Modification effect of ideal cardiovascular health metrics on genetic association with incident heart failure in the China Kadoorie Biobank and the UK Biobank

Abstract

Background: Both genetic and cardiovascular factors contribute to the risk of developing heart failure (HF), but whether idea cardiovascular health metrics (ICVHMs) offset the genetic association with incident HF remains unclear.

Objectives: To investigate the genetic association with incident HF as well as the modification effect of ICVHMs on such genetic association in Chinese and British populations.

Methods: An ICVHMs based on smoking, drinking, physical activity, diets, body mass index, waist circumference, blood pressure, blood glucose, and blood lipids, and a polygenic risk score (PRS) for HF were constructed in the China Kadoorie Biobank (CKB) of 96,014 participants and UK Biobank (UKB) of 335,782 participants which were free from HF and severe chronic diseases at baseline.

Results: During the median follow-up of 11.38 and 8.73 years, 1451 and 3169 incident HF events were documented in CKB and UKB, respectively. HF risk increased monotonically with the increase of PRS per standard deviation (CKB: hazard ratio [HR], 1.19; 95% confidence interval [CI], 1.07, 1.32; UKB: 1.07; 1.03, 1.11; P for trend < 0.001). Each point increase in ICVHMs was associated with 15% and 20% lower risk of incident HF in CKB (0.85; 0.81, 0.90) and UKB (0.80; 0.77, 0.82), respectively. Compared with unfavorable ICVHMs, favorable ICVHMs was associated with a lower HF risk, with 0.71 (0.44, 1.15), 0.41 (0.22, 0.77), and 0.48 (0.30, 0.77) in the low, intermediate, and high genetic risk in CKB and 0.34 (0.26, 0.44), 0.32 (0.25, 0.41), and 0.37 (0.28, 0.47) in UKB (P for multiplicative interaction > 0.05). Participants with low genetic risk and favorable ICVHMs, as compared with high genetic risk and unfavorable ICVHMs, had 56~72% lower risk of HF (CKB 0.44; 0.28, 0.70; UKB 0.28; 0.22, 0.37). No additive interaction between PRS and ICVHMs was observed (relative excess risk due to interaction was 0.05 [-0.22, 0.33] in CKB and 0.04 [-0.14, 0.22] in UKB).

Conclusions: In CKB and UKB, genetic risk and ICVHMs were independently associated with the risk of incident HF, which suggested that adherence to favorable cardiovascular health status was associated with a lower HF risk among participants with all gradients of genetic risk.

Keywords: Genetic risk; Heart failure; Ideal cardiovascular health metrics.

Fig. 1

Joint analysis for associations of ideal cardiovascular health metrics categories and genetic risk groups with risk of incident heart failure in CKB and UKB. ICVHMs, ideal cardiovascular health metrics. The categories were defined according to ICVHMs as favorable (6 to 8), intermediate (3 to 5), and unfavorable (0 to 2). PRS, polygenic risk score. The genetic risk categories were defined according to a weighted PRS as low (lowest tertile), intermediate (intermediate tertile), and high (highest tertile). The numbers above the curves represent the hazard ratios and the 95% confidence intervals. Cox proportional hazard regression in CKB adjusted for sex, education, marital status, and family histories of heart attack or stroke at baseline and stratified jointly by study area and age at baseline in the 5-year interval. Cox proportional hazard regression in UKB adjusted for sex, education, marital status, and socioeconomic status and first 20 principal components of ancestry and stratified jointly by age at baseline in the 5-year interval. HR, hazard ratio; CI, confidence interval

Fig. 2

Risk of incident heart failure according to genetic risk categories among different ideal cardiovascular health metrics categories in CKB and UKB. ICVHMs, ideal cardiovascular health metrics. The categories were defined according to ICVHMs as favorable (6 to 8), intermediate (3 to 5), and unfavorable (0 to 2). PRS, polygenic risk score. The genetic risk categories were defined according to a weighted PRS as low (lowest tertile), intermediate (intermediate tertile), and high (highest tertile). The numbers above the curves represent the hazard ratios and the 95% confidence intervals. Cox proportional hazard regression in CKB adjusted for sex, education, marital status, and family histories of heart attack or stroke at baseline and stratified jointly by study area and age at baseline in the 5-year interval. Cox proportional hazard regression in UKB adjusted for sex, education, marital status, and socioeconomic status and first 20 principal components of ancestry and stratified jointly by age at baseline in the 5-year interval. HR, hazard ratio; CI, confidence interval