Article Text
Abstract
This account of the condition now termed complex regional pain syndrome (CRPS) spans approximately 462 years since a description embodying similar clinical features was described by Ambroise Paré in 1557. While reviewing its historical origins, the text describes why it became necessary to change the taxonomies of two clinical syndromes with similar pathophysiologies to one which acknowledges this aspect but does not introduce any mechanistic overtones. Discussed at length is the role of the sympathetic component of the autonomic nervous system (ANS) and why its dysfunction has both directly and indirectly influenced our understanding of the inflammatory aspects of CRPS. As the following article will show, our knowledge has expanded in an exponential fashion to include musculoskeletal, immune, autoimmune, central and peripheral nervous system and ANS dysfunction, all of which increase the complexity of its clinical management. A burgeoning literature is beginning to shed light on the mechanistic aspects of these syndromes and the increasing evidence of a genetic influence on such factors as autoimmunity, and its importance is also discussed at length. An important aspect that has been missing from the diagnostic criteria is a measure of disease severity. The recent validation of a CRPS Severity Score is also included.
- reflex sympathetic dystrophy
- causalgia
- complex regional pain syndrome
- sympathetic
- central and peripheral nervous systems
- genetic
- immune
- autoimmunity musculoskeletal
- brain reorganization
- severity score
- taxonomy
- diagnostic criteria
- Autonomic Nervous System (ANS)
Statistics from Altmetric.com
- reflex sympathetic dystrophy
- causalgia
- complex regional pain syndrome
- sympathetic
- central and peripheral nervous systems
- genetic
- immune
- autoimmunity musculoskeletal
- brain reorganization
- severity score
- taxonomy
- diagnostic criteria
- Autonomic Nervous System (ANS)
a Prologue
In his classical account “Causalgia: A Centennial Review,” Richards1 referred to the outstanding description by Silas Weir Mitchell and his colleagues George Morehouse and William Keen2 of the lesion and severe and intractable burning pain that attended nerve wounds in a small number of soldiers. He also noted that Denmark3 had reported the amputation of a soldier’s arm because of severe burning pain after being wounded during the Peninsular War. One hundred years earlier, Paget4 described nutritional changes that are often seen in the fingers that are accompanied by “ distressing and hardly manageable pain and disability.” Ambroise Paré in 1557,5 however, is probably the first person to write about the clinical features of the syndrome that bears the name causalgia the term introduced by Mitchell. It should be noted that while Mitchell was not explicit in defining what constitutes the signs and symptoms of causalgia, a committee set up by the British Medical Research Council6 in 1929 to address the diagnosis and treatment of nerve injuries defined causalgic pain in the following terms: (1) it is spontaneous; (2) it is hot and burning in character, intense, diffuse, persistent, but subject to exacerbations; (3) it can be excited by stimulae which do not necessarily produce a physical effect on the limb; and (4) it tends to lead to profound changes in the mental state of the patient.
A working diagnosis of causalgia proposed by Kirklin et al7 at the end of World War II when he described an incidence of 1.8% of 2850 nerve injury cases, was “a syndrome associated with the lesion of a peripheral nerve containing sensory fibers manifested by pain in the affected extremity; this pain is usually of a burning character and is usually located in an area corresponding to the continuous distribution of the involved nerve. An integral characteristic of this pain, one whose presence is necessary in order to make the diagnosis, is its accentuation by certain disturbing features in the affected individual environment.”
Of importance to understanding the incidence of causalgia are data from Sunderland and Kelly,8 and Nathan,9 also from World War II, who described an incidence of 12.0% and 13.8% respectively out of a total 442 nerve injury cases involving primarily median and sciatic nerves. These are almost 2 to 3 times the incidence of causalgia reported by other observers but may reflect the higher proportion of sympathetic fibers in these 2 nerves.
Morbus Sudeck, a term adopted particularly by the German-speaking medical professions following Paul Sudeck’s description in 1900,10 was equated with causalgia but without a known nerve lesion. Similarly, Reflex Sympathetic Dystrophy (RSD) the name introduced by Evans11 at the end of World War II, has outlasted all the other acronyms that have been introduced during the past seven decades. Interestingly, both Sudeck and Evans felt that an underlying disturbance of the sympathetic nervous system was fundamental to expression of the clinical features that are observed in the syndrome. Notwithstanding that Sudeck’s primary interest was the bony inflammatory process and radiologic changes about which he wrote, he also described the motor dysfunction and difficulty in treating this condition. On the other hand, Evans was influenced by the description given by Lorente de No12 and Livingston,13 14 each of whom suggested that constant nociception from the injured extremity can be described as “a prolonged bombardment of pain impulses sets up a vicious circle of reflexes spreading through a pool of many neuronal connections upward, downward and even across the spinal cord”. Summation of this activity also incorporates synapses with sympathetic motor neuron cells in the ipsilateral intermedio-lateral column that controls vasomotor tone, sweating and visceral actions. The pathophysiology suggested by this activity could cause spasm in both arterioles and venues,’ raising filtration pressure, causing edema and clinical swelling. They also suggested that hypoxia would cause increased capillary permeability and filtration resulting in worsening edema.
John Bonica,15 who knew Livingston well,13 14 and from his own considerable experience gained from managing thousands of wounded troops from the Pacific arena during World War II, while Chief of Anesthesia at Madigan Army Hospital, Tacoma, Washington, was intimately familiar with conditions we now describe as complex regional pain syndrome types 1 and 2 (CRPS 1 and 2). He was convinced that a disturbance of the sympathetic nervous system (SNS) was fundamental to the expression of reflex sympathetic dystrophy (RSD), the term he adopted and included in his book The Management of Pain. Interestingly, during World War I, Leriche16 had already instituted the use of surgical sympathectomy by stripping sympathetic nerve fibers from the peripheral blood vessels of the injured region in soldiers to modify both the sensory response and pain. He was influenced by his mentor, Claude Bernard,17 the great French physiologist, who had extensively explored the role of sympathetic fibers to augment pain severity. As a prelude to the management of CRPS, Bonica15 promoted the idea of always performing a sympathetic block. The rationale was supported by Roberts,18 whose hypothesis that increased activity in wide dynamic range neurons in lamina 5 of the dorsal horn was responsible for initiating and maintaining what was at the time considered to be sympathetic hyperactivity. The term sympathetically maintained pain (SMP) was later introduced by him to describe those patients whose symptoms were relieved by a sympathetic block. This gave support to the contemporary view that a positive response to a sympathetic block, that is, SMP, was necessary to support a diagnosis of RSD.19 For further discussion, see the Sympathetic nervous system mechanisms section. The term sympathetically independent pain (SIP) was suggested by Campbell et al 20 to describe those patients with other features of RSD (eg, edema, vasomotor changes) who had a negative response to a sympathetic block, and therefore by definition did not manifest the SMP syndrome. While evidence in support of SNS activity in the genesis of CRPS is incomplete, there is increasing evidence for an indirect role. Furthermore, many studies have identified target receptors (α1 and α2 adrenoceptors) on nociceptors, sclerocytes, inflammatory cytokines, and macrophages at which SNS activity is realized.21 22 The fact that not all patients have a positive response to a sympathetic block is the reason why this component to satisfy the earlier diagnostic criteria was removed (see the Sympathetic nervous system mechanisms section).
A word should be said about the use of stages when describing the temporal changes familiar to CRPS. Steinbrocker et al 23 described the stages seen in shoulder hand syndrome. Based on his extensive clinical observations, Bonica incorporated three stages of the syndrome in his book, The Management of Pain.15 To the first stage which included burning dysesthesia, hot dry skin, edema, hyperesthesia, and muscle spasm, he assigned a time frame of a few weeks. In the second stage which would last 6 months, he described decreased movement, stiffness, spreading edema, decreased pain, cold moist skin, cyanosis, and dystrophic changes. The third stage is accompanied by decreasing function, atrophy, tendon shortening and contractions, and increasing integumentary changes that involve the nails, skin, and hair. While parallels can be made with contemporary views based on recent prospective controlled studies, the reason why stages are not useful to the diagnostic criteria is the implication of a strict temporal framework. In clinical practice the signs in one stage may randomly occur in other stages, and while structural tissue changes are not necessarily inevitable they may only be seen in a relatively small proportion of the CRPS population. Also, two quantitative cluster analysis studies failed to support time-dependent stages, Bruehl et al 24 and de Mos et al.25 The largest prospective longitudinal study by Veldman et al 26 has confirmed these points and begs the need to do a large multicenter longitudinal study that will define the natural history of CRPS. For the foregoing reasons, staging as a component in a diagnostic taxonomy was rejected at the 2000 IASP (Cardiff) Consensus Conference.27
Changing a name, developing a new taxonomy
Several earlier publications had attempted to categorize the diagnostic clinical features that represent what is now termed CRPS. Although none were validated, each group of authors has described signs and symptoms that are most frequently found in patients with CRPS. An early definition by Kozin et al,28 29 who was interested in scintigraphic testing, was too broad and in practice included too many non-RSD patients. Gibbons and Wilson30 used a numeric grading scale based on history, examination, and laboratory testing. SMP was not a requirement to fulfill the diagnosis. Blumberg et al 31 proposed a numeric grading scheme that had an autonomic-motor-sensory triad with a maximum numeric score of 18. Others including Veldman et al 26 have suggested diagnostic criteria, but like earlier attempts never undertook the rigorous scientific scrutiny as a necessary prelude to the validation of their results.
Out of frustration with the then-status quo and lack of uniform management of patients suffering from CRPS, I organized a workshop that included clinicians and basic scientists from nine countries was held at Schloss Rettershof, Kelkheim/Mainz in 198832 to discuss diagnostic criteria, epidemiology, and laboratory guidelines for future basic and clinical research to better understand the nature of this syndrome. An ulterior motive of this workshop was also to enlist the IASP in helping this group in two ways: (1) to form a Special Interest Group (SIG)-Pain and the Sympathetic Nervous System; and (2) to sponsor a second meeting, the purpose of which would be to develop a new taxonomy that could be presented to the Committee on Classification of Chronic Pain Terms of the IASP for their consideration. The list of participants is shown in box 1.
Participants: names of participants at the initial meeting held at Schloss Rettershof, Kelkheim, in September 1988 to consider diagnostic criteria and a new taxonomy for reflex sympathetic dystrophy
Wilfrid Jänig, Dr Med.
Christopher Glynn, MBBS.
Martin Zimmermann, Dr Med.
Terrence Murphy, MBChB.
Edmond Charlton, MBBS.
William Roberts, PhD.
Martin Kolzenberg, Dr Med.
Hans Nolte, Dr Med.
IImar Jurna, Dr Med.
Jennifer Kelly.
Hermann Kreuscher, Dr Med.
Peter Wilson, MBBS.
Karen McCann.
Gabor Racz, MD.
Stephen Butler, MD.
Erik Torbjörk, PhD.
Prithvi Raj, MBBS.
Ulf Egle, Dr Med.
Robert Boas, MBBS.
Helmut Blumberg, Dr Med.
Stephen Abram, MD.
David Haddox, DDS, MD.
JG Hannington-Kiff, MBBS.
Albert van Steenberge, Dr Med.
Hans Gebershagen, Dr Med.
O Nickel, Dr Med.
Michael Stanton-Hicks, MBBS.
Ronald Tasker, MD.
Box modified with permission from Stanton-Hicks et al.32
The conclusions and consensus drawn from this workshop are shown in box 2.
Proposed basis of a new terminology for reflex sympathetic dystrophy (RSD)
Definition: a syndrome of continuous diffuse pain, often burning in nature, and usually consequent to injury or noxious stimulus, and disuse, presenting with variable sensory, motor, autonomic, and trophic changes; causalgia represents a specific presentation of reflex sympathetic dystrophy associated with a peripheral nerve injury.
Clinical features: the symptoms and changes spread independently of both the source and site of the precipitating event, presenting with a glove and stocking anatomic distribution. Clinical findings include disturbances of:
Autonomic deregulation: alterations in blood flow, hypohidrosis/hyperhidrosis.
Sensory abnormalities: hypoesthesia/hyperesthesia, allodynia—cold/mechanical.
Motor dysfunction: weakness, tremor, joint stiffness.
Psychologic al reactive disturbances: anxiety, depression, hopelessness.
Diagnostic tests alluded to included autonomic testing, sensory testing, motor strength testing, sympathetic block, and bone scan.
Staging: Because clinical cases present with qualitative differences in pain intensity and clinical features that are not necessarily time-dependent or stimulus intensity-dependent, staging as a component of the definition should not be used. Patients should be graded according to the intensity of their presenting features, as being mild, moderate, or severe in each of the categories of sensory, autonomic, and motor changes.
Having established the SIG-Pain and the Sympathetic Nervous System, its first charge was to convene the Consensus Conference in Orlando that was held in conjunction with the American Pain Society Annual Meeting (1993). Participants of this meeting are shown in figure 1.
The object of this conference was to build on the previous consensus statement in box 2, ultimately leading to the proposal to replace the term RSD with the descriptor c omplex r egional p ain s yndrome (CRPS). Also agreed was an understanding that the term SMP would describe pain that is mediated by the SNS, a term that was introduced by Roberts.18 SMP is not synonymous with CRPS. These changes were accepted by the IASP and were published in the second edition of the Classification of Chronic Pain: descriptions of chronic pain syndromes and definitions of pain terms in 199433 and in Stanton-Hicks et al (box 3).34
Proposed new diagnostic criteria
General definition of the syndrome.
CRPS describes an array of painful conditions that are characterized by a continuing (spontaneous and/or evoked) regional pain that is seemingly disproportionate in time or degree to the usual course of any known trauma or other lesion. The pain is regional (not in a specific nerve territory or dermatome) and usually has a distal predominance of abnormal sensory, motor, sudomotor, vasomotor, and/or trophic findings. The syndrome shows variable progression over time.
There are two versions of the proposed diagnostic criteria: a clinical version meant to maximize diagnostic sensitivity with adequate specificity, and a research version meant to more equally balance optimal sensitivity and specificity.
Proposed modified clinical diagnostic criteria for CRPS.
Continuing pain, which is disproportionate to any inciting event.
Must report at least one symptom in three of the four following categories:
Sensory: reports of hyperesthesia and/or allodynia.
Vasomotor: reports of temperature asymmetry and/or skin color changes and/or skin color asymmetry.
Sudomotor/Edema: reports of edema and/or sweating changes and/or sweating asymmetry.
Motor/Trophic: reports of decreased range of motion and/or motor dysfunction (weakness, tremor, dystonia) and/or trophic changes (hair, nails, skin).
Must display at least one sign* at the time of evaluation in two or more of the following categories:
Sensory: evidence of hyperalgesia (to pinprick) and/or allodynia (to light touch and/or deep somatic pressure and/or joint movement).
Vasomotor: evidence of temperature asymmetry and/or skin color changes and/or asymmetry.
Sudomotor/Edema: evidence of edema and/or sweating changes and/or sweating asymmetry.
Motor/Trophic: evidence of decreased range of motion and/or motor dysfunction (weakness, tremor, dystonia) and/or trophic changes (hair, nails, skin).
There is no other diagnosis that better explains the signs and symptoms.
Proposed modified research diagnostic criteria for CRPS.
Continuing pain, which is disproportionate to any inciting event.
Must report at least one symptom in each of the four following categories:
Sensory: reports of hyperesthesia and/or allodynia.
Vasomotor: reports of temperature asymmetry and/or skin color changes and/or skin color asymmetry.
Sudomotor/Edema: reports of edema and/or sweating changes and/or sweating asymmetry.
Motor/Trophic: reports of decreased range of motion and/or motor dysfunction (weakness, tremor, dystonia) and/or trophic changes (hair, nails, skin).
Must display at least one sign* at the time of evaluation in two or more of the following categories:
Sensory: evidence of hyperalgesia (to pinprick) and/or allodynia (to light touch and/or deep somatic pressure and/or joint movement).
Vasomotor: evidence of temperature asymmetry and/or skin color changes and/or asymmetry.
Sudomotor/Edema: evidence of edema and/or sweating changes and/or sweating asymmetry.
Motor/Trophic: evidence of decreased range of motion and/or motor dysfunction (weakness, tremor, dystonia) and/or trophic changes (hair, nails, skin).
There is no other diagnosis that better explains the signs and symptoms.
Diagnostic criteria derived at the Consensus Meeting held in Orlando in 1993, in conjunction with the American Pain Society Annual Meeting. Box 3 was modified from CRPS: Current Diagnosis and Therapy. (Eds) P Wilson, M Stanton-Hicks, RN Harden. IASP Press, Seattle, 2005.
*A sign is counted only if observed at the time of diagnosis.
CRPS, complex regional pain syndrome.
Experience with the standardized IASP criteria has vindicated its value as a tool that would promote understanding and foster treatment, and more specifically it would become a catalyst for basic and clinical scientists to address its pathophysiology and most important the treatment of patients with CRPS. However, it was clearly recognized that the 1994 IASP criteria were inadequately specific, and as a result were responsible for overdiagnosis of the syndrome. To address this concern, the IASP criteria were then subjected to principal component analysis of factors that were associated with the signs and symptoms of CRPS. The data drawn from 123 patients with CRPS in several countries were then subjected to statistical analysis.35
Internal validation results of this factor analysis are shown in table 1.
External validation was tested using discriminant function analysis of 117 patients with CRPS meeting the IASP criteria and 43 patients with neuropathic pain.36 While sensitivity was high (0.98), specificity was extremely poor (0.36). After extensive discussions at two consensus meetings, one in Cardiff in 2000 and the second in Budapest in 2004, new criteria were developed which explicitly included both symptoms and signs that employed a change to the decision rule requiring 2 of 4 sign categories and 4 of 4 symptom categories to be positive. This change yielded a sensitivity of 0.70 and a specificity of 0.94. The change would also have the greatest likelihood of realizing a correct diagnosis and would, for research purposes, enhance patient selection given its high level of specificity. For the clinician, merely relaxing the criteria to 2 of 4 sign categories and 3 of 4 symptom categories still retained a reasonable sensitivity of 0.85 and a specificity of 0.69, a satisfactory compromise, a new gold standard!
These changes to the diagnostic criteria which would replace the original 1994 IASP criteria were submitted to the IASP. The proposal was accepted in 2007. However, before any change of the language describing the new diagnostic criteria could be envisaged for inclusion in the “ IASP Classification of Chronic Pain Terms, it would have to undergo further statistical validation. On successful completion of this additional validation, the diagnostic criteria now referred to as the Budapest (IASP) Criteria were accepted37 and were acknowledged by the IASP in 2012.31 38
Because of the heterogeneous nature of CRPS in terms of its multiple pathophysiologies and seemingly endless mechanisms underlying its clinical presentation, a diagnosis conforming to the above criteria could only be satisfied by strictly adhering to the adopted criteria. What was missing, however, was a simple tool that could measure the impact of the syndrome in terms of its severity and one which could be used to report change resulting from disease progression or the result of treatment. Certainly there are many validated psychosocial scores, measures of function, and pain scores that can be recorded at onset and during the course of CRPS. Many attempts have been made to develop a severity score. One of the most successful of these has been the Impairment Level SumScore.39 The score includes the Visual Analog Scale (VAS), McGill Pain Questionnaire, temperature asymmetry by infrared (IR) thermometry, edema by volume measurement, and goniometry. As an outcome measure, this score has merit. It does not however measure a number of diagnostic features that are included in the Budapest Criteria. Based on a multisite international study, using the Budapest Criteria as a template, a cross-sectional design was used to determine whether a continuous-type CRPS Severity Score (CSS) could be developed. The results yielded a scale showing high internal consistency and one which corresponded well to the Budapest diagnostic criteria.40
In order to validate these initial CSS results, 156 patients who met the IASP CRPS criteria in six countries were divided into two groups, one termed “stable” CRPS (n=49), representing patients already undergoing treatment, and a second group termed “new” CRPS (n=107), in which new treatment would be initiated after baseline measurements were made. The choice of these two quite different patient populations would challenge the CSS tool to distinguish a greater change in severity between the more dynamic “new” CRPS group and the likely slower rate of change in severity in the “stable” CRPS group.37 This assumption was borne out from the results. Scoring was based on the presence or absence of signs and symptoms—(coded 1/0); the IASP diagnostic criteria have eight symptoms and eight signs, representing two signs and two symptoms for each of the four diagnostic factors (table 1, box 4).
CRPS Severity Score (CSS)
Symptoms—patient report.
Continuing disproportionate pain.
Allodynia/Hyperalgesia.
Temperature asymmetry.
Skin color asymmetry.
Sweating asymmetry.
Asymmetric edema.
Trophic changes.
Motor changes.
Signs—on examination.
Hyperalgesia—pinprick.
Allodynia.
Temperature asymmetry.
Skin color asymmetry.
Sweating asymmetry.
Asymmetric edema.
Trophic changes.
Motor changes.
A maximum score of 16 (each sign and symptom being counted as 1) is possible.
A “smallest real CSS difference value” was calculated indicating that a change of ≤4.9 scale points meant a real change in CRPS severity. For example, in the CSS revalidation study, the mean CSS value across both groups at baseline was 11.4. Therefore, any intervention that resulted in a decrease of the CSS to 6.5 or lower would represent real clinical improvement.
With the limitations mentioned below, the authors observed more variation in CSS values over time in new patients compared with those in the stable treatment group. The change in CSS values was associated with changes in pain and functional status and agreed with the clinical impression that one has with such patients. These findings corroborate those from the earlier study by Harden et al 37 in 2010 and support the validity of using CSS as a monitor of disease progression as well as being an outcome measure. Limitations of the study relate to the labile course so often a feature of CRPS, vis-à-vis, temperature and color changes. The subjective nature of clinical data obtained from standardized assessment techniques is always a challenge. Strenuous attempts to promote the veracity of data collection were enhanced by written procedures and videos to help each investigator.
Epidemiology and risk factors
Two population-based studies conducted under quite different conditions have yielded a very different incidence of CRPS. The first, conducted in Olmsted County, Minnesota, by Sandroni et al 41 yielded an incidence of 5.5 cases per 100 000 person-years. The second, drawn from a heterogeneous population in the Netherlands of 100 000 by de Mos et al,42 found an incidence of 26.2 cases per 100 000 person-years. Women are three to four times more likely than men to develop CRPS, with expression of the condition in the upper extremity occurring in 60% of cases. Sprain and fracture are the most common causes.43 Migraine, asthma, and the concomitant use of ACE inhibitors are risk factors.44 In a ratio of 5 to 6:1, women may develop a more severe phenotype, and in addition the 20% who are more likely to present with cold CRPS also have a poorer prognosis in terms of progression and pain severity45 (see the Sympathetic Nervous system: mechanisms section ).
Wrist fracture as a source of CRPS has been studied by several investigators. The most recent longitudinal study by Moseley et al 43 determined an incidence of 3.8% in 1549 patients over a period of 4 months. The patients were managed non-surgically. Their initial assessment occurred within 1 week of their injury. Pain measured on a VAS was highly discriminative (c index 0.98). A pain rating of 5 or higher was highly predictive for the development CRPS.
Another risk factor for the development of CRPS is immobilization.45 Studies of immobilization in human volunteers have shown that they develop sensitivity to heat and pressure but without any onset of pain.
Many case reports of families with siblings who developed CRPS have appeared in the literature. A genetic predisposition involving the human leukocyte antigen (HLA) was first described by Mailis and Wade in 1994.46 Their study found a twofold increase in the A3, B7, and DR2(15) Major Histo Compatibility (MHC)antigens compared with control frequencies. Five of six DR2(15)-positive patients were resistant to treatment. These investigators noted that RSD (CRPS) is the third neuroimmune disorder besides multiple sclerosis and narcolepsy to be associated with the DR2(15) antigen. Two subsequent studies have identified an association with HLA-DR13 in patients with CRPS who have subsequently developed dystonia.47 48 A contribution of HLA alleles in 150 patients with CRPS who had fixed dystonia (HLA-A, HLA-B, and HLA-DQB1) was identified. The use of genome-wide expression profiling has found certain genes in CRPS that are differentially expressed,42 including HLA-related genes.49 While all of these reports suggest that there seems to be a genetic predisposition for some patients with CRPS, the observations have been obtained from small cohorts and some data have been contradictory. Nevertheless, the epidemiologic findings and accumulating genetic information are helping to explain just how complex is the pathophysiology exemplified by this syndrome.
Sympathetic nervous system mechanisms
As already discussed in the prologue, there is a large body of presumptive evidence supporting a predictive role of the SNS in the initiation and possible maintenance of CRPS. There is also good evidence that not only ipsilateral SNS dysfunction is observed but other body regions also demonstrate altered SNS activity.50 Concrete evidence is lacking and there are few prospective randomized controlled studies of high quality in a sufficiently large patient population to support the routine use of local anesthetic sympathetic blocks to achieve consistent, beneficial treatment outcomes. 43 The relief of pain after a sympathetic block, defined as SMP, implies a direct or indirect interruption of the SNS. Gradl et al have found support from their studies that dysfunction of the SNS is consistent with the onset of CRPS and involves the ipsilateral extremity and the entire body.51 52 While they did not allude to its continuing involvement in the ongoing disturbance at the time, their most recent paper provides more evidence supporting a possible fundamental abnormality of the autonomic nervous system in patients with CRPS.53 After the acute phase, vasoconstrictive activity tends to return to normal, but the extremity will in most cases become cold as we now understand from a number of factors including the upregulation and increasing density of alpha-1 adrenoceptors (α1-ARs).54 55 Beta-2 adrenoceptors activated by norepinephrine have also been shown to liberate interleukin-6 (IL-6) which sensitizes nociceptors, both amplifying CRPS symptoms and increasing vasoconstriction, another indirect effect of the SNS. It should be noted that skin temperature, while a reflection of vasomotor activity, is during the acute phase mostly warm but changes to cold skin with chronicity.22 Twenty percent of patients during early CRPS present with a cold extremity.56 There is evidence that α1-ARs expressed on nociceptors are also activated by circulating norepinephrine, causing increased pain that may be relieved by sympathetic blocks. Gibbs et al,57 Dawson et al,55 and Schattschneider et al 58 suggest that upregulation of α1-ARs might be the basis for SMP. The recent paper by Drummond et al,59 which described the upregulation of α1-ARs, also underscores an indirect role for the SNS in CRPS. In a study by Wasner et al,60 patients with CRPS 1 underwent maximum activation of the SNS by whole body cooling and experienced severe spontaneous pain and mechanical hyperalgesia that would be described as SMP if relieved by a sympathetic block.
In addition to coupling between SNS and somatic fibers innervating superficial tissues, the same coupling may also occur with somatic fiber innervation in deep tissues like bones, joints, tendons, and muscles.58 Central reorganization may also be responsible for the failure to relieve pain after a successful sympathetic block (a temperature rise to 35°C+). Such a situation would be described as Sympathetically Independent Pain (SIP).20 61 Obviously this consequence is independent from of the peripheral SNS.62 One should also bear in mind that vasomotor activity in the microcirculation is also a function of endogenous control. Endothelial dysfunction, particularly the nitric oxide:endothelin ratio, is reactive to changes in SNS activity and humoral responses to ischemia/hypoxia.63
To summarize the foregoing, a direct link between SNS and the primary afferent neuron (nociceptors) in patients with CRPS is not yet clarified. One study by Campero et al,64 who used a microneurographic technique, was not able to demonstrate a link between SNS activity and increased nociception. It is very likely that indirect mechanisms such as those already described above are responsible for the activation of nociceptors. In addition, inflammatory cytokines such as tumor necrosis factor (TNF)-alpha and IL-6, together with the activation of macrophages, can sensitize nociceptors.65
Mention should be made of the noradrenergic system in the central nervous system (CNS). This becomes particularly active in high-stress states and alerts the brain to potential threats. Activation of alpha-2 adrenoceptors in the dorsal horn inhibits the release of excitatory neurotransmitters from nociceptive afferents, thereby modulating pain. This antinociceptive system can be compromised after peripheral nerve injury.66 The noradrenergic system under such circumstances can facilitate pain, in complete contrast to its homeostatic function. Also, activation of α1-ARs will compromise inhibitory responses. In CRPS, the high level of nociceptive activity can therefore rapidly overwhelm inhibitory defenses. This topic will be reviewed in the CRPS and the Central Nervous System section, and in particular the function of the locus coeruleus.
Cytokines and the inflammatory response
Under normal circumstances, tissue trauma is associated with cytokine release. These inflammatory mediators in turn excite nociceptors, which depending on circumstances may initiate peripheral nerve sensitization with the release of nerve growth factor (NGF).66 67
This in turn is associated with the release of inflammatory neuropeptides such as substance P (SP) and calcitonin gene-related peptide (CGRP) from small fiber primary afferents. This set of events results in the classical signs of edema, redness, and an increase in skin temperature, hallmarks of neurogenic inflammation.68 In this regard, SP, endothelin-1, CGRP, bradykinin, and inflammatory cytokines are involved in the regulation of post-traumatic inflammation. Both SP and CGRP could be culprits behind the increase in temperature and edema in patients with “warm” CRPS. The recent review by Birklein et al 69 puts these aspects into focus.69 This may explain why ACE inhibitors, which among other actions play a role in the cleavage of neuropeptides, have been associated with the increased incidence of CRPS after bone fracture.45 The earlier observations of Schurmann et al (1996)51 using laser Doppler flowmetry demonstrated an increase in perfusion which they attributed to CGRP released from cytokine-sensitized nociceptors. Elevated serum concentrations of CGRP and substance P were found ipsilateral in patients with CRPS compared with healthy controls, while protein extravasation only occurred on the same side. It seems that in CRPS the exaggerated neurogenic inflammation results from either an impairment of neuropeptide inactivation, by an increase in target receptors, or both.70 We might certainly be nearer to one form of mechanism-based therapy if we just chose the mast cell as a target for example. If one were in a position to “turn-off” neuropeptide signaling (eg, CGRP and SP), or for instance prevent tyrosine activity in some proteins from acting as receptors on mast cells thereby initiating degranulation and the production of inflammatory cytokines, it might be possible to temper the inflammatory response in CRPS.71 A further effect of substance P is the increase in excitatory cytokines that are seen in CRPS. Substance P stimulates keratinocytes to produce inflammatory cytokines in vitro and in skin biopsies from rats and humans.72 73 The use of artificial skin blisters in patients and human volunteers has enabled us to understand the inflammatory response over the time-course of acute and chronic CRPS. These findings help to explain the clinical picture of enhanced extravasation, edema, and temporal measurement of cytokine concentrations: interleukins IL-6, IL-1Ra, the cytokine TNF-alpha, monocyte chemoattractant protein-1, and CC chemokine ligand 4 (CCL4) which are all elevated in chronic CRPS. While serum concentrations of TNF-alpha, IL-6, and IL-8 are greater ipsilateral to the lesion during the first 3 months of CRPS, anti-inflammatory cytokines such as IL-4, IL-10, and growth factor β-1 are decreased. Mechanical hyperalgesia as a symptom reflects sensitization of the CNS. Added to this, inflammatory cytokines may, in conjunction with the upregulation of the microglia, also contribute to central sensitization.74 The immune system is taking on an increasingly important role. Whether this is a predisposing factor or reactive to the inflammatory process of CRPS remains to be determined. The study by Ritz et al 75 in which a particular subgroup of monocytes (CD14+,CD16+) in long-standing CRPS were elevated but the anti-inflammatory cytokine IL-10 level was low addresses this point. Earlier work by Uçeyler et al 76 also noted this relationship. None of these interactions are reflective of a cellular response to injury. Laboratory values like blood count, T-cell concentration, autoimmune antibody concentration, sedimentation concentrations, and antigen concentrations are all normal. Biopsies show no inflammatory infiltration.77
Autoimmunity
An autoimmune hypothesis for CRPS has been a consideration during the past decade when serendipitous observations noted an improvement in CRPS after patients were treated with intravenous immunoglobulin (IVIG) for unrelated medical conditions. On the completion of a seemingly successful pilot study, Goebel et al 78 79 undertook a prospective randomized controlled study with IVIG in subjects with CRPS. The results exemplify how difficult it can be to undertake this type of study which did not demonstrate any significant response over placebo, a result that rests on many factors not least of which could be dose-related, study design, or patient-related, to name but a few.78 79 Translational studies using animals may be easier to control. Using an IgG-transfer-trauma model, Tékus et al 80 explored the response of animals which underwent mild tissue trauma after being exposed to IgG taken from humans with CRPS. The animals exhibited signs of central sensitization (hyperalgesia) and increasing pain behavior, an important role’ for IgG antibodies. In another animal (mouse) model in which CRPS symptoms had been engendered by fracture/cast immobilization, the response was less severe in those mice who were treated with anti-CD20 (polyclonal antibody) or in mice lacking B cells in comparison with wild-type (WT) mice that underwent the same procedure. However, deposits and complement activation were found in the skin and on the sciatic nerves of WT (fracture/cast) mice.81 The results suggest that CD20 B cells produce antibodies that promote CRPS-like changes. Several clinical observations have noted that a number of patients had IgG and IgM profiles concordant with past infections of chlamydia, parvovirus and Campylobacter. Some of these cases suggest that an autoimmune neuropathy may relate to cross-reactivity between infection-related antibodies and self-antigens.
A number of SNS neurons have been described as targets for autoantibodies in patients with CRPS, although there have been no reports of autoimmunity in patients with other types of peripheral neuropathy. Using immunohistochemical techniques and the special fluorescence-assisted cell sorting analysis, Kohr et al 82 not only explored the above targets, but using a beating cardiomyocyte preparation found that nearly half of the patients with CRPS had autoantibodies to the M2 muscarinic and β-2 adrenoceptor. Because of their immune function, sclerocytes are a target for gene expression studies with recombinant proteins. Using their murine model (fracture/immobilization) as a window for autoimmune reactions in the skin, Tajerian et al 83 employed liquid crystal mass spectrometry to quantitatively determine the levels of the target protein Krt16. This is a biomarker for other autoimmune conditions like rheumatoid arthritis and psoriasis. Krt16 was increased in both mRNA and in protein levels of mouse skin. The increased binding at the mouse skin suggests autoimmune reactivity in animals and humans with CRPS—it was highly reactive in sera with IgM from fracture in mice and humans. These findings also bolster those observations of associations with the HLA system already mentioned above.
Consideration of CRPS as an autoimmune syndrome is a daunting proposition that presents huge opportunities for experimental research and the development of specific immunotherapies, while at the same time adding an element of caution because of the inherent problems of immunosuppression and its long-term adverse consequences. Such therapies are unlikely to have a direct effect on analgesia, although by modifying the disease process there may be an indirect analgesic affect. At least there is for now a rational basis remaining to be confirmed, for employing novel therapies such as IVIG, and immune modulators like infliximab, rituximab and other biologics as adjuncts with current therapies that address CRPS-related autoimmunity.
CRPS and the central nervous system
There is cumulative evidence of functional and structural changes in the CNS of patients with chronic pain and CRPS.84 Central sensitization is one consequence of these changes, and while only incompletely understood it is accepted that disinhibition (of spinal and trigeminal neurons) and the promotion of nociceptive activity extend from the dorsal horn to the rostral medulla.85 86 Functional changes include motor, autonomic, and cortical representation. Motor symptoms that are found in a significant percent of patients with CRPS include paresis, tremor, dystonia, myoclonus, and exaggerated tendon reflexes. All or only some of these may be present at any one time.87 Other consequences of central sensitization are its impact on affective-motivational areas of the brain such as the amygdala. GABA-A receptors in the amygdala may be important for both sensory and affective pain. 88 Many CNS changes associated with CRPS are a distortion of normal cutaneous sensations in the CRPS-affected limb and the failure to inhibit or suppress nociception with the subsequent upregulation of thalamocortical nociceptive networks—a circulus vitiosusGlutamate receptors are intimately involved in the transmission of nociceptive activity and once sensitized can maintain this traffic even without further input from the source of injury source. So chronic pain, whether neuropathic or hyperalgesia, and allodynia resulting from central sensitization can expand to involve adjacent regions.89 A case report by Del Valle et al 90 describes the increased levels of glutamate in the cerebrospinal fluid (CSF) and spinal microgliosis in a cadaveric subject with CRPS. The increase of inflammatory cytokines, TNF-alpha, IL-1β, IL-6, CCL2, and NGF in the CSF 4 weeks after hind-limb fracture is a consequence of the upregulation of substance P and CGRP. When the animals received selective intrathecal antagonists, pain behavior receded, thereby supporting the hypothesis that facilitated neuropeptide signaling upregulates inflammatory spinal mediators.91
The rodent fracture model of CRPS continues to yield information regarding both structural and biochemical changes in the amygdala, perirhinal cortex, and hippocampus. The recently proposed mechanism, “hyperalgesia priming,” might possibly explain why some patients after a relatively small and transient insult will develop chronic pain. This neuroplastic change can induce a hyper-responsive state in nociceptors when confronted with future mild insults. The epsilon isoform of protein kinase C (PK Cepsilon) would seem to be the principle factor in this response.92 Recently, Golmirziae et al (2016),93 ,94 using t he validated (2011) Fibromyalgia Survey Criteria self-report as an instrument of widespread body pain in a cross-sectional analysis of 160 patients with CRPS, found a direct relationship between the duration of CRPS and onset of a central sensitization phenotype (widespread pain). Although the data are retrospective, including chart review, and are from a single center, their importance suggests that early diagnosis and pathology-centered treatment could obviate the development of centralized pain.
Intravenous ketamine is proving to be potentially useful as a therapy in patients with CRPS. As an N-methyl-D-aspartate (NMDA) receptor antagonist when given as a continuous infusion over a period of several days, a significant number of patients will have pain relief for at least 11 weeks.94 These results suggest that ketamine may provide extended relief of pain by desensitizing the NMDA receptor and by altered astrocyte activation.95 It is not uncommon for patients with CRPS to suffer cognitive impairment, labile personality, anxiety, and depression. While speculative, such CNS effects may result from the reduction in GABAergic or NMDA-mediated cortical neuroplasticity96, The foregoing symptoms are not necessarily permanent, but may merely be a part reflection of the acute phase of CRPS and in some cases do respond to multiple intravenous treatments with ketamine.
CRPS-associated changes are found in several brain centers, including the thalamus, S1 and S2 cortices (sensory processing), cingulate and amygdala (emotional function), and perirhinal and hippocampus (memory). Imaging studies in both adults and children have demonstrated similar changes in connectivity between these brain regions that are concordant with the clinical course of CRPS in response to different therapeutic interventions.97– ,100
Originally, the movement disorder associated with CRPS was not considered to be a frequent or reliable sign. Motor impairment already mentioned above is one of the most frequent components in this diagnosis.101 This prospective study by Lebel et al noted a frequency of tremor during the acute phase of 82%, which declined to 44% at 12 months. After tremor, dystonia is the most common movement disorder with posturing involving the wrist and fingers in the upper extremity, and in the lower extremity plantar flexion or inversion being the most common signs. Dystonia can occur in the acute phase but may take place throughout the course of this disease. Although the disease may spread to involve more than one extremity, so may dystonia involve more than one limb. This is another example of maladaptive neuronal plasticity that is so characteristic of this clinical disorder. Dystonia is not associated with inflammation and is more than likely a CNS disorder involving the basal ganglia. Further support for this idea is the positive response to baclofen, a gamma-aminobutyric acid (GABA) type B receptor agonist. Presynaptic GABAergic inhibition has been demonstrated in CRPS patients with dystonia. 102
Spread of CRPS from its original site was reported by Veldman et al in 1993.26 Until larer epidemiologic studies are done, the most recent study by van Rijn et al 103 in 185 patients suggests that spread occurs in up to 48% and contralateral spread is more likely than ipsilateral spread. The frequency of four-limb involvement was 29% and in most cases attended a subsequent injury.
Patients with CRPS or stroke have striking similarities in the way they perceive their affected extremity or side. For example, they may be able to feel touch on the ipsilateral extremity if they watch a mirror image of the unaffected side being touched. Similarly, if the unaffected limb is crossed to the ipsilateral side, any tactile information from the affected limb will be accepted as coming from the unaffected limb.104 This is the basis for mirror image therapy which has proven potentially useful as a rehabilitation tool. A recent study on 10 patients (6 women) with upper extremity cold-type CRPS showed that temperature as a reflection of vasomotor control is modulated by which side of the body relative to the body midline the arm is placed. That is, the disruption of thermoregulation of either hand is not determined by anatomic factors that apply to the location or configuration of the hand but rather to cortical representation of the hand’s location within a body-centered frame of reference. The investigators conducted their study using prism glasses that changed the field of view by 20°—in effect displacing the visual field sufficiently to cross the anatomic midline. The authors concluded that cortical mechanisms are responsible for encoding the perceived location of the limbs in space, and in a disease such as CRPS are responsible for the disruption of multiple efferent mechanisms. The study suggests that while further research is required, modulation of spatial perception could be an effective treatment for not just unilateral CRPS but for a wide range of neurologic disorders that are characterized by dysfunction in both spatial perception and thermal regulation.95
Hemisensory impairment ipsilateral to the affected CRPS extremity was already discussed almost 20 years ago.105 Hemisensory impairment is also a characteristic of the so-called “neglect syndrome.”106 107 Patients with CRPS, however, may exhibit neglect-like symptoms merely to avoid the pain of hyperalgesia or allodynia when touched. New magnetic source imaging techniques of the cortex have demonstrated marked reorganization of the somatotopic map of the primary somatosensory cortex (S1), contralateral to the affected extremity, of patients with CRPS.108 The S1 cortex representation was smaller than that from the contralateral (normal) extremity. Reversal of these changes occurs with treatment together with a reduction in pain and allodynia (figure 2).109
Similar contralateral cortical changes affecting the motor cortex (M1) have been demonstrated by transcranial magnetic stimulation in which decreased inhibitory activity was replaced by excitatory activity in patients with CRPS. Abnormalities were also noted on the ipsilateral motor cortex, confirming some collateral effect of CRPS from the affected extremity. A study in which finger tapping of the affected extremity demonstrated substantial reorganization of the primary motor circuitry was associated with greater activation of the primary motor and supplementary motor cortices than was found in controls.110 As a corollary, increased activation of the ipsilateral motor cortex and the magnitude of motor dysfunction coincided with the activation of parietal cortices, supplementary motor, and motor cortices, all together underscoring the fact that widespread changes of CNS function can occur in people with CRPS.
One important brain structure, the locus coeruleus is involved in hemilateral pain modulation and responds to nociceptive and alerting stimulae.111 The fact that sensitivity to painful pressure on the ipsilateral upper extremity was replicated on the ipsilateral forehead in patients with CRPS but not in other patients with neuropathic pain conditions suggests that the locus coeruleus may well be involved. After peripheral nerve injury, the spinal antinociceptive noradrenergic mechanism that projects from the locus coeruleus could transform into a pronociceptive mechanism.112 This study would suggest that treatment of the mechanism involving α1-ARs could benefit certain patients with CRPS.
The acute to chronic continuum of CRPS
When Richards wrote his centennial review on causalgia in 1967,1 his motivation was to highlight a medical event, namely the publication of the book “Gunshot Wounds and Other Injuries of Nerves” in 1864 by Silas Weir Mitchell, George Morehouse, and William Keen.2 This, at a time when many other events were being celebrated 100 hundred years after the Civil War, highlighted the fact that among other injuries, one such injury was most significant because it yielded severe, intractable, burning pain long after the initial injury. Since then, half a century later, we now have a far better understanding of the pathophysiology that underlies this and other inflammatory syndromes. The specific nature that distinguishes CRPS 1 and 2 from other trauma-related responses is the dynamic course, its many mechanisms, and the plethora of signs and symptoms with which it is associated. In the acute warm phase, the extremity is exquisitely sensitive, frequently edematous, and hot. Over a variable period of time, the condition will generally progress to a cold (chronic) phase, leaving only traces of its apparent inflammatory genesis while retaining pain that is usually burning, commonly severe, and with a variability that does not belie its origins. Among other symptoms, the persistent pain during the chronic phase is associated with the typical attributes of chronic pain that include depression, but also cognitive changes and the reactive symptoms of anxiety and hopelessness.113 114 The fact that wrist and ankle fractures are the most frequent cause of CRPS type 1, development of the tibial fracture model has given us an insight into the pathophysiology of these conditions and enabled us to explore the temporal changes in both acute and chronic phases. For example, mice only exhibit signs of edema, pain, and temperature increase during the acute phase. These changes are associated with increased chemokine signaling,115 whereas during the chronic phase, when neuroplastic mechanisms are dominant, the mouse model as a preclinical tool has demonstrated a therapeutic response to ketamine in the chronic phase, but not during the acute phase.116 These observations have a significant bearing on the selection of treatment. Knowing the nature of the inflammatory response during the acute phase, one could suggest that anti-inflammatory and immunologic measures might be the most efficacious therapies, perhaps with the singular advantage of arresting the process before it can induce central sensitization or take on other chronic characteristics. Similarly, with a greater understanding of its pathophysiology, it may be possible to abort the syndrome using selective treatments that anticipate the onset of CRPS not just from clinical features but from specific biomarkers. Also, a better characterization of the genetic underpinnings might eventually identify those individuals who are more likely to succumb to the syndrome after injury.
Conclusions
CRPS, once a “wastebasket” diagnosis, now represents a composite of SNS dysfunction, neurogenic inflammation, a novel form of immune and autoimmune disorder, likely with genetic overtones, and therefore one with many different mechanisms. The CNS changes associated with this syndrome obviously contribute to the chronic sensory features and behavioral symptoms, most of which are now amenable to appropriate treatment. Lastly, contemporary research is providing a more sophisticated window into the pathophysiology of the SNS and the manner in which it can be involved in the genesis of CRPS. Furthermore, given what we now know, it may not be necessary to maintain the distincion between the two types of CRPS. One should feel optimistic that 150 years after Mitchell, we are now on a path toward the successful management of CRPS.
References
- 1.↵
- 2.↵
- 3.↵
- 4.↵
- 5.↵
- 6.↵
- 7.↵
- 8.↵
- 9.↵
- 10.↵
- 11.↵
- 12.↵
- 13.↵
- 14.↵
- 15.↵
- 16.↵
- 17.↵
- 18.↵
- 19.↵
- 20.↵
- 21.↵
- 22.↵
- 23.↵
- 24.↵
- 25.↵
- 26.↵
- 27.↵
- 28.↵
- 29.↵
- 30.↵
- 31.↵
- 32.↵
- 33.↵
- 34.↵
- 35.↵
- 36.↵
- 37.↵
- 38.↵
- 39.↵
- 40.↵
- 41.↵
- 42.↵
- 43.↵
- 44.↵
- 45.↵
- 46.↵
- 47.↵
- 48.↵
- 49.↵
- 50.↵
- 51.↵
- 52.↵
- 53.↵
- 54.↵
- 55.↵
- 56.↵
- 57.↵
- 58.↵
- 59.↵
- 60.↵
- 61.↵
- 62.↵
- 63.↵
- 64.↵
- 65.↵
- 66.↵
- 67.↵
- 68.↵
- 69.↵
- 70.↵
- 71.↵
- 72.↵
- 73.↵
- 74.↵
- 75.↵
- 76.↵
- 77.↵
- 78.↵
- 79.↵
- 80.↵
- 81.↵
- 82.↵
- 83.↵
- 84.↵
- 85.↵
- 86.↵
- 87.↵
- 88.↵
- 89.↵
- 90.↵
- 91.↵
- 92.↵
- 93.↵
- 94.↵
- 95.↵
- 96.↵
- 97.↵
- 98.↵
- 99.↵
- 100.↵
- 101.↵
- 102.↵
- 103.↵
- 104.↵
- 105.↵
- 106.↵
- 107.↵
- 108.↵
- 109.↵
- 110.↵
- 111.↵
- 112.↵
- 113.↵
- 114.↵
- 115.↵
- 116.↵
Footnotes
Presented at This work was presented in part as the American Society of Regional Anesthesia and Pain Medicine’s 2017 Bonica Award lecture. The award is named for John J. Bonica, MD, who championed the collaboration of multidisciplinary specialists in the evaluation and treatment of patients with pain.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Patient consent for publication Not required.
Provenance and peer review Not commissioned; externally peer reviewed.