32oumal of Neurology, Neurosurgety, anid Psychiatri 1993;56:372-381

Nature and incidence of peripheral nerve

syndromes in HIV infection

Geraint N Fuller, Jean M Jacobs, Roberto J Guiloff

AbstractFifty four patients with peripheral nervesyndromes were seen during a 15 monthperiod in a population of about 1500 HIVinfected patients at all stages of the dis-ease. Distal symmetrical peripheral neu-ropathies were seen in 38 of the 54patients, (11.5% of AIDS patients) andcould be distinguished into two forms.The most common (n = 25) was a painfulperipheral neuropathy during AIDS,which is distinct clinically and pathologi-cally, having axonal atrophy, and is asso-ciated with cytomegalovirus infection atother sites. The 13 non-painful neu-ropathies seen were more heterogenous.Lumbosacral polyradiculopathy associat-ed with cytomegalovirus and lymphoma-tous mononeuritis multiplex occurred infewer than 1% of AIDS patients. Mono-neuropathies were seen in 3% of AIDSpatients. No patients with acute orchronic inflammatory demyelinatingpolyradiculoneuropathies were seen. Theannual incidence of neuropathies duringthe AIDS related complex stage was lessthan 1%; none were seen in asympto-matic HIV seropositive patients.

(3 Neurol Neurosurg Psychiatry 1993;56:372-38 1)

Department ofNeurology,Westminster Hospital,LondonG N FullerR J GuiloffDepartment ofNeuropathology,Institute of Neurology,Queen Square,LondonJ M JacobsCorrespondence to:R J Guiloff, Department ofNeurology, Westminster andChelsea Hospital, 369Fulham Road, LondonSW10 9NH, UKReceived 4 Februarv 1992and in revised form19 June 1992.Accepted 6 August 1992.

A variety of peripheral nerve syndromes havebeen described at different stages of HIVinfection, from seroconversion to AIDS.Most described are case reports or smallselected series which do not estimate the inci-dence. A large prospective study of HIVpositive patients found that peripheral neuro-

pathies were rare', and a cross sectional studyof 40 hospitalised patients with AIDS foundthat 33% had peripheral neuropathies.2 Wereport the nature and frequency of peripheralnerve syndromes seen during a 15 monthprospective, population based, referral seriesfrom a population of about 1500 HIV posi-tive patients3 and discuss the findings in rela-tion to current classifications of peripheralnerve syndromes.

MethodsPATIENTS

Patients were included if they had beenreferred to the neurology department withperipheral nerve syndromes between 1

September 1988 and 31 December 1989from the population of HIV infected patients

who attended hospitals and clinics in the EastRiverside area. Patients were excluded if theyhad been exposed to neurotoxins (drugs suchas vincrinstine, isoniazid, or thalidomide orregularly took more than 40 units of alcohol aweek) or had other systemic conditions, notrelated to HIV, associated with peripheralneuropathy (such as diabetes or uraemia) orhad a family history of peripheral neuropathy.Patients were classified as having AIDSaccording to the CDC criteria4; they wereclassified as having AIDS related complex ifthey were HIV positive and had persistentgeneralised lymphadenopathy, fever, oralhairy leukoplakia, oral candidiasis, herpeszoster radiculitis, and foliculitis but lackedAIDS defining opportunistic infections ortumours. At the time of the study the neuro-toxic nucleoside analogues DDI and DDCwere not being used.

At the start of the study (September 1988)1353 HIV positive patients were being fol-lowed up at these clinics. A total of 748 hadAIDS related complex and 261 had AIDS (CDalton, personal communication). BySeptember 1989 there were 1765 HIV posi-tive patients, of whom 778 had AIDS relatedcomplex and 297 had AIDS. Of the HIV pos-itive patients, 83% were homosexual or bisex-ual, 4-4% were drug misusers, and 2 5% wereheterosexual and did not misuse drugs. For10% the risk factors were unknown. Therewere no haemophilics. Between September1988 and 1 September 1989, 105 patientsdied with AIDS in this population.

Patients were assessed with a standardisedneurological examination. The five found tohave bilateral upper motor neuron signs atpresentation of peripheral neuropathy will beconsidered separately to allow the clinical fea-tures of the neuropathy to be studied in isola-tion. The following HIV related andnutritional data were collected in a standard-ised way: duration of AIDS, previous HIVrelated infections (using CDC criteria4), CD4count (ACSCAN TM, Becton Dickinson),body mass index (weight divided bv heightsquared), serum B- 12 concentration (DICO-PAC radioimmunoassays; normal range210-900 pg/ml), and serum albumin concen-tration. The same data were collected for theneurologically asymptomatic AIDS controlsand the series of sequential patients wvithAIDS admitted to hospital (see below).

Patients were classified according to thestage of the disease and the pattern of theperipheral nerve syndrome (distal symmetri-cal peripheral neuropathv, mononeuritis mul-

372 on S

eptember 7, 2020 by guest. P

rotected by copyright.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.56.4.372 on 1 A

pril 1993. Dow

nloaded from

Natzure anid itncidenice ofperipheral nerve syndronies in HIV infection

tiplex, polyradiculopathy, or mononeuro-

pathy). The distal symmetrical peripheralneuropathies were distinguished as those withpain and those without. Pain was defined as

an unpleasant sensory experience subjectivelyrecognised by the patient and mentioned inthe history. No formal pain scale was used.This distinction is further considered in thediscussion.

Standardised electrophysiological studiesand pathological studies of the sural nerve

were performed using previously describedmethods.

CONTROL GROUPSNeurologically asymptomatic AIDS patients-Patients with AIDS but without signs or

symptoms of peripheral neuropathy were

assessed clinically (n = 30) and were studiedelectrophysiologically (n = 30) and pathologi-cally (n = 5). Data related to HIV and nutri-tion were noted as described above. Thegroup has previously been described indetail;7 and it provided electrophysiologicaland pathological control data (see table 4).

Patients with AIDS admitted to hospitalsequentially (n = 30) were assessed separatelyand the standardised HIV related data were

collected. This data has been published previ-ously.8 Sequential patients, with or withoutneuropathies, were recruited to determine thefrequency of different opportunistic infectionsin patients with AIDS from the populationunder study, unselected in relation to theirneurological symptoms (see tables 1 and 2).

Sudden death controls-The sural nerves

from eight patients (mean age 42 (range21-58) years) without risk factors for HIVinfection and without history of peripheralneuropathy, all of whom died suddenly, were

examined morphometrically and used as con-

trols for pathological data. They have beendescribed previously.57Normal controls-Nerve conduction studies

have been carried out on 23 healthy men agematched (mean 31-2 (21-45) years) with theneurologically asymptomatic AIDS patients;the results (see table 4) have been publishedpreviously.7

STATISTICSNormally distributed variables were com-

pared by using one sample Student's t test.Non-normally distributed variables were

compared by using the Mann-Whitney Utest. Proportions were compared by using thetwo tailed Fisher's exact test. All statisticswere performed with the Statgraphics statisti-cal package (STSC Plusware, 1987).

ResultsFifty four patients with peripheral nerve syn-

dromes were seen. All were male. Thirty eighthad distal symmetrical peripheral neuro-

pathies; 25 (66%) of these were painful, 13painless. Eleven patients had mononeu-

ropathies, three had mononeuritis multiplex,and two had lumbosacral polyradiculopathy.Table 1 gives the HIV and nutritional vari-ables for the control groups and for thepatients with distal symmetrical peripheralneuropathies.

PAINIFUL PERIPHERAL NEUROPATHIESTwelve of 25 patients with painful peripheralneuropathies (PPN) were reported on previ-ously.8 The variables related to HIV infectionand nutrition are given in table 1 and the pre-

ceding infections in table 2. The duration ofAIDS, CD4 count, body mass index, albu-min concentration and percentage of patientswith abnormal vitamin B,2 did not differ fromthose in the control groups. Five patients hadclinical evidence of an associated myelopathy.

Clinical featuresPain was the most prominent symptom, com-

monly described as burning (10/20) or

intense pins and needles (9/20). Two patientsreferred to it as frostbite and one as havingthe soles of his feet beaten with bamboosticks. Four patients with painful dysaesthesiaalso had shooting pains to the feet. The painwas almost always made worse by contact(19/20), often by as little as the weight of thesheets on the bed. Most patients had to stopwearing their normal shoes and usually had toresort to slippers. One patient became wheel-

Table I HIV infection and nutrition in patients with peripheral neuropathies and controls with AIDS

Controls with AIDS

Patients wuith Sequential Neurologically Patiehts withPPN admissions asymptomatic OPN(?u = 25) (n = 30) (n = 30) (n = 13)

Mean(SD) age (years) 36 (6 5) 40 (9-4) 37 (9-1) 41 (7 3)

HIV infection:Median (interquartile range)

duration of AIDS 11-0 (3-17) 14-0 (3-23) 11-5 (3-20) 12-5 (3-24)Median (interquartile range)CD4 count (Oil) 4-0 (7-21) 18-5 (10-40) 20-5 (10-68) 12-0 (6-25)

Nutritional indicatorsMean(SD) body Mass Index 19-3* (2-8) 19 9** (2-1) 17-3 (2-5)

(kg/m2; normal range 21-3-29-0)Mean(SD) albumin concentration 34 (5-6) 34 (5 9) 35 (5-4)

(g/l; normal range 35-45 g/l)% of patients with abnormal 8-3 24 4 (31)serum B-12

PPN = painful peripheral neuropathy; OPN = other (non-painful) peripheral neuropathy.*p < 0-02 v other (non-painful peripheral neuropathy).**p < 0 005 v other (non-painful peripheral neuropathy).

373 on S

eptember 7, 2020 by guest. P

rotected by copyright.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.56.4.372 on 1 A

pril 1993. Dow

nloaded from

Fuller, Jacobs, Guiloff

Table 2 Percentage ofprior and current infections in patients with painful peripheral neuropathy (PPN),non-painful peripheral neuropathy (OPN) and controls with AIDS.

Controls with AIDS

Patients with Sequential Neurologically Patients withPPNt admissions asyniptonmatic OPN(n = 22) (n = 30) (n 30) (n = 13)

Pneumocystis carinii pneumonia 69 67 70 69

Cytomegalovirus:Any site 78** 37 40 31Retinitis 46 27 23 8Gastrointestinal 18 10 20 23Pneumonitis 23 6 3 8

Kaposi's sarcoma 32 27 27 46Visceral candidiasis 19 20 23 15Cryptosporidiosis 0 17 17 15Toxoplasmosis 6 13 13 0Lymphoma 6 0 3 0Cryptococcal meningitis 9 0 3 0Mycobacterium avium intracellulare infection 4 3 0 31 *

tExcludes the last three patients seen (see text).*p <005, **p < 001.Probability of proportion v (Fisher's exact test).

chair bound because of extreme contact pain.The pain was limited to the legs in all but twopatients, mostly affecting the feet (n = 17). Itwas limited to the toes in two patients and tothe soles in one. One patient had low backpain at the onset. The symptoms were sym-metrical in all but four patients.The onset of pain was subacute, usually

beginning with stiffness in the toes, later pro-gressing to become painful. Maximum painusually occurred within two to four weeks(13/20); in one patient the pain reached itspeak in less than a week; in another it tookover two months. The pain usually improvedspontaneously (in eight patients it resolvedcompletely) but in six patients it persistedunchanged. One patient had recurrence ofpain followed by further improvement. Somesymptomatic relief was obtained with carba-mazepine in seven patients and amitriptylinein eight, but both drugs failed to help in oneinstance. Two patients reported that marijua-na alleviated the pain. Two patients hadKaposi's sarcoma in the feet but it wasthought unlikely to be the cause of the painas neither the site nor the time course ofthe pain matched that of the Kaposi'ssarcoma.

All patients complained of numbness,which was again mostly in the feet (17/20), tomid-calf in one patient, only in the toes inone, and in the soles in another. Threepatients noted numbness in their fingertips.Paraesthesia was felt in the feet in 13, to mid-calf in one, in the toes in one, and the finger-tips in two. No patients complainedspontaneously of weakness. Three patientshad observed their toes moving spontaneous-ly, an observation confirmed in two. In twopatients, mononeuropathies had been the firstindication of peripheral nerve dysfunction(meralgia paraesthetica in one and a probablecommon peroneal palsy in the other). A fur-ther patient had an episode of meralgiaparaesthetica six months before the onset ofhis pain. Half the patients had noted that theybecame impotent at the onset of pain; in only

two was this volunteered spontaneously. Afurther three noted urinary symptoms; onehad difficulty with micturition and two hadfrequency.

Distal weakness with wasting of extensordigitorum brevis was found in five of 20patients. In a further patient, distal weaknessdeveloped 10 months after the onset of pain.At the first assessment the ankle reflexes werenormal in 10, impaired in five, and absent infive. There was a progressive loss of anklereflexes, with the reflex being absent at thefinal assessment in 14 patients and impairedin three. No patient had distal lower limboedema.

Sensory abnormalities were predominantlyof superficial sensation. All patients had alter-ation of light touch or pinprick (loss of lighttouch in 12 and pinprick in 15). Subjectivechanges extended to the ankle while the losswas usually in the toes. In all but threepatients the changes were limited to the legs.In these three the loss in the arms wasrestricted to the fingertips in two and to thelittle finger in one. Temperature sensationwas altered in 15; vibration sense was abnor-mal in 15 but only to the level of the toes.Position sense was impaired at the toes in six.In eight, the sensory signs did not alter. In 10there were changes: improvement in two, adeterioration in eight. The change was slightwith the exception of those patients who wereinitially assessed within one or two weeks ofthe onset of their symptoms.The clinical features of patients with

peripheral neuropathies are summarised intable 3. The clinical features of the fivepatients who also had signs of a myelopathyat presentation were similar, but in additionthey had bilateral pyramidal weakness andpathologically brisk reflexes in the arms andat the knee. The ankle reflexes were abnormalin all five at presentation.Twenty three patients were followed up

until death (median 7 months, range 1-21months); two were alive at review at 10 and14 months.

374 on S

eptember 7, 2020 by guest. P

rotected by copyright.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.56.4.372 on 1 A

pril 1993. Dow

nloaded from

Natuire atid incidence ofperipheral nerve syndromes in HIV infection

Table 3 Clinical features ofpatienits with paitnful (PPN\)and other peripheral nieuropathies (OPN) withoutevidenice of myelopathy, values are nunubers of patietnts

Pain:BurningDvsaesthesiaeShootingFrostbite

Associations:NumbnessParaesthesiaeWeaknessMoving toesBack pain

Onset:0-2 Weeks2-4 Weeks4-8 Weeks) 8 Weeks

Site:ToesFeetFingers

Wasting (distal)Weakness (distal)

Ankle reflex:NormalImpairedAbsent

Sensory:Cotton wool:

AlteredLost

Pinprick:AlteredLost

Vibration senseProprioceptionTemperature

PPN(ni = 20)

,Syvnptomls

10942

201703

131340

2183

Signs

55

1055

OPANOu = 9)

875

0

0

0

0

8

0

95

46

22

5

20 813 4

20 814 315 76 2

15 8Upper level of sensory signs:Toe 2Midfoot 4Ankle 10Midshin or above 4Fingers 3

ProgressPain:

ResolvedImprovedPersistedDeteriorated

Ankle reflex:ReducedLost

Sensory loss:ImprovedStableDeteriorated

I0

5

2

5

6950

6 19 0

2 110 48 4

Cerebrospinalfluid.Cerebrospinal fluid was studied in 12patients. All had five or less white cells, theprotein concentration was raised in seven

(mean 0-71 (range 0-2-1-3) g/l), in four toover 1 g/l.

ElectrophysiologyTable 4 shows that the reduction in the sen-

sory nerve action potentials was greater in thelegs than in the arms. Sensory conductionvelocities were significantly lower than inneurologically asymptomatic AIDS patients.Motor action potentials in the ulnar nerve didnot differ from those in this control group.The common peroneal compound muscleaction potential was significantly smaller.Distal motor latencies of the common per-oneal were slightly prolonged. There was no

difference in the ulnar nerve distal motor

latency. The F waves in extensor digitorumbrevis were not recordable in 46% of patientswith PPN, but when recordable the latencywas no different from that in controls. Thepatients with myelopathy were comparableelectrophysiologically to those without.

PathologyPathological study of eight sural nervesrevealed no inflammation with degeneratingfibres and a reduced myelinated fibre density(mean(SD) 4396 (931); fig 1). This is lowerthan the myelinated fibre density in the fiveneurologically asymptomatic AIDS patients(5 454 (631); p = 0 06). The neurologicallyasymptomatic AIDS patients had a meanreduction in fibre density of 30 5% comparedto the sudden death controls (mean(SD) fibredensity 7 845 (1414)). Axonal atrophy wasfound in the myelinated fibres in nerves fromall eight patients with painful peripheral neu-ropathy studied qualitatively (figs 1 and 2)and in the four of them studied quantita-tively.6 There were clusters of Schwann cellprocesses with a few unmyelinated axons sug-gesting unmyelinated axon loss (mean(SD)density 22 280 (3022)). This was not signifi-cantly different from density in the five (neu-rologically asymptomatic AIDS patients (22006 (2357)). The unmyelinated fibre densi-ties were lower than in the sudden death con-trols (30 900 (7117)) but this did not reachstatistical significance. These results havebeen reported in detail elsewhere.6

AssociationsA study of the first 12 of these patients withPPN revealed a temporal relation withcytomegalovirus infection with 75% havingprevious or current cytomegalovirus infectioncompared to 37% in the control group ofsequential admissions who had AIDS of simi-lar severity, as assessed by duration of infec-tion and CD4 counts (table 1; p < 0.05).8This association held when all 25 patientswere considered; the rate then was 78% or80%. In the 13 further cases, eight had clearevidence of cytomegalovirus infection (retini-tis in three, colitis in two, pneumonitis intwo, and oesophagitis in one); in three treat-ment for cytomegalovirus infection had beenstarted, partly on the basis of PPN. Of theselast three patients, two had PCP negativepneumonia with CMV and no otherpathogens found in induced sputum, and one(who declined biopsy) had a herpes simplexvirus negative perianal ulcer. Thus, the rate ofcytomegalovirus infection was 80% in allcases and 78% if the last three were excluded.The rate of other infections did not differfrom a control group of sequential admissionsnor from that found in neurologically asymp-tomatic controls (table 2).8The serum B- 12 concentration in patients

with PPN did not differ from that in the neuro-logically asymptomatic AIDS patients whowere at a similar stage in the disease. Only8-3% had abnormal concentrations of B2, andthe albumin concentrations were similar to theneurologically asymptomatic group (table 1).

375 on S

eptember 7, 2020 by guest. P

rotected by copyright.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.56.4.372 on 1 A

pril 1993. Dow

nloaded from

Fuller, J7acobs, Guiloff

Table 4 Nerve conduction studies in peripheral neuropathies in AIDS patients and control groups. Values are means(SD) or [nfedian (iinterquartile range)]

Patients wvith peripheral nieuiropathy Coutrolst

Neutrologicall/vPPN (u = 23) OPN (ui = 13) asy,mptoilatic AIDS(ui = ?0) Norial (i = 2.?)

Sensory conduction:Sensory nerve action potential (, V)

Ulnar 9 4 (4 5)* 6-5 (3.5)** 12-0 (4 2) 15 4 (6-0)Median 12 2 (6 2)* 11-3 (6 4)* 17 5 (6 6) 22-1 (8 4)Sural [1-0 (0-4)]* [0-8 (0-2 6)]** 9-9 (3-5) 15.0 (4 5)

Motor conduction:Compound muscle action potential (mV)

Ulnar 9 4 (3-9) 10-0 (2 4) 10 5 (2 6) 14 4 (298)Common peroneal [1 4 (0 6-2 0)]** [ 0 7 (0-1l5)]** 3 7 (2-0) 5-7 (2 8)

Distal motor latency (ms)Ulnar 3-1 (0 8) 3-2 (0-7)* 2 8 (0 7) 2-7 (0 4)Common peroneal 5-4(12)* 6 3 (1l8)* 4-6 (0 7) 45 (0-8)

Motor conduction velocity (m/s)Ulnar 52 2 (8 0)** 51 3 (4 8)** 57-3 (4 9) 60 2 (4 7)Common peroneal 36 6 (5-1)** 34-2 (4.7)** 43.9 (5*0 49 8 (4 8)

F waves (m/s)Abductor digiti minimi 68-0 (7-0) 67 3 (6 1) 70 2 (5 5) 73.9 (4 9)Extensor digitorum brevis 35-2 (5 5) 30-2 (2.5)** 36 6 (3 9) 42 3 (3 8)

*p = 0-05-0-01, **p < 0-01; comparison with neurologically asymptomatic controls. tA comparison of normal controls with ncu-rologically asymptomatic AIDS patients has been published. tF-waves corrected according to formula: height (F-wave latency -M-wave latency - 1). PPN = painful peripheral neuropathy; OPN = other (non-painful) peripheral neuropathies. SAP = SensorVnerve action potential, CMAP = Compound muscle action potential, DML = Distal motor latency, MCV = Motor conductionvelocity, ADM = Abductor digiti minimi, EDB = Extensor digitorum brevis

OTHER PERIPHERAL NEUROPATHIESThe characteristics of the 13 patients withnon-painful neuropathies (OPN) are given intable 1; the preceeding infections are given intable 2. The severity of AIDS in these neu-ropathies was not significantly different fromeither of the control groups, nor from patientswith painful neuropathies.

Clinical featuresSix patients had sensorimotor neuropathies,four had- additional bilateral upper motorneuron signs, and three had purely sensoryneuropathies. Five of the patients had symp-

toms or signs in both arms and legs; in fourthe symptoms were limited to the legs. Threepatients had clinical evidence of super-addedmononeuropathies. Eight of the patientscomplained of numbness, seven of paraesthe-sia, and five of weakness. The onset of theneuropathy was insidious, in most patientscoming on over two months, in one possiblyover years. Focal wasting distally was foundin four patients; in six there was distal weak-ness. Ankle reflexes were absent in fivepatients, impaired in two, normal in two.Superficial sensation was altered in all butone patient, being limited to the toes in one,

Figure I One micronaraldite sections of suralnerve x 470. (a) Froni apatient with painfulperipheral neuropathy:there is a reduced densityoffibres; nmanyfibres areatrophic and a few aredegenerating (arrows);(b) front a patient withnon-painful peripheralneuropathy: thefibredenlsity, is redulced aniddegeneratingfibres are seen(arrozvs); (c) controlnerve.

n

Se;i.

:w , 9.)e.C -iX b

* A.£2

376 on S

eptember 7, 2020 by guest. P

rotected by copyright.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.56.4.372 on 1 A

pril 1993. Dow

nloaded from

Nature and incidenice ofperipheral nierve synidronmes in HIV infectiotn

Figure 2 Electron micrograph of sural ierve from patient with paitnful peripheralnieuropathy, showing three atrophic nmyelinated fibres with axons that are snmall relativemyelin sheath thickness. Bar = 5 lin.

the foot in five, and the knees in two. Senscloss in the fingers was found in five patienposterior column loss was found in sevwThe symptoms and signs of the neuropatwere progressive in four patients, staticfour, and apparently improved in one. T'patients had pes cavus. Four patients wmyelopathy and neuropathy had similar clical pictures.

Twelve patients were followed until dee(median 8 months, range 1-17 months); owas reviewed nine months after onsetsymptoms.

ElectrophysiologyTable 4 summarises the electrophysiolcfindings. Sensory nerve action potentials wreduced, more in the legs than in the arnSensory conduction velocities were signcantly reduced when compared with thoseneurologically asymptomatic AIDS patierMotor action potentials in the ulnar nerve (

not differ from those in this control groiThe common peroneal compound musaction potential was significantly smallDistal motor latencies of the common poneal were slightly prolonged. There wasdifference in the ulnar nerve distal molatency. The F waves in extensor digitoribrevis were not recordable in 46% of 1

patients, although when they were recordathe latency was no different from that incontrol group. The patients with myelopawere comparable electrophysiologicallythose without.

CerebrospinalfluidCerebrospinal fluid was acellular inpatients studied. Four out of five patients Iraised concentrations of protein (mean 0g/l, range 0'4-1 7 g/l).

PathologyPathological findings have been reported indetail elsewhere.6 There was evidence ofmyelinated fibre loss with degenerating fibres(fig 1) but no evidence of axonal atrophy inthree sural nerves studied quantitatively.6Mean (SD) myelinated fibre density was4796 (414) and unmyelinated fibre densitywas 22 475 (5043).None of the five sural nerves of neurologi-

cally asymptomatic AIDS patients with sub-clinical neuropathy7 showed axonal atrophyqualitatively, nor was it found in one nervestudied quantitatively.6

AssociationsTwo associations were noted. Firstly, thebody mass index was significantly lower thanin either the asymptomatic AIDS controls(p < 0 005) or in patients with painful periph-eral neuropathy (p < 0 02) (table 1).Secondly, Mycobacterium avium intracellulareinfection was more common than in a groupof sequential AIDS controls (4/13 v 1/30;8 p< 0 03). The rate of other infections did not

to differ from that in the neurologically asymp-tomatic or sequential admission AIDS controlgroups (table 2). A total of 31% of thesepatients had abnormal concentrations of B,2,

zry a slightly higher rate, though not significantlyLts; so, than in the neurologically asymptomaticen. AIDS group.:hyin MONONEURITIS MULTIPLEX

wo One patient aged 57 with AIDS related com-ith plex presented with a rapidly progressiveni- monoplegia associated with pneumococcal

pneumonia. He was positive for hepatitis Bath surface antigen. He deteriorated rapidly andne at necropsy a large vessel vasculitis of theof femoral nerve was found.

Two patients were seen with mononeuritismultiplex related to lymphoma. One aged 41presented with fevers and weight loss together

)gy with a right femoral neuropathy. Hepaticere ultrasound showed multiple lesions. He diedns. shortly after investigation and was found toifi- have a non-Hodgkin B cell lymphoma inin nodes and liver. Lymphoma was found infil-

its. trating his femoral nerve (fig 3). Another

Figure 3 Paraffin section (x 140) offemoral nerveshowinig infiltration of the endoneuriumi by colunins oflyniphonia cells.

377 on S

eptember 7, 2020 by guest. P

rotected by copyright.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.56.4.372 on 1 A

pril 1993. Dow

nloaded from

Fuller, Jracobs, Guiloff

patient, aged 34, presented with multiple cra-nial nerve palsies and a cervical radiculopathyand was found to have lymphomatous menin-gitis.

ISOLATED MONONEUROPATHIESThree patients with isolated mononeu-ropathies had AIDS related complex andeight had AIDS. Their mean (SD) age was41-1 (6-7) years. The patients with AIDSrelated complex had a median CD4 count of160/,ul and a mean body mass index of 26&6.Those with AIDS had a median CD4 countof 55/,il and a body mass index of 21-6. Thenerve most commonly affected was the lateralcutaneous nerve of the thigh (n = 5), fol-lowed by the common peroneal (n = 3) andmedian (n = 3) nerves. The neuropathies pre-sented usually either during or after thepatients had been admitted to hospital (7/11)for a variety of infections or tumours. In mostno local cause was found. In one patient ameralgia paraesthetica occurred at the sametime and side as inguinal lymphadenopathydue to lymphoma; it resolved with treatmentof the lymphoma. In a further seven patientsthe mononeuropathy resolved spontaneously.In three (one with meralgia paraesthetica, onecommon peroneal and one median), thesymptoms persisted.

Results of electrophysiological studies innerves other than those affected clinically didnot differ from those in neurologically asymp-tomatic AIDS patients.

POLYRADICULOPATHYTwo patients with lumbosacral polyradicu-lopathy have been reported in detail else-where.9

DiscussionWe have described the range of peripheralnerve syndromes seen in a 15 month periodin a relatively stable outpatient population ofHIV positive patients at various stages of theirdisease. As the population was known at twopoints, an estimate of the population halfway

through the study can be obtained and thiscan be used to estimate the 15 month inci-dence of the various syndromes (table 5).Painful peripheral neuropathy was the mostcommon, occurring in 7 5% of AIDSpatients; non-painful peripheral neuropathieswere seen in 4%. Mononeuropathy occurredin 3% of AIDS patients, and mononeuritismultiplex and a lumbosacral polyradiculopa-thy were each seen in 06%. Neuropathiesduring the AIDS related complex stageoccurred in fewer than 1% of patients.

These estimates contain several possiblesources of inaccuracy. Firstly, the patientsmay not have been referred to the neurologyservice. Secondly, the population was chang-ing and patients may have sought treatmentelsewhere. Thirdly, patients were excludedwhen they had other potential causes for neu-ropathy, even though these causes may nothave produced the relevant neuropathy.These exclusions prevent this study fromassessing whether pre-existing neuropathypredisposes to the development of PPN.

In table 5 the incidence and relative ratesof different patterns of neuropathies are com-pared with those of other unselected series,including case series, prevalence studies, anda point prevalence study. In all studies, distalsymmetrical peripheral neuropathies were themost common and account for some 90% ofall neuropathies in AIDS. In early series theywere estimated to affect 2-5% of AIDSpatients.'01' In a study comparable to oursfrom Baltimore,'2 13% of AIDS patients haddistal symmetrical peripheral neuropathies. Inour series, lumbosacral polyradiculopathy andmononeuritis multiplex each were found infewer than 1% of patients with AIDS andAIDS related complex.

Other researchers have described a syn-drome similar to painful peripheral neuropa-thy, usually grouped together withnon-painful neuropathies, as "predominantlysensory neuropathy," "sensory motorneuropathy," "distal symmetrical polyneu-ropathy," or "painful neuropathy".'2-'5 Whenthe clinical features were similar to those

Table 5 Series ofperipheral nteuropathies of six or nmore cases

Aiuthorvear

Snider et al, 1983"'Levy e al, 1985"

Guiloff et al, 1988'Cornblath and

McArthur, 19882

**Miller et al, 1988'tSo et at, 1988

Lange et al, 1988' 5

***Leger et at, 19892***Chaunu et al, 198921

This series

Ceitre Population DSPN

New YorkSanFranciscoLondonBaltimore

SanFranciscoSanFrancisco

160 AIDS318 AIDS orARC122 AIDSEst 200 AIDS? ARC? HIV+NS

40 AIDS

New York NS AIDSHIV +

Paris NS AIDSARCHIV+

London 331 AIDS763 ARC565 HIV +

8 (5'S,)5 (2'S)24 (3' i,)

26 (13'S,)

18

13 (33%h)

158

38 (1 1 5'S.)

CIDP AIDP

1 (05'S>)17

103

235

AlAI

9 (3"/,)1 (O 8'!,)

LS PR AIX

1 (0 8"o)

2 3

1 1

2 (0 6N/,) 2 (0(6'"0) 8 (3 !0"o)1 (O "X.) 3 (0)4"'o)

DSPN = Distal symmetrical peripheral neuropathy. CIDP = Chronic inflammatory demyelinated polyradiculoneuropathy. AIDP= Acute inflammatory demyelinating polyradiculoneuropathy. MM = Mononeuritis multiplex. LSPR = Lumbosacral polyradicu-lopathy. MN = Mononeuropathies. NS = not stated. ARC = AIDS related complex. Percentages of given populations are inbrackets. **not classified by disease stage. ***2 overlapping series combined. tLimited to hospitalised patients.

378 on S

eptember 7, 2020 by guest. P

rotected by copyright.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.56.4.372 on 1 A

pril 1993. Dow

nloaded from

Nature and incidence ofperipheral tnerve syndronmes itn HIVinfection3

described here, pain was reported in about60% of patients, in all limited to the feet.Complaints of weakness are uncommon.

Mild distal weakness was found in some andthe ankle reflexes were usually lost or

impaired. Sensory loss was usually mild, mostoften affecting vibration sense and pinprickand often associated with contact dysaesthe-sia. ' Previous reports have noted that thepain has remained prominent,'5 although theneuropathy has progressed.'2 In 75% of our

patients with painful peripheral neuropathythe pain improved or resolved and patientsrequired less analgesia or pain modulatingagents such as tricyclics or carbamazepine.The sensory loss remained stable or improvedin 60%, although there was a progressive lossof ankle reflexes. Several possible explana-tions might account for these differentdescriptions: the entities may differ in some

fashion; patients who improve tend to be lostto follow up and this may eschew some serieswith limited follow up; and the treatmentwith anti-cytomegalovirus drugs may alter theclinical course of the syndrome.

Defining painful peripheral neuropathy on

the basis of a single subjective clinical symp-tom poses several difficulties. Firstly, individ-uals' pain thresholds vary. Secondly, in thesetting of HIV disease, central nervous systemdisorders such as myelopathy andencephalopathy may modulate the pain.Thirdly, the pain might result from more

than one peripheral nerve pathology in AIDS.In this study, as in our earlier report,8

active cytomegalovirus infection was found in80% of patients with painful peripheral neu-

ropathy as compared to 37% of sequentialunselected controls who were at a similarstage of the disease and who were assessed inthe same way as patients with neuropathy.This should have removed any diagnostic biasin identifying cytomegalovirus. The virus isknown to be an important pathogen in AIDSrelated lumbosacral polyradiculopathy'6 andin a multifocal sensory motor neuropathyassociated with AIDS.'@ On the basis of theclinical syndrome, we suggested that the painin painful peripheral neuropathy may berelated to a dorsal root ganglionitis,8 and,indeed, several case reports have found a dor-sal root ganglionitis at necropsy.'8I"

The electrophysiological findings in thisseries of patients with distal symmetricalperipheral neuropathy mirrored those fromprevious studies.""12' Sensory action poten-tials were considerably reduced, especially inthe legs, with a mild reduction of motor con-

duction velocities and compound muscleaction potentials. The F waves have beenreported as being prolonged,'5 although thecomparison was with normal controls ratherthan asymptomatic AIDS patients. Usingsuch AIDS patients as controls, we found no

difference. The findings were similar inpainful and other peripheral neuropathies andconsistent with axonopathy. The velocitieswere reduced but to a lesser degree than theamplitude of sensory motor action potentials,suggesting that demyelination is not a major

pathology. This is supported by the F wavevelocities not differing from those in our neu-rologically asymptomatic AIDS controls.Axonal loss was found in all our cases of dis-tal symmetrical peripheral neuropathy andhas been consistently reported by others.'3 '120-21 Myelinated fibre densities were reducedto a similar degree in this and other studies.'23 The myelinated fibre density in patientswith PPN was 20% lower than that in neuro-logically asymptomatic controls but no differ-ent from that in patients with OPN. Thedensity of unmyelinated axons was the sameas in the asymptomatic controls (22 260 v 22016). The unmyelinated fibre density and theunmyelinated axon density of the patientswith non-painful peripheral neuropathieswere similar to those of patients with painfulneuropathies.The only morphological finding that distin-

guished patients with a painful peripheralneuropathy from those with other peripheralneuropathies and the neurologically asympto-matic AIDS controls was axonal atrophy.Qualitatively, this was found in all eight suralnerves of patients with painful neuropathiesbut not in eight nerves of patients with non-painful neuropathies. Quantitatively, it wasseen in all four patients with painful neu-ropathies studied but in none of the threepatients with non-painful distal symmetricalneuropathies nor in any neurologicallyasymptomatic AIDS patients., A two tailedFisher's exact test on these data rejects thenull hypothesis of independence betweenPPN and axonal atrophy (p < 0.03).6 Therewas no concomitant atrophy of unmyelinatedaxons.6

These findings can be interpreted in sever-al ways. Firstly, it can be speculated that thesubclinical peripheral neuropathy found inneurologically asymptomatic AIDS,7 theother non-painful neuropathies, and thepainful peripheral neuropathy in AIDS repre-sent different points in a continuous spec-trum. The significantly longer duration ofOPN than PPN at presentation (table 3) isagainst the idea that a non-painful neuropa-thy progresses to a point when it becomespainful. Alternatively, the painful and thenon-painful peripheral neuropathies may rep-resent distinct entities which occur in the set-ting of a non-specific "axonal. loss of chronicdisorders." Evidence of this loss in our studyon AIDS related subclinical peripheral neu-ropathy,7 the relatively stereotyped clinicalpresentation of PPN, and the findings ofassociated axonal atrophy and CMV infectionin PPN would all point to PPN as a distinctentity rather than as part of a continuum andraise the question as to whether it should rep-resent a subdivision in DSPN.24The non-painful distal symmetrical periph-

eral neuropathies reported here seemed to bemore heterogeneous than PPN and twopatients had pes cavus, suggesting an inherit-ed neuropathy despite the lack of a family his-tory. In addition, this group had asignificantly lower body mass index, suggest-ing a possible role for nutritional factors.

379 on S

eptember 7, 2020 by guest. P

rotected by copyright.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.56.4.372 on 1 A

pril 1993. Dow

nloaded from

Fuller, _tacobs, Guiloff

These patients also had an increased inci-dence of Mycobacteriuni avium intracellulareinfection, suggesting a possible role for thisagent, which has been found in peripheralnerve.23 Vitamin B, was found to be abnor-mal in 31% of these patients compared with24% of the neurologically asymptomatic con-trols, but this difference did not reach signifi-cance. These associations must be treatedwith caution in the context of multiple statis-tical comparisons made between the twogroups.

Chronic inflammatory demyelinatingpolyradiculopathy was reported in only twopatients with AIDS in any series (table 5).Eighteen cases have been reported in patientswith AIDS related complex or asymptomaticHIV patients, nearly half from one centre.26 Afurther 10 were reported from another centre,but the disease stage was not specified.'4 Wesaw no cases in over a thousand patients dur-ing a 15 month period. No cases were report-ed in a point prevalence study of HIV positivemen.' Chronic inflammatory demyelatingpolyradiculopathy during the AIDS, AIDSrelated complex, and HIV seropositivitystages must therefore be recognised as rare.The Guillain-Barre syndrome is rare; only

four patients were reported in non-selectivepopulation studies (table 5). There are casereports of acute inflammatory demyelinatingpolyradiculopathy (AIDP) after seroconver-sion for HIV27 28 as for other types of infec-tion.2' In a selective series from Zimbabwe,59% of 29 consecutive patients with AIDPwere found to be HIV positive,30 but serocon-version and asymptomatic HIV positivitywere not distinguished. The expected rate ofco-occurrence of AIDP in asymptomatic HIVseropositivity by chance is 5-30 cases a year(1-2/100 000 population per year)3' in anestimated population of 0 5-1-5 million.32 Adefinite association between AIDP andasymptomatic HIV seropositivity rather thanseroconversion must wait for definitive confir-mation. The chronic condition is much rarerthan the acute form but accurate incidencefigures are not available; it may therefore beassumed that the association with HIVseropositivity is real.

Mononeuritis multiplex was also a rarepattern of peripheral nerve involvement inour study, secondary to lymphoma duringAIDS and vasculitis during AIDS relatedcomplex. Infiltration of the peripheral nerveby lymphoma is well recognised in HIV nega-tive lymphomas and in one retrospectivepathological series was found in 40% of 145cases of lymphoma.33 It has been reportedsporadically in AIDS.34The most frequent isolated mononeuropa-

thy was that involving the lateral cutaneousnerve of the thigh. This syndrome is commonin men in their 30s and 40s, and cases areoften identified only on direct questioning.33This suggests that this presentation may havebeen a reflection of heightened awareness ofphysical symptoms in patients with AIDSrelated complex and AIDS. The other mono-neuropathies seen affected nerves that are

often involved at entrapment sites, particular-ly in patients who have lost weight.36

Distal symmetrical peripheral neuropathiesaccount for 90% of neuropathies seen inAIDS. The aetiopathogenesis of these syn-dromes is currently poorly understood. Theform of painful peripheral neuropathy seen inthe AIDS patients described here seems to bea distinct clinical entity and may have a spe-cific aetiopathogenesis. Further detailedpathological studies of these patients areneeded.

We thank Drs B G Gazzard, A Lawrence, and D Hawkins forallowing us to study patients under their care, Dr J NHarcourt-Webster for access to postmortem material, andMiss J Roberts for secretarial assistance. GNF was an MRCresearch training fellow; this work formed part of an MD the-sis submitted to the University of London. We received finan-cial support from the Special Trustees of Westminster andRoehampton Hospitals.

1 McArthur JC, Cohen BA, Selnes OA, Kumar AJ, CooperK, McArthur JH, et al. Low prevalence of neurologicaland neuropsychological abnormalities in otherwisehealthy HIV-1-infected individuals: results from themulticenter AIDS cohort study. Ann Neurol1989;26:601-1 1.

2 So YT, Holtzman DM, Abrams DI, Olney RK. Peripheralneuropathy associated with acquired immunodeficiencysyndrome. Prevalance and clinical features from a popu-lation-based survey. Arch Neurol 1988;45:945-8.

3 Fuller GN. The peripheral nervous system in HIV infec-tion: a clinical, electrophysiological and pathologicalstudy [dissertation]. London: University of London,1990.

4 Centers for Disease Control. Revision of the CDC surveil-lance case definition for acquired immunodeficiencysyndrome. MMWR 1987;36:1-15S.

5 Jacobs JM, Love S. Qualitative and quantitative morphol-ogy of human sural nerve at different ages. Brain1985;108:897-924.

6 Fuller GN, Jacobs JM, Guiloff RJ. Axonal atrophy in thepainful peripheral neuropathy in AIDS. ActaNeuropathologica 1990;81: 198-203.

7 Fuller GN, Jacobs JM, Guiloff RJ. Subclinical peripheralnerve involvement in AIDS: an electrophysiological andpathological study. . Neurol Neurosurg Psychiatry1991;54:318-24.

8 Fuller GN, Jacobs JM, Guiloff RJ. Association of painfulperipheral neuropathy in AIDS with cytomegalovirusinfection. Lancet 1989;ii:937-41.

9 Fuller GN, Gill SK, Guiloff RJ, Kapoor R, Lucas SB,Sinclair E, et al. Ganciclovir for lumbosacral polyradicu-lopathy in AIDS. Latncet 1990;335:48-9.

10 Snider WD, Simpson DM, Nielsen S, Gold JW, MetrokaCE, Posner JB. Neurological complications of theacquired immune deficiency syndrome: analysis of 50patients. Anal Neurol 1983;14:403-18.

11 Levy RM, Bredesen DE, Rosenblum ML. Neurologicalmanifestations of the acquired immunodeficiency syn-drome (AIDS): experience at UCSF and review of theliterature. J Neurosurg 1985;62:475-95.

12 Cornblath DR, McArthur JC. Predominantly sensoryneuropathy in patients with AIDS and AIDS-relatedcomplex. Neurology 1988;38:794-6.

13 Bailey RO, Baltch AL, Venkatesh R, Singh JK, BishopMB. Sensory motor neuropathy associated with AIDS.Neurology 1988;38:886-91.

14 Miller RG, Parry GJ, Pfaeffl W, Lang W, Lippert R,Kiprov D. The spectrum of peripheral neuropathy asso-ciated with ARC and AIDS. Muscle anid Nerve1988;1 1:857-63.

15 Lange DJ, Britton CB, Younger DS, Hays AP. The neu-romuscular manifestations of human immunodeficiencyvirus infections. Arch Neurol 1988;45: 1084-8.

16 Miller RG, Storey JR, Greco CM. Ganciclovir in thetreatment of progressive AIDS-related polyradiculopa-thy. Neurology 1990;40:569-74.

17 Said G, Lacroix C, Chemouilli P, Goulon-Goeau C,Roullet E, Penaud D, et al. Cytomegalovirus neuropathyin acquired immunodeficiency syndrome: a clinical andpathological study. Ainn Neurol 1991 ;29: 139-46.

18 Robert ME, Geraghty JJ, Miles SA, Cornford ME, VintersHV. Severe neuropathy in a patient with acquiredimmune deficiency syndrome (AIDS). Evidence forwidespread cytomegalovirus infection of peripheralnerve and human immunodeficiency virus-likeimmunoreactivity of anterior horn cells. ActaNeuropathologica 1989;79:255-61.

19 Budzilovich GN, Avitabile A, Niedt G, Aleksic SN,Rosenblum MK. Polyradiculopathy and sensory gan-glionitis due to cytomegalovirus in acquired immunedeficiency syndrome (AIDS). Progress in Aids Pathology1989;1: 143-57.

20 Berger JR, Sheremata WA, Resnick L, Atherton S,

380 on S

eptember 7, 2020 by guest. P

rotected by copyright.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.56.4.372 on 1 A

pril 1993. Dow

nloaded from

Natuire and inicidence ofperipheral tnerve synidronmes in HIV inifection

Fletcher MA, Norenberg M. Multiple sclerosis-like ill-ness occurring with human immunodeficiency virusinfection. Neurology 1989;39: 324-9.

21 Gastaut JL, Gastaut JA, Pellissier JF, Tapko JB, Weill 0.Peripheral neuropathies in human immunodeficiencyvirus infection. A prospective study of 56 patients. RevNeuirol 1989;145:451-9.

22 Chaunu MP, Ratinahirana H, Raphael M, Henin D,Leport C, Brun-Vezinet F, et al. The spectrum ofchanges on 20 sural nerve biopsies in patients with HIV.Muscle anzd Nerve 1989;12:452-9.

23 Leger JM, Bouche P, Bolgert F, Chaunu MP, RosenheimM, Cathala HP, et al. The spectrum of polyneuropathiesin patients infected with HIV. Jf Neurol NeurosurgPsychiatry 1989;52: 1369-74.

24 Working Group of the American Academy of NeurologyAIDS Task Force. Nomenclature and research case def-initions for neurologic manifestations of human immun-odeficiency virus-type 1 (HIV- 1) infection. Neurology1991 ;41 :778-85.

25 Wrzolek MA, Rao C, Kozlowski PB, Sher JH. Muscle andnerve involvement in AIDS patient with disseminatedmycobacterium avivum intracellulare infection. Muscleanid Nerve 1989;12:247-9.

26 Comblath DR, McArthur JC, Kennedy PG, Witte AS,Griffin JW. Inflammatory demyelinating peripheral neu-ropathies associated with human T-cell lymphotropicvirus type III infection. Anzn Neurol 1987;21:32-40.

27 Piette AM, Tusseau F, Vignon D, Chapman A, Parrot G,Liebowitch J, et al. Acute neuropathy coincident withseroconversion for anti-LAV/HTLV-III. Lancet1986;i:852.

28 Hagberg L, Malmvall BE, Svennerholm L, Alestig K,

Norkrans G. Guillain-Barre syndrome as an early mani-festation of HIV central nervous system infection. Scand_7Infect Dis 1986;18:591-2.

29 Winer JB, Hughes RAC, Anderson MJ, Jones DM,Kangro H, Watkins RPF. A prospective study of acuteidiopathic neuropathy. II. Antecedent events. Jf NeurolNeurosurg Psychiatry 1988;51:613-8.

30 Thornton CA, Latif AS, Emmanuel JC. Guillain-Barresyndrome associated with human immunodeficiencyvirus infection in Zimbabwe. Neurology 1991 ;41 :812-5.

31 Arnason BGW. Acute inflammatory demyelinatingpolyradiculoneuropathies. In: Dyck PJ, Thomas PK,Lambert EH, Bunge R, eds. Peripheral neuropathy.Philadelphia: WB Saunders, 1984:2050-100.

32 Moss AR, Becchetti P. Natural history of HIV infection.AIDS 1989;3:55-61.

33 Jellinger K, Radaszkiewicz T. Involvement of the centralnervous system in malignant lymphomas. VirchozvsArchiv 1976;370:345-62.

34 Gold JE, Jimenez E, Zalusky R. Human immuno-deficiency virus-related lymphoreticular malignanciesand peripheral neurologic disease. A report of fourcases. Cancer 1988;61:2318-24.

35 Stephens H. Meralgia paraesthetica. Arch NeuirolPsychiatry 1957;77:557-74.

36 Sotaniemi KA. Slimmer's paralysis-peroneal neuropathyduring weight reduction. Jf Neurol Neurosurg Psychiatry1984;47:564-6.

37 Guiloff RJ, Fuller GN, Roberts A, Hargreaves M, GazzardB, Scaravilli F, et al. Nature, incidence and prognosis ofneurological involvement in the acquired immunodefi-ciency syndrome in central London. Postgrad Med Jf1988;64:919-25.

381 on S

eptember 7, 2020 by guest. P

rotected by copyright.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.56.4.372 on 1 A

pril 1993. Dow

nloaded from