Thyroid autoantibodies in adults with acquired binocular diplopia of unknown origin

Patients with acquired adult-onset strabismus mainly present with binocular diplopia. Although cranial nerve palsies are reportedly the most common cause of binocular diplopia in adults, thyroid disease can also cause diplopia. In patients with thyroid-associated ophthalmopathy, upper lid retraction and proptosis are the most common initial findings, but diplopia could be the first manifestation. So far, there has been little information on the diagnostic value of thyroid autoantibodies in patients with strabismus. Therefore, we examined adults with acquired binocular diplopia from 2008 to 2016 and evaluated the presence of thyroid autoantibodies and the relationship between thyroid autoantibody status and clinical characteristics in adults with acquired binocular diplopia. Thyroid autoantibody tests were performed for all patients, unless other causes of diplopia were identified. Fifty one (39%) of 132 patients were positive for thyroid autoantibodies. In the thyroid autoantibody-positive (TAb+) group, microsomal autoantibodies, thyroid-stimulating hormone receptor antibodies, thyroglobulin antibodies, and thyroid-stimulating antibodies were observed in 30, 27, 12, and 7 patients, respectively. The vertical deviation and grade of duction limitation were greater in the TAb+ group. The presence of ocular torsion was 15.5% and 39.5% in the TAb− and TAb+ groups, respectively. Thyroid autoantibody evaluation may be helpful in adults with idiopathic acquired binocular diplopia.

comparison of clinical characteristics of thyroid autoantibody-positive patients according to the thyroid function. In the TAb+ group, thyroid function was normal in 32 patients and abnormal in 19 patients. There was no difference in age, sex, duration of diplopia, angle of deviation, or in duction limitation or ocular torsion between the two groups (Table 3). No differences were observed in the clinical characteristics according to positivity or negativity of each of the four autoantibody types. changes in eye movements during follow-up in patients with binocular diplopia according to thyroid autoantibody status. Of a total of 132 patients, 109 were followed up for more than 3 months (67 in the TAb− group and 42 in the TAb+ group). During a mean follow-up of 15.6 ± 18.1 months, two patients in the TAb+ group were newly diagnosed with Graves' disease and one was diagnosed with Hashimoto's thyroiditis. The imaging scans of one patient are presented in Fig. 1. The change in the horizontal deviation angle was 3.3 ± 3.5 PD in the TAb− group and 4.4 ± 5.1 PD in the TAb+ group, while the change in the vertical deviation angle was 1.2 ± 1.8 PD in the TAb− group and 5.9 ± 8.1 PD in the TAb+ group (P = 0.227 and P = 0.001, respectively). The change in grade of duction limitation was 0.1 ± 0.1 in the TAb− group and 0.2 ± 0.3 in the TAb+ group (P < 0.001, Table 4).

Variable Values Range
Age, years, mean ± SD 53.

Discussion
The prediction and prevention of acquired binocular diplopia in adults is a challenging task often. Acquired binocular diplopia has several causes, among which cranial nerve palsies are the most common causes. TAO is also a cause of binocular diplopia in adults and is most commonly associated with Graves' disease. However, it is also occasionally observed in patients with Hashimoto's thyroiditis, primary hypothyroidism, and thyroid cancer, and in patients who have undergone radiotherapy for the neck region 13 . If a patient with acquired binocular diplopia has been previously diagnosed with any of these thyroid diseases, it would be helpful to identify TAO as the cause of diplopia. If there is no history of thyroid disease, thyroid function and thyroid autoantibody tests are usually performed for the diagnosis of TAO. Autoantibodies against thyroid gland antigens and the orbit might cross-react with both tissues. There are conflicting reports in the literature regarding the value of thyroid autoantibody tests in patients with thyroid disease 9,14-18 . Thyroid-stimulating antibody testing has been reported to be helpful for making a diagnosis of TAO in euthyroid patients and for the assessment of its severity 9,15 . In addition, some researchers asserted that new treatment modalities such as specific monoclonal antibodies, TSH-R antagonists, and other immunomodulatory agents show a promising outcome for Graves' ophthalmopathy patients 19 . However, another study showed that the thyroid-stimulating antibody was not a marker for Graves' ophthalmopathy 14 . Recently, Lin et al. 20 reported that there were no significant predictors of strabismus following a diagnosis of TAO. Despite these controversial reports, thyroid autoantibodies are still considered valuable in the diagnosis and management of Graves' disease 9,12,15,21 . However, most of the previous studies included patients with Graves' disease and patients who had already been diagnosed with TAO, and their main clinical manifestations were exophthalmos and lid retraction. www.nature.com/scientificreports www.nature.com/scientificreports/ Moreover, risk factors such as cigarette smoking 22,23 , and exposure to radioactive iodine have been linked to development of TAO, but none of the risk factors have been clearly associated with strabismus 24 .
TAO usually causes a restrictive, incomitant, and vertical strabismus. Therefore, in patients without other clinical sign of TAO, restrictive pattern could be the clue to suspect the strabismus caused by thyroid disease. In addition, the most common affected muscles in TAO are the inferior and medial recti. Since vertical rectus muscle involves the ocular torsion, measurement of ocular torsion is helpful to manage the patients with TAO 25,26 . The present study included patients with binocular diplopia. After excluding the patients with a clear etiology of diplopia, 38.6% of patients were positive for thyroid autoantibodies. These patients had a greater angle of vertical www.nature.com/scientificreports www.nature.com/scientificreports/ deviation and a higher duction limitation grade and were more likely to have ocular torsion. This finding suggests that patients who tested positive for thyroid autoantibodies could present the strabismus seen in TAO.
Abnormal thyroid hormone levels were also more common in the TAb+ group. However, 62.7% of these patients had normal thyroid function. Although there was no difference in the clinical characteristics among euthyroid patients and those with abnormal thyroid function in the TAb+ group, the vertical angle of deviation and duction limitation grade increased during the follow-up. This suggests that patients who tested positive for thyroid autoantibodies should be followed up regularly, even if their thyroid function tests appear normal, in order to detect a change in the deviation angle or duction limitation. For example, a change in the deviation angle and thickening of the extraocular muscles were observed in the patient shown in Fig. 1. Our findings suggest the possibility of changes in the clinical condition in patients who are positive for thyroid autoantibodies. Therefore, thyroid autoantibodies could be helpful markers for the differential diagnosis of binocular diplopia in adults.
This study had some limitations, resulting mainly from its retrospective design. For example, patients lacking medical records were excluded. We were also unable to obtain information on the smoking history. Although all the ophthalmologic examinations were performed by one experienced physician, duction was subjectively graded. Moreover, the possibility of a false-positive or false-negative result of thyroid antibody assays should be considered. In terms of types of Ab assay, we could not find the differences in the clinical characteristics among four types of thyroid autoantibodies. Prospective studies with larger populations, including patients with a clear etiology of binocular diplopia and long-term follow-up are needed in the future.
Nevertheless, to the best of our knowledge, this was the first study to screen a considerably large number of adults for acquired binocular diplopia and report the relationship of the thyroid autoantibody status with strabismus. The angle of vertical deviation and grade of duction limitation were greater in the TAb+ group. Ocular torsion and thyroid dysfunction were more common in the TAb+ group. We assumed that there would be changes in strabismus in the patients in the TAb+ group, which necessitates regular follow-up. In conclusion, thyroid autoantibody status could be helpful for the diagnosis of patients with binocular diplopia.

Methods
Subjects. We retrospectively reviewed the medical records of adults (aged above 17 years) who were diagnosed with acquired binocular diplopia at Seoul National University Hospital between January 2008 and July 2017. Patients who had undergone ocular surgery for cataract or refractive surgery, patients with a history of ocular trauma, and those with underlying disease, such as myasthenia gravis or chronic progressive external ophthalmoplegia were excluded. Patients with strabismus caused by a brain lesion, such as a brain tumour, brain haemorrhage, or microvascular lesion, patients who were diagnosed with strabismus before the age of 17 years, and patients in whom the age of onset of diplopia was not documented were also excluded. The patients enrolled underwent laboratory investigations and brain and orbit magnetic resonance imaging. The study was approved by the Institutional Review Board (IRB) of Seoul National University Hospital in South Korea and the study protocol followed the tenets of the Declaration of Helsinki. The IRB waived the need for informed consent as it was retrospective chart review study. thyroid autoantibody tests. Patients underwent thyroid function and thyroid autoantibody tests, unless other causes of diplopia were identified. Thyroid function tests included measurement of T3, T4, free T4, and TSH levels and were divided into four categories; hyperthyroidism, subclinical hyperthyroidism, euthyroidism, and hypothyroidism. Thyroid autoantibody tests were performed in all patients using the same assays: microsomal autoantibodies using chemiluminescent microparticle immunoassay (normal reference value 0-60 U/ml), TSH receptor antibodies using electrochemiluminescence immunoassay (normal reference value 0-1 IU/L), thyroglobulin autoantibodies using electrochemiluminescence immunoassay (normal reference value 0-100 U/ ml), and thyroid-stimulating antibodies using bioimmunoassay. In case of thyroid-stimulating antibodies, the specimen-to-reference ratio (SRR%) was calculated according to the following formula: (mean specimen TSAb/ mean reference TSAb)*100. Patient serum was considered positive if SRR was ≥ 140%. We divided patients Ophthalmologic examination. At the initial visit, all patients underwent manifest and cycloplegic refraction. The refractive error was converted to the spherical equivalent value [in dioptres (D)]. Myopia was represented as negative numbers and hyperopia as positive numbers. Amblyopia was defined as a difference of two lines or more in Snellen acuity and excluded in this study. The angle of deviation was measured by alternate prism cover test with accommodative targets for fixation both at near (0.33 m) and at distance (6 m) and indicated as prism diopters (PD). The horizontal and vertical deviation angles were analyzed separately after measurement. Limitation of duction was measured using a five-point scale (0 to −4), from "0" (indicating no limitation with full duction) through "−3" (indicating an eye, which could reach the midline) to a maximum of "−4" (representing no movement beyond the midline, with the eye fixed in the opposite position). Presence of inferior/superior oblique muscle dysfunction such as overaction and underaction was also evaluated. All ophthalmologic examinations were performed by an experienced ophthalmologist (SJ-K).
Objective ocular torsion was assessed using fundus photography. Patients were asked to look at an internal fixation target in order to align their eyes at the primary position. Based on the position of the fovea, we considered the eye to be in extorsion when the fovea was located below the horizontal line at the lower edge of the optic disc. We considered the eye to be in intorsion, when the fovea was located above the horizontal line passing through the center of the optic disc 27 . Physiological torsion was defined by the fovea laying in the area between the center and the lower edge of the optic disc. The present study analyzed the presence of pathological ocular torsion.

Statistical analysis.
As the data showed normal distribution by Kolmogorov-Smirnov test, parametric methods including the independence t-test, Fisher's exact test were used. Continuous variables are reported as mean ± standard deviation. P < 0.05 was considered statistically significant for all the tests. Statistical analyses were performed using SPSS software (version 23 for Windows; SPSS, Chicago, IL, USA).

Data availability
Data supporting the findings of the current study are available from the corresponding author on reasonable request.