Background: Exhaled nitric oxide (ENO) has been proposed as a noninvasive marker of airway inflammation in asthma. Objective: We investigated the relationships among ENO, eosinophilic airway inflammation as measured by induced sputum, and physiologic parameters of disease severity (spirometry and methacholine PC₂₀). We also examined the effect of corticosteroid treatment and atopy on ENO levels and eosinophil counts in induced sputum. Methods: Measurements were taken on one day in 22 healthy nonatopic subjects, 28 healthy atopic subjects, 38 asthmatic subjects not taking inhaled steroids, 35 asthmatic subjects taking inhaled steroids, and 8 subjects with eosinophilic bronchitis without asthma. Results: ENO levels showed significant but weak correlations with eosinophil differential counts in the steroid-naive asthmatic and healthy atopic groups (r_s < 0.05). ENO levels were significantly lower in the asthmatic subjects taking steroids compared with the asthmatic subjects not taking steroids, despite there being no difference in the sputum cell counts, and a tendency to increased airflow limitation. ENO levels and sputum eosinophil counts were equally good at differentiating from steroid-naive asthmatic subjects. ENO levels were consistently raised in subjects with eosinophilic bronchitis without asthma. Atopy had no effect on ENO levels in the healthy subjects. Conclusion: We conclude that ENO is likely to have limited utility as a surrogate clinical measurement for either the presence or severity of eosinophilic airway inflammation, except in steroid-naive subjects. (J Allergy Clin Immunol 2000;106:638-44.)