0021-972X/97/$03.

00/0 Journal of Clinical Endocrinology and Metabolism Copyright 1997 by The Endocrine Society

Vol. 82, No. 4 Printed in U.S.A.

The Study of Visual Evoked Potentials in Patients with Thyroid-Associated Ophthalmopathy Identifies Asymptomatic 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 Research Program, Allegheny-Singer Research Institute (J.R.W.), Pittsburgh, Pennsylvania 15212-4772
ABSTRACT In the present study we have recorded visual evoked cortical potentials (VECP) in 88 patients affected by autoimmune thyroid disease and thyroid-associated ophthalmopathy (TAO) without clinical signs of optic neuropathy. At the time of ophthalmological examination, 37 of these patients were hyperthyroid, 41 were euthyroid, and 8 were hypothyroid; 2 were not assessed. Twenty-nine normal subjects served as controls. We performed pattern reversal visual stimulation and recorded the amplitude and latency of the cortical electric response at 100 ms (P100 wave). There were no differences in the mean P100 amplitude of TAO patients and normal subjects. The mean P100 latency in patients was 105.6 0.5 ms, significantly higher than that in normal subjects (102.0 0.5 ms; P 0.00003). Latency in euthyroid patients did not differ from that in either hypo- or hyperthyroid patients. The VECP test was positive (latency, 110.0 ms) in 21 (23.8%) TAO patients. In patients with proptosis greater than 21 mm, latency was 106.7 0.7 ms, significantly higher than that in patients with normal Hertel measurements (104.3 0.6 ms; P 0.01). Latency was not increased in patients with acute inflammatory signs compared to those with inactive eye disease and in patients with altered extrinsic motility. In patients with an abnormal visual field study, the mean latency was 110.3 1.5 ms, significantly higher than that in patients with a normal visual field (104.7 0.4; by t test, P 0.000003). In conclusion, we observed a prolongation of the latency of the evoked cortical response in patients with TAO without subjective visual complaints and without optic nerve compression. We believe that the study of VECP in TAO is complementary to the study of the visual field in identifying early optic nerve dysfunction in the absence of decreased visual acuity. (J Clin Endocrinol Metab 82: 10271030, 1997)

CTIVE thyroid-associated ophthalmopathy (TAO) may progress to clinically evident optic neuropathy (ON) in 35% of patients. This complication presents with a sudden or progressive decrease in visual acuity, more frequently in patients with generalized congestive orbitopathy (1). It has been shown that ON in TAO is due to compression of the optic nerve at the orbital apex by the swollen extraocular muscles, as a consequence of the disproportion between the intraorbital content and the volume of the bony orbital space (2). Early optic nerve dysfunction in the absence of decreased visual 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),
Received March 27, 1996. Revision received June 21, 1996. Rerevision received October 14, 1996. Accepted October 18, 1996. Address all correspondence and requests for reprints to: Dr. Mario Salvi, 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 Nazionale delle Ricerche (Rome, Italy); the Allegheny-Singer Research Institute (Pittsburgh, PA), and Grant 421001. Presented in part at the 11th International Thyroid Congress, Toronto, Canada, September 10 15, 1995. Visiting Adjunct Professor, Department of Medicine, McGill University, Montreal, Canada. Recipient of a fellowship from Associazione Volontaria Promozione Ricerca Tumori (Parma, Italy).

electrical manifestations of brain response to an external stimulus, has provided great sensitivity and precision in the assessment of many disorders of the central nervous system (5). The study of pattern reversal visual evoked potentials measures the amplitude and latency of the transmission of the electric response along a complex central nervous system pathway after stimulation of the retina (6). Previous studies have considered VECP of preeminent importance in the assessment of ON impairment in TAO because electrophysiological abnormalities have been reported to be the most sensitive indicator of incipient ON (2, 7). In the present study, after conducting full ophthalmological assessment, we performed VECP in a consecutive series of patients with TAO with normal best-corrected visual acuity to identify those who might have asymptomatic optic nerve involvement.
Subjects and Methods
We performed the study of VECP in 88 consecutive patients with TAO, 9 men and 79 women, aged 14 77 yr (mean age, 45.3 1.1 yr). Eighty patients had Graves disease, 4 had Hashimotos thyroiditis, and 4 had primary myxedema. At the time of ophthalmological visit, thirtyseven patients were hyperthyroid, 41 were euthyroid, and 8 were subclinically hypothyroid, whereas two were not assessed. The mean duration of TAO was 9.6 3.7 months (median, 3.0 months) from the onset of thyroid disease. We excluded from the study patients affected by eye disease (severe myopia and astigmatism, cataract, glaucoma, maculopathy) that might affect the results of the test. Twenty-nine normal subjects,

1027

1028

SALVI ET AL.

JCE & M 1997 Vol 82 No 4

10 men and 19 women, aged 14 73 yr (mean age, 41.8 2.1 yr), were studied as controls.

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

Recording of VECP
The visual stimulus was a pattern reversal checkerboard displayed on a black and white monitor placed 105 cm from the patient, subtending a 10 visual angle, with each check subtending a 27 visual angle. This paradigm of visual stimulation provides a stimulation of the central (macular) part of the retina (8). This avoids contamination of the evoked cortical response that has its maximum positivity at 100 ms, with a response arising from paramacular stimulation with maximum positivity at 135 ms (9). The checkerboard had a 100% contrast and was alternated in time at 1 Hz (i.e. 2 reversal/s), with a space- and timeaveraged mean luminance of 70 candela/m2. Cortical responses were recorded from an electrode placed 2 cm above the inion, referenced to a midfrontal electrode, with ground placed at the right mastoid. All electrodes were Ag/AgCl. The signal was amplified 50,000 times and bandpass filtered between 0.1100 Hz. Responses to 100 reversals were averaged. The P100 component of the cortical response was considered for measurement. The latency of P100 was calculated as the time from stimulus reversal to the peak, and the amplitude was measured from the trough of the preceding N75 to the peak of P100.

(12.5%) with increased intraocular pressure (IOP) in the primary position or on upgaze and lateral gaze. By performing a cover test and drawing a Hesss screen, we observed that 40 patients (45.4%) had altered extraocular muscle function. Both the measurement of abnormal IOP and the finding of muscle dysfunction reflect an abnormality of extrinsic ocular motility and, therefore, were considered together for statistical analysis. We performed orbital CT scan in the presence of evident altered ocular motility and found increased eye muscles diameters in 41 of the 46 orbits studied, but not compression or abnormalities of the optic nerve at the orbital apex. Opthalmoscopy revealed a normal nerve head. Optic nerve function was studied by assessing color vision and performing a computerized visual field study. Only 1 patient had dyscromatopsia, whereas 23 (26.7%) had visual field defects. These were evidenced as paracentral scotomas (22 eyes; 19 patients) or as constriction of field isopters (7 eyes; 4 patients), without apparent significant distribution in the visual field. All patients had, at the time of the examination, normal best-corrected visual acuity.
VECP study

Statistical analysis
We used the t test for analysis of amplitude and latency values between the groups of patients and normal subjects and between the groups of patients with and without the various clinical ophthalmological signs. We compared the prevalence of a positive VECP test in the groups of patients with and without clinical signs by 2 analysis. Values are reported as the mean se.

Results Ophthalmological findings

At ophthalmological examination, eyelid signs, including lid lag and/or retraction (fissure, 11 mm) and lid edema were present in 70 patients (79.5%). Proptosis, with a Hertel measurement greater than 21 mm, was present in 54 patients (61.3%), of whom 43 had bilateral involvement. Signs of soft tissue inflammation, including lid edema, conjunctival injection, and/or chemosis, epiphora, caruncle edema, and corneal stippling were found in 39 patients (44.3%). Pupillary reflex was normal in all patients. We found 11 TAO patients

The group of patients with TAO and that of normal subjects did not differ in age (Table 1), but differed in the female to male ratio (8.7:1 vs. 1.9:1). There were no differences in the mean amplitude of the P100 wave of TAO patients (10.2 0.3 V) and normal subjects (11.3 0.6 V; by t test, P NS; data not shown). As shown in Table 1, the mean latency of the P100 wave in patients was 105.6 0.5 ms, significantly higher than that in normal subjects (102.0 0.5 ms; by t test, P 0.00003). The difference in the latency of the VECP was significant even when values were analyzed for each separate eye. To determine whether thyroid function would affect the results of the VECP recordings, we recalculated the mean latency values in TAO patients divided according to thyroid status. In hypothyroid patients, latency was 105.6 1.8 ms; in hyperthyroid patients, it was 106.1 0.7 (by t test, P NS; not shown). Latency in euthyroid patients was 105.2 0.7 ms and did not differ from that in either hypo- or hyperthyroid patients, but, again, did differ from that in normal controls (P 0.0004; data not shown). We calculated the upper limit of the normal range for the latency values recorded in our group of normal subjects as 109.2 ms (mean 2 sd), and we considered a VECP test positive when the latency was 110.0 ms 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 normal controls whose latency ranged from 93.0 107.0 ms. An abnormal visual field was found in 14 of 33 eyes (42.4%) with a positive VECP test (Table 2), of whom 1 also had impaired

TABLE 1. Mean (SE) age and latency of VECP in patients with thyroid-associated ophthalmopathy (TAO) and normal subjects
Eyes studieda TAO (no. of eyes) Normals (no. of eyes) P (by t test)

Mean age (yr) Latency (ms)

OS OD OS OD OS OD

45.3 1.1 (172) 105.6 0.5 (172) 105.8 0.7 (87) 105.5 0.6 (85)

41.9 2.1 (56) 102.0 0.5 (56) 102.5 0.7 (28) 101.5 0.7 (28)

NS 0.00003 0.01 0.001

OS, Left eye; OD, right eye.

VISUAL EVOKED POTENTIALS IN TAO

TABLE 2. Clinical signs and visual field examination of patients with thyroid-associated ophthalmopathy and a positive VECP test (latency, 110 ms)
Patient no. Visual acuitya OS OD Pupillary examination Fundus Color vision Visual fieldb OS OD Latency (ms) OS

1029

OD

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
a b

20/20 20/20 16/20 20/20 20/20 20/20 20/20 10/20 20/20 20/20 20/20 20/20 20/20 20/20 20/20 20/20 18/20 20/20 20/20 20/20 20/20

20/20 20/20 16/20 20/20 20/20 20/20 20/20 20/20 20/20 20/20 12/20 20/20 20/20 20/20 20/20 20/20 20/20 18/20 20/20 20/20 20/20

nc n n n n n n n n n n n n n n n n n n n n

n n n n n n n n n n n n n n n n n n n n n

n n n n n n n n n n n n n n n n n n n abnormal n

n para para para para n NE NE n isop isop n n n n n n para n para para

n n para para para n n n n isop NE n n isop n n n para n n n

110 112 110 121 114 112 105d NE 114 112 113 116 129 109d 106d 111 115 129 117 110 119

107d 114 114 123 117 115 111 110 117 108d NE 112 119 114 110 109d 111 127 111 105d 115

Best corrected visual acuity. para, Paracentral scotoma; isop, constriction of field isopters; NE, not examined. c n, Normal. d Normal VECP test. TABLE 3. Mean (SE) latency of VECP in relation to the ophthalmological signs of thyroid-associated ophthalmopathy
Clinical sign No. of eyes studied (OSOD) Mean latency (msec) P (by t test)

Proptosis (21 mm) No proptosis Soft tissue inflammation No inflammation Altered extrinsic motility Normal motility Abnormal orbital CT scana Normal orbit Visual field defects Normal visual field
a

96 76 74 98 58 114 41 5 29 142

106.7 0.7 104.3 0.6 106.6 0.8 104.9 0.5 106.6 0.9 105.2 0.5 107.6 1.3 102.2 0.6 110.3 1.5 104.7 0.4

0.012 NS NS NS 0.000003

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

CT scan performed on 46 eyes.

color vision. All of these patients had normal fundus examination. Next, we studied the variations in VECP latency in the patients in relation to the clinical signs of TAO. These results are shown in Table 3. Values recorded in both eyes were studied together. Although we recorded a significantly higher latency in patients with proptosis greater than 21 mm (P 0.01), no difference was observed in patients with soft tissue inflammatory signs, in patients with altered extrinsic motility, and in patients who underwent orbital CT scan in relation to the presence of eye muscle enlargement. In patients with an abnormal visual field, mean latency was significantly higher than that in patients with a normal test, and the latter was significantly higher than that in controls (P 0.0004; not shown). We then studied the prevalence of a positive VECP test in

In TAO patients affected by ON, vision loss occurs insidiously in the context of a congestive inflammatory orbitopathy. Crowding of the orbital apex, which causes optic nerve compression, is suspected in the presence of restricted eyeball movements. In studies of large groups of patients it has been reported that marked proptosis, palpable lacrimal glands, increase in intraocular pressure on upgaze, and restriction of extraocular muscles motility were predictive of the evolution of orbitopathy to ON (2, 10). In the absence of visual loss, other signs have been suggested as possible indicators of ON development, such as changes in color vision and in the optic nerve head on ophthalmoscopy (2) and abnormal visual field examination (3, 4). Indeed, none of these signs alone has proven to be specific for diagnosis. Electrophysiological studies, such as VECP, are considered the most objective and sensitive method of detecting early optic nerve abnormalities (5). Existing reports of VECP studies in TAO are not consistent because of the different techniques employed (11). Recordings of pattern reversal VECP are considered more reliable than those obtained after stimulation with either strobe or pattern flash, which produce a change in luminance (5). It has been reported that the latency of VECP increases with age after the fifth decade, usually by 2 ms/decade (5). In the present study, the difference in the latency of the evoked cortical response recorded in patients

1030

SALVI ET AL.

JCE & M 1997 Vol 82 No 4

TABLE 4. Prevalence of a positive VECP test (latency, 110 ms) in patients with thyroid-associated ophthalmopathy in relation to the different eye signs
Ophthalmological sign No. of positive VECP tests (% of total OS OD) P (by 2 test)

Proptosis (21 mm) Yes No Soft tissue inflammation Present Absent Visual field defects Present Absent Orbital CT scan Abnormal Normal

27/96 (28.1) 6/76 (7.9) 21/74 (28.4) 12/98 (12.2) 14/29 (48.2) 19/142 (13.4) 12/41 (29.3) 0/5 (0.0)

0.005 0.025 0.0005 NS

compression or to the presence of an increased ocular tone, factors such as ischemic damage due to narrowing and cellular infiltration of the nerve vessel walls (3) or a demyelinating-like neuritis (18) may be advocated to explain the pathophysiology of impaired optic nerve conduction. In conclusion, we have shown that the study of VECP in patients with TAO reveals asymptomatic optic nerve dysfunction in the absence of deterioration of visual acuity. The test should be used in addition to visual field examination in the ophthalmological assessment of the disease. VECP measurements are performed within a short time and require little collaboration by the patient. A positive test should suggest to the clinician additional intraorbital imaging and close follow-up of patients even in the absence of optic nerve compression.
Acknowledgment
We thank Prof. Marco Cordella for helpful critical revision of the manuscript.

with TAO without clinical ON and controls was unrelated to age, as age did not differ in the two groups. Furthermore, the greater female to male ratio of patients compared to controls was not expected to affect the results of the study because females are known to have a slightly shorter P100 latency than males (12). Hypothyroidism has been reported to prolong the latency of VECP, and restoration of euthyroidism with l-T4 has been shown to reverse the values to normal (1315). Most of our patients were euthyroid or hyperthyroid at the time of the study, and these conditions are not known to affect the results of the VECP test (16, 17). In the few patients who were subclinically hypothyroid, mean latency did not differ from that recorded in the other groups of TAO patients. Thus, as we have also excluded from the study patients with ocular diseases that might affect the specificity of the VECP test, we believe that the prolongation of latency observed in patients with TAO is related to the presence of autoimmune orbitopathy. The absence of amplitude changes in TAO patients is in agreement with the studies of Wijngaarde and Van Lith (18) and Seta la et al. (11), but not with that of Tsaloumas et al. (19), who, in patients with clinical ON, found a more significant abnormality of the amplitude than the latency of both flash- and pattern reversal-stimulated VECP. On clinical evaluation, none of the patients had clinical ON, although abnormalities of the visual field were recorded in 26.7% of the cases, suggesting asymptomatic optic nerve involvement. This finding is consistent with previous reports suggesting that visual field defects in TAO patients are an early sign of ON even in the presence of normal visual acuity (2, 3). We also found increased P100 latency in a proportion of patients (23.8%) who did not show eye muscle abnormalities on CT scan or increased IOP or congestive inflammatory signs of orbitopathy, which are usually indicative of ON (2). In about 50% of the eyes with increased latency there were visual field defects, and this association was significant. In a proportion of the eyes studied we observed a discrepancy between visual field and VECP measurements that may derive from the different areas and sensitivities of retinal stimulation in the two tests. Interestingly, optic nerve dysfunction would not have been diagnosed in 11 patients (12% of all cases) without the VECP test. As in the patients of this study optic nerve dysfunction was not due to intraorbital

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. 95:15151521. 3. Gasser P, Flammer J. 1986 Optic neuropathy of Graves disease. Ophthalmologica. 192:2227. 4. Trobe JD, Glaser JS, Laflamme P. 1978 Dysthyroid optic neuropathy. Arch Ophthalmol. 96:1199 1209. 5. Chiappa KH, Ropper AH. 1982 Evoked potentials in clinical medicine. N Engl J Med. 306:1140 1150. 6. Bobak P, Friedman R, Brigell M, Goodwin J, Anderson R. 1988 Visual evoked potentials to multiple temporal frequencies. Arch Ophthalmol. 106:936 940. 7. Batch JA, Lepre F. 1990 Early diagnosis of Graves optic neuropathy using visual 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; 557564. 9. Halliday AM, Barrett G, Blumhardt LD, et al. 1975 The macular and paramacular subcomponents of the pattern evoked response. In: Lehmann D, Callaway E, eds. Human evoked potentials. Applications and problems. New York: Plenum Press; 135151. 10. Feldon SE, Muramatsu S, Weiner JM. 1984 Clinical classification of Graves ophthalmopathy. Identification of risk factors for optic neuropathy. Arch Ophthalmol. 102:1469 1472. 11. Seta la K, Raitta C, Va limaki M, Katevuo V, Lamberg B-A. 1992 The value of visual evoked potentials in optic neuropathy of Graves disease. J Endocrinol Invest. 15:821 826. 12. Stockard JJ, Hughes JF, Sharbrough FW. 1979 Visually evoked potentials to electronic pattern reversal: latency variations with gender, age, and technical factors. Am J EEG Technol. 19:171204. 13. Osterweil D, Syndulko K, Cohen SN, et al. 1992 Cognitive function in nondemented older adults with hypothyroidism. J Am Geriatr Soc. 40:325335. 14. Abbott RJ, OMalley BP, Barnett DB, Timson L, Rosenthal FD. 1983 Central and peripheral nerve conduction in thyroid dysfunction: the influence of l-thyroxine therapy compared with warming upon the conduction abnormalities of primary hypothyroidism. Clin Sci. 64:617 622. 15. Ladenson PW, Stakes JW, Ridgway EC. 1984 Reversible alteration of the visual evoked potential in hypothyroidism. Am J Med. 77:1010 1013. 16. Mitchell KW, Wood CM, Howe JW. 1988 Pattern visual evoked potentials in hyperthyroidism. 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 hyperthyroid patients before and after achievement of euthyroidism. J Endocrinol Invest. 15:749 753. 18. Wijngaarde R, Van Lith GHM. 1979 Pattern evoked potentials in endocrine orbitopathy. Doc Ophthalmol. 48:327332. 19. Tsaloumas MD, Good PA, Burdon MA, Misson GP. 1994 Flash and pattern visual evoked potentials in the diagnosis and monitoring of dysthyroid optic neuropathy. Eye. 8:638 645.