Immunosuppressive treatment of thyroid-associated ophthalmopathy with high doses of oral prednisone has a response rate of 65% at the expense of major side effects. Retrobulbar irradiation is almost devoid of side effects and has an identical response rate of 65%. Newer treatment modalities have a similar response rate, except cyclosporin A which is effective in only 22% of patients. Intravenous pulses of methylprednisolone might be useful in optic neuropathy. Intravenous immunoglobulin and subcutaneous octreotide, although reasonably well tolerated, are laborious and expensive. Consequently, retrobulbar irradiation is still the treatment of choice in moderately severe cases; it is contra-indicated in diabetes mellitus.The efficacy of immunosuppression can be increased by combination therapy (e.g., prednisone + irradiation). Another way of increasing efficacy is restriction of immunosuppression to those patients who are likely to respond, i.e., patients with active eye disease. Disease activity can be assessed by the clinical activity score and various imaging techniques like orbital octreoscan and magnetic resonance. The accuracy of these methods in predicting outcome of immunosuppression still has to be determined in prospective studies in a sufficient number of patients.