Flunisolide nasal spray was compared to its propylene and polyethylene glycol vehicle in a randomized double-blind study of 20 adult patients with perennial rhinitis. After a two-week baseline period patients received either active flunisolide 50 microgram q.i.d. or placebo for four weeks. Laboratory studies included serum 8 a.m. cortisols, nasal air-flow measurements and nasal smears for eosinophils and fungi. Patients kept daily symptom diaries. There was no difference between active and placebo groups for sneezing, runny nose or nose blowing. Although post-nasal drip showed the greatest improvement in the active group, there was a trend for improvement in both groups. By the second week the percentage of eosinophils on nasal smear significantly decreased in both groups. Nasal air-flow measurements also showed improvement in both active and placebo groups. There was no change in serum cortisol levels compared to baseline. Side effects were similar in both active and placebo groups. Although no positive fungal cultures were obtained during the double-blind study. Candida was cultured during the long-term program in one patient.