Background: The prevalence of chronic obstructive pulmonary disease (COPD) varies in different countries, while the cut-off for airflow obstruction (AO) is still contested. No COPD prevalence data based on the Global Lung Initiative (GLI) 2012 equations are available in Russia. Objectives: This study aims to assess AO prevalence by the GLI lower limit of normal (GLI-LLN) and the fixed cut-off, to identify AO risk factors and to assess the diagnostic value of respiratory symptoms in north-western Russia. Methods: In a north-western Russian population-based sample of 2,974 adults aged 35-70 years, data on socio-demographics, smoking, occupational exposures and respiratory symptoms were collected, and spirometry was performed before and after bronchodilator (BD) administration. Results: The AO prevalence was 6.8% (95% CI 5.8-7.9) based on the fixed and 4.8% (95% CI 3.9-5.7) based on the GLI-LLN cut-off. 22.5% of the participants with post-BD AO had a positive bronchodilator test, and 1% showed a paradoxical BD response. Of the environmental factors studied, only smoking was independently associated with AO (odds ratio 2.47, 95% CI 1.60-3.82). The positive predictive value of respiratory symptoms for AO was 11% based on the fixed and 8% based on the GLI-LLN cut-off. Conclusions: In a sample of adults in north-western Russia, the AO prevalence by the GLI-LLN cut-off was lower than that by the fixed cut-off. The predictive value of respiratory symptoms was low.