Annals of the Rheumatic Diseases 1992; 51: 1188-1191

Role of the sympathetic nervous system in chronicjoint pain and inflammation

B L Kidd, S Cruwys, P I Mapp, D R Blake

A substantial body of evidence gathered from avariety of diseases points to a critical interactionbetween the sympathetic nervous system andthe inflammatory cascade. An importantsympathetic component to rheumatic disease isindicated by the observation that symptomsin both reflex sympathetic dystrophy andrheumatoid arthritis may be alleviated afterregional sympathectomy. 1

Clearly, attempts to explain pathophysio-logical mechanisms must take observations suchas these into account. This review summarisesthe evidence for a sympathetic influence onrheumatic disease and speculates on themechanism by which this may arise.

The Inflammation Group,ARC Building,London HospitalMedical College,LondonEl lAD,United KingdomB L KiddS CruwysPI MappD R BlakeCorrespondence to:Dr Kidd.

Clinical observationsIn 1864 Mitchell, Morehouse, and Keenreported the case of a soldier in the AmericanCivil War who sustained a gunshot wound tothe left upper arm. The second day after theinjury the soldier developed excruciating pain inthe arm such that light touch or even heavysteps in the room caused unbearable dis-comfort.4 This unfortunate state continued formonths after the original wound had healedand was termed 'causalgia' by Mitchell andcolleagues.4

Since then a number of other terms have beenused to describe disorders characterised bypersistent pain, vasomotor disturbance,and trophic change. These include Sudeck'satrophy, shoulder-hand syndrome, algodys-trophy, and reflex sympathetic dystrophy.5Pathological mechanisms have yet to be clearlyestablished, but three important clinical obser-vations suggest that the sympathetic nervous

system plays a part.6 As observed by Mitchell,pain in these disorders is exacerbated by stimulithat evoke sympathetic discharge, such as noiseand emotional arousal. Secondly, the disordersare associated with vasomotor and sweatingdisturbances, suggestive of abnormal sym-pathetic activity. Finally, patients with thesedisorders often have a dramatic and prolongedresponse to guanethidine, reserpine, and otheragents which alter sympathetic activity. '

Sympathectomy has also been used to treatrheumatoid arthritis.2 3 7 In 1927 Rowntree andAdson described a series of young patients withrheumatoid arthritis whose symptoms improvedafter sympathetic ganglionectomy.7 Flothow in1930, Leriche and Jung in 1933, and Young in1936 all reported relief of articular pain in smallseries of patients with rheumatoid arthritis.7A later study by Hertford ofpatients with intract-able hip or knee pain reported that six of seven

patients received substantial symptomatic reliefafter lumbar sympathectomy. Several of thepatients had reduction of joint swelling andimprovement of cutaneous ulcers in the sym-pathectomised limb. The relief of symptomswas observed for follow up periods lasting fromfour to 24 months.7More recently, Levine and colleagues inves-

tigated the effects of regional sympatheticblockade using guanethidine in 24 patients withactive rheumatoid arthritis.3 In a randomised,double blind study guanethidine was found todecrease pain and increase pinch strength overthe two week duration of the study. Nosignificant improvement was noted in gripstrength, joint tenderness, or duration ofmorming stiffness.

Reflex sympathetic dystrophyPersistent pain, often described as havinga burning quality, is the hallmark of thisdisorder.5 8 Vasomotor disturbances arecommon and include vasodilatation with warmerythematous skin, vasoconstriction with coolpallid skin, occasional Raynaud's phenomenon,and excessive sweating.8 Months or years afteronset of symptoms subcutaneous atrophyand irreversible flexion contractures maydevelop with a resultant pale, cold, and painfulextremity.5

Patchy osteoporosis may be present early inthe course of the disease, progressing to morediffuse change later.8 Kozin and colleaguesreported a high incidence of bilateral involve-ment with the articular areas being the mostaffected.9 Synovial biopsy specimens showvarying degrees of oedema, increased vascu-larity, and proliferation of synovial lining cells,an appearance similar to that found in frozenshoulder.9The pathogenesis of reflex sympathetic

dystrophy remains unclear, but, multiplefactors are probably involved both centrally andperipherally. In 1943 Livingstone proposed thatchronic stimulation of afferent fibres, whichrespond to noxious stimuli (nociceptors),produced excessive excitation in spinal inter-nuncial neurones causing abnormal sympatheticefferent activity.'0 Subsequently, Robertsspeculated that maintenance of the disorderdepends not so much on persistent nociceptorirritation but rather on abnormal central pro-cessing of apparently normal mechanoceptorinput into the spinal cord. "A number of investigators have focused on

peripheral mechanisms.'2-'5 It is known thationtophoretic application of noradrenaline is an

1188

on Novem

ber 18, 2021 by guest. Protected by copyright.

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.51.11.1188 on 1 N

ovember 1992. D

ownloaded from

Role ofthe sympathetic nervous system in chronic joint pain and inflammation

effective pain stimulus in patients with reflexsympathetic dystrophy'3 and that sympatheticblockade may result in symptomatic reliefwhether or not the block is performed centrallyor peripherally.'4 In contrast, catecholaminesdo not cause pain in normal subjects andsympathetic blockade does not impair painsensation.'2 These observations support theproposal by Janig that a vicious circle ofexcitation of primary afferent fibres by post-ganglionic sympathetic fibres maintains thepathophysiological process in this disorder. '5The best experimental evidence for peripheral

interactions between afferent and sympatheticfibres comes from observations made ofneuronal activity in stump neuromas.'6 In thisexperimental model transection and ligation of amajor nerve are followed by development of astump neuroma. Subsequently, damagedafferent fibres sprout within the neuroma and,unlike undamaged afferent fibres, respond tosympathetic stimulation and the application ofcatecholamines.'7 More recent models whichretain some nerve contact with the target tissueusing subtotal resection or ligature constrictionhave shown a similar result.'8

Present evidence supports a role for thesympathetic nervous system in the developmentand maintenance of reflex sympathetic dys-trophy. Preceding nerve or tissue trauma canbe demonstrated in over half the patientspresenting with this disorder.8 It seems likelythat in this situation continuing stimulation ofdamaged afferent fibres by sympathetic activityis at least partially responsible for the chronicpain state observed clinically.'2

Experimental arthritisOn a more general level a number of experi-mental studies have examined the contributionof neural components to the expression andoutcome of arthritis. The earliest, by Courtrightand Kuzell in 1965, suggested that priorsection of a sciatic nerve delayed the onset anddiminished the extent of rat adjuvant arthritis inthe operated limb when compared with theopposite side.'9 These results contrast markedlywith a later study by Rees et al, which reportedthat sciatic nerve section substantially exacer-bated disease as judged on clinical and radio-logical grounds.20There is accord, however, between studies by

Rees et al and Levine et al, which both show anattenuation of adjuvant disease after sympathec-tomy or selective destruction of small diametersensory nerves.20 21 These studies also show asignificant exacerbation of disease followingposterior nerve root section. Both Rees et al andLevine et al stress the importance of an intactsympathetic nerve supply on the severity ofarthritis and, interestingly, Levine has shownthat the enhanced disease seen in the rhizo-tomised limbs may be markedly reduced byprior sympathectomy. The importance ofsympathetic activity is further shown by theobservation that in spontaneously hypertensiverats, which have increased sympathetic tone,adjuvant arthritis is more severe than in normo-tensive control animals.2'

A series of studies by Levine and colleaguesexamined the relation between sympatheticpostganglionic nerve terminals and severity ofexperimental arthritis.22-24 Their early studiesindicated that blockade of the B2 receptor onthe sympathetic postganglionic nerve terminalsignificantly attenuated adjuvant arthritis andthat the effect was independent of the timingof inhibition relative to the onset of clinicaldisease.22 A similar attenuation was seen afteradrenal medullectomy, reversible by adminis-tration of adrenaline but only in the presence ofan intact sympathetic nervous system, suggest-ing mediation via the sympathetic postgang-lionic nerve terminal.23 From these studiesLevine suggested that catecholamines acting onsympathetic postganglionic nerve terminalsinduce the release of unknown compoundsresponsible for the deleterious effects of sym-pathetic activity on adjuvant disease.24The clear influence of the sympathetic

nervous system on experimental arthritis,coupled with the failure of adrenergic mediatorsto influence the outcome directly, suggests thatnon-adrenergic mechanisms may be involved.Postganglionic sympathetic nerves can bedivided into two broad groups: those inner-vating sweat glands and skeletal muscle bloodvessels containing acetylcholine, and the vastmajority of fibres containing noradrenaline. Asubgroup of the latter fibres also contains otherneurotransmitters, including neuropeptide Y,adenosine triphosphate, and nitric oxide.25The role of these agents in both normal andabnormal situations is currently the subject ofintensive speculation and investigation.

Neuropeptide Y is a 36 amino acid peptidewith a number of diverse physiological effects,including vasoconstriction and potentiation ofthe actions of other neurotransmitters such asnoradrenaline.26 It also acts presynaptically toinhibit its own release and that of noradrenalineand 5-hydroxytryptamine. In high doses neuro-peptide Y induces histamine release from mastcells, and in vitro studies suggest that it has arole in regulating osteoblast function.27 Agoniststudies indicate that there are at least threesubtypes of neuropeptide Y receptor and,recently, a number of receptor antagonists havebeen described.28

Edvinsson and colleagues have recently intro-duced an inositol derivative, D-myo-inositol-1 ,2,6-trisphosphate (PP56), as a functionalantagonist of vasoconstriction induced byneuropeptide y.29 Significantly, PP56 has beenshown to exert anti-inflammatory effects inseveral experimental models of arthritis,including reduction of oedema in both acutecarrageenin induced inflammation and chronicadjuvant arthritis.30

Rheumatoid arthritisSympathetic overactivity in reflex sympatheticdystrophy is associated with pain and juxta-articular disease, including synovitis, and thereappear to be important sympathetic influenceson the outcome of experimental adjuvant arth-ritis. How then might these observationsreflect pathogenic mechanisms in rheumatoidarthritis?

1189

on Novem

ber 18, 2021 by guest. Protected by copyright.

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.51.11.1188 on 1 N

ovember 1992. D

ownloaded from

Kidd, Cruwys, Mapp, Blake

Clearly, normal human synovium is richlyinnervated with both sympathetic and afferentnerve fibres.3' Immunocytochemical studiesusing antisera against protein gene product 9.5(PGP 9.5) as a marker for the overall innerva-tion have shown fibres distributed around bloodvessels and lying freely within synovium.Nearly all these nerves are immunoreactive forneuropeptides with neuropeptide Y immuno-reactive fibres being exclusively located aroundblood vessels, whereas substance P and calci-tonin gene related peptide immunoreactivefibres are located in both free and perivascularareas.3' The exact function of these peptidesremains uncertain, but in addition to vascularregulation they probably play a part in immuno-regulation and bone metabolism.32

In rheumatoid arthritis the more superficialsynovial tissues, including blood vessels and theintimal cell layer, are devoid of fibres immuno-reactive for PGP 9.5 or any of the individualneuropeptides.33 In deeper tissues the inner-vation is similar to normal tissues, but theimmunostaining is weaker and fibres take on abeaded appearance with small lengths appearingunstained. A similar loss of immunoreactivefibres has also been seen in experimentalmodels, including adjuvant arthritis.34The absence of immunoreactive fibres in the

superficial layers of rheumatoid synoviummight suggest that there is an increased releaseof substance P, calcitonin gene related peptide,and neuropeptide Y, which thereby reduces theneuronal stores to levels below that detectableby immunocytochemistry. As the immuno-staining for the neuronal marker PGP 9.5 isalso lost, however, it seems more probablethat nerve fibres in the superficial layers are

destroyed.33The loss of synovial nerves in rheumatoid

arthritis is potentially caused by products of theinflammatory cascade. In studies using hypo-xanthine/xanthine oxidase as a means of genera-

ting reactive oxygen species it is apparent thatneuropeptides are vulnerable to damage.35 Ourown unpublishedstudies have shown rapid lossof synovial nerve fibres in a glucose oxidasemodel of inflammation which is known todepend on production of hydrogen peroxide.Exercised rheumatoid joints are subjected to'intermittent' episodes of hypoxia, thus creatingthe conditions necessary for hypoxic-reperfusionevents and the subsequent production ofsimilarly destructive reactive oxygen species.36We speculate that a direct consequence of

nerve damage in diseased joints is that synovialafferent fibres become sensitised to continuingsympathetic activity. This is supported by theobservations made of the stump neuroma wheredamaged afferents respond to sympatheticstimulation and application of catecholamines.Within the heavily innervated human jointpersistent afferent stimulation would inevitablylead to chronic pain.The situation would be

compounded by the hypoxic nature of inflamedjoints37 as hypoxia induces both enhancedresponsiveness to sympathetic stimulationand reversible increases in the number ofreceptors on sympathetic postganglionic nerve

terminals.38

Persistent sympathetic stimulation of afferentnerves in rheumatoid synovium implies thatactivation of afferents responding to noxiousstimuli would no longer depend critically onprostaglandins and other hyperalgesic productsof inflammation. This is borne out clinically asthere is poor correlation between pain andclinical and laboratory indices of inflamma-tion,39 and basic clinical observation indicatesthat conventional anti-inflammatory drugs donot provide satisfactory analgesia in all patients.Sympathetic stimulation of nociceptors wouldalso explain the beneficial effects of sympathec-tomy described earlier.

It is apparent that sympathetic mechanismsmay have an important role in joint inflamma-tion. Existing treatment with adrenoceptorantagonists only partially inhibits sympathetictransmission within the joint, leaving non-adrenergic function intact. The increasing avail-ability of adrenergic and non-adrenergicinhibitors therefore offers intriguing potentialfor further exploration and eventual therapeuticuse.

The authors would like to acknowledge the financial support ofthe Arthritis and Rheumatism Council.

1 Hannington-KiffJ G. Relief of Sudeck's atrophy by regionalintravenous guanethidine. Lancet 1977; i: 1132-3.

2 Hannington-KiffJ G. Rheumatoid arthritis-interventionaltreatment with regionally applied drugs and the use ofsympathetic modulation: discussion paper. J7 R Soc Med1990: 83: 373-6.

3 Levine J D, Fye K, Heller P, Basbaum Al, Whiting-O'KeefeQ. Clinical response to regional intravenous guanethidine inpatients with rheumatoid arthritis. J Rheumatol 1986; 13:1040-3.

4 Mitchell S W, Morehouse G R, Keen W W. Gunshot woundsand other injuries of nerves, Philadelphia: Lippincott, 1864.

5 Swartzman R J, McLellan T L. Reflex sympathetic dys-trophy. A review. Arch Neurol 1987; 44: 555-61.

6Bennet G J. The role of the sympathetic nervous system inpainful peripheral neuropathy. Pain 1191; 45: 221-3.

7 Hertford R A. Extended sympathectomy in the treatment ofchronic rheumatoid arthritis. J Am Geriat Soc 1957; 5:904-15.

8 Chard M D. Diagnosis and management of algodystrophy.Ann Rheum Dis 1991; 50: 727-30.

9 Kozin F, McCarty D J, Sims J, Genant H. The reflexsympathetic dystrophy syndrome. 1. Clinical and histologicstudies: evidence for bilaterality, response to corticosteroidsand articular involvement. Am J Med 1976; 60: 321-31.

10 Livingstone W K. Pain mechanisms: a physiologic interpretationof causalgia and its related states. New York: Macmillan,1943.

11 Roberts W. A hypothesis on the physiological basis forcausalgia and related pains. Pain 1986; 24: 297-311.

12 Koltzenburg M, McMahon S B. The enigmatic role of thesympathetic nervous system in chronic pain. TrendsPharmacol Sci 1991;12: 399 402.

13 Wallin G, Torebiork E, Hallin R. Preliminary observationson the pathophysiology of hyperalgesia in the causalgic painsyndrome. In: Zotterman Y, ed. Sensory functions of the skinin primates: with special reference to man. Vol 27. (Wenner-Gren Center International Symposium Series.) Oxford:Pergamon Press, 1976: 489-502.

14 BonelliS, Conoscente F, Movilia P G, et al. Regionalintravenous guanethidine vs stellate ganglion block in reflexsympathetic dystrophies: a randomised trial. Pain 1983; 16:297-307.

15Janig W. The sympathetic nervous system in pain: physio-logy and pathophysiology. In: Stanton-Hicks M, ed. Painand thesympathetic nervoussvstem. Boston and London:Kluwer Academic, 17-89.

16 Wall P D, Gutnick M. Ongoing activity in peripheral nerves:the physiology and pharmacology of impulses originatingfrom a neuroma. Exp Neurol 1974; 43: 580-93.

17 Devor M. Nerve pathophysiology and mechanisms of pain incausalgia. J Auton Nerv Syst 1983; 7: 371-84.

18 Seltzer Z, Dubner R, Shir Y. A novel behavioural model ofneuropathic pain disorders produced in rats by partialsciatic nerve section. Pain 1990; 43: 205-18.

19 Courtright L J, Kuzell W C. Sparing effects of neurologicaldeficit and trauma on the course of adjuvant arthritis in rats.Ann Rheum Dis1965; 24: 360-7.

20 Rees R G, DixeyJ J, Lightman S L, et al. Selective neuralmanipulation can profoundly alter the course of adjuvantarthritis in rats. Brj Rheumatol 1989; 28 (suppl 2): 40.

21 Levine J D, Dardick S J, Roizen M F, Helms C. Contri-bution of sensory afferents and sympathetic efferents to

1190

on Novem

ber 18, 2021 by guest. Protected by copyright.

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.51.11.1188 on 1 N

ovember 1992. D

ownloaded from

Role ofthe sympathetic nervous system in chronic joint pain and inflammation

joint injury in experimental arthritis. J Neurosci 1986; 6:3423-9.

22 Levine J D, Coderre T J, Helms C, et al. P2-Adrenergicmechanisms in experimental arthritis. Proc Natl Acad SciUSA 1988; 85: 4553-6.

23 Coderre T J, Basbaum Al, Dallman M F, et al. Epinephrineexacerbates arthritis by an action at presynaptic B2-adreno-ceptors. Neuroscience 1990; 34: 521-3.

24 Coderre T J, Chan A K, Helms C, et al. Increasing sym-pathetic nerve terminal-dependent plasma extravasationcorrelates with decreased arthritic joint injury in rats.Neuroscience 1991; 40: 185-9.

25 Bernstock G. The changing face of autonomic neurotrans-mission. Acta Physiol Scand 1986; 126: 667-91.

26 Lundberg J M, Franco-Cereceda A, Hemsen A, Lacroix J S,Pernow J. Pharmacology of noradrenaline and neuropeptidetyrosine (NPY)-mediated sympathetic cotransmission.Fundam Clin Pharmacol 1990; 4: 373-91.

27 Larson J, Ekblom A, Henriksson K, et al. Concentration ofsubstance P, neurokinin A, calcitonin gene-related peptide,neuropeptide Y and vasoactive intestinal polypeptide insynovial fluid from knee joints in patients suffering fromrheumatoid arthritis. Scandj Rheumatol 1991; 20: 326-35.

28 Michel M C. Receptors for neuropeptide Y: multiplesubtypes and multiple second messengers. Trends PharmacolSci 1991; 12: 389-94.

29 Edvinsson L, Adamsson M, Jansen I. Neuropeptide Yantagonistic properties of d-myo-inositol-trisphosphate inguinea pig basilar arteries. Neuropeptides 1990; 17: 99-105.

30 Claxon A, Morris C, Blake D, et al. The anti-inflammatorveffects of D-myo-inositol-1.2.6-trisphosphate (PP56) onanimal models of inflammation. Agents Actions 1990; 29:68-70.

31 Kidd B L, Mapp P I, Blake D R, Gibson S J, Polak J M.Neurogenic influences in arthritis. Ann Rheum Dis 1990; 49:649-52.

32 Holzer P. Local effector functions of capsaicin-sensitive nerveendings: involvement of tachykinins, calcitonin gene-relatedpeptide and other neuropeptides. Neuroscience 1988; 24:739-68.

33 Mapp P I, Kidd B L, Gibson S J, et al. Substance P-,calcitonin-related peptide- and C-flanking peptide of neuro-peptide Y-immunoreactive fibres are present in normalsynovium but depleted in patients with rheumatoid arthritis.Neuroscience 1990; 37: 143-53.

34 Rees R G, Konttinen Y T, Hukkanen M, et al. The synovialinnervation of normal and arthritic rats. Br J Rheumatol1989; 28 (suppl 2): 41.

35 Hukkanen M, Konttinen Y T, Sergerberg M, et al. Neuro-peptides and inflammatory cells in synovium; possibleinvolvement of free radicals in neuropeptide degradation. BrJ Rheumatol 1991; 30 (suppl 1): 5.

36 Blake D, Merry P, Unsworth J, et al. 'lhe inflamed humanjoint is an example of hypoxic-reperfusion injury. Lancet1989; i: 289-93.

37 Stevens C R, Williams R B, Farrell A J, Blake D R. Hypoxiaand inflammatory synovitis: observations and speculation.Ann Rheum Dis 1991; 50: 124-32.

38 Heathers G P, Evers A S, Corr P B. Enhanced inositoltrisphosphate response to al-adrenergic stimulation incardiac myosites exposed to hypoxia. J Clin Invest 1989; 83:1409-13.

39 Rhind V M, Bird H A, Wright V. A comparison of clinicalassessments of disease activity in rheumatoid arthritis. AnnRheum Dis 1980; 39: 135-7.

In 1928 Egypt honoured Imhotep (he whocomes in peace). Imhotep was not only a grandvizier, architect, and personal physician of thePharoahs and chief medical officer of theKingdom, he was also deified by the Egyptians.He wrote that exercises are necessary 'to makethe joints limber'. He also stated that workingconditions of the slaves and peasants at thepyramids developed 'hardening of the limbs'.

BLYALE NDELMA

1191

YALE ANDELMAN

on Novem

ber 18, 2021 by guest. Protected by copyright.

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.51.11.1188 on 1 N

ovember 1992. D

ownloaded from