Declining Levels and Bioavailability of IGF-I in Cardiovascular Aging Associate With QT Prolongation-Results From the 1946 British Birth Cohort

Abstract

Background: As people age, circulating levels of insulin-like growth factors (IGFs) and IGF binding protein 3 (IGFBP-3) decline. In rat cardiomyocytes, IGF-I has been shown to regulate sarcolemmal potassium channel activity and late sodium current thus impacting cardiac repolarization and the heart rate-corrected QT (QTc). However, the relationship between IGFs and IGFBP-3 with the QTc interval in humans, is unknown.

Objectives: To examine the association of IGFs and IGFBP-3 with QTc interval in an older age population-based cohort.

Methods: Participants were from the 1946 Medical Research Council (MRC) National Survey of Health and Development (NSHD) British birth cohort. Biomarkers from blood samples at age 53 and 60-64 years (y, exposures) included IGF-I/II, IGFBP-3, IGF-I/IGFBP-3 ratio and the change (Δ) in marker levels between the 60-64 and 53y sampled timepoints. QTc (outcome) was recorded from electrocardiograms at the 60-64y timepoint. Generalized linear multivariable models with adjustments for relevant demographic and clinical factors, were used for complete-cases and repeated after multiple imputation.

Results: One thousand four hundred forty-eight participants were included (48.3% men; QTc mean 414 ms interquartile range 26 ms). Univariate analysis revealed an association between low IGF-I and IGF-I/IGFBP-3 ratio at 60-64y with QTc prolongation [respectively: β -0.30 ms/nmol/L, (95% confidence intervals -0.44, -0.17), p < 0.001; β-28.9 ms/unit (-41.93, -15.50), p < 0.001], but not with IGF-II or IGFBP-3. No association with QTc was found for IGF biomarkers sampled at 53y, however both ΔIGF-I and ΔIGF-I/IGFBP-3 ratio were negatively associated with QTc [β -0.04 ms/nmol/L (-0.08, -0.008), p = 0.019; β -2.44 ms/unit (-4.17, -0.67), p = 0.007] while ΔIGF-II and ΔIGFBP-3 showed no association. In fully adjusted complete case and imputed models (reporting latter) low IGF-I and IGF-I/IGFBP-3 ratio at 60-64y [β -0.21 ms/nmol/L (-0.39, -0.04), p = 0.017; β -20.14 ms/unit (-36.28, -3.99), p = 0.015], steeper decline in ΔIGF-I [β -0.05 ms/nmol/L/10 years (-0.10, -0.002), p = 0.042] and shallower rise in ΔIGF-I/IGFBP-3 ratio over a decade [β -2.16 ms/unit/10 years (-4.23, -0.09), p = 0.041], were all independently associated with QTc prolongation. Independent associations with QTc were also confirmed for other previously known covariates: female sex [β 9.65 ms (6.65, 12.65), p < 0.001], increased left ventricular mass [β 0.04 ms/g (0.02, 0.06), p < 0.001] and blood potassium levels [β -5.70 ms/mmol/L (-10.23, -1.18) p = 0.014].

Conclusion: Over a decade, in an older age population-based cohort, declining levels and bioavailability of IGF-I associate with prolongation of the QTc interval. As QTc prolongation associates with increased risk for sudden death even in apparently healthy people, further research into the antiarrhythmic effects of IGF-I on cardiomyocytes is warranted.

Keywords: IGF-I (insulin-like growth factor-I); IGF-I/IGFBP-3 molar ratio; IGF-II; IGFBP-3; QTc interval prolongation; cardiac repolarization.

Figure 1

Consort diagram of the recruitment process. The Medical Research Council National Survey of Health and Development (NSHD) consists of 5,362 individuals recruited in 1 week in March 1946 in Britain. The exposures of interest here were metabolic markers at 60–64 years while the outcome was QTc interval derived from resting 12-lead electrocardiography (ECG) during the same clinic visit. Both pieces of data were available for 1,513 out of the 5,362 participants. The number of participants involved int the study is presented in the figure below. ECG, electrocardiography; IGF-I, insulin-like growth factor 1; IGF-BP3, insulin-like growth factor binding protein 3.

Figure 2

Boxplots comparing circulating levels of IGF-I, IGF-II, IGFBP-3 and IGF-I/IGFBP-3 molar ratio at 53 and 60–64 years. Whiskers indicate variability outside the third and first quartiles [75th and 25th percentiles] represented as hinges around the median [bold midline]. p-Values derived from Mann-Whitney tests. ECG, electrocardiography; IGF-I, insulin-like growth factor 1; IGF-BP3, insulin-like growth factor binding protein 3.