Association of Bioavailable, Free, and Total Testosterone With Insulin Resistance

Influence of sex hormone-binding globulin and body fat

  1. Elaine C. Tsai, MD, MPH,
  2. Alvin M. Matsumoto, MD,
  3. Wilfred Y. Fujimoto, MD and
  4. Edward J. Boyko, MD, MPH
  1. From the Veterans Affairs Puget Sound Health Care System and the University of Washington, Seattle, Washington
  1. Address correspondence and reprint requests to Elaine Tsai, MD, MPH, VA Puget Sound Health Care System (152E), 1660 S. Columbian Way, Seattle, WA 98108. E-mail: tsaichin{at}


OBJECTIVE—Previous reports of an association between low testosterone levels and diabetes risk were often confounded by covariation of sex hormone-binding globulin (SHBG) and testosterone measurements. Measurements of bioavailable and free testosterone, more reliable indexes of biologically active testosterone, were examined for their associations with markers of insulin resistance and body fat measures in 221 middle-aged nondiabetic men.

RESEARCH DESIGN AND METHODS—Bioavailable and free testosterone were calculated from the concentrations of total testosterone, SHBG, and albumin, and they were not significantly correlated with SHBG (r = 0.07–0.1). In contrast, total testosterone correlated significantly with SHBG (r = 0.63). We evaluated the relationship between these measures of circulating testosterone and markers for insulin resistance (i.e., fasting insulin, C-peptide, and homeostasis model assessment for insulin resistance [HOMA-IR]) as well as total body fat (assessed by dual-energy X-ray absorptiometry [DEXA]) and abdominal fat distribution (assessed by single-slice computed tomography [CT]).

RESULTS—Bioavailable, free, and total testosterone and SHBG all correlated significantly with fasting insulin (age-adjusted r = −0.15 [P = 0.03], −0.14 [P = 0.03], −0.32 [P < 0.0001], and −0.38 [P < 0.0001], respectively), fasting C-peptide (r = −0.18 [P = 0.009] to −0.41 [P < 0.0001]), HOMA-IR (r = −0.15 [P = 0.03] to − 0.39 [P < 0.0001]), and body fat measures (r = −0.17 [P = 0.008] to −0.44 [P < 0.0001]). Only SHBG and total testosterone were significantly associated with fasting glucose (r = −0.20 [P = 0.003] to −0.21 [P = 0.002]). In multivariate analysis, bioavailable or free testosterone was significantly and inversely associated with insulin, C-peptide, and HOMA-IR, but this was not independent of total body or abdominal fat. SHBG was a significant determinant of insulin, C-peptide, and HOMA-IR, independent of body fat. The associations between total testosterone and insulin resistance were confounded by SHBG.

CONCLUSIONS—The inverse association between testosterone and insulin resistance, independent of SHBG, was mediated through body fat.


    • Accepted December 20, 2003.
    • Received August 4, 2003.
| Table of Contents