Is Total Testosterone Measurement an Accurate Method of Predicting Hypogonadism in Men ?

Total testosterone measurement to predict male hypogonadism
Clinical question
How accurate is the measurement of total testosterone for predicting hypogonadism in men?
Bottom line
A total testosterone level of 280 ng/dL or higher is fairly accurate in ruling out hypogonadism in men, with a false negative rate of only 2.1%. A low total testosterone level (< 280 ng/dL) is only moderately useful in predicting hypogonadism (false positive rate = nearly 62%). (LOE = 1b)
Reference
Anawalt BD, Hotaling JM, Walsh TJ, Matsumoto AM. Performance of total testosterone measurement to predict free testosterone for the biochemical evaluation of male hypogonadism. J Urol 2012;187(4):1369-1373. [PubMed® abstract]
Study design: Cohort (retrospective)

Funding source: Foundation

Setting: Outpatient (specialty)

Synopsis
Male hypogonadism affects an estimated 5% to 10% of men older than 30 years. The measurement of free testosterone level, considered the reference standard for the biochemical diagnosis of hypogonadism (< 34 pg/mL), is costly and time consuming and so it's not often used in clinical practice. These investigators reviewed the medical records of all 3672 men, aged 20 years to 90 years, who were evaluated for possible hypogonadism at a Veterans Administration health system from 1997 through 2007. All patients were tested for total testosterone (TT), sex hormone-binding globulin (SHBG), albumin, and calculated free testosterone (cFT). Approximately one third of the men had diabetes mellitus and nearly half were obese (BMI > 30 kg/m2). One individual abstracted and recorded all data. The prevalence of low cFT was 15.2%. The sensitivity of TT (lower limit of normal = 280 ng/dL) to identify a low cFT was 91% and the specificity was 73.7%. The positive predictive value (the likelihood that a positive result, meaning abnormally low TT, is truly positive) was 38.3% (false positive rate = nearly 62%). The negative predictive value (the likelihood that a negative result, meaning a TT > 280 ng/dL, is truly negative) was 97.9% (false negative rate = 2.1%). The positive likelihood ratio for hypogonadism, defined as a low cFT, was greater than 10 (which is considered moderately to strongly clinically useful) only when the TT level was less than 200 ng/dL. The negative likelihood ratio of a TT level 280 ng/dL or greater was 0.12, which is considered moderately useful. A more useful clinical decision rule which will calculate posttest prevalence for individual levels of TT will be available later this summer in Essential Evidence Plus.
David Slawson, MD
Vice Chair, Department of Family Medicine
University of Virginia
Charlottesville, VA