Association between serum total cholesterol levels and the risk of incident heart failure in patients with atherosclerotic cardiovascular disease

Abstract

To investigate the association between baseline total cholesterol (TC) levels and incident heart failure (HF) risk in patients with atherosclerotic cardiovascular disease (ASCVD). This retrospective cohort analysis was based on the Kailuan Study, a prospective cohort initiated in Tangshan, China. A total of 11,479 patients with first-onset ASCVD between 2006 and 2022 and available post-ASCVD TC measurements were included. Participants were classified into five TC groups: C1 (<3.100 mmol/L), C2 (3.100–<4.100 mmol/L), C3 (4.100–<5.200 mmol/L), C4 (5.200–<6.200 mmol/L), and C5 (≥6.200 mmol/L), with C3 as the reference. Cox proportional hazards models were used to assess HF risk. Restricted cubic spline analyses and Fine–Gray subdistribution hazard models were further performed to evaluate the nonlinear association and account for the competing risk of death. During 71,406.7 person-years of follow-up, corresponding to a mean follow-up duration of 6.22 years, 610 incident HF cases occurred, with an overall incidence proportion of 5.31% and an incidence density of 8.54 per 1,000 person-years. Incidence densities from C1 to C5 were 14.35, 8.46, 7.32, 7.89, and 12.34 per 1,000 person-years, respectively. In the most comprehensively adjusted model (Model 3), which accounted for demographic characteristics, comorbidities, smoking status, body mass index, laboratory indices, medication use, and follow-up TC-related variables, C1 and C5 were associated with higher HF risk than C3, with HRs of 1.578 ( 95% CI 1.160–2.148; p = 0.004) and 1.509 ( 95% CI 1.150–1.979; p = 0.003), respectively. Restricted cubic spline analysis showed a significant nonlinear, approximately U-shaped association, and competing-risk analyses yielded similar findings. Baseline TC levels showed a nonlinear association with incident HF risk in patients with ASCVD. High TC may reflect persistent atherogenic risk, whereas the association between low TC and HF may partly reflect residual confounding or reverse causation related to underlying comorbidities, inflammation, nutritional depletion, frailty, or lipid-lowering treatment history. Therefore, low TC should be interpreted as a potential risk marker rather than as a causal factor for HF.