View Abstract

A2955 - Sleep Apnea and Coronary Artery Calcium Density: Results from the Multi-Ethnic Study of Atherosclerosis (MESA)
Author Block: N. A. Shah1, M. Matsuzaki2, J. R. Kizer3, R. Kaplan3, Z. Fayad1, S. Shea4, M. Allison5, M. Criqui5, P. L. Lutsey6, R. McClelland2, S. S. Redline7; 1Icahn School of Medicine, New York, NY, United States, 2University of Washington, School of Public Health, Seattle, WA, United States, 3Albert Einstein College of Medicine, Bronx, NY, United States, 4Columbia University Medical Center, New York, NY, United States, 5Unversity of California, San Diego, La Jolla, CA, United States, 6University of Minnesota, Minneapolis, MN, United States, 7Brigham and Women's Hospital, Harvard Medical School, Boston, MA, United States.
Rationale: Sleep apnea (SA) has been associated with increased coronary artery calcium (CAC) score. Greater CAC density has been shown to be protective for future risk for coronary artery disease events. However, there is limited evidence on the association between SA and CAC density. We therefore sought to determine the independent association between SA and CAC density. We also explore whether statins modifies this association.
Methods: To assess the association of SA with CAC density, we utilized data from the fifth examination of MESA [April 2010 to February 2012]. CAC density was calculated by dividing the Agatston score by CAC area. Of the 3,200 participants that underwent CT coronary imaging, 2057 participants completed polysomnography of which 1041 have non-zero CAC data. Characteristics of the study sample were compared by SA status (defined as apnea hypopnea index ≥ 15). Mean CAC Agatston, volume, and density volume scores were reported for participants with and without SA (adjusted for age, sex, race/ethnicity via ANCOVA; Tukey’s post-hoc test was used to compare adjusted means). Linear regression models with multivariable adjustments (age, sex, race-ethnicity, HDL-C, total cholesterol, systolic blood pressure, statin-use, smoking and diabetes, sleep duration, slow-wave sleep) and formal test for interaction (SA*statin use) were built to evaluate the independent association between SA and CAC density.
Results: The prevalence of SA in our sample was 36.7% (n=383). The mean age was 70.4±9 and 70.7±9 years in the SA and non-SA groups, respectively. SA was associated with male sex, higher BMI, and diabetes. The age, sex, race/ethnicity adjusted CAC Agatston (457.9 vs. 410.9, p=0.92), CAC volume (357.8 vs. 313.6, p=0.87, mm3) and CAC density (3.1 vs. 3.1, p=0.14) scores were similar in the SA and non-SA groups. Fully adjusted linear regression models demonstrated that SA was weakly and inversely associated with CAC density (β=-0.09, 95% CI -0.18 to 0.00, p=0.04) but not CAC volume (β=0.02, 95% CI -0.09 to 0.13, p=0.70). The statin interaction was significant therefore after stratification by statin use, SA was inversely associated with CAC density in the non-statin group only (β=-0.17, 95% CI -0.31 to -0.04, p=0.01). Among statin users, SA was not associated with CAC density (β=0.00, 95% CI -0.11 to 0.12, p=0.97).
Conclusion: SA is associated with slightly lower CAC density. Statin use modifies this association such that SA only was associated with reduce CAC density in individuals not using statins.