|Home | About | Journals | Submit | Contact Us | Français|
Thoracic spine manipulation is commonly used by physical therapists for the management of patients with upper quarter pain syndromes. The theoretical construct for using thoracic manipulation for upper quarter conditions is a mainstay of a regional interdependence (RI) approach. The RI concept is likely much more complex and is perhaps driven by a neurophysiological response including those related to peripheral, spinal cord and supraspinal mechanisms. Recent evidence suggests that thoracic spine manipulation results in neurophysiological changes, which may lead to improved pain and outcomes in individuals with musculoskeletal disorders. The intent of this narrative review is to describe the research supporting the RI concept and its application to the treatment of individuals with neck and/or shoulder pain. Treatment utilizing both thrust and non-thrust thoracic manipulation has been shown to result in improvements in pain, range of motion and disability in patients with upper quarter conditions. Research has yet to determine optimal dosage, techniques or patient populations to which the RI approach should be applied; however, emerging evidence supporting a neurophysiological effect for thoracic spine manipulation may negate the need to fully answer this question. Certainly, there is a need for further research examining both the clinical efficacy and effectiveness of manual therapy interventions utilized in the RI model as well as the neurophysiological effects resulting from this intervention.
Thoracic spine manipulation is commonly used by physical therapists for the management of patients with neck and upper extremity pain syndromes. The theoretical construct for thoracic manipulation and its mechanism of action has traditionally been described as regional interdependence (RI). Regional interdependence has been defined by Wainner et al.1 as the ‘concept that seemingly unrelated impairments in a remote anatomical region may contribute to, or be associated with, the patient's primary complaint’. The authors originally described this as a mechanism solely functioning within the musculoskeletal system. However, Bialosky et al.2 argued that the RI concept is likely much more complex and could also be driven by a neurophysiological response including those related to peripheral mechanisms, spinal cord mechanisms or supraspinal mechanisms. More recently, the RI concept has been further refined by Suecki et al.3 as ‘the concept that a patient's primary musculoskeletal symptom(s) may be directly or indirectly related or influenced by impairments from various body regions and systems regardless of proximity to the primary symptom(s)’. This implies that impairments impacting one's musculoskeletal system may be greatly impacted through other systems, such as the peripheral and central nervous systems, via a neurophysiological or biopsychosocial response.
Descriptions of RI can be found in the literature for ~40 years, even prior to the coining of the familiar term, ‘RI’. In 1976, Murray-Leslie and Wright4 found that individuals who had carpal tunnel syndrome exhibited narrowing of the intervertebral discs of the cervical spine relative to the vertebral bodies. Furthermore, in 1976, Gunn and Milbrandt5 reported on the use of interventions directed to the cervical spine in individuals with tennis elbow. A majority of individuals who received intervention at the cervical spine experienced relief of distal symptoms. Early evidence of a relationship between the lumbar spine and lower extremities has also been documented. Cibulka et al.6 indicated the connection between passive hip internal and external range of motion (ROM) and sacroiliac joint pain. In more recent years, evidence has emerged demonstrating the effectiveness of utilizing manual therapy techniques that incorporate the concept of RI in various diagnoses, including spinal stenosis,7 knee osteoarthritis8 and patellofemoral syndrome.9 All of these studies found significantly greater improvements in patients who received manual therapy interventions both proximal and distal to the patients' primary symptoms. However, the main body of physical therapy literature using an RI approach describes treating the thoracic spine, often for individuals with neck pain and/or shoulder pain.10–18 The outcomes resulting from utilizing thoracic spine manipulation for individuals with neck pain and shoulder dysfunction have been very promising for reducing pain, increasing ROM and improving function.10–18
Bialosky et al.19 have proposed a model that identifies two categories of potential mechanisms by which manual therapy techniques exert their biomechanical and neurophysiological effects (Fig. 1). Biomechanical links can be made between different anatomical regions and concurrent symptoms such as the thoracic spine facet joint and referral of pain to the neck and shoulder.20 However, emerging evidence suggests that the neurophysiological effects of manual therapy play an important role, and it is likely that there is a complex inter-relationship between biomechanical and neurophysiological effects occurring in RI.19 The interplay between these theoretical mechanisms may maximize patient outcomes when the clinician includes treating anatomical areas sometimes remote from the site of the patient's symptoms. In addition, it is important to recognize that a third factor may come into play which is beyond the scope of the current manuscript; for example, outcomes may be further influenced by interaction between these effects combined with other non-specific factors including patient specific psychosocial factors and patient expectation.
Neurophysiological effects can be further sub-classified as peripheral, spinal and supraspinal mechanisms.19 Injuries to the musculoskeletal system may result in biochemical alterations in the periphery that may impact multiple body systems and initiate an inflammatory cascade. Preliminary evidence suggests that manual therapy techniques might have an influence on biochemical activity and perhaps even mediate the inflammatory process. A study by McPartland et al.21 demonstrated that manual therapy resulted in the production of endogenous cannabinoids when compared with a placebo. Another study found that thoracic manipulation reduced inflammatory cyotokines in the short term.22 Recently, Plaza-Manzano et al.23 randomly assigned individuals to either cervical manipulation, thoracic manipulation or control group. Both the cervical and thoracic spine manipulation groups experienced higher levels in neurotensin and the cervical manipulation group also experienced significantly greater increases in cortisol, suggesting that manipulation may positively influence biochemical processes.
Mechanisms associated with spinal cord activity may potentially be associated with hypoalgesia or altered motor neuron pool activity. For example, studies have demonstrated that spinal manipulation results in increased pressure pain thresholds in individuals with mechanical neck pain24 and increased thermal pain thresholds in asymptomatic individuals25 and in patients with low back pain.26,27 Dishman and Burke27 demonstrated that both cervical and lumbar manipulation may potentially alter motor neuron excitability.
Manual therapy has also been shown to exert effects on supraspinal structures and processes, including pain processing centres28 as well as endocrine and autonomic responses.19 Using functional magnetic resonance imaging (fMRI) to measure cerebral haemodynamics, a recent case series by Sparks et al.29 found thoracic spine manipulation resulted in reduced cerebral blood flow, as measured by the blood oxygenation level-dependent response, to areas associated with the pain matrix (insular cortex) in individuals with experimentally induced pain. In addition, the participants exhibited decreased subjective pain scores on the numeric pain rating scale. A recent systematic review found that all seven of the randomized clinical trials reported an increase in sympathetic nervous system activity, as measured by heart rate, respiratory rate, blood pressure or skin conductance following manual therapy.30 It should be recognized the aforementioned studies were all performed on asymptomatic individuals; therefore further research is needed to determine if these neurophysiological mechanisms occur with manual therapy when applied to individuals with musculoskeletal disorders and pain.
Thoracic spine manipulation for the treatment of neck pain has historically been thought to have less risk over cervical spine manipulation. Specifically, Masaracchio et al.31 suggest that clinicians have more commonly chosen to target the thoracic spine with manipulation instead of the cervical spine because of the controversy surrounding the perceived safety of cervical manipulation. Consequently, numerous studies have focused on the use of thoracic spine manipulation versus cervical spine manipulation for treating individuals with neck pain.13–15,31–36 In contrast, it has been suggested that cervical thrust manipulation may carry a more significant risk of vertebrobasilar artery injury and some purport that the benefits do not outweigh the risks37 despite research showing overall risk of complications to be low and potentially preventable.38–40
Several studies have utilized thoracic manipulation as an intervention on its own,33,41–44 compared it to another intervention or used it in conjunction with an alternate intervention.31,32,34,35,45–49
For those trials that compared thoracic manipulation to another intervention, the comparison intervention varied. In a randomized clinical trial, Cleland et al.32 compared the outcomes of a group of patients with mechanical neck pain receiving thoracic spine manipulation versus non-thrust manipulation. The results demonstrated that those who received manipulation experienced significantly greater improvements in pain and disability over those receiving non-thrust manipulation at short-term follow-up. Suvarnnato et al.44 also compared manipulation versus non-thrust manipulation and reported significant differences in pain and ROM immediately and 24 hours after intervention. One additional study compared supine thoracic manipulation to a placebo manipulation and those in the experimental group experienced an immediate and statistically significant reduction in pain.14 Puentedura et al.50 and Martinez-Segura et al.51 compared thoracic manipulation to cervical manipulation. The results of these two studies varied, with Puentedura et al.50 reporting statistically significant greater improvements in pain reduction and decreased neck disability with cervical manipulation whereas Martinez-Segura et al.51 showed similar reductions in pain for both the cervical and thoracic manipulation groups.
Several studies have examined the effects of thoracic manipulation in conjunction with alternative types of intervention, which is more reflective of clinical practice. Four of the studies combined thoracic manipulation with the use of modalities.34,35,45,46 Savolainen et al.48 used thoracic manipulation and a standard exercise program and in Cleland et al.'s13 derivation of a clinical prediction rule included thoracic manipulation and a home exercise programme for cervical active ROM. Saavedra-Hernández et al.49 and Dunning et al.47 looked at the combined use of cervical and thoracic manipulations, whereas Masaracchio et al.31 examined the use of thoracic manipulation and cervical non-thrust manipulation compared to only cervical non-thrust manipulation. All of the studies31,32,34,35,45–47 utilizing adjunct interventions combined with thoracic manipulation (multimodal approach) found statistically significant greater improvement and/or clinically meaningful differences for pain and disability scores. Five of the studies also found statistically significant differences in ROM measurements after intervention.34,35,45–47
Studies that have examined the effects of thoracic spine manipulation for individuals with neck pain have included a variety of technique positions, including supine, seated or prone, (Figs. 2 and 3) with limited evidence to suggest which specific thoracic manipulation technique is most effective. Karas and Hunt42 and Casanova-Mendez et al.43 both recently studied the use of specifically targeted thoracic manipulation techniques. Karas and Hunt42 compared the effects of a seated thoracic manipulation to a targeted supine thoracic technique. Those who received a targeted supine thoracic manipulation had statistically greater improvements in pain over those receiving a seated thoracic manipulation. In the Casanova-Mendez et al.43 study, a supine technique was compared to a prone technique and those receiving the prone technique had statistically greater improvements in cervical extension, right cervical side flexion and left cervical rotation. Overall, there is a large body of evidence supporting the use of thoracic manipulation via the RI approach for treating patients with neck pain.
In addition to treating mechanical neck pain, thoracic manipulation is a popular management strategy in treating patients with shoulder pain. According to a number of sources, of patients with a complaint of shoulder pain, more than 40% had associated impairment of the cervicothoracic spine and adjacent ribs.52–57 In addition Sobel et al.52,53 postulated that dysfunction in these adjacent areas may be a primary cause of some patients' shoulder pain. Furthermore, impairment of the cervicothoracic spine and ribs may increase an individual's risk of developing neck–shoulder pain and may contribute to an overall worse prognosis.10,52,54–56 The rationale for using thoracic spine manual therapy in individuals with shoulder pain has some face validity because of the increased prevalence of reduced mobility of the upper thoracic spine in these individuals.52,55,56,58 Furthermore, a study by Crosbie59 highlighted the inter-relationship between the thoracic spine and arm elevation, further supporting the concept that thoracic spine hypomobility may impair shoulder mechanics and ultimately patient function.
The RI model of Wainner et al.1 suggests that treating the thoracic spine in individuals with shoulder pain is warranted, and several research studies support this treatment approach.10–12 Studies have examined techniques directed at the thoracic spine to treat primary shoulder pain,10,12,60 shoulder impingement11,58,61,62 and rotator cuff tendinopathy.63 In a recent study by Haik et al.,58 individuals with shoulder impingement syndrome (SIS) had an immediate decrease in pain following a treatment of thoracic spine manipulation.
Studies investigating the use of thoracic spine manipulation for the treatment of shoulder pain have demonstrated improvement in shoulder pain, disability and shoulder ROM.11,12,58,63 A long-term follow-up study of patients with a primary report of shoulder pain found, at the follow-up periods of 12, 26 and 52 weeks, that the group that received manipulative therapy (UMC_MT) to the cervicothoracic spine and rib cage in addition to usual medical care (UMC) from their primary physician reported significantly greater improvements in severity of pain and disability, and demonstrated higher rates of perceived ‘full recovery’ than a group that solely received UMC.10 A study by Winters et al.64 investigated the use of manipulation, physiotherapy (exercise therapy, massage and physical modalities, excluding the use of thrust or non-thrust manipulation) and corticosteroid for the treatment of shoulder complaints. Patients were grouped into a ‘shoulder girdle group’ (pain and decreased movement proposed to originate from the cervical or thoracic spine and ribs) or the ‘synovial group’ (shoulder pain originating from subacromial, acromioclavicular or glenohumeral regions). The results of the study found that, in the shoulder girdle group, manipulation was superior to physiotherapy with 70% of patients in the manipulation group considering themselves cured.64 This suggests that there may exist a subgroup of individuals with shoulder pain that experience a dramatic response from thoracic manipulation.12
Additional effects on the shoulder girdle region after thoracic manipulation include increased middle trapezius activity in individuals with rotator cuff tendinopathy63 and increased lower trapezius strength in asymptomatic individuals.65 While scapular muscle strength may improve following thoracic manual therapy, changes in scapulohumeral rhythm and mechanics were not found to change significantly after application of thoracic manipulative interventions.58,63,66 These results further support a neurophysiological effect, as opposed to a local biomechanical effect.67 One could argue that this improved muscle function gives health care providers a therapeutic window to maximize therapeutic exercise interventions.
Certainly, there is a need for more research examining the clinical effectiveness of manual therapy interventions utilized in the RI model. It is also important to continue to examine the mechanisms of spinal manipulation as well as the interactions between the non-specific, neurophysiological and local mechanical effects. Although the RI approach of using thoracic manipulation for the management of neck and shoulder pain has shown positive outcomes, many of the aforementioned studies only assess immediate, short-term outcomes.14,41–43,45,51 Few studies have followed patients for longer time frames, such as 12 months.48 When considering the existing body of knowledge in this area, some important questions remain. How long do these effects actually last? What is the optimal dosage required? How many treatment sessions are appropriate to maximize patient response? Are there subgroups of patients who will respond more favourably? What treatments constitute the ideal ‘combined’ treatment approach?
Most clinical trials investigating specific interventions attempt to answer an explanatory question: Under ideal circumstances, can a certain intervention in ideal patients improve patients' outcomes?68 These explanatory, or prescriptive, trials explore efficacy and may be useful to expert clinicians with highly compliant patients but do not necessarily represent clinical reality. Pragmatic trials, on the other hand, ask a broader, more clinically relevant question: Does this intervention improve patient specific outcomes when applied by typical clinicians to typical patients? The goal of the pragmatic trial is to explore the effectiveness of a treatment in the ‘real,’ clinical world. In other words, prescriptive trials explore whether a given treatment works under ideal circumstances while pragmatic trials explore whether the treatment provides benefit under actual clinical practice circumstances. To this end, prescriptive trials frequently have strict inclusion and exclusion criteria to decrease the likelihood that extraneous factors, such as medical conditions, will impact the outcomes.69 Pragmatic trials usually attempt to be as inclusive as possible to increase the generalizability of the outcomes. As the clinic is not an ‘ideal’ world, and most of the patients the authors treat are not ‘ideal’ patients, the authors would argue that research in the area of thoracic manual therapy should shift to more pragmatic designs. In the clinic, the application of manual therapy includes the patient's feedback and modifications of multiple factors based on this feedback coupled with patient response. In a pragmatic trial, the researchers are not concerned with how an intervention works compared to a control, but they are simply interested in whether it works or not. Much of the research that has been conducted showing the clinical benefits of thoracic spine manual therapy has been prescriptive in nature, and several studies comparing low-velocity non-thrust manipulation to high-velocity manipulation have suggested that high-velocity techniques are superior to low velocity.32,47 Cook et al.70 argue that these outcomes may be because, in part, of the fact that the low-velocity techniques were prescriptively used, which is dissimilar to how the techniques are applied in clinical practice. In a recent systematic review of 14 studies by Young et al.17 looking at thoracic non-thrust manipulation versus thrust manipulation, the authors reported that there were improvements in pain, ROM and disability with the use of either technique, but they did not report if the studies were more pragmatic or prescriptive in nature. Future research should be conducted to directly compare the use of thoracic manipulation versus non-thrust manipulation and the use of pragmatic versus prescriptive applications of thoracic spine manual therapy in the treatment of musculoskeletal disorders.
Future studies should also include long-term follow-up since the majority of the studies in the literature only examined short-term outcomes. Ideally, researchers should strive to use multicentre, pragmatic, randomized controlled study designs with multiple treating clinicians to make the research more clinically applicable, generalizable and representative of the patients that we treat. Future studies should also investigate the effects of specific thoracic manipulation techniques, as many studies utilize ‘manual therapy packages’ which precludes the ability to determine if one technique is superior to another. In addition, future studies of the neurophysiological and non-specific effects of thoracic manipulation should investigate changes over multiple sessions and take into account a patient's desire for pain relief.71 Finally, continued research is needed to determine the ideal number of treatment sessions and the most effective dosage of the various manipulative techniques.
As we have illustrated above, Wainner et al.'s1 model of RI seems to have some validity in the physical therapy management of musculoskeletal conditions based on published research. This model seems particularly suitable to the application of manual therapy to the thoracic spine given the emerging evidence supporting a neurophysiological effect for this intervention.19,28,29,72,73 Given the reported lack of accuracy and specificity of manual therapy techniques,74–76 it seems even more plausible that the reported positive effects can be explained by non-specific and neurophysiological mechanisms. A systematic review by Coronado et al.77 concluded that manipulation is likely a non-specific technique acting on the pain modulating system, even though the exact mechanisms remain elusive. It is clear that neurophysiological effects occur following manual therapy, and that these effects are not necessarily specific to the region to which they are applied.19,67,78–80 Hypoalgesia following manual therapy has been reported in regions distant from the application.36,51,67,81 As described earlier, fMRI findings from a study by Sparks et al.29 marks a considerable paradigm shift in manual therapy clinical decision making. Rather than using manual therapy to treat a localized biomechanical impairment, today's clinician, armed with current best evidence, may decide to treat a patient with shoulder pain using thoracic manipulation based on a well-documented neurophysiological effect, as opposed to a local biomechanical effect. This decision would be weighed more heavily towards current best evidence over examination findings from clinical tests and measures that are limited by questionable reliability and validity.82,83 In addition, non-specific mechanisms, such as the placebo response and patient expectation, have been shown to have a significant effect on outcomes.71,84–87 As clinicians, we should embrace these findings to ultimately improve our treatment outcomes.
Contributors All authors contributed to the writing of the paper.
Conflicts of interest The authors declare no conflict of interest.
Ethics approval None required.