To better define the effect of systemic glucocorticoids on the cutaneous early and late phase response (LPR), nine atopic subjects were examined in a double-blind cross-over study using skin chambers fixed over denuded skin blisters. A challenge was carried out by placing allergen in the chamber for 60 min in subjects who received either a 3-day pretreatment with 60 mg/day of prednisone or placebo. Skin chamber cell counts and inflammatory mediators (histamine, PGD2, and leukotriene C4 (LTC4)) were measured at hourly intervals for 12 h. Prednisone pretreatment did not alter the immediate skin erythema or release of histamine but ablated the late secondary erythema and rise in histamine. The median histamine values during h 10, 11, and 12 in the placebo and prednisone pretreatment visits were 3.73 and 0.22 ng/ml, respectively (p less than or equal to 0.02). Prednisone did not alter PGD2 production; however, LTC4 production was suppressed during the LPR. The cumulative median LTC4 values during h 7, 8, and 9 were 5.6-fold (p less than or equal to 0.05) more after placebo than after prednisone pretreatment. Prednisone altered cellular traffic more dramatically than it did inflammatory mediators. The influx of eosinophils, which peaked during the 9th and 10th h in placebo-treated patients, was completely blocked by prednisone (p less than or equal to 0.02) for every h from 6 through 12. The influx of basophils, which started during the 9th h and peaked during the 12th h in placebo-treated patients, was suppressed at all time points (p less than or equal to 0.02) in prednisone-treated patients. There was no significant alteration in neutrophil transit into the skin chambers induced by prednisone. We suggest that the selective blockade of eosinophil and basophil influx by prednisone and the associated decrease in inflammatory mediators may contribute to the blockade of the clinical expression of the cutaneous LPR.