Eye Management
	A liberal use should be made of artificial tears and supportive measures like dark glasses and prisms: many symptoms are thus relieved in a simple manner. A further management plan can be laid out according to disease severity: In mild eye disease one may opt to wait and see in view of spontaneous improvement in a large proportion of patients (Perros et al 1995). In very severe eye disease ("malignant" GO, optic neuropathy) steroids are preferred because their efficacy and rapid reponse. Intravenous pulses of methylprednisolone (e.g. 1 gram on three successive days weekly for 2 weeks) followed by oral prednisone is an attractive scheme. If no improvement in optic nerve function is seen after two weeks, urgent decompressive surgery is required. In moderately severe eye disease immunosuppressive treatment is usually given. The mainstay of immunosuppression has been high doses of oral steroids, but retrobulbar irradiation (20 Gy delivered in ten daily doses of 2 Gy over a two-week period) has similar efficacy as oral steroids but is devoid of the many side effects of steroids (Prummel et al. 1993). Cyclosporin-A as monotherapy is not very effective, but potentiates the effect of low-dose steroids; it can be considered when irradiation is contraindicated like in diabetic patients (Polak and Wijngaarde 1995). Other immunosuppressive treatment modalities such as intravenous immunoglobulins and subcutaneous octreotide (table 2) are laborious and very expensive; their response rate is in the same order (around 65%) as steroids and irradiation (Wiersinga 1996). Retrobulbar irradiation has thus been advocated as the first choice. A recent abstract reports similar response rates of 20 Gy and 10 Gy (Kahaly et al. 1996). Table 2: Efficacy and side effects of immunosuppression in Graves' ophthalmopathy Immunosuppression Efficacy Side Effects Remarks*  Pharmacologic  glucocorticoids ++ ++ mainstay of immunosuppression  cyclosporine A ± + RCT, limited application  intravenous     immunoglobulin + ± RCT, very expensive  azathioprine - - RCT, not indicated  ciamexone - - RCT, not indicated  cyclophosphamide ± ± no RCT  subcutaneous octreotide + ± limited data, expensive  Nonpharmacologic  acupuncture - - RCT, not indicated  plasmapheresis ± ± no RCT, steroids needed to  prevent rebound effect  retrobular irradiation + - RCT, contraindicated in  diabetes mellitus  *RCT, randomized clinical trial  The efficacy of immunosuppression is good for relief of pain and improvement of visual acuity, moderate for improvement of appearance and correction of extra-ocular muscle dysfunction, and poor for reduction of proptosis. It follows that immunosuppression seldom cures the eye disease, and full restoration of eye function and appearance usually requires several surgical procedures. The results of rehabilitative surgery, however, can be lost when the eye disease is still active; it is therefore performed when the disease has been stable for six months. The main advantage of immunosuppression is thus inactivation of the inflammation of orbital tissues, thereby permitting corrective surgery at an earlier time. The efficacy of immunosuppression can be enhanced by combination therapy (like irradiation plus steroids). Alternatively, immunosuppression might be restricted to those patients who are likely to respond; age, sex, duration nor severity of GO, however, reliably predict outcome. The notion that patients with the same degree of severity of their eye disease can or cannot respond to immunosuppression has led us to hypothesize that disease activity is the main determinant of outcome of immunosuppression. This is illustrated in Figure 5. In the natural history of GO the early dynamic phase is characterized by an initial stage of ingravescence followed by a stage of remission with some - usually incomplete - regresion of eye signs; the dynamic phase may last from several months to a few years. The eye disease then enters a static phase in which the ocular findings do not change very much. Immunosuppression will not be effective in the static (fibrotic) phase; in contrast, much benefit may be expected if given in the dynamic phase when active inflammation is present. We propose that the natural history of GO is also characterized - besides Rundle's curve of disease severity - by a curve depicting disease activity. As illustrated in Figure 5, disease activity might have predictive value, useful in the selection of patients who are likely to respond to immunosuppression. How can disease activity be assessed? A variety of methods have been proposed including a clinical activity score (Mourits et al. 1989). T2 relaxation time on MRI (Just et al. 1991), internal eye muscle reflectivity on ultrasound (Prummel et al. 1993), orbital octreoscan (Postema et al. 1994; Moncayo et al. 1997) and urinary glycosaminoglycans (Kahaly et al. 1990). The validation of these parameters as reliable indicators of disease activity must be derived indirectly (because biopsies of retrobulbar tissues are seldom feasible) from their accurate prediction of outcome of immunosuppression. The results of such studies are eagerly awaited. Management of GO according to disease activity (Table 3) has the advantage that immunosuppression will be only applied in patients who are likely to respond, and that patients with inactive disease might proceed directly to surgery. Figure 5a: Figure 5b: Outcome of immunosuppressive treatment of Graves' ophthalmopathy as function of disease activity. The natural history of the activity and of the severity of the eye disease are depicted by two separate curves. Immunosuppression is given at half maximal disease severity: the response to treatment is negligible when given at zero disease activity (panel A) but substantial when applied at peak disease activity (panel B) (Wiersinga 1996). Table 3: Management of Graves' opthalmopathy according to diseaseactivity In any stage maintain euthyroidism  protect cornea (liberal use of artificial tears)  dark glasses and prisms as required  In active stage retrobulbar irradiation (not in diabetic patients)  steroids (for optic neuropathy)  urgent orbital decompression (in "malignant" GO with steroids failure)  In inactive stage rehabilitative surgery in the following sequence:  orbital decompression (for disfiguring proptosis)  eye muscle surgery (for double vision)