Abstract Body (Do not enter title and authors here): Background: Hypoxia-inducible factor-1 (HIF-1) is a heterodimeric transcription factor composed of a constitutively expressed β subunit (HIF-1β) and an oxygen-dependent α subunit (HIF-1α). HIF-1α has an important protective effect on the pathophysiology underlying coronary artery disease (CAD). However, the association between serum HIF-1α levels and incident heart failure (HF) in patients with stable CAD is unknown.
Methods and Results: Serum levels of HIF-1α were measured in 1,308 patients with stable CAD with no history of HF. The primary outcome was HF hospitalization. The secondary outcomes were cardiovascular (CV) death and a composite of HF hospitalization or CV death. Patients were divided into 5 groups according to HIF-1α levels; below the sensitivity (Q0, <30 pg/mL) and quartiles of measurable HIF-1α levels (Q1, 31-119; Q2, 123-285; Q3, 291-765; Q4, 774-4131 pg/mL). Patients were followed up over a 6-year period. During the follow-up, 121 (9.3%) patients developed HF hospitalization, 89 (6.8%) died of CV diseases, and 181 (13.8%) developed a composite of HF hospitalization or CV death. After adjustment for potential clinical confounders and established CV biomarkers (i.e., N-terminal pro-brain natriuretic peptide, high-sensitivity cardiac troponin I, and high-sensitivity C-reactive protein), the associations between HIF-1α levels and outcomes were not linear: an adjusted hazard ratio (aHR) was consistently low and similar in Q2 and Q3 (v.s. Q0); there were no significant differences in aHRs between Q0 and Q1 for all outcomes; and an aHR (v.s. Q0) was significantly higher in Q4 for HF hospitalization, but not for CV death or a composite of HF hospitalization or CV death (Fig.1). Thus, we combined Q2 and Q3 into the reference group, and combined Q0 and Q1 thereafter. After the reanalysis, serum HIF-1α levels exhibited J-shaped associations with HF hospitalization and a composite of HF hospitalization and CV death, but not with CV death: the high HIF-1α level (Q4) was significantly associated with HF hospitalization and a composite of HF hospitalization and CV death, but not with CV death, while the low HIF-1α level (Q0-1) was significantly associated with a composite of HF hospitalization and CV death, but not with HF hospitalization or CV death (Fig. 2).
Conclusions: We first demonstrated that serum HIF-1α levels exhibited a J-shaped association with the risk of incident HF in stable CAD.