Association of cardiovascular health at old age with all-cause mortality: a prospective cohort study in China

Figures

• 

  Fig. 1. Mosaic plot of each ideal cardiovascular health metrics by age, gender and residence. Mosaic plot was drawn by R package “vcd” to depict the percentage of each ideal CVH in age and gender group. Female participants age < 75 years had the highest percentage of ideal CVH in smoking metric and male participants living in rural area did well in BMI metric. Compared with averaged expectation of ideal fasting plasma glucose, male participants who aged less than 75 years and lived in rural area had an excess percentage, while female participants who were at the same age and lived in urban area had less percentage

• 

  Fig. 2. Percentage bar plot for the distribution of CVH scores by age, gender and residence. Percentage bar plot was drawn by R package “ggplot2” to present the distribution of CVH scores by the combination of age, gender and residence. It shows that only 8% study participants had 11–12 points (only 1.6% of participants obtained all the 6 metrics of ideal CVH) and the same centralized pattern was observed in subgroup

• 

  Fig. 3. RCS and Survival curve for the association between CVH score category and all-cause mortality with unadjusted cox regression model. Restricted cubic spline was drawn by R package “rms” to test if the relationship between CVH score and all-cause mortality fits linear model. Survival curve was drawn by R package “survival” with cumulative hazard was generated to confirm whether CVH score fulfill the proportion hazard assumption. RCS: restricted cubic spline; ICH: ideal cardiovascular; The CVH score of 0–4 was set as the reference point with 4 knots. The result fits cox regression model

• 

  Fig. 4. Subgroup analysis for the association of CVH with all-cause mortality. Forest plots were draw by R package “forestplot” to conduct subgroup analysis. Cox regression model adjusted for age, gender, residence, ethnic group, marital status, education, alcohol drinking, previous coronary heart disease and previous stroke. P for trend was computed by including CVH score as continuous variable. VH: cardiovascular health. In age-stratified analysis, each one-unit higher CVH score was similarly associated with a smaller likelihood of all-cause mortality among the participants aged under 75 years (P < 0.001). The association was similar across residence-stratified analysis, the increasing trend wasn’t observed (P = 0.146). In both male and female groups, the higher CVH score, the lower risk of all-cause mortality (P male = 0.008, P female = 0.042)