ASSOCIATION OF INSULIN LEVELS WITH LIPIDS AND LIPOPROTEINS IN ELDERLYJAPANESE-AMERICAN MEN

PURPOSE: Elevated insulin levels have been associated with cardiovascu lar disease, but the relationship of insulin with other risk factors a nd its position in the atherosclerotic pathway is uncertain, A cross-s ectional study was conducted to determine whether insulin concentratio ns were associated with lipids and lipoproteins independently of adipo sity and other cardiovascular rick factors. METHODS: Subjects included 3417 Japanese-American men from the Honolulu Heart Program who comple ted a follow-up examination between 1991 and 1993 and were 71-93 years of age. Men were categorized by quintiles of fasting and 2-hour insul in concentration. RESULTS: Age-adjusted mean high density lipoprotein (HDL) cholesterol and triglyceride levels varied significantly across quintiles of fasting and 2-hour insulin (P < 0.001, tests for trend), but insulin was not related to total cholesterol and low-density lipop rotein (LDL) cholesterol. HDL cholesterol decreased from 59.3 to 43.7 mg/dL and triglycerides increased from 95.6 to 175.8 mg/dL comparing l owest to highest quintiles of fasting insulin, respectively. These ass ociations were slightly stronger in lean than obese subjects and in no ndiabetic versus diabetic individuals particularly for 2-hour insulin levels. Multiple linear regression analysis adjusting for several adip osity measures separately (body mass index (BMI), subscapular skinfold thickness, waist circumference, and waist/hip ratio) and other cardio vascular risk factors attenuated associations slightly but they still remained statistically significant. Estimated differences in HDL chole sterol across extreme quintiles of fasting insulin were reduced slight ly from 15.6 mg/dL with adjustment for age to 12.5 mg/dL with adjustme nt for age and BMI, and to 11.3 mg/dL with adjustment for age, BMI, an d cardiovascular risk factors. CONCLUSIONS: Insulin concentration was strongly and independently associated with HDL cholesterol and triglyc erides in this cohort of elderly Japanese-American men. Since this stu dy was cross-sectional, further investigation is required to determine whether elevated insulin levels are causally related to dyslipidemia. Published by Elsevier Science Inc.