelo Branco, Portugal ABSTRACT

Freitas, SR, Vaz, JR, Gomes, L, Silvestre, R, Hila´rio, E, Cordeiro, N, Carnide, F, Pezarat-Correia, P, and Mil-homens, P. A new tool to assess the perception of stretching intensity. J Strength Cond Res 29(9): 2666–2678, 2015—This study aimed to develop a valid and reliable scale to assess the perception of stretching intensity below and above the maximal range of motion. Experiments were conducted through a passive leg extension angle-torque assessment to healthy population (n = 90). In the study’s first phase, the visual, numerical, and description of the stretching intensity scale (SIS) components were developed. The visual analog scale (VAS) score, absolute magnitude estimation (AME) score, and verbal stretching intensity symptom descriptors were assessed for different stretching intensities. In the second phase, the SIS was tested for validity, reliability, scale production, and estimation properties as well as responsiveness to stretching. In the first phase, a high correlation was found between SIS score and range of motion (ROM), as well as SIS and torque in both submaximal (intraclass correlation coefficient [ICC] = 0.89–0.99, r2 = 0.88–0.99) and supramaximal (ICC = 0.75–0.86, r2 = 0.68– 0.88) stretching intensities. The AME and VAS scores fitted well in an exponential model for submaximal stretching intensities (y = 14.829e0.0187x, ICC = 0.97 [0.83–0.99], r2 = 0.98), and in a linear model for supramaximal stretching intensities (y = 0.7667x 25.751, ICC = 0.97 [0.89–0.99], r2 = 0.9594). For the second phase, a high correlation was found between SIS score and ROM (r = 0.70–0.76, ICC = 0.76– 0.85), as well as SIS and torque (r = 0.62–0.88, ICC = 0.57– 0.85). The interday reliability was high to produce (r = 0.70, ICC = 0.70 [0.50–0.83]) or estimate (r = 0.89, ICC = 0.89 [0.82– Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s Web site (http://journals.lww.com/nsca jscr). Address correspondence to Sandro R. Freitas, [email protected]. 29(9)/2666 2678 Journal of Strength and Conditioning Research Ó 2015 National Strength and Conditioning Association

2666

the

M

Journal of Strength and Conditioning Research

0.93]) stretching intensities. The acute stretching effects on ROM and passive torque were detectable using the SIS. It is expected a high application in assessing the stretch intensity using the SIS in future studies and practical interventions.

KEY WORDS joint flexibility, passive torque, range of motion, reliability, validity INTRODUCTION

I

t is known that humans have the ability to rank perceived sensations from mechanical stimulus (3,23). Stretching is practiced worldwide in both sports and rehabilitation; however, the methods and instruments used to measure stretching intensity are divergent and lack consensus (1,2,5,26). “Stretching intensity” has been defined as the degree of muscle tendon lengthening induced by a change in joint range of motion (ROM), that is controlled by subjective assessment of human tolerance to stretch using the criteria of pain or discom fort (18,21). When stretching intensity is increased, a greater passive joint torque is observed due to the increase in tissue tension (17). Stretching intensity has conventionally been set by using pain thresholds (2,5,18) without using a validating scale (2); or as a percentage of the joint maximal ROM (mROM) (1,26). Previous studies have examined muscle stretching and joint ROM by measuring joint passive angle torque responses (13,16). However, the relation between flexibility performance (i.e., mROM), joint mechanical response (i.e., torque), and per ceived exertion has never been studied. It is known that during stretching, the mROM increases following subsequent repetitions due to a greater tolerance of muscle tendon lengthening (16). This means that the stretching intensity can be acutely increased above its initial maximum after a repetition of stretching. Thus, if we con sider the initial mROM (first repetition), it is possible to define a supramaximal domain (i.e., above the initial maxi mum). However, it is unknown how the perception of the stretching intensity will change and how these changes will correlate with the physiological responses. Such a correlation may help suggest that the perceived stretching exertion is a valid outcome for assessing stretching intensity. Nonethe less, there is no specific instrument constructed to assess stretching for intensities below (i.e., submaximal) and above

the

TM

Journal of Strength and Conditioning Research (i.e., supramaximal) the mROM. The current perceived exer tion scales (e.g., Borg scale) has not been validated to mea sure stretching intensity. A scale designed for this purpose would help to determine the perception of stretching inten sity for different intervention settings. This study aimed to develop, validate, and determine the reliability of a specific scale for assessing the perception of stretching intensity based on Borg’s continuum model (3). Instead of validating an existing scale (e.g., OMNI scale), it was decided to develop a new instrument. The reasons for that were as follows: (a) previous scales did not take a stretching construct into consideration; (b) no specific stretching instruction has been developed; (c) the scale met rics, anchors, and descriptors have not been correlated with stretching physiological variables; and (d) previous scales did not contemplate scale properties to assess supramaximal stretch intensities (i.e., supramaximal ROM). We hypothe sized that the stretching perception scale outcome would be related to the joint angle torque response, as a consequence of tissue deformation. A passive stretch targeting the leg flexors, by means of knee extension, was used for the pur pose of this study.

METHODS Experimental Approach to the Problem

This study included 2 phases (Figure 1A). The first phase aimed to develop the scale (i.e., determine the scale’s visual, numerical, and descriptive components) and to systematize the scales instruction to be used in phase II. The second phase was designed to validate the scale for different items (i.e., content, construct, face, and criterion), test the reliability of the scales properties (i.e., stretch intensity prediction and production), determine intra and inter test reliability, and to test the scale’s responsiveness to acute changes induced by stretching. Four sessions were performed in phase I, and 3 in phase II. All sessions were performed on different days, with at least 24 hours between sessions. As detailed below (i.e., section Conceptual Scale Develop, Validation, and Reliability), the present scale was developed to determine submaximal and supramaximal stretching intensities. The supramaximal intensities were considered as those per formed to a joint angle greater than the angle determined in the initial (i.e., before static stretching) mROM test, which did not elicit pain. This angle was determined at the begin ning of each session (please see in section Procedures). Sub maximal intensities were performed below the initial mROM. Thus, in phase I, the first session was performed for familiarization purposes, the second determined the sub maximal components of the scale, and the third and fourth sessions determined the supramaximal scale components. In phase II, the first session was performed for familiarization, and the second and third sessions aimed to determine the validity and reliability of the stretching intensity scale (SIS). The details of each session are presented in the Procedures section. A passive leg extension test (i.e., to stretch the leg

| www.nsca.com

flexors) was used for the purpose of this study (Figure 1B) (8). A dynamometer (Biodex system 3 research, Shirley, NY, USA) was used to impose the stretch with a velocity of 28$s 1 in all tests. The primary investigator of the study was not involved in data collection, so that the scale administration could be blinded. Thus, 3 experienced researchers were in structed and trained in the study’s procedures, equipment measurement, administration of the scales, and questionnaires. Subjects

Two groups of participants were recruited for the study. All participants were older than 18 years. Informed consent was obtained from all the participants. Sixty participants (age = 21.1 6 1.8 years; height = 1.77 6 0.07 m; body mass = 71.4 6 11.8 kg) were involved in phase I, and 30 participants (age = 26.4 6 4.6 years; height = 1.75 6 0.07 m; body mass = 72.4 6 10.0 kg) were included in phase II. Only 30 of the subjects in phase I (age = 21.3 6 1.9 years; height = 1.76 6 0.07 m; body mass = 67.7 6 8.8 kg) participated in the third and fourth sessions of phase I. Only participants with active knee extension flexibility lower than 1608 (hip flexed at 908 and the full knee extension corresponding to 1808) partici pated in the study to ensure that the highest knee angle would be achieved before full knee extension was reached. All participants were Portuguese, physically active, injury free, and mostly academic students with an academic degree. They reported having no injuries or orthopedic problems in their lower limbs and lower back at the time of the study and were all familiar with the concept of stretching. The right lower limb of all participants was tested. Only male subjects were recruited to eliminate any potential uncertainty due to sex difference (11,14). The local Ethics Council approved the study (#1/2013). Outcomes

Joint Angles. The ankle, knee, and hip angles were assessed in all stretching tests using a digital camera (JVC, GR DVL9800U, Yokohama, Japan) placed parallel to the sagittal plane and operated at 50 Hz (8). Reflective markers were placed over the head of the first metatarsal, the medial fem oral condyle, and the medial malleolus of the right lower limb; markers were also placed over the greater trochanter of the left femur and on the left side of the trunk at the intersection of a transverse line passing over the spinous process of the first lumbar vertebra and a line linking the greater trochanter and the midaxillary point (Figure 1B). The Ariel Performance Analysis System software was used to analyze the kinematic data. Passive Torque. Resistance to the passive leg extension force was measured at 50 Hz, since we previously observed no significant differences compared to the force measured at 1000 Hz, during slow stretching maneuvers (8). All stretch ing tests were performed with the equipment shown in Fig ure 1B. This device had a force sensor (Platform Load Cell 1042; Sensor Techniques Ltd., Cowbridge, UK) and was VOLUME 29 | NUMBER 9 | SEPTEMBER 2015 |

2667

Perception of Stretching Intensity Measurement

Figure 1. A) Design of the study. B) Experimental passive leg extension setup to induce a stretching for the leg flexors (reprinted with from Freitas et al. (8). Institute of Physics and Engineering in Medicine. Reproduced by permission of IOP Publishing. All rights reserved). C) Passive torque response during a maximal range of motion test with the indication of the point at which the participant started feeling the onset of stretching symptoms. D) Example for 1 participant of rest interval (left) and non rest interval (right) stretching protocols. In this example, the participant has performed 5 repetitions with a rest interval, and 4 stretching repetitions without a rest interval. Adaptations are themselves works protected by copyright. So in order to publish this adaptation, authorization must be obtained both from the owner of the copyright in the original work and from the owner of copyright in the translation or adaptation.

used to ensure that the distance from the measurement site to the knee axis was not affected by misalignment of the dynamometer shaft with the knee axis. Knee passive torque was then determined by multiplying the passive resistance to knee extension by the distance between the knee axis and the site where the resistance to stretch was measured (8). Stretching Intensity Scale Descriptors. A list of verbal descrip tors was used to determine the term that best described the

2668

the

M

Journal of Strength and Conditioning Research

participant’s perceived stretch intensity degree. The descrip tors were written in Portuguese: nenhum (i.e., none), muito pouco (i.e., very few), pouco (i.e., few), moderado (i.e., moder ate), muito (i.e., high), quase ma´ximo (i.e., almost maximal), and ma´ximo (i.e., maximal) for the submaximal component of the scale; and ma´ximo (i.e., maximal), pouco supra ma´ximo (i.e., a few supramaximal), quase supra ma´ximo (i.e., almost supra maximal), e supra ma´ximo (i.e., supramaximal), for the supra maximal scale component.

the

TM

Journal of Strength and Conditioning Research Stretching Intensity Absolute Magnitude Estimation. The abso lute magnitude estimation (AME) method (24) was used in sessions of phase I to determine the numerical scaling within the stretch repetitions performed at different intensities. The AME allowed us to determine the magnitude of cognitive perception of the stretching intensity to a physical stimulus, by attributing a number to classify the stimulus intensity (24). Participants were instructed to consider 100 as correspond ing to the “maximal stretching intensity they could perform without feeling pain.” Such intensity was previously deter mined in the first repetition of each experimental session. Participants were informed that they could attribute “any number below or above 100, based on the stretching inten sity they perceived.” The AME score was determined imme diately after each stretching repetition. Pain and Stretching Intensity Visual Analog Scale. A 100 mm visual analog scale (VAS) was used to assess both the stretching intensity (8) in all sessions, and pain in the ses sions of phase I (7). The VAS is used to visually quantify the intensity of a perceived mechanical stimulus, in this case the perception of stretching intensity. The VAS scale was anchored with the words “no stretch” and “maximum pos sible stretching” as the left and right anchors, respectively. For pain, the scale was anchored with the words “no pain” and “worst imaginable pain” as the left and right anchors, respectively. The linear distance from the left anchor to the subject’s mark determined the VAS score. The VAS stretch ing results were used to determine the AME score position in the scale and to determine the position of scale descriptors. Body Chart. The body region where subjects felt the stretching symptoms were also registered for the participants of the phase I. For this, a body map adapted from a previous study (4) was used. Participants were instructed to point the sites that best represented the stretching perception. Onset of Stretching Sensation. The point at which participants of phase I started feeling the onset of stretching symptoms (OS) during the stretching repetition (Figure 1C) was as sessed using a trigger that was synchronized with others outcomes (4). The participants held the trigger in hand and pressed the button when they felt the first symptoms of stretching in all repetitions. The angle corresponded to OS allowed us to determine 2 ranges of motion (from OS to mROM and from starting test position to mROM), whereas the mechanical and psychometric data were used for valida tion purposes (see in detail in the Conceptual Scale Develop, Validation, and Reliability Section). Electromyography. To ensure a passive condition of the stretching tests, the average amplitude of the surface electro myography (sEMG) was measured in the semitendinosus and quadriceps vastus medialis during the stretching tests. Surface

| www.nsca.com

bipolar electrodes (Plux, Lisbon, Portugal, gain of 1000) were placed over the midportion of each muscle with a 20 mm center to center distance, in accordance with SENIAM guidelines (10). The skin was previously shaved, slightly roughened using abrasive sandpaper, and cleaned with alco hol. The ground electrode was fixed over the left patella. The sEMG signals were amplified (input impedance .100 MV; CMRR = 110 dB) and A/D converted (MP100 Biopac Systems, 16 bits, Shirley, NY) with a sample rate of 1000 Hz. Procedures

Demographic and anthropometric data were collected in the first session for all participants of each phase (Figure 1A). Familiarization with the experimental setup, a description, and a demonstration of the protocol procedures were pro vided before testing. Familiarization involved an explanation of the angle torque equipment, study protocol, basic con cepts of stretching definition and symptoms, the generic body region that participants would theoretically feel stretch, and how to respond to the questionnaires and scales used in this study. Before testing, participants were also asked to rank items in a list of verbal descriptors in an ascending order to ensure recognition of the words meaning. Several proce dures were taken until the participants reported confidence and understanding about the protocol of the upcoming ses sions. Visual inspection was done in muscle sEMG during the passive leg extension testing to ensure that no artifacts were created by the testing environment. For phase I, one mROM repetition was initially performed followed by 6 sub maximal repetitions in a balanced order. Three of these rep etitions were determined as a percentage of the starting position (SP) to mROM range; and the other 3 repetitions from an OS to mROM range (Figure 1C). The submaximal intensities were calculated at 40, 60, and 80% of both ranges. A different frame of reference (SP vs. OS) was used to test which range would have a higher correlation with the phys iological responses (i.e., ROM and torque). The submaximal stretching repetitions of OS mROM (RO repetition in an OS mROM range; intensities R040, RO60, and RO80) and SP mROM (R repetition in an SP mROM range; intensities R40, R60, and R80) ranges were determined after the initial repetition to mROM. An examiner stopped the dynamometer at the target submaximal ROM and at mROM for z3 seconds before the limb was returned to the SP. Five minutes were given before performing the submaximal repetitions so that the percentage of submaximal intensities could be determined and to ensure the dissipation of stretch ing effects (19,20). To determine the OS mROM range, the participants were instructed to indicate “the moment they feel the first OS” by pressing a trigger that they held during the stretching maneuvers, and to say “OK” when obtaining the “maximum range of motion without pain” (mROM). This procedure was replicated in all stretching repetitions, to confirm OS detection reliability. Two minute rests were given between each submaximal stretching repetition. VOLUME 29 | NUMBER 9 | SEPTEMBER 2015 |

2669

Perception of Stretching Intensity Measurement During this interval time, participants were asked to report stretch symptom body location(s), to classify the stretch intensity (using stretch VAS score and AME), and to indicate the scale anchor from a list of words that best described the intensity. In the third and fourth sessions of phase I, 2 stretching protocols were performed with a balanced order: a rest interval (RI) and a non rest interval (NRI) stretching pro tocol (Figure 1D). It was previously observed that these pro tocols induce distinct stretching supramaximal intensities (i.e., above mROM of first repetition) among repetitions (9). Protocols were administered at the same time of day with a 1 week interval between the 2 sessions. Participants were briefly instructed about the protocol procedures at the start of the second session. The stretching protocols con sisted of the following: (a) 5 stretching repetitions with a 30 second rest interval, performed at mROM without pain (RI) (b) and a maximal number of repetitions without rest intervals between repetitions (NRI). Each stretching repeti tion lasted 90 seconds in a static position for both protocols. The maximal number of NRI repetitions was determined when subjects reported that they could not withstand a fur ther repetition without feeling pain (Figure 1C). For all rep etitions, an examiner stopped the apparatus on the subject’s signaling of mROM. Stretching repetition tests started from the resting position. The VAS scale and AME method were applied at the beginning of the static phase of each repeti tion. At the end of each session, a pain VAS scale was applied to determine the degree of pain felt during the stretching protocol. Using the results of phase I, the SIS (Figure 4) was constructed (please see Conceptual Scale Develop, Validation, and Reliability Section). In the sessions of phase II, the 2 stretching protocols previously used (i.e., NRI and RI) were performed in a balanced order to determine the scale validity and reliability. A systematic approach was used to instruct all participants how to quantify the stretching intensity (Supplemental Digital Content, http://links.lww.com/JSCR/A12). Each repetition of both protocols also lasted 90 seconds in a static position. Before and after each protocol, 4 repetitions were performed to a target stretch intensity. Four repetitions were used to have a minimal number of data points to establish a statistical correlation. The first repetition was performed until the mROM without feeling pain was established, and the remaining 3 repetitions were completed in a balanced order by either producing a stretch intensity or estimating the stretch intensity. Half of the subjects (n = 15) produced the stretch intensities for an SIS score of 80, 60, and 40. The other half estimated the SIS score at 80% (R80), 60% (R60), and 40% (R40) of the mROM that was previously determined in the first repetition. After the stretching, the same pre stretching repetitions were performed in the same order. The dynamometer was held at the target stretching intensity for 3 seconds in the pre and post testing repetitions before the limb was returned to the starting position. The scale was

2670

the

M

Journal of Strength and Conditioning Research

applied immediately after achieving the target stretch intensity of each repetition. To determine intra and inter examiner assessment reliability, half of the participants (n = 15) were assessed by the same examiner in the 2 stretching sessions, and the other half by 2 different examiners. At the end of all sessions, participants performed 3 maximal voluntary isometric muscle contraction (MVIC) repetitions of 5 second duration with a 10 second break for both knee extension and flexion, with the hip and knee flexed at 908, for the purposes of sEMG signal normalization. Conceptual Scale Develop, Validation, and Reliability

The SIS was developed to have 3 components: a visual, a numerical, and a descriptive portion (Figure 4). The visual component was determined based on the VAS score. The numerical component was determined based on the AME score. The descriptive component was determined based on the verbal descriptors chosen from the list by the partici pants during the stretching trials. Scale construction and validation was based on the concept of Borg’s continuum model (3). According to Borg’s model, physiological re sponses should be observed during physical performance. Thus, the joint passive torque and angle were chosen as the stretching physiological outcomes. The study design as well as the content (i.e., words used in the scale), construct (i.e., the concepts of submaximal and supramaximal), and validity were determined before all the stretching test sessions by a stretching specialist professional and an experienced specialist in scale development with more than 10 years of experience in their respected areas of expertise. None of these professionals were involved in the stretching tests. The face validity was confirmed by asking the participants if they think that the SIS would measure the perception of stretching intensity. The criterion validity was determined following the stretching tests by confirming the relation between the SIS score and the VAS and AME scores. Content validity was determined based on previous studies and participants’ understanding (3,8). The words used to qualify the nature of the symptoms during the stretching were used as per Boyd et al. (4), who found in a straight leg raise test that the terms used by individuals were essentially “stretching” and “tension.” Consequently, we used the term “stretching” to refer to symptoms felt during the exercise. The descriptors used to characterize the degree of stretching intensity were chosen based on words from other scales (3,7) translated and adapted to the Portuguese language. The participants’ understanding of the ranking order of the submaximal and supramaximal terms was tested by asking subjects to sort a group of terms in order of increasing intensity. The metric chosen for the scale using the AME method ranged from 0 to values above 100. The number 100 was considered mROM without pain. Such a metric range was chosen because it was previously observed in a pilot test (not published) to have a comparable concordance with physiological responses as relative (i.e.,

the

TM

Journal of Strength and Conditioning Research normalized to maximum values) joint passive torque, angle, and area under the angle torque curve. The values and posi tion of the scale numbers were based on the VAS score and AME phase I results for both submaximal (i.e., below the maximal ROM) and supramaximal intensities. The VAS has been shown to be a reliable instrument to determine supra maximal measures (3). Construct validity was based on the assumption that the perception of stretching intensity varies with tissue deformation by changing the joint angle, which in turn induces a change in joint passive torque. The SIS criterion validity was determined by ensuring 4 conditions: (a) that the SIS score would be related to the physiological measures of joint passive torque and angle (i.e., concurrent validity), for submaximal (in both OS mROM and SP mROM ranges) and supramaximal intensities; (b) that the SIS would be reliable to predict the ROM and torque (i.e., predictive validity) (3,6); (c) that the SIS would be reliable to produce a certain submaximal stretch intensity of passive torque and ROM (i.e., productive validity) (3,6); and (d) that the acute effects induced by stretching on ROM and passive torque could be detectable using the SIS score. Mechanical and sEMG Data Processing

All data were synchronized and recorded using the BIOPAC MP100 Acquisition System (Santa Barbara, CA, USA), except in the case of the joint angles, which were obtained by a digital camera. A manual trigger was sent to the A/D converter to synchronize the joint angle data. The data were processed by an automatic routine using MATLAB v12.0 software (The Mathworks Inc., Natick, MA, USA). The routine first filtered the force from the sensors using a Butterworth second order, low pass filter (10 Hz). Then, the passive knee torque was calculated and corrected to the gravity data (8). To eliminate the mechan ical torque artifacts during the stretching protocols, raw torque was fitted to mathematical models detailed else where (8). Once the sEMG data were recorded and visually inspected, the raw sEMG signals were digitally filtered (25 490 Hz), fullwave rectified, and smoothed with a low pass filter (Butterworth fourth order, 12 Hz). All these procedures were according to the International Society of Electrophysiology and Kinesiology. The average amplitude of the sEMG signal was measured over a 100 millisecond window during the MVIC knee extension and flexion tests for sEMG normalization. The sEMG values are reported as a percentage (%) of the MVIC. Psychometric Data Processing

Phase I data were used for scale development. An exponen tial model (for submaximal data) and a linear model (for supramaximal data) were used because we previously observed in a pilot study (data not showed) a higher adjustment to the psychometric data. All the submaximal repetitions were first ordered according to the ROM performed. When normal distribution was not found for

| www.nsca.com

the AME and relative VAS (i.e., normalized to R100 value) variables, median values were calculated for each submax imal intensity (25). Then, repetitions with a within AME score difference of less than 15% were excluded to avoid the bias due to a ceiling effect among the AME scores (22), except for R100. The remaining repetitions were adjusted to an exponential function (1):

y ¼ a3e bx ;

(1)

where y is the relative AME score, a and b are mathematical parameters, and x is the VAS score. After modeling for equa tion 1, the values of relative VAS score (i.e., scale number positioning) for the AME percentiles values from a to 100 were determined. After confirming the descriptors’ order initially set by the participants, the most frequent descriptors chosen for each stretch intensity were determined. When a similar descriptor was observed for 2 successive stretch intensities, the average relative VAS score value was deter mined for that descriptor. Descriptors that did not follow the initial order set by the participants were excluded. Conse quently, the percentile numbers (i.e., AME values) and stretch intensity anchors (i.e., descriptors) were positioned based on their respective relative VAS score values in a 100 mm vertical line. In respect to the supramaximal data, all repetitions of the RI protocol and the number of maximal NRI repetitions performed by each participant were used for data analysis. When normal distribution was not found for AME and VAS variables, median values were determined for each repetition in both stretching protocols (25). The most frequent descrip tor was determined for each stretching intensity. Then, all repetitions of both protocols were ordered according to the maximal ROM performed. Repetitions with a within AME difference of less than 15% (22) and with a different order than the initial order set by the participants were excluded. The average relative VAS score value was determined when a similar descriptor was observed for 2 successive stretch intensities. Data were then fitted to a linear mathematical function (2):

y ¼ mx þ b;

(2)

where y is the relative AME score, m and b are mathematical parameters, and x is the VAS score. Based on equation 2 model fitting results, the VAS score values (i.e., number posi tioning) were determined for every 10 AME score points from the 100 AME value to the maximal AME observed (i.e., 150). Consequently, the AME numbers (i.e., median AME values) and stretch intensity anchors (i.e., descriptors) were positioned based on their respective relative VAS score values in a vertical line above the submaximal scale line. Statistical Analyses

All data were analyzed using IBM SPSS Statistics 19.0 (IBM Corp., New York, NY, USA). Normal distribution was VOLUME 29 | NUMBER 9 | SEPTEMBER 2015 |

2671

Perception of Stretching Intensity Measurement confirmed using Shapiro Wilk test. Correlations between physiological (i.e., torque and ROM) and SIS score for both submaximal and supramaximal intensities were determined using Pearson coefficient (r) and intraclass correlation coefficient (ICC2,1) at a 95% confident interval. When nor mality was not observed, the Spearman’s rank correlation

coefficient (rho) was used. Reliability was determined ICC2,1 and the calculation of the SEM (27). The ICCs were classified as follows: “little” (0.00 0.25), “low” (0.26 0.49), “moderate” (0.50 0.69), “high” (0.70 0.89), and “very high” (0.90 1.00) (15,22,27). Systematic differences between reli ability variables were examined using the paired t tests.

Figure 2. Correlations between relative (normalized to maximal range of motion values) values of (A) VAS peak torque, (C) VAS maximal range of motion (mROM), (B) AME peak torque, (D) AME mROM, and (E) VAS AME for values obtained in the submaximal stretching repetitions. In the graphs, (A D) scatters are plotted for both SP mROM (j) and OS mROM (3) ranges. The VAS AME data were fitted to an exponential model and the VAS score was estimated for the percentiles of AME. E) The most common descriptors for different intensities are shown. VAS = visual analog scale; SP = starting position; OS = onset of stretching symptoms.

2672

the

M

Journal of Strength and Conditioning Research

the

TM

Journal of Strength and Conditioning Research Intra and inter examiner assessment reliability, OS detection reliability, and scale properties (i.e., production and estima tion) reliability were determined using r and ICC. The pre post stretch effects of both RI and NRI protocols on mROM, passive peak torque (i.e., maximum passive torque obtained at the mROM of each repetition) and passive torque at a given

| www.nsca.com

angle for scale estimation assessment, and SIS score for scale production assessment were determined using paired t tests. The effect size for the changes induced by stretching was determined for the difference using Cohen’s d effect sizes. SEM was determined for reliability and pre and post outcomes. Statistical significance was set at p # 0.05.

Figure 3. Correlations between relative (normalized to maximal range of motion values) values of (A) VAS peak torque, (C) VAS maximal range of motion (mROM), (B) AME peak torque, (D) AME mROM, (E) and VAS AME for supramaximal values obtained in rest interval and non rest interval stretching protocols. The VAS AME data were fitted to a linear model, and the VAS score was estimated for every 10 AME points above 100 until 150. E) The most common descriptors for different intensities are shown. VAS = visual analog scale.

VOLUME 29 | NUMBER 9 | SEPTEMBER 2015 |

2673

Perception of Stretching Intensity Measurement

Figure 4. Stretch intensity scale, composed by 2 intensity dimensions (submaximal and supramaximal). The number 100 represents the maximal range of motion without pain. The font used for the lettering and numbering was Tiresias Signfont Regular.

RESULTS Phase I—Scale Construction

The EMG activity was below 4% of MVIC in all repetitions of all sessions. All participants ordered all submaximal and supramaximal descriptors in an ascending manner. The body regions reported to be stretched in session 2 of phase I were the posterior thigh (93.6%), posterior leg (7.6%), and the anterior thigh (3.4%). The average values of pain VAS score in supramaximal stretching protocols of phase I were below 20 mm (NRI = 9.7 6 15.0 mm and RI = 4.7 6 7.7 mm, p , 0.001), thus indicating that no pain was felt during their stretching. The detection of OS was found to have a very high reliability (r = 0.96, ICC = 0.93 [0.87 0.97]). A normal distribution was not found for the AME score, but it was for relative VAS; thus, the median (for AME) and mean (for VAS) values were used for all submaximal repetitions. A very high correlation was obtained between the physical (i.e., ROM) and physiological (i.e., knee passive torque) outcomes and the perceptual variables (i.e., relative VAS mean score and for AME median score) for submaximal intensities of both SP mROM and OS mROM ranges (Figures 2A D).

2674

the

M

Journal of Strength and Conditioning Research

Because of the 15% AME within difference criterion, the RO80, RO60, and R60 repetitions were excluded, and re maining repetitions were fitted to equation 1 (Figure 2E). The equation 1 parameters obtained were a = 14.829 and b = 0.0189 (r2 = 0.98). Consequently, the relative VAS score values (i.e., numbers positioning) were determined for the percentile numbers ranging from 20 to 100 (Figure 2E). In respect to the submaximal descriptors, the same term was observed for R80 and RO80 repetitions (i.e., “almost maximal”). Thus, the relative VAS score average value was determined for that descriptor (i.e., descriptor positioning). The descriptor “much” was also excluded, because it did not follow the initial order set by the participants. The most common descriptors for each stretching intensity can be observed in Figure 2E. The resultant submaximal scale component that was constructed can be observed in Figure 4. For supramaximal data, normal distribution was not observed for the relative VAS score or AME score, thus, median values were used. Participants performed different numbers of maximal NRI repetitions (2R, n = 3; 3R, n = 12; 4R, n = 10; 5R, n = 5). No differences were observed between protocols in the first repetition in maximal ROM or peak torque (NRI = 49.2 6 15.0 vs. RI = 46.3 6 13.8, p = 0.20). A very high correlation was obtained between the physical (i.e., ROM) and physiological (i.e., knee passive torque) outcomes and the perceptual variables (i.e., relative VAS mean score and AME score) and torque in submaxi mal intensities (Figures 3A D). The relative VAS score and AME score fitted to equation 2 produced parameter values of m = 0.7667 and b = 25.751 (Figure 3E). The estimated relative VAS score for every 10 AME score points from 100 to 150 (i.e., maximal median AME value observed) is shown in Figure 3E. The fifth repetition of NRI protocol was excluded from the supramaximal scale data analysis, because only 2 participants performed 2NRI repetitions. After order ing all repetitions according to the maximal ROM, and using the 15% within difference criterion, the repetitions R3 and R5 of RI protocol and R4 of the NRI protocol were considered for the descriptors’ positioning. Consequently, the supramax imal scale component was designed (Figure 4). Phase II—Scale Validation

The scale constructed in phase I and used in phase II for measuring the perception of stretching intensity can be observed in Figure 4. Participants performed different numbers of maximal NRI repetitions (2R, n = 3; 3R, n = 12; 4R, n = 10; 5R, n = 5). No differences were observed between protocols in the first repetition in maximal ROM (NRI = 43.2 6 9.28 vs. RI = 45.0 6 9.08, p = 0.09) or peak torque (NRI = 30.9 6 12.2 Nm vs. RI = 32.8 6 13.1 Nm, p = 0.21). A typical example of 1 participant’s SIS, torque, and angle response during 2 stretching protocols can be observed in Figure 5.

the

TM

Journal of Strength and Conditioning Research

| www.nsca.com

Figure 5. An example for 1 participant’s maximal range of motion (mROM), peak torque, and SIS score (A) in rest interval and (B) non rest interval stretching protocols. Peak torque and ROM are normalized to the value of the first repetition. This participant performed 4 non rest interval stretching repetitions.

A moderate to high correlation was found between SIS score and torque (r = 0.88, ICC = 0.85 [0.81 0.88] and r = 0.62, ICC = 0.57 [0.47 0.65] for submaximal and supra maximal intensities, respectively) and between SIS score and ROM (r = 0.85, ICC = 0.85 [0.81 0.88] and r = 0.76, ICC = 0.70 [0.63 0.76] for submaximal and supramaximal intensities, respectively). The reliability for both intra and inter examiner assessment, and for the scale production and estimation properties at different submaximal angles is shown in Table 1. For all angles, intraexaminer (r = 0.88,

ICC = 0.88 [0.79 0.92]) and interexaminer (r = 0.93, ICC = 0.93 [0.88 0.96]) assessment showed high to very high reliability results. Scale estimation property showed an inferior reliability result (r = 0.70, ICC = 0.70 [0.50 0.83], for all angles) as compared with the scale produc tion property (r = 0.89, ICC = 0.89 [0.82 0.93], for all angles). The pre and post measurements of both production (torque and ROM) and estimation (SIS score) methods are shown in Table 2.

TABLE 1. Reliability outcomes for intra- and inter-examiner assessment, and for SIS properties of estimation and production for repetitions at 40% (R40), 60% (R60), 80% (R80), and 100% (R100) of maximal tolerable torque.* Intensity Intraexaminer

Interexaminer

Estimation

Production

R40 R60 R80 R100 R40 R60 R80 R100 R40 R60 R80 R100 R40 R60 R80 R100

r

ICC (95% CI)

p

SEM

0.69 0.92 0.90 0.99 0.89 0.95 0.93 0.98 0.06 0.47 0.41

0.63 0.92 0.89 0.99 0.88 0.95 0.93 0.98 0.06 0.47 0.36

(0.19 to 0.86) (0.78 to 0.98) (0.71 to 0.96) (0.97 to 0.99) (0.68 to 0.96) (0.86 to 0.98) (0.80 to 0.98) (0.95 to 0.99) ( 0.47 to 0.56) ( 0.06 to 0.79) ( 0.19 to 0.74)

0.10 0.17 0.52 0.72 0.11 0.32 0.37 0.45 0.49 0.46 0.77

8.41 5.48 6.74 2.81 6.59 5.41 7.53 3.91 11.06 6.76 9.42

0.81 0.77 0.82 0.75

0.81 0.78 0.81 0.74

(0.52 (0.45 (0.52 (0.38

0.35 0.58 0.25 0.69

4.03 4.38 4.27 4.75

to to to to

0.93) 0.92) 0.93) 0.90)

*SIS = stretching intensity scale; ICC = intraclass correlation coefficient; CI = confidence interval.

VOLUME 29 | NUMBER 9 | SEPTEMBER 2015 |

2675

Perception of Stretching Intensity Measurement

TABLE 2. Values (mean 6 SD) of ROM and torque when using the estimation scale method, and SIS score when using the production scale method, before and after the stretching intervention.* Pre Production

ROM (degrees)

Torque (N$m)

Estimation

SIS

R40 R60 R80 All angles R40 R60 R80 All angles R40 R60 R80 All angles

15.8 21.5 28.5 22.0 9.9 15.4 22.8 16.0 47.2 67.0 77.2 63.5

6 6 6 6 6 6 6 6 6 6 6 6

10.2 11.7 14.2 13.1 5.8 9.3 13.7 11.3 11.6 9.1 11.6 16.5

Post

p

6 6 6 6 6 6 6 6 6 6 6 6

0.001 0.003 0.004 ,0.001 0.01 0.03 0.01 ,0.001 0.19 0.04 0.77 0.03

18.2 24.3 31.5 24.7 11.3 17.0 25.6 18.0 44.0 62.7 77.7 60.6

11.4 11.9 14.0 13.5 6.6 8.9 13.6 11.6 16.6 16.0 18.5 21.5

d 0.23 0.23 0.21 0.20 0.24 0.18 0.21 0.17 0.03 0.33 0.21 0.16

SEM 2.41 2.83 3.30 3.95 2.06 2.75 3.15 3.74 8.29 9.46 6.72 6.95

*ROM = range of motion; SIS = stretching intensity scale.

DISCUSSION In this study, a new instrument to assess the perception of stretching intensity was developed and was shown to be valid and reliable during a slow passive leg extension maneuver. The results found in this study supports Borg’s (3) findings. Borg’s theory assumes that the ratio scales of perceived exertion were based on the assumption that the exertion perceived during physical exercise is stimulated by several biological systems, and consequently that diverse physiological parameters correlate with perceived exertion. The SIS score was accompanied by changes in joint ROM and passive torque in both submaximal and supramaximal stretching intensities with a surprisingly high correlation when compared with the validity results from other studies (3). Passive joint torque is mainly changed by tissue defor mation (either tension or compression), and joint angle af fects tissue length. Changes in both variables affect afferent drive to the central nervous system, thereby mediating the pain pathways (12). In this study, for the purpose of concur rently validating the SIS, the joint passive torque was chosen instead of the muscle tendon tension measurement. This was chosen because it was conceptualized that stretching perception would be generated as a result of the net forces caused by the deformation of the tissues crossing the joint, and not a single tissue; however, it remains unknown whether a higher correlation could be obtained to a direct measure of a specific tissue’s passive tension. The passive tension of tissues could be assessed using supersonic shear wave elastography in vivo during a stretching maneuver (17). Thus, future studies should extend this validation by corre lating the SIS score with the passive tension of tissues. The flexibility reference used in this study to classify the perception of stretching intensity was the maximal ROM without pain. Thus, at the beginning of the stretching test

2676

the

M

Journal of Strength and Conditioning Research

sessions, the joint angle corresponding to the maximal ROM without pain was determined, and the participants were asked to memorize such performance when classifying further stretching repetitions. This criterion is often used in clinical and sports contexts (1,4,26). Based on this reference, participants scored the stretching intensity perception on SIS, for both submaximal and supramaximal intensities. It is important to note that all stretching repetitions classified as supramaximal were performed without pain. (i.e., stretches performed above the initial mROM). This occurred due to acute sensory adaptations that allowed a greater stretch tolerance (2,16). However, to get higher SIS assess ment validity, we strongly suggest that the participants should be instructed to report the SIS score in respect to the initial mROM. This procedure ensures the reliability of the scale, as was demonstrated in this study for both intra and inter examiner assessment. In addition, to ensure that mROM is well understood and memorized by the partici pants, we also suggest that when determining the mROM, the examiners should ask the participants in an understand able way whether they believe that mROM was achieved; if needed, the participants should repeat the trial. This is an important procedure, because the SIS scoring will depend on the accuracy of this reference. However, other criteria to determine mROM have been used by other researchers (18). In addition, it should be noted that since mROM was performed to a point free of pain, the perceived exertion might be more in a discomfort than a pain domain. Future studies should examine the validity, reliability, and relevance of different criteria in determining the joint maximal range of motion. The validity of SIS was also confirmed by the scale properties to either produce or to estimate a certain stretching intensity (3,6). We found a high reliability in SIS

the

TM

Journal of Strength and Conditioning Research when producing a stretching intensity (ICC = 0.75 0.82), but a low to moderate reliability when estimating the different submaximal intensities tested (ICC = 0.06 0.47). However, when reliability was determined for all the intensities tested together (i.e., R40, R60, and R80), both SIS production (ICC = 0.89) and estimation (ICC = 0.70) methods showed high reliability. This suggests that SIS is better for producing a stretching intensity than estimating the SIS score for certain stretching intensities. Because no previous studies have as sessed stretching intensity using an estimation method, we are unable to compare these results. Although the results obtained in estimating submaximal intensities were low, we assume that SIS estimation properties’ reliability might improve with a longer familiarization (3). Thus, we suggest that future studies explore this property of the scale, since there are few studies comparing the estimation and produc tion methods. A specific characteristic of SIS is the supramaximal component (Supplemental Digital Content, http:// links.lww.com/JSCR/A12), which allows the detection of stretching intensities above the maximal ROM with a higher sensitivity. The term supramaximal (i.e., supra = more) was chosen and used in this study because it suggests that the stretching is being performed above the maximal range of motion; and, considering that the ROM increases among stretching repetitions in respect to the initial mROM, the stretch is considered to be in a supramaximal domain. There are many methodological procedures that facilitate the increase of stretch intensity above the initial mROM ob tained in the first repetition (23). However, when performing stretching training, it is difficult to assess the magnitude of ROM and torque changes using only the perceived exertion feedback from the participants. The scale presented in this study allowed to determine these changes above the initial mROM (i.e., in a supramaximal domain). The participants performed 2 different stretching protocols that led to 2 dis tinct levels of stretching intensity; whereas the RI protocol produced a smaller increase in torque and ROM among repetitions than the NRI stretching method. Although dif ferent changes of ROM and torque were produced by the 2 protocols, the SIS proved capable of discriminating and fol lowing up on these changes (Figure 4). Thus, a high corre lation between the SIS score and the physiological changes was observed. This means that SIS can be used to detect changes in ROM and peak torque above the maximal ROM when performing stretch training. The SIS was also shown to be capable of detecting acute angle torque changes induced by stretching. It is known that stretching induces an acute decrease in passive joint torque and an increase in mROM (11,16). It was seen that using a produc tion intensity method, the ROM and the torque performed for the SIS scores of 80, 60, and 40 were higher after the stretching protocol. Using the estimation method, the SIS score was also higher at angles of R80, R60, and R40 after stretching. This reinforces the validity of SIS in assessing stretching intensity.

| www.nsca.com

The present SIS is expected to be useful in future studies for various different reasons, such as observing acute and chronic adaptations induced by different types of stimuli or to compare different types of populations. However, it must be considered that the SIS was only validated and tested for reliability in 1 human joint (i.e., knee), for a slow stretching maneuver, in a specific population (i.e., men with low knee extension flexibility). In addition, it must be considered that the SIS was validated using a constant angle protocol; meaning that application to other stretching protocols (e.g., constant torque) may need further validation. The SIS should be tested under other stretching conditions (e.g., different joints) and extended to other populations. In addition, the SIS should be also compared (i.e., concurrent validation) to other existing scales (e.g., Borg CR 100).

PRACTICAL APPLICATIONS A new and valid instrument with high reliability was developed to measure the perception of stretching intensity. Since stretch ing with different intensities induces distinctive physiological and performance effects, this instrument allows individuals to assess the stretching intensity below and above the initial maximal range of motion. Accordingly, this scale is thought to be relevant to assessing stretch intensity in flexibility training interventions, since the SIS score relates to the relative changes of joint ROM and passive torque. We expect that this instrument will be useful for field interventions and research settings.

ACKNOWLEDGMENTS The authors acknowledge the support of the Portuguese Scientific Foundation. All the authors declare that they have no conflicts of interest regarding this article.

REFERENCES 1. Behm, DG and Kibele, A. Effects of differing intensities of static stretching on jump performance. Eur J Appl Physiol 101: 587 594, 2007. 2. Bjorklund, M, Hamberg, J, and Crenshaw, AG. Sensory adaptation after a 2 week stretching regimen of the rectus femoris muscle. Arch Phys Med Rehabil 82: 1245 1250, 2001. 3. Borg, G. Borg’s Perceived Exertion and Pain Scales (1st ed.). Champaign, IL: Human Kinetics, 1998. 4. Boyd, BS, Wanek, L, Gray, AT, and Topp, KS. Mechanosensitivity of the lower extremity nervous system during straight leg raise neurodynamic testing in healthy individuals. J Orthop Sports Phys Ther 39: 780 790, 2009. 5. Branco, V, Negra˜o Filho, R, Padovani, C, Azevedo, F, Alves, N, and Carvalho, AR. Relationship between tension applied and sensations of discomfort in the hamstring muscle during stretching. Braz J Phys Ther 10: 465 472, 2006. 6. Coquart, B, Garcin, M, Gaynor, P, Tourny Chollet, C, and Eston, R. Prediction of maximal or peak oxygen uptake from ratings of perceived exertion. Sports Med 44: 563 578, 2014. 7. Ferreira Valente, M, Pais Ribeiro, J, and Jensen, M. Validity of four pain intensity rating scales. Pain 152: 2399 2404, 2011. 8. Freitas, S, Vaz, J, Bruno, P, Valamatos, M, and Mil Homens, P. Comparison of different knee extension passive torque angle assessments. Physiol Meas 34: 1483 1498, 2013. VOLUME 29 | NUMBER 9 | SEPTEMBER 2015 |

2677

Perception of Stretching Intensity Measurement 9. Freitas, SR, Vaz, JR, Bruno, PM, Valamatos, MJ, Andrade, RJ, and Mil Homens, P. Are rest intervals between stretching repetitions effective to acutely increase range of motion? Int J Sports Physiol Perform 10: 191 197, 2015. 10. Hermens, H, Freriks, B, Disselhorst Klug, C, and Rau, G. Development of recommendations for SEMG sensors and sensor placement procedures. J Electromyogr Kinesiol 10: 361 374, 2000. 11. Hoge, K, Ryan, E, Costa, P, Herda, T, Walter, A, Stout, J, and Cramer, J. Gender differences in musculotendinous stiffness and range of motion after an acute bout of stretching. J Strength Cond Res 24: 2618 2626, 2010. 12. Holdcroft, A and Jaggar, S. Core Topics in Pain. New York, NY: Cambridge University Press, 2005. 13. Kubo, K, Kanehisa, H, and Fukunaga, T. Effect of stretching training on the viscoelastic properties of human tendon structures in vivo. J Appl Physiol (1985) 92: 595 601, 2002. 14. Kubo, K, Kanehisa, H, and Fukunaga, T. Gender differences in the viscoelastic properties of tendon structures. Eur J Appl Physiol 88: 520 526, 2003. 15. Kwah, LK, Pinto, RZ, Diong, J, and Herbert, RD. Reliability and validity of ultrasound measurements of muscle fascicle length and pennation in humans: A systematic review. J Appl Physiol (1985) 114: 761 769, 2013. 16. Magnusson, S, Simonsen, E, Aagaard, P, and Kjaer, M. Biomechanical responses to repeated stretches in human hamstring muscle in vivo. Am J Sports Med 24: 622 628, 1996. 17. Maı¨setti, O, Hug, F, Bouillard, K, and Nordez, A. Characterization of passive elastic properties of the human medial gastrocnemius

2678

the

M

Journal of Strength and Conditioning Research

muscle belly using supersonic shear imaging. J Biomech 45: 978 984, 2012. 18. McHugh, M and Cosgrave, C. To stretch or not to stretch: The role of stretching in injury prevention and performance. Scand J Med Sci Sports 20: 169 181, 2010. 19. Mizuno, T, Matsumoto, M, and Umemura, Y. Decrements in stiffness are restored within 10 min. Int J Sports Med 34: 484 490, 2013. 20. Nordez, A, McNair, PJ, Casari, P, and Cornu, C. Static and cyclic stretching: Their different effects on the passive torque angle curve. J Sci Med Sport 13: 156 160, 2010. 21. Page, P. Current concepts in muscle stretching for exercise and rehabilitation. Int J Sports Phys Ther 7: 109 119, 2012. 22. Portney, L and Walkins, M. Foundations of Clinical Research: Applications to Practice (3rd ed.). Upper Saddle River, NJ: Prentice Hall, 2009. 23. Sharman, M, Cresswell, A, and Riek, S. Proprioceptive neuromuscular facilitation stretching: Mechanisms and clinical implications. Sports Med 36: 929 939, 2006. 24. Stevens, SS. On the psychophysical law. Psychol Rev 64: 153 181, 1957. 25. Szklo, M and Neto, J. Epidemiology: Beyond the Basics (3rd ed.). Mississauga, ON: Jones & Bartlett Learning, 2012. 26. Walter, J, Figoni, SF, Andres, FF, and Brown, E. Training intensity and duration in flexibility. Clin Kinesiol 50: 40 45, 1996. 27. Weir, JP. Quantifying test retest reliability using the intraclass correlation coefficient and the SEM. J Strength Cond Res 219: 231 240, 2005.