Relation of insulin to left ventricular geometry and function in African American and white hypertensive adults: The HyperGEN study

Background: It has been suggested that the trophic effects of insulin may contribute to left ventricular (LV) hypertrophy in hypertension, but few population-based data exist to assess the potential impact of insulin level on LV structure or systolic function. Methods: Fasting plasma insulin levels (log-transformed) in 1542 nondiabetic African American or white hypertensive participants in the Hypertension Genetic Epidemiology Network (HyperGEN) study were compared to indices of LV geometry and function, with adjustment for potential confounders (age, sex, ethnicity, blood pressure (BP), body mass index, height and antihypertensive drugs). Results: In simple correlation analysis, a weak positive relation (r = 0.078, P = .002) was found between LV mass and insulin, principally due to positive correlations (r = 0.22 and 0.50) of both variables to body mass index. Multivariate analysis, adjusting for age, sex, ethnicity, body size, systolic BP, and antihypertensive drugs revealed a modest negative relation between insulin and LV mass (r = -0.08, P = .001) due to a negative relationship between insulin with LV chamber size (r = -0.10, P < .001) and no significant relation to LV wall thicknesses. These results were confirmed in additional analysis controlling for a positive relation (r = .19, P < .0001) of insulin to heart rate, and the relation with insulin became slightly more negative when LV mass was indexed for height^2.7 (r = -0.13, P < .001). After adjustment for covariates, there were no significant relations of insulin to LV ejection fraction or stress-corrected midwall shortening; however, insulin was negatively related to stroke volume (r = -0.12, p < 0.001) and was weakly related positively to relative wall thickness (r = .053) (P = .04). Conclusions: After adjustment for body mass index and other covariates, insulin in nondiabetic hypertensive individuals has weak negative relations to LV chamber size, mass, and stroke volume, and also has weak positive relations to relative wall thickness but not to measures of LV systolic function. Thus, native plasma insulin level may not play a major independent role in the pathogenesis of hypertensive LV hypertrophy or dysfunction.