Ranking age-specific modifiable risk factors for cardiovascular disease and mortality: evidence from a population-based longitudinal study

Abstract

Background: Cardiovascular disease (CVD) remains a paramount contemporary health challenge. This study examined age-specific effects of 14 risk factors on CVD and mortality in different age groups.

Methods: We analyzed data from 226,759 CVD-free participants aged 40 years and older in the UK Biobank during the period from baseline time (2006-2010) to September 30, 2021. The primary CVD outcome was a composite of incident coronary artery disease, heart failure, and stroke. We calculated age-specific hazard ratios (HRs) and population-attributable fractions (PAF) for CVD and mortality associated with 14 potentially modifiable risk factors.

Findings: During 12.17-year follow-up, 23,838 incident CVD cases and 11,949 deaths occurred. Age-specific disparities were observed in the risk factors contributing to CVD, and the overall PAF declined with age (PAF of 56.53% in middle-age; 49.78% in quinquagenarian; 42.45% in the elderly). Metabolic factors had the highest PAF in each age group, with hypertension (14.04% of the PAF) and abdominal obesity (9.58% of the PAF) being prominent. Behavioral factors had the highest PAF in the middle-aged group (10.68% of the PAF), and smoking was the leading behavioral factor in all age groups. In socioeconomic and psychosocial risk clusters, low income contributed most among middle-aged (3.74% of the PAF) and elderly groups (3.66% of the PAF), while less education accounted more PAF for quinquagenarian group (4.46% of the PAF). Similar age-specific patterns were observed for cardiovascular subtypes and mortality.

Interpretation: A large fraction of CVD cases and deaths were associated with modifiable risk factors in all age groups. Targeted efforts should focus on the most impactful risk factors, as well as age-specific modifiable risk factors. These findings may inform the development of more precise medical strategies to prevent and manage CVD and related mortality.

Funding: The work was supported by the Bill & Melinda Gates Foundation (grant number: INV-016826 to Hualiang Lin) and the National Natural Science Foundation of China (grant number: 82373534 to Hualiang Lin).

Keywords: Age; Cardiovascular disease; Epidemiology; Modifiable factor; Prospective study.

Fig. 1

The proportion of cardiovascular disease events and mortality stratified by age groups in overall population (A), female (B) and male (C). Abbreviations: CVD, cardiovascular disease; CAD, coronary artery disease; HF, heart failure.

Fig. 2

Associations between 14 modifiable factors and incident cardiovascular disease in overall, middle-aged (38 to <50 years), quinquagenarian (50 to <60 years) and the elderly (≥60 years) groups. Models were adjusted for age, sex, ethnicity, region, family history, and mutually adjusted for individual risk factors. P for interaction was estimated with the use of likelihood ratio test. P values were false discovery rate (FDR) corrected. Abbreviations: HR, hazard ratios; HDL, high-density lipoprotein; SEP, socioeconomic and psychosocial risk factors; PM_2.5, fine particulate matter with diameter <2.5 μm.

Fig. 3

Hazard ratios and 95% CIs of incident cardiovascular disease associated with each one-point increase in behavioral, metabolic and SEP risk scores by age groups. Models were adjusted for age, sex, ethnicity, region, family history, and risk scores of other categories. P for interaction was estimated with the use of likelihood ratio test. P values were false discovery rate (FDR) corrected. Abbreviations: HR, hazard ratios; CIs, confidence intervals; SEP, socioeconomic and psychosocial risk factors.

Fig. 4

Population attributable fractions (PAFs) for incident cardiovascular disease associated with the modifiable risk factors. (A) PAFs of each risk factor on cardiovascular disease in overall population; (B) PAFs of each risk factor on cardiovascular disease among age groups; (C) PAFs of risk factor clusters on cardiovascular disease among age groups. Models were adjusted for age, sex, ethnicity, region, family history, and mutually adjusted for individual risk factors. Estimated PAFs for modifiable risk factors were truncated at a lower limit of 0, as this is the lowest threshold to show a relationship with increased risk. Abbreviations: Abdo obesity, abdominal obesity; non-HDL cholesterol, non-high-density lipoprotein cholesterol; SEP, socioeconomic and psychosocial risk factors.