Diagnostic Accuracy of a Single Serum Testosterone Measurement

Presentation Number: THR-121
Date of Presentation: March 5th, 2015

David Muram*1 and Xiao Ni2
1Eli Lilly and Company, 2Eli Lilly and Company, Indianapolis, IN


There are inherent biological and technical difficulties in measuring serum total testosterone (TT) levels that may reduce the diagnostic accuracy of a single TT determination. Consequently, evidence-based guidelines suggest that although the measurement of morning TT level is the initial diagnostic test for hypogonadism, confirmation of the diagnosis by repeating the TT measurement is recommended. While repeat measurements are recommended for the diagnosis of hypogonadism, a single testosterone determination is required for dose titration in hypogonadal men receiving testosterone replacement therapy. This post-hoc analysis evaluated the diagnostic accuracy of a single serum TT measurement in men receiving testosterone replacement therapy who had a serum average concentration (Cavg) of TT in the normal range. In an open-label, multi-center, titration trial, androgen-deficient men (N=155) were started on a daily morning dose of a 60-mg testosterone 2% solution (Axiron®) applied to axillae (30 mg/axilla). Serum testosterone Cavg was determined on Days 15, 60, and 120. If necessary, the dose was adjusted to maintain Cavg in the normal range (300-1050 ng/dL). This analysis included subjects (n=85) whose Cavg was within the normal range on Days 15, 60, and 120. A generalized linear mixed model was fitted to assess the visit and hour effects with indicator of hourly concentration within the normal range as the outcome variable (assuming a Bernoulli distribution with a logit link function), including visit, hour, and visit-by-hour interaction as the fixed effects and subject as random effects in the model. Main outcome measures were the proportion of men with normal serum testosterone levels at 2, 4, 8, 12, 16, or 20 hours post-dose on Days 15, 60, and 120. Between 79% and 92% of subjects had testosterone serum levels within the normal range at 2, 4, 8, 12, 16, or 20 hours post-dose. No significant differences were found in the proportion of men whose TT levels were in the normal range in the samples obtained at the three different days, whereas there were significant hour effects on the diagnostic accuracy (p=0.01), with the accuracy peaking in the 4- to 8-hour window and tapering at around 16 hours. The diagnostic accuracy of a single TT measurement in men whose Cavg was within the normal range was up to 93%, varying at different time points post-dose, which reflects the pharmacokinetic profile and may be related to the circadian rhythm of endogenous testosterone. Because all of these men had a normal Cavg and required no dose adjustment, relying on a single TT determination would have resulted in an unnecessary dose adjustment in up to 20% of men.


Disclosure: DM: Employee, Eli Lilly & Company. XN: Employee, Eli Lilly & Company.