PurposeThe longitudinal association between changes in depressive symptoms (improvement/worsening) and arthritis is unclear.MethodsStudy data were obtained from the China Health and Retirement Longitudinal Study (CHARLS) 2011–2018. The 10-item Center for Epidemiological Studies Depression Scale (CES-D-10) was used to examine participant depressive symptoms and data on self-reported history of arthritis were collected. Depressive symptoms improving is defined as depression at baseline and no depression at follow-up. Similarly, depressive symptoms worsening is defined as no depression at baseline and depression at follow-up. Cox proportional hazards models were used to evaluate the effects of improvement or deterioration in depressive symptoms on arthritis. Participants with missing data on depression and arthritis, having arthritis in 2011 CHARLS and lost to follow-up was excluded.ResultsA total of 8556 participants free of arthritis were included from 2011 to 2018. After adjustment for confounders, depressive symptoms were associated with a 54% increased risk of developing arthritis. Each 1-point increase in CES-D-10 score was associated with a 4% higher risk of arthritis. Participants with depressive symptoms at baseline but improved symptoms (without depressive symptoms) at follow-up had a 25% lower rate of arthritis, and a 1-point reduction in CES-D-10 score during 8 years of follow-up was associated with a 5% lower risk of developing arthritis. Participants with no depressive symptoms at baseline but depression at follow-up had a 66% higher rate of arthritis, and a 1-point increase in CES-D-10 score during 8 years of follow-up was associated with a 5% higher risk of arthritis.ConclusionsImprovement in depressive symptoms was associated with lower risk of arthritis and worsening of depression was associated with higher risk of arthritis. These findings suggest that the relationship between depression and arthritis is complex.