Complex Regional Pain Syndrome

Michael T. Massey

Robert Norman Harden

Complex regional pain syndrome (CRPS) is the current taxonomy for the syndrome previously known by various names, including reflex sympathetic dystrophy (RSD), causalgia, Sudeck’s atrophy, shoulder-hand syndrome, neuro-algodystrophy, and reflex neurovascular dystrophy, etc. It was originally recognized as a distinct pain syndrome seen among Union veterans of the War Between the States following traumatic nerve injury (“causalgia”).¹ It is an extremely heterogeneous disease with inflammatory, autoimmune, sympathetic, and neuropathic features that change over time. It is usually a chronic disease that involves a full measure of biopsychosocial features, and it can be significantly disabling.

Epidemiology

Epidemiologic data regarding CRPS in the general population are limited, although three large-scale studies are available. Sandroni et al.² reported an incidence of 5.46 new cases of CRPS type I per 100,000 annually (see Table 25.1 for distinction between type I and type II).² A larger study reported an incidence as high as 26.2 new cases per 100,000 annually.³ Based on this reported incidence, over 50,000 new cases of CRPS type I could be anticipated annually in the United States alone.⁴ However, both of these studies used the 1994 International Association for the Study of Pain (IASP) diagnostic criteria.⁵ These results may be inflated as the new 2012 IASP criteria, being more objective, have been demonstrated to reduce diagnostic rates by about 50% when compared to the old 1994 IASP criteria.⁶ Most recently, a retrospective analysis from a Nationwide Inpatient Sample Database during 2007 to 2011 demonstrated 22,533 patients with a discharge diagnosis of CRPS type I sampled from 33,406,123 patients. Based on these data, the general prevalence of CRPS type I in an inpatient environment could be 67.4 cases per 100,000 patients. However, these numbers are limited to an inpatient environment and should not be generalized to patients or healthy individuals in an outpatient setting. Further, there was no relevant data indicating what criteria were used to make the diagnosis leaving validity of the diagnostic coding uncertain.⁷ For physicians making pain diagnoses, the incidence of CRPS in relevant at-risk populations (e.g., postfracture) may be more clinically relevant. Three recent large-scale prospective studies examine the incidence of CRPS postfracture.⁸^,⁹^,¹⁰ Beerthuizen et al.⁸ reported an incidence of 7.0% from 596 patients after acute fracture of the wrist or ankle and receiving conventional treatment. The highest rate of incidence occurred at about 3 months after the fracture.⁸ Moseley et al.⁹ reported an incidence of 3.8% within 4 months of a wrist fracture in 1,549 patients. Bullen et al.¹⁰ reported an incidence of 0.3% within 3 months of ankle fracture or postoperative treatment of ankle fracture in 300 patients. Several smaller prospective studies suggest that acute CRPS type I may develop in up to 11% to 18% of patients following fracture or total knee arthroplasty, although most cases resolve relatively quickly with conservative care.¹¹^,¹²^,¹³

TABLE 25.1 1994 IASP Diagnostic Criteria for Complex Regional Pain Syndrome

The presence of an initiating noxious event or a cause of immobilization
Continuing pain, allodynia, or hyperalgesia with which the pain is disproportionate to any inciting event
Evidence at some time of edema, changes in skin blood flow, or abnormal sudomotor activity in the region of pain
This diagnosis is excluded by the existence of conditions that would otherwise account for the degree of pain and dysfunction.

Type I: without obvious nerve damage (aka “reflex sympathetic dystrophy”)

Type II: with obvious nerve damage (aka “causalgia”)

Modified from Merskey H, Bogduk N. Classification of Chronic Pain: Descriptions of Chronic Pain Syndromes and Definitions of Pain Terms. 2nd ed. Seattle, WA: IASP Press; 1994.

Based on available epidemiologic data, fractures and sprains may be the most common events triggering CRPS. CRPS appears to be more common in the upper extremities, is more common in females, and is most likely to occur in the 50 to 70 years age range.²^,³

Pathophysiology

CRPS remains one of the most enigmatic and difficult to treat of all pain conditions. Although good clinical characterizations started to appear in the literature in the late 1800s,¹ a definitive pathophysiology remains to be determined, and it may be that there are a collection of distinct pathophysiologies, early or late, that are currently grouped under the rubric of CRPS. This lack of a precise, singular mechanism partially explains the relative lack of clinical progress to date.¹⁴ There is a great deal of discussion now of identification of clinically relevant subsets, both pathophysiologically and diagnostically. In the following section, animal and human models that have relevance to understanding CRPS are first described and then several pathophysiologic mechanisms that may contribute are reviewed.

ANIMAL MODELS

There are several animal models that may shed light on the mechanisms of CRPS, albeit indirectly. However, as always, one must be cautious extrapolating from animal models to human syndromes. The most useful animal model, best paralleling CRPS type II (causalgia) in which major nerve injury is a key feature, is the chronic constriction injury model in rat that was first described by Bennett and Xie.¹⁵ This model produces some behavioral features (e.g., guarding, disuse, vasomotor disturbance) that mimic some of the features of human CRPS (e.g., allodynia, hyperpathia, spontaneous pain) as well as some of the sympathetic abnormalities.¹⁵ This model has been utilized in the preclinical screening of pharmacologic interventions for CRPS. Another model that may be useful is the spinal nerve injury model,¹⁶ although there is some controversy as to whether this model actually produces “sympathetically maintained pain” (SMP) and how similar the resulting pain syndrome is to human CRPS.¹⁴^,¹⁷ A third model, involving partial ligation of the sciatic nerve, may also have some relevance to human CRPS type II.¹⁸ All three models likely generate ectopic activity at the site of injury and/or the dorsal root ganglion (DRG) which may (or may not) be responsive to sympathetic outflow and which may cause/maintain central sensitization.¹⁹
These models also result in various other phenomena putatively related to CRPS, such as sprouting of fibers in lamina II of the dorsal horn.²⁰^,²¹ All of the aforementioned models appear most relevant to understanding CRPS type II, although two less widely used models may be relevant to understanding CRPS type I (no major nerve injury evident). Models using tetanic electrical stimulation²² and an ischemic reperfusion injury²³ appear to produce a syndrome that mimics some of the features of CRPS type I in the absence of signs of major nerve injury.²² These animal models may help somewhat in understanding mechanisms of CRPS, but their ultimate value may be in screening pharmaceutical interventions.

HUMAN MODELS

Intracutaneous injection of capsaicin induces burning pain and cutaneous mechanical hypersensitivity (allodynia) and as such may represent a useful model to study certain features of CRPS. The capsaicin model demonstrates the impact of intense nociceptive stimulation of normal skin with the rapid development of areas of primary and secondary hyperalgesia, allodynia, wheal, and flare peripheral sensitization.²⁴^,²⁵^,²⁶ This model results in hyperalgesia to suprathreshold heat within the capsaicin-induced secondary hyperalgesic skin despite the absence of changes in heat pain threshold.²⁷ Another interesting human model of CRPS-like pain is the controlled heat injury model, which also leads to a zone of secondary hyperalgesia to suprathreshold heat.²⁸ These models and other lines of evidence corroborate a primary role of central sensitization in CRPS²⁹^,³⁰^,³¹; however, the overlapping regions of flare/vasodilatation and hyperalgesia may also suggest peripheral sensitization.³² Peripheral sensitization likely has a significant role in clinical CRPS. Polymodal C receptors may mediate the temperature allodynia,²⁷^,³³ polymodal A receptors the pinprick hyperalgesia,³⁴ and Aβ receptors the mechanical allodynia²⁶ in these models. Whole-body cooling (to induce sympathetic activation) and sympathetic block have shown no effects on features of either the capsaicin or heat injury models,²⁸^,³⁵ although phentolamine (an adrenergic antagonist) reduces the area of allodynia in the capsaicin model.³⁶ Given these mixed findings, the value of these models in understanding interactions between sympathetic nervous system (SNS) activity and pain in CRPS remains uncertain. Although some features of these models are concordant with signs and symptoms of CRPS, the usefulness of these normal nociception models in unraveling disease specific mechanisms in CRPS is limited to date. As with animal models, a primary use of these human models may be in the testing of certain interventions for CRPS.²⁷^,³⁷

INFLAMMATION

Clinical evidence indicates that some features of acute/early CRPS often attributed exclusively to sympathetic hypofunction (i.e., vasodilatation, swelling, and edema) could perhaps be better explained by an exaggerated localized inflammatory process.³⁸ Sudeck³⁹ was the first to propose this, along with a “patchy inflammatory osteoporosis,” and Goris⁴⁰ revitalized this idea. Consistent with inflammatory mechanisms, one experimental study demonstrated that 82% of “acute” CRPS patients exhibited a progressive accumulation of immunoglobulin in the affected extremity relative to the unaffected extremity compared to only 17% of “chronic” CRPS patients (in this study, Oyen et al.⁴¹ defined chronic as 5 months or more since onset). Analyses of joint fluid and synovial biopsies in CRPS patients have shown an increase in protein concentration, synovial hypervascularity, and neutrophil infiltration.⁴²^,⁴³^,⁴⁴ Free radicals, suspected to play a prominent role in inflammation and ischemic damage, recreate acute CRPS symptomatology in a rat model (vasomotor abnormalities, edema, and pain behavior), and this is the partial rationale for some investigators recommending free radical scavengers in acute CRPS patients.⁴⁵ Eisenberg et al.⁴⁶ demonstrated an increase in antioxidant levels in saliva and serum for 31 CRPS type I patients compared to 21, and they suggest these lab markers could be utilized as a potential screening tool for CRPS. From these lines of evidence, it is logical to conclude that an inflammatory component is likely in CRPS type I, particularly in the early phase.⁴⁵^,⁴⁷ A shift toward a pro-inflammatory cytokine profile in patients with CRPS suggests a potential pathogenic role for these compounds in the generation of pain and other symptoms.⁴⁸ In plasma, no alterations in inflammatory mediators were observed in CRPS patients. However, in “blister fluid” obtained in the region of CRPS pain, significantly higher levels of interleukin (IL)-6 and tumor necrosis factor-α (TNF-α) were observed in the involved extremity relative to the uninvolved extremity.⁴⁹ Wesseldijk et al.⁵⁰ demonstrated the differences in levels of TNF-α and IL-6 of 12 CRPS type I patients are significantly decreased in blister fluid 6 years, compared to levels taken at 6 and 30 months, after their initial event. In addition, significantly elevated CSF levels of pro-inflammatory cytokines (IL-1β, IL-6) have been shown in CRPS patients compared to healthy controls and those with other types of pain.⁴⁸ Results to date do not support genetic factors as determinants of the cytokine profile that may contribute to CRPS.⁵¹ Mast cells might also be involved in inflammatory reactions observed in CRPS type I and probably play a role in the production of proinflammatory cytokines such as TNF-α.⁴⁹

Some evidence suggests that “neurogenic inflammation” is facilitated in CRPS patients. Transcutaneous electrical stimulation caused increased axon reflex vasodilatation and provoked protein extravasation only in CRPS patients, whereas it resulted in decreased axon reflex vasodilatation in healthy controls.⁵² Animal studies demonstrate that the SNS can influence the intensity of an inflammatory process.⁵³^,⁵⁴^,⁵⁵^,⁵⁶ In human pain models, sympatholytic procedures can reduce pain, inflammation, and edema.⁵⁷^,⁵⁸

IMMUNOLOGIC FACTORS

There has been a recent emphasis on immunologic factors being associated with the pathophysiology of CRPS.⁵⁹^,⁶⁰^,⁶¹ A possible autoimmune etiology of CRPS in some cases has also been proposed, and autoantibodies against nervous system structures have been described in some patients.⁶² Kohr et al.⁶³ demonstrated autoantibodies from the serum of CRPS patients binding to autonomic neurons including sympathetic and myenteric plexus neurons compared to controls. They subsequently reported CRPS autoantibody binding against β2-adrenergic receptors and muscarinic-2 receptors.⁶⁴ Dubuis et al.⁶⁵ conducted a study reporting binding of long-standing (at least 6 months) CRPS patient autoantibody binding to α-1A receptors compared to controls. Interestingly, they noted that CRPS patients who had pain relief from an earlier trial of immunoglobulin treatment all had the binding CRPS autoantibody.⁶⁵

Autoantibodies from CRPS serum incubated with endothelial cells, smooth muscle cells, and osteoblasts have been demonstrated to produce changes that could explain the clinical signs of CRPS.⁶⁶ Tékus et al.⁶⁷ conducted a study where they injected mice with immunoglobulin G (IgG) from CRPS patients prior to making a hind limb muscle incision and reported the mice developing CRPS clinical signs including mechanical hyperalgesia and edema compared to mice injected with IgG from healthy subjects of saline. Reilly et al.⁶⁸ demonstrated CRPS IgG binding to DRGs when they are incubated with inflammatory mediators representing an “inflammatory soup.” However, this did not occur without preincubation from inflammatory mediators.⁶⁸

Other immunologic factors besides immunoglobulins have been demonstrated to possibly contribute to CRPS clinical signs.
Li et al.⁶⁹ conducted a study that demonstrated the depletion of CD20+ B cells from anti-CD20 antibodies reduces the amount of CRPS changes seen in mice in a fracture/cast CRPS model compared to controls. Further, pro-inflammatory monocytes CD14+ and CD16+ have been found to be elevated in CRPS patients compared to controls along with a decreased IL-10, an anti-inflammatory cytokine IL typically low in CRPS patients.⁷⁰

Afferent Dysfunction

CRPS is characterized by “disproportionate” spontaneous and evoked pain (e.g., hyperalgesia, thermal and mechanical allodynia). Whether these sensory symptoms are due to peripheral and/or central sensitization is the subject of considerable debate and investigation. Sensory impairment occurs in more than 70% of CRPS patients.⁷¹^,⁷² Although traditionally these sensory changes were thought to occur in a distal, regional distribution along with characteristic autonomic disturbances, some data suggest that the afferent disturbances may be hemilateral or quadrantic.⁷³ Study observations corroborate that CRPS may be associated with more generalized sensory impairments.⁷⁴^,⁷⁵ The types of remote sensory abnormalities described include not only the allodynia and hyperalgesia typical of CRPS but also hypoesthesia and other dysesthesias. Psychophysical technologies such as temperature quantitative sensory testing reveal that ipsilateral “hypoalgesia” is common, and contralateral hypoesthesia may occur in some cases.⁷⁵ Patients found to have generalized sensory impairment were also found to have ipsilateral weakness.⁷⁴ Another study showed that patients with generalized sensory impairment had ipsilateral hemibody increases in thresholds for touch, thermal sensation, and heat pain, whereas those with sensory deficits limited to the distal affected limb showed touch threshold elevations only in the affected limb.⁷⁵ In this trial, 46% of patients showed abnormalities in nerve conduction testing, and 24% abnormalities in somatosensory evoked potentials, although these abnormalities did not correspond to the extent of sensory impairments observed on quantitative sensory testing.⁷⁵ These studies and others suggest that the sensory impairments in CRPS often extend beyond the area affected by pain, and up to 50% of subjects show hypoesthesia and hypoalgesia in a quadrantic or hemibody distribution ipsilateral to the pain. These effects are more likely to occur with greater CRPS chronicity, suggesting centralization of the pathology over time.

Skin biopsies have revealed decreased C-fiber and Aδ-fiber axonal densities in the affected limb of CRPS type I patients compared to the unaffected side.⁷⁶ It is not known whether such changes are primary to the disease or a consequence of nutritional changes and ischemia caused by chronic vasoconstriction and inflammation.³⁹ These changes could help account for some of the sensory alterations described earlier that may occur in CRPS type I (as well as CRPS type II).⁷⁷^,⁷⁸

A recent study examined patterns of sensory signs for upper limb CRPS type I, CRPS type II, and peripheral nerve injury (PNI) patients.⁷⁹ Recruitment included 416 patients from the German Research Network on Neuropathic Pain. Quantitative sensory testing demonstrated several overlapping somatosensory abnormalities between CRPS and PNI patients. This suggests quantitative sensory testing cannot, alone, differentiate between CRPS and PNI patients. In all syndromes, most patients (66% to 69%) were determined to have a combination of sensory loss and hyperalgesia/allodynia. Increased pressure pain in CRPS patients was the only obvious differentiating sign between CRPS and PNI sensory abnormalities. Loss of mechanical detection was greater in CRPS type II than CRPS type I and was considered to be the greatest differentiator between these two syndromes in this study.

CENTRAL DYSFUNCTION

There is considerable evidence supporting centralization of the pathology in chronic CRPS.⁷³^,⁸⁰^,⁸¹^,⁸² Further support comes from evidence suggesting presence of sympathetic and motor dysfunction in the region of pain (see discussion in the following section). An acute “nociceptive barrage” from an inciting trauma, or due to peripheral sensitization and/or neurogenic inflammation, can cause rapid changes in the central nervous system (brain and spinal cord), a process commonly referred to as central sensitization.⁸³ As a corollary, it has been hypothesized that normalization of afferent activity will reset and/or dampen this sensitization (e.g., increased “functional” input on large fiber tracts may modulate or normalize activity of small fiber tracts or “shut the pain gate”⁸⁴).

Evidence for supraspinal central mechanisms in CRPS also derives from studies using neuroimaging techniques that permit exploration of changes in information processing in the brains of CRPS patients. Studies using magnetoencephalography (MEG), quantitative electroencephalography (EEG), functional magnetic resonance imaging (fMRI), and positron emission tomography (PET) techniques all indicate that alterations of afferent input lead to cortical and thalamic plasticity and reorganization of sensory representations in patients with CRPS type I.⁸⁵^,⁸⁶^,⁸⁷^,⁸⁸^,⁸⁹ Increased brain responsiveness among CRPS patients in some imaging studies also supports the presence of central sensitization.⁸⁸ Some studies suggest a reduced size of the motor cortex devoted to the affected limb in unilateral CRPS type I patients compared to the unaffected side,⁸⁸ and the relevance of such brain changes to clinical symptoms is supported by the fact that degree of this “shrinkage” correlates with the degree of hyperalgesia to pinprick.⁹⁰ fMRI studies indicate different patterns of cortical activation to pinprick in the CRPS-affected side (S1, S2, insula, frontal, anterior cingulate cortex) compared to the unaffected side (S1, S2, insula only).⁹¹ Moreover, prospective research using MEG indicates that stimulation of early, unilateral CRPS type I subjects leads initially only to contralateral activation, whereas after 3 years of CRPS, the same stimulation leads to bilateral activation.⁸⁸ Such results suggest one possible brain mechanism by which reported contralateral “spreading” of CRPS symptoms may occur. A meta-analysis that included six studies investigated S1 representation in upper extremity CRPS.⁹² They determined, from these limited data, that S1 representation is smaller in the CRPS-affected upper limb compared to the unaffected limb. A follow-up study comparing 17 patients with upper extremity CRPS and 16 healthy controls determined that the S1 representation was the same size in the affected limb of CRPS subjects and controls.⁹³ Further, they found S1 representation in the unaffected limb of CRPS patients was larger than controls. This is consistent with the previous meta-analysis and provides interesting insight on the possible neuroplasticity changes that occur in CRPS pathology. Further studies are needed to explore these findings.

Imaging studies suggest that CRPS-related brain changes like those described earlier may reverse after successful treatment; thus, a reduction or resolution of CRPS pain may correlate with “resetting” of the cortical reorganization associated with the disorder.⁹⁰^,⁹⁴

SYMPATHETIC DYSFUNCTION

Although autonomic dysfunction has long been implicated as a key mechanism involved with CRPS pathology, the actual role of the SNS is incompletely characterized.⁹⁵ A seminal role of the SNS in CRPS is pivotal in most diagnostic criteria, emphasizing signs and symptoms of autonomic disturbance (e.g., vasomotor, sudomotor, and fluid regulation changes⁷¹^,⁹⁶^,⁹⁷). The crucial role of the SNS in at least some cases of CRPS is also suggested by the fact that sympatholytic blocks cause
substantial pain relief in a subset of patients (those with SMP; e.g., Wasner et al.,⁹⁸ and see following discussion). Because of the frequent beneficial response empirically seen to sympathetic blocks in chronic, cold, blue, sweaty CRPS, logically, sympathetic hyperactivity was originally thought to be a primary pathophysiologic mechanism.⁹⁹ However, an analysis of the current available data provides evidence that sympathetic vasoconstrictor activity is inhibited rather than enhanced, at least in early CRPS.¹⁰⁰^,¹⁰¹ The clinical presentation of CRPS appears to take two distinct forms, which may be sequential. Acutely, vasodilatation and sudomotor dysfunction (hot, red, occasionally dry) is characteristically observed; in contrast, patients with chronic CRPS often exhibit signs of vasoconstriction and hyperhidrosis (cold, blue, sweaty).¹⁰² This apparent temporal progression of vasomotor dysfunction from hypoactive to hyperactive is in accord with the sequential changes in rat paw temperature observed in the chronic constriction injury animal model of CRPS.¹⁰³ In a series of human studies examining thermoregulation and sympathetic reflexes in response to whole-body warming or cooling using a thermal suit, Wasner and colleagues⁹⁸^,¹⁰⁰ have corroborated the temporal progression from acute, relative sympathetic hypofunction, through an intermediate stage, to a chronic state of relative hyperfunction. Whole-body cooling, a very effective stimulus to tonically activate cutaneous vasoconstrictor pathways, induced a much lower level of vasoconstriction in the affected side as compared with the healthy side in acute CRPS patients.⁹⁸ Wasner et al.¹⁰⁰ also reported a significant negative correlation (P < .001) between the duration of the disease and the maximal temperature difference between the affected and unaffected sides achieved during this thermoregulatory testing. The “cold” symptom pattern often seen in chronic CRPS is most likely related to adrenergic receptor supersensitivity, perhaps resulting from early sympathetic hypofunction due to local sympathetic nerve damage,¹⁰⁴ decreased central drive, or both. This hypothesis is supported by studies examining plasma catecholamines in CRPS and studies using PET scanning techniques.¹⁰⁵^,¹⁰⁶^,¹⁰⁷ These mechanisms do not exclude processes at the capillary and venular site of the vascular bed that may involve the sympathetic and unmyelinated afferent neurons¹⁰⁸ that could contribute to neurogenic inflammation and edema.⁵⁸

Results of thermoregulatory sweat tests and quantitative sudomotor axon tests often reveal increased sudomotor activity in acute CRPS; however, only results of the former are increased in chronic CRPS.⁹⁸^,¹⁰⁹ Although these sudomotor changes are not entirely in line with the pattern of reported vasomotor changes in CRPS, they do support a central sympathetic dysregulation in this acute period that is in accord with known thermoregulatory mechanisms.⁹⁸^,¹⁰⁹ The dysfunctional effectors of the SNS that are most prominent in CRPS (vasomotor and sudomotor) are thermoregulatory and thought to be primarily under hypothalamic control.⁹⁵^,¹¹⁰ It is important to note that these sympathetic signs and symptoms are highly variable between patients and even within patients over time.¹¹¹^,¹¹²

During the 1980s and early 1990s, many clinicians and researchers considered the presence of SMP to be a central feature of CRPS. SMP was corroborated by reports of significant analgesia when the efferent sympathetic nerve supply to the affected area was blocked.¹¹³^,¹¹⁴^,¹¹⁵ The concept of SMP remains potentially useful clinically, as it suggests an intervention that may provide a relatively pain-free window of opportunity so that responsive patients might begin a functional restoration course. Physiologically, it is well established that nociceptive afferents are not influenced by sympathetic fibers under normal conditions,¹¹⁶^,¹¹⁷ and the specific role of the SNS in perpetuating pain in pathologic states such as CRPS is not clear.¹¹⁸ However, evidence from several studies suggests potential direct interactions between SNS activity and CRPS pain. Drummond¹¹⁹ introduced a small dose of norepinephrine into capsaicin-treated skin by iontophoresis and showed markedly increased thermal hyperalgesia. Moreover, injection of adrenergic agonists into the symptomatic limb of CRPS patients often provokes or enhances pain, even if the limb has been sympathectomized.¹¹⁴^,¹²⁰^,¹²¹ These findings are consistent with the clinical observation that CRPS pain often increases in cold weather or in response to psychological stress, when catecholamine secretion would be expected to increase.⁹⁸ Experimentally evoking the startle response also causes increased pain intensity in CRPS type I patients, presumably via sympathetic activation, and these pain changes are paralleled by significantly greater vasoconstriction on the affected side versus the unaffected side.¹²² Many of the motor abnormalities of CRPS have also been reported to improve with sympatholytic blocks, implying that some of these motor symptoms may also be sympathetically maintained.¹²³ Despite the hyperalgesic impact of increased SNS activity on CRPS pain in experimental studies, and the common clinical assumption that SNS hyperactivity contributes directly to CRPS pain, unilateral fracture patients who later developed CRPS showed impaired SNS shortly after injury (recorded with laser Doppler fluxmetry).¹²⁴ In addition, presence of impaired SNS function shortly after fracture prospectively predicted which later developed acute CRPS type I.¹²⁴ In a small trial, 33.5% of CRPS patients exhibited asymmetric vasomotor responses compared to a homologous symmetric pattern of response in all healthy controls when both groups view ambiguous visual stimuli, which is meant to illicit an autonomic response.¹²⁵ Further, 61% of the CRPS patients had intense pain within seconds of viewing ambiguous visual stimuli.

A theoretical synthesis of this SNS data: There is acute damage to small efferent sympathetic fibers with the original trauma (such as a fracture or crush injury) resulting in relative sympathetic hypofunction (hot, red, dry). Soon thereafter, cholinergic receptor upregulation occurs in the target tissues (vessels, sweat glands), and importantly, adrenergic receptors may become activated/sensitized on afferent pain fibers as well. Ultimately, there may be regeneration of transformed sympathetic fibers into this pathologically altered region (or restitution/upregulation of central sympathetic drive), giving the clinical appearance of sympathetic hyperfunction (cold, blue, sweaty) and providing direct stimulation of noradrenergically sensitized nociceptors.¹⁰⁰^,¹⁰²^,¹⁰⁴^,¹⁰⁵

TROPHIC, DYSTROPHIC, AND NUTRITIONAL ABNORMALITIES

Trophic/dystrophic changes to skin (thin and glossy, or thickened), nails (thickened, striated), and hair growth (increased or decreased) are ubiquitous in reports of chronic CRPS symptomatology but are of unknown etiology. These changes may simply be due to relative hypoxia with chronic vasoconstriction,⁷⁷^,¹²⁶^,¹²⁷ and a similar process could also lead to the observed nerve dropout⁷⁶ and osteopenia³⁹ reported in CRPS. There is evidence of cellular hypoxia or impaired oxygen utilization in the CRPS affected side compared to the unaffected side based on magnetic resonance spectroscopy.¹²⁸ Skin capillary hemoglobin oxygenation (HbO₂) has also been shown to be decreased on the affected side in CRPS patients as determined by microlight guide spectrophotometry.¹²⁷ Skin, but not venous lactate, was also reported to be increased on the affected side as measured by dermal microdialysis, suggesting decreased oxygenation.¹²⁶ The decreased range of motion often seen in CRPS could be due in part to trophic changes of joints, tendons, or ligaments. It is also possible that many of these changes could be caused by inflammation as originally suggested by Sudeck³⁹ or by hormonal changes.¹²⁹ Small-fiber dropout may be a nutritional/relative hypoxia type of “trophic” change or may be an essential etiologic factor in some cases.⁷⁶^,¹³⁰^,¹³¹^,¹³²

Since Sudeck¹³³ first described osteopenia in 1901, this bone loss has been shown to be a prominent feature of most chronic CRPS patients.¹³⁴ Whether this is due to disuse or is an elemental change in the osteoclast/osteoblast relationship is unclear, and elements of both may be present in most subjects. Certainly, the correction of these aberrations (e.g., bisphosphonates, graded exercise) may provide treatment targets.

MOTOR AND MOVEMENT DISORDERS

Although older diagnostic criteria do not mention motor or movement disorders,⁵ the new empirically derived criteria and clinical experience feature motor system abnormalities prominently.⁹⁷^,¹³⁵ Most CRPS patients show weakness, spasm, tremor, bradykinesia, and/or range of motion abnormalities at some time in the disease,¹²³ whereas a minority (˜10%) may show more dramatic aberrations such as dystonia.¹³⁶ Weakness of skeletal muscles of the affected distal extremity is common, being either true weakness (motor dysfunction) or weakness due to pain/kinesophobia.¹³⁶ Tremor may occur in as many as 70% of patients with upper extremity CRPS, with most represented by an apparent increase in physiologic tremor.¹²³^,¹³⁵ This tremor may decrease with sympathetic block or sympathectomy, suggesting that some motor features of CRPS may be sympathetically maintained.¹²³ Muscle spindles have extensive adrenergic innervation¹³⁷ and may become sensitized similar to the vasomotor system. Inflammatory mediators, especially cytokines, may also sensitize muscle spindles.¹³⁸ Bradykinesia is a common abnormality in CRPS and is likely a central abnormality at a spinal and/or cortical level.¹²³^,¹³⁵ Cerebral motor processing abnormalities have been shown in CRPS with kinematic and grip force analysis.¹³⁹ Increased reactivity of the motor cortex has been shown by MEG⁸⁸ and transcranial magnetic stimulation (TMS) studies.¹⁴⁰ In sum, numerous findings suggest clinically important motor abnormalities in CRPS that may be of central origin. A systematic review and meta-analysis attempted to examine the literature to determine the extent the motor cortex may be altered in CRPS.¹⁴¹ There were 18 studies included in the systematic review that included 14 unique data sets using TMS, fMRI, MEG, and PET modalities with 8 studies included in the meta-analysis. The study determined there was limited evidence of bilateral M1 disinhibition in CRPS of the upper limb and better designed studies were needed to reexamine other claims about motor cortex involvement in CRPS. There is some evidence that there may be an incongruity between sensory input and motor output, and this hypothesis has been used as the rationale for a treatment that involves “normalizing” the sensory input with mirror therapy.¹⁴² An apparent motor neglect syndrome could also contribute to significant motor dysfunction in some patients, further suggesting a central etiology of the motor dysfunction of CRPS.¹⁴³ CRPS-related motor dysfunction may worsen problems with disuse (see following discussion). An area of interest in motor dysfunction is the basal ganglia, as this is an area of somatosensory and motor processing physiologically positioned to explain motor aberrations in CRPS.¹⁴⁴ Chronic nociceptive input to the basal ganglia may lead to abnormal programming of motor tasks in CRPS patients.

IMMOBILIZATION AND DISUSE

Patients with prolonged casting/immobilization of a limb present with many signs and symptoms considered characteristic of CRPS (e.g., sensory disturbances, vasomotor and sudomotor asymmetry, atrophic and dystrophic changes including osteopenia¹⁴⁵). In rats casting of hind paw for 4 weeks led to warmer limbs, edema, enhanced cellular extravasation, allodynia, and periarticular osteopenia that reversed after removing the casts, and rats casted after tibial fracture showed vasomotor and nociceptive abnormalities that persisted for several months longer than casting alone.¹⁴⁶ A follow-up study from Guo et al.¹⁴⁷ demonstrated glucocorticoid treatment had no effect on periarticular osteopenia. The potential contribution of disuse to CRPS symptomatology provides a primary rationale for functional restoration as treatment.¹⁴⁸^,¹⁴⁹

Genetics

One of the unsolved questions in human pain is why only a minority of patients develop chronic pain after identical inciting events (e.g., nerve lesions).¹⁵⁰^,¹⁵¹ In contrast, in animal models of nerve lesion, almost all animals develop neuropathic pain behavior (e.g., Bennet and Xie,¹⁵ Kim and Chung,¹⁶ Selzer et al.,¹⁸ Blenk et al.,¹⁵² Decosterd and Woolf¹⁵³). In CRPS specifically, it is interesting to consider why some individuals who have frequent, sometimes severe injuries considered etiologic never develop CRPS (e.g., soccer players, American football players); yet in others, trivial injuries may lead to full-blown CRPS (e.g., intravenous starts, minor sprains).¹⁵⁰ As always, there may be environmental considerations that predispose, but genetic factors will likely ultimately prove to be important. The human leucocyte antigen (HLA) molecules encoded by genes of the major histocompatibility complex (MHC) may contribute to several neurologic disorders including CRPS.¹⁵⁴ A study evaluated 150 phenotypically homogenous CRPS patients with fixed dystonia and found a significant association with HLA alleles HLA-B62 and HLA-DQ8.¹⁵⁵ The same research group conducted a follow-up study to evaluate whether a phenotypically homogenous CRPS group without dystonia has the same associations, and they found HLA-DQ8 to be the only significantly associated allele.¹⁵⁶ However, these results were not replicated by another group that interrogated about 83% of common single-nucleotide polymorphisms in the human genome, finding they were not associated with CRPS phenotype in their cohort.¹⁵⁷ An early report described three families with two or more members in each affected by CRPS, suggesting the possibility that CRPS risk was heritable.¹⁵⁸ In small studies, the HLA loci A3, B7, DQ1(06), DR2(15), and DR13 are specifically implicated.¹⁵⁹^,¹⁶⁰^,¹⁶¹ One Japanese study suggests an association between CRPS and the (non-MHC axis) angiotensin-converting enzyme gene and notes elevated angiotensin II in CRPS patients.¹⁶² At present, data regarding possible genetic contributions to CRPS are rather limited but are sufficient to justify further exploration of this very plausible contribution to pathophysiology.

A Convergent Pathophysiologic Theory

These seemingly divergent factors that are proven, theorized, or hypothesized to be involved in the pathophysiology of CRPS can be synthesized.¹⁴ The peripheral factors such as nerve and tissue damage and inflammation (regional and neurogenic) cause an afferent barrage that begins the process of peripheral and central sensitization. Central effects such as changes in the dorsal horn, brainstem, limbic areas, and cortex acutely and chronically occur. An efferent response evolves as feedback from all these areas along the motor (corticospinal) and sympathetic (limbic-hypothalamic-brainstem) systems. This efferent outflow may then substantially impact the peripheral dynamic by affecting the nutritional and inflammatory stasis in the damaged region. All of these sites interconnect, for instance, by ephapses and short loops in the CNS (e.g., the sensory to sympathetic ganglion, dorsal to lateral horn, brainstem). Overlying all of this is variance in response on the basis of genetics (nature) and psychosocial factors (nurture). Thus, nested feedback and feed forward loops and cascades within the neuraxis develop. Importantly, these feedback and feed forward loops can be self-maintaining and can “spread” (come to involve more and more elements of the central nervous) (Fig. 25.1).

FIGURE 25.1 Complex regional pain syndrome maintained and reinforced by nested, reverberating, feed forward (afferent) and feedback (efferent) loops. Overlaying this are the vectors of genetics (endogenous) and sociologic operant paradigms (exogenous). NE, norepinephrine. (Modified from Merskey H, Bogduk N. Classification of chronic pain: descriptions of chronic pain syndromes and definitions of pain terms. 2nd ed. Seattle: IASP Press; 1994. This table has been reproduced with permission of the International Association for the Study of Pain^® [IASP]. The table may not be reproduced for any other purpose without permission.)

Diagnosis

In the past, CRPS was diagnosed using a variety of nonstandardized and idiosyncratic diagnostic systems (e.g., Gibbons and Wilson,⁹⁶ Bonica,¹⁶³ Kozin et al.,¹⁶⁴ Wilson et al.¹⁶⁵), each of which was derived from the authors’ clinical experiences and none of which achieved wide acceptance. This lack of a common diagnostic criteria hindered clinical progress for decades.

THE INTERNATIONAL ASSOCIATION FOR THE STUDY OF PAIN CRITERIA

After much debate, the name for the syndrome was ultimately changed to complex regional pain syndrome (CRPS) at a consensus workshop in Orlando, Florida, in 1994,⁵⁴^,¹⁶⁶ with the new name and diagnostic criteria codified by the IASP taskforce on taxonomy (see Table 25.1).⁵ This new diagnostic entity was meant to be descriptive, general, and not to imply any pathophysiology (including any direct role for the SNS). According to these criteria, CRPS can be diagnosed regardless of whether the pain is SMP (SNS block responsive) or sympathetically independent pain (SIP). These IASP-endorsed criteria had the potential to lead to improved clinical communication and greater generalizability across research samples.⁵⁴ However, realization of this potential has been somewhat limited by the fact that these rather vague criteria were derived solely by consensus. Also, since publication in 1994, utilization of the criteria has been sporadic (although increasing) in the literature,¹⁶⁷ and certain groups had resisted the change (especially certain advocacy groups and payors). As a consequence, the full benefits of common criteria have not been completely realized.

Experience gained in developing diagnostic criteria in headache and psychiatric disorders indicates that consensus-based criteria can be significantly improved by systematic empirical validation¹⁶⁸ and consensus-derived criteria that are not subsequently validated may lead to over- or underdiagnosis, thus reducing the ability to provide timely, optimal treatment. Results of validation studies to date suggest that the 1994 IASP criteria are adequately sensitive (i.e., rarely miss a case of actual CRPS); however, both internal and external validation research suggest potential problems with specificity (overdiagnosis).⁹⁷^,¹⁶⁹^,¹⁷⁰ The 1994 IASP criteria implicitly assume that signs and symptoms of vasomotor, sudomotor, and edema-related changes provide redundant diagnostic information; that is, presence of any one of these is sufficient to meet criterion 3 (see Table 25.1). This combination of multiple distinct elements of the syndrome into the same diagnostic criterion and allowing patient-reported symptoms are likely elements contributing to overdiagnosis in the 1994 IASP scheme.⁹⁷^,¹⁷¹ An additional weakness of these criteria was the failure to include motor/trophic signs and symptoms, which could lead to important information being ignored that may help discriminate CRPS from other syndromes.

These conclusions are supported by the results of factor analysis (a statistical pattern recognition technique) that was conducted in a series of 123 CRPS patients. These results indicated that signs and symptoms of CRPS clustered into four statistically distinct subgroups.⁹⁷ The first of these subgroups was a set of signs and symptoms indicating abnormalities in pain processing (e.g., allodynia, hyperalgesia). Skin color and temperature changes characterized the second subgroup, indicative of vasomotor dysfunction. Edema and sudomotor dysfunction (e.g., sweating changes) combined to form a third unique subgroup. The finding that vasomotor signs and symptoms were statistically distinct from those reflecting sudomotor changes/edema is in contrast to the 1994 IASP criteria, which treats all three of these as diagnostically equivalent. A fourth and final subgroup was identified that included motor and trophic signs and symptoms. Numerous studies have described various signs of motor dysfunction (e.g., dystonia, tremor) as important characteristics of this disorder,¹⁴³^,¹⁶⁵^,¹⁷² and trophic changes have frequently been mentioned in historical clinical descriptions. The absence of both of these features from the 1994 IASP criteria is therefore notable, especially given factor analytic findings that this subgroup of signs and symptoms does not overlap significantly with other characteristics of CRPS currently used in diagnosis.

External validity, which addresses the ability of the diagnostic criteria to distinguish CRPS patients from those with other types of pain conditions, is also an important issue. In the absence of a definitive pathophysiology of CRPS and thus the absence of a definitive objective test to serve as a “gold standard,” providing evidence for the external validity of diagnostic criteria is challenging.¹⁷³ However, the upper limit of external validity can be evaluated by using the original criteria themselves as a reference point.⁹⁷^,¹⁶⁹^,¹⁷³ In this methodology, a CRPS patient group is identified using a strict application of the 1994 IASP/CRPS criteria with a comparison group of non-CRPS neuropathic pain patients defined by independent diagnostic information (e.g., diabetic neuropathy diagnosed by presence of chronic diabetes with ascending symmetrical pain and corroborated by electrodiagnostic studies). Existing criteria and modifications to these criteria can then be evaluated with regard to their ability to distinguish between these two groups based on patterns of signs and symptoms. This model was used to test the utility of the 1994 IASP/CRPS criteria for discriminating between 117 patients meeting 1994 IASP criteria and 43 neuropathic pain patients with established non-CRPS etiology. The 1994 IASP/CRPS criteria and decision rules (e.g., “evidence at some time” of edema or color changes or sweating changes satisfy criterion 3) did discriminate significantly between the CRPS and non-CRPS groups. However, closer examination of the results indicated that although diagnostic sensitivity (i.e., being able to detect the disorder when it is present) was quite high (.98), specificity (i.e., minimizing false-positive diagnoses) was poor (.36), and a positive diagnosis of CRPS was likely to be correct in as few as 40% of cases.⁹⁷^,¹⁷⁰

For clinical purposes, sensitivity is extremely important. On the other hand, the issue of specificity is quite important for selection of research samples as well as for minimizing unnecessary, potentially invasive treatments. The clinical implication of high sensitivity at the expense of specificity is that CRPS may be overdiagnosed and, ultimately, overtreated. It also has the very significant downside of identifying pathophysiologically heterogeneous groups for research, contributing potentially to negative results in clinical trials.⁹⁷

THE BUDAPEST CRITERIA

A meeting of international researchers and clinicians with expertise in CRPS (the “Budapest Group”) was held in Budapest, Hungary, in 2003 to make consensus recommendations on proposed revisions to the 1994 IASP diagnostic criteria. The criteria ultimately proposed by the Budapest group were consensus-based modifications of the criteria that were statistically derived from the validation studies, as in the earlier discussion.⁹⁷^,¹⁷⁰ These modified criteria assessed CRPS characteristics within each of the four statistically derived factors described earlier. Given evidence that objective signs on examination and patient-reported symptoms both provide useful but nonidentical information, the modified criteria required the presence of signs and symptoms of CRPS for diagnosis.⁹⁷^,¹⁶⁹^,¹⁷⁴ A test of these modified criteria regarding their ability to discriminate between the CRPS and non-CRPS neuropathic pain groups indicated that they could increase diagnostic accuracy.⁹⁷^,¹⁷³ Results indicated that a decision rule requiring that two of four sign categories and three of four symptom categories be positive for a diagnosis to be made resulted in a sensitivity of .85 and a specificity of .69. This decision rule represented a good compromise between identifying as many patients as possible in the clinical context while substantially reducing the high level of false-positive diagnoses associated with 1994 IASP criteria. This decision rule was therefore adopted in a set of proposed clinical diagnostic criteria endorsed by the Budapest group (summarized in Table 25.2).⁹⁷^,¹⁷⁴

The proposed clinical diagnostic criteria described reflected an improvement over 1994 IASP criteria for clinical purposes but still suffered from less than optimal specificity for use in the research context.¹⁷⁴ Tests of the aforementioned modified CRPS criteria indicated that modifying the decision rules to require that two of four sign categories and four of four symptom categories be positive for diagnosis to be made resulted in a sensitivity of .70 and a specificity of .94. Of all the permutations tested, this decision rule resulted in the greatest probability of accurate diagnosis for both CRPS and non-CRPS patients (approximately 80% and 90% accuracy, respectively⁹⁷^,¹⁷⁰^,¹⁷¹^,¹⁷⁴). This high level of specificity was considered desirable in the research context by the Budapest consensus group and therefore was adopted as part of a set of proposed research diagnostic criteria.¹⁷¹^,¹⁷⁴ Current distinctions between CRPS type I and CRPS type II subtypes, reflecting respectively the absence and presence of evidence of PNI, were ultimately retained despite ongoing questions by the consensus group as to whether such distinctions have clinical utility.¹⁷⁴ The Budapest criteria were adopted by the IASP in 2012 after they consistently demonstrated a better way to maintain sensitivity and specificity when diagnosing CRPS.¹⁷⁵

Thus, CRPS is a clinical diagnosis made with simple bed side testing techniques known by all physicians. Technical testing procedures are sometimes used to corroborate or objectively document clinical impressions (e.g., thermography, bone scans⁴⁷). Recently, there was a study that suggested osteoprotegerin, an inhibitor of osteoclastic function, may aid as a biomarker for diagnosis.¹⁷⁶ To date, none of these have been formally validated.

SEQUENTIAL STAGES AND SUBSETS OF COMPLEX REGIONAL PAIN SYNDROME

Although not part of clinical diagnosis per se, it is lore among clinicians that CRPS entails three sequential stages that differ in patterns of signs and symptoms. This traditional staging model is summarized by Bonica.¹⁶³ The early, acute stage of CRPS (stage I) was characterized by pain/sensory abnormalities (e.g., hyperalgesia, allodynia), vasomotor and sudomotor dysfunction, and prominent edema. Stage II (dystrophic stage) was proposed to occur 3 to 6 months after onset and to be characterized by more marked pain/sensory dysfunction, with continued evidence of vasomotor dysfunction and development of significant motor/trophic changes. Stage III (atrophic stage) was characterized by decreased pain/sensory disturbance, continued vasomotor disturbance, and markedly increased motor/trophic changes. Although until recently, there had been only limited empirical tests of this hypothesized staging of CRPS, the concept has frequently been accepted as fact in the CRPS literature.¹⁶³^,¹⁷⁷^,¹⁷⁸

TABLE 25.2 Budapest Diagnostic Criteria for Complex Regional Pain Syndrome (Officially Adopted by IASP in 2012)

General Definition of the Syndrome

Complex regional pain syndrome 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) but may spread and usually has a distal predominance of abnormal sensory, motor, sudomotor, vasomotor, and/or trophic findings. The syndrome shows variable progression over time.

To make a clinical^a diagnosis, the following criteria must be met:

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, nail, skin)
Must display at least one sign at 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, nail, skin)
There is no other diagnosis that better explains the signs and symptoms.

^aFor research purposes, diagnostic decision rule should be at least one symptom in all four symptom categories and at least one sign observed at evaluation in two or more sign categories.

Modified from Harden RN, Bruehl S, Stanton-Hicks M, et al. Proposed new diagnostic criteria for complex regional pain syndrome. Pain Med 2007;8(4):326-331. Reproduced by permission of American Academy of Pain Medicine.

Only gold members can continue reading. Log In or Register to continue