We reasoned that a prospective assessment of glucocorticoid withdrawal in subjects with asthma would provide insight into the basis for flares of the disease. We therefore enrolled 25 subjects with moderate persistent asthma and treated them for 30 days with inhaled fluticasone propionate (1,760 microg/day) followed by a withdrawal period that lasted until peak expiratory airflow decreased by 25% and FEV(1) by 15% or 6 weeks elapsed. After glucocorticoid withdrawal, 13 of 25 subjects reached the target, whereas 12 subjects did not. The number of eosinophils in bronchial biopsies was increased by glucocorticoid withdrawal in both groups, but increases in airway T cells were found in only those with exacerbation. T-cell accumulation was a reflection of similar increases in both CD4(+) and CD8(+) T cells and was accompanied by increased expression of chemokine CCL5 (regulated upon activation, normal T cell expressed and secreted) in the airway epithelium without activation of the transcription factor nuclear factor-kappaB. The pattern of glucocorticoid-sensitive inflammation during an asthma exacerbation is more reminiscent of an antiviral response than an eosinophil-predominant response to allergen and implies an independent role for airway T cells in mediating asthma flares and in determining glucocorticoid efficacy in the treatment of this disease.