Skip to main content
. Author manuscript; available in PMC: 2009 Apr 8.
Published in final edited form as: Asian J Androl. 2007 Dec 20;10(2):193–200. doi: 10.1111/j.1745-7262.2008.00332.x

Figure 1.

Figure 1

Cross-sectional association (at T1) between hormones and high density lipoprotein cholesterol (HDL-C) by cardiovascular disease (CVD) status (n = 1 661). Lines represent adjusted regression of log HDL-C on hormone. (A): total testosterone (TT); (B): free T (fT); (C): log sex hormone-binding globulin (SHBG); or (D): log dihydrotestosterone (DHT). All models contain the following T1 variables: hormone, CVD, interaction between CVD and hormone, log waist to hip ratio (WHR), smoking, alcohol intake, age and hormone medications. For the purposes of display, the covariates were set to the following values: log waist (log median = log 96.5 cm), smoking (non-smoker), alcohol intake (< 1 drink/day), age (mean = 55 years), and hormone medications (none). P-values for the interaction between CVD and hormone are shown in the upper right corner of each panel. Massachusetts Male Aging Study, 1987–1989.