Joint associations of cumulative uric acid-to-HDL cholesterol ratio and cumulative depressive symptom burden with incident cardiometabolic multimorbidity: a landmark analysis of CHARLS

Abstract Background Cardiometabolic multimorbidity (CMM) is a growing health concern in ageing populations. Although the uric acid-to-high-density lipoprotein cholesterol ratio (UHR) and depressive symptoms have each been associated with cardiometabolic risk, their cumulative joint associations remain unclear. Whether depressive symptom burden modifies the UHR–CMM association also warrants investigation. Methods We conducted a prospective landmark analysis in the China Health and Retirement Longitudinal Study (CHARLS). UHR and depressive symptoms were assessed in 2011 and 2015. Cumulative UHR (cuUHR) and cumulative depressive symptom burden (cuCESD) were calculated as the mean of the two measurements multiplied by the 4-year interval. Participants with CMM by the 2015 landmark were excluded. Cox models assessed individual and joint associations with incident CMM and estimated standardized 3-year absolute risks. Effect modification was examined on multiplicative and additive scales, with spline models allowing a nonlinear cuUHR association. Results Among 5,617 participants, 1,030 developed CMM. In adjusted models, each standard deviation increase in cuUHR was associated with a hazard ratio (HR) of 1.198 (95% confidence interval [CI] 1.142–1.258). Each 5-score-year increase in cuCESD was associated with an HR of 1.059 (95% CI 1.045–1.074). Compared with participants with low levels of both exposures, those with high levels of both had the highest hazard (HR 2.550, 95% CI 2.063–3.152). Their standardized 3-year risk was 15.21% (95% CI 13.62%–16.92%), compared with 6.30% (95% CI 5.17%–7.46%) in the low/low group. The continuous multiplicative interaction indicated a weaker relative cuUHR–CMM association at higher cuCESD (interaction HR 0.989, 95% CI 0.981–0.998; likelihood-ratio P = 0.006). Evidence for this interaction weakened after adjustment for baseline cardiometabolic disease burden. Additive-interaction estimates were not statistically significant. Conclusions Higher cuUHR and cuCESD were individually and jointly associated with incident CMM. In the primary model, the relative cuUHR–CMM association was weaker at higher depressive symptom burden, while the high/high group had the highest estimated absolute risk. These findings warrant replication and evaluation of the prognostic value of repeated metabolic and depressive symptom assessments.