Limb muscle dysfunction in patients with COPD may be associated with local muscle and/or systemic inflammation, and therefore, we investigated whether exercise training altered markers of inflammation and oxidative stress. We obtained vastus lateralis muscle biopsies and venous blood samples from patients with COPD (n=30) before and after 8 weeks of resistance training (RT) (n=15) or endurance training (ET) (n=15). Healthy age-matched subjects were included as baseline controls (n=8). Inflammatory markers in muscle and systemically were determined by interleukins (IL), tumour necrosis factor alfa (TNF-α), leukocyte concentration together with immunohistochemical staining for macrophages. Muscle oxidative stress and antioxidant capacity were determined by NADPH oxidase (NOX) and superoxide dismutase 2 (SOD2), respectively. Before exercise training, COPD patients had a higher muscular NOX protein content and circulating IL-8, IL-18, CRP and leukocyte levels but a similar number of muscle-infiltrating macrophages compared with controls. Eight weeks of ET or RT increased muscle SOD2 content with no difference between groups. Plasma TNF-∝, increased (P < 0.05) after ET and tended to (P = 0.06) increase after RT, but had no effect on muscular NOX protein content, number of muscle-infiltrating macrophages or systemic levels of other pro-inflammatory cytokines or leukocytes. In patients with COPD, we found no evidence for muscular inflammation and no effect of exercise training. However, systemic inflammation was elevated in COPD and both training modalities induced an upregulation of muscle antioxidant capacity. This article is protected by copyright. All rights reserved.