Abstract: Objective To investigate the sleep characteristics of elderly patients with late-life depression (LLD) in different seasons and meteorological conditions, and to explore the relationship of seasons and meteorology with sleep quality. Methods A total of 183 LLD patients who were hospitalized in Affiliated Psychological Hospital of Anhui Medical University from 2019 to 2022 and completed polysomnography (PSG) examination were selected. The general data, clinical information and PSG data of the subjects were collected. Based on the Northern Hemisphere meteorological seasonal division method, the subjects were divided into spring group (March to May), summer group (June to August), autumn group (September to November) and winter group (December to February of the following year). The daily meteorological data of Hefei station was obtained through the National Environmental Information Center (NCEI) of National Oceanic and Atmospheric Administration (NOAA). Chi-square test was used to compare the general data and the clinical information of subjects in different seasons, and Kruskal-Wallis test was used to compare the overall distribution differences of sleep characteristics among different seasons, Bonferroni test was used for multiple comparison test. Spearman correlation analysis was used to analyze the correlation of the above data, and Hierarchical linear regression models (HLM) was constructed to analyze the relationship between meteorology and sleep quality in LLD patients. Results There were no differences in the general data and clinical information among the four groups. There were significant differences in total sleep time, sleep efficiency and awakening time among the four groups (P<0.05). Correlation analysis showed that time in bed, total sleep time, non-rapid eye movement sleep phase 2(N2) duration and percentage of N2 to total sleep time(N2%) in LLD patients were positively correlated with sunshine intensity, and percentage of non-rapid eye movement sleep phase 1 to total sleep time (N1%) was negatively correlated with sunshine intensity. While time in bed, total sleep time and N2 were negatively correlated with relative humidity, and N2 duration was negatively correlated with precipitation (P<0.05). The results of the HLM showed that sunlight intensity had a positive effect on time in bed in LLD patients (P<0.05). Conclusions The sleep quality of LLD patients is related to season and meteorology. Compared with cold weather, sufficient sunshine and warm weather is conducive to night sleep of LLD patients.

Key words: late-life depression, sleep, season, meteorology, hierarchical linear regression model