Annals of the Rheumatic Diseases 1994; 53: 828-832

Clinical experience with thalidomide in themanagement of severe oral and genital ulcerationin conditions such as BehSet's disease: use ofneurophysiological studies to detect thalidomideneuropathy

J MM Gardner-Medwin, N J Smith, R J Powell

AbstractObjective-To examine the efficacy, dose,and safety profile, including neurophysio-logical testing of thalidomide used in 59patients (including 23 with Behcet'sdisease) to treat severe oral or genitalulceration (OGU).Methods-We identified prospectivelysubjects (including women ofchildbearingpotential) who had persistent OGU overperiods lasting one to 40 years and whoseactive ulceration was not controlled byother therapies. They were treated withthalidomide. Retrospectively, we identi-fied the number of subjects with completeresolution of the ulcers at one and twomonths of thalidomide therapy, and thedose required to maintain that improve-ment in those individuals who relapsedafter stopping thalidomide. The decreasefrom the baseline sensory nerve actionpotential (baseline SNAP) amplitudevalue (derived from median, radial andsural nerve SNAPs) at which thedevelopment ofparaesthesiae was likely tooccur was also determined.Results-Complete resolution of theulcers occurred in 81% of patients withinone month ofthalidomide therapy at dosesof200 mg/day. No further thalidomide wasrequired by 20% of patients respondingand in the remainder improvement wasmaintained with smaller doses (7-200 mg/day). Using an approximate 50% decreasefrom baseline SNAP as an indication todiscontinue thalidomide, the incidence ofsymptomatic neuropathy was 13*5%. Nopatients with a decrease of less than 42%developed neuropathy, and a further13 5% were asymptomatic with a decreasein SNAP between 42 and 69%. Other sideeffects were seen in 44% ofpatients. Therewere no pregnancies and no requirementfor urgent pregnancy testing.Conclusions-Thalidomide provided a usefiutherapeutic option in severe oral and genitalulceration which had not responded toother therapies. The physician mustremain vigilant to the continuing dangerof axonal neuropathy and teratogenesis atal times during thalidomide therapy.

(Ann Rheum Dis 1994; 53: 828-832)

A report from Turkey in 1982' led to our ownpreliminary studies2 3 with thalidomide,confirming its therapeutic value in severe oralor genital ulceration (OGU) including patientswith Behcet's disease. Thalidomide has beenshown to inhibit tumour necrosis factor (xsynthesis,4 5 and this is likely to be themechanism of action in severe OGU.Although the licence for the use of thalido-

mide in the United Kingdom was revoked in1961 after recognition of its teratogenic6 andneuropathic complications,7 its use on a"named patient" basis has continued inaccordance with Section 9(1) of the MedicinesAct 1968.8 The axonal neuropathy associatedwith thalidomide therapy is notable as it ispainful and irreversible.9 We present ourexperience using thalidomide in severe OGU,and in particular the use of sensory nerveaction potential amplitude measurements inthe detection of thalidomide neuropathy.

Patients and methodsPATIENTSCollection ofdataSixty three patients with OGU have beentreated with thalidomide by one of us (RJP) inNottingham during the past decade.Prospectively, sensory nerve action potential(SNAP) data on these patients were collectedand thalidomide dosages and duration oftherapy were recorded on a named patientbasis by the Pharmacy Department at Queen'sMedical Centre. Retrospectively, the patient'smedical records were available in 59 cases (infour, the notes were untraceable or destroyed)and allowed confirmation and retrieval ofdiagnosis, previous treatment regimens,contraception, therapeutic efficacy and sideeffects in each case. In 50 of the 59 patients,electrophysiological data had been recorded inNottingham. The anticipated technicalvariability of such data between centresexcluded the remaining results from thataspect of the analysis.

Indication for therapyAll patients had suffered severe oral or genitalulceration, or both, which had failed torespond to therapies which included topicaltreatments, oral and pulse intravenousprednisolone, and in some patients even

Clinical ImmunologyUnit, Department ofImmunology, Queen'sMedical Centre,Nottingham NG7 2UH,United KingdomJ MM Gardner-MedwinR J PowellDepartment of ClinicalNeurophysiologyN J SmithCorrespondence to:Dr R J Powell.

Accepted for publication29 July 1994

828

on October 24, 2021 by guest. P

rotected by copyright.http://ard.bm

j.com/

Ann R

heum D

is: first published as 10.1136/ard.53.12.828 on 1 Decem

ber 1994. Dow

nloaded from

Thalidomide in the management of ukeration: neurophysiological detection of neuropathy

cytotoxic agents. The duration of history ofOGU before receiving thalidomide rangedfrom one to 40 years (mode 10 years). Twentyfive of the 59 patients were clinically diagnosedas having severe idiopathic orogenitalulceration (one patient was also diabetic) and23 had Behcet's disease. The remaining 11patients included two with systemic lupus ery-thematosus (one complicated by tuberculosisand metastatic breast carcinoma), two(cousins) with chronic granulomatous disease,and one case each of cyclical neutropenia, C4deficiency/vasculitis, mixed essential cryo-globulinaemia, pemphigus, AIDS, pyodermagangrenosum with underlying Hodgkin'slymphoma, and erythema multiforme.

Consent and information given to patientsBefore treatment, patients were warned of thepossible side effects of thalidomide, particu-larly of peripheral neuropathy, and the need forregular neurophysiological studies was empha-sised. Women of childbearing age werecounselled on the teratogenic effects of thalido-mide. Thalidomide was refused unless thewoman agreed to use adequate reliablecontraceptive measures, and demonstratedthat she understood the possible consequencesof contraceptive failure.

ContraceptionThirty eight of the 59 patients were female, ofwhom eight were postmenopausal at the timeof treatment and one had not reached themenarche. The 29 women of childbearing agewho were treated with thalidomide practisedthe following contraceptive measures: contra-ceptive pill (CP) (seven), vasectomy of partner(six), hysterectomy (seven), coil (one),sterilised (five), condom (two), CP followed bysterilisation (one). There were no pregnanciesin any women taking thalidomide or occasionson which urgent pregnancy testing wasrequired.

THALIDOMIDE DOSAGE AND DURATION OFTREATMENT

Although the dose of thalidomide was initiallystandardised in all patients, it became clearthat smaller doses of the drug were effective.From 1982 until 1986 the dosage of thalido-mide was taken from the work of Saylan andSaltik1 and was 400 mg for five days, followedby 200 mg/day to complete a four week course(6-6 g). Further thalidomide was administeredonly if the ulcers recurred, according to one oftwo basic regimens. If patients continued tohave chronic persistent ulceration, theyreceived thalidomide 200 mg daily for onemonth, then gradually reduced to the smallestdose that controlled the majority of ulcers(7-100 mg/day). However, if there weresignificant periods of relative freedom fromulcers between severe exacerbations,intermittent monthly courses were given asrequired. From 1986 to the present, the initial28 days of thalidomide treatment was at thereduced dose of 200 mg/day (5-6 g). A fewpatients with marked drowsiness were given

smaller doses of 100 mg/day until the ulcerscleared. The total doses of thalidomidereceived were in the range 0-56-152 g overperiods of 8-3369 days.

NEUROPHYSIOLOGICAL STUDIESTo detect early changes suggestive of axonalneuropathy, SNAP amplitudes of the median,radial, and sural nerves were recorded bystandard techniques1012 in a temperaturecontrolled environment by the same observerfor all readings in 46 subjects. After baselinereadings, studies were repeated every sixmonths or after each 10 g of thalidomide if thisoccurred earlier.

In the early phase of our thalidomide usage,nerve conduction studies were not performedbefore treatment in 14 patients. Although all14 had received less than 7 g before their firststudy, it became apparent that this could bemisleading, as one patient had evidence ofaxonal neuropathy on his first recordingdespite only receiving 2-8 g of thalidomide.Subsequently, recordings of SNAP beforetreatment was begun became mandatory.

In normal subjects the radial nerve has asignificantly larger SNAP amplitude than boththe median and sural nerves; consequently,when the baseline SNAP value was calculatedfor each patient, the amplitudes of the threenerves (median, radial, and sural) requiredequal weighting. For this purpose the SNAPamplitude for each individual nerve at theinitial reading was deemed to be 100%, and thethree summated. A decrease of approximately50/O from this baseline total percentage SNAP(baseline SNAP) was prospectively selected asan indication of subclinical neuropathy, andthalidomide was discontinued.

SUBSET ANALYSISData from a group of 12 patients who hadrequired long term thalidomide treatmentnecessitating more than four SNAPmeasurements during their therapy, wereanalysed further. The amplitudes of theindividual nerves and their total SNAPs werecalculated and plotted against the cumulativedose of thalidomide, the daily dose of thalido-mide, and the duration of thalidomide therapyto determine if the clinical development ofperipheral neuropathy could be predicted.

ResultsTHERAPEUTIC EFFICACYThe efficacy ofthalidomide in the whole cohortwas such that 81 3% were ulcer free at onemonth, and 84-7% ulcer free at two months.Of those with idiopathic OGU, 91-6% wereulcer free at one and two months, and inBehcet's disease 73-9% were ulcer free at onemonth and 82-6% at two months. For theother diagnoses, 72-7% patients were ulcer freewithin two months. A further two patientsimproved but never became ulcer free, andseven stopped thalidomide because of sideeffects before improvement was noted. Of the

829

on October 24, 2021 by guest. P

rotected by copyright.http://ard.bm

j.com/

Ann R

heum D

is: first published as 10.1136/ard.53.12.828 on 1 Decem

ber 1994. Dow

nloaded from

Gardner-Medwin, Smith, Powell

Table 1 Relationship between dose and duration of thalidomide therapy (time between thefirst and last dose of thalidomide) and the degree of deterioration in sensory nerve actionpotential amplitudes (SNAP)

Decrease from total SNAP

<50% (n = 34) >50% (n = 16)

Cumulative dose (g) 50-46 (15 2-89-8) 47-82 (17-3-115)Duration of therapy (days) 464 (8-2555) 1079-5 (28-3369)Cumulative dose with time (g/day) 0-147 (0-004-0 378) 0-08 (0-013-02)

Values are mean (range); n = number of patients.

50 patients who responded to thalidomide, 36(Behcet's 16, idiopathic OGU 14, other six)required either a single further course, or lowdose continuous thalidomide therapy forrecurrence of the ulcers. The range of dosesrequired to maintain control of the ulcers was7-200 mg/day. Of the five patients who couldnot tolerate thalidomide because of sideeffects, three nevertheless showed a sustainedreduction in both severity and duration ofulcers after receiving thalidomide. Thalido-mide had no obvious effect on the otherfeatures of these patients' diseases.

Table 1 shows the doses of thalidomide.

SIDE EFFECTS

Symptomatic neuropathyEight of 59 individuals (13-5%) developedparaesthesiae of hands and feet; only oneof them developed a painful neuropathyclassically associated with thalidomide (patientwith SLE, tuberculosis and metastatic carci-noma of the breast). Four of the eight hadBehcet's disease (two of whom were the onlypatients taking cyclosporin A concurrently withthalidomide), a single patient had idiopathicOGU and the remaining three comprised the

Table 2 Individual percentage decreases from the totalsensory nerve action potential amplitudes (SNAP) at thetime ofstopping thalidomide because of deterioration in thenerve conduction tests, in eight patients who wereasymptomatic and eight who showed symptoms ofneuropathy

Decrease in total SNAP (%o)

Asymptomatic Symptomatic

69-0 51 449-8 48-457.4 62-753.7 53.7426 61-759.4 47.452 6 64-354-3 51 4

Table 3 Percentage decreasefrom the baseline total sensory nerve action potentialamplitudes (SNAP) in all patients with a >40% decreasefrom baseline, in whomsubsequent SNAP measurement was made after stopping thalidomide, and time intervalbetween the two measurements, in patients with and without paraesthesiae

Patient No. Decreasefrom baseline total SNAP Time betweenmeasurement

At time of stopping At subsequent (days)thalidomide (%) measurement (%)

Symptomatic1 51*4 26-0 7272 48-4 38-6 2523 62-7 42-2 5924 53.7 28-0 304

Asymptomatic1 69-0 31*2 8352 49-8 25-8 4133 42-6 13.0 206

individuals with SLE and metastatic carcinomaof the breast, mixed cryoglobulinaemia, andHodgkin's disease with pyoderma gangren-osum. In patients who developed neuropathicsymptoms the total SNAP amplitudes haddecreased by 47.4-64.3% from the initialbaseline readings (table 2).Four of these eight patients with

paraesthesiae had further measurements afterstopping thalidomide and these showed animprovement in baseline SNAP, albeit withinthe 30% variability (table 3). No furtherdeterioration was noted in the total SNAPamplitude of any patient after thalidomidewithdrawal. Six patients reported an improve-ment in their symptoms, one indicated nochange and one died (patient with SLE,tuberculosis and metastatic carcinoma of thebreast) before reassessment.The patient with Behcet's disease and co-

existent diabetes mellitus had no demonstrableperipheral neuropathy at any time.

Asymptomatic deterioration in baseline SNAP:putative subclinical neuropathyEight individuals without paraesthesiae (fivewith Behcet's disease and three with idiopathicOGU) had a decrease of 42-6-69% frombaseline SNAP (table 2) and their thalidomidewas discontinued. Subsequently, three of theseindividuals had further SNAP recordings afterstopping therapy; two showed improvement of>30%, and one showed an improvement of lessthan 30% (table 3). None of these patientsbecame symptomatic after stopping thalido-mide.Comparison of the 16 patients with clinical

and putative subclinical neuropathy with the34 patients with a decrease of less than 42%showed that the duration of thalidomidetreatment was longer in those with 'neuro-pathy', but there was no difference between thetotal cumulative doses of the two groups.These data are summarised in table 1. Thedaily average dose per patient was less in theneuropathic group, reflecting the low doses ofthalidomide in those receiving long termtherapy, as opposed to the greater dose in thosereceiving intermittent courses. Within thegroup of 16 neuropathic individuals,asymptomatic subjects (n = 8) received a largercumulative dose for longer (mean 51 42 g overa mean of 1221.1 days) than did symptomaticsubjects (n = 8) (mean 44-23 g over a mean of937 9 days). For the complete cohort, the rateof deterioration for a given patient could notbe predicted.

Serial SNAP in long term treatmentTwelve patients received thalidomide treat-ment for longer than two years andconsequently had undergone more than fournerve conduction studies; their data wereanalysed further. In seven of these patientsthalidomide was withdrawn because of a >50%reduction in amplitude from baseline SNAP(range 51-69%, mean 57.24%); two of theseven developed symptoms suggestive ofperipheral neuropathy. All seven stoppedthalidomide. Only three of them subsequently

830

on October 24, 2021 by guest. P

rotected by copyright.http://ard.bm

j.com/

Ann R

heum D

is: first published as 10.1136/ard.53.12.828 on 1 Decem

ber 1994. Dow

nloaded from

Thalidomide in the management of ukeration: neurophysiological detection of neuropathy

40- \ _

20- b

O 0-0 200 4004434

24

°I Days0

0 200 400Days

Sensory nerve action potential amplitude (;expressed as a percentage of the baseline recmedian (X), radial (0), and sural (td) ne?SNAP (A), against days of thalidomide t/cumulative dose (Dose) of thalidomide ove

period in a single patient.

had further SNAP measurementsshowed improvement in the amp62-7 and 42.6% decreases fro:baseline measurement on stop:mide, increasing to 31-2, 42-2, ax

than the baseline, respectivel)continuing thalidomide).The remaining five of these

showed a decrease in SNAP (0-6treatment, yet none developed 1The figure shows the changes icumulative dose of thalidomideone of these patients.

Non-neuropathic side effectsNon-neuropathic side effectsmented in 26 of 59 (44%)Behcet's, 12 idiopathic OGU, 1

Drowsiness was a common featsevere enough to necessitate a

dose or cessation of therapy,patients. Other side effects includoedema (two), constipation (tlgain (seven), dizziness (three),headache (three), and tinnitus (o:

DiscussionThese data represent our practicover 10 years in the use of thpatients with conditions suchdisease characterised by longstaOGU; the study should not be:trial.

In patients with a long histouncontrolled oral and genital u

have confirmed'3 14 the therapthalidomide, and demonstrated tof severe ulcers in more than 80after one month of treatment.thalidomide required to heal themaintain improvement were 4lower than previously described.'

SNAP amplitudes were measured asto indication of axonal neuropathy because

conduction velocities would not be expected toshow significant early changes.9 In 1991 thepoor reproducibility of SNAP in documentingprogressive neuropathy was described byChaudry et al'5 and detailed by Bleasel et al.'6The latter report demonstrated a wide range of

'. variability in the individual nerve sensoryamplitudes between serial measurements innormal subjects at constant temperature:median 26/9%, ulnar 32- 1%, sural nerves

F 31 2%. We noted a similar variability in theamplitude for the individual nerves (figure)

600 800 and chose to use the sum of the amplitudes ofthe three nerves, to minimise the dominantcontribution of the radial nerve. If patients hada decrease of >30% from baseline SNAP, we

600 800 repeated the test sooner than planned.Our observations and a study of thalidomide

SNAP) in discoid lupus erythematosus by Knop et al"7cording, in the highlight the importance of baseline studies inrves, and total the interpretation of changes in SNAP. Ourzerapy, and the experience suggests that a single baseline,r the same time SNAP recording may create difficulties in the

interpretation of the significance of SNAPchanges, and that two studies before the

s and all three commencement of treatment are needed toolitudes (69-0, improve the interpretation of subsequentm the initial change, by allowing the calculation of confi-ping thalido- dence limits for each patient. Alternatively, tond 13-0% less control for the analytical and biologicaly, after dis- variation, a large normal population should be

studied so that critical differences can be12 patients established.'8

A42-6%) with All patients demonstrated an overallparaesthesiae. decrease from baseline SNAP values whilstn SNAP and taking thalidomide. Prospectively, we decidedwith time for to repeat SNAP measurements every six

months or after 10 g of thalidomide if this wassooner. We also selected a decrease ofapproximately 50% as an indication to stop

were docu- thalidomide. In retrospect, this was apatients (11 reasonable estimate of the critical level ofthree others). deterioration in SNAP associated withture, but was paraesthesiae; however, it was too high, as itreduction in failed to predict symptoms in all patients. Noin only 14 patient with a decrease in total SNAP of less

led dependent than 42% developed symptoms. Thus anyhree), weight patient with a >30% decrease in SNAP fromrash (five), baseline (which therefore exceeds the expected

ne). variability) should be considered to havedemonstrated a significant deterioration inSNAP, and should have their treatment withthalidomide monitored even more closely.

:al experience In the literature the reported incidence oflalidomide in peripheral neuropathy in patients takingas Behcet's thalidomide appears to vary between

inding severe conditions. In prurigo nodularis it approachesregarded as a 1000%,' contrasting with discoid lupus

erythematosus in which the incidence isry of severe, approximately 28%.17 Leprosy may be a specialilceration, we case, confounded by lepromatous neuropathy;eutic role of however, the incidence does seem genuinely tothe resolution be low.2" In idiopathic OGU/Beh9et's disease,% of patients the reported incidence of neuropathy, albeit inThe doses of small numbers, lies between 23 and 50%.21 22

ulcers and to Careful monitoring with regular neurophysio-demonstrably logical studies has resulted in lower incidence

of symptomatic neuropathy (13-5%) in our

a)

-0%.001-

zU)

0)(0)0

a

831

on October 24, 2021 by guest. P

rotected by copyright.http://ard.bm

j.com/

Ann R

heum D

is: first published as 10.1136/ard.53.12.828 on 1 Decem

ber 1994. Dow

nloaded from

Gardner-Medwin, Smith, Powell

patients. Even in those patients with para-esthesiae, an improvement in symptoms wasseen after thalidomide was stopped.

Differences between the OGU patientsdeveloping neuropathy and those who did notincluded concurrent cyclosporin A therapy andunderlying malignancy. We suggest thatpatients with additional risk factors for thedevelopment of neuropathy should bemonitored with increased vigilance, includingmore frequent SNAP measurements. Theimportance ofwarning patients to stop thalido-mide with the development of paraesthesiaemust be stressed, as electrophysiologicaltesting in our protocol failed to anticipate theclinical symptoms in eight of the 59 patients.The need to stop thalidomide before thedevelopment of symptoms, or immediatelythey develop, is emphasised. After stoppingthalidomide, six of the eight patients withsymptoms of peripheral neuropathy showed animprovement in their symptoms, but these didnot resolve in all. This contrasts with the workof Fullerton et al,9 who observed individualswith longstanding painful neuropathy sec-ondary to thalidomide and noted someimprovement or recovery in 50/O and nochange in the remainder. In addition, we haveshown that those patients considered to havesubclinical neuropathy also demonstrated animprovement in the SNAP after stoppingthalidomide.The rate of deterioration from baseline

SNAP for a given patient could not bepredicted accurately from their dose orduration of therapy. Consequently, individualpatients must be monitored regularly for theduration of thalidomide therapy with baselineand regular electrophysiological testing.Prospectively we chose the criteria 10 g ofthalidomide or six months of therapy, which-ever was earlier, as the intervals for repeatingnerve testing. This was not sufficientlyfrequent in all patients as neuropathy diddevelop, and shorter periods between testswould be required to prevent the onset ofneuropathy in all patients.We observed a lower incidence of clinical

neuropathy, particularly painful neuropathy,than previously described with thalidomidetherapy,22 and suggest that regular monitoringwith SNAP contributed to this outcome.However, peripheral neuropathy did occur inthis cohort, and we present our observations tohighlight the complex issues arising from themanagement of patients treated with thalido-mide, which would benefit from further clinicalstudies.

The continued, albeit limited, use of thalido-mide must recognise the significant neuro-pathic and teratogenic potential of this drug.Special clinical vigilance is demanded tomaintain the safest possible use of thalidomidefor disabling conditions such as severe OGUwhich have failed to respond to conventionalmedication. Its safe use should be assisted bythis description of our clinical experience over10 years, and the recent development of aclinical guideline to promote the safest possibleuse of thalidomide.23

1 Saylan T, Saltik I. Thalidomide in the treatment of Behcet'ssyndrome. Arch Dermatol 1982; 118: 536.

2 Bowers P W, Powell R J. Effect of thalidomide on orogenitalulceration. BM3r 1983; 287: 799-800.

3 Jenkins J S, Powell R J, Allen B R, Littlewood S M, MauriceP D L, Smith N J. Thalidomide in severe orogenitalulceration. Lancet 1984; 2: 1424-6.

4 Sampaio E P, Samo E N, Galilly R, Cohn Z A, Kaplan G.Thalidomide selectively inhibits tumour necrosis factor (xproduction by stimulated human monocytes. .7 Exp Med1991; 173: 699-703.

5 Sampaio E P, Moreira A L, Sarno E N, Malta A M, KaplanG. Prolonged treatment with recombinant interferon -yinduces erythema nodosum leprosum in lepromatousleprosy patients. J7Exp Med 1992; 175: 1729-37.

6 McBride W G. Thalidomide and congenital abnormalities[letter]. Lancet 1961; 2: 1358.

7 Fullerton P M, Kremer M. Neuropathy after intake ofthalidomide (Distaval). BMJ 1961; 2: 855-8.

8 The Medicines Act 1968. London: HMSO, 1968.9 Fullerton P M, O'Sullivan D J. Thalidomide neuropathy:

a clinical electrophysiological and histological follow-upstudy. _7 Neurol Neurosurg Psychiatny 1968; 31: 543-51.

10 Gilliatt R W, Sears T A. Sensory nerve action potentials inpatients with peripheral nerve lesions. _7 Neurol NeurosurgPsychiatry 1958; 21: 109-18.

11 Downie A W, Scott T R. An improved technique for radialnerve conduction studies. .7 Neurol Neurosurg Psychiatry1967;30: 332-6.

12 Burke D, Skuse N F, Lethlean A K. Sensory conduction ofthe sural nerve in polyneuropathy. .7 Neurol NeurosurgPsychiatry 1974; 37: 647-52.

13 Mascaro J M, Lecha M, Torras H. Thalidomide in thetreatment of recurrent, necrotic and giant mucocutaneousaphthae and aphthosis. Arch Dermatol 1979; 115: 636-7.

14 Grinspan D J. Significant response of oral aphthosis tothalidomide treatment. AmAcad Derm 1985; 12: 85-90.

15 Chaudhry V, Comblath D R, Mellitis E D, et al. Inter- andintra-examiner reliability of nerve conductionmeasurements in normal subjects. Ann Neurol 1991; 30:841-3.

16 Bleasel A F, Tuck R R. Variability of repeated nerveconduction studies. Electroenceph Clin Neurophysiol 1991;81: 417-20.

17 Knop J, Bonsmann G, Happle R, et al. Thalidomide in thetreatment of sixty cases of chronic discoid lupuserythematosus. BrJ7 Dermatol 1983; 108: 461-6.

18 Fraser C G, Fogarty Y. Interpreting laboratory results:Analytical and biological variation must be taken intoaccount. BM7 1989; 298: 1659-60.

19 Wulff C H, Hoyer H, Absoe-Hanson G, Brodthagen H.Development of polyneuropathy during thalidomidetherapy. BrJ7 Derm 1985; 112: 475-80.

20 Sheskin J. Therapeutische Erfahrungen uber den Einflussdes Thalidomids bei der Lepra-Reaktion. Hautarzt 1975;26: 1-5.

21 Ochonisky S, Verrous J, Bestuji-Garin S, Gheradi R, RevuzJ. Thalidomide neuropathy: incidence and clinico-electro-physiological features in 42 patients. Rev Med Interne1993; 14(suppl 1): 109.

22 Hamza M, Oueslati M, Hamida M B. Peripheralneuropathy induced by thalidomide in Behcet's disease.Rev Med Interne 1993; 14(suppl 1): 110.

23 Powell R J, Gardner-Medwin J M M. Guideline for theclinical use of thalidomide. Postgrad Med.7 1994; 70. Inpress.

832

on October 24, 2021 by guest. P

rotected by copyright.http://ard.bm

j.com/

Ann R

heum D

is: first published as 10.1136/ard.53.12.828 on 1 Decem

ber 1994. Dow

nloaded from