Rationale: Predicting corticosteroid response in COPD is important but difficult. Response is more likely to occur in association with eosinophilic airway inflammation, for which the fraction of exhaled nitric oxide (FENO) is a good surrogate marker. Objectives: We aimed to establish whether FENO levels would predict the clinical response to oral corticosteroid in COPD. Methods: We performed a double-blind, crossover trial of steroid in patients with COPD. After a 4-week washout of inhaled steroids, patients received prednisone 30 mg/d or matching placebo, in random order, with an intervening 4-week washout. The predictive values of FENO for clinically significant changes in 6-minute-walk distance (6MWD), spirometry (FEV1), and St. George's Respiratory Questionnaire (SGRQ) were calculated. Measurements and Main Results: A total of 65 patients (mean FEV1 = 57% predicted) were randomized. With prednisone, there was a net increase of 13 m in 6MWD (P = 0.02) and 0.06 L in postbronchodilator FEV1 (P = 0.02) compared with placebo. The change in SGRQ was not significant. Using receiver operator characteristic analysis, the area under the curve for an increase of 0.2 L in FEV1 was 0.69 (P = 0.04) with an optimum FENO cut-point of 50 ppb. The positive and negative predictive values were 67 and 82%, respectively. FENO was not a significant predictor for changes in 6MWD or SGRQ. Conclusions: FENO is a weak predictor of short-term response to oral corticosteroid in COPD, its usefulness being limited to predicting increase in FEV1.