Dropout or lack of response is an important issue in pulmonary rehabilitation (PR), which underlines the need to identify predictors of dropout and response. Acute exacerbations (AEs) of COPD may influence dropout rates and PR response. We aimed to assess differences in dropout and outcomes of PR between COPD with and without AEs. Clinically stable patients with moderate-to-very severe COPD (age: 64.1 ± 9.1 years, 55.6% males, forced expiratory volume in 1 s (FEV1 ): 48.6 ± 20.0% predicted) were assessed during PR (inpatient and outpatient). Mild-to-moderate AEs were defined as 'the prescription of systemic glucocorticosteroids and/or antibiotics, following an acute increase in respiratory symptoms'. Severe AEs were defined as 'a hospital admission due to an AE'. Health status was measured by COPD Assessment Test (CAT), COPD-specific version of the St George's Respiratory Questionnaire (SGRQ-C) and Clinical COPD Questionnaire (CCQ). Symptoms of anxiety and depression were measured by Hospital Anxiety and Depression Scale (HADS). Exercise capacity was measured with the 6-min walking test (6MWT) and constant work rate test (CWRT). A total of 518 patients were assessed during a pre-rehabilitation assessment. Four hundred and seventy-six patients started PR, of whom 419 (88.0%) completed it. A larger proportion of patients who dropped out had a severe AE during PR (20.0% vs 3.5%, P < 0.001). Completers with severe AE showed a deterioration in 6MWT, while completers without AE and with mild-to-moderate AE improved (-24.8 (95% CI: -94.0 to 44.5) vs 24.2 (95% CI: 16.0 to 32.5) vs 25.1 (95% CI: 14.0 to 36.3) metres, P = 0.042). No other significant differences were observed in outcomes comparing completers with and without AE during PR. Mild-to-moderate AEs do not affect dropout or response of PR, although severe AEs are associated with dropout. AEs should not lead to discontinuation of PR, as response is in general not affected.