The Study of Visual Evoked Potentials in Patients withThyroid-Associated Ophthalmopathy IdentifiesAsymptomatic Optic Nerve Involvement*

MARIO SALVI†, ELIO SPAGGIARI, FABRIZIO NERI, CLAUDIO MACALUSO,ELIANA GARDINI, FRANCESCO FERROZZI, ROBERTA MINELLI‡, JACK R. WALL,AND ELIO ROTI

Centro per lo Studio, Prevenzione, Diagnosi e Cura delle Tireopatie, Istituto di Oftalmologia (E.S.,F.N., C.M.) and Radiologia (F.F.), Universita di Parma, Parma, Italy; and Thyroid-Eye ResearchProgram, Allegheny-Singer Research Institute (J.R.W.), Pittsburgh, Pennsylvania 15212-4772

ABSTRACTIn the present study we have recorded visual evoked cortical po-

tentials (VECP) in 88 patients affected by autoimmune thyroid dis-ease and thyroid-associated ophthalmopathy (TAO) without clinicalsigns of optic neuropathy. At the time of ophthalmological examina-tion, 37 of these patients were hyperthyroid, 41 were euthyroid, and8 were hypothyroid; 2 were not assessed. Twenty-nine normal sub-jects served as controls. We performed pattern reversal visual stim-ulation and recorded the amplitude and latency of the cortical electricresponse at 100 ms (P100 wave). There were no differences in themean P100 amplitude of TAO patients and normal subjects. The meanP100 latency in patients was 105.6 6 0.5 ms, significantly higher thanthat in normal subjects (102.0 6 0.5 ms; P , 0.00003). Latency ineuthyroid patients did not differ from that in either hypo- or hyper-thyroid patients. The VECP test was positive (latency, $110.0 ms) in

21 (23.8%) TAO patients. In patients with proptosis greater than 21mm, latency was 106.7 6 0.7 ms, significantly higher than that inpatients with normal Hertel measurements (104.3 6 0.6 ms; P ,0.01). Latency was not increased in patients with acute inflammatorysigns compared to those with inactive eye disease and in patients withaltered extrinsic motility. In patients with an abnormal visual fieldstudy, the mean latency was 110.3 6 1.5 ms, significantly higher thanthat in patients with a normal visual field (104.7 6 0.4; by t test, P ,0.000003). In conclusion, we observed a prolongation of the latency ofthe evoked cortical response in patients with TAO without subjectivevisual complaints and without optic nerve compression. We believethat the study of VECP in TAO is complementary to the study of thevisual field in identifying early optic nerve dysfunction in the absenceof decreased visual acuity. (J Clin Endocrinol Metab 82: 1027–1030,1997)

ACTIVE thyroid-associated ophthalmopathy (TAO) mayprogress to clinically evident optic neuropathy (ON)

in 3–5% of patients. This complication presents with a sud-den or progressive decrease in visual acuity, more frequentlyin patients with generalized congestive orbitopathy (1). It hasbeen shown that ON in TAO is due to compression of theoptic nerve at the orbital apex by the swollen extraocularmuscles, as a consequence of the disproportion between theintraorbital content and the volume of the bony orbital space(2). Early optic nerve dysfunction in the absence of decreasedvisual acuity may be revealed by color vision impairment (2)in combination with an abnormal visual field examination(3, 4).

Measurement of visual evoked cortical potentials (VECP),

electrical manifestations of brain response to an externalstimulus, has provided great sensitivity and precision in theassessment of many disorders of the central nervous system(5). The study of pattern reversal visual evoked potentialsmeasures the amplitude and latency of the transmission ofthe electric response along a complex central nervous systempathway after stimulation of the retina (6). Previous studieshave considered VECP of preeminent importance in the as-sessment of ON impairment in TAO because electrophysi-ological abnormalities have been reported to be the mostsensitive indicator of incipient ON (2, 7).

In the present study, after conducting full ophthalmolog-ical assessment, we performed VECP in a consecutive seriesof patients with TAO with normal best-corrected visual acu-ity to identify those who might have asymptomatic opticnerve involvement.

Subjects and Methods

We performed the study of VECP in 88 consecutive patients withTAO, 9 men and 79 women, aged 14–77 yr (mean age, 45.3 6 1.1 yr).Eighty patients had Graves’ disease, 4 had Hashimoto’s thyroiditis, and4 had primary myxedema. At the time of ophthalmological visit, thirty-seven patients were hyperthyroid, 41 were euthyroid, and 8 were sub-clinically hypothyroid, whereas two were not assessed. The mean du-ration of TAO was 9.6 6 3.7 months (median, 3.0 months) from the onsetof thyroid disease. We excluded from the study patients affected by eyedisease (severe myopia and astigmatism, cataract, glaucoma, maculopa-thy) that might affect the results of the test. Twenty-nine normal subjects,

Received March 27, 1996. Revision received June 21, 1996. Rerevisionreceived October 14, 1996. Accepted October 18, 1996.

Address all correspondence and requests for reprints to: Dr. MarioSalvi, Cattedra di Endocrinologia, Universita di Parma, via Gramsci 14,43100 Parma, Italy.

* This work was supported by Grants 93.00413.CT04, 95.00877.CT04,and 93.00405.CT04 and 95.00940.CT04 from the Consiglio Nazionaledelle Ricerche (Rome, Italy); the Allegheny-Singer Research Institute(Pittsburgh, PA), and Grant 421001. Presented in part at the 11th Inter-national Thyroid Congress, Toronto, Canada, September 10–15, 1995.

† Visiting Adjunct Professor, Department of Medicine, McGill Uni-versity, Montreal, Canada.

‡ Recipient of a fellowship from Associazione Volontaria PromozioneRicerca Tumori (Parma, Italy).

0021-972X/97/$03.00/0 Vol. 82, No. 4Journal of Clinical Endocrinology and Metabolism Printed in U.S.A.Copyright © 1997 by The Endocrine Society

1027

10 men and 19 women, aged 14–73 yr (mean age, 41.8 6 2.1 yr), werestudied as controls.

Ophthalmological assessment

Ophthalmological examination included 1) evaluation of eyelid andsoft tissue inflammation with measurement of the lid fissure, 2) Hertelexophthalmometry, 3) color vision (Ishihara tables), 4) cover test andHess’s screen, 5) best-corrected visual acuity, 6) tonometry in primaryposition and in upgaze and lateral gaze, 7) fundus examination, 8)computerized visual field examination (a scotoma was defined as $2adjacent points of $5 decibels sensitivity loss for each point or as $1point of $10 decibels sensitivity loss), and 9) orbital computed tomog-raphy (CT) scan with measurement of the muscle volume and evaluationof apical crowding and optic nerve compression. The CT scan wasperformed in only 23 patients, who had evident altered ocular motility.

Recording of VECP

The visual stimulus was a pattern reversal checkerboard displayed ona black and white monitor placed 105 cm from the patient, subtendinga 10° visual angle, with each check subtending a 27° visual angle. Thisparadigm of visual stimulation provides a stimulation of the central(macular) part of the retina (8). This avoids contamination of the evokedcortical response that has its maximum positivity at 100 ms, with aresponse arising from paramacular stimulation with maximum posi-tivity at 135 ms (9). The checkerboard had a 100% contrast and wasalternated in time at 1 Hz (i.e. 2 reversal/s), with a space- and time-averaged mean luminance of 70 candela/m2. Cortical responses wererecorded from an electrode placed 2 cm above the inion, referenced toa midfrontal electrode, with ground placed at the right mastoid. Allelectrodes were Ag/AgCl. The signal was amplified 50,000 times andbandpass filtered between 0.1–100 Hz. Responses to 100 reversals wereaveraged. The P100 component of the cortical response was consideredfor measurement. The latency of P100 was calculated as the time fromstimulus reversal to the peak, and the amplitude was measured from thetrough of the preceding N75 to the peak of P100.

Statistical analysis

We used the t test for analysis of amplitude and latency valuesbetween the groups of patients and normal subjects and between thegroups of patients with and without the various clinical ophthalmolog-ical signs. We compared the prevalence of a positive VECP test in thegroups of patients with and without clinical signs by x2 analysis. Valuesare reported as the mean 6 se.

ResultsOphthalmological findings

At ophthalmological examination, eyelid signs, includinglid lag and/or retraction (fissure, .11 mm) and lid edemawere present in 70 patients (79.5%). Proptosis, with a Hertelmeasurement greater than 21 mm, was present in 54 patients(61.3%), of whom 43 had bilateral involvement. Signs of softtissue inflammation, including lid edema, conjunctival in-jection, and/or chemosis, epiphora, caruncle edema, andcorneal stippling were found in 39 patients (44.3%). Pupillaryreflex was normal in all patients. We found 11 TAO patients

(12.5%) with increased intraocular pressure (IOP) in the pri-mary position or on upgaze and lateral gaze. By performinga cover test and drawing a Hess’s screen, we observed that40 patients (45.4%) had altered extraocular muscle function.Both the measurement of abnormal IOP and the finding ofmuscle dysfunction reflect an abnormality of extrinsic ocularmotility and, therefore, were considered together for statis-tical analysis. We performed orbital CT scan in the presenceof evident altered ocular motility and found increased eyemuscles diameters in 41 of the 46 orbits studied, but notcompression or abnormalities of the optic nerve at the orbitalapex. Opthalmoscopy revealed a normal nerve head. Opticnerve function was studied by assessing color vision andperforming a computerized visual field study. Only 1 patienthad dyscromatopsia, whereas 23 (26.7%) had visual fielddefects. These were evidenced as paracentral scotomas (22eyes; 19 patients) or as constriction of field isopters (7 eyes;4 patients), without apparent significant distribution in thevisual field. All patients had, at the time of the examination,normal best-corrected visual acuity.

VECP study

The group of patients with TAO and that of normal sub-jects did not differ in age (Table 1), but differed in the femaleto male ratio (8.7:1 vs. 1.9:1). There were no differences in themean amplitude of the P100 wave of TAO patients (10.2 60.3 mV) and normal subjects (11.3 6 0.6 mV; by t test, P 5 NS;data not shown). As shown in Table 1, the mean latency ofthe P100 wave in patients was 105.6 6 0.5 ms, significantlyhigher than that in normal subjects (102.0 6 0.5 ms; by t test,P , 0.00003). The difference in the latency of the VECP wassignificant even when values were analyzed for each sepa-rate eye. To determine whether thyroid function would affectthe results of the VECP recordings, we recalculated the meanlatency values in TAO patients divided according to thyroidstatus. In hypothyroid patients, latency was 105.6 6 1.8 ms;in hyperthyroid patients, it was 106.1 6 0.7 (by t test, P 5 NS;not shown). Latency in euthyroid patients was 105.2 6 0.7 msand did not differ from that in either hypo- or hyperthyroidpatients, but, again, did differ from that in normal controls(P , 0.0004; data not shown). We calculated the upper limitof the normal range for the latency values recorded in ourgroup of normal subjects as 109.2 ms (mean 1 2 sd), and weconsidered a VECP test positive when the latency was 110.0ms or more. The test was positive in 33 eyes for a total of 21(23.8%) TAO patients (Table 2) and was negative in all nor-mal controls whose latency ranged from 93.0–107.0 ms. Anabnormal visual field was found in 14 of 33 eyes (42.4%) witha positive VECP test (Table 2), of whom 1 also had impaired

TABLE 1. Mean (6SE) age and latency of VECP in patients with thyroid-associated ophthalmopathy (TAO) and normal subjects

Eyes studieda TAO (no. of eyes) Normals (no. ofeyes) P (by t test)

Mean age (yr) OS 1 OD 45.3 6 1.1 (172) 41.9 6 2.1 (56) NS

Latency (ms) OS 1 OD 105.6 6 0.5 (172) 102.0 6 0.5 (56) ,0.00003OS 105.8 6 0.7 (87) 102.5 6 0.7 (28) ,0.01OD 105.5 6 0.6 (85) 101.5 6 0.7 (28) ,0.001

a OS, Left eye; OD, right eye.

1028 SALVI ET AL. JCE & M • 1997Vol 82 • No 4

color vision. All of these patients had normal fundusexamination.

Next, we studied the variations in VECP latency in thepatients in relation to the clinical signs of TAO. These resultsare shown in Table 3. Values recorded in both eyes werestudied together. Although we recorded a significantlyhigher latency in patients with proptosis greater than 21 mm(P , 0.01), no difference was observed in patients with softtissue inflammatory signs, in patients with altered extrinsicmotility, and in patients who underwent orbital CT scan inrelation to the presence of eye muscle enlargement. In pa-tients with an abnormal visual field, mean latency was sig-nificantly higher than that in patients with a normal test, andthe latter was significantly higher than that in controls (P ,0.0004; not shown).

We then studied the prevalence of a positive VECP test in

TAO patients grouped according to the clinical presentationof eye disease. The results are summarized in Table 4. Again,a positive VECP test was significantly more prevalent in theeyes of patients with visual field defects (14 of 29, 48.2%) thanin those with a normal visual field (19 of 142, 13.4%; P ,0.0005).

Discussion

In TAO patients affected by ON, vision loss occurs insid-iously in the context of a congestive inflammatory orbitopa-thy. Crowding of the orbital apex, which causes optic nervecompression, is suspected in the presence of restricted eye-ball movements. In studies of large groups of patients it hasbeen reported that marked proptosis, palpable lacrimalglands, increase in intraocular pressure on upgaze, and re-striction of extraocular muscles motility were predictive ofthe evolution of orbitopathy to ON (2, 10). In the absence ofvisual loss, other signs have been suggested as possible in-dicators of ON development, such as changes in color visionand in the optic nerve head on ophthalmoscopy (2) andabnormal visual field examination (3, 4). Indeed, none ofthese signs alone has proven to be specific for diagnosis.Electrophysiological studies, such as VECP, are consideredthe most objective and sensitive method of detecting earlyoptic nerve abnormalities (5). Existing reports of VECP stud-ies in TAO are not consistent because of the different tech-niques employed (11). Recordings of pattern reversal VECPare considered more reliable than those obtained after stim-ulation with either strobe or pattern flash, which produce achange in luminance (5). It has been reported that the latencyof VECP increases with age after the fifth decade, usually by2 ms/decade (5). In the present study, the difference in thelatency of the evoked cortical response recorded in patients

TABLE 3. Mean (6SE) latency of VECP in relation to theophthalmological signs of thyroid-associated ophthalmopathy

Clinical signNo. of eyes

studied(OS1OD)

Mean latency(msec) P (by t test)

Proptosis (.21 mm) 96 106.7 6 0.7 ,0.012No proptosis 76 104.3 6 0.6

Soft tissue inflammation 74 106.6 6 0.8 NSNo inflammation 98 104.9 6 0.5

Altered extrinsic motility 58 106.6 6 0.9 NSNormal motility 114 105.2 6 0.5

Abnormal orbital CT scana 41 107.6 6 1.3 NSNormal orbit 5 102.2 6 0.6

Visual field defects 29 110.3 6 1.5 ,0.000003Normal visual field 142 104.7 6 0.4

a CT scan performed on 46 eyes.

TABLE 2. Clinical signs and visual field examination of patients with thyroid-associated ophthalmopathy and a positive VECP test(latency, $110 ms)

Patientno.

Visual acuityaPupillary

examination Fundus Colorvision

Visual fieldb Latency (ms)

OS OD OS OD OS OD

1 20/20 20/20 nc n n n n 110 107d

2 20/20 20/20 n n n para n 112 1143 16/20 16/20 n n n para para 110 1144 20/20 20/20 n n n para para 121 1235 20/20 20/20 n n n para para 114 1176 20/20 20/20 n n n n n 112 1157 20/20 20/20 n n n NE n 105d 1118 10/20 20/20 n n n NE n NE 1109 20/20 20/20 n n n n n 114 117

10 20/20 20/20 n n n isop isop 112 108d

11 20/20 12/20 n n n isop NE 113 NE12 20/20 20/20 n n n n n 116 11213 20/20 20/20 n n n n n 129 11914 20/20 20/20 n n n n isop 109d 11415 20/20 20/20 n n n n n 106d 11016 20/20 20/20 n n n n n 111 109d

17 18/20 20/20 n n n n n 115 11118 20/20 18/20 n n n para para 129 12719 20/20 20/20 n n n n n 117 11120 20/20 20/20 n n abnormal para n 110 105d

21 20/20 20/20 n n n para n 119 115a Best corrected visual acuity.b para, Paracentral scotoma; isop, constriction of field isopters; NE, not examined.c n, Normal.d Normal VECP test.

VISUAL EVOKED POTENTIALS IN TAO 1029

with TAO without clinical ON and controls was unrelated toage, as age did not differ in the two groups. Furthermore, thegreater female to male ratio of patients compared to controlswas not expected to affect the results of the study becausefemales are known to have a slightly shorter P100 latencythan males (12). Hypothyroidism has been reported to pro-long the latency of VECP, and restoration of euthyroidismwith l-T4 has been shown to reverse the values to normal(13–15). Most of our patients were euthyroid or hyperthyroidat the time of the study, and these conditions are not knownto affect the results of the VECP test (16, 17). In the fewpatients who were subclinically hypothyroid, mean latencydid not differ from that recorded in the other groups of TAOpatients. Thus, as we have also excluded from the studypatients with ocular diseases that might affect the specificityof the VECP test, we believe that the prolongation of latencyobserved in patients with TAO is related to the presence ofautoimmune orbitopathy. The absence of amplitude changesin TAO patients is in agreement with the studies of Wijn-gaarde and Van Lith (18) and Setala et al. (11), but not withthat of Tsaloumas et al. (19), who, in patients with clinical ON,found a more significant abnormality of the amplitude thanthe latency of both flash- and pattern reversal-stimulatedVECP.

On clinical evaluation, none of the patients had clinicalON, although abnormalities of the visual field were recordedin 26.7% of the cases, suggesting asymptomatic optic nerveinvolvement. This finding is consistent with previous reportssuggesting that visual field defects in TAO patients are anearly sign of ON even in the presence of normal visual acuity(2, 3). We also found increased P100 latency in a proportionof patients (23.8%) who did not show eye muscle abnormal-ities on CT scan or increased IOP or congestive inflammatorysigns of orbitopathy, which are usually indicative of ON (2).In about 50% of the eyes with increased latency there werevisual field defects, and this association was significant. In aproportion of the eyes studied we observed a discrepancybetween visual field and VECP measurements that may de-rive from the different areas and sensitivities of retinal stim-ulation in the two tests. Interestingly, optic nerve dysfunc-tion would not have been diagnosed in 11 patients (;12% ofall cases) without the VECP test. As in the patients of thisstudy optic nerve dysfunction was not due to intraorbital

compression or to the presence of an increased ocular tone,factors such as ischemic damage due to narrowing and cel-lular infiltration of the nerve vessel walls (3) or a demyeli-nating-like neuritis (18) may be advocated to explain thepathophysiology of impaired optic nerve conduction.

In conclusion, we have shown that the study of VECP inpatients with TAO reveals asymptomatic optic nerve dys-function in the absence of deterioration of visual acuity. Thetest should be used in addition to visual field examination inthe ophthalmological assessment of the disease. VECP mea-surements are performed within a short time and requirelittle collaboration by the patient. A positive test shouldsuggest to the clinician additional intraorbital imaging andclose follow-up of patients even in the absence of optic nervecompression.

Acknowledgment

We thank Prof. Marco Cordella for helpful critical revision of themanuscript.

References

1. Trobe JD. 1981 Optic nerve involvement in dysthyroidism. Ophthalmology.88:488–492.

2. Neigel JM, Rootman J, Belkin RI, et al. 1988 Dysthyroid optic neuropathy.Ophthalmology. 95:1515–1521.

3. Gasser P, Flammer J. 1986 Optic neuropathy of Graves’ disease. Ophthalmo-logica. 192:22–27.

4. Trobe JD, Glaser JS, Laflamme P. 1978 Dysthyroid optic neuropathy. ArchOphthalmol. 96:1199–1209.

5. Chiappa KH, Ropper AH. 1982 Evoked potentials in clinical medicine. N EnglJ Med. 306:1140–1150.

6. Bobak P, Friedman R, Brigell M, Goodwin J, Anderson R. 1988 Visual evokedpotentials to multiple temporal frequencies. Arch Ophthalmol. 106:936–940.

7. Batch JA, Lepre F. 1990 Early diagnosis of Graves’ optic neuropathy usingvisual evoked responses. Postgrad Med J. 66:664–666.

8. Holder GE. 1991 Chiasmal and retrochiasmal lesions. In: Heckenlively JR,Arden GB, eds. Principles and practice of clinical electrophysiology of vision.St. Louis: Mosby Year Book; 557–564.

9. Halliday AM, Barrett G, Blumhardt LD, et al. 1975 The macular and para-macular subcomponents of the pattern evoked response. In: Lehmann D,Callaway E, eds. Human evoked potentials. Applications and problems. NewYork: Plenum Press; 135–151.

10. Feldon SE, Muramatsu S, Weiner JM. 1984 Clinical classification of Graves’ophthalmopathy. Identification of risk factors for optic neuropathy. Arch Oph-thalmol. 102:1469–1472.

11. Setala K, Raitta C, Valimaki M, Katevuo V, Lamberg B-A. 1992 The value ofvisual evoked potentials in optic neuropathy of Graves’ disease. J EndocrinolInvest. 15:821–826.

12. Stockard JJ, Hughes JF, Sharbrough FW. 1979 Visually evoked potentials toelectronic pattern reversal: latency variations with gender, age, and technicalfactors. Am J EEG Technol. 19:171–204.

13. Osterweil D, Syndulko K, Cohen SN, et al. 1992 Cognitive function in non-demented older adults with hypothyroidism. J Am Geriatr Soc. 40:325–335.

14. Abbott RJ, O’Malley BP, Barnett DB, Timson L, Rosenthal FD. 1983 Centraland peripheral nerve conduction in thyroid dysfunction: the influence ofl-thyroxine therapy compared with warming upon the conduction abnormal-ities of primary hypothyroidism. Clin Sci. 64:617–622.

15. Ladenson PW, Stakes JW, Ridgway EC. 1984 Reversible alteration of thevisual evoked potential in hypothyroidism. Am J Med. 77:1010–1013.

16. Mitchell KW, Wood CM, Howe JW. 1988 Pattern visual evoked potentials inhyperthyroidism. Br J Ophthalmol. 72:534–537.

17. Avramides A, Papamargaritis K, Mavromatis I, Saddic G, Vyzantiadis A,Milonas I. 1992 Visual evoked potentials in hypothyroid and hyperthyroidpatients before and after achievement of euthyroidism. J Endocrinol Invest.15:749–753.

18. Wijngaarde R, Van Lith GHM. 1979 Pattern evoked potentials in endocrineorbitopathy. Doc Ophthalmol. 48:327–332.

19. Tsaloumas MD, Good PA, Burdon MA, Misson GP. 1994 Flash and patternvisual evoked potentials in the diagnosis and monitoring of dysthyroid opticneuropathy. Eye. 8:638–645.

TABLE 4. Prevalence of a positive VECP test (latency, $110 ms)in patients with thyroid-associated ophthalmopathy in relation tothe different eye signs

Ophthalmological sign No. of positive VECP tests(% of total OS 1 OD) P (by x2 test)

Proptosis (.21 mm)Yes 27/96 (28.1) ,0.005No 6/76 (7.9)

Soft tissue inflammationPresent 21/74 (28.4) ,0.025Absent 12/98 (12.2)

Visual field defectsPresent 14/29 (48.2) ,0.0005Absent 19/142 (13.4)

Orbital CT scanAbnormal 12/41 (29.3) NSNormal 0/5 (0.0)

1030 SALVI ET AL. JCE & M • 1997Vol 82 • No 4