High insulinlike growth factor binding protein 1 level predicts incident congestive heart failure in the elderly

Abstract

Background: Low levels of insulinlike growth factor 1 (IGF-I) may influence the development of age-related cardiovascular diseases including congestive heart failure (CHF). Insulinlike growth factor binding protein 1 (IGFBP-1), which increases during catabolic states and inhibits anabolic IGF-I effects, is increased in patients with CHF and has been associated prospectively with increased mortality among older adults and survivors of myocardial infarction. We investigated the association between fasting plasma levels of IGF-I, IGFBP-1, IGFBP-3, and insulin and risk of incident CHF in the prospective Cardiovascular Health Study.

Methods: From among 5,888 adults 65 years old and older in the Cardiovascular Health Study, we studied 566 incident CHF cases and 1,072 comparison subjects after exclusion of underweight individuals (body mass index <18.5 kg/m(2)) and insulin users. Hazard ratios (HRs) with 95% CIs for CHF were estimated after adjustment for age, race, sex, hypertension, systolic blood pressure, lipid levels, left ventricular hypertrophy, coronary disease, C-reactive protein, health status, diabetes, and body mass index.

Results: High baseline IGFBP-1 level was a significant predictor of CHF, independent of established CHF risk factors and inflammation markers. The HR per SD of IGFBP-1 was 1.22 (95% CI 1.07-1.39, P < .01). Relative to the lowest IGFBP-1 tertile, the HR was 1.29 (95% CI 0.96-1.74, P = .09) for the second IGFBP-1 tertile and 1.47 (95% CI 1.06-2.04; P = .02) for the highest IGFBP-1 tertile (tertile cut points 19.5 and 35.8 ng/mL). Total IGF-I, IGFBP-3, or insulin levels had no association with CHF after adjustment for CHF risk factors.

Conclusions: High circulating IGFBP-1 level may be a CHF risk factor among older adults.

Figure

Risk of CHF per SD of IGFBP-1 level, by subgroups defined by diabetes/IGF status, gender, and age. Hazard ratio is indicated by boxes (with box size proportional to sample size in subgroups) and 95% confidence intervals are indicated by horizontal lines. *Adjusted for age, race, gender, hypertension, systolic BP, HDL-c, LDL-c, total cholesterol, ECG-LVH, CAD, CRP, general health, diabetes/IFG and BMI.