127

THE INVESTIGATION ANDTREATMENT OF VERTIGO

By D. RANGER, F.R.C.S.Assistant Ear, Nose and Throat Surgeon, Middlesex Hospital

The investigation of a patient with vertigoshould start with a careful and complete history,for in most cases this plays a very important part inmaking an accurate diagnosis and often it is quiteimpossible to interpret the findings on examinationwithout a knowledge of the symptomatology.

Terminology on this subject varies a lot evenamong medical men and one can never be surewhat patients mean by terms such as vertigo,giddiness and dizziness except by asking them totry and describe in simple words the exact sensa-tions they experience. Even then it is oftennecessary to ask leading questions such as whetherthe apparent movement is rotatory or swaying orsome other type.The severity of the vertigo is no indication of

the severity of the underlying lesion and in manycases the most serious cases have only minorsymptoms. There does not appear to be anysignificant difference between the apparent rotationof objects around the patient (objective vertigo)and the apparent rotation of the patient in relationto his surroundings (subjective vertigo). Likewisethere does not seem to be any practical value indiscovering whether rotation is clockwise oranti-clockwise.

It should be ascertained whether the vertigooccurs in bouts with periods of complete freedomin between episodes or whether there is a more orless constant sensation of unsteadiness. In general,bouts of vertigo occurring at intervals in patientsotherwise normal suggest a peripheral lesion in thelabyrinth whereas constant unsteadiness is morelikely to occur in central lesions.

It should be determined whether the attacksoccur in any particular position. Although changeof position may make any vertigo worse, certaintypes of vertigo occur only when the patientassumes a particular posture, such as lying in bed,and these cases of positional vertigo are discussedmore fully later.

In addition the duration of the attacks should bedetermined. A severe bout of vertigo lastingseveral days which keeps the patient in bed for aweek or so and keeps him off work for a few weeksis very suggestive of a sudden complete vestibular

failure such as occurs, without any loss of hearing,in vestibular neuronitis. In Meniere's disease theattacks often vary enormously in severity andduration; although they may be momentary mostattacks last for many minutes or for hours or evenfor one or two days.Enquiry about associated symptoms such as

nausea, vomiting and diarrhoea may give someindication of the severity of the vertigo and of thepatient's reaction to the vertigo. Diarrhoea is arare accompaniment and tends to occur only in theemotionally unstable. In patients who havemany attacks of vertigo, associated nausea andvomiting often diminish or disappear in laterattacks.

Definite loss of consciousness can occur inassociation with vertigo from a peripheral labyrin-thine lesion, but this symptom must always raisethe possibility of epilepsy. One difference whichmay be helpful is that a patient with a labyrinthinelesion will usually still have vertigo on regainingconsciousness whereas this is seldom. the case inepilepsy. In addition an epileptic will usuallyhave lost consciousness on other occasions withouthaving any sensation of vertigo, whereas a patientwith a labyrinthine lesion will probably have hadother attacks of vertigo without loss of con-sciousness.As most cases of vertigo are due to a cause in

the ear it is important to ask about other symptomssuch as aural discharge, deafness and tinnitus.The presence of any aural discharge at the timeof the vertigo immediately suggests the possibilityof a mastoiditis with a peri-labyrinthitis and,possibly, an erosion of the bony labyrinth. Auraldischarge at some time in the past with no dis-charge at the time of the vertigo is unlikely to beconnected with the vertigo. Patients can be verymisleading about the question of discharge, foroccasionally they regard wax as discharge whileothers strongly deny the presence of discharge,even when it is quite obvious on inspection.While enquiring about the presence of aural

discharge, enquiry should also be made about thepresence of any spots in and around the ear.Although a herpes of the vestibular ganglion is

B2

copyright. on 3 M

ay 2018 by guest. Protected by

http://pmj.bm

j.com/

Postgrad M

ed J: first published as 10.1136/pgmj.32.365.127 on 1 M

arch 1956. Dow

nloaded from

128 POSTGRADUATE MEDICAL JOURNAL March I956usually associated with a geniculate herpes and anobvious facial paralysis, some cases of herpes occurwith vertigo (and deafness) and with an herpeticeruption in the ear but without any facial nerveinvolvement.

If there is any deafness enquiry should be madeabout its time of onset in relation to the vertigoand also about any fluctuation in the degree ofdeafness, particularly in relation to the attacks ofvertigo. It is also important to ask about thepresence of any distortion of hearing or anysounds which jar on the ear. In Meniere's diseasethe distortion of hearing may sometimes be somarked as to be the predominant feature and ahistory of distortion and of marked fluctuation ofthe deafness, whether during the attacks ofvertigo or not, can be regarded as almost diagnosticof Meniere's disease.

It is important to enquire about tinnitus.Cawthorne (I953) has pointed out that althoughtinnitus may be a very troublesome symptom tothe patient, it may be a great help to the diagnostic-ian by giving an indication of the side of the lesionin difficult cases.Other illnesses and also drugs may well have

some bearing on the problem and the historyshould include details of any other illnesses oraccidents. Particular attention should be paid tothe question of any head injury and to the possi-bility of the patient having had streptomycin orlarge doses of quinine. It should be rememberedthat even quite small doses of streptomycin cancause vertigo, especially in the presence of a renallesion. Large doses of quinine are sometimestaken in an attempt to produce a miscarriage, and ahistory may be difficult to obtain. In one caseseen personally, permanent deafness and severevertigo followed the taking of enormous doses ofquinine for this purpose, without having anyeffect on the pregnancy.

Other symptoms may suggest some neurologicallesion. An episode of mistiness of vision ordiplopia, possibly occurring a long time before theonset of vertigo, immediately raises the possibilityof multiple sclerosis while numbness in the facesuggests an acoustic neuroma.Headache may occur in patients with Meniere's

disease but persistent headache must always arousethe suspicion of an intracranial neoplasm.Clinical Examination

Ideally every patient with vertigo should have acomprehensive clinical examination which can beconsidered most conveniently under the headingsof general, neurological and otological:General

In most cases of vertigo a ' general' examina-

tion will not reveal any significant abnormality butobviously some findings can be of the utmostimportance. The discovery of a malignantneoplasm anywhere will immediately raise thequestion of cranial metastases for secondaries oftenproduce a very complicated and puzzling picturefrom the neurological and otological point of view.Again the findings of arteriosclerosis or hyper-tension may help a lot in the diagnosis. Anyevidence to suggest chronic alcoholism or syphiliswill be of value and the finding of neuro-fibromataanywhere will immediately raise the possibility ofan auditory neuro-fibroma. Any pallor indicativeof anaemia should be noted and the patient shouldbe assessed for the possibility that the symptomsare neurotic.

NeurologicalVertigo may occur in association with neuro-

logical lesions in such diverse areas as the temporallobes, the brain stem and the cerebellum, soobviously a complete neurological examination iscalled for.

Particular attention should be paid to thequestion of nystagmus. In purely peripherallabyrinthine lesions nystagmus persists for only afew weeks even after complete loss of vestibularfunction and the persistence of nystagmus forlonger than this must indicate a more centrallyplaced lesion. A vertical nystagmus indicates acentral lesion.

Acoustic neuromata usually have otologicalsymptoms and signs for some time, often years,before other manifestations occur and it isobviously desirable to arrive at a diagnosis at thatstage if possible. However, later the fifth andseventh nerves may be involved and must beexamined carefully. Impairment of the cornealreflex may be a sign of such involvement. Minimalfacial weakness may often be revealed on rapidblinking of the eyes.Examination of the ocular fundus must obviously

form part of the examination and in many casescharting of the visual fields will be required.Romberg's test should be carried out and the gaitnoted.

OtologicalThe first essential preliminary in all otological

examinations is a careful inspection of the externalauditory meatus and tympanic membranes. Allcerumen must be removed to allow a completeview of the tympanic membranes and to preventany interference with subsequent tests.Note should be made as to whether the tympanic

membranes are intact or whether there is anyperforation and, if so, whether the perforation is

copyright. on 3 M

ay 2018 by guest. Protected by

http://pmj.bm

j.com/

Postgrad M

ed J: first published as 10.1136/pgmj.32.365.127 on 1 M

arch 1956. Dow

nloaded from

March 1956 RANGER: The Investigation and Treatment of Vertigo 129

dry or discharging. Particular attention should bepaid to any perforation in the attic region or atthe posterior margin of the membrane as these arelikely to be associated with a cholesteatoma withpossible erosion of the lateral semicircular canal orwith a temporal lobe or cerebellar abscess. If aperforation is present the tragus should be pressedsharply into the meatus with a finger so as to causesudden compression of the air in the meatus andmiddle ear, and note made as to whether anynystagmus is produced. If nystagmus does occurunder these circumstances the ' fistula test'is said to be positive and it implies erosionof bone surrounding the inner ear. The erosionis nearly always into the lateral semicircular canalwhich is most exposed to pressure from a chole-steatoma. A negative fistula sign does notnecessarily exclude the possibility of a fistula.While examining the ear careful inspection

should be made of the meatus and pinna andadjacent areas of skin to see if there are anyherpetic vesicles or residual scars from previousvesicles.

HearingIt is obviously essential to examine the hearing

carefully in patients with veitigo and it is oftenmost convenient to do this before proceeding totests of vestibular function. Tuning fork testsgive much valuable information and for routinetesting it is best to use a fork of 512 cycles fittedwith a base. A sounded tuning fork placed withthe base on the midline of the head is heardcentrally, or equally in both ears, by persons withnormal hearing, but tends to be lateralized to theworse hearing ear in patients with a conductive(middle ear) deafness and to the side of the betterhearing ear in patients with a perceptive (cochlearand nerve) deafness. However, in many patients,especially when the deafness has been present fora long time, no localization occurs and the tuningfork appears to be heard in the centre of the headeven though there is a marked difference in hearingbetween the two ears.

Normally a tuning fork is heard better when thelimbs of a vibrating fork are placed opposite theexternal auditory meatus (air conduction) thanwhen the base is placed on the mastoid process(bone conduction). The difference is quite markedand tuning forks of most patterns can be heard byair cbnduction for a matter of 15 to 20 secondsafter they have ceased to be heard by bone con-duction. In cases with deafness which is entirelyperceptive in type the same relationship is found.However, when there is some interference withthe sound conducting mechanism as, for instance,in middle ear disease and in otosclerosis, then the

relationship is altered and in most of these casesbone conduction is better than air conduction.However, in some slight cases of conductive deaf-ness conduction by air and bone may be equal orthere may merely be a reduction of the normal15 to 20 second difference between the two.

In nearly all cases in which the deafness is dueto a lesion in the cochlea or in the auditory nerve,there will be found the typical signs of a perceptivedeafness as outlined above. Obviously, however,these signs may be very much altered if there isalso a deafness due to some other cause such as anold otitis media. In addition, confusion sometimesarises in cases with moderate or severe unilateralperceptive deafness because of the ready trans-mission of sound from one mastoid process acrossthe skull to the cochlea of the opposite side whenbone conduction is being tested. In this way boneconduction may appear to be better than air con-duction but the true state of affairs can be demon-strated if the test is repeated with the better ear'masked' by another sound.

Sometimes it is found that with a neurofibromaof the auditory nerve, true bone conduction isbetter than air conduction and Dix and Hallpike(1950) have explained this phenomenon as beingdue to an exudate in the inner ear interfering withthe conduction of sounds through the fluids of thecochlea to the organ of Corti.

Hearing acuity may be assessed to some extentby whisper and voice tests, but a more accurateassessment by audiometry is desirable in order todetect slight degrees of deafness and to allowaccurate comparison between the hearing ondifferent occasions. Audiometry can be used forassessing the hearing of pure tones or of speechand often it is desirable to carry out both tests.It is important when testing the worse hearingear that the better hearing ear should be adequatelymasked to prevent any confusion caused by theunappreciated conduction of sound to the betterear. In addition to assessing the level of hearing,a speech audiogram may have diagnostic value indeciding on the cause of the deafness. In mostcases of deafness the percentage of words orsyllables understood increases progressively withgreater amplification, but in some cases increasingamplification beyond an optimum may actuallyreduce the percentage recognition and this isespecially liable to occur in Meniere's disease.

In cases with a perceptive deafness it is valuableto carry out tests for loudness recruitment. Thisphenomenon is one in which the deafness forthreshold sounds is greater than the deafness forlouder sounds and, in fact, when full recruitment ispresent there is no deafness for loud sounds evenwhen there is marked deafness for soft sounds. If

copyright. on 3 M

ay 2018 by guest. Protected by

http://pmj.bm

j.com/

Postgrad M

ed J: first published as 10.1136/pgmj.32.365.127 on 1 M

arch 1956. Dow

nloaded from

130 POSTGRADUATE MEDICAL JOURNAL March I956

one ear is normal and there is a moderate differencebetween the two ears at threshold the test can becarried out by a loudness balance test in whichnotes of different intensities are relayed altern-ately into each ear and note made of the intensitieswhich seem equally loud in the two ears. Inpatients without any recruitment the differencebetween the two ears will remain the same at allintensities, whereas when recruitment is presentthe difference at threshold levels will be reducedas the intensities are increased. If there is anyreduction of this nature there is said to be ' partialrecruitment,' whereas if a stage is reached inwhich the same sound seems equally loud in thetwo ears 'full recruitment' is present. In somecases the process is carried even further and' overrecruitment ' occurs.When the loudness recruitment phenomenon

was first described by Fowler (1936) it was re-garded as a characteristic of a perceptive (nerve)deafness. Later, Dix, Hallpike and Hood (1948)applied the test to different types of perceptivedeafness and found that full recruitment waspresent in 30 cases of unilateral deafness due toMeniere's disease, whereas in 20 cases of degenera-tion of the auditory nerve due to neurofibromata orto other tumours of the cerebello-pontine angle,recruitment was absent in 14 cases and partiallypresent in the remaining six. In a later publicationby the same authors (i949) these findings werefurther substantiated, but they recorded one caseof an acoustic neurofibroma in which full recruit-ment was present. To sum up the diagnosticvalue of the loudness balance test it may be saidthat it is applicable only in cases of unilateraldeafness when there is a moderate difference ofhearing between the two ears, and although thepresence of full recruitment is very much in favourof a diagnosis of Meniere's disease, this phe-nomenon can occur in other disorders and doesnot provide an infallible distinction betweenMeniere's disease and an acoustic neuroma.On theoretical grounds it would seem possible

to devise a test for the recruitment phenomenonbased on one ear only and thus provide a testapplicable to cases with a bilateral deafness.However, although several tests have beendescribed none of them at present seems to becompletely satisfactory.

Vestibular FunctionThe question of spontaneous nystagmus has

already been discussed but in addition it is im-portant to test patients for the presence of positionalnystagmus, especially in patients who havemomentary bouts of vertigo and if the historysuggests that the vertigo is related to posture. The

patient should sit on a couch with the headturned to one side and the eyes fixed on some pointsuch as the bridge ofthe observer's nose. Then withthe head supported by the examiner's hands thepatient should lie down. If any nystagmus de-velops note should be made of its time of onset,its duration and its direction. If there is nonystagmus within a minute the patient should situp and then repeat the manoeuvre with the headturned to the opposite side. Finally, the patientshould lie down with the head hanging straightback over the end of the couch.

Although some intermediate reactions occur,in most cases any positional nystagmus whichoccurs falls into one of two main types. In onetype, by far the commoner, after the horizontalposition has been assumed there is a latent periodof some seconds followed by vertigo and nystag-mus for about 5 to 5 seconds and then no furthernystagmus. On sitting up there may be a littlevertigo and possibly slight nystagmus in theopposite direction. On lying down again theremay be no nystagmus or vertigo at all or else it isless than on the first occasion and then none onthe third or fourth occasion. The phenomenonof diminishing reaction with repetition is usuallyreferred to as fatiguability and this type ofpositional vertigo is almost certainly due to a peri-pheral lesion in the labyrinth and tends to run abenign paroxysmal course. Some cases areinitiated by a head injury but most occur for noapparent reason. In the second type of positionalvertigo, the nystagmus occurs immediately thehorizontal or some intermediate position isassumed and continues so long as that position ismaintained. On sitting up the nystagmus stopsand on lying down a second (or third or fourth,etc.) time the nystagmus occurs again with un-diminished vigour. This type is nearly alwaysassociated with a lesion in the posterior fossa,commonly a space occupying lesion, but it canoccur in the course of a demyelinating process.

In addition, the nystagmus induced in responseto a rotating object should be observed-opto-kinetic or optomotor nystagmus. In generalterms peripheral labyrinthine lesions do not pro-duce any interference with the normal optokineticnystagmus although in some patients sudden com-plete vestibular failure, such as occurs following alabyrinthectomy, produces some impairment ofoptokinetic nystagmus to the same side andaccentuation to the opposite side. However, apartfrom this, any abnormality of optokinetic nystag-mus is an indication of the site of the lesion inthe brain stem or cerebral cortex.

Provided the tympanic membranes and externalauditory meati are intact the caloric test pro-

copyright. on 3 M

ay 2018 by guest. Protected by

http://pmj.bm

j.com/

Postgrad M

ed J: first published as 10.1136/pgmj.32.365.127 on 1 M

arch 1956. Dow

nloaded from

March 1956 RANGER: The Investigation and Treatment of Vertigo 131

vides a simple and effective method of assessingvestibular function. It is best carried out by thetechnique described by Fitzgerald and Hallpike(1943). The lateral semicircular canal is the mostreadily stimulated by thermal changes andaccordingly is the one chosen for routine testing.In order to place it in a vertical position, andtherefore in the optimum position for the settingup of convection currents, the head should beflexed 30° from the horizontal. The ears are thenirrigated with water at 3o°C. for 40 seconds andafterwards at 44°C. for the same length of time.Nystagmus is timed from the moment of startingthe syringing until the nystagmus stops. If nonystagmus is elicited by this procedure the test isrepeated using water at 20°C. for a period of oneminute and the nystagmus timed in the same way.If no nystagmus is produced by this stimulus itcan be assumed that there can be very little, ifany, remaining function.A normal result of this caloric test is shown by

responses of equal length, to within five or tenseconds, from both ears at 3o°C. and 44°C.Provided the responses are all equal, reactions ofone-and-a-half to three minutes are within normallimits but inequality in the responses is a sign ofabnormality even if all the reactions lie withinthese periods of time. Reactions longer than threeminutes are suggestive of a central cortical lesionand values of less than one-and-a-half minutesindicate impaired function. The caloric test alonecannot differentiate between impaired functiondue to a lesion in the labyrinth and a lesion of thevestibular nerve or nucleus.A galvanic test of vestibular function and also

rotation tests may help in the investigation ofspecial cases but are not used as routine tests andcannot be considered in detail here. They havethe advantage that they can be used whatever thestate of the tympanic membranes but have thedisadvantage that they are more difficult to carryout and to interpret the results obtained.

Special InvestigationsA blood Wasserman reaction or Kahn test or

both are obvious investigations of importance,especially in bilateral cases and particularly if therehas been a rapid and progressive impairment ofcochlear and vestibular function. A haemoglobinestimation should be made if anaemia is suspectedand a leucocyte count and sedimentation rate inpossible cases of infection.

Plain X-rays of the skull may help in thediagnosis and if there is any possibility of anacoustic neuroma, views should be taken of theinternal auditory meatuses. These may sometimesbe shown up best by tomography (Denny, 1955).

An electro-encephalogram may also give valu-able information, especially in cases of epilepsywith a vertiginous aura.A lumbar puncture is indicated in any central

lesion of doubtful etiology and air encephalo-graphy may be necessary if there is any possibilityof a space occupying lesion.

TreatmentThis paper can deal only briefly with the

question of treatment and for that reason it isproposed to limit the discussion to two conditions-Meniere's disease and benign paroxysmalpositional vertigo. Even so it is possible to giveonly brief and rather dogmatic views.

Meniere's Disease (Endolymphatic Hydrops)This is such a variable condition that it is

difficult to assess accurately the value of some ofthemethods of treatment. It would be ideal, of course,to have a therapeutic measure which would restorethe labyrinth to normal again and this is the aimof most forms of treatment. Unfortunately thewide variety of methods employed demonstratesonly too clearly that there is no treatment kiownat present which can be relied on to achieve thisand as a result it is sometimes necessary to resortto a destructive procedure which will cut off allimpulses from the affected labyrinth and whichcan be relied on to stop the attacks of vertigo inthat way.

In the present state of our knowledge it wouldseem that in most cases it is worth trying the effectof a restricted fluid and sodium intake andpossibly also a diuretic. In addition there is someevidence that vasodilation is of value andnicotinic acid has been used for this purpose formany years. More recently other oral vaso-dilators have become available and they also havetheir advocates. Injection of histamine, eitherintramuscularly or intravenously, can be employedand procaine can also be used intravenously. Inview of this use of histamine it may seem surprisingthat there are also many advocates for the use ofthe antihistamine group of drugs. In addition,vasodilation can be produced by injection ofthe stellate ganglion or by cervical sympathectomyand this may help some patients (Lewis, I954).Whichever of the above treatments is adopted

many of the patients are also helped by sedationwith a barbiturate and a combination of pheno-barbitone and theobromine often seems useful.Also, it is comforting to the patient to be reassuredthat the cause of his trouble lies in the ear andnot in the brain.Although the above remarks may suggest 'that

the methods of treatment already mentioned are

copyright. on 3 M

ay 2018 by guest. Protected by

http://pmj.bm

j.com/

Postgrad M

ed J: first published as 10.1136/pgmj.32.365.127 on 1 M

arch 1956. Dow

nloaded from

132 POSTGRADUATE MEDICAL JOURNAL March 1956

of little value it must be pointed out that mostpatients with Meniere's disease are able to con-tinue living a reasonably normal life when treatedon the above lines. If the attacks of vertigo docontinue to interfere with the patient's life and ifthe condition is unilateral the attacks can bestopped by a destructive procedure as mentionedabove. Impulses can be cut off from the affectedlabyrinth either by destruction of the labyrinthitself or by division of the vestibular portion of theauditory nerve. Destruction of the labyrinth alsoinvolves complete loss of hearing on that side andfor that reason division of the vestibular portion ofthe nerve might seem more desirable. However,this latter operation is a much more formidableprocedure with a greater morbidity and a definitemortality and in most cases there is no value to begained in preserving the distorted remnant ofhearing which remains. As a result division of thevestibular portion of the nerve is restricted tothose cases which have troublesome attacks ofvertigo but who still have good hearing.Benign Paroxysmal Positional Vertigo

For this condition there is no specific cure butfortunately most cases settle down spontaneously.In the period when the condition is active the

patients are helped most by appreciation of thecircumstances under which an attack develops andby realization that once an attack has occurredthere is usually a period of several hours duringwhich the offending position can be adopted withimpunity. Some patients find it possible to avoidany posture which provokes the vertigo but in theothers it is often best deliberately to produce anattack at a convenient moment and then enjoy aperiod of freedom. In this way patients with thiscomplaint can often carry on with an occupationwhich involves climbing ladders, etc.

In addition, many of these patients are con-vinced that they are helped by those antihis-taminics which seem most active in motionsickness.

BIBLIOGRAPHYCAWTHORNE, T. E. (I953), Proc. Roy Soc. Med., 46, 833.DENNY, W. R. (I955), J. Laryng., 69.DIX, M. R., and HALLPIKE, C. S. (1950), Proc. Roy. Soc. Med.,

43, 291.DIX, M. R., HALLPIKE, C. S., and HOOD, J. D. (1948), Ibid.,

41, 5i6.DIX, M. R., HALLPIKE, C. S., and HOOD, J. D. (x949), Ibid.,

42, 527.FITZGERALD, G., and HALLPIKE, C. S. (1942), Brain, 65,115.FOWLER, E. P. (1936), Arch. Otolaryng., Chicago, 24, 731.LEWIS, ROLAND S. (I954), J. Laryng., 68, 636.

LLOYD-LUKEBooks that enshrine profound thought

OPHTHALMOLOGY STUDIES ON THE CEREBRAL PRACTICALA Textbook for Diploma Students CORTEX OBSTETRIC PROBLEMS

by P. D. TREVOR-ROPER by S. RAMON y CAJAL by IAN DONALDM.B., B.Chir.(Camb.), F.R.C.S., (Translation from the Spanish by M.B.E., M.D.(Lond.), F.R.C.O.G.

D.O.M.S.(Eng.) Lisbeth M. Kraft, Yaje University)(1955) 75s. net ( 1955) 24s. 6net

LECTURES ON . FLUID BALANCEGENERAL PATHOLOGY IN SURGICAL PRACTICE

Edited by SIR HOWARD FLOREY by L. P. LE QUESNEM.D., F.R.C.P., F.R.S. D.M.(Oxon.), F.R.C.S.(Eng.)

(1954) 63s. net (1954) 17s. 6d. net

Descriptive catalogue available on request

LLOYD-LUKE (MEDICAL BOOKS) LTD., 49 Newman Street, W.I

copyright. on 3 M

ay 2018 by guest. Protected by

http://pmj.bm

j.com/

Postgrad M

ed J: first published as 10.1136/pgmj.32.365.127 on 1 M

arch 1956. Dow

nloaded from