Sputum colour is regarded as a good marker of bacterial involvement in acute exacerbations of chronic obstructive pulmonary disease (COPD) and guides many physicians in deciding on antibiotic treatment. Although most doctors rely on the sputum colour that is reported by patients, it can also be assessed using a validated colour chart. In this study, reported sputum colour and assessed sputum colour were compared as markers of the presence of bacteria, bacterial load, and systemic inflammation. Data on 257 exacerbations in 216 patients hospitalized with an acute exacerbation were analysed (mean age, 72 years; mean forced expiratory volume in 1 s, 44.8% + or - 17.8% (+ or - standard deviation)). Sputum colour was reported by the patients and assessed at the laboratory with a colour chart. Subsequently, quantitative sputum cultures were performed. C-reactive protein was measured as a marker of systemic inflammation. A sputum sample was obtained in 216 exacerbations (84%), of which 177 (82%) were representative. A pathogen was identified in 155 patients (60%). Assessed sputum colour was a better marker of the presence of bacteria (OR 9.8; 95% CI 4.7-20.4; p <0.001) than reported sputum colour (OR 1.7; 95% CI 1.0-3.0; p 0.041). The sensitivity and specificity were 73% and 39% for reported sputum colour, and 90% and 52% for assessed sputum colour. Assessed sputum colour was clearly related to sputum bacterial load and C-reactive protein levels, whereas reported sputum colour was not. It is concluded that sputum colour reported by patients is an unreliable marker of the presence of bacteria in acute exacerbations of COPD. Assessed sputum colour is clearly superior and is also related to bacterial load and systemic inflammation.