3.3.10 Graves’ ophthalmopathy and dermopathy

The many and often disfiguring features of a typical patient with Graves’ ophthalmopathy are obvious at first glance (Fig. 3.3.10.1). The changed appearance of the patient has a profound effect on their emotional and social status. The various signs and symptoms can be described according to the NO SPECS classification (1) (Box 3.3.10.1). Class 1 signs can be present in any patient with thyrotoxicosis regardless of its cause. Upper eyelid retraction causes stare and lid lag on downward gaze (the latter is the well-known von Graefe’s sign). Soft tissue involvement (class 2) comprises swelling and redness of eyelids, conjunctiva, and caruncle. Symptoms are a gritty sandy sensation in the eyes, retrobulbar pressure, lacrimation, photophobia, and blurring of vision. Proptosis (class 3) can be quite marked. Upper eyelid retraction by itself may already give the impression of exophthalmos. Extraocular muscle involvement (class 4) may result in aberrant position of the globe, or fixation of the globe in extreme cases. More common is limitation of eye muscle movements in certain directions of gaze, especially in upward gaze; it is usually associated with diplopia. Diplopia will not occur if the vision of one eye is very low (e.g. in amblyopia), or if the impairment of eye muscle motility is strictly symmetrical. Patients may correct for double vision by tilting the head, usually backwards and sideways; the ocular torticollis often leads to neck pain and headache. Corneal involvement (class 5) occurs through overexposure of the cornea due to lid lag, lid retraction, and exophthalmos, easily leading to dry eyes and keratitis. Lagophthalmos is often noted first by the patient’s partner because of incomplete closure of the eyelids during sleep. Sight loss (class 6) due to optic nerve involvement is the most serious feature, often referred to as dysthyroid optic neuropathy (DON). Besides the decrease of visual acuity, there may be loss of colour vision and visual field defects. Visual blurring may disappear after blinking (caused by alteration of the tear film on the surface of the cornea due to lacrimation or dry eyes) or after closing one eye (attributable to eye muscle imbalance). Visual blurring that persists is of great concern as it may indicate optic neuropathy (2).

The frequency of the various eye changes among patients with Graves’ ophthalmopathy is as follows (3): von Graefe’s sign 59%, upper eyelid swelling 75%, proptosis of 21 mm or higher 63%, diplopia 49%, impairment of elevation 49%, impairment of abduction 32%, impairment of depression 17%, corneal involvement 16%, and optic nerve involvement 21%. Predisposing factors for DON are male sex, old age, smoking, and diabetes mellitus (4, 5) (Fig. 3.3.10.2). Diabetes is present in 9%, and glaucoma or cataract in 14%. Unilateral eye disease is observed in about 10% of patients. Eye changes are similar to those of patients with bilateral eye disease, but unilateral cases are more often euthyroid. Eye muscle enlargement of the fellow eye can be detected by imaging in about one-half of cases. Progression to bilateral eye disease is common (6). Unilateral Graves’ ophthalmopathy may thus represent an early stage of the disease that already is or will develop shortly into a bilateral disease. Unknown local factors must be involved in the unilateral expression of Graves’ ophthalmopathy, which essentially is a bilateral and fairly symmetrical eye disease.

A population-based cohort study in Olmsted County, Minnesota, USA reports an overall age-adjusted incidence rate of Graves’ ophthalmopathy of 16.0 women and 2.9 men/100 000 inhabitants per year (7). The incidence rate exhibits an apparent bimodal peak in the fifth and seventh decades of life. Male sex and older age are associated with more severe ophthalmopathy (4, 5). Childhood Graves’ ophthalmopathy is rare (8, 9). Clinical manifestations are less severe in paediatric patients: exophthalmos is seen in 75%, but impaired muscle motility only in 11%.

Smoking greatly increases the risk for Graves’ ophthalmopathy (odds ratio 7.7, 95% CI 4.3 to 13.7) (10). Smokers have more severe eye disease than nonsmokers (Fig. 3.3.10.3). A trend to a lower incidence rate of Graves’ ophthalmopathy is reported. In a single centre in the UK, the proportion of patients with Graves’ ophthalmopathy among all referred patients with Graves’ hyperthyroidism decreased from 57% in 1960 to 35% in 1990; there was also a decline in the prevalence of severe Graves’ ophthalmopathy (diplopia or DON) from 30% to 21% (11). In a European questionnaire study in 1998, 43% of respondents thought Graves’ ophthalmopathy was decreasing in frequency, 42% thought it unchanged, and 12% thought it to be increasing (12). In this respect it is noteworthy that all responders from Hungary and Poland, where the prevalence of smoking had increased since 1990, indicated an increased incidence of Graves’ ophthalmopathy. The trend to a lower incidence rate of Graves’ ophthalmopathy might therefore be causally linked to a secular decrease in the prevalence of smoking. Alternatively, earlier diagnosis and treatment of Graves’ hyperthyroidism could be involved in view of the introduction of sensitive thyroid-stimulating hormone (TSH) assays in the late 1980s.

Graves’ ophthalmopathy is usually but not invariably associated with Graves’ hyperthyroidism. Graves’ hyperthyroidism is present in about 90% of patients with Graves’ ophthalmopathy, autoimmune hypothyroidism in 3%, and euthyroidism in 7% (3). The euthyroid and primarily hypothyroid patients have milder and more asymmetrical Graves’ ophthalmopathy than the hyperthyroid Graves’ ophthalmopathy patients (13). It is not unusual that hypothyroid Graves’ ophthalmopathy patients proceed to Graves’ hyperthyroidism, linked to a shift from TSH-receptor blocking to stimulating antibodies. Euthyroid Graves’ ophthalmopathy patients develop hyperthyroidism in due time in about 20%, but it is unknown why others remain euthyroid although TSH-receptor stimulating antibodies can be detected in almost everyone (14).

Among patients with both Graves’ ophthalmopathy and Graves’ hyperthyroidism, the eye disease becomes manifest before the onset of hyperthyroidism in about 20% (the interval can be months to years), concurrent with hyperthyroidism in 40%, and after the onset of hyperthyroidism in 40% (2). Autoimmune thyroid disease is therefore present in most if not all patients with Graves’ ophthalmopathy. Conversely, Graves’ ophthalmopathy is present in the majority if not all patients with Graves’ hyperthyroidism, although over one-half of patients with Graves’ hyperthyroidism have no clinically appreciable ophthalmopathy. Evidence for subclinical ophthalmopathy is, however, found in most patients with Graves’ hyperthyroidism without apparent eye changes, encompassing a shift to higher proptosis values in this group as compared with healthy controls, an abnormal increase in intraocular pressure on upgaze in 61%, and enlarged extraocular muscles on ultrasound or CT scan in 70–100% (2).

The available data strongly support the view that the eye and thyroid manifestations belong to the same disease entity, Graves’ disease, a multiorgan autoimmune disorder. Factors determining the differential expression of the disease remain largely unknown, but smoking is definitely involved.

Graves’ ophthalmopathy is characterized by enlargement of extraocular muscles and retro-ocular connective/adipose tissue. The increase in tissue volume and the associated rise of retrobulbar pressure can explain the various signs and symptoms (2). The swollen retrobulbar tissues will impair the venous drainage of the eyelids and conjunctiva, resulting in oedematous swelling of the eyelids and chemosis. Upper and lower eyelid swelling can also be caused by herniation of retrobulbar fat through openings in the orbital septum (Fig. 3.3.10.4a, b). The only other outlet for the increased orbital content, in view of the confinement within the bony walls of the orbit, is pushing forward the eyeball, resulting in exophthalmos (Fig. 3.3.10.4a, b). The volume of the orbital cavity is approximately 26 ml, and of the combined extraocular muscles about 3.3 ml. A three- to fourfold increase in size of extraocular muscles is observed in severe cases, and an increase of 4 ml in muscle volume will cause a proptosis of 6 mm (15).

The enlargement of extraocular muscles impairs muscle relaxation, not the ability for muscle contraction. Limited motion of eye muscles is due to impaired relaxation of the antagonist upon contraction of the agonist. Impaired elevation is thus primarily the result of insufficient relaxation of the rectus inferior muscle. This can be appreciated by the forced duction test: by actually grasping the globe and attempting to move it in the affected direction, mechanical resistance is felt. Restricted eye muscle motility may cause diplopia. Upper eyelid retraction (due either to an increased adrenergic activity in hyperthyroidism or to swelling of the levator muscle) and proptosis contribute to overexposure of the cornea, which may become dry and inflamed.

Marked swelling of the extraocular muscles in the apex of the orbit (known as apical crowding), close to the entrance of the optic nerve in the optic canal, may damage the optic nerve either via direct pressure or via impairment of the blood supply to the nerve. The resulting optic neuropathy causes loss of visual functions. The degree of proptosis in patients with optic neuropathy, despite a greater mass of extraocular muscles, is less than in patients without optic neuropathy (16). This is remarkable because the retrobulbar pressure (in the normal orbit 3.0–4.5 mmHg) is greatly elevated in patients with DON up to values between 17 and 40 mmHg, much higher than the pressure of 9–11 mmHg measured in orbits of patients with exophthalmos but no optic neuropathy (17). A well-developed tight orbital septum might preclude proptosis, resulting in a very high retrobulbar pressure and optic neuropathy (Fig. 3.3.10.4c,d).

The macroscopic appearance of the orbital content in Graves’ ophthalmopathy is dominated by enlargement of extraocular muscles and to a lesser extent of retrobulbar fat and connective tissue. Microscopy demonstrates lymphocytic infiltration, oedema, and fibrosis. The swelling of tissues is largely caused by an increase of ground substance consisting of collagen and glycosaminoglycans. Glycosaminoglycans, because of their hydrophilic nature, attract water, resulting in oedematous swelling. The ground substance accumulates in the endomysial space between the muscle fibres. There is no increase in the number of muscle fibres and no ultrastructural damage to the muscle cells themselves (except in very advanced cases when some damage may be seen). An increased number of fibroblasts is found in the endomysial space and in the connective and adipose tissues. The fibroblasts are responsible for the excessive production of glycosaminoglycans.

The lymphocytic infiltrate is often focal, and consists of T helper cells, suppressor/cytotoxic T cells, many macrophages, and relatively few B cells (18). There is abundant expression of HLA-DR. Many of these cells are activated memory cells (CD45RO⁺), frequently located adjacent to blood vessels. The lymphocytic infiltration diminishes in the late inactive stage of the disease. The infiltrating immunocompetent cells produce cytokines capable of remodelling orbital tissues. The cytokine profile in the early stages is predominantly derived from T helper 1 cells, whereas, in patients with a duration of Graves’ ophthalmopathy longer than 2 years, cytokines are mostly derived from T helper 2 cells (19). The data suggest Graves’ ophthalmopathy is primarily a T-cell mediated disease. In keeping with this notion is the observation that, in contrast to neonatal Graves’ hyperthyroidism, no single convincing case of neonatal Graves’ ophthalmopathy has been reported (whereas immunoglobulins cross the placenta, T cells do not). The cytokines induce expression of immunomodulatory proteins on orbital endothelial cells and fibroblasts, such as HLA-DR, heat shock protein 72, and several adhesion molecules. Cytokine-activated orbital fibroblasts synthesize chemoattractants IL-16 and RANTES, generating T-cell migration. More T cells migrate to the orbit, perpetuating the immune attack. Macrophages may present antigen to T cells (CD40L) through provision of costimulatory signals and proinflammatory cytokines. Activated T cells may bind to CD40⁺ fibroblasts inducing hyaluronan synthesis, cytokines, COX2, and PGE2. Orbital fibroblasts are considered the target cells of the autoimmune attack. Retrobulbar T cells from patients with Graves’ ophthalmopathy recognize autologous orbital fibroblasts (but not eye muscle extracts) in a major histocompatibility complex (MHC) class I restricted manner, and proliferate in response to autologous proteins from orbital fibroblasts (but not from orbital myoblasts). Conversely, orbital fibroblasts proliferate in response to autologous T cells dependent on MHC class II and CD40-CD40L signalling (18).

Orbital fibroblasts have site-specific characteristics. Orbital fibroblasts expressing Thy-1 (present in orbital fat and muscles) produce more PGE2, whereas fibroblasts not expressing Thy-1 (present only in orbital fat) may differentiate into mature adipocytes (e.g. when incubated with IL-1 or PPARγ agonists). The process of adipogenesis contributes to volume expansion, but interestingly is also associated with increased TSH-receptor expression (20). This brings us to the nature of the autoantigen in the orbit.

The TSH receptor is presently viewed as the major autoantigen in Graves’ ophthalmopathy. Orbital fibroblasts express full-length functional TSH receptors; the expression is more abundant in active than in inactive disease and is directly related to IL-1β (21). Graves’ immunoglobulins recognize TSH receptors on orbital fibroblasts as evident from increased cAMP and hyaluronan production in cell cultures of differentiated orbital fibroblasts (18). Clinical studies support the role of TSH receptors. The serum concentrations of TSH-receptor antibodies are higher in patients with Graves’ ophthalmopathy than in Graves’ patients without eye changes (22), and are related to the severity and activity of the ophthalmopathy (whereas thyroid peroxidase and thyroglobulin antibodies are not) (23, 24). The higher the level of TSH-receptor antibodies, the higher the risk for an unfavourable course of the eye changes (25). However, experimental studies in which animals were immunized against the TSH receptor, observed hyperthyroidism in some animals but so far never ophthalmopathy. Another candidate is the IGF-1 receptor. Older studies reported inhibition of [¹²⁵I] IGF-1 binding to orbital fibroblasts by Graves’ IgG, and more recent studies indicate Graves’ IgG recognize and activate IGF-1 receptors on orbital fibroblasts inducing synthesis of IL-1β and hyaluronan (26). Further studies propose colocalization of TSH and IGF-1 receptors to cell membranes (27). Expression of IGF-1 receptors is increased in orbital fibroblasts and peripheral blood T cells of patients with Graves’ disease (28). These intriguing data remain to be confirmed, but may well indicate involvement of fibrocytes in the extrathyroidal manifestations of Graves’ disease. Antibodies against eye muscle antigens (such as calsequestrin) are likely secondary responses to tissue destruction and release of sequestered proteins (29, 30).

Taken together, Graves’ ophthalmopathy starts with the accumulation of immune cells in the orbit, giving rise to cytokine release and up-regulation of TSH receptors on orbital fibroblasts, the target cells of the autoimmune attack. The cellular response of fibroblasts includes glycosaminoglycan production and differentiation of preadipocytes into mature adipocytes (adipogenesis). Consequently, the intraorbital pressure rises due to volume expansion, and mechanical trauma by itself will further attract immune cells (Fig. 3.3.10.5). This sequence of events has aptly been called the cycle of disease (31).

Many questions, however, remain unanswered. It is difficult to explain why many patients with Graves’ hyperthyroidism, despite high titres of TSH-receptor antibodies, do not develop Graves’ ophthalmopathy. The different phenotypes of Graves’ disease might be related to genetic and environmental factors. Graves’ ophthalmopathy is more prevalent in white patients than in Asian patients (32). Whereas several susceptibility genes for Graves’ disease have been identified, there is no difference in the frequency of particular polymorphisms in these genes (HLA, CTLA4, PTPN22, CD40, FRCL3, TSHR) between Graves’ hyperthyroid patients with and without ophthalmopathy (33). Some polymorphisms in genes encoding for intercellular adhesion molecule 1, interferon-γ, and tumour necrosis factor are more prevalent in Graves’ ophthalmopathy, enhancing slightly the risk for eye changes. In contrast, smoking increases greatly the risk of Graves’ ophthalmopathy by incompletely understood mechanisms. Orbital fibroblasts when cultured under hypoxic conditions produce more glycosaminoglycans (34). Exposure of orbital fibroblasts in vitro to cigarette smoke extract dose-dependently increases adipogenesis and hyaluronan production (35).

Graves’ ophthalmopathy has a tendency towards spontaneous improvement. There have been few studies on the natural history of the eye disease. The most extensive ones were carried out in the 1940s and 1950s by Rundle (36) (Fig. 3.3.10.6). He described a stage of ingravescence, characterized by the development of exophthalmos (by 0.5 mm monthly, up to an average extent of 2–5 mm) and limitation of elevation; 4–5 degrees elevation is lost for each millimetre of protrusion. Thereafter a stage of remission occurs, which is slower and less complete than ingravescence. Recovery from restricted eye muscle motility precedes that from proptosis. This dynamic phase is succeeded by a static phase, in which exophthalmos and eye muscle disturbance remain unchanged in 75% of patients. The time period in which the stable endstage is reached varies considerably between patients, ranging from several months up to 5 years. Despite spontaneous improvement, the eye changes do not completely disappear in about 60% of patients (37). Recent studies confirm these earlier observations: during a 1-year follow-up in patients whose ophthalmopathy did not require immediate treatment, substantial improvement occurred in 22%, slight improvement occurred in 42%, the disease remained stable in 22%, and the disease progressed in 14% (38).

The few histological studies support Rundle’s observations. In the early active stage of the disease there is usually a lymphocytic infiltrate, oedema, and activated fibroblasts; in the endstages there is only fibrosis. The data imply that the natural history of Graves’ ophthalmopathy can also be described according to the activity of the eye disease, as well as Rundle’s curves depicting the severity of the eye disease (39) (Fig. 3.3.10.7). Assessment of the activity of the eye disease may influence the treatment plan. Immunosuppression is unlikely to be effective when given in the fibrotic inactive endstage of the disease, but might be of much benefit in the early active stage with ongoing inflammation. Likewise, the results of eye muscle and lid surgery might be lost when performed in the active stage.

The initial work-up is aimed at delineating the optimal treatment plan (40). Treatment will depend on the severity and activity of the eye disease and its influence on the quality of life of the patient. Timing of a particular treatment option may depend on thyroid function. If the diagnosis of Graves’ ophthalmopathy is uncertain, several other conditions must be considered.

Close monitoring of thyroid function is relevant because the eye changes are more severe in patients who still have an abnormal thyroid function (41). Serum concentrations of TSH-receptor antibodies are usually high in Graves’ ophthalmopathy, and their presence may support the diagnosis of Graves’ ophthalmopathy in euthyroid patients.

The patient is usually reluctant to voice cosmetic complaints and will at first focus on symptoms such as gritty feeling, photophobia, excessive lacrimation, and blurred vision. Diplopia may be present only when fatigued. Typically, the patient will report difficulty in reading or watching television (subtitles) in the evening. Diplopia can be present only in certain directions of gaze, e.g. hampering driving because of double vision when looking in the rear-view mirror. When diplopia is present in the primary position of gaze, it will severely affect daily activities. Severe asymmetrical motility disorders can be masked by a ocular torticollis. Typically, the head is tilted backwards, which often results in headaches originating from tense neck muscles. Positioning of the head in the default position will unmask diplopia.

The NO SPECS system is a useful memory aid in the physical examination of eye signs (Table 3.3.10.1). Quantitative measurements are preferred whenever possible. Lid aperture is measured with the ruler centred on the pupil while the facial muscles are relaxed and gaze is directed straight ahead. The degree of soft tissue swelling is difficult to assess objectively, but using a comparative photographic colour atlas is very helpful in this respect (42). Proptosis is readily quantified with an exophthalmometer. The normal range in white people is 12–20 mm, in African-Caribbean people is 14–22 mm, and in Asian people is 10–18 mm. Significant exophthalmos may develop without proptosis readings exceeding the upper normal limit; consequently, a normal reading does not always exclude exophthalmos. Comparison with pictures of the patient before the onset of disease can be helpful in this respect.

Motility disturbances are recognized by asking the patient to move the eyes upwards, downwards, and from side to side. Ductions in the four directions of gaze (elevation, depression, abduction, and adduction) are age dependent, and can be measured by a Goldman or modified hand perimeter (43). Diplopia can be graded subjectively. A more objective description of diplopia in the various positions of gaze is obtained by the Hess chart or Lancaster red–green test and by measuring the field of single binocular vision on a Goldman perimeter.

The cornea is investigated by split lamp and application of dyes. Stippling of the cornea may be seen, and in more severe cases ulceration, clouding, necrosis, and perforation.

Visual acuity can be measured using the Snellen chart; the use of a pinhole may correct for refraction disorders and sight loss due to keratitis. Not all patients with optic neuropathy have decreased visual acuity. Further investigation is warranted if optic nerve involvement is suspected, e.g. in patients complaining of persistent blurred vision or greyish vision. In patients with DON, decreased visual acuity is present in 80%, reduced colour vision in 77%, visual field defects in 71%, afferent pupillary defect (Marcus Gunn’s phenomenon) in 45%, optic disc oedema in 56%, and disc pallor in 4% (44). Choroidal folds, caused by impression of the globe by retrobulbar tissues, are rare.

In some patients, such as the one depicted in Fig. 3.3.10.2, one glance is sufficient to conclude that the eye disease is in the active phase. However, if it is not self-evident whether the eye disease is active or inactive, an assessment of disease activity provides guidance in selecting the most appropriate treatment (Fig. 3.3.10.7). The gold standard for activity could be histological examination of retrobulbar tissues, but retrobulbar biopsies are not feasible in daily practice. Consequently, the response to immunosuppression has been introduced as a surrogate standard. Any parameter proposed to indicate disease activity should accordingly be validated by its predictive value for the outcome of immunosuppression, thereby serving its clinical purpose.

No change of the eye signs over a period of 6 months is generally considered proof that the eye disease is inactive. A duration of the ophthalmopathy of less than 18 months (as obtained from the history of the patient) is likely to be associated with active disease (45). A clinical activity score of 4 or above, based on the classic signs of inflammation, indicates active disease (46) (Table 3.3.10.2, Fig. 3.3.10.8).

Tissue oedema prolongs the T₂ relaxation time of MRI. A prolonged T₂ relaxation time in extraocular muscles (>130 ms) is associated with a favourable response to immunosuppression (47). A-mode echography depicts the internal echogenicity of eye muscles, which may be low in oedematous and high in fibrotic muscles. A value below 30% in the muscle with the lowest reflectivity has modest predictive value (45). ¹¹¹In-labelled octreotide scintigraphy may reveal significant uptake in the orbital region. It may be explained by the binding of octreotide to somatostatin receptors expressed on activated lymphocytes and fibroblasts. A positive orbital octreotide scan would thus indicate active eye disease. An orbital/occipital skull uptake ratio of more than 1.85 is associated with a favourable response to immunosuppression (48, 49).

Serum TSH-receptor antibodies interpreted in relation to the duration of eye changes, help to predict a favourable or unfavourable course of Graves’ ophthalmopathy (25). Serum concentrations of particular cytokines, cytokine receptors, and adhesion molecules differ between Graves’ ophthalmopathy patients and controls, and between patients with active and inactive eye disease. However, the overlap between the groups is too large for meaningful application in individual patients. The same holds true for urinary glycosaminoglycans.

None of the clinical and radiological parameters can be used on their own to make therapeutic decisions about the value of administering immunosuppression. Most parameters have a good positive but rather low negative predictive value for the outcome of immunosuppression, except MRI which has a high negative but low positive predictive value. A combination of tests might optimize distinction between active and inactive eye disease. The drawback of orbital Octreoscan (besides being very expensive) and ultrasonography is the high operator dependency of these methods. For the time being, the combination of duration of the eye disease, clinical activity score, and T₂ relaxation time on orbital MRI seems to be the most reliable and cost-effective manner for assessing disease activity.

The degree of swelling of extraocular muscles and orbital fat can be evaluated using CT scans or MRI. Coronal sections are preferred in view of the pear-shaped orbit with its axis directing backwards and medially. The bony structures are best evaluated by CT scan and are of relevance in case of surgical decompression. MRI has the advantage of providing an activity parameter in addition to imaging. The muscles are swollen typically at the belly, leaving the tendons uninvolved (Fig. 3.3.10.4). For unknown reasons the inferior and medial rectal muscles are most frequently enlarged, followed by the superior rectus; the lateral rectus muscle is least affected (2). The extraocular muscles originate in Zinn’s annulus, which surrounds the optic canal and thus the optic nerve. Muscle swelling at this location (apical crowding) and intracranial fat prolapse are risk factors for optic neuropathy (44, 50) (Figs. 3.3.10.4 and 3.3.10.9).

Eye changes adversely affect a patient’s self-image and daily functioning. The overall health-related quality of life of patients with moderately severe Graves’ ophthalmopathy is lower than for patients with other chronic conditions such as diabetes mellitus, emphysema, or heart failure (51). Because the goal of treatment in Graves’ ophthalmopathy is to improve daily life and to make patients feel better rather than to prolong life, clinical measures are often surrogate outcomes for what we really want to measure, the effect of treatment on patients’ lives. Consequently, self-assessment of the eye condition is recommended. A disease-specific Graves’ ophthalmopathy quality of life questionnaire has been developed, the GO-QOL (52, 53). It contains eight questions about problems with visual functioning and eight questions about the psychosocial consequences of changed appearance. The answers are summarized to one score for visual functioning and one score for appearance. The GO-QOL might be useful not only in evaluation of treatment, but also in reconciling priorities of the patient and of the physician in delineating a management plan. The psychosocial burden imposed by Graves’ ophthalmopathy is considerable, and support by mental health care professionals might be needed (54).

The diagnosis of Graves’ ophthalmopathy can be quite easy in patients with typical bilateral eye signs and, past or present, Graves’ hyperthyroidism. However, it can be difficult in euthyroid patients or in unilateral eye disease. The finding of thyroid autoantibodies may provide circumstantial evidence for the autoimmune nature of the ophthalmopathy. Orbital imaging in unilateral eye disease is mandatory to exclude other conditions, although Graves’ ophthalmopathy is the single most common cause of unilateral proptosis representing 15% of cases.

None of the eye signs is pathognomonic for Graves’ ophthalmopathy. Lid retraction can be due to non-Graves’ thyrotoxicosis, contralateral ptosis, or cocaine use. Diplopia is common in myasthenia gravis, and bilateral proptosis can be caused by orbital fat accumulation (Cushing’s syndrome, obesity), lithium therapy, liver cirrhosis, Wegner’s granulomatosis, arteriovenous malformations, lymphoma, metastatic tumours, or severe myopia (pseudoproptosis).

Proptosis, motility disturbances, and optic nerve compression can be caused by the ill-defined disease entity of orbital pseudotumour, an idiopathic unilateral focal or diffuse fibroinflammatory orbital lesion (55). The clinical presentation is characterized by acute or subacute signs of inflammation (pain, redness, oedema) and mass effects; CT shows a focal or diffuse mass which is poorly demarcated. Orbital myositis is, after Graves’ ophthalmopathy, the second most common cause of extraocular muscle enlargement, due to nonspecific inflammation perhaps of autoimmune aetiology. The cardinal clinical feature is acute orbital pain exacerbating on eye movements. The disease is bilateral in 50% of patients. One or more muscles may be enlarged and, in contrast to Graves’ ophthalmopathy, involve the anterior muscle tendons, as evident from CT scans. Orbital pseudotumour and orbital myositis respond quickly to corticosteroids, but recurrences occur in 50% of patients.

Management of Graves’ ophthalmopathy requires close consultation between the patient, the endocrinologist, and the eye physician, but in this respect notable differences in the delivery of care exist (56). Combined thyroid–eye clinics are most appropriate to delineate the treatment plan best suited for a particular patient. The timing and mode of thyroid treatment, immunosuppressive treatment, and surgical treatment should be coordinated in a multidisciplinary approach. The patient should be reassured of the possibilities for improvement of eye changes, both functionally and cosmetically, but at the same time be informed that it may require 1–2 years before full rehabilitation is reached. In this respect the experience of other patients with Graves’ ophthalmopathy can be quite informative, and contact with patient self-help groups is very useful. The European Group on Graves’ Orbitopathy (EUGOGO) has published a consensus statement on management (57). An overview is given in Box 3.3.10.2.

The advice to stop smoking should be given repeatedly. Progression of eye changes after ¹³¹I therapy is more frequent in smokers than in nonsmokers (58, 59). Improvement of eye changes after prednisone or retrobulbar irradiation is less frequent in smokers (60, 61). Smoking increases the risk of recurrence in Graves’ hyperthyroidism (62). Some evidence exists that passive smoking is also a risk for developing Graves’ ophthalmopathy (63).

The ophthalmopathy is more severe in patients who still have an abnormal thyroid function despite antithyroid treatment; it improves slightly after thyroid function has returned to normal (41, 64). Restoration and maintenance of euthyroidism is thus relevant for the eyes. Prolonged treatment with antithyroid drugs (preferably in combination with thyroxine, the so-called block and replace regimen) until full rehabilitation of the ophthalmopathy is obtained is a feasible option. When the eye disease has become inactive and needs no further treatment, antithyroid drugs can be discontinued; if Graves’ hyperthyroidism recurs, it can be treated with ¹³¹I without adverse effects on the eyes (65).

Iodine-131 therapy is associated with a risk of about 20% for developing or worsening of eye changes in patients with no or minimal ophthalmopathy before treatment. This conclusion is based on three large randomized clinical trials (58, 59, 66) and a systematic review (67). Risk factors for developing or worsening of Graves’ ophthalmopathy after ¹³¹I therapy are a pretreatment triiodothyronine value above 5 nmol/l (Fig. 3.3.10.10), pre-existent active ophthalmopathy, high TSH-binding inhibiting immunoglobulin (TBII) values, and smoking. High TSH levels after ¹³¹I therapy also constitute a risk for ophthalmopathy and should be avoided (68, 69). The eye changes after ¹³¹I therapy usually occur within 6 months and are mostly transient and mild in nature. They can be prevented by steroids (Table 3.3.10.3), but to expose all patients selected for radio-iodine to the side effects of prednisone to prevent persistent eye changes in 5% is not warranted. It might be better to restrict steroids to patients with one or more risk factors (70). Preventive steroids do not interfere with the efficacy of ¹³¹I therapy (71), but uncertainty exists about the most appropriate dosage schedule: oral prednisone 0.2 mg/kg per day given for 6 weeks might be as effective as 0.3–0.5 mg/kg per day given for 3 months. A causal relationship between ¹³¹I therapy and the eye changes is plausible in view of the release of thyroid antigens following radiation injury. The resulting T-cell activation and prolonged increase in serum TBII (which lasts for 5 years, in sharp contrast to the fall in serum TBII observed after treatment with antithyroid drugs or thyroidectomy) (72) may trigger autoimmune reactions in the orbit. A similar mechanism would explain the occurrence of ophthalmopathy after neck irradiation for nonthyroidal neoplastic disease (73).

Total thyroidectomy has been recommended in Graves’ ophthalmopathy in order to remove all thyroid antigens and thyroid-directed T lymphocytes. This approach is logical assuming cross-reactivity between thyroid and orbital antigens is involved in the immunopathogenesis of Graves’ ophthalmopathy, although orbital autoimmunity may proceed independently once the eye disease is well established. In a case–control study, however, development or worsening of eye changes after near-total thyroidectomy occurred in 3.3%, exactly the same as in carefully matched patients treated with methimazole (74). Another study found no difference in the course of ophthalmopathy between subtotal and total thyroidectomy (75). A recent randomized trial in Graves’ ophthalmopathy found total thyroid ablation (total thyroidectomy + ¹³¹I ablation) associated with a slightly better outcome at 9 months than total thyroidectomy alone, but improvement was limited to lid aperture and proptosis and was of minor clinical relevance (76).

In summary, how hyperthyroidism should be treated in the presence of Graves’ ophthalmopathy needs to be judged on an individual basis. Antithyroid drugs and thyroidectomy seem to be neutral with respect to the course of the eye changes. If ¹³¹I therapy is chosen, preventive steroids should be given in high-risk patients, particularly in patients who smoke and still have active eye disease.

General recommendations are aimed at protecting the cornea and relieving symptoms such as photophobia and diplopia (Box 3.3.10.2). Prisms can be helpful when the angle of strabismus is constant and not too large. While awaiting definitive treatment, botulinum toxin injections can be given into the upper eye lid to relieve exposure problems, or into the glabellar muscles to alleviate frowning. Further specific recommendations depend on the severity and activity of the ophthalmopathy (57).

Patients with DON require immediate intervention. The effect of retrobulbar irradiation develops slowly, and consequently radiotherapy is not the treatment of choice in DON. The effect of surgical decompression is observed within days, and that of glucocorticoids within weeks. There is only one randomized controlled trial in DON, which, although its sample size is limited, indicates that intravenous pulses of methylprednisolone are associated with better outcome than immediate surgical decompression (77). Our current scheme therefore is to start with methylprednisolone pulses (1000 mg daily, given as a 60-min intravenous infusion, on three successive days in week 1 and repeated in week 2). If visual functions have not improved after 2 weeks, we do an urgent surgical decompression. Otherwise we continue with oral prednisone (40 mg/day for 2 weeks, 30 mg/day for 4 weeks, 20 mg/day for 4 weeks, and then tapering to zero dose by 2.5 mg/week).

Mild Graves’ ophthalmopathy is defined as having only one or more of the following features: lid retraction less than 2 mm, mild soft tissue involvement, exophthalmos less than 3 mm above normal, and intermittent or no diplopia. The impact on daily life is minor and insufficient to justify immunosuppressive or surgical treatment (57). Spontaneous improvement can be expected in about 30% of patients after 1 year. Therefore, in general, a ‘wait-and-see’ policy is recommended. Sometimes the mild eye changes have such a negative impact on a patient’s life that intervention is warranted. Retrobulbar irradiation in mild ophthalmopathy has a response rate of 50–60%, twice as high as that of sham irradiation, but it does not prevent progress to the more severe ophthalmopathy that occurs in about 15% of patients (78, 79). Recent data suggest selenium treatment improves the outcome of mild ophthalmopathy.

Immunosuppressive treatment modalities in moderately severe Graves’ ophthalmopathy are listed in Table 3.3.10.4. Immunosuppression, in general, is effective for the relief of pain and restoring visual acuity; its effectiveness is moderate for improvement of soft tissue involvement and extraocular muscle dysfunction, and poor for reduction of proptosis. Immunosuppression consequently seldom cures the ophthalmopathy, and most patients still require rehabilitative surgery afterwards. The main accomplishment of immunosuppression seems to be inactivation of the eye disease, thereby permitting earlier corrective surgery.

Glucocorticoids are better than placebo (80) and generally considered to be the mainstay of immunosuppression in Graves’ ophthalmopathy. Intravenous pulses of methylprednisolone have a higher response rate than oral prednisone (74–88% vs 51–63%), and have fewer side effects (17–56% vs 51–85%) (81, 82). Intravenous pulsed methylprednisolone is thus presently the treatment of choice (57, 83). However, four deaths due to acute liver failure occurred in patients receiving large cumulative doses (8–15 g) of intravenous methylprednisolone with an estimated incidence of 0.04%. This has not been observed at a lower cumulative dose of 4.5 g which is currently recommended and can be considered relatively safe as long as liver function tests and hepatitis serology are checked (84). Whereas the combination of oral prednisone and radiotherapy is more effective than oral prednisone alone (83), addition of radiotherapy to intravenous pulsed methylprednisolone may have no extra benefit (85). If the eye changes worsen after discontinuation of glucocorticoids, a combination of low-dose oral prednisone (20 mg/day) with either orbital irradiation or ciclosporin can be tried.

The rationale of orbital radiotherapy is the radiosensitivity of lymphocytes and fibroblasts. Improvement is seen in about 50–60% of patients, predominantly in soft tissue swelling and eye muscle ductions (57, 83, 86). The efficacy of orbital radiotherapy administered in 20 divided fractions of 1 Gy weekly over 20 weeks is slightly better than that of 10 fractions of 1 or 2 Gy daily over 2 weeks (87). Even lower doses than 20 or 10 Gy may be effective. Side effects of retrobulbar irradiation are few. A transient increase of conjunctival irritation is seen in 15% of patients. Radiation-induced retinopathy is extremely rare; the reported cases are, with few exceptions, associated with errors in dosage calculation and radiation technique, or diabetes (86). Although long-term follow-up studies did not detect serious complications after 21 years (88, 89), there exists a theoretical risk of about 0.5% for radiation-induced cancer. The attributable life-time risk from small radiation doses is considerably larger at a younger age than later in life. It is prudent to avoid radiotherapy in patients under 35 years of age, as well as in patients with diabetes or undergoing chemotherapy.

Improvement after ciclosporin monotherapy occurs in 22% of patients, probably reflecting the natural course of the disease (90). The combination of ciclosporin with a low dose of prednisone (20 mg/day) can be effective in patients not responding to steroids or in whom radiotherapy is contraindicated (90, 91). For other modalities, randomized controlled trials have shown no benefit of azathioprine, ciamexone, or acupuncture, but intravenous immunoglobulins can be effective. Somatostatin analogues such as octreotide and lanreotide decrease slightly the clinical activity score but otherwise have little effect (83). The new analogue pareotide is promising because it binds with higher affinity to somatostatin receptor subtypes 1, 2, 3, and 5, all expressed on orbital fibroblasts. Antibodies against tumour necrosis factor (etanercept and infliximab) or against CD20 (rituximab) have shown some beneficial effects but so far have not been tested in randomized controlled trials (92, 93).

Once Graves’ ophthalmopathy has become inactive, rehabilitative surgery can be carried out to improve visual functions and appearance. If surgery is performed while the disease is still active, the benefits might be lost because of ongoing disease. Most orbital surgeons therefore require stable eye disease for 6 months before surgery.

Orbital decompression, achieved by removal of part of the bony orbital wall, is very effective in reducing exophthalmos. The more orbital wall that is removed, the greater the reduction in proptosis. Careful fat removal during bony decompression is increasingly done. Photographs of how the patient looked before the onset of Graves’ ophthalmopathy are helpful in determining the required extent of proptosis reduction. Transeyelid or transconjunctival incisions leave a barely visible scar. Postoperative de novo or worsening of diplopia occurs in 10–30% of patients; corrective eye muscle surgery for diplopia should therefore not be done before decompressive surgery. Single binocular vision can be obtained in about 80% of patients, but more than one surgical session is often required to reach this goal. Lastly, eyelid surgery can be performed in order to correct upper or lower eyelid retraction; blepharoplasty can reduce any remaining eyelid swelling.

Graves’ dermopathy or localized myxoedema occurs usually in the pretibial area, but can occasionally occur at other sites associated with a history of trauma. The most frequent form (49%) is nonpitting oedema with violet discolouration and induration of the skin and prominent hair follicles, so that the lesions have the appearance and texture of orange peel (peau d’orange). Other clinical forms are plaques (27%), nodules (18%), and elephantiasis (5%) (94). Graves’ dermopathy occurs almost always in conjunction with Graves’ ophthalmopathy and high serum concentrations of TSH-receptor antibodies. It is observed in 4% of patients with clinically evident ophthalmopathy, and develops mostly about 1 year after the onset of ophthalmopathy (94). The postulated pathogenesis of dermopathy is remarkably similar to that of ophthalmopathy (Fig. 3.3.10.5): cytokine-induced glycosaminoglycan production by dermal fibroblasts, up-regulated TSH receptor expression on dermal fibroblasts, and a contributory role of local mechanical pressure (95).

The natural course of Graves’ dermopathy is not well known. One-half of the patients do not require any therapy, and in such cases 50% have complete remission within 17 years. If treatment is necessary because of functional or cosmetic complaints, night-time occlusive dressings of 0.05–0.1% triamcinolone acetonide in a cream base induce partial remissions in 38% of patients. Usually a trial of 4–10 weeks is needed, followed by intermittent maintenance therapy. The use of compressive bandages or stockings during the day provides additional benefit. Treatment of the coexistent ophthalmopathy with systemic glucocorticoids may cause regression of the skin lesions as well.