| JOURNAL HOME | CME HOME | THIS MONTH | PAST ISSUES | ETOC | COLLECTIONS |
| AUTHORS | REVIEWERS | EDITORIAL BOARD | FEEDBACK | RSS | HELP |
| ||||||||||||||
| ||||||||||||||
|
|
|
|
||||||||||||
|
||||||||||||||
,
*Center for Pain Studies, Chicago, Illinois; Rehabilitation Institute of Chicago, Chicago, Illinois; Northwestern University Medical School, Chicago, Illinois; Department of Orthopedics and Rehabilitation Medicine and Pain Treatment and Research Center, University of Wisconsin Medical School, Madison, Wisconsin; ||Vanderbilt University School of Medicine, Nashville, Tennessee; ¶Medical University of South Carolina, Charleston, South Carolina; #University of Pennsylvania, Philadelphia, Pennsylvania; and **Chronic Pain and Fatigue Research Program, University of Michigan Health System, Ann Arbor, Michigan
Address correspondence and reprint requests to R. Norman Harden, MD, Rehabilitation Institute of Chicago, Center For Pain Studies, 446 E. Ontario, Ste. 1011, Chicago, IL 60611. Address e-mail to nharden{at}rehabchicago.org
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
IMPLICATIONS: This pilot study explored systemic venous plasma catecholamine levels in complex regional pain syndrome and the correlation with scores on psychometric measures. Our results suggest that increased catecholamine levels may result from pain or may reflect premorbid adrenergic hyperactivity, which could predispose these individuals to develop the syndrome.
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
CRPS is often associated with apparent dysfunction in the sympathetic nervous system; it clinically presents as dysautonomic changes (e.g., edema, increased sweating, and decreased temperature) in the affected limbs, which superficially appear to reflect a state of chronic sympathetic hyperactivity. However, sympatholytic treatments such as nerve blocks (3) and surgical sympathectomies often have incomplete or temporary effects, and a significant portion of CRPS patients do not respond to them at all. Raja et al. (4) classify these nonresponders as having sympathetically independent pain (SIP), whereas those whose symptoms improve after chemical or surgical sympatholysis have sympathetically maintained pain (SMP) (5).
If the sympathetic nervous system is not "overdriven," then the dysautonomia so prominent in chronic CRPS may suggest some form of adrenergic supersensitivity (6,7). Studies using models of peripheral nerve injury (8,9) support this concept, suggesting that tissue injury can trigger adrenosensitivity in injured afferent axons and surrounding vessels independently of sympathetic activity. Therefore, after nerve injury, circulating catecholamines may directly trigger nociceptive firing (10) and the vasomotor changes characteristic of the syndrome (6,7). Drummond et al. (11) found that CRPS patients experience more pain during states of sympathetic arousal, lending further support to this hypothesis.
We previously compared venous pool plasma catecholamine levels between affected and unaffected limbs of CRPS patients with SMP, all of whom manifested the dysautonomic changes classic in this syndrome, and we observed a small but significant decrease in venous plasma norepinephrine (NE) on the affected side (7). Drummond et al. (6) reported similar findings. Although the reasons for the decrease require further exploration, these results further contradict the hypothesis that the dysautonomia of CRPS is mediated by sympathetic hyperactivity in the affected limb. Instead, the more likely mechanism is peripheral adrenosensitivity, with circulating catecholamines causing an exaggerated peripheral autonomic response (6,7).
In this previous work, we noted that many CRPS patients had systemic plasma NE and epinephrine (E) levels in excess of prototypical normal ranges (derived from only five control subjects in that study) (7). In individuals without chronic pain, plasma NE and E levels increase during dysphoric emotional states such as anxiety, hostility, and depression (12,13) and may be reduced by relaxation techniques (14). Depression and anxiety are reported to be increased in CRPS patients (15), and these individuals’ increased systemic NE and E may be a consequence of the pain or the affective distress accompanying the syndrome. In a patient with peripheral adrenosensitivity in the affected limb, increased plasma NE and E would be expected to result in dysautonomia and, hypothetically, pain.
No study has formally compared systemic plasma catecholamine levels in CRPS patients with those in healthy controls, which is the purpose of this pilot work. CRPS patients with SMP were specifically chosen because their responsiveness to sympathetic blockade suggested direct or indirect involvement of sympathetic nervous system function in the production or maintenance of symptoms. This study also reports our preliminary investigation of interactions between psychological variables and plasma catecholamine levels with a series of standardized psychometric tests. If affective distress is positively correlated with catecholamine levels in CRPS patients and if our hypothesis that chronic CRPS subjects have peripheral adrenosensitivity is correct, then this relationship would suggest one mechanism by which psychological factors might influence the development, maintenance, and exacerbation of the syndrome.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Thirty-three individuals with upper limb CRPS were recruited from among the patients enrolled in the MUSC chronic pain treatment program. Each met published International Association for the Study of Pain (IASP) criteria for CRPS (1). All patients were also classified as having SMP, as indicated by at least a 50% decrease in visual analog pain ratings in response to paravertebral sympathetic ganglion blockade. All subjects were at least 1 wk out from their last block. This response to block was required for inclusion in the data set, with the assumption that this defined a more homogeneous group than the IASP diagnostic criteria alone—a group that was responsive to gross sympathetic manipulation and therefore exemplary of the hypothesis. The study was not designed to assess the effect of sympathetic blockade on pain, affect, or catecholamine levels. Twenty-three patients were subjects in our previous catecholamine investigation (7). In this pilot study, no attempt was made to control for medications.
Control subjects were 30 healthy volunteers who were recruited from the faculty and support staff of MUSC and the Rehabilitation Institute of Chicago. Criteria for inclusion as a control were 1) absence of any major physical or psychiatric disorder (by bedside history/assessment); 2) no regular psychoactive, endocrine, or antihypertensive medication (diuretics were permissible); 3) no known allergy or sensitivity to heparin; 4) age between 18 and 70 yr; 5) no history of exaggerated bleeding tendency or thrombocytopenia; and 6) ability to provide informed consent. Efforts were made to match ages of patients and control subjects, but these attempts were limited by control subject availability.
Subjects were asked to refrain from using any medication or caffeine on the day of blood collection and to refrain from smoking for 4 h before collection time. For patients, the limb affected by CRPS was identified, and blood was drawn from the contralateral limb. Twenty-three of the patients were preparing to undergo therapeutic IV regional (Bier) block of the affected limb, as previously described (7). For control subjects, limb laterality for blood collection was determined by subject preference.
An IV catheter was started by using sterile technique in a distal vein of the chosen limb, usually in the dorsal venous arch of the hand or wrist. After confirmation of catheter placement by observation of blood return, an intermittent infusion cap was attached, and the catheter was flushed once with heparin sodium (100 U in 1 mL of solution) to maintain catheter patency.
Efforts were made to minimize the noxious or visual effects of stimuli related to blood collection that might artificially increase plasma catecholamine levels. To reduce the immediate effects of pain and anxiety surrounding the needlestick, each subject rested alone in the supine position in a quiet, dimly lit room for 30 min after IV insertion. The investigator then re-entered the room and asked each subject to remain supine, facing away from the IV site. The infusion lock was removed, and 3 mL of blood was withdrawn through the IV and discarded to minimize the risk of heparin contamination. A second syringe was used to draw an additional 10 mL of blood for analysis. Either the IV catheter was then maintained for access during the Bier block, or it was removed and a sterile dressing was applied.
The 10 mL of blood was immediately transferred into 2 sterile 5-mL vacuum collection tubes containing EDTA. The tubes were labeled with patient initials and the date of collection and were immediately transported on ice to a centrifuge, where plasma was separated, extracted, and stored at –70°C.
A subset of 18 CRPS patients completed standardized psychometric questionnaires before blood collection. Instruments included the Beck Depression Inventory (BDI) (16) and Beck Anxiety Inventory (BAI) (17), which are well validated measures of depression and anxiety states, respectively, and the Minnesota Multiphasic Personality Inventory-2 (MMPI-2) (18), an extensively validated measure of personality functioning.
Frozen plasma samples were transported on dry ice to the MUSC chemistry laboratory for analysis. NE and E were extracted from plasma onto activated alumina and assayed with high-pressure liquid chromatography with electrochemical detection, with dihydroxybenzylamine as the internal standard. The sensitivity of the assay was 15 pg/mL for both compounds. NE levels were determined on all samples submitted for analysis. E levels were determined for 25 (83%) of 30 controls and for 30 (91%) of 33 patients.
All analyses were conducted with SPSS statistical software (SPSS Inc., Chicago, IL). Age matching between groups was tested with an independent-samples Student’s t-test. Sex matching was tested with nonparametric 2 x 2 
2 analysis. Control ranges for plasma NE and E were calculated on the basis of the upper and lower limits of the 95% confidence interval for the control sample. Because NE values were normally distributed, control and patient plasma NE distributions were compared by using simple factorial analysis of variance. E values did not approach a normal distribution in either patients or controls; groups were therefore compared by using the nonparametric Mann-Whitney U-test.
Although the available sample size was small, we were able to conduct a preliminary test of the relationships between catecholamine levels and scores on the BDI, BAI, and MMPI-2. NE distribution was normal within the subset of CRPS patients who completed psychometric instruments, although the distribution of E was nonnormal. Therefore, correlations between the psychometric measures and NE levels were examined by using parametric correlations (Pearson’s r), whereas correlations with E levels were analyzed with nonparametric correlations (Spearman’s 
).
Fourteen (47%) of 30 CRPS patients and 14 (56%) of 25 controls had E concentrations less than the lower limit of detection of the assay, and these were coded as 0 for statistical analyses. Sample sizes for analyses of psychometric measures in this pilot study varied because of missing data (ranging from 12 for the BDI and BAI to 16 for the MMPI-2). The maximum number of subjects available for each variable was used in all analyses.
A priori directional hypotheses regarding relationships between depression/anxiety and catecholamines were tested by using one-tailed probability values to increase statistical power, given the limited sample size for these variables. Probability values for all other analyses were two tailed.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
2 = 3.442; P < 0.08). Age distributions across the two groups were not significantly different (P > 0.10). Within-group analyses indicated that mean age was not significantly different across sexes in either the CRPS or control groups (P > 0.10).
|
|
= 0.05;P > 0.10) or controls (Spearman’s = –0.04; P > 0.10).
|
27 = –0.41; P < 0.05). In contrast, there was a significant relationship between NE and sex among controls, with higher plasma NE levels in female patients (28 = 0.64; P < 0.001). Male controls displayed a trend toward higher E levels, but this did not reach statistical significance (23 = –0.38; P < 0.1). Grossly, no significant differences were found between men and women on any of the psychometric tests (P <0.1); however, this sample was too small to make any definitive statements as to sex/psychometric effects.
Age and E were significantly related in the control group, with higher E levels in younger individuals (Spearman 23 = –0.46; P < 0.05). In contrast, neither NE nor E was significantly correlated with age in the CRPS group (all correlations: P > 0.10).
Table 4 presents exploratory correlations between psychometric measures (BDI, BAI, and MMPI-2) and plasma E and NE levels for a subset of CRPS patients. Because correlations were not based on a specific directional hypothesis, P values are given strictly for descriptive reasons. Plasma NE was not significantly correlated with any of the psychometric scores. However, several significant relationships were observed between psychometric results and plasma E levels.
|
Given the small sample size available for this analysis, we considered the possibility that extreme cases could bias the results. However, exclusion of statistical outliers for both E and NE did not appreciably alter the pattern of significant relationships observed.
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The original theoretical concept of reflex sympathetic dystrophy (also known as CRPS) implicated sympathetic nervous hyperactivity in the affected limb as the mechanism driving the syndrome. However, current research indicates that regional venous plasma NE (and, indirectly, sympathetic function) may be decreased in the affected limb relative to the contralateral unaffected limb (6,7). If tissue injury sensitizes peripheral nociceptors to catecholamines, then the pain and dysautonomic phenomena seen in CRPS could be caused by normal (or high) systemic levels of catecholamines affecting supersensitive tissue, rather than by high regional levels of catecholamines produced by an overactive sympathetic system (6,7). Indeed, the decrease in NE seen in CRPS patients may actually reflect local sympathetic hypofunction that is perhaps related to damage to sympathetic efferents (7), and this is consistent with the clinical observation that many patients go through a period of apparent sympathetic hypofunction acutely (hot, red, and dry) before developing the characteristic chronic presentation (cold, blue, and sweaty) (7).
Although the sympathetic nervous system efferents may not be overdriven in CRPS, experimental evidence does suggest functional sympathetic/sensory coupling at the dorsal root ganglion (DRG) in animal models of neuropathic pain. The nociceptive barrage of pain signals from sensitized afferent terminals induces windup of wide dynamic range neurons in the DRG, manifested by increased firing frequency and decreased discrimination between noxious and normally innocuous stimuli (28). Postganglionic sympathetic activation has been shown to potentiate ectopic discharges from primary DRG neurons during windup (29). This relationship may have a structural basis; sympathetic postganglionic axons form basket-like skeins around the somata of DRG sensory neurons in animal and human neuropathic pain states (30). There are thus sound arguments for both peripheral noradrenergic sensitization (8,9) and for direct sympathetic nervous involvement (31) in the maintenance of neuropathic pain states. Both mechanisms may be involved in CRPS to varying degrees.
In this pilot study we found that levels of circulating NE were higher than normal control levels in 78% of our CRPS sample, a statistically significant difference. If CRPS patients are regionally sensitized to NE at the tissue level and if upregulation of peripheral NE receptors is tied to abnormal nociception and vasomotor function, then increased circulating NE levels would have marked clinical significance. Patients with high plasma NE would be expected to experience greater regional pain and manifest increased edema, vasoconstriction, and other hyperadrenergic manifestations of CRPS. Subsequent investigations should assess these variables and correlate them with NE levels.
This finding raises questions about the clinical distinction between SMP and SIP (4,32). Patients with SMP respond to sympatholytic therapies, whereas SIP patients do not. This difference in response could reflect differing regional or circulating NE levels, differing tissue sensitivities to NE, or both. All CRPS patients involved in this study had documented SMP. Further research comparing SMP and SIP patients and the relationship of these subsets of CRPS to catecholamine levels, clinical presentation, and psychometric testing is indicated.
We found levels of circulating E to be higher than normal control values in 52% of CRPS patients, with concentrations up to 90 times larger than normal in one patient. (None of the individuals with increased E levels had known pheochromocytomas or other endocrine, gastrointestinal (GI), or kidney pathology, nor were any found on subsequent workup.) However, the difference in plasma E levels between groups was not statistically significant. Whether the marked E increases in certain individuals reflect preexisting emotional distress, anxiety over the sampling technique, some other premorbid or comorbid hyperadrenergic state, or another process entirely (such as an artifact) is unexplained. One source of variance not controlled for in this pilot study was the fact that some of the subjects were taking antidepressant drugs for pain or depression, and these drugs theoretically may influence blood catecholamine levels. This factor should be considered in a definitive trial.
One might hypothesize that increased E and NE levels in CRPS are a consequence of the pain and distress that accompany the syndrome, rather than a manifestation of endocrine pathology or an autonomic or psychological state peculiar to CRPS. For this hypothesis to be correct, chronic pain patients as a group would be expected to have increased catecholamine levels. A review of the literature fails to support this hypothesis. NE and E levels are similar to control values in patients with fibromyalgia (32,33), painful diabetic neuropathy (34), and cluster headaches (35). Plasma NE was less than control levels in patients with migraine (36), and concentrations of NE, E, and dopamine were decreased in patients with tension-type headaches (37). The only studied pain states besides CRPS that were accompanied by increases of plasma NE and E were acute pain conditions, i.e., venipuncture (34) and early myocardial infarction (38,39). Please note our efforts to minimize the acute effect of venipuncture, in the Methods section.
As noted, one possible reason for plasma NE and E increases is affective distress. The mechanisms underlying this relationship include both central activation of the sympathetic nervous system (primarily accounting for systemic NE) and pituitary/adrenal activation (primarily accounting for E). Several studies suggest that CRPS patients may have more emotional distress than patients with other types of chronic pain conditions (15,40–45), but they also may experience "more" pain. Of course, there is no way to compare absolute pain levels across diagnoses.
Anxiety, stress, and pain all increase circulating catecholamine levels (34,46–48). In our small sample, we found that plasma E was positively correlated with scores on psychometric measures that reflected depression, high somatic focus, avoidant coping style, and emotional sensitivity/suspiciousness ("paranoia/hostility"). Our findings support the hypothesis that psychological distress and increased plasma catecholamines are related in CRPS patients (41,49–51), although no causal connection with CRPS development can be assumed, given the correlational nature of these data and the small sample size. We did not correlate pain severity with depression, anxiety, or E levels, and these relationships should be explored carefully in future investigations.
The kidneys and GI tract are also significant sources of peripheral circulating NE and E. The differential contribution of various organs to circulating NE and E was not addressed in this pilot investigation.
By contrast, no significant correlations were demonstrated between NE levels and psychometric scores, and the concentration of one measured catecholamine did not predict that of the other. These were unexpected findings; one possible explanation is separate roles for E and NE in the development or maintenance of CRPS. Whereas NE has been experimentally linked to peripheral sensitization, allodynia, and hyperalgesia, E may be more involved in the realm of emotional distress and the stress response to pain. Further investigation is necessary to clarify these relationships.
Do increased E and NE levels predispose individuals to develop CRPS (52)? Our data provide no ready answers. Whether increased NE and E predate the development of CRPS is unknown. No published studies have documented psychological status and catecholamine levels before the onset of the syndrome. Whatever the ultimate cause of CRPS, our findings are consistent with the theoretical effects of many of the treatments recommended for the syndrome (53). Early and aggressive treatment of stress, anxiety, and depression by methods such as relaxation training (14) may reduce plasma NE and E levels, theoretically decreasing peripheral sensitization and pain and thereby alleviating emotional distress. Pharmacologic management of pain, depression, and anxiety may also help to reduce affective distress, the adrenergic stress response, and circulating catecholamine levels, leading to further beneficial effects.
The connection between plasma catecholamines, psychological state, and CRPS certainly deserves exploration beyond this pilot study. A group of patients with SIP should be studied and compared with SMP patients to examine any differences in plasma catecholamine levels or psychometric scores. Enlargement of sample size, expansion of the psychometric battery, and use of a control group (perhaps one with a different chronic pain condition) will increase the statistical power and robustness of the study. We did not control for the duration of CRPS symptoms, time since the last sympathetic block (except that there was at least one week since the last block), type of block, comorbid medical conditions, or the roles of any long-acting medications taken by the patients on the days before sampling. These variables must be considered in further investigations.
Even as we understand more about the complex pathophysiology of neuropathic pain, many aspects of CRPS remain a mystery, and each new piece of knowledge tends to raise more questions. The exact role of catecholamines in the genesis and maintenance of CRPS is still incompletely understood. In this study, CRPS patients with SMP demonstrated increases in plasma NE and (to a lesser extent) E compared with normal control subjects, and increased E was correlated with psychometric scores that indicated affective distress and certain personality characteristics. Whether these psychological characteristics originated before or after the onset of CRPS remains unclear. Further investigation may lead us toward techniques for the identification of individuals at high risk for CRPS and may eventually lead to effective preventive measures and more effective treatments.
|
Acknowledgments |
|---|
The authors gratefully acknowledge the support and assistance of the following individuals: Felice Borisy Rudin, PhD, for critical review of the manuscript; and John Cate for the use of the MUSC laboratory facilities.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  F. Wesseldijk, D. Fekkes, F. J. Huygen, E. Bogaerts-Taal, and F. J. Zijlstra Increased Plasma Serotonin in Complex Regional Pain Syndrome Type 1 Anesth. Analg., June 1, 2008; 106(6): 1862 - 1867. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
