Clinically evident Graves' ophthalmopathy (GO) develops in 25-50 percent of patients with Graves' hyperthyroidism (1). While some patients experience only mild ocular discomfort, approximately 5% have severe ocular disease and are at risk for visual loss. Subclinical GO appears to be present in the vast majority of Graves’ patients who are without clinically evident eye involvement; sensitive orbital imaging studies suggest that approximately 90% of patients with Graves’ hyperthyroidism have orbital changes suggesting ocular involvement (2). It is not understood why some patients with Graves’ disease develop severe ocular disease while others are spared this complication. While it is possible that there exist GO susceptibility genes, no unique gene associations, beyond those known to predispose to Graves' disease itself, have been convincingly identified in the subset with severe GO. In contrast, recent laboratory and clinical studies, to be discussed in this review, have begun to identify environmental factors that interact with the existing orbital autoimmune milieu to contribute to the progression of GO in some patients with Graves’ disease.

THE ORBITAL AUTOIMMUNE MILIEU
Cellular constituents of the autoimmune response within the GO orbit include the myocytes of the extraocular muscles, connective tissue cells (fibroblasts, adipocytes, and intercellular matrix), as well as “professional” immune effector cells (3). Tissue histology shows largely intact muscle fibers and an expanded fat compartment containing macrophages,T lymphocytes, and to a lesser extent, B lymphocytes and natural killer cells. Further characterization reveals increases in both CD4⁺ and CD8⁺ T cells with restriction in the T cell receptor repertoire. The particular T cell-derived cytokines found within GO tissues appears to depend on the “stage” of disease; Th1 cells (secreting IL-2, IFN-β, TNF-α) are predominant in tissues from patients with early disease, while Th2 cells (producing IL-4, IL-5, IL-10) appear in later stages (4). Enlargement of the connective tissue compartment results both from an increase in the volume of the orbital fat and from tissue edema caused by the presence of hydrophilic glycosaminoglycans (GAG). The former is thought to result from de novo adipogenesis occurring within the orbital tissues (5), the stimulus for which is at present unclear. GAG production by orbital fibroblasts appears to be stimulated by inflammatory cytokines present within the orbit (6).
It is well accepted that an autoimmune response against the TSH receptor (TSHR) expressed on thyrocytes is responsible for the hyperthyroidism of Graves' disease. Because of the close clinical relationship between Graves’ hyperthyroidism and GO, it has long been postulated that TSHR might also serve as an autoantigen within the orbit, thus linking Graves' hyperthyroidism with its ocular complications. Most investigators seeking evidence of orbital TSHR expression have reported finding at least low level TSHR gene expression in orbital fibroblasts, preadipocytes or orbital fat, and either intact TSHR protein, or an antigenically related protein within the cells (7-12). Further evidence that orbital TSHR expression may play a role in GO stems from studies showing higher levels of TSHR gene expression in orbital adipose tissues from patients with active GO than in normal orbital tissues or tissues from inactive GO patients (13). Recent studies by Terry Smith and colleagues have demonstrated autoantibodies against IGF-1 receptor in the sera of patients with GO, suggesting that this receptor may represent another important orbital autoantigen (14, 15).

ENVIRONMENTAL INFLUENCES

Mechanical factors and trauma

CT scans of the orbits of patients with Graves’ ophthalmopathy (GO) show increased volume of orbital tissues. While this appears to be due to enlargement of both the orbital fat and the extraocular muscles in the majority of patients, some patients appear to have predominantly either adipose tissue or extraocular muscle expansion. (16). Because the bony orbital socket is unyielding in response to the pressure generated by increased tissue volume, forward displacement of the globe (proptosis) may ensue as a means of orbital decompression. The expanded orbital tissue volume may impact venous and lymphatic outflow from the orbit, leading to periorbital and conjunctival edema. Orbital tissue trauma generated by tissue expansion within the confines of the bony orbit might further aggravate the disease process by releasing inflammatory cytokines and factors that facilitate antigen presentation and T cell activation (17). In addition, it is possible that individual anatomic variability, such as the shape or size of the orbits, or variations in venous or lymphatic drainage, may aggravate the intraorbital process and predispose to the development of severe GO. Similarly, it has been postulated that mechanical factors and trauma to soft tissues of the lower extremities is involved in the pathogenesis of the dermal complications of Graves' disease, termed pretibial dermopathy (18).

Tobacco Smoking
The association between smoking and GO is striking, representing the major risk factor known for this condition. The odds ratio, relative to controls, has been reported to be as high as 20.2 for current smokers, and 8.9 for ex-smokers, suggesting a direct and immediate effect of smoking (19). In addition, studies have shown that among patients with GO, smokers have more severe eye disease than non-smokers, that smoking is associated with aggravation of eye disease following radioiodine therapy, that that it adversely influences the course of GO during treatment with corticosteroids and orbital radiotherapy (19). That smoking is linked to other autoimmune diseases, including rheumatoid arthritis and Crohn's disease, suggests there may be a generalized stimulation of autoimmune processes in smokers. Although mechanisms underlying this association remain unclear, effects of orbital hypoxia, free radicals contained in tobacco smoke, or the low levels of interleukin-1 receptor antagonists found in smokers may be involved (19).

Therapy for Thyrotoxicosis
An area of considerable controversy in the past concerned the impact of the choice of therapy for hyperthyroidism on the subsequent course of GO. Several retrospective studies examined this topic, often with conflicting results. More recently, however, randomized, prospective trials have focused on this area and have helped to clarify the issues involved (20-22). A study by Bartalena and colleagues compared eye changes in 443 patients with moderately severe and active GO, prospectively treated with either radioiodine, methimazole, or radioiodine and prednisone (0.4 to 0.5 mg/kg body weight, starting two to three days after radioiodine therapy and continuing for one month, followed by a 2 month taper). Patients were monitored for 1 year and assessed by objective criteria, an activity score, and patient self-assessment. The groups were similar with regard to percentages of smokers or patients with preexisting GO. The investigators found worsening of eye disease within 6 months after radioiodine therapy in 15% of patients treated with radioiodine alone, in 2.7% of patients receiving antithyroid drugs, and in no patients receiving both radioiodine and corticosteroids. The majority (74%) of the patients who experienced worsening eye status after radioiodine therapy had preexisting GO; the eye changes that occurred were largely mild and returned to baseline within 2 to 3 months in 65% of cases. However, 8 patients (5%) in the radioiodine group required additional treatment for their GO, compared with 1 patient in the methimazole group. Patients with preexisting GO and smokers were more likely to have progression after radioiodine administration.
A recent study by Perros and colleagues examined the effects of radioiodine in patients with minimally active GO, and found no association between this treatment and ocular disease progression when post-radioiodine hypothyroidism is prevented (22). In composite, these studies suggest that patients with Graves' hyperthyroidism who have pre-existing and at least moderately active GO have a slightly increased risk of ocular disease progression following radioiodine therapy. When ocular worsening occurs, it is generally mild and may be prevented with concurrent steroid treatment. This does not appear to be the case in patients with minimally active GO. Mechanisms responsible for this mild worsening are unclear, but may be related to the hypothetical release of autoantigen from the thyroid gland, the elevated TSHR autoantibody production known to occur post-radioiodine, or to destruction of radiosensitive suppressor T cells in the thyroid. It is also possible that the effect is primarily due to the induction of hypothyroidism by radioiodine ablation of the thyroid, as ocular changes were not seen in the Perros study in which hypothyroidism was prevented.

SUMMARY
Ocular involvement see figure,

whether clinically evident or subclinical, appears to be almost ubiquitous in patients with Graves' disease. While studies from several laboratories have identified cells and cellular responses involved in orbital autoimmunity, it remains unclear why some patients with Graves' disease develop clinical GO. To date, no unique genetic associations have been identified in Graves' patients with severe GO. Recent studies suggest that environmental factors may interact with the orbital autoimmune milieu to worsen the ocular disease. Some of these, including variations in orbital anatomy and pressure-related trauma to intraorbital tissues, do not appear amenable to intervention short of orbital surgery. In this context, it would seem prudent to avoid intraorbital steroid injections. Attention to some of the other environmental factors may favorably impact the disease course. Certainly, the advice to stop smoking forms the centerpiece of patient counseling. In addition, some patients with active GO might benefit from a tapering course of prophylactic corticosteroids if radioiodine is to be given to treat hyperthyroidism. This could be considered especially in smokers or patients with severe thyrotoxicosis, weighing the potential benefit against the known side-effects of these medications. Further, it seems wise to avoid significant post-radioiodine hypothyroidism in these patients. Future investigations will focus on identifying therapeutic agents or clinical interventions that will either prevent the disease or interrupt disease progression at a point proximal to the development of serious ocular complications.