Objective   To investigate the relationship between sleep duration and depressive symptoms in older people in China, and to explore whether there are gender differences in the relationship.

  Methods   Accessing the data from China Health and Retirement Longitudinal Study (CHARLS) for 2015 and 2018, we covered in the study a total of 2898 respondents, including 1684 males (58.1%) and 1214 females (41.9%). The 10-item form of Center for Epidemiological Studies Depression Scale (CES-D) was used to measure the depressive symptoms of the older people. The effects of sleep duration on depressive symptoms of older people in China were analyzed with the logistic model. Based on the depressive and non-depressive subgroups in 2018, the differences of health factor variables and sleep duration between 2015 and 2018 were studied. Adjustments were made for health factors presenting statistical differences in 2018 to further explore the stability of the analysis results.

  Results   Among the female older people, those with short sleep duration and those with long sleep duration had increased risks of developing depressive symptoms odds ratio (OR)=1.815, 95% confidence interval (CI): 1.357-2.429 and OR=1.364, 95% CI: 1.001-1.854, respectively in comparison to those with normal sleep duration. Among the male older people, no connection was identified between sleep duration and risks for depressive symptoms (P>0.05). Residing in rural areas, suffering from chronic diseases and having limited ability for Instrumental Activities of Daily Living (IADL) were risk factors for developing depressive symptoms (P<0.05). The sleep duration for 2015 and that for 2018 were found to be statistically different in the depressive older people (P<0.001), but no no statistically significant difference was found between the sleep duration for 2015 and that for 2018 among the non-depressive older people. Findings for chronic diseases and IADL for 2015 and those for 2018 were statistically different in the depressive older population (P<0.001). In the non-depressive older population, findings for chronic diseases, IADL and tobacco smoking status for 2015 and those for 2018 showed statistical difference (P<0.05). Further sensitivity analysis showed that in the female older population, short sleep duration increased the risks for depressive symptoms (OR=1.819, 95% CI: 1.356-2.440, P<0.001), while long sleep duration did not affect the risks for depressive symptoms. In the male older population, short or long sleep duration were not associated with the risks for depressive symptoms.

  Conclusion  The association between sleep duration and the risks for depressive symptoms was different for men and women. For the female older population, the possibility of depressive symptoms should be considered for those with short or long sleep duration. For the male older population, short or long sleep duration had little impact on the risks for depressive symptoms.