This is an intense, usually persistent, and burning regional pain. The hyperpathic pain was constant in 81% of our 824 patients. Blümberg and Jänig reported the pain constant in 75%, and intermittent in 25% of their patients. The pain is out of proportion to the severity of trauma.

Simple tactile stimulation of the involved area originating the hyperpathia may be accompanied by objective signs of rise in pulse and BP. In addition, the hyperpathic pain is accompanied by a regional mild hypoesthesia to touch and pain. This hypoesthesia is not in the sensory distribution of somatic nerve roots.

By virtue of exclusive function of thermal regulation, these thermal sensory nerve fibres have an affinity to the anatomical structures of arterioles and arteries (heat source). As a result, dysfunction of these sensory nerve fibres does not show a dermatomal, but a thermatomal sensory nerve distribution. This is a sensory loss usually in the distribution of brachial, femoral, carotid, or mesenteric arteries.

There are actually three different types of sensory loss:

1. Dermatomal.

2. Thermatomal.

3. Glove and stocking distribution .

The dermatomal sensory loss is seen in radiculopathies, and other forms of somatic sensory nerve damages. The thermatomal, neuropathic (microvascular) pain distribution is usually seen in CRPS, diabetic neuropathy, and post herpetic neuralgia. This type of thermatomal sensory loss should not be mistaken for the classical glove and stocking sensory loss - due to primary or secondary gain (malingering or conversion reaction) - which limits itself to joints such as the wrist, knee, or shoulders in a glove and stocking distribution. Only a careful examination of touch and pain sensations separates each of these three sensory types.

The regional hyperpathia (protopathia) is mainly due to summation of repetitive stimulation of the thermoreceptors, which results in a tendency for spread of hyperpathia to proximal and distal portions of the extremities. In advanced, complicated cases, it may lead to horizontal and vertical spread to other extremities. This spread may play a major role in aggravation of CRPS. This phenomenon is due to stimulation of the uninhibited c-fibres perpetuating the sensitization of afferent receptor nerves in the spinal cord . Other factors such as sympathectomy and surgical procedures are also instrumental in the spread of CRPS.

Activation of thermal c nociceptor sensory pain fibres plays a major role in hyperpathic pain. The afferent small c-fiber system has a tendency to be inhibited by the larger A-fiber myelinated somatosensory nervous system. Lack of such inhibition (i.e., lack of simple touch or avoidance of tactile contacts due to hyperpathic pain) can result in increased firing of the afferent neuropathic pain fibres with secondary pathological efferent sympathetic discharge, and further sensitization of the layers I and II of the gray matter of the spinal cord. This is not a simple up-regulation, but a dysfunction of the sympathetic system. This dysfunction explains the reason for failure of sympathectomy. Sympathectomy aggravates the already existing sympathetic dysfunction.

Allodynic pain is elicited by a stimulus that usually does not cause pain: e.g., a simple breeze, bed sheet contact, and other types of mild tactile stimulation. In CRPS, there is a tendency for sensitization of the involved skin surface. Typically, the patient avoids any type of sensory stimulation, and protects the allodynic area with the help of wrapping the extremity with a cloth or a glove. In extreme cases, the patient may avoid taking a shower or bathing for months. Mechanoallodynia is usually mediated by A-beta low-threshold mechanoreceptors which are small myelinated nerve fibres. They do not respond successfully to pure sympathetic ganglion blocks. They are more likely to respond to epidural or paravertebral blocks.

Application of ice exaggerates vasoconstriction, coagulates and damages the myelinated nerve fibres, aggravates the nerve damage and enlarges the area of Mechanoallodynia resulting in bias toward SIP rather than SMP. Procedures such as carpal tunnel surgery for inflammation due to CRPS, arthroscopy, or exploratory surgical procedures, also aggravate the sensory damage and reproduce a similar allodynic phenomenon.

In our study the application of ice for 2 months or longer resulted in a higher percentage of stage III-IV as compared to no ice or less than 2 months treatment(P<0.001). As ice selectively damages and coagulates the myelinated nerve fibers(which are rich in lipids), the allodynia is gradually transformed to thermatomal hyposthesia augmented by ice induced hypothermia.

The causalgic pain in CRPS II syndrome, is characterized by attacks of "lightning," "stabbing," "electric shock," "prickling," "deep itching, " and "jerking " type of pain. In later stages, the pain is accompanied by myelogenic response of myoclonic jerks of the extremity or of the trunk - as well as atonic (akinetic) falling attacks - due to myelopathic sensitization.

Recently, Blümberg , Jänig and Koltzenburg have discovered a new source of pain. It originates from the deep chemoreceptor c-fibres in muscle and bone. These chemoreceptors become activated with inactivity. Blümberg and Jänig reported incidence of pain with inactivity in 65% of their patients. The deep spontaneous pain associated with inactivity showed a higher incidence in 75% of their patients. In our series of 824 patients, the deep pain sensation on arousal in the morning (associated with inactivity) was noted in 79% of patients vs pain with activity in 63% of patients. The patients describe this type of pain as deep, itching, and intolerable. With increasing inactivity (e.g., use of wheelchair), this deep pain arouses the patient 2 ½ times more frequently than the ambulatory patients. Intermittent walking reduces the incidence of deep pain.

2. EFFERENT (MOTOR) RESPONSE

The second diagnostic principle is efferent motor dysfunction in the form of vasoconstriction, flexor spasm, movement disorder including dystonia and tremor. In more advanced cases, myoclonic jerks due to spinal cord sensitization and de-afferentation develop. These myelogenic myoclonic jerks are due to the enlargement of peripheral receptive fields of central pain projecting neurons in superficial laminae I and II of the dorsal horn. The sensitization and the enlargement of laminae I and II excitatory fields is due to relatively long-term, uninhibited dysfunctional afferent sensory nerve input to the deeper layers 4 and 5 of dorsal horn of the spinal cord. The deeper layers 4 and 5 exert inhibitory function on superficial laminae I and II excitatory internuncial cells. The small granular primary sensory neurons of the superficial layers 1 and 2 normally possess small receptor fields that respond mainly to the c-nociceptors and A-delta fibres. These neurons synthesize neuropeptides that are transported through the afferent fibres centrally and peripherally. The chronic repetitive stimulation of these sensory nerve fibres is accompanied by release of chemicals such as calcitonin G-related peptide (CGRP), substance P and Nitric Oxide (NO). Excessive somatostatin and substance P release can potentially damage and reduce the inhibitory function of the granular cells in the deeper layers 4 and 5 of dorsal horn neurons leading to sensitization, deafferentation, and aggravation of hyperpathic pain and allodynia. The same phenomenon leads to myelogenic myoclonus. CGRP exerts an inhibitory effect on the excitatory neurokines substance P and somatostatin. Dynorphin activation and break down by inflammation contributes to sensitization of the dorsal horn. Lack of inhibition of the larger afferent nerve fibres and secondary disturbance of inhibition of wide dynamic range function of the layers 4 and 5 results in increased excitation of the efferent spinal cord nerve cells. The end result is disturbance of spinal cord plasticity, deafferentation, attacks of myelogenic myoclonic jerks, akinetic attacks, as well as tremor and other forms of movement disorder. Simple somatic peripheral nerve injury is not enough to cause tremor. The above- mentioned central sensitization at spinal cord level is required to lead to dystonia and tremor.

The motor dysfunction may manifest itself in the form of dystonic flexor spasm, flexor deformity , toe walking, and pronation of the hand or foot (equino varus). The dystonic flexor spasm seems to be due to the primitive withdrawal response to the pain source. Tremor is not uncommon. Blümberg and Jänig have reported tremor and other movement disorders in over 80% of CRPS patients. Veldman, et al, have noted movement disorder in 95% of 829 patients. In our series of 824 patients, the incidence was 78%. Application of cast causes immobilization and stimulation of the deep mechanoreceptors. These "silent sleeping nociceptors" become activated with rest and inactivity. This in turn, leads to pain, edema and movement disorder. Cardoso and Jankovic have reported 11 cases of patients suffering from CRPS who developed Parkinsonian type of tremor following application of cast to the extremity. The cast becomes harmful if the extremity is edematous and inflamed due to neuroinflammation of the original trauma or surgery.

Myoclonic jerks may be a manifestation of de-afferentation and sensitization of spinal cord due to long-term afferent cytokines damage to the inhibitory granular cells in layers I and II. As such, they develop in later stages of the disease. Any form of immobilization (cast, wheelchair, etc.) contributes to this phenomenon. The myoclonic jerks are seen in patients undergoing withdrawal of opioids (rebound phenomenon).

In 38 of our 824 patients suffering from CRPS due to spinal cord injury myoclonic jerks were invariably observed by the examiner. Yet, more than 3/4 of the CRPS patients has a history of myoclonic jerks- not observed at the time of examination. In addition, myoclonic jerks were present in 44 of 63 CRPS patients secondary to electrical injury. This may be due to electricity going through the path of least resistance (afferent c-fibers) and secondarily originating spinal cord dysfunction. Myoclonic jerks are a long term complication of limb amputation (10 of 11 amputees among our 824 patients).

3. DISTURBANCE OF IMMUNE SYSTEM

(NEUROGENIC INFLAMMATION)

The third diagnostic principle is neuropathic pain, including CRPS I, is complicated by inflammation in varying degrees. This inflammation was first reported by Mitchell (1864)as "shiny skin" and, later on, by Sudeck (1942). The neurogenic inflammation results in bullbous lesions, sterile abscess, edema and impingement of the nerves at the wrist, elbow, thoracic outlet and ankle areas - resulting in the disease being mistaken for conditions such as carpal tunnel, thoracic outlet (TOS), tarsal tunnel syndromes, and myofascial syndrome. The well-intended surgical procedure to relieve such entrapment neuropathies may in turn aggravate the inflammation by virtue of surgical trauma becoming a new source of neuropathic pain.

The inflammation is another manifestation of dysfunctional sympathetic system. The sympathetic system is responsible for immune system regulation (Arnason, 1993). As a result, the patient may develop bouts of unexplained fever, edema, attacks of subcutaneous bleeding, neurodermatitis, bulbous lesions , pelvic inflammatory disease (PID), or interstitial cystitis. Inflammation may cause development of subcutaneous skin nodules, pulmonary nodules, laryngitis, difficulty with phonation, attacks of hacking cough and hematemesis. In late stages it can cause elephantiasis, subcutaneous bleeding, bullous deep ulcerative lesions involving the skin as a manifestation of disturbance of the immune system. It can be mistaken for infection, osteomyelitis, dermatitis, and cystitis. Treatment with antibiotics provides no relief.

The inflammation is usually intermittent, and is not consistently present. Only a careful history taking can document previous attacks of inflammation.

4. EMOTIONAL ASPECTS OF CRPS:

LIMBIC SYSTEM DYSFUNCTION

The forth and final diagnostic principle is emotional disturbance in CRPS. In contrast to somatic sensory nerves, the sensory neuropathic nerve fibres responsible for the development of CRPS do not end up in the contralateral neocortical parietal sensory cortex. Instead, according to Bennarroch, over 90% of these sensory nerve impulses terminate in the limbic system. More over, positron emission tomography (PET) demonstrates a significant cerebral insular and limbic activation during painful stimulation of neuropathic pain. The vicious circle of chronic neuropathic pain leading to disturbance of plasticity, as well as inflammation, causes further long term potentiation (LTP) of pain impulse and nerve stimulation in higher centers in the limbic system. This leads to insomnia, agitation, depression, poor memory and poor judgment. The above neurophysiological observations explain the fact that practically every patient suffering from CRPS demonstrates some degree of limbic system disturbance. In our study of 824 patients, one or more of the limbic system dysfunctions were present in every case except three. These consisted of insomnia (92%), irritability, agitation, anxiety (78%), (depression (73%), poor memory and concentration (48%), poor judgment (36%), and panic attacks (32%). Understanding the nature of emotional components of RSD spares the patient from misdiagnosis and improper treatment.

The majority of chronic pain patients suffer from the somatic type of nerve damage or dysfunction, with no neurovascular involvement, such as carpal tunnel syndrome (CTS), thoracic outlet syndrome (TOS), tarsal tunnel syndromes, rotator cuff injury, disc herniation, Morton’s neuroma, fibromyalgia (FM), or myofascial syndrome(MFS). In a small minority of cases, the patients suffer from similar syndromes caused by neuropathic/ neurovascular damage or dysfunction mimicking the above syndromes. The clinician has a tendency to try to explain the manifestation of such neuropathic pain (such as CRPS) by categorizing the disease as somatic syndromes mentioned above. Even normal EMG/NCV tests do not convince the surgeon to cancel the carpal tunnel surgery. Even the purplish skin discoloration or asymmetrical sweating does not bring up a question about the presumptive presurgical diagnosis of CRPS. The loser is the patient who has to cope with the new source of neuropathic pain at the surgical scar area as well as the referred pain and the spread of CRPS to other parts of the body due to the trauma of surgery.

The main hurdle in diagnosis is the fact that majority of physicians do not consider CRPS in their list of differential diagnosis. This syndrome is commonly over-or under- diagnosed. In our series of 824 patients, CRPS was over-diagnosed in 134 (16%) of cases, and under diagnosed in 173 (21%) of cases (Table 2). The 134 over-diagnosed cases have already been excluded from this study.

A syndrome as complex - and as potentially variable in symptomatology and temporal course as RSD- cannot be expected to be diagnosed with a single laboratory test. CRPS is a syndrome and should be diagnosed by inclusion (the above- outline 4 principles) rather than by exclusion.

1. Scintigraphic triphasic bone scanning (SBS) has been the popular test of choice for the diagnosis of CRPS in the past three decades. Whereas earlier literature has described the SBS as highly sensitive and specific in establishing the diagnosis of CRPS, a recent review of medical literature by Lee and Weeks has shown this test to be positive in approximately 55% of the cases, which is quite close to a random statistical yield. Chelimsky et al found this test abnormal in no more than 25% of CRPS patients.

2. Diagnostic nerve blocks phentolamine and guanethidine are usually positive in the early stages and gradually lose their sensitivity.

3. EMG and NCV: A. CRPS I (RSD): Realizing that CRPS I is due to dysfunction of poorly myelinated or unmyelinated sensory nerve fibres, EMG and NCV cannot be expected to show any abnormality. NCV measures the velocity and function of the large myelinated fibres, which are not usually involved in CRPS I. EMG/NCV cannot identify disturbance of small sensory or autonomic nerve fibres . Diagnosing CRPS with the help of EMG and NCV is similar to diagnosing a viral infection with a standard- rather than an electron microscope.

4. Computed tomography and magnetic resonance imaging (MRI) are not expected to detect the damage to the microscopic nerve fibres in the wall of blood vessels, and usually do not show any abnormalities in CRPS.

5. Quantitative sudomotor axon reflex test (QSART) studies the cholinergic sudomotor function of the sympathetic system. It does not address the norepinephrine dysfunction. It has a high degree of sensitivity and specificity in detecting post-ganglionic dysfunction of cholinergic (parasympathetic) sudomotor nerves.

6. Infrared thermal imaging (ITI): The infrared imaging has a limited application in neurology. It can study and compare subtle temperature changes in different parts of the body. Like any other test outlined above, it cannot "diagnose" CRPS, but can identify areas of damage (hyperthermia) versus irritation (hypothermia) of sympathetic nerves. The ITI is quite sensitive in pointing to the function of skin temperature which is the exclusive domain of sympathetic system. Cold stress-ITI may provide additional useful information. Limitations of ITI: The thermal imaging shows any old or new sympathetic nerve damage or dysfunction, thus confusing the examiner and demanding careful and proper clinical correlation. In addition, as the disease becomes chronic and the sympathetic dysfunction becomes bilateral, the ITI shows identical bilateral temperature changes causing difficulty in diagnosis. The same phenomenon causes confusion in interpretation of other tests in CRPS. Infrared thermal imaging is useful in identifying the area of maximal damage as follows:

A. In the area of original nerve damage (e.g., hand or foot), the hyperthermia points to damage and paralysis of vasoconstrictive function of sympathetic system (the central hyperthermic area). The central hyperthermia usually points to the apex of damaged tissue resulting in heat leakage, as well as accumulation of substance P and nitric oxide. This is an important therapeutic clue to avoid further trauma. Traumatic procedures such as surgical exploration, nerve blocks, Clonidine Patch, Capsaicin, or EMG needle insertion should not be applied to the damaged hyperthermic area in the extremity which may lead to further damage and aggravation of the condition. In acute stage, the damaged area is hyperthermic. After a few weeks, the hyperthermic area shrinks. In some cases the hyperthermia persists due to permanent damage to sympathetic nerve fibers. This bodes a poor prognosis.

B. Hyperthermia in referred pain areas (e.g., paravertebral nerves ) is due to SP and NO accumulation, but does not necessarily point to the origin of the injury.

C. Virtual Sympathectomy: After more than a dozen stellate, or lumbar ganglion nerve blocks, the repetitive needle insertion traumatizes the ganglion enough to result in permanent hyperthermia in the extremity ("Virtual Sympathectomy"). Infrared imaging identifies this phenomenon, and spares the patient from further damage.

7. Laser Doppler Flow Study is a sensitive test for the study of capillary circulation. It studies a small area of the body limiting its overall extent of information. This test has demonstrated sympathectomy to be ineffective in providing increased circulation in extremity after exposure to cold.

8. Quantitative thermal sensory evoked response test (QST) is sensitive and useful in studying the functions of c-thermoreceptors and A-beta mechanoreceptors in CRPS. This test identifies the threshold of cold and heat touch and pain sensations. This test may be abnormal in CRPS patients (cold hyperalgesia) and in erythromelalgia (heat hyperalgesia). The test has been well standardized.

The CRPS has been divided into different stages. Depending on nature of injury, the stages vary in their duration. In the 17 patients suffering from venipuncture CRPS in our series, deterioration from stage I to stage III was measured in a few weeks up to less than 9 months. This is in contrast with CRPS in children in whom stages would stagnate, reverse or improve slowly.

The STAGE I , is a sympathetic dysfunction with typical thermatomal distribution of the pain . The pain may spread in a mirror fashion to contralateral extremity or to adjacent regions on the same side of the body. In stage one, the pain is usually SMP in nature.

In STAGE II, the dysfunction changes to dystrophy manifested by edema, hyperhidrosis, neurovascular instability with fluctuation of livedo reticularis and cyanosis - causing change of temperature and color of the skin in matter of minutes. The dystrophic changes also include bouts of hair loss, ridging, dystrophic, brittle and discolored nails, skin rash, subcutaneous bleeding, neurodermatitis, and ulcerative lesions. Due to the confusing clinical manifestations, the patient may be accused of factitious self-mutilation and "Münchausen syndrome." All these dystrophic changes may not be present at the same time nor in the same patient. Careful history taking is important in this regard.

In STAGE III, the pain is usually no longer SMP and is more likely a sympathetically independent pain(SIP). Atrophy in different degrees is seen. Frequently, the atrophy is overshadowed by subcutaneous edema. The complex regional pain and inflammation spread to other extremities in approximately 1/3 of CRPS patients. At stage II or III it is not at all uncommon for CRPS to spread to other extremities. At times, it may become generalized. The generalized CRPS is an infrequent late stage complication. It is accompanied by sympathetic dysfunction in all four extremities as well as attacks of headache, vertigo, poor memory, and poor concentration. The spread through paravertebral and midline sympathetic nerves may be vertical, horizontal, or both. The original source of CRPS may sensitize the patient to later develop CRPS in another remote part of the body triggered by a trivial injury. The ubiquitous phenomenon of referred pain to remote areas (e.g., from foot or hand to spine) should not be mistaken for the spread of CRPS.

At STAGE III, inflammation becomes more problematic and release of neuropeptides from c-fiber terminals results in multiple inflammatory and immune dysfunctions. The secondary release of substance P may damage mast cells and destroy muscle cells and fibroblasts.

STAGE IV:

1. Failure of the immune system, reduction of helper T-cell lymphocytes and elevation of killer T-cell lymphocytes.

2. Intractable hypertension changes to orthostatic hypotension.

3. Intractable generalized edema involving the abdomen, pelvis, lungs, and extremities.

4. Ulcerative skin lesions which may respond to treatment with  I.V. Mannitol, I.V. Immunoglobulin, and ACTH treatments. Calcium channel  blockers such as Nifedipine may be effective in treatment.

5. High risks of cancer and suicide are increased.

6. Multiple surgical procedures seem to be precipitating factors for development of stage IV.

The stage IV is almost the flip side of earlier stages, and points to exhaustion of autonomic and immune systems. Ganglion blocks in this stage are useless and treatment should be aimed at improving the edema and the failing immune system. Sympathetic ganglion blocks, alpha blockers, including Clonidine, are contraindicated in stage IV due to hypotension. Instead, medications such as Proamantin (midodrin) are helpful to correct the orthostatic hypotension.

Dogmatic reliance on staging is somewhat artificial in nature. Each patient follows a different course. In children and teenagers, the prognosis is excellent and stages need not develop with passage of time due to the fact that their rich cerebral growth hormone, sex hormone and endorphin formation prevent deterioration. The same logic applies to pregnant women. With early treatment, the disease is frozen at stage one. Even patients suffering from stages II or III revert to stage I with proper treatments. The reverse is true: unnecessary surgery, as an stressor can cause rapid regression from stage I to III, as well as spreading the disease to other extremities.

The main goal of treatment is reversal of the course, amelioration of suffering, return to work if at all possible, avoiding surgical procedures, and improvement of quality of life. The key to success is early diagnosis and early assertive treatment. Lack of proper understanding and proper diagnosis leads to improper treatment with poor outcome. There is a desperate need for future research in the treatment of CRPS. Delay in diagnosis is a factor in therapeutic failure. According to Poplawski, et al, treatment, and its results, are hampered by delay in diagnosis. Early diagnosis (up to 2 years) is essential for achieving the goal of successful treatment results. Simple monotherapy with only nerve block, only Gabapentin, or otherwise, is not sufficient for management of CRPS. Treatment should be multidisciplinary and simultaneous: effective analgesia, proper antidepressants to prevent pain and insomnia, physiotherapy, nerve blocks, proper diet, when indicated channel blockers, and anticonvulsant therapy should be applied early and simultaneously. Administration of piece-meal, minimal treatments is apt to fail.

Proper physical therapy is at the top of the list for proper treatment. In this regard, in neuropathic pain, "no pain is all gain" - not the opposite. Any activity that aggravates the neuropathic pain, should be avoided. Distress of pain aggravates the sympathetic dysfunction. The patient is instructed to frequently change positions. Usually, the major aggravators of the pain are inactivity, distressful overdoing of exercise, or repetitive strain injury (RSI).

Basbaum, and others have demonstrated extensive lesions affecting large myelinated axons secondary to ice exposure. These lesions are in the form of Valerian degeneration and segmental demyelination. Cold injuries, frost bites and heat burns are common iatrogenic causes of peripheral neuropathic pain. Heat or cold therapy with warm or cold water should not be mistaken for freezing ice or boiling water exposure. Obviously, ice and hot water are damaging and should be avoided. Temporary use of ice is the treatment of choice for acute but not chronic pain. The repetitive application of ice in chronic pain patients causes cold skin due to vasoconstriction followed by vasodilation usually lasting about 15 minutes. In our study of 824 patients, 34 patients were exposed to ice treatment for less than 2 months versus 226 patients exposed to ice treatment for more than 2 months. The group with over two months exposure to ice 52% ended up in Stages III-IV versus 30% in the less than 2 month exposer to ice (P<0.001). Conversely, only 7% of the group with longer exposer were in Stage I versus 38% in the group with shorter exposure(P<0.001).

If at all possible, the CRPS patient should not be hospitalized unless it is absolutely necessary (such as for emergency surgery). The usual hospital policy of enforced bed rest aggravates the CRPS. The inactivity results in up regulation and activation of the sleeping nociceptors (deep chemoreceptors in bone and muscle), with secondary deep, intolerable pain. The patient is instructed to stay in bed no more than 8-9 hours a night and to try to walk before going to bed. If sitting or laying down cause pain, the patient is instructed to get up and move around. If walking or any type of exercise causes pain, the patient is to rest frequently. Treatment of osteopenia requires ambulation and weight bearing. The use of wheelchair, walker and other assistive devices should be discontinued.

Opioids play a major role in management of pain and inflammation in peripheral and central nervous system. The endogenous ligands-opioid peptides (endorphins) are expressed by resident immune cells in peripheral tissues. Depriving the patient of proper pain medication can aggravate the immune system dysfunction. The selection of proper opiates for treatment of CRPS is quite critical. Both opioid agonists and mixed opioid agonist-antagonists have been used for treatment of pain in such patients. Opiates are considered effective in treatment of neuropathic pain. However, due to the complexity and multiple origins of the pain in CRPS, in some patients the opioid agonists are not as effective. Morphine does not consistently reduce the neuropathic pain.

Morphine (0.1-1mg per kg IV) may increase the localization threshold of lesioned limb pressure and may decrease the chronic pain score. Morphine may decrease mechanoallodynia in the diabetic rat, but the effective doses have to be quite high in the range of 2-4mg per kg IV which are too high for human application. Long term use of opioid agonists has the potential of tolerance and dependence, impairment of physical function, and depression. Yet, 83% of pain specialists have been reported in 1992 to maintain chronic non-cancer pain patients on these medications. This percentage has grown far higher since then: of 824 patients in this study, only 36 (4.3%) had not receive long term opioid agonists therapy. Moreover, the present trend is for poly- pharmacy of opioids in high doses. Such high doses exceed the optimal therapeutic window for analgesia.

The therapeutic window refers to the fact that opiates, similar to anticonvulsants, are most effective in their therapeutic range. Above and below this window they are ineffective.

The opioid agonists such as morphine, fentanyl, etc, have been found ineffective against the abnormally fluctuating reaction to thermal allodynia (neurovascular instability), while retaining anti-nociceptive activity against painful thermal stimuli (heat hyperalgesia). Long term use of Morphine suppresses many specific functions of the immune system. Both acute and chronic application of Morphine strongly suppress the T-cell immune functions. Morphine may interfere with the development of antibody - antigen immune function. Due to the fact that many cells and organs related to the immune system have shown opiate receptors, Morphine has the potential of directly affecting and altering many immune processes. Morphine may affect and suppress noxious stimulus-evoked fos protein-like immunoreactivity. Morphine and other similar opioid agonists bind to opioid receptors in limbic system (temporal lobe), affecting memory and mood.

Long term application of opioid agonists (e.g. morphine) suppresses the formation of endorphins (Table 2).

Contrary to the common concept, large doses of opiates usually disrupt the natural sleep pattern. It is true that opiates induce excessive sedation in 24 hours. However, the nocturnal sleep pattern is interrupted every few hours due to withdrawal phenomenon , leaving the patient tired and sleepy during the day. The use of proper antidepressants and adherence to the above mentioned therapeutic window help correct this problem.

Table 2*

b. Acid rain effect: alcohol as well as exorphins flood the brain cells and hamper their ability to form the dirunal hormones needed for alertness, sleep, tranquility, and antidepressant effects.

Table 2*- From:Chronic Pain: Reflex Sympathetic Dystrophy: Prevention and Management. CRC Press, Boca Raton, Florida 1993.

The above side effects of long-term treatment with opioid agonists leave the door open to search for more effective opiates. Buprenorphine, an opiate agonist-antagonist, is a strong analgesic without causing dysphoria, or dependence. Sublingual Buprenorphine has been used successfully for detoxification from Cocaine, Heroin and Methadone dependence. Buprenorphine is a Class V narcotic in contrast to Morphine, Methadone or Fentanyl, which are Class II. Within the proper therapeutic window, Buprenorphine (2-6mg/day) and Butorphanol (up to 14 mg/day), act as opioid antagonists by occupying only mu and delta receptors. In higher than therapeutic doses, they fill the Kappa receptors as well, changing said drugs to pure opioid agonists and resulting in problems of rebound and tolerance. Within 2-6mg per day, Buprenorphine occupies mu and delta opioid receptors, but kappa receptor is not occupied and is capable of receiving endorphins. When all 3 opioid receptors are occupied, endorphins cannot bind to them. Consequently, endorphins formation is ceased, leading to dependence and tolerance.

The Harvard group and others have found Buprenorphine to act as an antidepressant leading to "clinically striking improvement in both subjective and objective measures of depression." This is in contrast to the common depressive effect of opioid agonists.

Antidepressants, similar to Carbamazepine, block the NMDA receptors and improve cell membrane function. Antidepressants are important in improving the eventual recovery, immune system function, and reduction of mortality and morbidity in chronic pain patients.

Antidepressants possess pure analgesic properties. Examples: Doxepin (Zonalon) topical cream is an excellent topical analgesic for neuropathic pain. The analgesic effect of tricyclics is reversed by Naloxone. The analgesic property makes the therapeutic use of antidepressants essential for treatment of neuropathic pain.

Antidepressants with properly balanced serotonin and norepinephrine (Nor Ep) reuptake inhibition provide maximal analgesia. Antidepressants, similar to Morphine pump, provide naloxone -reversible endorphin type pain relief . Such drugs as desipramine, imipramine and trazodone are superior to mainly serotonin inhibitors such as Mitrazepine (Remeron) and fluoxetine. Remeron is a good hypnotic, but in our patients it has shown no significant analgesic value. On the other hand, Venlafaxine (Effexor) is a weak inhibitor of serotonin and a strong inhibitor of nor ep reuptake-aggravating hypertension and sympathetic vasoconstriction by augmenting norepinephrine function. Venlafaxine has a high profile of adverse drug interaction with P450 and CYP2D6 Isoenzymes inhibitors (which comprise a long list of medications). It is best not to use this drug in CRPS. Buproprion (Wellbutrin) aggravates seizure disorder. Myoclonic jerks (see Movement Disorders) being a common complication of CRPS is aggravated by this drug. Its use is contraindicated in CRPS.

Certain antidepressants such as tricyclics and Trazodone, increase the synaptic serotonin and nor ep concentrations. This balanced phenomenon provides effective analgesia, natural sleep, and antidepressant effect. Trazodone provides analgesic effect in less than 24 hours in contrast with five to seven days for the same effective result with tricyclics. Trazodone does not cause weight gain when compared to amitriptyline(see below).

Of the tricyclics, Amitriptyline has been the most widely used analgesic, but it has strong anticholinergic and sedative side effects, and my cause paranoid and manic symptoms. More importantly, it has a tendency to cause weight gain. In our study of 824 CRPS patients, 612 had already been tried on Amitriptyline. In the first year, these patients gained an average of over 7kg, and, in the following year, an additional 3.6kg. Trial of Desipramine or Trazodone did not cause any significant weight gain. Weight gain in a CRPS patient who already has difficulty with ambulation is quite harmful. In addition, tricyclics have adverse cholinergic and muscarinic properties resulting in complications of orthostatic hypotension and ECG changes.

Anticonvulsant treatment is helpful in CRPS for two types of symptoms: 1. Spinal cord sensitization leading to myoclonic and akinetic attacks, and 2. In patients who suffer from ephaptic or neuroma type of nerve damage characterized by stabbing, electric shock, or jerking type of pain secondary to damage to the nerve fibres. In such cases, anticonvulsants, especially Tegretol (non-generic), Depakene, Gabapentin, and Klonopin (non-generic), are quite effective. The ephaptic, causalgic CRPS II is best managed with combination of an effective anticonvulsant, antidepressant, and analgesics.

Clonazepam is effective in control of myoclonic jerks. Decades of experience with Klonopin and Tegretol in neurology have taught the lesson that brand Klonopin and Tegretol are superior to their generic forms (Clonazepam and Carbamazepine) in controlling epileptic seizures. The American Academy of Neurology has recommended that generic antiepileptic drugs not be prescribed. Gabapentin (Neurontin) which is an adjunctive anticonvulsant, provides relief for burning type of neuropathic pain. Similar to Tegretol, Gabapentin is also neuroleptic. Carbamazepine, similar to Mexiletine, is an effective sodium channel blocker. It is far better tolerated than Mexiletine.

(NSAIDS)

The inflammatory complications of CRPS respond properly to NSAIDS. The beneficial effects of NSAIDS may be related to correcting the immune inflammatory damages in nerve death-be it neuropathic inflammation of CRPS, nerve death due to Alzheimer, or cerebrovascular disease (e.g., benefits from aspirin therapy). In Alzheimer, immune factors such as "membrane attack complex" play a role in nerve death-this may explain the benefits of NSAIDS. Cox inhibitors (e.g., Celebrex or Vioxx) are very helpful for pain relief and detoxification from opioid dependence.

The alpha-1 blockers Phenoxybenzamine (Dibenzyline) and Hytrin (Terazocin) are effective systemic nerve blocking agents. Forty soldiers suffering from CRPS type II were treated with phenoxybenzamine with excellent results, eliminating the need for sympathectomy. Clonidine in oral, intrathecal, or cutaneous patch forms, Clonidine is quite effective as an alpha-2 blocker. Application of Clonidine patch to the area of original damage in the extremity may aggravate the pain. It is effective when applied topically to paravertebral area in cervical or lumbar region corresponding to the referred pain of sensory nerve roots. After completion of sympathetic nerve block injection, application of Clonidine patch is a complementary treatment and may prevent the need for further invasive nerve block. Another effective alpha-2 blocker, Yohimbine, is not as potent as alpha-1 blockers.

There are two different forms of edema: 1. extracellular hypervolemia such as seen with congestive heart failure. This is a pitting edema due to increased plasma volume and 2. intracellular edema such as seen with glaucoma and pseudotumor cerebri. This form of edema causes intracellular cytoplasmic water retention leading to cerebral edema and indurated edema associated with neurovascular instability (fluctuating rusty, reddish, or pale discoloration). Normally, the perineurium is impenetrable to water. In inflamed tissue the peripheral nerve terminals increase by ("sprouting"). As a result, edema sensitizes the tissue to opioid peptides and to pain. Standard diuretics such as HCTZ (Hydrochlorothiazide) or Furosemide dehydrate and reduce the volume of the extracellular space. They are most effective in cardiovascular diseases. The osmotic diuretics such as Mannitol, chlordiazepoxide or magnesium salts reduce the intracellular volume and reduce neurogenic edema. The edema of CRPS due to normal cell membrane dysfunction leads to rise in intracellular Na+ and Ca++ . Correction of sodium potassium pump with the help of NMDA inhibitors such as mexiletine, Carbamazepine, and MK801also help reduce edema.

Magnesium sulfate (Epsom salt), in oral, IV, enema, or bathing form, effectively reduces the edema. It acts similar to calcium channel blockers which are also effective in neuropathic pain and headache . For the complication of Neurogenic Bladder and Interstitial Cystitis, Nifedipine may be helpful.

Osteopenia, usually transient, is a common complication of CRPS. Most commonly, it affects the hip, foot, shoulder, and wrist areas. Treatment consists of combination of weight bearing, mobilization, estrogen replacement, biphosphonates, and diet rich in calcium (e.g., cabbage, and dairy products). Mobilization is the most indispensable form of treatment.

"Ovarian steroids produce measurable cognitive effects after ovariectomy and during aging" (McEwen, et,al). Hormone replacement improves the cerebral cognitive functions. Estrogen plays a major role in formation of new excitatory synapses and NMDA regulation both in male and in female formative brains. Realizing that female CRPS patients, regardless of age, have a tendency for menopausal symptoms (hot flashes and excessive sweating), serum estrogen levels were measured in 60 of these patients in this study. The serum estrogen(mid-cycle estradiol) was in the 87 to 195 PG/ml range as compared to the normal 100 to 395 PG/ml range. Estrogen replacement therapy improved the cognitive function and reduced the tendency for hyperhydrosis on these patients.

In 43 patients who underwent infusion pump therapy for CRPS , a more significant drop in serum estrogen and testosterone levels were noted. Forty one patients required hormone replacement therapy which improved pain reduction by an average of 1.7 (on a basis of 0-10), and reduced or cleared up the edema of lower extremities.

Nerve blocks may be diagnostic, therapeutic, or both. The two types of blocks are not identical and interchangeable. The diagnostic nerve blocks with simple local anesthetic injection last a few hours to a few days. Unfortunately diagnostic ganglion nerve blocks are commonly mistaken to be a form of therapy. The meta - analysis studies by Kozin , Carr, et al, and Schott, have emphasized such blocks are indistinguishable from placebo. Simple pain relief from blocks (SMP) does not prove CRPS.

There are three main forms of diagnostic blocks:

1. Sympathetic ganglion block.

2. Local anesthetic nerve block.

3. Compression regional block.

Diagnostic nerve blocks are hampered by a significant incidence of false-positive and false-negative results, even in the best hands. The ganglion nerve blocks with local anesthetics are mainly diagnostic. The local anesthetic effect doesn’t last more than two hours to maximum 1- 14 days. Ganglion nerve block should be complimented with therapeutic nerve blocks such as brachial plexus, regional, and epidural nerve blocks. Where as ganglion nerve blocks temporally improve the circulation and relieve the pain, they do not improve flexor spasm and deformity of the hands and feet. The brachial plexus and regional blocks are more beneficial in correcting such movement disorders.

The relief from epidural, paravertebral, regional and brachial plexus blocks with combination of Bupivacaine and Prednisolone lasts about 8-12 weeks. It is effective for somatic radiculopathy and for neuropathic pain . Repeated stellate ganglion blocks can permanently damage the sympathetic nerve cells, and result in "Virtual Sympathectomy." In addition, such repetitive trauma may be complicated by migraine headaches.

According to Cheema, paravertebral nerve block provides effective pain relief for both SMP and SIP. This is in contrast with articular facet (zygoapophyseal) blocks (ZAB) which are fraught with painful joint injuries.

The paravertebral nerve blocks are technically similar to zygoapophyseal blocks (ZAB) but should not be mistaken for each other. The ZAB invades the zygoapophyseal (ZA) joint. Insertion of needle, or radiofrequency treatment of ZA joint is traumatic to the joint, and has the potential of adding new pathology with additional source of pain. Bogduk’s team, et al, have reported only 40% pain relief in patients undergoing such ZA joint neurotomy. The paravertebral nerve block does not invade any joint structure, and should not be mistaken for ZA injection or neurotomy.

In our series of 824 CRPS patients, 22 had undergone surgical sympathectomy with temporary partial relief of 6 days to 38 weeks in 9 patients: up to 54 weeks in 10 patients: and no relief in 3. Chemical sympathectomy was done (prior to referral to our medical center) on 13 patients with temporary relief of 3 days to 29 weeks in 4 patients, no relief in 5 , and rapid deterioration of CRPS in 4 patients. Surgical, radiofrequency and chemical or (neurolytic), sympathectomies, have been applied in treatment of CRPS since 1916. Sympathectomy may provide temporary pain relief, but after a few weeks to months it loses its effect. The success has been limited to the series that have had short-term patient follow-ups of a few months after surgery. Sympathectomy and application of neurolytic agents should be limited to patients with life expectancies measured in weeks or months - e.g., cancer and end stage advanced occlusive vascular disease patients. On the other hand, CRPS patients usually have 3 to 5 decades of life expectancy ahead of them. They should not be exposed to aggravation of pain due to sympathectomy. The sympathectomized patients developed post operative spread of CRPS in 12 of our 35 patients (37%). This high incidence of spread is in contrast to the 18% incidence in the rest of 824 cases.

1. Laser Doppler-vascular studies have revealed the temporary benefits of vasoconstrictor reflexes lasting no more than four weeks. The neurovascular instability in late stage RSD is not expected to respond to sympathectomy. Sympathectomy is aimed at achieving vasodilation. The neurovascular instability refers to vacillation and instability of vasoconstrictive function.

2. At no time the sympathectomy can be complete - unless it is done at the time of autopsy with complete removal of the chains of sympathetic ganglia.

3. Due to extensive interconnection of chains of sympathetic ganglia, removal of a short chain of sympathetic ganglia is easily compensated by rerouting of the sympathetic impulses through the horizontal and vertical connections of sympathetic nerve fibres in paraspinal chain of sympathetic ganglia as well as through the midline connection of sympathetic ganglia via the midline sympathetic plexi such as cardiac plexus, superior mesenteric plexus, etc.

4. The wide dynamic range of spread of pain impulse in adjacent levels of spinal cord causes the spread of pain to adjacent levels above and below the sympathectomized area.

5. The alpha-1 adrenoreceptor in the sympathectomized area are hypersensitive to the smallest concentrations of circulatory or adjacent tissue noradrenaline.

6. By the time the sympathectomy is undertaken, the disease is usually too advanced and in late stages. The pain is mainly SIP and is expected to be non-responsive to ablation of the sympathetic ganglia. Sympathectomy aggravates the already existing sympathetic dysfunction.

7. Chemical and radiofrequency sympathectomy cause chemical damage and scarring of the adjacent tissues. This is especially true in the case of alcohol or phenol chemical sympathectomy. The scar of such chemicals becomes a new source of neuropathic pain. Chemical or radiofrequency ablation surgical procedures are justifiable as an act of mercy for advanced cancer patients, but CRPS patients usually have a few decades of life expectancy and cannot be expected to live with the pain due to the scar of such destructive procedures for several years.

8. The beneficial effects of sympathectomy are reported in surgical series of patients followed for a few months - as short as one to four months. The follow up of up to 5 years reveals a high incidence of recurrence of symptoms and signs after sympathectomy.

Elective surgery for presumptive conditions such as carpal tunnel, tarsal tunnel, and thoracic outlet syndrome (TOS)- in spite of normal nerve conduction studies - only adds a new source of neuropathic pain at the surgical scar. According to Cherington, et al , there is a tendency for unnecessary TOS surgery, elective surgery is the strongest predictor (P<0.001) of poor treatment outcome (Please see " Treatment Outcome").

According to Rowbotham, "amputation is not to be recommend as pain therapy." All 11 patients in our series who underwent amputation showed marked deterioration post-op. The surgical stump was the source of multiple neuromas with sever CRPS II type of intractable pain. Amputation should be avoided by all means due to its side effects of aggravation of pain and tendency for spread of CRPS.

 SURGERY AND IMMUNE SYSTEM

Surgical procedures in neuropathic pain patients, in general, are sources of stress and produce characteristic neuro-endocrine and metabolic responses, local inflammation, and can cause disturbance of immune system function.

The body responds in opposite direction to surgery for somatic versus neuropathic pain. An acute appendicitis or cholecystitis responds quite nicely to surgery. On the other hand, surgery in the area of the extremity involved with neuropathic pain has the potential of aggravating the condition. Tissue damage from the surgical procedures results in the local release of inflammatory neurokines. This biochemical and cellular chain of events leads to up-regulation of the immune system and nervous system activation by releasing Substance P, histamine, serotonin, CGRP, bradykinin, prostaglandins, and other agents. This leads to a local vasodilation response in the area of the surgical scar, increased capillary permeability, and sensitization of the peripheral afferent nerve fibers resulting in allodynia and hyperpathia. Surgery can cause suppression of immune function aggravating the manifestations of neuropathic pain. Post-operatively, there is a tendency for dysfunction of the lymphocytic role in immune regulation. This is manifested by a decrease in number of T-cell lymphocytes and the function of the T-cell lymphocytes. The disturbance and suppression of the immune system due to surgery enhances the malignant tumor growth and metastasis . Surgery "results in a perturbation of nervous, endocrine and immune system as well as their interregulatory mechanisms leading to compromised immunity." This disturbance of immunity may manifest itself in skin ulcerations noted in 2 of 11 amputees referred to our clinic during 1990-1995 period. A similar case of amputee with skin ulcers has been recently reported .

There are times that surgery is unavoidable. Examples: tear of ligament or cartilage in the knee joint that would preclude weight bearing. In such patients, epidural nerve block with a combination of Bupivacaine and 20 to 30 mg Prednisolone before, during, and after surgery (with the help of epidural catheter) helps reduce the side effects of surgical trauma. Another example is extensor deformity of a finger causing useless hand which in turn aggravates CRPS.

Four potential variants influencing the treatment outcome were studied:

1. Modes of therapy.

2. The nature of pathology.

3. Patient’s age.

4. Delay in diagnosis and treatment.

1. The type of treatment was the critical predictor of outcome. For example, in patients younger than 21 years, the surgical treatment reversed the beneficial prognostic value of youth (P<0.001). Moreover, application of ice over 2 months (P<0.001), application of ice less than 2 months (P<0.001), amputation (P=0.025), and sympathectomy, were the strong predictors of poor prognosis.

2. The nature of pathology was an accurate predictor of outcome. Examples: Chemical Sympathectomy with neurolytic agents (alcohol, phenol, or other agents) was done in 13 patients. Chemical sympathectomy was the third group of poor prognosis after venipuncture and amputation regardless of other risk factors. One probable reason explaining such a poor prognosis may be the fact that the lytic agent may infiltrate beyond the target area of injection.

The venipuncture CRPS group of 17 patients showed the worst prognosis and the most rapidly deteriorating course regardless of age at the onset, any delay in diagnosis or mode of therapy. The venipuncture CRPS is the purest form of selective sensory nerve injury in the wall of skin and subcutaneous microvasculature. The poor prognosis may be due to lack of simultaneous somatic sensory nerve stimulation which would over-shadow the neuropathic microvascular sensory nerve irritation.

3. In regard to age at onset, prognosis was good among the 138 patients with onset at 2-22 years. The exception to this rule was 32 patients in this group who had undergone surgical procedures with poor results, and 5 other patients ages 3-21 years having been accused of being Münchausen syndrome, followed by years of no treatment.

4. Delay in diagnosis: as Poplawski, et al have emphasized, early diagnosis is important in the management of CRPS. However, if the disease is not diagnosed early, it is of no value if the patient is treated with stressors such as ice, surgery, or cast application. Additionally, if the original pathology is severe and irreversible in nature, early diagnosis is of little value.

The group with the best prognosis was typified by CRPS patients with mainly cold extremity treated with no surgery or ice. The patients with permanently warm extremity, due to sympathetic nerve damage, fared poorly.

CONCLUSION

CRPS/RSD is a complex chronic pain syndrome with four main features of hyperpathic/allodynic pain, vasomotor dysfunction and flexor spasms, inflammation, and limbic system dysfunction. Elective surgery, and amputation are at the top of the list of aggravating factors. CRPS is usually caused by a minor injury, and requires proper evaluation and multi-disciplinary treatment addressing the multifaceted pathological processes that evolve during its chronic course. Patient’s age, the nature of pathology, and mode of therapy influence the outcome of treatment. If at all possible, surgery, ice and cast applications should be avoided. There is a desperate need for research in proper management of CRPS.

Acknowledgment: We are grateful to Mr. Eric Phillips for his enormous contributions in organization of patient materials and references in this review.