Corticosteroids were the mainstay of therapy for idiopathic pulmonary fibrosis (IPF) for more than four decades, but their efficacy is unproven and toxicities are substantial. The course of IPF is characterized by progressive respiratory insufficiency, leading to death within 3 to 8 years from the onset of symptoms. Although a subset (10-20%) of patients survives more than 10 years, there is no evidence that any form of therapy alters the natural history of the disease. Nonetheless, given the poor prognosis, a trial of corticosteroids is often given. Because of the rarity of IPF, randomized, placebo-controlled therapeutic trials have not been done. Further, no studies have compared differing dosages or duration of corticosteroid in matched patients. Interpretation of therapy efficacy is obscured by several factors including heterogeneous patient populations, inclusion of patients with histologic entities other than usual interstitial pneumonia, lack of objective, validated endpoints, different criteria for "response." We review published data regarding corticosteroid therapy for IPF and present a rationale for stratifying therapy based on host, demographic, and clinical factors that influence prognosis as well as risk for corticosteroid complications.