ObjectiveTo analyze the sensitivity of peak flow meter screening in different subgroups of chronic obstructive pulmonary disease (COPD). MethodsA total of 156 outpatients with COPD from Peking Union Medical Hospital from May 2013 to December 2014 were recruited in the study. Each patient's symptoms,history of exposure to risk factors,and the times of exacerbation in last year was recorded. All patients completed CAT,mMRC,the St George's Respiratory Questionnaire (SGRQ),6 minutes walking test,spirometry,and peak expiratory flow (PEF) by peak flow meter. ResultsUsing the cut-off of PEF%pred=80%,the PEF detected 120 COPD patients in 156 subjects. The predictive factors of abnormal PEF%pred in COPD was FEV1%pred and the total score of SGRQ (P<0.05). PEF screening could identify 76.9% of COPD patients,30.0%-60.0% of patients of less symptoms (mMRC<2 or CAT<10 or SGRQ<25),83.3%-90.9% of COPD patients with more symptoms (mMRC ≥ 2 or CAT ≥ 10 or SGRQ ≥ 25),27.7% of COPD patients with mild airflow limitation,68.5% of COPD patients with moderate airflow limitation,83.3% of COPD patients with moderate to very severe airflow limitation. When grouped by GOLD combined assessment method,PEF screening could identify 35.2% of patients of group A,75.0% of patients of group B,and 95.9% of patients of group C and D. The cut-off value of PEF% pred=80% showed low sensitivity to early stage of COPD,but when using the cut-off value of PEF% pred=95%,that sensitivity increased signifcantly. ConclusionsPeak flow meter may be used as a tool to screen COPD. It can identify part of COPD patients especially for those patients with more symptoms,requiring regular treatment,with deteriorated pulmonary function and high risk of exacerbation.