BACKGROUND: Limited evidence concerns the associations between ambient temperature and lung function and the results are mixed. OBJECTIVE: To evaluate the associations between temperature variations and daily fluctuations in pulmonary function in chronic obstructive pulmonary disease (COPD) patients. METHODS: We designed a time-series panel study of 28 male urban COPD patients with repeated daily lung function measurements from December 2012 to May 2013 in Shanghai, China. We used a linear mixed-effect model combined with a distributed lag model to estimate the cumulative effects of temperature on morning/evening pulmonary function tests (PFTs), including peak expiratory flow (PEF) and forced expiratory volume in 1-s (FEV1), while adjusting for within-subject correlations, individual characteristics, time trends and air pollution levels. RESULTS: We obtained a total of 8618 pairs of morning PFTs and 8528 pairs of evening PFTs. The associations between daily mean temperature and PEF were inverted U-shaped with both low and high temperatures significantly reducing morning and evening PEF. Compared with the referent temperature (16 degrees C), the low temperature (1st percentile, -1 degrees C) would result in cumulative decreases of 32.20L/min in morning PEF and 21.15L/min in evening PEF over lags of two weeks. The corresponding decrements at the same lag associated with high temperature (99th percentile, 25 degrees C) were 38.10L/min in morning PEF and 27.08L/min in evening PEF. There were no statistically significant changes in morning or evening FEV1. CONCLUSIONS: This time-series panel study provided robust evidence that both low and high temperatures were significantly associated with decrements in pulmonary function, particularly in PEF.