Most pathologists assume that a diagnosis of Churg-Strauss syndrome (CSS) requires the finding of necrotizing vasculitis accompanied by granulomas with eosinophilic necrosis in the setting of asthma and eosinophilia. However, recent data indicate that this definition is too narrow and that adherence to it leads to cases of CSS being missed. CSS has an early, prevasculitic phase that is characterized by tissue infiltration by eosinophils without overt vasculitis. Tissue infiltration may take the form of a simple eosinophilia in any organ, and a fine-needle aspirate showing only eosinophils may suffice for the diagnosis in this situation. The prevasculitic phase appears to respond particularly well to steroids. Even in the vasculitic phase of CSS, many cases do not show a necrotizing vasculitis but often only an apparently nondestructive infiltration of vessel walls by eosinophils. In modern biopsy materials, granulomas frequently cannot be found. In the postvasculitic phase of CSS, healed vascular lesions resemble organized thrombi but typically show very extensive destruction of elastica and, often, an absence of eosinophils. The widespread use of steroids as therapy for asthma has led to the peculiar and confusing situation in which the steroid therapy accidentally suppresses CSS and changes in steroid treatment uncover the disease; this type of "formes frustes" CSS is now well recognized with leukotriene receptor antagonist treatment and will be seen with increasing frequency as other steroid-sparing therapies for asthma are introduced.