Stretch Exercises Increase Tolerance to Stretch in Patients With Chronic Musculoskeletal Pain: A Randomized Controlled Trial

Design Overview and Randomization

A randomized controlled trial using a within-subject design was undertaken. One leg of each participant was randomly allocated to an experimental condition, and the other leg was randomly allocated to a control condition. In this way, each participant acted as his or her own control, reducing variability and increasing statistical power. To ensure concealed allocation, a blocked random allocation schedule with an equal number of right and left legs was generated by computer and placed in a series of consecutively numbered, opaque, sealed envelopes by a person external to the study. The envelopes were kept off-site and opened after each participant's initial assessment, indicating his or her inclusion into the trial.

Setting and Participants

Thirty adult participants were recruited from patients enrolled in a multidisciplinary pain management program based at a hospital in Sydney, Australia. To be eligible for inclusion, patients had to have pain of musculoskeletal origin persisting for at least 3 months, be likely to participate in a hamstring muscle stretch regimen as part of the pain management program, and be over 18 years of age. They were excluded if they were unable to tolerate the testing procedure; had excessive hamstring muscle extensibility (able to place palms on the floor in a standard toe-touch test)³⁰; required further medical, surgical, or psychiatric investigations or interventions; or had a history of drug or alcohol abuse. A treatment effect of 5 degrees of hip flexion was determined a priori to be clinically important, based on the recommendations of previous researchers.^{19,20,22,23,31} We estimated that 30 participants would provide a 95% probability of detecting a between-group difference of 5 degrees. This power calculation was based on a predicted standard deviation of 5 degrees,²⁰ a dropout rate of 15%, and an alpha level of .05.

Intervention

The stretch intervention was supervised by trained physical therapists and embedded within a multidisciplinary pain management program (ADAPT). This program is a well-established method of chronic pain management using a multidisciplinary team approach.³² Cognitive-behavioral principles form the basis of the program, incorporating the following components: exercise and stretch, pacing, education, drug reduction, relaxation, sleep management, relapse prevention, and family involvement.³³ The program is conducted over 3 weeks, with at least 3 hours of physical rehabilitation daily aimed at improving muscle extensibility, fitness, strength, and posture. Physical therapists supervise the exercise sessions and use cognitive-behavioral strategies throughout to provide education on behavioral modifications and other issues.³⁴

Stretch Treatment

The hamstring muscles of the experimental leg were stretched for 1 minute per day for 3 weeks over 18 consecutive days, and the hamstring muscles of the control leg received no specific stretch treatment during this period. Stretches were self-administered. Participants sat on the ground, reaching forward with both hands over the treatment leg while maintaining full knee extension (Fig. 1).

Figure 1.

The stretch treatment administered to the hamstring muscles. Participants sat on the ground, reaching forward with both hands over the treatment leg while maintaining full knee extension. Image used with permission from www.physiotherapyexercises.com.

Participants were instructed to refrain from stretching the hamstring muscles of the control leg during the course of the study. Fourteen of the 18 stretch sessions were performed in the pain management program during weekdays and supervised by physical therapists. The other 4 stretch sessions were unsupervised and performed independently by the participants on weekends. Adherence to the intervention was carefully monitored, and all stretch sessions were recorded on an exercise sheet. This adherence was reviewed at each subsequent supervised stretch session.

Outcome Measures

The 2 primary outcome measures were hamstring muscle extensibility and stretch tolerance. These measures were assessed on both legs of each participant before and after the 3-week stretch intervention.

Measurement Device

A device specifically designed to measure passive hip flexion and hip flexor torque was used.³⁵ Its key feature is that it enables the application of known stretch torques after counteracting the torque due to the weight of the leg. The device consists of a wheel connected to the side of a physical therapy bed (Fig. 2). A leg splint was attached to the wheel and both rotated simultaneously. The leg splint ensured full knee extension and restricted any hip abduction or rotation. Adjustable counterweights attached to a long rod were used to counteract the torque produced by the weight of the leg and splint. The long rod was connected to the wheel apparatus and extended proximally from the splint toward the head of the participant. Weights were hung tangentially from the rim of the wheel. The weights generated a hip flexor torque that rotated the leg splint and leg together. The torque was a product of the mass of the weights and the radius of the wheel (28 cm). The device has good reliability (intraclass correlation coefficient=.97, 95% confidence interval [CI]=.96 to .98).³⁵

Figure 2.

The testing device. The wheel, leg splint, and rod rotated simultaneously. Image used with permission from www.physiotherapyexercises.com.

Testing Procedure

Participants attended a trial session at least 3 days prior to the commencement of the study to allow for familiarization with the testing procedure. Initial measurements were taken on the first day of the 3-week program, prior to the first stretch. Final measurements were taken at least 24 hours and no more than 48 hours after the last stretch to ensure that the results reflected long-lasting changes in extensibility rather than mere transient effects of stretch. This distinction is important, because we were not interested in the immediate effects (due to viscous deformation) known to quickly dissipate upon removal of the stretch.^10–14 By taking measurements more than 24 hours after the last stretch, we could be sure that our results reflected something more than just viscous deformation.⁸ One blinded assessor was used for all testing, and participants were instructed not to discuss any aspect of the intervention with the assessor. The success of blinding was confirmed by asking the assessor to guess the treatment allocation for each participant.

All testing was performed in the same format, with measurements of the right leg taken prior to measurements of the left leg. Participants were positioned supine on a physical therapy bed with the measured leg strapped to a splint. The contralateral leg and pelvis were stabilized with straps. Participants were positioned with their hip joint aligned to the center of the wheel. Passive hip flexion angle was measured with a digital inclinometer aligned on the long axis of the leg splint (Fig. 2). All verbal instructions and explanations were standardized across participants.

Assessment of Muscle Extensibility

The extensibility of the hamstring muscles was reflected by the angle of passive hip flexion with the application of a standardized torque. This standardized torque was the same for both legs of each participant, but not across participants. This procedure was appropriate because the effects of the stretch intervention were being compared within, rather than between, participants. The standardized torque corresponded to the highest common torque tolerated on both legs at both preintervention and postintervention assessments.

Prior to measurement, a stretch torque of 18 N·m was applied for 2 minutes. This 2-minute prestretch exhausted most viscous deformation, reducing the effect of “creep” on subsequent measurements and diminishing any reflex muscle activity around the hip and knee joints.^{11,14,36–38}

Assessment of Stretch Tolerance

Stretch tolerance was reflected by the angle of passive hip flexion with the application of a nonstandardized torque, that is, one that corresponded to the highest stretch torque participants were willing to tolerate. This torque varied between legs, testing sessions, and participants. A series of gradually increasing stretch torques, applied at increments of 6.1 N·m every 30 seconds, was used to determine stretch tolerance. Hip flexion angle at each increment was measured with a digital inclinometer.

Participants were asked to indicate when the stretch felt “very uncomfortable.” At this point, the torque was slightly reduced. Subsequently, the torque was further increased again but in smaller increments of 1.53 N·m at a faster rate. This continued until the participant indicated a second time that the stretch was “very uncomfortable.” At this point, final hip flexion was measured.

Pain intensity also was used to reflect stretch tolerance. Participants were asked to rate their pain on an 11-point numerical rating scale (with 0 being “no pain” and 10 being the “worst pain you can imagine”). Numerical rating scales are sensitive to changes in pain and have high validity and reliability.^39–41 Participants were blindfolded throughout testing to minimize the influence of visual feedback on their tolerance to stretch.

Data Analysis

Changes in hip flexion angles between initial and final measurements with and without a standardized torque were calculated for both the experimental and control legs. Mean between-group differences and their corresponding 95% CIs were then calculated. The t-distribution was used to estimate the 95% CIs for between-group differences in hip angle (posttest score minus pretest score). Paired t tests were used to test for significant differences. A positive change in hip flexion angle with a standardized torque reflected an increase in hamstring muscle extensibility following the stretch intervention. Positive changes in hip flexion angle and pain rating score at the highest-tolerated torque reflected an increase in stretch tolerance. In addition, paired-samples t tests were performed on data collected from 4 pain questionnaires^42–45 routinely used in the ADAPT pain management programs. A P value of <.05 was considered significant. Data were analyzed by intention-to-treat.⁴⁶

Participant Characteristics

All 30 participants completed the study, with no dropouts or withdrawals (Fig. 3). The demographic characteristics of the participants are shown in Table 1. The site of pain was diverse, affecting the neck, back, arm, and leg, with more than half (60%) of the participants reporting pain in 2 or more major sites. Mean initial values on the Depression Anxiety and Stress Scales,⁴² modified Roland-Morris Disability Questionnaire,⁴³ Multidimensional Pain Inventory,⁴⁵ and Pain Self-Efficacy Questionnaire⁴⁴ were consistent with recently developed normative data in people with chronic pain,⁴⁷ indicating that our participants were representative of this population.

Figure 3.

Flow of participants through the trial. Primary outcomes measured on the first and final days of the 3-week program.

Adherence to Protocol

The protocol required participants to perform 18 stretch sessions over 18 consecutive days, with 14 sessions supervised and 4 sessions unsupervised. Minor deviations from the protocol occurred, with 12 stretches missed out of a total of 540 (2%) for all participants.

Effects of Intervention

At the commencement of the trial, there was no difference in the extensibility of the hamstring muscles between the experimental and control legs. The initial mean (SD) hip flexion angles with the application of an 18-N·m torque for the experimental and control legs were 47 (12) and 47 (13) degrees, respectively. Similarly, there was no difference in participants’ tolerance to the discomfort associated with stretch between their experimental and control legs. The initial mean (SD) hip flexion angles at the highest-tolerated torque were 69 (18) and 69 (19) degrees, respectively (Tab. 2).

Primary Analyses

Effects of Intervention

This trial investigated the effects of stretch on muscle extensibility and stretch tolerance in patients with chronic musculoskeletal pain. The results indicated no real change in hamstring muscle extensibility, despite the apparent increase in hip flexion. This apparent change was due to participants’ improved willingness to tolerate the discomfort associated with stretch, rather than any underlying change in the passive mechanical properties of the hamstring muscles.

The participants of this study were all adults aged between 19 and 68 years with chronic musculoskeletal pain originating from a variety of different sources. The most common location of pain was the back (77% of the sample group) followed by the legs (53% of the sample group; Tab. 1). This study examined the effects of stretch to the hamstring muscles regardless of the site of pain. It is possible that participants with leg pain responded differently to hamstring muscle stretches than participants with arm or neck pain. There were insufficient participants to explore this possibility in post hoc subgroup analyses, although the narrow 95% CI associated with the between-group difference suggests that all participants responded to the stretch intervention in a consistent way.

Important design features were used in this study to minimize bias, including randomization and concealed allocation. Although participants were not blinded to the intervention, their legs were carefully screened from view during testing to block visual cues of leg position. Nevertheless, it is possible that the increased tolerance to stretch reflects the use of unblinded participants, with strong expectations about the therapeutic benefits of stretch. That is, participants anticipated that the stretch intervention would improve muscle extensibility and, therefore, inadvertently tolerated larger torques during the final testing of their experimental legs, leading to an apparent increase in ROM.

There is consistent evidence from studies of animals to indicate that sustained stretch changes the passive mechanical properties of muscles, leading to an increase in real extensibility.^{16,17,48–50} This increased extensibility is attributed to the highly adaptable nature of soft tissues. Evidence from studies of humans is less consistent,^8,51 due in part to the poor definition of extensibility²⁶ and the failure of some investigators to measure joint ROM with a standardized torque.^28,52–54 Without accompanying measures of torque, it is impossible to distinguish between increases in ROM from changes in the passive mechanical properties of soft tissues (indicating real extensibility) and increases in ROM from changes in stretch tolerance (indicating apparent extensibility).

Numerous stretch techniques have been developed, applied, and used by physical therapists, coaches, and trainers.⁵⁵ These stretch techniques include isometric, ballistic, and dynamic ROM techniques; proprioceptive neuromuscular facilitation; and passive and static stretches.⁵⁶ A static stretch is a slow, sustained muscle lengthening that can easily be self-administered and is commonly used across clinical and community settings. Static stretches are usually held for durations of 15 to 60 seconds. We chose a 1-minute static stretch for our intervention protocol, as this duration and type of stretch are typical and representative of current clinical practice in pain management programs.

Muscle Extensibility

Our results demonstrated that stretch does not increase hamstring muscle extensibility. This finding contradicts anecdotal evidence and the results of some clinical trials indicating the effectiveness of stretch for increasing extensibility in individuals without disabilities.^{36,52,54,56–59} Our findings, however, are consistent with those of 2 similar studies of our own in individuals without disabilities.^20,23 They also are comparable to the findings of multiple clinical trials involving people with neurological disabilities.^19,60,61 Interestingly, these trials examined stretch interventions that were administered for much longer than a few minutes a day, yet still failed to demonstrate any real changes in muscle extensibility, despite good statistical power. It is possible that the muscles of people with neurological disabilities respond differently to stretch than the muscles of individuals without neurological disabilities. Nevertheless, these trials increasingly support the view that stretch, as typically administered in the clinical setting, does not induce lasting changes in extensibility. No other study to date has investigated the role of stretch in people with chronic pain.

A possible explanation for the failure to increase real muscle extensibility is that 1 minute of stretch per day is subtherapeutic. Perhaps if each stretch had been administered for a longer duration (that is, more than 1 minute a day), the results may have been different. Similarly, it is possible that 3 weeks of stretch is insufficient and if the intervention had been performed over a longer period (that is, more than 3 weeks), different outcomes may have been attained. Nonetheless, the dosage of stretch used in this trial is similar to that typically applied in clinical practice.⁹ It is unlikely that patients with chronic pain would tolerate more-intensive stretch regimens administered for longer than a few minutes per muscle. In addition, recent evidence in individuals without disabilities indicate no added benefit from stretch administered for up to 20 to 30 minutes a day.^20,23

It is possible that the stretch intervention might have increased extensibility if only participants with limited hamstring muscle extensibility had been included. However, it is difficult to define limited hamstring muscle extensibility, because joint ROM is a direct function of applied torque. If the applied torque were ignored, most clinicians would argue that a mean maximum of 69 degrees hip flexion, as attained by our participants on entry to the trial, is indicative of limited hamstring muscle extensibility. However, our participants tolerated lower levels of torque than normal. For example, on average, our participants tolerated 36 N·m of stretch torque at the commencement of the trial, whereas individuals without disabilities tolerate considerably more (ie, 55 N·m).²⁰ It is difficult, therefore, to make direct comparisons.

We were aware of the problems associated with defining limited extensibility when designing the research protocol and decided to adopt a pragmatic approach to overcome the problems. That is, we mimicked clinical practice and included all participants who would have received hamstring muscle stretches as part of their pain management program unless they had excessive extensibility (ie, could place the palm of their hands on the floor). The results of this trial do not rule out the possibility that the response to stretch may be a function of initial extensibility.

The failure to detect a treatment effect on muscle extensibility was not due to an insufficient sample size. This finding is evidenced by the relationship between the 95% CIs and the predetermined minimally worthwhile treatment effect.

Stretch Tolerance

This study demonstrated that a stretch program can increase a person's tolerance to stretch. Similarly, Halbertsma and Goeken²⁵ and Magnusson et al²⁶ also reported altered stretch tolerance following stretch interventions. Both groups of authors investigated the hamstring muscles with relatively intensive regimens of 10 minutes of stretch twice a day for 4 weeks²⁵ and 225 seconds of stretch twice a day for 20 days.²⁶ Bjorklund et al²⁷ similarly showed sensory adaptations in the rectus femoris muscle after a 2-week stretch protocol. Interestingly, although Bjorklund et al²⁷ used a relatively mild stretch regimen (80 seconds of stretch, 4 times a week for 2 weeks) compared with the 2 previous studies,^25,26 the intervention was still sufficient to alter stretch tolerance. Recent evidence from studies by Folpp et al²⁰ and Ben and Harvey²³ add further support for this proposition. These investigators reported increases in stretch tolerance but not extensibility following fairly intensive stretch protocols (20 minutes a day, 5 days a week for 4 weeks,²⁰ and 30 minutes a day, 5 days a week for 6 weeks²³). Our study suggests that even a shorter and less-intensive regimen (1 minute daily over 3 weeks) can increase stretch tolerance and improve apparent muscle extensibility. These findings have important implications for clinical practice if the aim of stretch is to achieve a greater joint ROM regardless of the underlying mechanism.

The explanation for changes in stretch tolerance is not known. Stretch tolerance may be influenced by nociceptive nerve endings, mechanoreceptors, or proprioceptors.^26,62 Alternatively, stretch may change some other aspect of the sensory neural pathways.^26,63 For example, afferent input from muscles and joints during a stretch maneuver may interfere with signals from nociceptive fibers (stretch discomfort), subsequently inhibiting an individual's perception of pain. This explanation is consistent with the gate control theory of pain.⁶⁴

Alternatively, changes in stretch tolerance may be psychologically mediated. It is possible that participants anticipated the positive effects of stretch and, therefore, their perception of the stretch discomfort was dampened. According to the gate control theory,⁶⁴ sensations of pain and discomfort are affected by descending modulatory influences from higher centers. Prior experiences of stretch, motivation to stretch (possibly from supervision), and elevated mood and confidence from positive expectations of stretch benefits are all potential psychological contributors explaining the participants’ altered perception of the discomfort and willingness to tolerate greater stretch over time.

It is interesting to note that although the participants tolerated a larger testing torque on their experimental leg following 3 weeks of stretch, there were no changes in their preintervention and postintervention pain intensity scores at the highest-tolerated torque. Similarly, there were no changes in preintervention and postintervention pain scores on the Multidimensional Pain Inventory,⁴⁵ with a mean (SD) change of 0 (1; P=.16). These results suggest that it was not the pain itself that changed but rather the participants’ willingness to tolerate it. Perhaps repeated exposure to stretch produced a desensitizing effect, accustoming the individual to the sensation. The participants progressively tolerated larger stretch torques for the same level of experienced pain. This finding is consistent with the concepts of pain acceptance⁶⁵ and self-efficacy,⁴⁴ both commonly referred to when discussing coping strategies in the management of chronic pain. Interestingly, there was a significant improvement in mean (SD) Pain Self-Efficacy Questionnaire⁴⁴ ratings during the ADAPT program from 24 (12) to 40 (10) points (P<.001). As participants’ acceptance of the stretch discomfort progressively improved over the 3 weeks, they began to feel more confident in their ability to perform the stretch.

Participants’ increased willingness to tolerate the discomfort associated with stretch was found to be specific to the experimental leg, and there was no crossover to the control leg. That is, administering stretch on one leg did not produce any corresponding effect on the opposite leg. This was an interesting finding and merits further investigation. Furthermore, the effects of stretch on the experimental leg occurred on top of general improvements in the participants’ overall depression, anxiety, and physical disability. For example, the mean (SD) changes between the initial and final scores on the Depression and Anxiety Stress Scales⁴² and the modified Roland-Morris Disability Questionnaire⁴³ were 5 (6) (P=.002) and 3 (4) points (P=.002), respectively. The overall improvements in mood and physical disability presumably were influenced by other components of the ADAPT program. These findings together suggest that the stretch intervention itself did not provide a generalized benefit. Rather, the effects were specific to the leg that was stretched. This finding is consistent with reports on fear-avoidance interventions in patients with chronic pain, showing that exposure to one particular movement produced specific improvements in that movement only and did not generalize toward a different movement.⁶⁶

Clinical Implications

This study provides support for the ongoing incorporation of stretch into pain management programs, provided the aim of stretch is to improve joint ROM via increased tolerance to the discomfort associated with stretch, rather than via changing the passive mechanical properties of tissue. Stretch, therefore, may be conceptualized as a graded exposure to movement, increasing tolerance to stretch and ROM. Stretch exercises are particularly relevant in the rehabilitation of patients with chronic musculoskeletal pain, who may have physical deconditioning due to fear of movement or reinjury.

Directions for Future Research

Further research is needed to investigate the mechanisms underlying changes in tolerance and perception of discomfort associated with stretch. This study raised questions about the possible contributions of neurophysiological and psychological factors accounting for the observed increase in stretch tolerance, and further investigations are needed to address this issue more closely.

Future studies also should be directed at ascertaining the relative merits of targeting stretch to specific muscle groups and joints in patients with pain originating from different parts of the body. It is possible that patients with shoulder pain will respond better to specific shoulder stretches, as opposed to a generic hamstring muscle stretch. Similarly, patients with low back pain may respond significantly better to hamstring muscle stretches than patients complaining of neck pain. It is important, therefore, to prescribe specific stretches to target specific muscles and joints. The effectiveness of specific stretch interventions to target specific areas needs to be established by further research.