rheumatoid arthritis and surgery · joint damage and fixed deformity appears solomon has...

Annals ofthe Rheunatic Diseases 1990; 49: 837-844

Rheumatoid arthritis and ankle surgery

J Kirkup

Twenty years' experience in a foot clinic com-

bining the interests of a rheumatologist withthose of an orthopaedic surgeon indicates thatthe foot disabled by rheumatoid arthritis com-monly presents with (a) multiple joint damageat different stages of evolution and also ofasymmetrical distribution between the two feet;(b) a commonly progressive and yet often erraticpathology; and (c) management problems posedby proximal joint disease, especially at the kneeand in the hip.

Thus rheumatoid disease can affect many andoccasionally all of some 25 closely interrelatedjoints in the foot, including the ankle joint,which is rarely attacked in isolation and isusually influenced by the other joints. Forexample, damaged subtaloid and midtarsaljoints may cause a valgus or, less frequently, avarus attitude of the foot, leading to a valgus orvarus strain, or, indeed, a subluxation at theankle joint. Thus intertarsal correction takesprecedence over ankle surgery. Similarly,severe destruction of the metatarsophalangealjoints combined with skin breakdown andinfection in the sole demands their prior solu-tion. Knee valgus induces a valgus strain on thefoot, and if the knee is painful and unstable itmay need surgical attention before the foot,which can often be stabilised with orthoses. Afixed flexion attitude of the hip and particularlyof the knee induces a dorsiflexed attitude of thefoot under load, and thus arthrodesis of theankle in the classical position of moderateequinus may prove a disservice to the patient.

Royal National Hospitalfor Rheumatic Diseases,Upper Borough Walls,Bath BAI 1RLJ Kirkup

Statistical backgroundIn a survey of 200 consecutive patients withdefinite rheumatoid arthritis' 104 were noted tohave painful deformity of one or both feet. Fourof these feet had been subjected to previoussurgery, leaving 204 feet for analysis. Radio-logical changes were noted in 176 forefeet(metatarsophalangeal and toe joints) and 133hindfeet involving 124 midtarsal, 64 subtaloid,and 52 ankle joints.A further survey of 150 consecutive patients

attending the foot clinic of the Royal NationalHospital for Rheumatic Diseases, Bath, withhindfoot pain or deformity, and usually otherfoot complaints, also confirmed that the anklejoint was relatively resistant to attack. All thesepatients had attended the clinic for at least oneyear and had had rheumatoid arthritis for fromtwo to 47 years (average 15). A radiographicsurvey of the 300 limbs (table 1) showed jointdisease or previous surgery in 276 (92%) of theforefeet, 263 (88%) of the intertarsal joints, 155

(52%) of the ankles, 153 (51%) of the knees, and68 (23%) of the hips. Of the damaged joints ineach anatomical category, the operative rateswere as follows: 89/276 (32%) of the forefeet,28/263 (11%) of the intertarsal joints, 22/155(14%) of the ankles, 64/153 (42%) of the knees,and 39/68 (57%) of the hips.Of the 300 feet, some two thirds presented a

valgus attitude, principally owing to intertarsalsubluxation but also to tilting of the talus in theankle mortise, or both. Varus and equinus wasuncommon, the latter always being associatedwith ankle joint disease (table 2).

Although the hip was often spared disease,the chances of surgical intervention relative todamaged joints was high, occurring in 39/68(57%) cases (table 1). In contrast with the hip,although pathological changes were oftenpresent in the feet, operative surgical interven-tion was less common and, in the case of thehindfeet the lowest of all, occurring in only28/263 (11%) of the intertarsal joints and 22/155(14%) of the ankle joints.

Natural history of the rheumatoid ankleInitially, pain, joint effusion, and tendon sheathswelling present without radiological changes.The tendon sheaths often communicate with theeffusion2 and may become chronically enlargedwith synovium infiltrating the tendons them-selves, leading to attrition rupture.3 If thearthritis progresses, radiological changes nowdevelop and generally follow two patterns.

Firstly, provided that the foot remains inneutral or in slight valgus or varus, the jointnarrows, erosions appear, cysts develop, and

Table I Radiological changes and previous surgery in 150consecutve rhewnatoid patiens with hindfoot symptoms.Mean disease duration 15 years. Number ofjoints affectedis shown

joint No x ray xRay Operationschanges changes

Forefeet 24 187 89Intertarsal 37 235 28Ankle 145 133 22Knee 147 89 64Hip 232 29 39

Table 2 Radiographic analysis of foot attitude in 150consecutive rhewnatoid patients with hindfoot symptoms.Mean disease duration 15 years. Number ofjoints is shown

Foot aitude Intertarsal Ankle

Valgus 165 43Varus 17 15Equinus 2 8

837

on January 18, 2021 by guest. Protected by copyright.

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.49.S

uppl_2.837 on 1 October 1990. D

ownloaded from

http://ard.bmj.com/

Kirkup

Figure I Standingradiographs ofadvancedrheumatoid arthritis ofhindfoot. Valgus deformityis minimal and the talus hasimpacted into the mortise,leading to equinus andsecondary osteophytes.

Figure 2 Severe hindfootvalgus affecting both thesubtaloid and the anklejointon the left.

Figure 3 Standing radiographs showing early valgus tilt ofthe right talus and secondarychanges in the left ankle with more pronounced tilt and diastasis ofthe inferor tibiofibularjoint.

the talus may collapse proximally into themortise; in longstanding cases osteophytesappear and the foot tends to assume an equinusposition (fig 1).

Secondly, in those patients with markedvalgus, often associated with subtaloid jointdisease or joint laxity, or both, the ankleassumes a valgus position (fig 2). Initially, very

little is seen on routine radiographs, but stand-ing anteroposterior views are more revealingand disclose tilting of the talus into a valgusposition relative to the mortise (fig 3). At acertain stage of this subluxation bony changesdevelop at the point of contact of the supero-lateral corner of the talus and the tibial surface(fig 3); as the valgus progresses the lateralsurface of the talus makes contact against the tipof the lateral malleolus and bending of the lattermay be seen. At a more advanced stage thelower fibula may undergo stress fracture (fig 4)and, rarely, the distal tibia may fracture and re-align the valgus. Ultimately, general changes ofjoint damage appear, mainly owing to mech-anical forces acting on an incompetent mortise,leading to increasing pain and instability, thoughsurprisingly good movement is often retained.On weight bearing the foot assumes a valgusattitude combined with dorsiflexion, whichinduces a flexed attitude at the knee (fig 5).Severe valgus due to both ankle and subtaloidchanges can result in the medial aspect of themidfoot taking abnormal pressure on skin ill-adapted to such stress, resulting in painfulcallouses opposite the navicular or even thehead of the talus and ultimately skin breakdownand local infection (fig 6).

If the foot remains in neutral or close toneutral the sequence of destruction may be veryslow, often over many years. Proximal migrationof the talus in the mortise leads to profoundstiffness, though actual bony fusion is rare (fig1); occasionally even badly destroyed joints maypresent evidence of imperfect repair (fig 7).Although the ankle joint can be the first site

of rheumatoid change in the foot, this isunusual, and in such cases there is often ahistory of previous trauma, usually a significantfracture before the onset of the arthritis. Moreoften disease of the ankle follows establishedsubtaloid and forefoot pathology and the changesare usually bilateral, though frequently one jointcauses disability before the other. At a later datethis joint may ease and the second joint thenbecomes the major source of disability.

Surgical treatmentAs hinted above, most ankles attacked byrheumatoid arthritis can be treated conserva-tively and only some 14% (22/155) in our clinicrequired surgical intervention (table 1).The prime indication for surgery is persistent

pain. In some instances, abnormal gait due tosevere valgus, or more especially equinus, orinstability due to varus are factors in confirmingan operative approach, but deformity withoutpain is rarely an indication. Occasionally skinnecrosis medially may demand intervention,though it is unwise to intervene before localinfection is controlled and scabs have separated.When intertarsal disease is also significant it

may requiire an operative solution before theankle is considered. Sometimes both ankle andintertarsal joints have to be tackled together, aswhen pantalar fusion is necessary. If bilateraldisease causes disability, a staged programme ofsurgery on the two sides may be needed.Clearly, adequate arterial circulation and sound

838

t..


http://ard.bmj.com

/A

nn Rheum



ownloaded from

http://ard.bmj.com/


- ' _ W.Figure 4 Standing radiographs showing late effects ofsevere valgus, including anklejointchanges, bending ofthefibula, and stress fractures ofbothftbula and tibia.

Figure S Severefoot valgus and dorsiflexion combined withknee varus andflexton.

Figure 6 Severe leftfoot valgus, culminating in skinnecrosis over the head ofthe talus. The rightfoot remains inneutral but shows severeforefoot disorganisation.

U -..

Figure 7 Radiographs ofthe same ankle taken at intervalsbetween 1961 and 1975 showinggradual deterioration andfinally some attempt at repair (inferiorfilm on right). Notethat the talus has remained in neutral.

skin are vital prerequisites for safe intervention.Even normal feet have sluggish capillary circu-lation and thin skin over the anterior aspect ofthe ankle. Rheumatoid skin, especially in thosetreated with steroids, is more vulnerable andhas to be handled gently. Undue thumb pres-sure or an exsanguinating Esmarch bandage caneasily produce a degloving injury (fig 8).The patient needs to appreciate the constraints

and limitations of surgery. For example, painrelief by ankle arthrodesis is achieved at theexpense of loss of movement in a major jointand a permanently fixed heel height for foot-wear. Further, the rehabilitation phase afteroperation is often difficult and dependent on thepatient's active participation. Unilateral stagedoperations on the ankle require a period ofdiminished or complete non-weightbearing onthe operated side and hence increased load onthe other side, which may deteriorate as aconsequence. In certain instances the arms areincapable of controlling crutches and non-weightbearing may be impossible for six weeksafter an arthrodesis procedure, at which point aweightbearing cast becomes possible; thuspatients may need to be in hospital for aprolonged period. After ankle arthroplastyactive movement is required within 48 hours ofsurgery if not sooner, and a personal exerciseprogramme should continue for at least threemonths.The operative procedures available are syno-

vectomy, tendon repair and transfer, osteotomy,decompression of the lateral compartment,arthrodesis including pantalar arthrodesis, andankle joint replacement.

Synovectomy and tenosynovectomySwelling of the peroneal and tibialis posteriortendon sheaths is common and can lead torheumatoid synovial infiltration of the tendonsthemselves. This may lead to rupture of thelatter, resulting in a sudden increase of valgus

839


http://ard.bmj.com

/A

nn Rheum



ownloaded from

http://ard.bmj.com/

Kirkup

Figure 8 Degloving type injury offragile rheumatoid skin and wound edge necrosis afterankle arthroplasty. Both lesions healed uneventfully.

and weakness of the foot. Wilkinson in 1965and Marmor in 1967 advised ankle synovectomyfor rheumatoid arthritis,4 5 but Vahvanen,Jakubowski, and Gschwend were the first toreport a significant series of patients.S8 Allrecommended early tenosynovectomy ofaffectedsheaths through lateral and medial incisions,often combining this procedure with partialankle joint synovectomy through the lateralwound. Gschwend reported 20 patients, ofwhom 16 underwent tenosynovectomy; on

follow up at 4-4 years 15 patients were free frompain, the results being better when the articularcartilage was intact. There has been littleenthusiasm for ankle synovectomy in Britain,perhaps because intervention has been offeredtoo late in the disease. In addition, patients are

reluctant to accept what is an essentiallyprophylactic operation when there is hope thatthe condition may resolve. Indeed the response

to treatment, including local steroid injections,is often excellent (A St J Dixon, personalcommunication). On two occasions we haveundertaken synovectomy, including tenosyno-vectomy, as a secondary procedure when per-

forming intertarsal fusion, with the resolutionof ankle symptoms.

If tendons have ruptured, direct repair isimpossible and either a tendon graft is necessary5or the ends are motored to adjacent tendons. Inthe case of tibialis posterior, Jahss has recom-

mended suture to flexor digitorum longus as the

Figure 9 BathAnkle arthroplasty before (left) and after (nght) resection of the distalfibula.

best available procedure,9 though the footremains in valgus.

Tendon transferIn an endeavour to control valgus deformitybefore tendon rupture and before significantjoint damage and fixed deformity appearsSolomon has recommended correction bytendon transfer of flexor digitorum longus intothe navicular tubercle (personal communi-cation). This procedure requires at least sixweeks in a plaster cast, and patients are oftenreluctant to accept surgery when pain, valgus,and swelling are minimal.

Lateral compartment decompressionSome patients with mobile valgus deformitieshave pain well localised to the tip of the lateralmalleolus when weight bearing. If the joint isradiologically intact, except between the mal-leolous and the lateral facet of the talus or the oscalcis, then resection of the distal one centi-metre of the fibula may resolve symptoms, atleast temporarily.'0 It is a minor procedure,enabling the patient to resume activity within 48hours or so. In five instances after a Bath anklearthroplasty pain developed and localised inthe lateral malleolus; exploration showedabnormal contact of its tip with the talus in fourankles, which were helped by resection of thedistal extremity of the malleolus (fig 9).

Ankle arthrodesisSound fusion is the best method of relievingankle pain in unilateral disease and particularlywhen the midtarsal joint remains intact, for atthis joint significant dorsiplantar flexion of thefoot may develop to compensate for true anklemovement. These circumstances are usual withosteoarthritis of the ankle but regrettably rare inrheumatoid arthritis.

In the presence of bilateral stiff or spontane-ously fused subtaloid and midtarsal joints it hasto be recognised that successful fusions of bothankles increase loading and wear of the meta-tarsophalangeal and knee joints which, if alreadyattacked by disease, will deteriorate further. Ifthe hallux and tarsus are already ankylosed, anot uncommon situation, then the knee is themost distal mobile joint, perhaps tolerable onone side but not on both. To rise from a chair orlavatory seat some dorsiflexion in one foot isrequired unless the hands and arms are strongand pain free; an unlikely situation in rheumatoidarthritis. The problem is compounded if theankles are fused in the classical equinus position.For these reasons fusion has to be approachedwith caution.

Nevertheless, a number of authors, includingGschwend," Hamblen,'2 Bolton-Maggs et al,'3Cracchiolo,'4 Smith et al,'5 and Smith andWood'6 are firmly of the opinion that thealternative of ankle arthroplasty is not viable inthe longer term and thus arthrodesis is prefer-able. The few published reports of ankle fusionfor rheumatoid arthritis show that it is not a

simple alternative and that complications of

840


http://ard.bmj.com

/A

nn Rheum



ownloaded from

http://ard.bmj.com/


surgery are common. Thus Smith S R et al,commenting on the results of 30 ankle fusionsobserved that 12 (40%) developed wound break-down and infection and 12 (40%) non-union, sixpatients requiring further attempts at fusion.'5Of 25 operations reviewed after an average offour years' follow up, 52% were consideredgood or excellent, 38% fair, and 100/o poor.Although emphasising the technical difficultiesand high morbidity, they still concluded thatarthrodesis was the procedure of choice.Similarly, Smith and Wood reporting on 11fusions by the Charnley compression techniquenoted that four (36%) suffered pin track sepsisand two (18%) delayed union, though eventually100% fusion resulted. 16 Neither of these papersdiscusses the significant problem of the disabledpatient having to avoid weight bearing for sixweeks-during which, in our experience, theweightbearing ankle and foot often deteriorate,nor do they discuss difficulties posed by fusionof both ankles. Further, we have found thatCharnley clamps inhibit independent mobilityin these frail patients and often damage thefragile skin of their other leg. For these reasonswe have abandoned this method.

Cracchiolo has noted the extensive range offusion techniques published and divides theminto four groups'4: (a) using compressionclamps, (b) using the fibula as a strut graft, (c)using the tibia as a transposed graft, and (d) adowel technique and others. He emphasises thatthe final choice depends on the actual disease,previous incisions and internal fixation devices,and individual preference. Latterly, we havefavoured an anterolateral approach, dividing thefibula at the level of the joint, which providesgood access to clear the joint surfaces and tocorrect deformity; a curved gouge is useful inpreparing the medical malleolous. Satisfactorycancellous bony contact is usual and can besupplemented with cancellous bone from thefibula, tibia, or os calcis. The position isadjusted and held with three or more wirestaples at tibiotalar level while the distal fibula isstapled into the talus and beneath the tibia,

Figure 10 Spontaneous tarsalfusion followed by surgical ankle arthrodesis in slightdorstflexmn.

though the latter is not essential. A plaster backslab is applied and changed to a full cast at twoweeks. Weight bearing is avoided for six weeksafter surgery and then a weightbearing cast isfitted until there is radiological fusion.The angle of fusion is highly critical. Buck et

al studied the optimum position by gait analysisin 19 patients after fusion for trauma.'7 Theyconcluded that the best position was neutral,0-5° of valgus, and 5-10° of external rotation,and noted that 50 of dorsiflexion was bettertolerated than 50 of plantar flexion. We are incomplete agreement that plantar flexion mustbe avoided at all costs and find that mostrheumatoid patients are best in slight dorsiflexion(fig 10), though patients with a knee flexioncontracture need at least 100 of dorsiflexion.

Ankle joint replacementIf both ankles are severely disabled the questionof ankle arthroplasty arises, on at least one side,especially if there is bilateral tarsal ankylosisand fixed great toes. For these patients arthro-plasty offers potential advantages comparedwith arthrodesis, including a quick return toweight bearing often without splintage orapparatus.

Since Buchholz'8 and Lord'9 first reportedankle arthroplasty with modern materials in1973 a variety of prostheses have becomeavailable. Most of these are constrainedprostheses designed to reproduce dorsiplantarflexion in the saggital plane only; the Lord,Smith,20 and Bath2' joints, however, have aspherocentric configuration, which permits un-constrained motion, including supination andpronation. The rationale for using such a 'balland socket' joint is derived from observationsthat, when the intertarsal joints are fused, theankle joint can develop polyaxial motion tocompensate for the loss of subtaloid and mid-tarsal motion. Thus children with congenitaltarsal fusion (fig 11), children after Still'sdisease, and adults after longstanding rheuma-toid arthritis may form ball and socket ankleconfigurations and present polyaxial foot move-ment.

In 1979 Demottaz et al studied 21 total anklereplacements, of which 16 were for rheumatoidarthritis, with an average follow up of 14-7months.22 Seven of the joints were Smith orsimilar multiple axis joints, while the remainderwere single axis articulations, principally of theMayo design. The follow up was very compre-hensive and included gait analysis and electro-myographic studies. Pain relief was consideredcomplete in only four instances, while radio-lucent lines were present in 19 joints, of whichtwo were loose. In this small series with a shortfollow up no significant difference between thetwo groups was observed. They concluded thatankle arthrodesis was the preferred operation,except for elderly patients with limited midtarsalmotion. In 1985, Bolton-Maggs et al followedup 41 of 62 uniaxially designed Imperial CollegeLondon Hospital ankle prostheses insertedbetween 1972 and 1981 (mean follow up 5 5years), during which the prosthesis was modifiedseveral times and both anterior and posterior

841


http://ard.bmj.com

/A

nn Rheum



ownloaded from

http://ard.bmj.com/

Kirkup

Figure 11 Congenital tarsalfusion with 'ball and socket' ankle providing universal motion.

approaches were used.'3 Of 34 arthroplasties forrheumatoid arthritis, seven were lost to followup, five underwent arthrodesis for loosening,and 22 were reviewed. Pain was absent in seven,mild in seven, moderate in five, and severe inthree. The average range of movement changedfrom 18° to 230 and walking ability was generallyimproved. Only six rheumatoid arthroplastieswere considered fully satisfactory, and it wasconcluded that total ankle replacement couldnot be recommended as a long term solution.They also noted, however, that arthrodesis inrheumatoid arthritis might be less satisfactorythan in osteoarthritis.

In a review editorial stimulated by the latterpaper Hamblen considered that the ankle couldnot be replaced reliably by current prosthesesbut expressed the view that improved un-cemented designs might reverse that con-clusion.'2

Other reports have been more optimistic. In1982 Herberts et al reporting on 18 ICLHprostheses, 13 for rheumatoid arthritis, at amean of 36 months, concluded that anklearthroplasty has a definite place in the treatmentof severe arthritis in rheumatoid patients.23They considered that the osteoarthritic ankledid less well and found a high incidence ofloosening and radiolucent zones. In 1984Lachiewicz et al reporting on 15 uniaxial arthro-plasties (14 of the Mayo type) all for rheumatoidarthritis, at a mean of 39 months, notedgratifying pain relief and rated seven anklesexcellent and eight good.24 Nevertheless, 11ankles developed radiolucent lines and sixcomponents showed evidence of subsidence.

Our own experience supports the view thatarthroplasty benefits the severely disabledrheumatoid patient with bilateral hindfootdisease. Of 20 polyaxial Smith prostheses (figs12 and 13) inserted in 17 rheumatoid patientsbetween 1975 and 1979, 15 were reviewed at amean of seven years. Six components becamepainfully loose, two of which were revised andone of which developed infection at 6-8 yearsand was removed, while nine ankles had little orno pain, though two of these had ankylosedtotally. Skin healing was often delayed (fig 8)but ultimately sound in all cases. This experi-ence indicated that the polyaxial concept couldprovide adequate movement in all planes in thebest cases, and was not prone to dislocation.The single thickness design of the Smith joint,however, restricted the surgeon's ability to

Figure 12 (A) RichardSmith anklejoint with steeltibial and polyethylene talarimplant. (B)Bath andWessex anklejoint withpolyethylene tibial and steeltalar implant.

842


http://ard.bmj.com

/A

nn Rheum



ownloaded from

http://ard.bmj.com/

Rhewnatoid arthritis and ankle surgery

Figure 15 Bath and Wessex implant in situ showingmovement at the implant interface under load, and alsoradiolucent lines. Note the previous talonaviculararthrodesis. Afterfiveyears the patient is pleased and able tocycle.

Figure 13 Smithjoint 16 months after operation; in addition, the talonavicularjoint wasstapled. After seven years thepatent was very pleased and able to dance.

Figure 14 A sekction ofBath and Wessex ankle trials and implants. Upper row: talartrials. Middle row: tibial trial and polyethylene implant cut to length in the anteroposteriorplane. Lower row: talar implants.

accommodate variations of vertical joint spacedue to individual variations of collateral liga-

mentous tension. We therefore designed a

prosthesis with a polyethylene tibial implant tobe cemented in situ under pressure with a

special clamp, and a choice of steel talarimplants of 2, 3, 4, 5, and 6 mm thicknessinserted during os calcis pin traction (figs 12and 14).A preliminary report2' of these Bath and

Wessex prostheses, inserted in 20 patientsbetween 1980 and 1982, at a minimum followup of three years, was encouraging (fig 15). Afuller account was presented in 1987 at the 16thcongress of the International College of FootMedicine and Surgery, Vienna (Kirkup, un-

published data), when 66 Bath prostheses(1980-85), as well as 24 Smith prostheses(1975-79) were analysed. Of the Bath joints, 57were for rheumatoid arthritis, including 10patients with bilateral replacements. Delayed

wound healing was seen in 14 patients three ofwhom while receiving long term steroidsdeveloped a low grade infection and sinusformation; one ankle developed a late infectionafter a trouble free 13 months. All four infectedprostheses were removed leading to resolutionof the infection and a stable ankylosis after aprolonged period in a cast. At a mean of 4-2years five patients had died for reasons uncon-nected with their ankle and one underwent mid-thigh amputation for an infected knee pros-thesis.Of the remaining 47 joints, most had radio-

lucent lines without symptoms (fig 15). Six wereunder observation for loosening, while two wererevised and one underwent arthrodesis forloosening. Dorsiflexion and plantar flexionimproved in 24, was the same in 16, anddiminished in seven ankles. Prosupinationimproved in 20, was the same in 20, anddiminished in seven feet. Thirty two ankleswere free from pain or experienced discomfortafter walking or standing for half an hour, 15ankles experienced moderate pain on weightbearing, and none experienced severe pain.

Thus, of 57 Bath prostheses, six could not beassessed owing to death or amputation; of theremainder, 44 (86%) were in situ at an averageof 4-2 years, and 32 (63%) were either free frompain or experienced discomfort after activity.Although these results are modest and do notcompare favourably with current hip and kneereplacement arthroplasties, they resemble thoseof the earlier versions of knee prostheses.Clearly, modifications of ankle prostheses andoperative techniques can be expected and mayparallel the evolution of knee replacement.We conclude that for severely disabled rheu-

matoid patients with bilateral tarsal ankylosisand crippling ankle pain, spherocentric jointreplacement is justified on one side and some-times on both. The advantages include thepreservation of hindfoot movement, rapid re-habilitation often without splintage, andimprovement of the patients' independence.The disadvantages include a risk of early or lateinfection, and implant loosening in osteoporotic

843


http://ard.bmj.com

/A

nn Rheum



ownloaded from

http://ard.bmj.com/

Kirkup

bone. Careful selection of patients is essential,and those receiving steroids, those with fragileskin or with severe foot valgus may have toaccept either the limitations and risks of arthro-desis or continued conservative management.

1 Vidigal E, Jacoby R K, Dixon A St J, Ratliff A H, Kirkup J.The foot in chronic rheumatoid arthritis. Ann Rheum Dis1975; 34: 292-7.

2 Dixon A St J, Graber J. Local injection therapy. Basle:EULAR, 1981.

3 Tilhman K. The rheumatoidfoot. Stuttgart: Thieme, 1979: 59.4 Wilkinson M C, Lowry J H. Synovectomy for rheumatoid

arthritis. Bone Joint Surg [Br] 1%5; 47: 482-8.5 Marmor L. Surgery ofrheumatoid arthritis. London: Kimpton,

1967: 216.6 Vahvanen V. Synovectomy of the talocrural joint in rheuma-

toid arthritis. Annales Chirurgiae et Gynaeceologiae Fenniae1968; 57: 576-82.

7 Jakubowski S. Synovectomie des oberen Sprunggelenkes.Orthopade 1973; 2: 79.

8 Gschwend N. Surgical treatment of rheumatoid arthritis. Stutt-gart: Thieme, 1980: 260-3.

9 Jahss M H. Spontaneous rupture of the tibialis posteriortendon: clinical findings, tenographic studies and a newtechnique of repair. Foot Ankle 1982; 3: 158-66.

10 Benjamin A, Helal B. Surgical repair and reconstruction inrheumatoid disease. London: Macmillan 1980: 204.

11 Gschwend N. Stable fixation in hindfoot arthrodesis, avaluable procedure in the complex R.A. foot. In: HagenaF-W ed. Rheumatoid arthritis surgery of the complex hand andfoot. Basel: Karger, 1987: 114-25.

12 Hamblen D. Can the ankle joint be replaced?JBoneJointSurg[Br] 1985; 67: 689-90.

13 Bolton-Maggs B G, Sudlow R A, Freeman M A R. Totalankle arthroplasty: a long-term review of the LondonHospital experience. J. Bone Joint Surg [Br] 1985; 67:785-90.

14 Cracchiolo A. Operative technique of the ankle and hindfoot.In: Helal B, Wilson D, eds. The foot. Edinburgh: ChurchillLivingstone, 1988: 1205-44.

15 Smith S R, Morgan C G, Pinder I M. Results andcomplications of ankle arthrodesis [Abstract]. Bone JointSurg [Br] 1990; 72: 530.

16 Smith E J, Wood P L R. Ankle arthrodesis in the rheumatoidpatient [Abstract]. Bone joint Surg [Br] 1990; 72: 530.

17 Buck P, Morrey B F, Chao E Y S. The optimum position ofarthrodesis of the ankle. Bone Joint Surg [Am] 1987; 69:1052-62.

18 Buchholz H W. Totale Sprunggelenk-endoprosthese'Model St Georg'. DerChirurgie 1973; 44: 241-4.

19 Lord G, Marotte J H. Prothese totale de cheville: techniqueet premier resultats. A propos de 12 observations. Rev ChirOrthop 1973; 59: 139-51.

20 Kirkup J R. Richard Smith ankle arthroplasty. R Soc Med1985; 78: 301-4.

21 Marsh C H, Kirkup J R, Regan M W. The Bath and Wessexankle arthroplasty.J BoneJointSurg [Br] 1987; 69: 153-4.

22 Demottaz J D, Mazur J M, ThomasW H, Sledge C B, SimonS R. Clinical study of total ankle replacement with gaitanalysis. J Bone3Joint Surg [Am] 1979; 61: 976-88.

23 Herberts P, Goldie I F, Korner L, Larsson U, Lindborg G,Zachrisson B E. Endoprosthetic arthroplasty of the anklejoint: a clinical and radiological follow-up. Acta OrthopScand 1982; 53: 687-96.

24 Lachiewicz P F, Inglis A E, Ranawat C S. Total anklereplacement in rheumatoid arthritis. J Bone Joint Surg[Am] 1984; 66: 340-3.

844


http://ard.bmj.com

/A

nn Rheum



ownloaded from

http://ard.bmj.com/

rheumatoid arthritis and surgery · joint damage and fixed deformity appears solomon has...

Documents