Relative fat mass and cardiovascular risk in Chinese adults: a nationwide prospective cohort study

Fig. 1. Study flowchart. This flowchart outlines the participant selection process for investigating the association between relative fat mass (RFM) and incident cardiovascular diseases (CVDs), using data from the China Health and Retirement Longitudinal Study (CHARLS). At Wave 1 (2011), 17,708 participants were initially recruited. First, 1,499 participants with CVD at or before baseline and 1,254 participants without CVD information during 2013–2020 were excluded, yielding 14,955 participants without baseline CVD. Next, 556 participants aged < 45 years were excluded, leaving 14,399 adult participants. Further exclusions were applied for unavailable data: height and waist circumference (n = 3,886), blood samples (n = 3,237), demographic information (n = 55), and smoking/drinking status (n = 194). Ultimately, 7,027 eligible participants were enrolled in the final analysis and followed up across 2011, 2013, 2015, 2018, and 2020

Fig. 2. Kaplan-Meier plot of cumulative incidence of new-onset CVDs based on RFM quartiles. This Kaplan-Meier curve illustrates the cumulative hazard of incident cardiovascular diseases across four RFM quartiles (RFM_Q1 to RFM_Q4) from 2011 to 2020. The y-axis represents the cumulative hazard ranging from 0% to 30%, while the x-axis denotes the follow-up time in years. The number at risk for each quartile is indicated below the curve, showing a gradual decrease over time. A highly significant difference (P < 0.0001) exists among the quartiles, indicating that RFM levels are strongly associated with the risk of cardiovascular events, with higher RFM quartiles corresponding to greater cumulative hazards. RFM, relative fat mass

Fig. 3. Dose-response relationship of RFM with the risk of cardiovascular diseases. This figure presents the dose-response relationship between RFM and the adjusted hazard ratio (HR) of incident cardiovascular diseases. The x-axis represents RFM values ranging from 20 to 50, while the y-axis shows the adjusted HR on a logarithmic scale from 0.1 to 5.0. The count of participants at each RFM level is displayed at the top. The P-value for non-linearity is 0.327, suggesting a roughly linear relationship between RFM and the risk of cardiovascular events. As RFM increases, the adjusted HR for incident cardiovascular diseases also rises, indicating a dose-dependent association. The red solid line represents the HR modeled using a restricted cubic spline with 4 knots of RFM. RFM, relative fat mass

Fig. 4. Forest plot of subgroup analyses for the effect of RFM on the risk of cardiovascular diseases. This forest plot presents adjusted hazard ratios (HRs) and 95% confidence intervals (CIs) for the association between relative fat mass (RFM) and incident cardiovascular disease (CVD) across subgroups, with RFM quartiles recalculated within each subgroup. Subgroups include age (< 60 years, ≥ 60 years), hypertension (no/yes), diabetes (no/yes), dyslipidemia (no/yes), and kidney disease (no/yes). For each subgroup, the first RFM quartile (Q1) serves as the reference category; reported data include subgroup sample size, incident CVD rate (%), adjusted HR (95%CI) for each RFM quartile, and P-values for interaction. HRs for continuous RFM are given per 1-unit increase