Introduction

Physical therapists have historically used manual therapy to treat musculoskeletal pain. Manual therapy was introduced to the profession by physical therapists who trained with osteopathic, chiropractic, and orthopedic physicians. James Cyriax (United Kingdom) seems to be the most commonly mentioned name associated with early mentoring of physical therapists in the practice of manual therapy.

Physical therapists such as Gregory Greive (United Kingdom), Freddy Kaltenborn (Norway), and Geoffrey Maitland (Australia) would eventually adopt their own manual therapy techniques and philosophies and begin widespread teaching to other therapists. Grieve, Kaltenborn, and Maitland also worked to develop a governing body for physical therapists practicing manual therapy, and in the late 1960s each was involved in the creation of the International Federation of Orthopaedic Manipulative Therapy (IFOMT). Stanley Paris (New Zealand) was the founding chairman of IFOMT and is credited with accelerating the use of manual therapy in the United States, through his participation in the Orthopedic Section of the American Physical Therapy Association (APTA), the founding of a physical therapy school, and continuing education programs.

As is typical in manual medicine, the use of manual therapy by physical therapists is linked to specific philosophies and approaches. For example, therapists trained in Maitland techniques would be expected to commonly utilize oscillatory joint mobilization techniques, and much less frequently to use high-velocity, low-amplitude thrust techniques. Although this culture of trainer-specific approaches still exists in physical therapy, recent trends indicate a shift away from this emphasis in the use of manual therapy.

Standardization of manual therapy instruction

The standardization of manual therapy teaching is one example of a move away from trainer-specific approaches. Historically in the United States, there was considerable variation in whether manual therapy was taught in entry-level physical therapy programs and, if taught, what philosophies and techniques were included. One model of standardization is evidenced in the Guide to Physical Therapist Practice (2001), which outlines broad utilization parameters for manual therapy. For example, the guide describes the use of mobilization and manipulation and outlines practice patterns which are appropriate for treatment with manual therapy. Another source of standardization provided by the APTA is the Manipulation Education Manual. This provides information on specific examination skills and manual therapy techniques in which physical therapists should be proficient.

It is appropriate to note that a majority of the language provided by the APTA describes physical therapists’ use of manual therapy as focused to ‘joint and soft tissue’ structures. The application of manual therapy to visceral targets is not explicitly addressed, although the language related to ‘soft tissue’ is vague enough to allow for alternate interpretations. Anecdotally, visceral manipulation is rarely taught in entry-level education programs, but can be accessed by physical therapists through continuing education formats. Visceral manipulation techniques attempt to effect visceral structures through external body contact and are more common in massage and osteopathic practice (Barral & Mercier 1998). When visceral approaches are used by physical therapists the focus remains on the treatment of chronic pain, instead of directly affecting visceral function.

Clinical implementation of manual therapy

Clinical decision making is another recent trend affecting the use of manual therapy by physical therapists. Clinical decision making has undergone a paradigm shift over the past 10 years, and much of the supporting research for this has been led by physical therapists. Traditionally, appropriateness for manual therapy hinged on the identification of key structural or movement faults believed to be pathological or abnormal. After the pathology or abnormality was identified, manual therapy was utilized to ‘correct’ the fault, resulting in less pain and improved function for the patient. It is beyond the purpose of this chapter to exhaustively review all manual therapy approaches to demonstrate this common link; instead, a specific example will be provided.

In 1995, a treatment-based classification (TBC) system was proposed for low back pain (Delitto et al. 1995). Within this system there were two treatment categories appropriate for manual therapy. One hinged on the identification of static or dynamic pelvic asymmetry through palpation of key bony landmarks such as the anterior and posterior superior iliac spine (ASIS and PSIS, respectively). Palpation of these landmarks and positional tests served to determine whether asymmetry or pain existed, and whether the patient was appropriate for manual therapy. For example, an anterior innominate could be detected by comparing bilateral ASIS and PSIS positions. In the original article, either a pelvic manipulation technique applied through the ASIS or a pelvic muscle energy technique was suggested to correct the rotation (Delitto et al. 1995). Another treatment category hinged on the identification of lumbar movement faults consistent with ‘blocked’ facet joints. These movement faults were identified by recognizing opening and closing movement patterns during lumbar movement testing, as well as complementary findings when passive intervertebral mobility of the lumbar spine was tested. In this treatment category an appropriately directed lumbar manipulation technique (posterior/inferior for closing restriction and anterior/superior for opening restriction) was recommended to correct the movement fault (Delitto et al. 1995).

These ‘correction’ approaches, and others like them, have been widely taught to physical therapy students and are commonly implemented in physical therapy clinical settings. However, research reported in the peer-reviewed literature questions the validity of treatment approaches based solely on the identification of structural pathology as normally performed by manual therapists (i.e., through palpation, pain provocation, movement testing, and observation). Again, it is beyond the purposes of this chapter to exhaustively review this literature. Instead, the focus will be on examples that represent general trends in the literature suggesting that correction-focused application of manual therapy is a questionable treatment strategy.

One apparent trend in the literature is that the reliability of identification of structural pathology by palpation is poor. Two studies (Levangie 1999a, Riddle & Freburger 2002) have reported that the inter-rater reliability of the identification of pelvic asymmetry, with either a static or a dynamic approach, is too low for use in most clinical settings. Furthermore, there is no strong association between pelvic asymmetry and low back pain, which also questions the use of these techniques in clinical settings (Levangie 1999b). Another trend noted in the literature is that manual therapy is applied with forces that are variable (Herzog et al. 2001, Hessell et al. 1990), nonspecific (Herzog et al. 2001 Ross et al. 2004), and does not result in measurable positional changes in targeted joints (Tullberg et al. 1998). Specifically, in a study of 10 consecutive patients identified with sacroiliac joint dysfunction and treated with an appropriate spinal manipulation technique, clinicians noted an improvement in sacroiliac joint dysfunction on clinical examination, although the reference standard (roentgen stereophotogrammetric analysis) did not detect any change in joint motion that exceeded measurement error (Tullberg et al. 1998).

These trends de-emphasizing the importance of identifying structural pathology continue in studies with a clinical focus. A study of 140 subjects with low back pain suggested that although pain relief was an expected outcome, there was no difference in outcomes if a specific lumbar manual therapy technique was selected by experienced clinicians, or whether the technique was randomly determined (Chiradejnant et al. 2003). In another study, subjects underwent an extensive physical examination including static, dynamic, and provocation tests for commonly utilized pelvic assessment (Flynn et al. 2002). Patients were then treated with a specific manipulation technique and range of motion exercises, regardless of whether the examination results indicated that treatment was indicated. This study design may seem counterintuitive at first; however, it must be stressed that this is an appropriate methodology to determine which baseline examination characteristics predict a favorable response to a selected intervention. Essentially, this design allows determination of how variability in patient baseline status affects variability in outcome. The results of the study indicated something very interesting: namely, that the presence of pelvic asymmetry (dynamic or static) was not predictive of a successful outcome from the manipulation (Flynn et al. 2002).

These studies highlight a shift in philosophy that de-emphasizes the importance of identifying specific structural pathology before the application of manual therapy. Instead, the identification of clinical factors that indicate a favorable response has been emphasized (Fig. 14.1). Essentially, the identification of responder subgroups is now a treatment priority, because it appears that larger treatment effects are observed if focused manual therapy is applied to those likely to respond (Brennan et al. 2006, Childs et al. 2004, Fritz et al. 2003). In the case of lumbar spinal manipulation, a clinical prediction rule consisting of five factors (duration of symptoms, location of symptoms, lower score on psychological distress questionnaire, presence of lumbar hypomobility at any segment, and hip internal range of motion >35°) has been identified and validated (Childs et al. 2004). Derivation clinical prediction rules for a variety of other musculoskeletal pain conditions have been reported by physical therapy researchers (Table 14.1). This philosophical shift is still in its professional infancy, but it is quite clear that is has an immediate impact on physical therapy practice, and it is highly likely that the profession will de-emphasize the application of manual therapy based solely on biomechanical theory and continue to develop and validate clinical prediction rules.

Figure 14.1 Shift in clinical decision making of utilization of manual therapy.

Table 14.1 Clinical prediction rule studies for utilization of manual therapy by physical therapists that have been reported in the peer-review literature

Selected mechanism of pain inhibition for manual therapy

From the previously cited examples it appears that manual therapy has great potential to be effective for musculoskeletal pain. However, many questions remain regarding the mechanisms of manual therapy for pain relief. Neurophysiological effects have always been associated with manual therapy, but over the last 15–20 years mechanistic studies have emphasized the importance of reflex mechanisms, as reviewed by Pickar (2002). It is beyond the scope of this chapter to review neurophysiological mechanisms involving muscle spindles, Golgi tendon organ, paraspinal muscle reflexes and/or motor neuron excitability. This is not meant to diminish these topics, but rather to allow for focus on a reflex that directly involves the processing of pain.

Surprisingly, manual therapy’s direct effect on pain has not been extensively covered in the literature, as indicated in a qualitative review (Vernon 2000). Prior studies consistently reported hypoalgesia, or a lessening of pain, reports in response to a standard stimulus. Therefore, hypoalgesia following manual therapy is commonly reported in the literature. However, previous methodologies have only studied general aspects of hypoalgesia and have not been able to discern specific mechanisms.

Manual therapy’s effect has been hypothesized as a ‘counterirritant’ stimulus to nociceptive input received by dorsal horn cells (Boal & Gillette 2004). Specifically, manual therapy activates joint and muscle spindle mechanoreceptors, with the potential of activating low (Aβ) and high (Aδ, C) threshold afferents. This afferent input converges on the spinal cord, where inhibition of dorsal horn cells is physiologically plausible (Boal & Gillette 2004).

In the author’s opinion a likely explanation for the effectiveness of manual therapy could be that it has an inhibition effect that is specific to C-fiber-mediated pain perception. Numerous animal studies have suggested that central sensitization of nociceptive systems leading to pain is a specific neurophysiological mechanism associated with the development and maintenance of chronic pain syndromes (Dickenson & Sullivan 1987, 1990, Price et al. 1978, 1994, Vierck et al. 1997, Woolf & Thompson 1991). ‘Wind-up’ is a specific example of central sensitization within dorsal horn cells which results from tonic, peripheral nociceptive C-fiber input. The tonic, nociceptive input activates N-methyl-d-aspartate (NMDA) and substance P receptors in a wide dynamic range and nociceptive-specific dorsal horn cells. The activation of these cells induces a central hyperalgesia mediated at the spinal cord level, such that subsequent evoked pain stimuli are relayed from dorsal horn cells as increasing in intensity, despite their being of standard amplitude. In animal models, the temporal parameter (frequency of nociceptive input) is a primary factor in eliciting wind-up (Price et al. 1978).

Direct measurement of wind-up is not feasible in humans, but temporal summation of thermal stimuli is an accepted behavioral measure of wind-up (Price 1999). The use of temporal summation as a proxy measure of wind-up is supported by human studies, which consistently demonstrate that an increase in the frequency of standard nociceptive input is associated with an increase in pain perception (Price et al. 2002, Robinson et al. 2004, Staud et al. 2001). Specifically, thermal input at 0.33 Hz or higher frequencies tends to induce temporal summation in humans, whereas input at 0.20 Hz or slower frequencies does not (Price 1999). It should be noted that this explanation is currently speculative, as only one study directly links temporal summation with disability from chronic LBP (George et al. 2006b).

If manual therapy has the potential to inhibit pain directly, it would have a measurable hypoalgesic effect on pain perception. As was previously mentioned, manual therapy hypoalgesia has been observed by decreased cutaneous receptive field from pin-prick (Glover et al. 1974), tolerance from electrical current (Terrett & Vernon 1984), and mechanical pressure (Vernon et al. 1990). Collectively, these results demonstrate manual therapy’s potential for general dorsal horn-mediated pain inhibition. There are, however, several important, unresolved issues regarding manual therapy hypoalgesia, and these provided the impetus for a novel line of research that will be highlighted in this chapter.

The previously cited studies reported a hypoalgesic response in anatomical areas with the same or overlapping dermatomes as those treated with manual therapy (Glover et al. 1974, Terrett & Vernon 1984, Vernon et al. 1990), for example assessing hypoalgesic response to cervical manipulation only in anatomical areas innervated by cervical nerve roots, such as the upper extremity (Vernon et al. 1990). As a result, these studies were unable to determine whether the observed hypoalgesia was a general effect or a specific effect local to the spinal levels involved with the manipulation (Vernon, 2000). Previous studies utilized pain induction protocols assessing general peripheral pain perception, instead of distinguishing manual therapy’s separate effect on Aδ- and C-fiber-mediated pain perception (Price 1999).

Manual therapy’s potential for specific inhibition of C-fiber-mediated pain perception has not been widely tested in human studies. Previously cited data support the relevance of investigating manual therapy’s effect on wind-up (or temporal summation) in commonly encountered chronic pain states such as fibromyalgia and low back pain. However, it is also important to note that we do not know whether this mechanism of acute central sensitization generated by stimulation of nociceptors and the mechanisms of central sensitization underlying chronic pain conditions are the same.

In our proposed line of research we hypothesized that the effectiveness of manual therapy was through inhibition of dorsal horn wind-up. This would be detectable in experimental settings by measuring decreased temporal summation following the application of manual therapy. If this hypothesis was confirmed, manual therapy could be viewed as a treatment with the potential to reduce the development of chronic pain syndromes, at least those associated with dorsal horn wind-up. This hypothesis was further supported by a previous cross-sectional study in which increased temporal summation was associated with increased self-reporting of disability in patients with chronic low back pain (George et al. 2006b). Therefore, the overriding purpose of this line of research was to determine whether: