Background Antibiotics do not reduce mortality or short-term treatment non-response in patients receiving treatment for acute exacerbations of COPD in an outpatient setting. However, the long-term effects of antibiotics are unknown. The aim of this study was to investigate if the antibiotic doxycycline added to the oral corticosteroid prednisolone prolongs time to next exacerbation in patients with COPD receiving treatment for an exacerbation in the outpatient setting. Methods In this randomised double-blind placebo-controlled trial, we recruited a cohort of patients with COPD from outpatient clinics of nine teaching hospitals and three primary care centres in the Netherlands. Inclusion criteria were an age of at least 45 years, a smoking history of at least 10 pack-years, mild-to-severe COPD (Global Initiative of Chronic Obstructive Lung Disease [GOLD] stage 1–3), and at least one exacerbation during the past 3 years. Exclusion criteria were poor mastery of the Dutch language, poor cognitive functioning, known allergy to doxycycline, pregnancy, and a life expectancy of shorter than 1 month. If a participant had an exacerbation, we randomly assigned them (1:1; with permuted blocks of variable sizes [ranging from two to ten]; stratified by GOLD stage 1–2 vs 3) to a 7 day course of oral doxycycline 100 mg daily (200 mg on the first day) or placebo. Exclusion criteria for randomisation were fever, admission to hospital, and current use of antibiotics or use within the previous 3 weeks. Patients in both groups received a 10 day course of 30 mg oral prednisolone daily. Patients, investigators, and those assessing outcomes were masked to treatment assignment. The primary outcome was time to next exacerbation in all randomly allocated patients except for those incorrectly randomly allocated who did not meet the inclusion criteria or met the exclusion criteria. This trial is registered with the Netherlands Trial Register, number NTR2499. Findings Between Dec 22, 2010, and Aug 6, 2013, we randomly allocated 305 (34%) patients from the cohort of 887 patients to doxycycline (152 [50%]) or placebo (153 [50%]), excluding four (1%) patients (two [1%] from each group) who were incorrectly randomly allocated from the analysis. 257 (85%) of 301 patients had a next exacerbation (131 [87%] of 150 in the doxycycline group vs 126 [83%] of 151 in the placebo group). Median time to next exacerbation was 148 days (95% CI 95–200) in the doxycycline group compared with 161 days (118–211) in the placebo group (hazard ratio 1·01 [95% CI 0·79–1·31]; p=0·91). We did not note any significant differences between groups in the frequency of adverse events during the first 2 weeks after randomisation (47 [31%] of 150 in the doxycycline group vs 53 [35%] of 151 in the placebo group; p=0·54) or in serious adverse events during the 2 years of follow-up (42 [28%] vs 43 [29%]; p=1). Interpretation In patients with mild-to-severe COPD receiving treatment for an exacerbation in an outpatient setting, the antibiotic doxycycline added to the oral corticosteroid prednisolone did not prolong time to next exacerbation compared with prednisolone alone. These findings do not support prescription of antibiotics for COPD exacerbations in an outpatient setting. Funding Netherlands Organization for Health Research and Development.