Testosterone treatment and cognitive function in older men with low testosterone and age-associated memory impairment

Susan M. Resnick, Alvin M. Matsumoto, Alisa J. Stephens-Shields, Susan S. Ellenberg, Thomas M. Gill, Sally A. Shumaker, Debbie D. Pleasants, Elizabeth Barrett-Connor, Shalender Bhasin, Jane A. Cauley, David Cella, Jill P. Crandall, Glenn R. Cunningham, Kristine E. Ensrud, John T. Farrar, Cora E. Lewis, Mark E. Molitch, Marco Pahor, Ronald S. Swerdloff, Denise CifelliStephen Anton, Shehzad Basaria, Susan J. Diem, Christina Wang, Xiaoling Hou, Peter J. Snyder

Research output: Contribution to journalArticlepeer-review

102 Scopus citations


Importance: Most cognitive functions decline with age. Prior studies suggest that testosterone treatment may improve these functions. Objective: To determine if testosterone treatment compared with placebo is associated with improved verbal memory and other cognitive functions in older men with low testosterone and age-associated memory impairment (AAMI). Design, Setting, and Participants: The Testosterone Trials (TTrials)were 7 trials to assess the efficacy of testosterone treatment in older men with low testosterone levels. The Cognitive Function Trial evaluated cognitive function in all TTrials participants. In 12 US academic medical centers, 788 men who were 65 years or older with a serum testosterone level less than 275 ng/mL and impaired sexual function, physical function, or vitality were allocated to testosterone treatment (n = 394) or placebo (n = 394). A subgroup of 493 men met criteria for AAMI based on baseline subjective memory complaints and objective memory performance. Enrollment in the TTrials began June 24, 2010; the final participant completed treatment and assessment in June 2014. Interventions: Testosterone gel (adjusted to maintain the testosterone level within the normal range for young men) or placebo gel for 1 year. Main Outcomes and Measures: The primary outcomewas the mean change from baseline to 6 months and 12 months for delayed paragraph recall (score range, 0 to 50) among men with AAMI. Secondary outcomes were mean changes in visual memory (Benton Visual Retention Test; score range, 0 to -26), executive function (Trail-Making Test B minus A; range, -290 to 290), and spatial ability (Card Rotation Test; score range, -80 to 80) among men with AAMI. Tests were administered at baseline, 6 months, and 12 months. Results: Among the 493 men with AAMI (mean age, 72.3 years [SD, 5.8]; mean baseline testosterone, 234 ng/dL [SD, 65.1]), 247were assigned to receive testosterone and 246 to receive placebo. Of these groups, 247 men in the testosterone group and 245 men in the placebo completed the memory study. Therewas no significant mean change from baseline to 6 and 12 months in delayed paragraph recall score among men with AAMI in the testosterone and placebo groups (adjusted estimated difference, -0.07 [95%CI, -0.92 to 0.79]; P = .88). Mean scores for delayed paragraph recallwere 14.0 at baseline, 16.0 at 6 months, and 16.2 at 12 months in the testosterone group and 14.4 at baseline, 16.0 at 6 months, and 16.5 at 12 months in the placebo group. Testosteronewas also not associated with significant differences in visual memory (-0.28 [95%CI, -0.76 to 0.19]; P = .24), executive function (-5.51 [95%CI, -12.91 to 1.88]; P = .14), or spatial ability (-0.12 [95%CI, -1.89 to 1.65]; P = .89). Conclusions and Relevance: Among older men with low testosterone and age-associated memory impairment, treatment with testosterone for 1 year compared with placebo was not associated with improved memory or other cognitive functions.

Original languageEnglish (US)
Pages (from-to)717-727
Number of pages11
JournalJAMA - Journal of the American Medical Association
Issue number7
StatePublished - Feb 21 2017

Bibliographical note

Publisher Copyright:
© 2017 American Medical Association. All rights reserved.


Dive into the research topics of 'Testosterone treatment and cognitive function in older men with low testosterone and age-associated memory impairment'. Together they form a unique fingerprint.

Cite this