Complex regional pain syndrome (CRPS) is characterized by the presence of regional pain and sensory changes following a predominantly traumatic noxious event. The pain is associated with abnormal skin color, temperature change, abnormal sudomotor activity, and edema. Two types of CRPS can be distinguished: type I, formerly called reflex sympathetic dystrophy, occurs without a definable nerve lesion, and type II, formerly called causalgia, refers to cases where a definable nerve lesion is present (Stanton-Hicks et al., 1995).
The diagnosis of CRPS is clinical but there is still a lack of systematic investigations on many clinical aspects of CRPS (Scola and Schliack, 1991). Sensory impairment has been reported to occur in more than 70% of cases (Boas, 1996). Studies already conducted, mainly concentrated on the quality of sensory abnormalities in the affected limb (Price et al., 1992; Gracely et al., 1996). Price et al. (1992) studied the incidence of cutaneous mechanical allodynia, temporal summation of mechanical allodynia, and thermal hyperalgesia in 17 patients with CRPS and 14 patients with persistent limb pain following trauma. The sensory abnormalities occurred to varying extents, were not reliably associated with each other and could often be completely eliminated by sympathetic block. The authors concluded that, with the exception of persistent pain and mechanical allodynia, no single sensory abnormality can be used as a criterion for the diagnosis of CRPS (Price et al., 1992; Gracely et al., 1996).
In contrast to the detailed studies on the quality of sensory abnormalities in the painful area of the affected limb, there is only little information in the literature on proximal spreading of sensory impairment in patients with CRPS. Sensory abnormalities in patients with CRPS have been reported to extend in a glove or stocking, quadratic or hemilateral distribution (Mascher, 1950; Trostdorf, 1956; Rosen and Graham, 1957).
These phenomena have only been described in case reports and they are regarded to be less specific, subjective, variable and often difficult to validate with clinical testing. Furthermore, sensory abnormalities extending past the painful area of the limb are often regarded as a psychosomatic response since they seem to be inconsistent with known patterns of dermatomal or peripheral distribution (Boas, 1996).
The aims of the present study were (1) to determine the proximal extent of sensory impairment in a larger sample of patients with CRPS type I (2) to determine the frequency of mechanical allodynia and hyperalgesia in these patients; and (3) to investigate other associated clinical features such as motor abnormalities or autonomic changes.
Twenty-four patients suffering from CRPS type I (19 female, five male; mean age: 53 years; range: 27 to 73 years) were examined (clinical data are summarized in Table 1). The diagnosis of CRPS was based on the criteria outlined by Stanton-Hicks et al. (1995)and Boas (1996). All patients had permanent or intermittent pain following an initiating traumatic noxious event. In 15 patients the precipitating event was a bone fracture, in eight patients minor trauma (sprain or strain), and one patient developed CRPS following an operation. Thus, in the majority of patients, deep somatic afferents were affected. The spontaneous pain occurred beyond the territory of a single peripheral nerve, and, in all cases, was disproportionate to the inciting event. All patients reported that they had edema in the affected limb in the early phase after trauma. In two patients, the trauma was associated with severe psychological strain. The first patient had been raped and the second patient was in a severely depressed mood (death of her husband) when she developed CRPS after a minor trauma.
The duration of CRPS varied between 4 and 192 months (mean 49 months). In nine patients, the affected body part was the right hand, in ten patients the left hand. Two patients suffered from CRPS of the right foot, three from CRPS of the left foot.
All patients were treated in the Department of Pain Therapy of our hospital with regional sympathetic blockades and/or plexus anesthesia. Six patients reported persistent improvement of symptoms following sympathetic blockades, five patients reported that pain was temporarily (more than 1 day) ameliorated by sympathetic blockades whereas 13 patients reported no improvement or improvement for less than 1 day. In addition, all patients received physiotherapy and individual pharmacological treatment (Table 1). At the time of our examination, no invasive treatment had been performed. None of the patients had a previous medical history of cerebrovascular or other neurological or psychiatric diseases. Although, in two patients the development of CRPS was associated with severe psychological strain, none of the patients needed psychiatric intervention.
All patients underwent a complete bed-side neurological examination including examination of the cranial nerves, motor system and coordination. To evaluate impairment of sensation, the entire body surface was tested for the perception of light touch using smooth blotting-paper (comparable to cotton wool). In the painful area of the affected limb particular attention was given to mechanical allodynia to light pressure (defined as provocation of pain by a brief and normally innocuous stimulus). Then pinprick sensation was tested using the sharp and blunt side of a wood stick (discrimination between sharp and blunt) and in the affected limb particular attention was given to mechanical hyperalgesia. Temperature testing was performed using a standard device made of metal and plastic. Normal subjects perceive the metal side as being cooler than the plastic side, although there is no major temperature difference between the two sides. In some patients with abnormalities of thermal sensation, additional investigations using test tubes filled with heated (38°C) or cold water (22°C) were carried out. Additionally, stereognosis (objects (key, coin) were placed in the hand and had to be recognized, graphesthesia (different numbers traced bilaterally on the arms, back of hands, trunk, legs and dorsal foot had to be recognized), as well as vibratory sense (using a tuning fork (knuckle's of both hands; ankle of both feet) with a vibration scale fixed at the tip of the arms; measurement in n/8: 0/8, vibration is not perceived; 8/8, vibration is perceived up to the final swinging) were examined. Quantitative measurements were not undertaken.
All patients were independently evaluated by two examiners to see if the findings were consistent. Both examinations were performed on the same day within a few hours. Neither examiner was aware of the colleague's results. In six patients with distinct sensory impairment repeated clinical examinations were performed over 3–6 months.
3.1. Features of spontaneous pain
Seven patients complained of temporary and 17 patients of persistent pain in the affected limb. Fourteen patients (58%) complained of burning pain, ten patients (42%) of aching pain, and two patients (8%) complained of throbbing pain. In the case of 17 patients, pain was mainly localized in the forearm/hand or lower leg/foot. Five patients described pain in the whole limb affected by CRPS and two patients reported spreading of the pain over the ipsilateral half of the body.
3.2. Sensory impairment
Twenty patients (83%) had sensory impairment attributable to CRPS and in 16 of these patients (67%) sensory changes extended past the painful area of the affected limb. Four patients had no sensory deficits (Table 1, patients B2).
Sensory changes observed were (1) hemisensory impairment (2) sensory impairment in the upper quadrant of the body, and (3) sensory impairment limited to the limb affected by CRPS. Sensory impairment mainly affected light touch, pinprick and temperature sensation, whereas, graphesthesia, as well as stereognosis, were preserved in all but one patient and only five patients revealed a mild impairment of vibration sense in the affected limb (Table 1).
3.2.1. Hemisensory impairment
In eight patients (33%), a hemisensory deficit including the face, trunk and limbs ipsilateral to the affected limb could be observed (Table 1, patients A1) (Fig. 1). Seven patients had left-sided and one patient had right-sided CRPS. In six patients, pinprick and temperature sensation were diminished and in two patients completely abolished on the whole half of the body ipsilateral to the limb affected by CRPS and with a border in the midline of the body.
Four patients were aware or partially aware of hemisensory impairment and described spreading of symptoms over the ipsilateral half of the body within a few weeks following trauma. Four patients were unaware of hemisensory deficit and localized major symptoms to the painful affected limb. Although the majority of patients with hemisensory impairment revealed severe CRPS with a high incidence of motor impairment or mechanical allodynia/hyperalgesia, hemisensory impairment could also be demonstrated in one patient who had recovered almost completely from CRPS.
3.2.2. Sensory impairment in the upper quadrant of the body
Sensory impairment of the upper quadrant of the body could be observed in four patients (17%) (Table 1, patients A2) (Fig. 2). In two patients this included the ipsilateral half of the face, in two other patients sensory impairment started below dermatome C2. The lower border of sensory impairment varied between dermatomes T4 and T7. Three patients described that hypesthesia and pinprick as well as temperature discrimination were impaired within this area, whereas one patient described mechanical allodynia and hyperalgesia in the whole upper quadrant of the body. In two patients, sensory impairment with mechanical allodynia in the ipsilateral lower limb (dermatome S1 following lumbar disc prolapse; lateral part of the foot of unknown etiology) could also be found.
3.2.3. Sensory deficits limited to the affected limb
In eight patients (33%), sensory deficits were limited to the affected limb (Table 1, patients B1) (Fig. 3). Four patients had hypesthesia and decreased pinprick as well as temperature sensation in the whole arm affected by CRPS. In four other patients, sensory impairment was limited to the forearm and hand (n=3) or lower leg and foot (n=1). In all patients, sensory changes did not conform to a specific dermatome or to peripheral nerve distribution.
3.2.4. Body side and sensory impairment
Hemisensory impairment and sensory impairment in the upper quadrant were seen in ten out of 13 patients when the initiating noxious event affected the left side of the body, but only in two out of 11 patients when the noxious event was on the right side. This difference is highly significant (P<0.005, chi-square-test).
3.3. Mechanical allodynia/hyperalgesia
Eleven out of twelve patients with hemisensory deficits or sensory impairment in the upper quadrant revealed mechanical allodynia and ten of these patients additionally demonstrated mechanical hyperalgesia in the limb affected by CRPS. In eight out of these patients, allodynia/hyperalgesia was limited to the affected limb. In three other patients, allodynia/hyperalgesia extended past the limb affected by CRPS: in the first patient, with hemisensory impairment, it involved the upper quadrant of the body and the ipsilateral leg and in the second patient, the whole arm, as well as the ipsilateral trunk below T 6 and the leg (following lumbar sympathectomy). One patient with sensory impairment in the upper quadrant revealed mechanical allodynia and hyperalgesia in the whole upper quadrant of the body. In contrast, only two out of 12 patients with sensory impairment limited to the affected limb or without sensory impairment revealed mechanical allodynia, as well as hyperalgesia, in both cases limited to the forearm and hand.
This shows that the incidence of mechanical allodynia and hyperalgesia is significantly higher in patients with generalized sensory impairment than in patients with spatially restricted sensory impairment (11/12 vs. 2/12 patients, p<0.005, Fisher's exact-test).
3.4. Motor impairment and edema
Movement disorders observed consisted of diminution of motor strength of the affected limb (n=10), contractures and trophic changes (n=8), dystonic posturing of the affected limb (n=4), slight differences in tendon reflexes between sides (n=3), myoclonic jerks (n=1), and cogwheel rigidity (n=1). Four patients had a flexion-extension tremor with a low rate of about 4–7/s, which was most prominent when maintaining a posture and sometimes showed features of asterixis (flapping tremor) on clinical examination, affecting not only the hands and fingers but the whole arm of the patient. Only three patients revealed marked edema of the affected limb at the time of our examination. All three patients had hemisensory deficits or sensory deficits in the upper quadrant (Table 1).
The incidence of motor impairment was significantly higher in patients with generalized sensory impairment than in patients with spatially restricted sensory impairment (10/12 vs. 5/12 patients, P<0.05, Fisher's exact-test). Furthermore, motor impairment was highly correlated with allodynia/hyperalgesia, i.e. 12/15 patients with motor impairment had allodynia/hyperalgesia whereas only 1/9 patients without motor impairment had these features (P<0.005, Fisher's exact-test) (Table 2).
3.5. Interrater reliability
When sensory impairment was divided into four groups (hemisensory deficit; sensory impairment in the upper/lower quadrant; sensory impairment in the arm or leg; sensory impairment in the forearm and hand or lower leg and foot), interrater reliability (the percentage of agreement between the two investigators) was 92%. Differences between the two investigators could be observed in two patients (hemisensory deficit vs. sensory deficit in the upper quadrant; sensory impairment in the whole arm vs. forearm and hand). In six patients with repeated examinations, hemisensory impairment persisted over 3–6 months.
3.6. Effect of medication
While there was no correlation between sensory impairment and pharmacotherapy (eight out of 12 patients with hemisensory impairment/sensory deficits in the upper quadrant and seven out of 12 patients with localized or without sensory impairment received medicaments) motor impairment could be observed in a higher percentage of patients receiving medicaments than in drug-free patients (12/15 vs. 3/9) and in all patients treated with opioids (8/8).
In the present study, we found that the area of sensory impairment in patients with chronic CRPS frequently extends past the area affected by pain. In 50% of patients, sensory abnormalities in the upper quadrant of the body or a hemisensory impairment could be observed which mainly affected temperature and pain sensation, predominantly in left-sided CRPS. This may indicate that the development of the more generalized sensory impairment may be related to cortical dominance. Patients with hemisensory impairment or sensory deficits in the upper quadrant demonstrated a significantly higher frequency of mechanical allodynia/hyperalgesia, as well as movement disorders, than patients with sensory impairment restricted to the affected limb. On the other hand, hemisensory impairment could still be demonstrated in one patient who had recovered almost completely from CRPS.
The somatosensory examination performed in our study relies on a subjective response to given stimuli. The presence of pain may alter cutaneous sensation and may influence detailed sensory examination (Nathan, 1960). Therefore, it must be questioned if the sensory impairment described reflects the dysfunction of peripheral or central somatosensory pathways or whether psychological influences might play a greater role.
There is no doubt that psychological processes may be responsible for a significant component of CRPS. Depression, anxiety, or stress may influence CRPS through alpha-adrenergic mechanisms (sympathetic stimulation) and may increase nociception (Covington, 1996). However, apart from this, there is no convincing evidence that CRPS is a psychogenic condition or that certain personality traits predispose one to develop CRPS (Lynch, 1992; Covington, 1996). Patients in our study had no previous history of psychiatric disease and only two patients had severe psychological strain associated with the initiating trauma. Hemisensory deficits were observed in our group of patients following injuries associated with severe psychological stress (i.e. following rape), as well as in patients with major trauma and without obvious psychological problems.
Hemisensory deficits, as observed in the present study, have been described after large lesions of the contralateral parietal cortex or after contralateral thalamic lesions. In particular, the thalamus has a pivotal role in sensory processing and receives sensory input from the whole contralateral part of the body, including the face. Lesions in the ventral posterior nuclear group of the thalamus are connected with a contralateral hemisensory deficit which mainly affects tactile, temperature and pain sensation and may be associated with a central pain syndrome (Dejerine and Roussy, 1906; Bogousslavsky et al., 1988; Bowsher, 1995). Sensory impairment in the upper quadrant of the body has been described after median nerve injury of the upper arm and quadrant hypohydrosis is known to occur following stellate ganglion lesion (Pette, 1927; Bonica, 1959; Gross, 1974).
Several authors have suggested that pathophysiological mechanisms in the central nervous system play an important part in producing CRPS (Livingston, 1943; Sunderland, 1976; Schott, 1986; Baron et al., 1996; Jänig, 1996). Current pathophysiological concepts on central disturbances in CRPS concentrate on abnormalities of pain and sensory processing in the spinal cord (Baron et al., 1996). However, more centrally located structures are likely to play an important part in the pathophysiology of CRPS. The hemisensory deficit observed in our patients with CRPS might indicate functional disturbances in noxious event processing in the thalamus. The increased frequency of hemisensory impairment in patients with left-sided CRPS might be due to the specialization of the right hemisphere for somatic state monitoring, including mediation of pain (Nasreddine and Saver, 1997). Right-hemispheric lesions are associated with a syndrome of ‘neglect’ (Heilman and Van Den Abell, 1980). Due to the excessive sensory input from the painful area of the affected limb, in left-sided CRPS a functional ‘neglect’ syndrome with decreased perception of the sensory input from the remaining contralateral half of the body might be the consequence.
Motor impairment in patients with CRPS has been described by several authors (Marsden et al., 1984; Schott, 1985; Schwartzman and Kerrigan, 1990; Deuschl et al., 1991). The main features reported were also observed in our patient group. Motor impairment usually has the same distribution as the pain but, as for sensory abnormalities, it can spread proximally and may involve even the ipsilateral side of the body. The high frequency of movement disorders in patients with more generalized sensory impairment, as well as mechanical allodynia, indicates that alterations in central sensory and pain processing might also be responsible for the motor impairment. However, as movement disorders could be observed in a higher percentages in patients receiving medicaments, in particular opioids, an effect of medication can not be ruled out.
In conclusion, motor as well as sensory impairment observed in patients with CRPS most probably relies on functional alterations in central processing of noxious events. This may explain the variability of the motor and sensory symptoms observed. Further quantitative studies have to clarify whether sensory findings are influenced by therapeutic procedures, i.e. sympathetic blockades, and if sensory abnormalities observed have prognostic implications.
Baron, R., Blumberg, H. and Jänig, W., Clinical characteristics of patients with CRPS in Germany with special emphasis on vasomotor function. In: W. Jänig and M. Stanton-Hicks (Eds.), Reflex Sympathetic Dystrophy: A Reappraisal, IASP Press, Seattle, WA, 1996, pp. 25–48.
Boas, R.A., Complex regional pain syndromes: symptoms, signs and differential diagnosis. In: W. Jänig and M. Stanton-Hicks (Eds.), Reflex Sympathetic Dystrophy: A Reappraisal, IASP Press, Seattle, WA, 1996, pp. 79–92.
Bogousslavsky, J., Regli, F. and Uske, A., Thalamic infarcts: cinical syndromes, etiology and prognosis, Neurology, 38 (1988) 837-848.
Bonica, J.J., Clinical applications of diagnostic and therapeutic nerve blocks. Charles C. Thomas, Springfield, IL, 1959.
Bowsher, D., The management of central post-stroke pain, Postgrad. Med. J., 71 (1995) 598-604.
Covington, E.C., Psychological issues in reflex sympathetic dystrophy. In: W. Jänig and M. Stanton-Hicks (Eds.), Reflex Sympathetic Dystrophy: A Reappraisal, IASP Press, Seattle, WA, 1996, pp. 191–216.
Dejerine, J. and Roussy, G., Le syndrome thalamique, Rev. Neurol., 14 (1906) 521.
Deuschl, G., Blumberg, H. and Lücking, C.L., Tremor in reflex sympathetic dystrophy, Arch. Neurol., 48 (1991) 1247-1252.
Gracely, R.H., Price, D.D., Roberts, W.J. and Bennett, G.J., Quantitative sensory testing in patients with complex regional pain syndrome (CRPS) I and II. In: W. Jänig and M. Stanton-Hicks (Eds.), Reflex Sympathetic Dystrophy: A Reappraisal, IASP Press, Seattle, WA, 1996, pp. 151–172.
Gross, D., Pain and autonomic nervous system, Adv. Neurol., 4 (1974) 93-103.
Heilman, K.M. and Van Den Abell, T., Right hemisphere dominance for attention: the mechanism underlying hemispheric asymetries of inattention (neglect), Neurology, 30 (1980) 327-330.
Jänig, W., The puzzle of ‘reflex sympathetic dystrophy’: mechanisms, hypotheses, open questions. In: W. Jänig and M. Stanton-Hicks (Eds.), Reflex Sympathetic Dystrophy: A Reappraisal, IASP Press, Seattle, WA, 1996, pp. 1–24.
Livingston, W.K., Pain mechanisms: A physiologic interpretation of causalgia and its related states. The Macmillan Company, New York, 1943, pp. 83–127.
Lynch, M.E., Psychological aspects of reflex sympathetic dystrophy: a review of the adult and paediatric literature, Pain, 49 (1992) 337-347.
Marsden, C.D., Obeso, J.A. and Traub, M.M., Muscle spasms associated with Sudeck's atrophy after injury, Br. Med. J., 288 (1984) 173-176.
Mascher, W.L., Über die Rolle des zentralen Nervensystems für die Entstehung des Sudeck-Syndroms, Nervenarzt, 21 (1950) 67-74.
Nasreddine, Z.S. and Saver, J.L., Pain after thalamic stroke: right diencephalic predominance and clinical features in 180 patients, Neurology, 48 (1997) 1196-1199.
Nathan, P.W., Improvement in cutaneous sensibility associated with relief of pain, J. Neurol. Neurosurg., 23 (1960) 202-206.
Pette, H., Quadrantensyndrome, Dtsch. Zschr. Nervenheilk., 100 (1927) 143-148.
Price, D.D., Long, S. and Huit, C., Sensory testing of pathophysiological mechanisms of pain in patients with reflex sympathetic dystrophy, Pain, 49 (1992) 163-173.
Rosen, P.S. and Graham, W., The shoulder-hand syndrome: historical review with observations on 73 patients, Can. Med. Assoc. J., 77 (1957) 86-91.
Schott, G.D., The relationship of peripheral trauma and pain to dystonia, J. Neurol. Neurosurg. Psychiatry, 48 (1985) 698-701.
Schott, G.D., Mechanisms of causalgia and related clinical conditions: the role of the central and of the sympathetic nervous system, Brain, 109 (1986) 717-738.
Schwartzman, R.J. and Kerrigan, J., The movement disorder of reflex sympathetic dystrophy, Neurology, 40 (1990) 57-61.
Scola, E. and Schliack, H., Das posttraumatische Sudeck-Syndrom, Dtsch. Ärztebl., 88 (1991) 1888-1890.
Stanton-Hicks, M., Jänig, W., Hassenbusch, S., Haddox, J.D., Boas, R. and Wilson, P., Reflex sympathetic dystrophy: changing concepts and taxonomy, Pain, 63 (1995) 127-133.
Sunderland, S., Pain mechanisms in causalgia, J. Neurol. Neurosurg. Psychiatry, 39 (1976) 471-480.
Trostdorf, E., Die Kausalgie: Ein Beitrag zur Frage der Schmerzen bei peripheren Nervenverletzungen. Thieme, Stuttgart, 1956, pp. 62—81.