Psychological morbidity is common in chronic respiratory diseases. The diagnostic accuracy of the Hospital Anxiety and Depression Scale (HADS) and risk factors for comorbid depression in chronic obstructive pulmonary disease (COPD) are addressed. Consecutive COPD patients (GOLD stage I-IV, 40-75 years old) were enrolled in a multicentre, cross-sectional cohort study. Diagnosis of depression was ascertained through clinical records. Lung function, HADS score, 6-minute walking test (6-MWT), MRC dyspnoea score, and COPD Assessment Test (CAT) were evaluated. Two hundred fifty-nine COPD patients (mean age 62.5 years; 32% female; mean FEV1 48% predicted) were included. Patients diagnosed with depression (29/259; 11.2%) had significantly higher HADS-D and HADS-Total scores than nondepressed patients (median (quartiles) HADS-D 6 [4; 9] versus 4 [2; 7], median HADS-Total 14 [10; 20] versus 8 [5; 14]). Receiver-operating characteristic plots showed moderate accuracy for HADS-D, AUC 0.662 (95%CI 0.601-0.719), and HADS-Total, AUC 0.681 (95%CI 0.620-0.737), with optimal cut-off scores of >5 and >9, respectively. Sensitivity and specificity were 62.1% and 62.6% for HADS-D compared to 75.9% and 55.2% for HADS-Total. Age, comorbidities, sex, and lower airflow limitation predicted depression. The HADS exhibits low diagnostic accuracy for depression in COPD patients. Younger men with comorbidities are at increased risk for depression.