theafferent pupillary defect acute optic neuritis · theafferent pupillary defectin...

Journal ofNeurology, Neurosurgery, and Psychiatry, 1979, 42, 1008-1017

The afferent pupillary defect in acute optic neuritisC. J. K. ELLIS

From the Department of Medical Ophthalmology, St Thomas' Hospital, London

SUMMARY Twenty-two patients with acute optic neuritis were studied by the techniques ofinfrared pupillometry and visual evoked responses (VER) to pattern reversal. A relative afferentpupillary defect was found in all cases and the magnitude of this defect was found to be relatedto the amplitude, but not to the latency, of the VER. During follow-up the afferent defect wasfound to remain persistently abnormal while other methods of clinical evaluation could notdemonstrate abnormality reliably. The amplitude of the VER also remained low.

Pupillary reactions have been recorded for cen-turies and have been used as an indication ofpathology in the anterior visual pathways. Earlyinterpretations of pupillary abnormalities wereinaccurate because of misunderstanding of basicvisual physiology. However, with increasing aware-ness of the true nature of vision, pupillary reac-tions were widely used as a guide to visualprognosis after cataract couching.

Saint-Yves (1742) described in detail a patientblind in one eye in whom the pupil dilated onclosing the other eye and contracted when it wasopened again. He also related the degree of im-pairment of vision to the reduction in amplitudeof the pupil constriction saying "if the iris has onequarter of its movement, we judge that quarter ofthe sight remains." Similar findings indicating thatpupillary reactions show characteristic abnormal-ities in association with conditions that affectvision have been reported by other authors(Gerold, 1846; Hirschberg, 1884). Kestenbaum(1946) in describing the alternate cover test namedthese findings after Marcus Gunn (1904) who hadreported his frequent use of pupillary signs indiagnosis. Thus when Levatin (1959) introducedthe swinging light test into clinical practice he wasnot describing any previously unrecognised pupil-lary abnormality, but rather was facilitating theassessment of pupillary function by comparing theresponse of the two eyes when alternately stimu-lated with light of the same intensity. Levatindescribed two main abnormalities. On the side ofa partial optic nerve lesion the direct light re-action was reduced in amplitude and under con-

Address for reprint requests: Dr C. J. K. Ellis, Department of MedicalOphthalmology, St Thomas' Hospital, London SEI 7EH.Accepted 10 April 1979

tinuing light stimulation there was a more markedpupillary redilatation (or "escape") than in theother eye. Levatin stressed that the presence of anabnormality of this test did not imply that theother eye was normal, but only that the two eyeswere not affected equally. Thus the affected eyeis said to show a relative afferent defect.

It was not until the advent of infrared pupil-lometry (Lowenstein and Friedman, 1942) thatdetailed analysis of pupillary function in lesions ofthe afferent pupillary pathways became possible.This technique, by using infrared illumination,allows pupillary reflexes to be recorded continu-ously and in darkness. Lowenstein (1954) foundthat in optic neuritis the pupillary responses in aneye with an afferent defect were reduced in ampli-tude, poorly sustained, and had a prolongedlatency. He named these responses "low-intensity"reactions because of their similarity to reflexeselicited by low intensity light stimulation in normaleyes. Thompson (1966) described similar abnor-malities in patients with optic nerve compression.Fison et al. (1979) quantitated the relative afferentdefect and showed a close inverse relation betweenvisual acuity and the magnitude of the relativeafferent defect.

Abnormalities of the occipital pattern evokedresponse in acute optic neuritis have been studiedin detail (Halliday et al., 1972). The characteristicprolonged latency appears to be caused, at leastin part, by the presence of demyelination in theoptic nerve (McDonald and Sears, 1970; Mc-Donald, 1977a).There has been no study correlating pupillary

responses with the visual evoked responses in acuteoptic neuritis. In this paper 1 report the results ofsuch a study of 22 patients investigated by infra-

1008

guest. Protected by copyright.

on Decem

ber 27, 2019 byhttp://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.42.11.1008 on 1 N

ovember 1979. D

ownloaded from

http://jnnp.bmj.com/

The afferent pupillary defect in acute optic neuritis

red pupillometry, visual evoked responses (VER)to pattern reversal, and clinical assessment. Thepatients were studied several times during andafter the acute phase of their illness. The pupillaryabnormalities, their relation to the VER, and theiruse as a means of assessment in optic neuritis arediscussed.

Subjects and methods

Nineteen healthy subjects were investigated byinfrared pupillometry. They were aged between18 and 48 years, and 12 were female.

Visual evoked responses were performed on 20healthy subjects between 20 and 35 years of age.Fifteen of these subjects were female. Refractiveerrors were corrected by spectacles.Twenty-two patients with unilateral acute optic

neuritis were studied. The age range of the patientswas 16 to 48 years, and 16 were female. Thepatients were followed up for a mean of fivemonths (range 1-8 months) after the onset ofvisual symptoms. The diagnosis of acute opticneuritis was made on clinical grounds (McDonald,1977b) and supported by the subsequent clinicalcourse. Other causes of optic neuropathy wereexcluded by appropriate investigations in eachcase.During follow-up seven patients showed clinical

and VER evidence of a recurrent attack of opticneuritis. Two patients were lost to follow-up afterthe first examination. Four patients had evidenceof disseminated lesions on presentation.

PUPILLOMETRYA Whittaker series 1800 infrared television pupil-lometer was used in this study. The pupillometeranalyser system measured the maximum verticalpupillary diameter and this was recorded continu-ously on a paper recorder and on magnetic tapefor computer analysis. The characteristics of thetelevision system imposed a limit on resolution oftimed events of 20 ms and on pupillary diameterof approximately 0.03 mm.A 100 ms white light stimulus, focused to a

2.0 mm beam at the midpupillary point was de-livered every eight seconds from an angle 70 31'lateral to visual axis. An electronic shutter gavethe stimulus almost square wave characteristics.Movements of the shutter were also recorded onthe magnetic tape for latency analysis. Six lightintensities were employed by filtering the twoSylvania concentrated arc lamps with neutraldensity filters. The intensity of the two stimuluslamps was measured regularly with a photometerto ensr ie equal stimulus intensity in the two eyes.

All subjects were dark-adapted in red gogglesfor 30 minutes before pupillometry. Recordingswere made in darkness with the patient's headsupported comfortably while fixating a distantobject by means of a mirror system. Between sixand 10 stimuli were delivered first to the affectedeye, at each level of light intensity. A pause be-tween increments of stimulus intensity alloweddark adaptation to be maintained. Care was takento ensure that the non-stimulated eye was kept indarkness.The output from the pupillometer was recorded

on a Yasec CD1 10 FM cassette data taperecorderand replayed into a Varian V72 digital computerwith compensation for variation in tape speed byutilising a reference input channel. The computerwas programmed to average the direct pupillaryresponses at each level of stimulus intensity and togive accurate numerical values for reflex ampli-tude, latency from stimulus to onset of pupillaryconstriction, and maximum rate of pupillary con-striction. Details of the computer programme willbe published elsewhere.

VISUAL EVOKED RESPONSESThe visual evoked responses were recorded topattern reversal stimuli between a midoccipitalelectrode 50 mm above the inion and a midfrontalreference electrode.

Electrodes were 9 mm silver chlorided platesand were attached to the scalp with collodion.Electrode resistance was maintained below 1.5 kfQ.A slide of a black and white checkerboard

pattern was back-projected onto a translucentscreen. The pattern was displaced through onesquare width every 500 ms, and each pattern re-versal was completed in 6 ms.The luminance of the black squares was 20 cd/

m2 and of the white squares 215 cd/M2. Theoverall luminance of the pattern did not change.The pattern subtended 300 of field and the indi-vidual squares 44'.

Routinely, 128 pattern reversals were averagedin two successive recordings and these resultsaveraged to give values for 256 reversals. TheVER was amplified and averaged by a MedelecDAV6 averager (band width of amplifier 0.16 Hz-0.8 Hz) and was recorded on photosensitive paper.The subjects viewed the pattern from one metreand recordings were made from each eye separ-ately. Refractive errors were corrected with spec-tacles. Patients with reduced visual acuity anddifficulty fixating the centre of the pattern werehelped by a centrally placed cross or more per-ipherally placed marker that lay within theirvisual field.

1009


on Decem


j.com/

J Neurol N


ovember 1979. D

ownloaded from


C. J. K. Ellis

Measurements of latency were made to the peakof the major positive potential, and the amplitudewas measured from the preceding negative peak.

CLINICAL ASSESSMENTThe pupils were assessed clinically by the swinginglight test. Visual acuity was measured to distantand near objects using Snellen charts and standardtest types.

Colour vision was assessed with H-R-R pseudo-isochromatic plates under standard lighting con-ditions. Visual fields were tested on the FriedmannCentral Field Analyser and on the Goldmannperimeter. Subtle defects demonstrated on thesetests were confirmed on the two metre tangentscreen. Macular threshold was recorded from theFriedmann Analyser. The irides of each patientwere examined with a Haag-Streit 900 slit lamp toexclude local pathology.Where possible patients were examined at ap-

proximately weekly intervals during the acutephase of the illness and then at longer intervalsonce visual acuity had returned to normal. Alltests were performed at one recording session. Therelationships between stimulus intensity, reflexamplitude, latency, and maximum rate of pupil-lary constriction were expressed mathematicallyby second order polynomials using standard tech-niques: 95% confidence limits were calculated inorder to define the limits of the normal range.

Results

PUPILLOMETRYLight reflex amplitudeIn all control subjects the amplitude of the changein pupil diameter during the direct light reflexincreased with increasing stimulus intensity. Thiswas not a linear relationship at the higher stimulusintensities as shown by the normal range in Fig. 1.There was a wide variation of reflex amplitudeamong the control subjects indicated by the 95%confidence limits. All but two of the 104 valuesfrom the control subjects lay within these limits.

In any individual subject at any given level ofstimulus intensity, the difference in amplitude ofthe direct light reflex from the two eyes was gen-erally small. However, there was a wider variationin this interocular difference at the lower stimulusintensities than at the three highest intensities, atwhich levels the difference was consistently small.Thus for each subject a single numerical value ofinterocular difference was given by the mean ofthe values at the three highest stimulus intensities.These values from the 19 control subjects aregiven in Fig. 2.

3-

E

x t~~~**

0i12 3 4 5 53

Stimulus intensity (log units)

Fig. 1 Relationship between light reflex amplitudeand stimulus intensity. The lines indicate the meanand the 95% confidence limits from the controlsubjects determined in this and all subsequent figuresby second order polynomials. The black dots indicatethe values from the affected eye of the 22 patients atthe same levels of stimulus intensity on their firstexamination.

1 4

12-EE

-DD 10-

Ea

t 08-a)

YE.91

-06-Z)-e

5)'0

0

5o-2

0

Controlsubjects

I

A

Patients

Fig. 2 Interocular difference in amplitude of thedirect light reflex. These values are determined bytaking the mean interocular difference at the threehighest stimulus intensities. The bar indicates themean +2 SD. Values from the patients relate to theirfirst examination.

1.010

0

0

0

0

0


on Decem


j.com/

J Neurol N


ovember 1979. D

ownloaded from



The mean interocular difference was 0.09 mm(SD 0.06 mm), and the upper limit of the normalrange was defined as the mean +2 SD, namely0.21 mm. One subject lay outside this normalrange.The light reflex amplitude was reduced in all

22 patients. This reduction could be identifiedeither by comparison with the control range or bycomparison with the reflex amplitude from theasymptomatic eye. Figure 1 shows the values fromthe affected eye of the patients. The number ofvalues that lay outside the normal range variedwith the level of stimulus intensity, being 50% ofall values at the three highest levels. At the lowerstimulus intensities fewer pupils can be shown tohave a small reflex amplitude. The most severelyaffected patients with visual acuity of 6/60 or lessshowed a reflex amplitude of less than 1.50 mmin the affected eye at the highest stimulus intensity.Patients with less reduction in visual acuity allshowed values in excess of 1.50 mm. However, themost intense stimulus available was not sufficientto elicit a maximal response.When light reflex amplitude in the two eyes was

compared, the interocular difference was found toincrease with increasing stimulus intensity. Avalue for this interocular difference at the threehighest stimulus intensities was calculated as forthe control subjects. Figure 2 shows the valuesfrom the 22 patients on their first examination.All lay outside the normal range and varied fromthe severely affected patient with an interoculardifference of 1.28 mm down to 0.23 mm. Thisvalue was used as an expression of the magnitudeof the relative afferent pupillary defect.The magnitude of the relative afferent defect

assessed on the first examination in each patientwas found to correlate with the visual acuity atthat time (Fig. 3). The afferent defects of mag-nitude in excess of about 0.8 mm were all associ-ated with visual acuity of 6/60 or less. The smallerafferent defects were associated with correspond-ingly less impairment of visual acuity.

LatencyThe latency from stimulus to onset of response incontrol subjects was found to decrease with in-creasing stimulus intensity. Figure 4 shows the lineof best fit of this relationship and the 95% con-fidence limits. Three of the 106 values lay outsidethese confidence limits. The minimum latency re-corded from the control subjects was 220 ms.The latency was also analysed in relation to the

corresponding light reflex amplitude. Figure 5shows that latency decreases with increasing lightreflex amplitude: again 95% confidence limits are

131- 2 -

C 1'0-a,

0 0-8-

>

° E 06-SE-

m 02-a

0 -

* :* :

0*

6/60 6/36 6/24 61t8 6/126/9 6/6Visual acuity

50 40 30 20 10 5Angle subtended

Fig. 3 Relationship between the magnitude of therelative afferent defect and visual acuity in theaffected eye of the 22 patients on their firstexamination. Acuity of 6/60 also includes all patientswith vision worse than 6/60.

500

400-

> 300-

200-

100.

* 03

0** 0 S~~~~~~~~~~

* ;- m

1 2 3Stimulus intensity ( lag units)

5 53

Fig. 4 Relationship between latency and stimulusintensity. The lines indicate the mean and the 95%confidence limits from the control subjects. The blackdots indicate the values from the affected eye of thepatients at the same levels of stimulus intensity ontheir first examination.

shown. Two of the 106 values lay outside thisrange.The latency of the pupillary responses in the

affected eyes of patients was prolonged in mostcases. This abnormality was less obvious than theassociated reduction in light reflex amplitude, poss-ibly because of the low resolution of thz pupil-lometer and the difficulties of defining exactly theonset of the pupillary constriction. However, Fig.

1011

.

.

.


on Decem


j.com/

J Neurol N


ovember 1979. D

ownloaded from


C. J. K. Ellis

00

.m 00O

'..0 0

... S S

. .00.. 0

0 . 0.. 0 0.. *.... I

* . *0.000 0000

*o.000 0.0*-.

.. 0 @0 @0

8-

7-

6-

a

a 5-

I, E 4-o'-a

E . 3-

'x

O ' 2-

0-

I I0 1Light retlex amplitude (mm)

2 25

Fig. 5 Relationship between latency and amplitudeof the light reflex. The lines indicate the mean and the95% confidence limits from the control subjects. Theblack dots indicate the values from the affected eyeof the patients on their first examination.

*

0~~~~~~~~~~

0~~~~~~~~~~

a~~~~~~

0

I~~~~~~~~~~

I1 2 3Stimulus intensity ( log units )

4 5 53

Fig. 6 Relationship between maximum rate ofpupillary constriction and stimulus intensity. The linesindicate the mean and 95% confidence limits from thecontrol subjects. The dots indicate the values from theaffected eye of the patients at the same level ofstimulus intensity on their first examination.

4 shows a number of affected eyes to lie outsidethe normal range of latency at each level ofstimulus intensity. The longest latency recordedwas 480 ms, and upper limit of the normal rangewas 420 ms at the lowest stimulus intensity. Whenlatency was plotted against light reflex amplitude(Fig. 5) all but eight of the values from thepatients lay within the normal range, indicatingthat although the latency was abnormal with re-spect to stimulus intensity, it was normal withrespect to its corresponding light reflex amplitude.

Maximum rate of pupillary cons.trictionThe maximum rate of pupillary constriction incontrol subjects was found to increase with in-creasing stimulus intensity. Figure 6 shows themean and 95% confidence limits of this relation-ship. Four of the 113 values from the controlsubjects lay outside these confidence limits.

Figure 7 shows the maximum rate of pupillaryconstriction plotted against the correspondinglight reflex amplitude. This shows the constrictionrate increasing with increasing light reflex ampli-tude. None of the 113 values from the controlsubjects lay outside the 95% confidence limitsshown.The maximum rate of pupillary constriction in

the patients is shown in Fig. 6 plotted againststimulus intenvity. Although many of the valueslie outside the normal range at the higher stimulus

8

7.

6.

a-a._.44_ 3.E-

2'c 2

1-

O0u- I

0 1-0Light reflex amplitude (mm)

Fig. 7 Relationship between maximum rate ofpupillary constriction and light reflex amplitude. Thelines indicate the mean and 95% confidence limitsfrom the control subjects. The dots indicate the valuesfrom the affected eye of the patients at their firstexamination.

1012

500-

400

E 300.>1

200.

100-

20 25


on Decem


j.com/

J Neurol N


ovember 1979. D

ownloaded from



intensities, the majority are inside and can beshown to be abnormal only by comparison withthe other eye.When plotted against light reflex amplitude all

except one of these values lay within the normalrange (Fig. 7). Thus when a pupil was found toconstrict slowly, that rate of constriction was onlyabnormal with respect to stimulus intensity andnot to reflex amplitude.

VISUAL EVOKED RESPONSES (VER)A mplituderhe amplitude of the major positive potential wasmeasured from the preceding negative peak. Themean peak amplitude from the 40 eyes of the 20control subjects was 11.0 p.V (SD 2.6 15V). Therewas a consistently small interocular difference inamplitude in the control subjects. This differencewas expressed as a percentage of the larger ampli-tude, and the mean was found to be 4.3% (SD3.1%/). The upper limit of the normal range wasdefined as the mean +2 SD, namely 10.5%. Noneof the control subjects lay outside this range.The amplitude of the major positive peak was

reduced when evoked from all the affected eyesof patients. Visual acuity of 6/60 or worse wasassociated with abolition of any recognisable posi-tive peak. The degree of reduction of the VERamplitude relative to the other eye was found tocorrelate with the magnitude of the relative affer-ent defect such that the larger the afferent defect,the greater the reduction in the VER amplitude

100 -

.( 80-

CE 60-

>40-

C

'- 20-cc

0-

0 0 0 a 0 0000

00

0

00 0*.00. 0

*

0% 0

0 *

so

0

00

0

1013

(Fig. 8). A reduction of 100% in VER amplitudedenotes abolition of any recognisable positivepotential, and this was seen with afferent defectsof magnitude greater than about 0.6 mm. How-ever, it was a significant finding that in patientswith abolition of the VER and visual acuity ofworse than 6/60, pupillary responses were stillpresent. The values in Fig. 8 are taken both fromthe first visit of the patients and during follow-up.

LatencyThe latency was measured to the peak of themajor positive wave. The mean from the 40 con-trol eyes was 104.0 ms (SD 3.5 ms). The upperlimit of normal was defined as the mean +3 SD(115 ms). This was the criterion used by Asselmanet al. (1975). None of the control eyes gave alatency outside this range.The latency to the peak of the major positive

potential was prolonged from all affected eyes.Five of the 22 patients had prolonged VER laten-cies from their asymptomatic eye, at the firstexamination. Four of these had clinical evidenceof further lesions in the nervous system. The inter-ocular difference in latency, in the patients inwhom there was a detectable positive peak fromthe affected eye, was not related to the magnitudeof the afferent defect.

CLINICAL ASSESSMENT

On presentation all patients had clinically detect-able relative afferent pupillary defects in theaffected eye. The magnitude of the defect variedin relation to visual acuity (Fig. 3). All patientsshowed a central scotoma on Goldmann andFriedmann field testing. The macular thresholdcorrelated with visual acuity at presentation.Vision of less than 6/36 was associated with norecordable macular threshold on the Friedmannanalyser. Fourteen of the 22 patients had discswelling at presentation. The presence of discswelling did not correlate with severity as assessedby any other of the tests performed. Colour visionwas affected in all cases, both red-green and blue-yellow defects being associated with the severecases, and only red-green defects being associatedwith the mildly affected cases.

0 02 04 06 08 10 1 2Magnitude of relative afferent defect ( mm )

Fig. 8 Relationship between VER amnplitude andmagnitude of the relative afferent defect. VERreduction of 100% indicates abolition of theresponse. The values are taken from the patients atpresentation and during follow-up.

FOLLOW-UP STUDIES

During follow-up, seven of the 22 patients devel-oped some evidence, either clinically or on VERexamination, of involvement of the other opticnerve. Assessment of the relative afferent defect inthese cases was unreliable. Two patients did notattend for follow-up. Most information was gained


on Decem


j.com/

J Neurol N


ovember 1979. D

ownloaded from


Time after V isualI evoked response Size of Visual Macular Visual Colouronset of afferent acuity threshold f ield visionyseptoms Pupillary (left) (left) (left) (left)symptoms ~Right eye Left eye defe'ct (mm)

8 days -'.-| 083 3/60 0 Central Plates_____________1Q~~~~~~~~~~~~~~Ls~~~ scotoma not

AL50ms ~~~~~~~~~seen16 days | 053 6/9 1 2 Ceirntrcal blue-yellow

____________________________ ~~~~~green defect

Fibre Mild1 month I /N|. 0 37 6/5 2 4 bund le red -green

______________ ~~~~~~~~~~detectdefect

6weeks ~~~~~~~~~~~~~~~~~~~~~~~~~~FibreMild|6D weeks |-\/ \x~/ | ~P~\~ | 045 6/5 2 4 bundle red-greendefect defectFibre Mild

9 weeks -j--Vj 00°39 6/5 2 6 bundle red-greendefect defect

Fibre Plate 33 months 0-36 6/5 2-4 bundle missed

_____________________________ ________deiect ______

Fibre Plate 35 morThs 040 6/5 2 4 bundle missed

defect

Fig. 9 Example of typical results during follow-up. Patient TO'R aged 36 years. Left optic neuritis.Results during five month follow-up. Colour vision was measured with H-R-R pseudoisochromatic plates.Shows persistent abnormality of VER, pupil reaction, and subtle defects in visual fields and colour vision.Positive potentials are indicated by an upward deflection.

from the remaining 13 cases in whom an unchang-ing VER latency from the asymptomatic eyetestified to the reliability of its use as a control.The results from these 13 cases were similar and

were typified by patient TO'R (Fig. 9). Two phasescould be identified during follow-up. An initialrecovery phase, which lasted until visual acuityreturned to normal, was followed by a static phasein which no change in clinical state was observed.This static phase persisted for the length of follow-up of this study.During the recovery phase visual acuity im-

proved in association with an increase in VERamplitude, decrease in the magnitude of the rela-tive afferent defect, reduction in scotoma size,decrease in macular threshold, reduction in theseverity of the colour defect, and resolution ofoptic disc oedema.At the end of the recovery phase visual acuity

was normal in all 13 cases. Colour testing showedpersistent mild red-green defects in five patients.Macular threshold remained abnormal in twocases. Visual fields tested on the Goldmann per-

imeter showed a persistent abnormality in sevenof the 13 cases. These abnormalities consisted ofconstriction of the field to a small test object,mild relative central or paracentral scotomata, ormultiple arcuate defects compatible with retinalnerve fibre atrophy. These field defects were subtleand were only detectable consistently on successivetesting in the most cooperative and alert patients.A relative afferent pupillary defect was present

in each of these 13 patients throughout the lengthof follow-up. The magnitude of this defect did notalter significantly after the reduction in magnitudeseen during the recovery phase was complete. Thegreater the magnitude of the defect the easier wasits detection by the swinging light test.The persistent afferent defect was associated

with a persistent reduction in VER amplituderelative to the other eye. The degree of this VERreduction was related to the magnitude of therelative afferent defect (Fig. 8).

In the seven cases with subsequent involvementof the other optic nerve, the magnitude of therelative afferent defect was decreased, abolished or,

C. J. K. Ellis1014


on Decem


j.com/

J Neurol N


ovember 1979. D

ownloaded from


The aflerent pupillary defect in acute optic neuritis

in one case, reversed. However, the relationshipbetween the percentage reduction of VER ampli-tude and the magnitude of the relative afferentdefect was undisturbed in these cases.Two of the patients developed unequivocal optic

disc pallor during follow-up.In none of the 13 unilateral cases did the

latency of the VER in the affected eye changeduring follow-up.

RELATION OF PUPILLOMETRY TO THE SWINGINGLIGHT TESTThe swinging light test was performed by at leastone other independent observer ignorant of thecondition of the patient. It was found that relativeafferent defects of magnitude in excess of 0.8 mmshowed little or no detectable initial pupil con-striction as the light fell onto the affected eye, butonly rapid dilatation. Afferent defects of thismagnitude were seen only in the initial, acutephase when the patient had other obvious clinicaldeficit. As the magnitude of the afferent defectdecreased it was found that abnormalities werestill detected readily if asymmetry of the amplitudeof the initial pupillary constriction and subsequentpupillary escape were assessed. On this basis eventhe smallest significant afferent defect wasdetectable by the swinging light test.

Discussion

Results of this study have shown that pupillaryabnormalities are a constant feature of acute opticneuritis. There was a wide physiological range ofall parameters studied in the control subjects, andsome affected eyes could only be shown to be ab-normal by comparison with the asymptomatic eye-that is, they only showed a relative afferentdefect. However, 50% of the affected eyes laybelow the 95% confidence limits of light reflexamplitude at the higher stimulus intensities andthus showed an "absolute" afferent defect. Thismethod of definition may thus be used to detectpupillary abnormality even in bilateral disease.

It is well recognised that, in any individual,variation in amplitude of the pupillary light re-actions may occur under constant experimentalconditions. This variation is thought to be theresult of fluctuation in supranuclear inhibition ofthe Edinger-Westphal nucleus (Lowenstein et al.,1963).The pupillary responses to light in an affected

eye were of smaller reflex amplitude, longerlatency, and slower maximum constriction velocitythan in the response from the other eye at thesame level of stimulus intensity. However, if re-

flexes of equal amplitude in both the affected andasymptomatic eye were compared, the latency andmaximum constriction velocity did not differ sig-nificantly and the only abnormality was that moreintense light stimulation was required to elicit sucha reflex in the affected eye. This finding impliesthat the afferent defect seen in optic neuritis isrelated to an apparently decreased stimulus in-tensity in the affected eye. Thus Lowenstein'sdescription of these reflexes as "low intensity re-actions" would seem entirely appropriate.When the pupils are examined by the swinging

light test two parameters of pupillary function areassessed. These are the amplitude of the initialpupillary constriction when the light falls on theeye and the pupillary redilatation, or escape, seenunder continuing light stimulation. In optic neur-itis reduction in the amplitude of the initial con-striction and increase in pupillary escape occurtogether in the affected eye and constitute what istermed the relative afferent defect or MarcusGunn sign. This study has shown that small differ-ences in the amplitude of the initial constrictionare significant and that the swinging light test pro-vides a sensitive means of assessment of asym-metric involvement of the optic nerves.

In order to correlate pupillary abnormality withthe visual evoked response and with clinical find-ings it has been necessary to quantify the relativeafferent defect. The usual method is to reducestimulus intensity in the normal eye with neutraldensity filters until the reflex amplitude in the twoeyes is similar and to express the magnitude of therelative afferent defect as the value of the logunits of filter required. In this study it has beenshown that in the control subjects a reflex of anygiven amplitude has a latency and maximum con-striction velocity that lie within statistically de-fined limits. In the patient's affected eye thisfundamental relation between reflex amplitude,latency, and maximum constriction velocity wasnot disturbed, and values for latency and con-striction velocity lay within the normal rangewhen plotted against reflex amplitude (see Figs. 5and 7). Therefore any one of the measured par-ameters of pupillary function could be used torepresent the magnitude of the relative afferentdefect. In this study reflex amplitude was used asit is most easily observed in clinical pupillaryassessment. The differences between the reflexamplitudes from the two eyes at the three higheststimulus intensities were averaged to give a singlenumerical value in millimetres that representedthe magnitude of the relative afferent defect.Lowenstein et al. (1964) found that in normal sub-jects latency did not vary with reflex amplitude

1015


on Decem


j.com/

J Neurol N


ovember 1979. D

ownloaded from


C. J. K. Ellis

but rather was related to stimulus intensity, andsuggested that latency was, therefore, the mostreliable parameter of the low intensity reaction tolight. The present study using averaged pupillaryresponses, did not confirm this and found latencyto vary with both reflex amplitude and stimulusintensity.The visual evoked response (VER) to pattern

reversal has become established as a reliable andsensitive means of detecting abnormalities in opticnerve conduction in optic neuritis (Halliday et al.,1972). In this study the commonly employed mid-line occipital response was measured. It is anobvious course to attempt to relate VER andpupillary latencies. However, there are difficultiesin this approach. The pupillary latency is long,the minimum value in this study was 220 ms. Thisis due primarily to the delay in initiating irissmooth muscle contraction. Loewenfeld (1966)found that in the pigeon-with a striated musclesphincter-a minimum latency of around 60 mscould be recorded. She also found that directelectrical stimulation of the ciliary ganglion incats-with a smooth muscle iris sphincter-resultedin a minimum latency no shorter than 180-200 ms.It has been demonstrated repeatedly that pupillarylatency decreases with increasing stimulus intensity(Alpern et al., 1963; Lowenstein et al., 1964), andin this study a range of 220 to 480 ms was re-corded. This variability makes comparison of VERand pupillary latency difficult, and this difficulty iscompounded by the resolution of the pupillometerbeing low in relation to that of the VER.That the delay in the latency of the VER is

not related to the magnitude of the afferent defectis suggested by two of my findings. The differencebetween the VER latencies from the two eyes wasnot related to the magnitude of the relative affer-ent defect. Further, during the recovery phasewhile the magnitude of the pupillary defect wasdecreasing, there was no corresponding decreasein the VER latency in that eye.The VER amplitude from the normal controls

was variable. However, the interocular differencein amplitude showed a good correlation with themagnitude of the afferent defect such that largeafferent defects were associated with a greaterrelative reduction of VER amplitude and the mostmarked afferent defects with abolition of the VER.This relation implies that from the presence of arelative afferent pupillary defect may be inferreda reduction in VER amplitude in that eye. Noreliable information regarding VER latency isgiven by pupillary examination.Any discussion of the relation between VER

and pupillometry must acknowledge one particular

field of ignorance. The identity of the pupillomotorafferent nerve fibres in the optic nerve, chiasm,and tract is not known. It is now widely acceptedthat at the retinal level there is no distinction be-tween pupillary and visual rods and cones(Lowenstein and Loewenfeld, 1969). However,the argument has never been resolved satisfactorilyas to whether these afferent fibres are separatefrom ganglion cell onwards or whether the pupil-lary afferent fibres are collaterals of the visualafferent fibres. This study has shown a degree ofconcordance between pupillary and visual deficitin optic neuritis which suggests that, even if thetwo groups of fibres are separate, they are affectedequally by the demyelinating process.

Halliday and McDonald (1977) have explainedsymptoms and signs in optic neuritis in relation toreduction in amplitude rather than prolongation oflatency of the VER. They have attributed the re-covery during the acute phase to resolution ofnerve swelling and oedema with gradual reductionin the number of blocked fibres. The presentresults suggest that this may also be the explan-ation of the conduction disturbance underlyingthe afferent pupillary defect. Large afferent defectsare seen with the more reduced VER amplitudeimplying a larger number of optic nerve fibres inwhich conduction is blocked and a consequentlyreduced synchronous afferent pupillary volley.

Halliday et al. (1973) also found that visualacuity and VER amplitude returned to normaltogether. In that study normal VER amplitudewas defined in absolute terms. In the present studynormal VER amplitude has been defined in rela-tive terms, and on this basis none of the patientsshowed a normal amplitude during follow-up. Thispersistently reduced VER amplitude was associatedwith a persistent relative afferent defect in allpatients. Although some of these patients alsoshowed subtle visual field defects and mild abnor-malities of colour vision, the pupillary abnormalitywas the only consistent, objective, clinical sign ofthe previous episode of optic neuritis. It is im-portant to stress that during the recovery phasethere is some reduction in magnitude of theafferent defect and consequently the defects seenduring follow-up are less obvious than those inthe acute phase. To facilitate the clinical detectionof these subtle defects, Thompson (1976) has sug-gested that the sensitivity of the swinging lighttest be increased by performing it in a dark roomwith a bright torch so as to increase the pupillaryexcursions. The patient should fix a distant objectin order to avoid the miosis of a near reaction andfluctuating fixation. The light should be switchedfrom eye to eye every three to five seconds, and

1016


on Decem


j.com/

J Neurol N


ovember 1979. D

ownloaded from



any asymmetry of the amplitude of the intialpupillary constriction or of the degree of pupillaryescape should be observed. The timing of theswing may be adjusted to accentuate asymmetriesin pupillary escape, but both eyes should bestimulated for the same length of time to avoidunequal bleaching of the retinae.

I am indebted to Mr P. Bourne of the RayneInstitute, St Thomas' Hospital who devised thecomputer programme. I would also like to thankDr S. E. Smith of the Department of Pharma-cology, St Thomas's Hospital Medical School,Dr R. W. Ross Russell of the Department ofNeurology, St Thomas' Hospital, Dr Hisako Ikedaof the Rayne Institute, St Thomas' Hospital, andProfessor W. I. McDonald of the Institute ofNeurology for their help in devising the projectand their comments on the manuscript, and MissJ. Lace for typing the manuscript. This work wasundertaken during tenure of the position of Re-search Fellow sponsored by the Prevention ofBlindness Research Fund.

References

Alpern, M., McCready, D. WV., and Barr, L. (1963).The dependence of the photopupil response onflash duration and intensity. Journal of GeneralPhysiology, 47, 265-278.

Asselman, P., Chadwick, D. W., and Marsden, C. D.(1975). Visual evoked responses in the diagnosis andmanagement of patients suspected of multiple scler-osis. Brain, 98, 261-282.

Fison, P. N., Garlick, D. J., and Smith, S. E. (1979).Assessment of unilateral afferent pupillary defectsby pupillography. British Journal of Ophthalmology,63, 195-199.

Gerold, H. (1846). Die Lehre vom schwarzen Staar unddessen Heilung, p. 30. Rubach: Magdeburg.

Gunn, R. M. (1904). Retrobulbar neuritis. Lancet, 4,412.

Halliday, A. M., and McDonald, W. I. (1977). Patho-physiology of demyelinating disease. British MedicalBulletin, 33, 21-27.

Halliday, A. M., McDonald, W. I., and Mushin, J.(1972). Delayed visual evoked responses in opticneuritis. Lancet, 1, 982-985.

Halliday, A. M., McDonald, W. I., and Mushin, J.(1973). Delayed pattern-evoked responses in opticneuritis in relation to visual acuity. Transactionsof the Ophthalmological Society of the UnitedKingdom, 93, 315-324.

Hirschberg, J. von (1884). Neuritis Retrobulbaris.Zentralblatt fur Praktische Augenheilkunde, 8, 185.

Kestenbaum, A. (1946). Clinical Methods of Neuro-ophthalmological Examination, pp. 282-291. Gruneand Stratton: New York.

Levatin, P. (1959). Pupillary escape in disease of theretina or optic nerve. Archives of Ophthalmology,62, 768-779.

Loewenfeld, 1. (1966). Recent Developments in VisionResearch. National Research Council Publication,no. 1272, pp. 17-105. Edited by M. A. Whitcomb.Washington.

Lowenstein, 0. (1954). Clinical pupillary symptomsin lesions of the optic nerve, optic chiasm, optictract. Archives of Ophthalmology, 52, 385-403.

Lowen3tein, O., and Friedman, E. D. (1942). Thepresent state of pupillography: its method anddiagnostic significance. A rchives of Ophthalmology,27, 969-993.

Lowenstein, O., and Loewenfeld, I. (1969). In TheEye, pp. 255-337. Edited by Hugh Davson. Aca-demic Press: New York.

Lowenstein, O., Feinberg, R., and Loewenfeld, I.(1963). Pupillary movements associated with acuteand chronic fatigue. Investigative Ophthalmology,2, 138-157.

Lowenstein. O., Kawabata, H., and Loewenfeld, I.(1964). The pupil as indicator of retinal activity.American Journal of Ophthalmology, 57, 569-596.

McDonald, W. I., and Sears, T. A. (1970). The effectsof experimental demyelination on conduction in thecentral nervous system. Brain, 93, 583-598.

McDonald, W. I. (1977a). In Vi-ual Evoked Potentialsin Man: New Devolopments, pp. 427-437. Edited byJ. E. Desmedt. Clarendon Press: Oxford.

McDonald, W. I. (1977b). Acute optic neuritis. BritishJournal of Hospital Medicine, 18, 42-48.

Saint-Yves, C. de (1742). Un nouveau traite6 desmaladies des yeux, p. 291. Pierre-Augustin leMercier: Paris.

Thompson, H. S. (1966). Afferent pupillary defects.American Journal of Ophthalmology, 65, 860-873.

Thompson, H. S. (1976). Pupillary signs in the diag-nosis of optic nerve disease. Transactions of theOphthalmological Society of the United Kingdoin,96, 377-381.

1017


on Decem


j.com/

J Neurol N


ovember 1979. D

ownloaded from


theafferent pupillary defect acute optic neuritis · theafferent pupillary defectin...

Documents