Nonpharmacological Treatments for Musculoskeletal Pain

Abstract

Background: 

Several types of physical therapy are used in the management of painful musculoskeletal disorders. These treatment modalities can be broadly categorized as electrotherapy modalities (e.g., transcutaneous electrical nerve stimulation), acupuncture, thermal modalities (e.g., moist heat, ultrasound), manual therapies (e.g., manipulation or massage), or exercise. Within each of these broad categories significant variations in treatment parameters are possible.

Objective: 

To consider the evidence base for each of these main categories of physical therapy in the management of musculoskeletal pain.

Method: 

To consider the available evidence related to clinical effectiveness and then to review evidence from basic science studies evaluating potentially therapeutic effects of the various therapies.

Results: 

There seems to be evidence from basic science research to suggest that many of the therapies could have potentially therapeutic effects. However, there appears to be limited high-quality evidence from randomized clinical trials to support the therapeutic effectiveness of several of the therapies.

Conclusions: 

There is some preliminary evidence to support the use of manual therapies, exercise, and acupuncture in the management of some categories of musculoskeletal pain. Limitations of the existing research base are discussed and recommendations for areas of future research are provided.

A number of physical therapies are used to manage pain of musculoskeletal origin and to address sensorimotor dysfunction in patients with musculoskeletal disorders. These can be broadly categorized as electrotherapy modalities (e.g., transcutaneous electrical nerve stimulation [TENS]), acupuncture, thermal modalities (e.g., moist heat, ultrasound), manual therapies (e.g., manipulation or massage), or exercise. In practice, physical therapists use a combination of modalities to address the treatment needs of each patient based on the findings of clinical examination. In the past, some therapies have been particularly espoused in different countries or by different groups of therapists. Increasingly, however, physical therapists are adopting an evidence-based approach to patient management and are therefore selecting those therapeutic modalities for which there is scientific evidence indicating effectiveness. Each of the aforementioned broadly outlined therapeutic categories has been subject to scientific investigation in recent decades. However, there are still many specific therapies or specific dosage regimes that have not been subjected to scientific investigation. Given the variety of physical therapies that exist, the variety of ways in which they can be applied, and the lack of a major funding source such as the pharmaceutical industry, it may be several decades before all physical therapies have been thoroughly investigated. The aim of this review is to provide a current perspective on research related to the mechanism of action and therapeutic effectiveness of each of the broad therapeutic categories listed above, specifically related to the management of musculoskeletal pain.

OVERVIEW OF MODALITIES

Electrotherapy modalities

Electrical stimulation can be used in a variety of ways to achieve therapeutic outcomes. In general, it can be used to produce sensory effects or to facilitate motor function. By varying parameters such as frequency, waveform, pulse duration, electrode configuration, and duration of stimulation, it is possible to produce a range of therapeutic effects. Transcutaneous electrical nerve stimulation and interferential therapy are two forms of electrical stimulation most commonly used for pain modulation. Transcutaneous electrical nerve stimulation is usually used as a self-administered therapy in which the patient is given initial instruction on safe and appropriate use of the TENS device and then self-administers the treatment according to a predefined schedule. Conversely, interferential therapy usually is used in a physical therapy clinic setting.

Transcutaneous electrical nerve stimulation

Transcutaneous electrical nerve stimulation is noninvasive, inexpensive, safe, and easy to use. Clinically, TENS is administered at varying frequencies of stimulation. These frequencies can be broadly classified as high-frequency (>50 Hz), low-frequency (<10 Hz), or burst TENS (bursts of high-frequency stimulation administered at a much lower frequency). Intensity is determined by the response of the patient as being sensory level or motor level TENS. For sensory level TENS, the voltage (i.e., amplitude) is increased until the patient feels a comfortable tingling (perceived with high frequency) or tapping (perceived with low frequency) sensation without motor contraction. This amplitude is referred to as low intensity. For motor level TENS, the intensity is increased to produce a motor contraction. Usually the intensity is increased to the maximal tolerable level but is not noxious. This is referred to as high-intensity TENS. High-frequency TENS is administered at low intensities and is referred to as conventional TENS. In contrast, low-frequency TENS usually is administered at high intensities so that a motor contraction is produced. This mode of stimulation is referred to as strong, low-rate, or acupuncturelike TENS. Stimulus strength duration curves for administration of TENS to the skin show that sensory level TENS occurs with the lowest amplitude, followed by motor contraction and then noxious sensation (see review^1,2).

Transcutaneous electrical nerve stimulation is used for pain relief in a variety of patient populations, including postoperative pain, acute and chronic pain, musculoskeletal pain, and neuropathic pain conditions.^3-6 With regard to treatment of musculoskeletal pain, clinical studies show mixed results, and few studies compare the effects to a placebo treatment. Many studies do not standardize the stimulus parameters of TENS, which makes interpretation difficult between groups and studies; therefore, this review will highlight a few of the better-quality studies. Deyo et al.⁷ compared the effects of TENS with those of placebo TENS with and without exercise in patients with chronic low back pain. The effects of TENS were the same as those of placebo TENS. However, the stimulation parameters were not standardized for TENS, and other modalities, such as hot packs, were used. In contrast, Marchand et al.⁸ showed that high-frequency TENS treatment significantly reduced the affective component of pain. The effects of TENS were cumulative across treatments.⁸ However, when treatments were terminated, no long-term effects were observed (1 month or 6 months) when compared with placebo TENS. The patients in this study did not receive other therapies. Thoresteinsson et al.⁹ showed short- and long-term relief of pain after treatment using TENS when compared with placebo TENS for patients with chronic low back pain. However, several studies showed no significant difference between placebo TENS and TENS for pain relief in patients with acute low back pain,¹⁰ chronic low back pain,¹¹ temporomandibular joint disorders,¹² or osteoarthritis.¹³ Therefore, the current data are conflicting. However, it is difficult to draw conclusions regarding effectiveness because different pain syndromes, different stimulation parameters, and different outcome measures were used in each study.

Transcutaneous electrical nerve stimulation alone may not provide complete inhibition of hyperalgesia and pain and thus will probably not be the only method used clinically for pain relief. However, as an adjunct to other pain-relief methods, TENS may have several benefits. Knowledge of the mechanism of TENS enables clinicians to determine better which patients will benefit from TENS treatment based on the type of medication the patient is using for pain control. Effective pain relief using TENS in combination with other therapies will allow the patient to increase activity level and return to work more quickly. Treatment with TENS increases joint function in patients with arthritis^4-6,14 and improves the physical and mental component summary on the Short Form-36 quality of life survey in patients with chronic low back pain.¹⁵ Improving physical function allows the patient to tolerate other therapies and activities, which results in an improved quality of life.

Animal models of inflammation show that TENS partially reduces hyperalgesia at the site of injury¹⁶ and completely reduces hyperalgesia outside the site of injury.^17,18 This reduction in primary hyperalgesia (at the site of injury) is dependent on frequency of stimulation, such that high-frequency but not low-frequency stimulation reduces the hyperalgesia.¹⁶ However, high- and low-frequency TENS are equally effective in reducing secondary hyperalgesia.^17,18

Two different theories have been proposed for the mechanism of pain relief using TENS. The most popular theory for the mechanism of action of TENS is the gate control theory of pain.^6,19-21 This theory suggests that stimulation of large-diameter afferent fibers inhibits second-order neurons in the dorsal horn and prevents pain impulses carried by small-diameter fibers from reaching higher brain centers. It is most commonly used to explain the relief of pain using high-frequency TENS. Alternatively, release of endogenous opioids is thought to underlie the actions of low-frequency TENS.^22-25 Increasing evidence also supports a role of opioid receptors in pain relief using high-frequency TENS. Selective blockade of μ-opioid receptors in the spinal cord prevents the antihyperalgesia produced by low-frequency TENS, whereas blockade of δ-opioid receptors prevents the antihyperalgesia produced by high-frequency TENS.²⁶ This opioid-mediated inhibition could be segmental or supraspinal. High- and low-frequency TENS both reduce dorsal horn neuron responses to noxious stimuli in normal^27,28 and arthritic animals,²⁹ which supports a role for spinally mediated inhibition. However, low- or high-frequency TENS antihyperalgesia is prevented using blockade of μ-opioid or δ-opioid receptors in the rostral ventral medulla, a component of the opioid-mediated descending inhibitory pathways.

Interferential therapy

Interferential therapy involves a different form of electrical stimulation than TENS, although the physiologic and therapeutic effects of interferential therapy potentially are similar to those of TENS. It is based on the principle that when two medium-frequency (KHz) currents are applied to the skin, a low-frequency current will be induced in the deep tissues that is equivalent to the difference in frequency between the two medium-frequency currents. Therefore, interferential therapy allows effective stimulation of deep tissues, whereas TENS is predominantly a cutaneous or superficial stimulus. In the British Isles, interferential therapy is used more frequently than TENS for the management of low back pain,³⁰ and it is also a popular treatment modality in Canada.³¹

Treatment usually is administered using two pairs of electrodes, and most treatment units allow variation in waveform, stimulus frequency and stimulus amplitude, or intensity. The interference frequency will usually range from 2 to 200 Hz (similar to TENS), and treatments may involve low-frequency or high-frequency stimulation. Many treatments also involve a sweep frequency in which the frequency is systematically changed over a predetermined range (e.g., 50-100 Hz). Intensity will normally be adjusted to produce either a strong but comfortable sensation or a motor response. Similarly to TENS, high-frequency stimulation is normally adjusted to produce a sensory response and low-frequency stimulation will often be adjusted to produce a motor response.

Interferential therapy is thought to have similar mechanisms of action to TENS and is considered to act through segmental inhibition or through activation of descending pain-inhibitory systems. It is also thought to have a positive influence on blood flow, which may contribute to improved tissue healing.³²

The evidence base to determine the clinical effectiveness of interferential therapy is less adequate than that for TENS. One study did not show additional benefit for interferential therapy and exercise beyond that produced by exercise therapy alone in the management of patients with musculoskeletal pain in the shoulder region at 12-month follow-up.³³ A study comparing interferential therapy with traction in the management of patients with low back pain did not show any difference between the outcomes for the two treatments.³⁴ Unfortunately, no control group was included in the study and so it is impossible to determine whether the treatments were equally effective or equally ineffective.

Although it is likely that interferential therapy may trigger physiologic mechanisms similar to those of TENS, and it has the advantage of potentially providing more effective stimulation of deep musculoskeletal structures, the evidence base to allow us to properly determine the effectiveness of this treatment modality is inadequate. Those studies that have been reported do not suggest significant therapeutic benefit. There is a clear need for more high-quality trials investigating this form of treatment.

There are many other forms of electrical stimulation (e.g., diadynamic current, H-wave therapy) but there is little evidence to suggest that any of these treatments are effective, or that there is likely to be any difference in therapeutic outcome between different forms of electrical stimulation.

In summary, there is compelling evidence from basic science studies to suggest that electrical stimulation has important physiologic effects on pain modulatory systems and modulation of blood flow. Although these findings suggest potential therapeutic benefit for electrical stimulation, this has yet to translate into data from randomized, controlled clinical trials that provide convincing evidence of either short-term or long-term therapeutic effectiveness. In part, this is because of the lack of research studies and because of the inadequacies and limitations of many of the studies that have been completed. In addition, there is no evidence to suggest that any one form of electrical stimulation is superior to any other form of electrical stimulation, although differences in frequency appear to underlie differences in the physiologic effects of different treatment regimes. Much more research is required to provide evidence for or against the therapeutic effects of electrotherapy modalities. There is insufficient evidence to allow us to make an evidence-based determination of therapeutic effectiveness or lack of effectiveness in the management of musculoskeletal pain.

Acupuncture

Acupuncture is administered via insertion of needles into designated acupuncture points, and subsequently the needles are manually manipulated or electrical current is administered via the needles. When electrical current is administered via the needles, it is similar to TENS in regard to parameters and mechanisms, and it is more effective than manual acupuncture (see review³⁵). A number of clinical trials have been performed over the years that show the effectiveness of acupuncture in the treatment of musculoskeletal conditions.^36-47

Manual acupuncture produces a short-term improvement in pain and function in patients with osteoarthritis of the knee when compared with a no-treatment control group³⁹ but not with a placebo acupuncture group.⁴⁸ Similar results were observed for osteoarthritis of the cervical spine⁴⁷ and temporomandibular joint disorders.³⁸ However, in a group of patients with myofacial neck pain, acupuncture administered to sites relevant to the cervical spine produced significantly increased pain relief when compared with irrelevant sites outside the cervical spine or with no treatment.⁴⁴ Kleinhenz et al.⁴² showed in a single-blind clinical trial that manual acupuncture significantly improved function scores in patients with rotator cuff tendonitis. Electroacupuncture shows similar results, with improvements in pain and function compared with placebo acupuncture in patients with fibromyalgia⁴⁰ or no treatment in patients with temporomandibular disorders.⁴¹ Long-term effects were observed in patients with chronic low back pain when treated using low-frequency electroacupuncture but not when using manual or high-frequency electroacupuncture.⁴⁶ However, when compared with physical therapy that consisted of massage, cryotherapy, TENS, and relaxation exercises, electroacupuncture was similar to physical therapy in reducing pain. In fact, physical therapy also reduced opioid intake.³⁷

Acupuncture is thought to activate endogenous opioid pathways in human subjects and animals.^49-54 Ulett et al. have systematically examined the effects of electroacupuncture and found similar mechanisms to TENS (see review⁴⁹). Specifically, blockade of opioid receptors with naloxone in the habenula, nucleus accumbens, amygdala, or periaqueductal gray matter prevents the analgesia produced using electroacupuncture (4-15 Hz).⁵⁵ Further, microinjection of antisera against β-endorphin into the PAG reduced the analgesia produced using low-frequency electroacupuncture (4-15 Hz).⁵⁶ Increased concentrations of methionine enkephalin and dynorphin A were observed in the lumbar cerebrospinal fluid after treatment of patients using low- or high-frequency electroacupuncture, respectively.⁵⁷ Further, rats made tolerant to a μ-opioid agonist were also tolerant to 2 Hz but not to 100 Hz electroacupuncture, which suggests that low-frequency stimulation activates μ-opioid receptors.⁵⁸ However, rats tolerant to a δ-opioid receptor agonist were also tolerant to 2 Hz but not to 100 Hz electroacupuncture,⁵⁸ whereas rats tolerant to a κ-opioid agonist were tolerant to 100 Hz electroacupuncture.⁵⁸ Supraspinal (intracerebroventricular administration) blockade of μ- or δ-, but not κ-opioid, receptors significantly prevented the analgesia produced using low-frequency (2 and 30 Hz) but not high-frequency (100 Hz) electroacupuncture.⁵⁹ Conversely, supraspinal blockade of κ-opioid receptors prevented the analgesia produced using high-frequency but not low-frequency electroacupuncture.⁵⁹ These data suggest that analgesia from high-frequency electroacupuncture is not mediated through μ- or δ-opioid receptors but through κ-opioid receptors, and that analgesia from low-frequency electroacupuncture is mediated by μ- or δ-opioid receptors. Further, descending inhibitory pathways involving the amygdala and PAG, opioid peptides and opioid receptors are involved in electroacupuncture analgesia.

In summary, manual acupuncture, low-frequency electroacupuncture, and high-frequency electroacupuncture may produce analgesia via different mechanisms. In particular, low-frequency electroacupuncture seems to involve descending systems and activation of μ- or δ-opioid receptors. Conversely, high-frequency electroacupuncture is more dependent on κ-opioid receptor activation. There is some encouraging evidence from clinical trials to support the use of acupuncture in the management of musculoskeletal pain; however, it is difficult to provide a satisfactory placebo control for acupuncture and, therefore, the quality of some research studies is less than satisfactory. Although there is good evidence to support the analgesic effect of acupuncture in basic science studies, further research is required to investigate the clinical outcomes of acupuncture therapy in the management of musculoskeletal pain.

Thermal modalities

A variety of methods produce heating and cooling of the tissues to manage musculoskeletal pain in acute musculoskeletal injuries and in chronic musculoskeletal disorders. The effects and effectiveness of superficial heating, deep tissue heating, ultrasound therapy, and cryotherapy are discussed.

Superficial heat

The use of heat to relieve pain of musculoskeletal origin is common. Heat can be applied superficially by application of moist hot packs, immersion in hot water baths, use of infrared light, or paraffin wax application. The most common method utilized for treating musculoskeletal conditions is moist hot packs. Moist hot packs are applied over the area of pain for 20-30 minutes and only heat the superficial tissues.

Many of the local physiologic effects of heat have been studied thoroughly. For instance, heat increases skin and joint temperature and blood flow, and decreases joint stiffness.^60,61 Activity of local cartilage-degrading enzymes is influenced by joint temperature. As found in patients with rheumatoid arthritis, when temperature increases from the normal of 33 °C to 36 °C these enzymes are considerably more active.^62,63 Temperatures above 41 °C decrease the activity of these enzymes.^62,63 However, the data on superficial heat treatment are not consistent with respect to decreasing pain or increasing function (see review^61,64). Some studies show decreases, some show increases, and some show no change in arthritic pain or associated symptoms^65-68 (see review^64,69).

The use of superficial heat is common and has been studied by several different groups with little support for its use in the treatment of painful conditions. Superficial heat increased the pressure pain threshold over approximately 50% of trigger points immediately after treatment in patients with myofascial pain who had at least one trigger point. However, the patient population was not specified.⁷⁰ Increases in pressure pain thresholds were also noted in patients with temporomandibular joint dysfunction.⁷¹ Hoyrup and Kjorvel⁷² compared either whirl-pool or paraffin wax baths alone or with exercise, and showed significant pain relief in all groups with no differences between groups. Toomey et al.⁷³ studied patients with Colles' fractures and compared warm whirl-pool baths with exercise to exercise alone and found no difference between groups, although both groups had increased range of motion and decreased pain. No difference between groups was observed when superficial heat was compared to cold treatment.^67,74 Thus, there is little evidence to support the use of superficial heat in the treatment of pain. However, few types of painful conditions have been studied, the outcome measures have been limited, and the quality of trials could be improved.

We can only speculate on potential mechanisms of relief of pain by heating modalities. Muscle spasms or guarding can result in a pain response or contribute to the pain experienced. Muscle spasms can cause local ischemia activating nociceptive afferent fibers. Reducing muscle spasms and guarding would then be expected to reduce pain by reducing ischemia and preventing activation of nociceptors. Elevating muscle temperature to about 42 °C decreases the firing frequency of Type II muscle spindle afferent fibers and increases the firing frequency of Type Ib Golgi tendon organ afferent fibers.⁷⁵ However, Type Ia muscle spindle afferent fiber firing frequencies are also increased in response to elevation of muscle temperature.⁷⁵ Type II muscle spindle afferent fibers are tonically active and respond to muscle length. Type Ia muscle spindle afferent fibers respond dynamically and respond to velocity of change in muscle length. When activated, the Type Ia and Type II muscle spinal afferent fibers cause an excitation of the agonist muscle and inhibition of the antagonist muscle.

Golgi tendon organs respond to muscle stretch and when activated inhibit the agonist muscle and excite the antagonist muscle.⁷⁶ Thus, increasing muscle tissue temperature could reduce muscle spasm by decreasing activity of Type II muscle spindle afferents and increasing activity of Type Ib Golgi tendon organ afferents. This may be a viable explanation for the use of deep heating modalities. Alternatively, heat could decrease pain indirectly. Increasing skin temperature or deep tissue temperature would cause vasodilation of the tissue, increased metabolism, and increased blood flow, all of which assist with healing and repair. Improving healing and repair would result in increased removal of inflammatory compounds, known to activate and sensitize primary afferent fibers. This would result in less input being transmitted to the spinal cord and higher brain centers, and thus decreased perception of pain.

Deep tissue heating

Several methods exist to produce heat in the deeper tissues and these treatments are often used in the management of painful musculoskeletal disorders (e.g., osteoarthritis of the hip). One commonly used method is short-wave diathermy, which involves the application of high-frequency oscillating current (27 MHz). The rapidly alternating electromagnetic field induces a rapid to and fro motion of ions, generating heat within the tissues. Short-wave diathermy may be administered in either a continuous or a pulsed mode and a variety of different electrodes can be used to heat specific regions of the body.

Such deep heating methods are likely to induce similar physiologic effects to more superficial heating but they have the advantage of specifically heating deep musculoskeletal tissues that may be the source of musculoskeletal pain. They may influence muscle spindle sensitivity in line with the mechanism outlined above.

A few studies have investigated the therapeutic effectiveness of short-wave diathermy in the management of painful musculoskeletal disorders. Foley-Nolan et al. showed short-term beneficial effects of pulsed short-wave diathermy in the management of patients with acute whiplash injury⁷⁷ and individuals with chronic neck pain.⁷⁸ These effects were not sustained with long-term (12-week) follow-up.

Ultrasound

Therapeutic ultrasound consists of inaudible acoustic vibrations delivered at a frequency between 0.75 and 3.0 MHz and intensity between 0.5 and 3 W/cm². The lower the ultrasound frequency, the deeper the penetration of sound waves. Ultrasound as a therapeutic modality for pain relief is applied through the skin overlying the painful area (see review⁷⁹). Ultrasound preferentially heats deeper tissues and the effects are not normally perceived by the patient, making placebo controlled studies relatively easy. Heating occurs predominantly at tissue interfaces.

Several studies have examined the effects of ultrasound versus either no ultrasound or sham ultrasound for a variety of musculoskeletal conditions (see review⁸⁰). Downing and Weinstein,⁸¹ in a double-blind randomized controlled trial in patients with shoulder pain, showed no difference in pain relief, range of motion, or functional activities between sham ultrasound and ultrasound. In contrast, two nonrandomized studies with blind evaluators showed differences between ultrasound and either no ultrasound^82,83 or sham ultrasound⁸³ for patients with shoulder pain or low back pain, respectively. When compared to other noninvasive physical modalities such as ice,⁸⁴, TENS,^85,86 superficial heat,⁸⁵ or massage,⁸⁷ there was no difference between groups, although significant pain relief occurred with any of these treatments. For the treatment of trigger points, a program of exercise and massage with ultrasound or sham ultrasound showed significant differences between groups in pain relief or analgesic intake, although both groups demonstrated a significant improvement over controls who did not receive treatment.⁸⁸ Thus, in contrast to the use of superficial heat, there is more evidence to support the use of ultrasound for pain relief. However, all heating modalities show similar improvements in reducing pain.

Cryotherapy

The use of cold for the treatment of inflammatory pain has been recognized for hundreds of years. Cold is applied in a variety of ways to reduce pain and swelling, including ice bags, ice baths, cold packs, cold baths, and ethyl chloride spray to the skin.

Cold treatment decreases skin and joint temperature, decreases blood flow, and increases joint stiffness.^60,61 In addition, it is quite clear that cold is analgesic.^66,69,89-95 Topical application of cold decreases skin, muscle, and intraarticular temperature.^60,90,96

Measuring local pain thresholds after treatment with ice gives varying results, with the effectiveness lasting from 30 minutes to 12 hours.^{67,89,90,92,93} Cold also slows the conduction velocity of peripheral nerves.^97,98 This being the case, decreased nociceptive information transmitted through primary afferents centrally to the spinal cord would result in a decrease in behavioral signs, and a decrease in neuronal activity of dorsal horn neurons. Because an ice bath decreases secondary hyperalgesia,⁹⁹ it is inferred that there would be a decrease in activity of central neurons and a reduction of the expanded receptive fields. Alternatively, Williams et al.⁶⁷ suggest that application of cold to an arthritic joint serves as a counterirritant by bombarding central pain pathways with painful cold impulses and activating descending inhibitory pathways. Support for this is based on the observation that cooling the skin alone, as with ethyl chloride spray, increases pain threshold.¹⁰⁰

Several studies demonstrate immediate or short-term effects of ice on pain in patients with rheumatoid arthritis^74,92 and low back pain.⁹³ Williams et al.⁶⁷ compared heat packs in combination with exercise to ice packs in combination with exercise in patients with rheumatoid arthritis of the shoulder. Significant reductions in pain occurred for both groups but there was no difference between groups. Melzack et al.⁹³ compared the effects of ice massage for low back pain to TENS and found a significant reduction in pain with both techniques, with TENS showing longer lasting relief (23 hours) compared to ice (12 hours). Curkovic et al.¹⁰¹ demonstrated an analgesic effect from both heat and cold treatment in patients with rheumatoid arthritis, without differences between groups. Thus, short-term effects of ice are clearly analgesic and there may be some long-term benefits. However, cold treatment does not appear superior in effect to other noninvasive treatments. It may be that application of cold, like heat, is useful as an adjunct therapy to allow a patient to exercise with reduced pain.

Manual therapy

Manual therapy techniques include a vast array of treatment procedures intended to promote motion and relieve pain in musculoskeletal structures. The most common forms of treatment are joint manipulation, joint mobilization, and various forms of massage, although techniques specifically intended to mobilize nerve tissue¹⁰² and muscle¹⁰³ are also commonly used. There are also a large number of approaches that have been developed by particular practitioners or groups of practitioners. Research to date has focused on joint mobilization, joint manipulation, and massage. Many other manual therapy approaches have yet to be subjected to scientific scrutiny.

Joint manipulation and mobilization

Joint manipulation is one specific form of joint mobilization. Mobilization techniques are procedures designed to increase the range of movement in a joint. They involve specific positioning of a joint and then oscillatory movement of that joint, either in the mid range or at the limit of the available range of movement. The movement may involve either an accessory glide of the joint or a physiologic movement of the joint. Determination of dosage involves either changing the position in the range of movement or modifying the duration of the mobilization. Typically the duration of treatment will vary from 30 seconds to several minutes. Joint manipulation involves similar positioning of the joint but the movement performed is a low amplitude thrust beyond the normal range of movement.

Manipulation of spinal joints has been the subject of a substantial number of randomized clinical trials. To date, there have been fewer randomized trials specifically investigating joint mobilization. There is also a noticeable lack of studies addressing the use of manipulation or mobilization techniques to treat peripheral joints.

A sufficient number of randomized clinical trials for spinal manipulation and mobilization have been reported to allow researchers to conduct structured reviews or meta-analyses to summarize the findings.^104-111 An early meta-analysis suggested beneficial effects of spinal manipulation and mobilization on measures of pain, flexibility, and physical activity, although the effects were considered to be short term.¹⁰⁸ Subsequent structured reviews evaluating the effect of manipulation on back pain have failed to find convincing evidence of the effectiveness of manipulation,^106,107 although it was concluded that spinal manipulation might be beneficial for subgroups of patients with chronic back pain.¹⁰⁷ Other researchers have used meta-analytical techniques to evaluate the outcomes of multiple randomized controlled trials and concluded that spinal manipulation is efficacious in the management of acute low back pain (less than 3 weeks) but that there is insufficient evidence to support or refute its effectiveness in the management of chronic low back pain.¹¹⁰ Anderson et al.¹¹² concluded that spinal manipulative therapy is consistently more effective than any comparison treatment used in the studies analyzed. The findings of different groups are therefore not consistent, even though the groups of randomized clinical trials evaluated were similar. One reason for this is the low quality of some of the published studies and the variability in both the quality and outcomes of the studies.

A major government-appointed group in the United States reviewed the literature and concluded that manipulation produces short-term beneficial effects in the management of acute low back pain.¹¹³ The report states that "manipulation can be helpful for patients with acute low back problems without radiculopathy when used within the first month of symptoms."¹¹³ The authors also suggest that manipulation can be used safely with patients who have had pain for more that 1 month but that there is insufficient evidence to determine whether such treatment is efficacious. Their conclusions were based on studies that included both mobilization and manipulation techniques. In general, it appears that there is evidence to support the use of manipulation to promote pain relief in the management of low back pain. It is clear, however, that the duration of effect is limited.

The use of manipulation and mobilization in the management of neck pain has also been investigated using structured reviews and meta-analysis.^104,105,111 Following a structured review of randomized clinical trials evaluating manipulation and mobilization in the management of neck pain and headache, Hurwitz et al.¹¹¹ concluded that both mobilization and manipulation were probably of short-term benefit in the management of acute and subacute neck pain. They also concluded that manipulation and mobilization are beneficial in the management of muscle tension headache.¹¹¹ The authors suggested that there was a clear need for further better-quality studies and those future studies should address different doses of manual therapy and comparisons between manipulation and mobilization. They also indicated the need for studies to clearly report any complications that arise in the course of treatment. Other researchers, however, have been more cautious in their support of manual therapies for the management of neck pain. Aker et al.¹⁰⁴ concluded that although there is "early evidence of support [for] manual treatments in combination with other treatments, conclusions must be made cautiously because of the small numbers of trials on which they are based." However, when presenting the same analysis in a different context, Gross et al.¹⁰⁵ concluded that "manual therapies have been demonstrated to be effective for mechanical neck pain in the short term when used in combination with other treatments." Part of the reason for the difference in conclusions between studies may be that although there is clearly a positive effect of manual therapies, the average effect size is relatively small (12-16 mm on a 100-mm scale) and there are not sufficient studies to determine whether the effect is different for patients with acute, subacute, or chronic neck pain.

In summary, it appears that both manipulation and mobilization are beneficial in the management of neck pain, low back pain, and muscle tension headache, but much more research is needed to determine the magnitude of that effect and the relative effectiveness for patients with pain of different durations. There is a paucity of research investigating the effects of manual therapy treatments on peripheral joint problems. There is also a lack of studies evaluating manual therapy treatments that are directed toward structures other than joints and there is a clear need for studies directly comparing joint mobilization and joint manipulation.

Multifactorial models have been presented to explain the effect of manual therapy treatments on pain.^105,114,115 It is believed that these procedures may have beneficial effects on local tissues and that they can suppress pain by activating neurophysiologic mechanisms at either spinal or supraspinal levels. Emerging evidence suggests that manual therapy techniques applied to the cervical spine elicit concurrent changes in pain perception, autonomic function, and motor function in patterns that are similar to the patterns of change elicited by direct stimulation of the periaqueductal gray region of the midbrain.^116-118 Interestingly, manual therapy treatments appear to exert a predominant influence on mechanical nociception rather than thermal nociception.¹¹⁸ See Wright¹¹⁶ for a review of research detailing the neurophysiologic effects of mobilization treatments. This emerging evidence therefore supports the concept that manual therapy techniques exert important neurophysiologic effects that may contribute to the ability of these treatments to reduce pain. Further research is required to provide a detailed understanding of these mechanisms. To date, no studies have attempted to determine if there are distinctions among the neurophysiologic effects produced by electrical stimulation, acupuncture, and manual therapy.

Massage

Massage techniques include a range of procedures to move and mobilize soft tissues, particularly skin and underlying muscle tissue.

Several randomized clinical trials have been published evaluating the therapeutic effectiveness of massage. Recently, these studies have been the subject of two meta-analyses.^119,120 In 1999, Ernst reviewed four studies that used massage in the treatment of low back pain.¹¹⁹ The quality of studies reviewed was poor and it was not possible to make any reliable evaluation of effectiveness. One study suggested that massage is superior to no treatment; two suggested that it might be as effective as TENS and spinal manipulation; the other clearly demonstrated that spinal manipulation was superior to massage. A further meta-analysis¹²⁰ reviewed studies that had specifically investigated the effect of massage on delayed onset muscular soreness (DOMS). The studies reviewed exhibited significant methodological flaws and produced conflicting results. However, it was concluded that "massage may be a promising intervention for the reduction of DOMS."

It appears that massage may be of limited benefit, although much more research is required to determine the magnitude and duration of any therapeutic effect. At this stage there is no evidence to suggest that massage produces effects that are superior to those of other physical treatments.

Massage is considered to have a number of beneficial physiologic effects that may contribute to tissue repair, pain modulation, relaxation, and improved mood state (see review¹²¹). Goats¹²¹ suggested that massage has beneficial effects on arterial and venous blood flow and edema. For example, vigorous massage increases local blood flow and cardiac stroke volume¹²²; it has also been shown to improve lymph drainage¹²³ and it appears to have an anticoagulant effect.¹²⁴ Massage is also thought to activate segmental inhibitory mechanisms to suppress pain and it is considered that some techniques such as connective tissue massage may activate descending pain inhibitory systems.¹²¹ It is clear that massage may have a number of beneficial physiologic effects that vary depending on the massage technique used. In common with a number of other physical therapies, the available research does not clearly demonstrate that these effects translate into beneficial clinical outcomes.

Therapeutic exercise

Exercise is used very extensively in the physical therapy management of musculoskeletal disorders. Although there may be different rationales for using exercise, it is increasingly recognized that exercise can have a relatively direct influence on pain perception. There are a variety of different forms of exercise that can be used in the treatment of painful musculoskeletal disorders. These can include moderate aerobic exercise, intense aerobic exercise, strengthening exercises, isometric exercises, mobility exercises, and exercises to promote specific activation and re-education of key muscle groups.

The effects of exercise interventions have been most extensively investigated for three main groups of musculoskeletal disorders: painful spinal disorders, arthritic diseases, and fibromyalgia.

A variety of exercise approaches have been used in the management of patients with back pain. These range from low intensity aerobic exercise programs to high intensity strengthening programs. Findings from a recent meta-analysis of randomized controlled trials¹²⁵ suggest that exercise is not more effective than other treatments such as manual therapy, back school and usual care by a family physician, or inactive treatments of acute back pain (less than 12 weeks' duration). There appears to be little evidence to support the use of either flexion or extension exercises in the management of acute back pain, although the evidence suggests that extension exercises (such as the McKenzie approach) seem to be more effective than flexion exercises in patients with discogenic back pain.

The situation is rather different for chronic back pain (greater than 12 weeks' duration). In that situation exercise interventions appear to be superior to usual care by a family physician, although exercise alone is not superior to a general physical therapy intervention. It is not possible to differentiate between different forms of exercise.

Many studies have adopted a general exercise approach in the management of acute low back pain. Studies such as that by Faas et al.¹²⁶ suggest that this approach has no beneficial effect on pain or pain recurrence. Some physical therapists have adopted a much more specific approach to exercise in the treatment of low back pain.^127-129 This approach is based on specific re-education of motor control and muscle function for those muscles that contribute to stabilization of the lumbar spine (e.g., multifidus and transverses abdominis).¹³⁰ An interesting study that was not included in the meta-analysis conducted by Tulder et al.¹²⁵ suggests that although specific exercise may have a minimal impact on pain in patients with acute low back pain, they may significantly reduce subsequent recurrence of back pain.^128,129 Hides and colleagues adopted a highly specific approach to exercise that involved using ultrasound imaging as a biofeedback technique to encourage specific activation and improved endurance of the segmental lumbar multifidus at the level of injury. Although both the treatment group and the control group showed a significant reduction in pain after 4 weeks, recurrence rates for the treatment group were significantly lower at both 1-year and 3-year follow-up.^128,129 This very specific form of exercise therefore seems to offer a protective effect against subsequent recurrence when used in the period after an acute back injury.

A similar approach to specific re-education of the transversus abdominis muscle has been shown to have a beneficial effect on patients with chronic low back pain due to spondylolysis or spondylolisthesis.¹²⁷ In this case, the exercise group showed significant reduction in pain and functional disability relative to a control group that received normal care under the direction of their family physician. These differences were maintained at 30-month follow-up. A more general exercise program, in the form of Norwegian medical exercise therapy, has been shown to be more effective and more cost effective than a self-exercise, walking program in the management of chronic back pain.¹³¹ An interesting study by Taimela et al.¹³² points to the importance of patients maintaining their exercise program even after the end of any formal exercise intervention. Subjects in this study completed a 12-week supervised exercise program. Those who maintained a significant level of exercise after the program ended had significantly fewer recurrences of pain and less absence from work in the 2-year period after treatment.

Exercise has also been shown to be beneficial in the management of patients with neck and shoulder pain affecting the trapezius muscle. In this case the particular exercise approach seemed to be of limited importance because improvements occurred in three treatment groups whose exercise programs focused on improving strength, endurance, and coordination, and there were no differences among the groups.

Although it appears that general exercise may be of limited value in the treatment of acute back pain, there is preliminary evidence to support the use of a highly specific exercise intervention in the management of acute back pain. General active exercise and more specific activation of muscles contributing to spinal stability appear to be useful in the management of chronic back pain. Potential benefits may include reduced pain, reduced recurrence, and reduced costs. More research is required to determine the optimum approach to exercise therapy for both acute and chronic back pain.

Exercise programs have also been recommended for patients with arthritis. Whereas in the past the emphasis has been on range of movement exercises and exercises to strengthen specific muscle groups, more recently there has been an increased emphasis on moderate aerobic exercise as a means to improve function and self-efficacy in patients with arthritis. A recent structured review¹³³ concluded that moderate aerobic exercise, sufficient to improve aerobic capacity, resulted in improved strength, improved joint mobility, and small improvements in functional ability (e.g., time required to walk 50 feet). The studies evaluated did not suggest any specific improvement in pain or joint inflammation as a result of the interventions. Importantly, however, they did not suggest any increase in pain or inflammation as a result of relatively vigorous exercise. Other studies have suggested that pain may be improved with aerobic exercise programs, however.^134,135 A study by O'Reilly et al.¹³⁵ investigated the effect of a moderate intensity home exercise program in patients with osteoarthritis of the knees. This program resulted in improvements in pain scores and physical function scores after a 6-month exercise program. The control groups received no intervention. Significant improvements in quadriceps strength were also noted. Similar results have been reported for both an aerobic exercise program and a strengthening program.¹³⁶ These programs resulted in improved function and decreased pain relative to a group that received health education information only. There were no significant differences between the two forms of exercise. Interestingly, this study also evaluated patient X-rays as a means of determining any deterioration in joint status. No significant differences were noted for joint X-ray scores, suggesting that the exercise interventions did not hasten deterioration in arthritic joints.

A similar study by Stenstrom¹³⁴ evaluated a home exercise program in the management of patients with rheumatoid arthritis. In this case the program lasted for 12 weeks and was followed up after an additional 12 weeks. Significant improvements in pain, function, Ritchie articular index scores, and joint mobility occurred as a result of the intervention. Relatively short durations of exercise (15 minutes per week) appear to result in significant improvements in pain and function and Ritchie articular index in patients with rheumatoid arthritis.¹³⁷

It would appear that moderately intensive exercise programs are beneficial for patients with rheumatoid arthritis and osteoarthritis. Improvements in physical function are apparent and at a minimum pain ratings are not increased by the exercise. There are data from some studies suggesting that exercise interventions may reduce pain in these patient populations.

Exercise has also been recommended as a component of the multidisciplinary management of patients with fibromyalgia.¹³⁸ A number of studies have shown positive effects on physical fitness and well being.^139,140 It has proven more difficult to demonstrate reductions in pain report as a result of exercise programs. However, there is evidence that well-controlled programs may bring about a reduction in pain report.^141,142 One problem with exercise programs seems to be poor compliance in this patient population with consequent loss of any improvement in the period after completion of a formal exercise program.¹⁴¹ It seems that exercise is beneficial for patients with fibromyalgia but there are difficulties in terms of determining the optimum intensity of exercise and ensuring patient compliance in the long term.

A beneficial effect of moderately intense exercise on pain in clinical populations is not entirely unexpected because there is good evidence from basic science research to suggest that exercise is capable of activating endogenous pain control systems (see review¹⁴³). In humans, the analgesic effects of walking and cycling have been extensively tested. A variety of models have been used to evaluate the effect of exercise on pain perception. One approach has been to evaluate the effect of cycling on dental pain thresholds, determined using electrical stimulation.^144,145 It seems to require moderately intense exercise with workloads in excess of 200 Watts for periods of at least 15 minutes before demonstrable analgesia is apparent in this model. Running (10 km) has been shown to have an analgesic influence on thermal pain and ischemic pain but no influence on cold pressor pain.¹⁴⁶ Running has also been shown to increase pressure pain thresholds and this effect was reversed by the administration of naloxone (10 mg). Similar increases in pressure pain threshold have been demonstrated following moderately intense cycling for 30 minutes.¹⁴⁷ Pressure pain thresholds are also elevated following resistance exercise training, although the duration of the effect is less than 15 minutes.¹⁴⁸

There is still considerable debate as to the mechanism of action of exercise-induced analgesia because studies evaluating naloxone reversal of this effect have produced very variable results.¹⁴³ It appears that exercise may be an adequate stimulus to activate the endogenous opioid system although many analgesic effects appear to involve a mix of opioid and nonopioid mechanisms.

There is a substantial body of literature demonstrating analgesia following swimming in water of various temperatures in rats and mice. Analgesia has been demonstrated after 3.5-minute swims at 2 °C,¹⁴⁹ 3-minute swims at 20 °C,¹⁵⁰ and 3-minute swims at 32 °C.¹⁵¹ On this basis, it would appear that swimming is the main stimulus producing analgesia; however, when cold water and warm water swimming are compared there are distinct differences in the mechanism of the induced analgesia. O'Connor and Chipkin¹⁵² showed that cold water swim analgesia was not reversed by the administration of naloxone, whereas the effect produced by swimming in warm water was significantly reduced by the administration of naloxone. Spontaneous running has also been shown to produce an analgesic effect in rats that is correlated with the amount of running and is partially reversed by the administration of naloxone.¹⁵³ Exercise in humans can cause elevation of β-endorphin levels in the peripheral circulation, predominantly related to increased lactate concentration.¹⁵⁴ Both incremental graded exercise and acute anaerobic exercise will produce a substantial increase in circulating β-endorphin concentration.¹⁵⁴ Changes induced by aerobic exercise appear to be more variable.¹⁵⁴ It is also of interest that exercise can have a significant influence on natural immunity, which appears to be linked to activation of endogenous opioid systems.^155,156 Moderate exercise induces increased concentration and activity of natural killer cells, potentially augmenting the immune response.^155,156

There is good evidence from basic science studies to suggest that exercise can activate endogenous pain modulation systems in both animals and humans. The endogenous opioid system appears to play a role in this analgesia although the mechanism is relatively complex and the exact nature of the analgesia appears to depend on various parameters of the exercise stressor. Evidence is emerging from clinical studies to support the beneficial effects of exercise in the management of chronic musculoskeletal pain although there is much less evidence in favor of exercise as a means of treating acute musculoskeletal pain. Further clinical research is required to distinguish between different forms of exercise.

CONCLUSION

It is interesting that for all of the major groupings of physical therapies discussed in this review, there is good evidence from basic science studies to suggest that the therapeutic approaches may have potentially beneficial effects on musculoskeletal pain. So far, however, this has failed to translate into sound evidence of effectiveness in the clinical environment for many of the treatment modalities. In part, this may be due to inadequacies of the clinical trials or because effects of individual treatments are relatively modest and in some cases short lasting. Stronger evidence to support the use of some treatments such as manual therapy, acupuncture, and exercise is beginning to emerge, however. Judicious combinations of treatments (as occurs in the clinical situation) may be required to produce a significant therapeutic effect. Interestingly, in a pragmatic trial, usual physical therapy care (judicious combination of all modalities at the discretion of the physical therapist) produced the best and most cost-effective outcome for patients with chronic back pain.¹³¹ The other issue that has yet to be resolved in relation to physical therapies is whether these treatments represent multiple ways of accessing the same endogenous pain control mechanism or whether there are real differences in the mechanism of action of different therapies that would allow for judicious combinations of therapies to maximize the endogenous analgesic effect produced. There is clearly a need for more research in this area, which should be a priority for national and international funding agencies.