Global population-level concurrence and directional associations between rheumatoid arthritis and depressive...

Rheumatoid arthritis (RA) and depressive disorders (DEP) frequently co-occur in clinical populations, but their population-level temporal patterns and variation across developmental contexts remain insufficiently characterized. Using Global Burden of Disease (GBD) 2021 data, this study assessed population-level spatial concurrence, temporal predictability, and projected combined estimated incident cases of RA and DEP across 204 countries and territories, with particular attention to the role of the Sociodemographic Index (SDI).

Countries were grouped by SDI quintile. Annual age-standardized incidence rates (ASIRs) for RA and DEP from 1992 to 2021 were analyzed. ARIMAX models were used to describe single-disease trends, whereas VARX models jointly modelled RA and DEP while adjusting for SDI. Granger causality tests assessed whether the past burden of one condition improved prediction of the other, and impulse response functions (IRFs) summarized model-estimated responses over subsequent years. Projections to 2030 were interpreted as conditional model-based estimates.

RA and DEP incidence rates were positively correlated globally. After SDI adjustment, the number of countries showing bidirectional temporal predictability increased from 67 to 104. IRF analyses indicated larger and more heterogeneous model-estimated RA-to-DEP responses than DEP-to-RA responses, while supporting bidirectional temporal associations. Conditional projections suggested that some middle-SDI countries may experience changes in population-scaled combined estimated incident cases by 2030, whereas several higher-SDI settings showed stable or declining projected rankings under the model assumptions.

Developmental context shaped the detectability and magnitude of country-level RA-DEP temporal associations. These findings provide population-level context for interpreting RA and DEP patterns across settings; clinical screening and treatment decisions require evidence from individual-level studies and guidelines. Key Points • SDI adjustment increased the number of countries with detectable bidirectional temporal predictability between rheumatoid arthritis (RA) and depressive disorders (DEP) from 67 to 104. • Granger analyses of both-sex country-level annual ASIR series indicated bidirectional temporal predictability, while impulse response analyses showed larger and more heterogeneous model-estimated RA-to-DEP responses than DEP-to-RA responses. • Conditional time-series projections suggested that some middle-SDI settings may experience changes in population-scaled combined estimated incident cases by 2030 under historical-trend and model assumptions. • The findings provide population-level context for interpreting RA and DEP patterns across settings.